Comparative Effectiveness Review

Number 14

Comparative Effectiveness, Safety, and Indications of Insulin Analogues in Premixed Formulations for Adults With Type 2 Diabetes

Appendices

Appendix A: Detailed Electronic Database Search Strategies

MEDLINE Strategy

Terms	Returns
(("diabetes mellitus, type 2"[mh] OR "type 2 diabetes"[tiab] OR ((diabetes[tiab] OR diabetics[tiab] OR diabetic[tiab]) AND ("non-insulin dependent"[tiab] OR "type 2"[tiab] or type-2[tiab] OR "type II"[tiab]))) AND ("insulin/analogs and derivatives"[mh] OR "BIAsp 30"[tiab] OR "BIAsp30"[tiab] OR (Humalog[tiab] AND (Mix[tiab] OR 25[tiab] OR 50[tiab])) OR (NovoLog[tiab] AND (Mix[tiab] OR 70[tiab] OR 30[tiab])) OR (insulin[tiab] AND ((biphasic[tiab] OR premixed[tiab] OR "pre-mixed"[tiab] OR protamin*[tiab] OR Mix[tiab] OR mixture[tiab]) OR (aspart[tiab] OR lispro[tiab] OR analogue[tiab] OR analogues[tiab] OR analog[tiab] OR analogs[tiab] OR Humalog[tiab])))))NOT (animals[mh]NOT humans[mh])	1149

Terms

Returns

(("diabetes mellitus, type 2"[mh] OR "type 2 diabetes"[tiab] OR ((diabetes[tiab] OR diabetics[tiab] OR diabetic[tiab]) AND ("non-insulin dependent"[tiab] OR "type 2"[tiab] or type-2[tiab] OR "type II"[tiab]))) AND ("insulin/analogs and derivatives"[mh] OR "BIAsp 30"[tiab] OR "BIAsp30"[tiab] OR (Humalog[tiab] AND (Mix[tiab] OR 25[tiab] OR 50[tiab])) OR (NovoLog[tiab] AND (Mix[tiab] OR 70[tiab] OR 30[tiab])) OR (insulin[tiab] AND ((biphasic[tiab] OR premixed[tiab] OR "pre-mixed"[tiab] OR protamin*[tiab] OR Mix[tiab] OR mixture[tiab]) OR (aspart[tiab] OR lispro[tiab] OR analogue[tiab] OR analogues[tiab] OR analog[tiab] OR analogs[tiab] OR Humalog[tiab])))))NOT (animals[mh]NOT humans[mh])

1149

EMBASE Strategy

(('non insulin dependent diabetes mellitus'/exp OR 'type 2 diabetes':ti,ab OR ((diabetes:ti,ab OR diabetics:ti,ab OR diabetic:ti,ab) AND ('non-insulin dependent':ti,ab OR 'type 2':ti,ab or type-2:ti,ab OR 'type II':ti,ab))) AND ('biphasic insulin'/exp OR 'BIAsp 30':ti,ab OR 'BIAsp30':ti,ab OR (Humalog:ti,ab AND (Mix:ti,ab OR 25:ti,ab OR 50:ti,ab)) OR (NovoLog:ti,ab AND (Mix:ti,ab OR 70:ti,ab OR 30:ti,ab)) OR ((insulin:ti,ab AND (biphasic:ti,ab OR premixed:ti,ab OR 'pre-mixed':ti,ab OR protamin*:ti,ab OR Mix:ti,ab) OR (aspart:ti,ab OR lispro:ti,ab OR analogue:ti,ab OR analogues:ti,ab OR analog:ti,ab OR analogs:ti,ab OR Humalog:ti,ab))))) NOT ([animals]/lim NOT [humans]/lim)

1344

Cumulative Index to Nursing and Applied Health Literature (CINAHL)

(((MH "Diabetes Mellitus, Non-Insulin-Dependent") OR (TX "type 2 diabetes") OR (((TX "diabetes") OR (TX "diabetics") OR (TX "diabetic")) AND ((TX "non-insulin dependent") OR (TX "type 2") or (TX "type-2") OR ( TX "type II")))) AND ((MH "Insulin/AA") OR (TX "BIAsp 30") OR (TX "BIAsp30") OR ((TX "Humalog") AND ((TX "Mix") OR (TX "25") OR (TX "50"))) OR ((TX "NovoLog") AND ((TX "Mix") OR (TX "70") OR (TX "30"))) OR ((TX "insulin") AND (((TX "biphasic") OR (TX "premixed") OR (TX "pre-mixed") OR (TX "protamin*") OR (TX "Mix") OR (TX "mixture")) OR ((TX "aspart") OR (TX "lispro") OR (TX "analogue") OR (TX "analogues") OR (TX "analog") OR (TX "analogs") OR (TX "Humalog"))))))

299

The Cochrane Central Register of Controlled Trials (CENTRAL)

#1	(type 2 diabetes):ti,ab,kw in Clinical Trials
#2	(diabetes):ti,ab,kw or (diabetics):ti,ab,kw or (diabetic):ti,ab,kw in Clinical Trials
#3	(non-insulin dependent):ti,ab,kw or (type 2):ti,ab,kw or (type-2):ti,ab,kw or (type II):ti,ab,kw in Clinical Trials
#4	(#2 AND #3)
#5	(#1 OR #4)
#6	(BIAsp 30):ti,ab,kw or (BIAsp30):ti,ab,kw in Clinical Trials
#7	(Humalog):ti,ab,kw in Clinical Trials
#8	(Mix):ti,ab,kw or (25):ti,ab,kw or (50):ti,ab,kw in Clinical Trials
#9	(#7 AND #8)
#10	(NovoLog):ti,ab,kw in Clinical Trials
#11	(Mix):ti,ab,kw or (70):ti,ab,kw or (30):ti,ab,kw in Clinical Trials
#12	(#10 AND #11)
#13	(insulin):ti,ab,kw in Clinical Trials
#14	(biphasic):ti,ab,kw or (premixed):ti,ab,kw or (pre-mixed):ti,ab,kw or (protamin*):ti,ab,kw or (mix):ti,ab,kw in Clinical Trials
#15	(mixture):ti,ab,kw in Clinical Trials
#16	(#14 OR #15)
#17	(aspart):ti,ab,kw or (lispro):ti,ab,kw or (analogue):ti,ab,kw or (analogues):ti,ab,kw or (analog):ti,ab,kw in Clinical Trials
#18	(analogs):ti,ab,kw or (Humalog):ti,ab,kw in Clinical Trials
#19	(#17 OR #18)
#20	(#16 OR #19)
#21	(#13 AND #20)
#22	(#6 OR #9 OR #12 OR #21)
#23	(#5 AND #22)

654

Appendix B: Hand Searched Journals

All Journals Hand Searched

June 2007 – September 2007

Acta Diabetologica
Annals of Internal Medicine
Clinical Therapeutics
Diabetes Care
Diabetes, Obesity & Metabolism
Diabetic Medicine
Diabetologia
European Journal of Internal Medicine
Experimental and Clinical Endocrinology and Diabetes
Hormone and Metabolic Research
JAMA
Journal of Diabetes and its Complications
New England Journal of Medicine

Appendix C: List of Excluded Articles

Anonymous. 1-2-3: study results and clinical application: the ''Start & Stay'' approach. Journal of Diabetes Nursing 2006;(1):3p.
No original data
Anonymous. The 1-2-3 study: achieving glycaemic goals in type 2 diabetes. Journal of Diabetes Nursing 2006;(1):1p.
No original data
Anonymous. Key abstract: the EUROMIX study. Journal of Diabetes Nursing 2005;[3].
No original data
Anonymous. Rapid acting insulin analogue effective in a range of body types launched. Pharm. J. 2005;275(7369):401.
No original data
Anonymous. DTB questions first-line use of insulin analogues. Pharm. J. 2004;273(7321):552.
No original data
Anonymous. Lispro, a rapid-onset insulin. Med. Lett. Drugs Ther. 1996;38(986):97-98.
No original data
Anonymous. The why and how of early intervention with insulin analogs. Diabetes Educator 2007;3352S-75s.
No original data
Anonymous. CDC Fact Book 2000/2001. Department of Health and Human Services Centers for Disease Control and Prevention, 2000 Sep (138 p)
No original data
Anonymous. Diabetes Overview. US Department of Health and Human Services National Institute of Diabetes and Digestive and Kidney Diseases, 2000 Sep (8 p)
No original data
Anonymous. Is biphasic, prandial, or basal insulin best for poorly controlled type 2 diabetes? J Fam Pract 2008;57(2):84.
No original data
Abraham M R, Al-Sharafi B A, Saavedra G A et al. Lispro in the treatment of insulin allergy. Diabetes Care 1999;22(11):1916-1917.
No original data
Akram J. Prevention of hypoglycaemia in insulin-treated patients during Ramadan: results from a multicentre study: 2. Practical Diabetes International 1998;15(1):S19.
Did not evaluate a premixed insulin analogue
Aristides M, Weston A R, FitzGerald P et al. Patient preference and willingness-to-pay for Humalog Mix25 relative to Humulin 30/70: a multicountry application of a discrete choice experiment. Value Health 2004;7(4):442-54.
Does not apply to a key question
Bain S C, Kamal A D. Safety and side effects of the insulin analogues. Expert Opin. Drug Saf. 2006;5(3):349-350.
No original data
Bell D, Bode B, Clements R S et al. Premixed vs. self-mixed insulin in the treatment of type II diabetes mellitus: A randomized trial. Today's Ther. Trends 1991;9(1):63-73.
Did not evaluate a premixed insulin analogue
Bolli G B, Di Marchi R D, Park G D et al. Insulin analogues and their potential in the management of diabetes mellitus. Diabetologia 1999;42(10):1151-1167.
No original data
Bullano M F, Fisher M D, Grochulski W D et al. Hypoglycemic events and glycosylated hemoglobin values in patients with type 2 diabetes mellitus newly initiated on insulin glargine or premixed insulin combination products. Am J Health Syst Pharm 2006;63(24):2473-82.
Did not evaluate a premixed insulin analogue
Calle-Pascual A L, Bagazgoitia J, Calle J R et al. Use of insulin lispro in pregnancy. Diabetes Nutr Metab 2000;13(3):173-7.
No original data
Cappelleri JC, Cefalu WT, Rosenstock J et al. Treatment satisfaction in type 2 diabetes: a comparison between an inhaled insulin regimen and a subcutaneous insulin regimen. Clinical therapeutics 2002;24(4):552-64.
Did not evaluate a premixed insulin analogue
Chan W B, Chow C C, Yeung V T F et al. Effect of insulin lispro on glycaemic control in Chinese diabetic patients receiving twice-daily regimens of insulin. Chin. Med. J. 2004;117(9):1404-1407.
Did not evaluate people with type 2 diabetes
Clements M R, Tits J, Kinsley B T et al. Improved glycaemic control of thrice-daily biphasic insulin aspart compared with twice-daily biphasic human insulin; a randomized, open-label trial in patients with type 1 or type 2 diabetes. Diabetes Obes Metab 2008;10(3):229-37.
Did not evaluate people with type 2 diabetes
Cobden D, Lee W C, Balu S et al. Health outcomes and economic impact of therapy conversion to a biphasic insulin analog pen among privately insured patients with type 2 diabetes mellitus. Pharmacotherapy 2007;27(7):948-62.
Did not compare a premixed insulin analogue to another antidiabetic agent
Coscelli C, Calabrese G, Fedele D et al. Use of premixed insulin among the elderly. Reduction of errors in patient preparation of mixtures. Diabetes Care 1992;15(11):1628-30.
Did not evaluate a premixed insulin analogue
Culy C R, Jarvis B. Management of diabetes mellitus: Defining the role of insulin lispro mix75/25 (Humalog(registered trademark) Mix75/25(trademark)). Dis. Manage. Health Outcomes 2001;9(12):711-730.
No original data
Currie C J, McEwan P, Poole C et al. Comments on Long-term clinical and cost outcomes of treatment with biphasic insulin aspart 30/70 versus insulin glargine in insulin-naive type 2 diabetes patients: cost-effectiveness analysis in the UK setting. Curr Med Res Opin 2006;22(5):967-9; author reply 968-9.
No original data
Davidson M B. Twice-Daily NPH or mixture insulins versus triple therapy: apples versus oranges: response to Poulsen et al. Diabetes Care 2004;27(7):1846; author reply 1847-8.
No original data
DeWitt D E. Case study: Treating new-onset catabolic type 2 diabetes with glargine and lispro. Clin. Diabetes 2006;24(4):180-181.
No original data
Dunbar JM, Madden PM, Gleeson DT et al. Premixed insulin preparations in pen syringes maintain glycemic control and are preferred by patients. Diabetes care 1994;17(8):874-8.
Did not evaluate people with type 2 diabetes
Ebeling P, Tuominen J A, Koivisto V A. Insulin analogues and carcinoma of the breast. Diabetologia 1996;39(1):124-125.
No original data
Edelman S. Does a patient-administered titration algorithm of insulin glargine improve glycemic control? Nat Clin Pract Endocrinol Metab 2006;2(2):78-9.
Did not evaluate a premixed insulin analogue
Ejskjaer N, Rasmussen M, Kamp N et al. Comparison of thrice daily 'high' vs. 'medium' premixed insulin aspart with respect to evening and overnight glycaemic control in patients with type 2 diabetes. Diabetes Obes Metab 2003;5(6):438-45.
Did not compare a premixed insulin analogue to another antidiabetic agent
Gale E, Del Prato S. Emerging clinical uses for insulin lispro. Pract. Diabetes Int. 1997;14(4 SUPPL.):S4-S10.
No original data
Garber A J. Assessing the role of biphasic insulin aspart 30 as an effective and tolerable front-line therapy for type 2 diabetes. Clin Ther 2005;27 Suppl 2S39-41.
No original data
Garg S K. New insulin analogues. Diabetes Technol Ther 2005;7(5):813-7.
No original data
Groop L, Harno K, Tolppanen EM. The combination of insulin and sulphonylurea in the treatment of secondary drug failure in patients with type II diabetes. Acta Endocrinol 1984;106(1):97-101.
Did not evaluate a premixed insulin analogue
Hamid Z, Simmons D L. Triple therapy in type 2 diabetes: insulin glargine or rosiglitazone added to combination therapy of sulfonylurea plus metformin in insulin-naive patients: response to Rosenstock et al. Diabetes Care 2006;29(10):2331; author reply 2332.
No original data
Herz M. Clinical update on Humalog Mix25 a novel pre-mixed formulation of insulin lispro and NPL. Int J Clin Pract Suppl 1999;1048-13; discussion 18-20.
No original data
Home PD, Bailey CJ, Donaldson J et al. A double-blind randomized study comparing the effects of continuing or not continuing rosiglitazone + metformin therapy when starting insulin therapy in people with Type 2 diabetes. Diabetic medicine : a journal of the British Diabetic Association 2007;24(6):618-25
Did not evaluate a premixed insulin analogue
Home P D. Comment on: Nauck MA, Duran S, Kim D et al (2007) A comparison of twice-daily exenatide and biphasic insulin aspart in patients with type 2 diabetes who were suboptimally controlled with sulfonylurea and metformin: a non-inferiority study. Diabetologia 50:259-267. Diabetologia 2007;50(7):1561-2.
No original data
Ishii H, Yamamura A, Malone J K. Quality-of-life (QOL) assessment of type 1 and type 2 diabetes mellitus patients in regard to insulin lispro mixture-25 and mixture-50 twice daily therapy. J. Jpn. Diabetes Soc. 2005;48(8):607-616.
Non-English article
Iwamoto Y, Kawamori R, Kadowaki T, et al. Clinical study on insulin lispro Mixture-25 and Mixture-50 administered twice daily in insulin requiring patients with type 1 and 2 diabetes mellitus. Rinsho-Iyaku 2002;18(3):395-409.
Non-English article
Janka H U, Hogy B. Economic evaluation of the treatment of type 2 diabetes with insulin glargine based on the LAPTOP trial. Eur J Health Econ 2007.
Did not evaluate a premixed insulin analogue
JiXiong X, Jianying L, Yulan C et al. The human insulin analog aspart can induce insulin allergy. Diabetes Care 2004;27(8):2084-5.
Does not apply to a key question
Jungmann E, Bolle J, Schmitz C et al. Intensified insulin therapy (IIT) in type II Diabetes mellitus: pre- or postprandial injection of aspart insulin? Medizinische Klinik 2004;99:109.
Non-English article
Katahira M, Hara I, Nishizaki T. Insulin allergy decreased by Humulin S (Humulin R) and not by insulin aspart or Actrapid Penfill (Penfill R). Diabetic Med. 2005;22(10):1455-1457.
No original data
Kazda C M, Forst T, Gierhake C et al. Improving blood glucose and reducing incidence of hypoglycemia in type 2 diabetics using insulin lispro 25%/NPL 75%: VERBESSERUNG DER BLUTGLUKOSEEINSTELLUNG UND SENKUNG DER HYPOGLYKAMIERATE BEI TYP-2-DIABETIKERN UNTER INSULIN LISPRO 25. Diabetes Stoffwechsel 2003;12(5):233-238.
Non-English article
Kitowicz A, Criswell D F. Question: is insulin glargine more effective? J Okla State Med Assoc 2007;100(1):26-7.
No original data
Kitzmiller J L, Buchbinder A, Khoury J et al. Insulin lispro and the development of proliferative diabetic retinopathy during pregnancy (multiple letter) [4]. Am. J. Obstet. Gynecol. 2001;185(3):774-775.
No original data
Kitzmiller J L, Main E, Ward B et al. Insulin lispro and the development of proliferative diabetic retinopathy during pregnancy. Diabetes Care 1999;22(5):874-6.
Did not evaluate a premixed insulin analogue
Koivisto V A. International experience with insulin lispro. Pract. Diabetes Int. 1998;15(1 SUPPL.):S15-S17.
No original data
Koivisto V A, Tuominen J A, Ebeling P. Lispro Mix25 insulin as premeal therapy in type 2 diabetic patients. Diabetes Care 1999;22(3):459-62.
Does not apply to a key question
Kokic S, Bukovic D, Radman M et al. Lispro insulin and metformin versus other combination in the diabetes mellitus type 2 management after secondary oral antidiabetic drug failure. Coll Antropol 2003; 27(1):181-7.
Did not evaluate a premixed insulin analogue
Lee W C, Balu S, Cobden D et al. Medication adherence and the associated health-economic impact among patients with type 2 diabetes mellitus converting to insulin pen therapy: an analysis of third-party managed care claims data. Clin Ther 2006;28(10):1712-25.
Did not compare a premixed insulin analogue to another antidiabetic agent
Levinson P D. Premixed or self-mixed insulin for elderly patients. Ann. Intern. Med. 1993;118(SUPPL. 3):80.
No original data
Lindholm A, Jensen L B, Home P D et al. Immune responses to insulin aspart and biphasic insulin aspart in people with type 1 and type 2 diabetes. Diabetes Care 2002;25(5):876-82.
Does not apply to a key question
Luddeke H J. Improving post-prandial control with Humalog and Humalog mixtures. Int J Clin Pract Suppl 2000;(112):23-8.
No original data
McCormack J, Bassett K. The evidence for insulin lispro. CMAJ 1998;159(11):1353-5.
No original data
Mikhail N. The combined effect of triple therapy with rosiglitazone, metformin, and insulin aspart in type 2 diabetic patients: response to Poulsen et al. Diabetes Care 2004;27(7):1846-7; author reply 1847-8.
No original data
Mohn A, Marcovecchio M, Chiarelli F et al. Insulin analogues [5] (multiple letters). New Engl. J. Med. 2005;352(17):1822-1824.
No original data
Moisey R S, Jenkins R, Nagi D. The use of basal insulin (NPH) compared with pre-mixed biphasic insulin in patients with type 2 diabetes mellitus. A single centre experience. Pract. Diabetes Int. 2007;24(4):212-216.
Exclude other reason
Nathan J P, Rosenberg J M. How are insulin glargine and insulin aspart different from the ''older'' insulins? Drug Topics 2000;144(22):41.
No original data
Nauck M A, Trautman M, Brodows R et al. Response to comment on: Nauck MA, Duran S, Kim D et al (2007) A comparison of twice-daily exenatide and biphasic insulin aspart in patients with type 2 diabetes who were suboptimally controlled with sulfonylurea and metformin: a non-inferiority study. Diabetologia 50:259-267. Diabetologia 2007;50(7):1563-4.
No original data
Ning G, Xiang K, Gao Y, et al. Comparison of post-prandial blood glucose excursions between insulin lispro 75/25 and human insulin 70/30 in Chinese patients with type 1 or type 2 diabetes. International Journal of Medicine 2005;9(2):14-22.
Did not compare a premixed insulin analogue to another antidiabetic agent
Oosthuizen H. Insulin therapy in type 2 diabetes mellitus. J. Endocrinol. Metab. Diabetes S. Afr. 2003;8(3):72-78.
No original data
Panczel P, Hosszufalusi N, Horvath M M et al. Advantage of insulin lispro in suspected insulin allergy. Allergy 2000;55(4):409-10.
Exclude other reason
Peragallo-Dittko V. Insulin therapy for type 2 diabetes. Diabetes Self-Management 2003;20(5):17.
No original data
Plog T. Insulin analogs plus oral antidiabetic drug therapy for older patients with type 2 diabetes. Long-Term Care Interface 2007; 7.
No original data
Renner R, Vocke K, Hepp KD. Blood Glucose Profiles in Type I and Type II Diabetic Patients under Different Insulin Mixtures of BHI-Regular and BHI-NPH. Munchener Medizinische Wochenschrift 1983;125(Suppl 1):57-62.
Non-English article
Reviriego J, Herz M, Roach P, and the Humalog® Mix25 Spanish Study Group. Improved glycaemic control without increased risk of hypoglycaemia with a 25% insulin/lispro 75% NPL mixture twice daily compared with NPH twice daily in patients with type 2 diabetes. J Appl Ther Res 2004:4(4):3-9.
Exclude other reason
Roach P, Strack T, Arora V et al. Improved glycaemic control with the use of self-prepared mixtures of insulin lispro and insulin lispro protamine suspension in patients with types 1 and 2 diabetes. Int J Clin Pract 2001;55(3):177-82.
Did not evaluate a premixed insulin analogue
Robertson D. Achieving fasting and postprandial blood glucose control in type 2 diabetes. Br J Hosp Med (Lond) 2006;67(10):518-22.
No original data
Rubin R R, Peyrot M. Quality of life, treatment satisfaction, and treatment preference associated with use of a pen device delivering a premixed 70/30 insulin aspart suspension (aspart protamine suspension/soluble aspart) versus alternative treatment strategies. Diabetes Care 2004;27(10):2495-7.
Exclude other reason
Schmoelzer I, de Campo A, Pressl H et al. Biphasic insulin aspart compared to biphasic human insulin reduces postprandial hyperlipidemia in patients with Type 2 diabetes. Exp Clin Endocrinol Diabetes 2005;113(3):176-81.
Does not apply to a key question
Schreiber S A, Russmann A. Insulin glargine and educational intervention in patients with type 2 diabetes in clinical practice: long-term improvement in glycaemic control without weight gain. Exp Clin Endocrinol Diabetes 2006;114(1):41-2.
No original data
Shichiri M, Kishikawa H, Ohkubo Y et al. Long-term results of the Kumamoto Study on optimal diabetes control in type 2 diabetic patients. Diabetes Care 2000;23 Suppl 2B21-9.
Did not evaluate a premixed insulin analogue
Sridhar G R. Two regimens of twice-daily premix insulin analogue: an observational study. Diabetes Res Clin Pract 2006;71(1):105-7.
Did not compare a premixed insulin analogue to another antidiabetic agent
Swenson K, Brackenridge B. Lispro for type 2? Diabetes Forecast 2000;53(7):81-83.
No original data
Swenson K, Brackenridge B. Lispro insulin for improved glucose control in obese patient with type 2 diabetes. Diabetes Spectrum 1998;11(1):13-15.
No original data
Thaware P, Howe J, Lawrence I G et al. Use of the rapid acting insulin analogue lispro and its protamine retarded from (Humalog Mix 25) in a clinical setting. Pract. Diabetes Int. 2004;21(9):329-333.
Did not compare a premixed insulin analogue to another antidiabetic agent
Valentine W J, Palmer A J, Lammert M et al. Long-term clinical and cost outcomes of treatment with biphasic insulin aspart 30/70 versus insulin glargine in insulin naive type 2 diabetes patients: cost-effectiveness analysis in the UK setting. Curr Med Res Opin 2005;21(12):2063-71.
No original data
Walczak I M. Lantus reduces blood glucose levels, less hypoglycemia in treatment of type 2 diabetes. Diabetes Technol Ther 2002;4(5):735-6.
No original data
Warren M L, Conway M J, Klaff L J et al. Postprandial versus preprandial dosing of biphasic insulin aspart in elderly type 2 diabetes patients. Diabetes Res Clin Pract 2004;66(1):23-9.
Did not compare a premixed insulin analogue to another antidiabetic agent
White J R. Insulin glargine clinical trials. Clin Ther 2004;26(7):1179-81; discussion 1182-3.
No original data
Yamada S, Watanabe M, Funae O et al. Effect of combination therapy of a rapid-acting insulin secretagogue (glinide) with premixed insulin in type 2 diabetes mellitus. Intern Med 2007;46(23):1893-7.
Did not evaluate a premixed insulin analogue
Yasuda H, Nagata M, Moriyama H et al. Human insulin analog insulin aspart does not cause insulin allergy. Diabetes Care 2001;24(11):2008-9.
Did not evaluate a premixed insulin analogue
Zinman B. The pharmacokinetics of insulin analogues and pumps. Pract. Diabetes Int. 2001;18(5 SUPPL.):S3-S4.
No original data

Appendix D: Data Abstraction Forms

Appendix E. Evidence Tables

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome
	Premixed vs. long- acting insulin analogues	Premixed vs. rapid- acting insulin analogues	Premixed vs. rapid- acting + long-acting insulin analogues	Premixed insulin analogues vs. premixed human insulin	Premixed insulin analogues vs. NPH	Premixed insulin analogues vs. noninsulin antidiabetic agents
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
	Fasting glucose
Quantity of evidence: Number of studies	11	2	2	9	2	10
Range of sample sizes	20-469	107-473	145	25-177	93-403	49-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Moderate	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	-1	0	-1	0	-1
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	-1	0	0	0	0	-1
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	-1	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	0	0
Overall grade of evidence (high, moderate, low)	Moderate	Low	Low	Moderate	Low	Moderate
	Pre-dinner glucose
Quantity of evidence: Number of studies	8	3	1	7	1	8
Range of sample sizes	20-469	106-474	374	25-187	394	49-501
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Moderate	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	0	0	0	0	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	0	0
Overall grade of evidence (high, moderate, low)	Moderate	Moderate	Moderate	Moderate	Moderate	Moderate
	2-hour postprandial glucose after breakfast
Quantity of evidence: Number of studies	9	1	1	11	2	10
Range of sample sizes	20-315	107	374	23-177	140-403	143-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Moderate	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	NA	0	0	NA	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	-1	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	-1	0	0	-1	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	1	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	0	0
Overall grade of evidence (high, moderate, low)	High	Low	Low	Moderate	Low	Moderate
	2-hour postprandial glucose after dinner
Quantity of evidence: Number of studies	10	2	2	8	2	9
Range of sample sizes	20-469	107-473	145-374	25-177	140-143	49-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Moderate	High	High	high
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	0	NA	0	-1	-1
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	-1	-1	0	-1	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	0	0
Overall grade of evidence (high, moderate, low)	High	Moderate	Low	High	Low	Moderate
	HbA1c
Quantity of evidence: Number of studies	9	2	1	7	2	6
Range of sample sizes	20-708	159-708	145	40-177	140-403	129-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Low	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	-1	0	0	0
Did the studies have important inconsistency? (-1)	0	0	0	0	0	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	-1	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	+ 1	0	0	+ 1	0	0
Did the studies have evidence of a dose-response gradient? (+1)	+ 1	0	0	+ 1	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	0	0
Overall grade of evidence (high, moderate, low)	High	Low	Low	High	Low	Moderate
	Hypoglycemia
Quantity of evidence: Number of studies	11	2	2	16	2	10
Range of sample sizes	20-708	159-708	145-374	13-187	140-403	49-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Moderate	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	0	0	0	0	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	+1	0	0	0
Overall grade of evidence (high, moderate, low)	High	Low	Low	High	Low	High
	Weight change
Quantity of evidence: Number of studies	10	2	2	7	2	10
Total number of patients studied	20-469	98-473	145-374	30-151	93-403	49-597
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	Medium	High	High	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	0	0
Did the studies have important inconsistency? (-1)	0	0	0	0	0	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	0	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	0	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	0	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	+1	0	0	0
Overall grade of evidence (high, moderate, low, very low)	High	Low	Low	High	Low	High

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome (continued)
	Premixed vs. long-acting insulin analogues	Premixed vs. exenatide	Premixed vs. another premixed insulin analogue	Premixed vs. premixed human insulin	Other comparisons	Premixed vs. oral antidiabetic agents
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
	All-cause mortality
Quantity of evidence: Number of studies	2	1	1	2	0	2
Total number of patients studied	804	501	133	167	NA	926
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), mostly non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	High	High	NA	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	0	NA	0
Did the studies have important inconsistency? (-1)	0	0	0	-1	NA	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	0	NA	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	-1	-1	-1	-1	NA	-1
Did the studies have high probability of reporting bias? (-1)	-1	-1	-1	-1	NA	-1
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	0	NA	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	0	NA	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	0	NA	0
Overall grade of evidence (high, moderate, or low)	Low	Low	Low	Low	Insufficient	Low

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome (continued)
	Premixed vs. long-acting insulin analogues	Premixed vs. premixed human insulin	Other comparisons	Premixed vs. oral antidiabetic agents
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; insufficient = no data.
	Cardiovascular mortality
Quantity of evidence: Number of studies	2	1	0	1
Total number of patients studied	804	186	NA	329
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	NA	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	NA	0
Did the studies have important inconsistency? (-1)	0	0	NA	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	NA	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	-1	-1	NA	-1
Did the studies have high probability of reporting bias? (-1)	-1	-1	NA	-1
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	NA	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	NA	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	NA	0
Overall grade of evidence (high, moderate, low, very low)	Low	Low	Insufficient	Low

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome (continued)
	Premixed vs. long-acting insulin analogues	Premixed vs. exenatide	Premixed vs. premixed human insulin	Other comparisons	Premixed vs. oral antidiabetic agents
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; insufficient = no data.
	Cardiovascular morbidity
Quantity of evidence: Number of studies	2	1	2	0	3
Total number of patients studied	456	501	368	NA	530
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	High	NA	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	0	0	NA	0
Did the studies have important inconsistency? (-1)	0	0	0	NA	0
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	0	0	NA	0
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	-1	-1	-1	NA	-1
Did the studies have high probability of reporting bias? (-1)	-1	-1	-1	NA	-1
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	0	NA	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	0	NA	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	0	NA	0
Overall grade of evidence (high, moderate, low, very low)	Low	Low	Low	Insufficient	Low

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome (continued)
	Premixed vs. long-acting insulin analogues or oral agents	Other comparisons
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; insufficient = no data.
	Nephropathy
	Premixed vs. long-acting insulin analogues or oral agents	Other comparisons
Quantity of evidence: Number of studies	3	0
Total number of patients studied	1223	NA
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	NA
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	NA
Did the studies have important inconsistency? (-1)	0	NA
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	NA
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	-1	NA
Did the studies have high probability of reporting bias? (-1)	-1	NA
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	NA
Did the studies have evidence of a dose-response gradient? (+1)	0	NA
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	NA
Overall grade of evidence (high, moderate, low, very low)	Low	Insufficient

Evidence Table 1. Grading of the body of evidence of the effects of premixed insulin analogues compared to other diabetes medications for each outcome (continued)
	Premixed vs. long- acting insulin analogues	Premixed vs. rapid- acting insulin analogues	Premixed vs. rapid- acting + long-acting insulin analogues	Premixed insulin analogues vs. premixed human insulin	Premixed insulin analogues vs. NPH	Premixed insulin analogues vs. oral antidiabetic agents
High = further research is very unlikely to change our confidence in the estimates; moderate = further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low = further research is likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; insufficient = no data.
	Quality of life
Quantity of evidence: Number of studies	3	1	0	1	0	2
Total range in number of patients studied	45 to 708	159	NA	160	NA	451
Quality and consistency of evidence: Were study designs mostly RCTs (high quality), non-RCTs (medium quality), observational studies (low quality), or about a 50:50 mix of experimental and observational (medium quality)?	High	High	NA	High	NA	High
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	0	-1	NA	0	NA	-1
Did the studies have important inconsistency? (-1)	0	0	NA	0	NA	-1
Was there some (-1) or major (-2) uncertainty about the directness (i.e. extent to which the people, interventions and outcomes are similar to those of interest)?	0	-1	NA	0	NA	-1
Were data imprecise or sparse? (-1) (i.e. lack of data or very wide confidence intervals that may change conclusions)	-1	-1	NA	-1	NA	-1
Did the studies have high probability of reporting bias? (-1)	0	-1	NA	0	NA	-1
Did the studies show strong evidence of association between intervention and outcome? ("strong" if significant relative risk or odds ratio > 2 based on consistent evidence from 2 or more studies with no plausible confounders (+1); "very strong" if significant relative risk or odds ratio > 5 based on direct evidence with no major threats to validity (+2)) - use your clinical judgment for absolute differences.	0	0	NA	0	NA	0
Did the studies have evidence of a dose-response gradient? (+1)	0	0	NA	0	NA	0
Did the studies have unmeasured plausible confounders that most likely reduced the magnitude of the observed association? (+1)	0	0	NA	0	NA	0
Overall grade of evidence (high, moderate, low, very low)	Low	Low	Insufficient	Low	Insufficient	Low

Evidence Table 2. Characteristics of the studies comparing a premixed insulin analogue to other diabetes treatments
Author, year, Country	Study design	Study duration	Exclusion criteria	Target glucose values
* The study population for Boehm 20029 was patients with either type 1 or type 2 diabetes. The type 2 diabetic population was the same study population used for Boehm 2004.⁴⁵ The study duration was 12 weeks. A1c = Hemoglobin A1c; BG = blood glucose; BHI = biphasic human insulin; BIAsp = biphasic insulin aspart; BID = twice daily; BMI = body mass index; CRP = C-reactive protein; CVD = cardiovascular disease; DBP = diastolic blood pressure; dl = deciliter; FPG = fasting plasma glucose; g/day = gram per day; GAD = glutamic acid decarboxylase; hr = hours; IU = international unit; kg = kilogram; kg/m² = kilogram per square meter; L = liter; lbs = pounds; m = meter; mg = milligram; mmHg = millimeter of mercury; mmol = millimole; ng/mL = nanograms per milliliter; nmol = nanomole; NPH = Neutral Protamine Hagedorn; NR = not reported; OA = oral antidiabetic; PPG = postprandial glucose; SBP = systolic blood pressures; SU = sulfonylurea; T1DM = Type 1 diabetes mellitus; T2DM = Type 2 diabetes mellitus; TZD = thiazolidinedione; U = unit; ULN = upper limit of normal
Abrahamian, 2005⁵³ Austria	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	A1c < 7%, no T2DM, BMI > 40 kg/m², history of insulin use or OA agent plus insulin and under good control	Target glucose was according to local practice
Bebakar, 2007⁴⁶ Western Pacific	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	Age < 18 years, any liver disease, any kidney disease, A1c < 7 and > 12%, no T2DM, BMI < 18 and > 30 kg/m², duration of diabetes < 24 and > 60 months, OA agents for < 4 months (SU, biguanide, glinide, alpha-glucosidase inhibitor, or combination (more than two not permissible), CRP < 0.33 nmol/L, TZD therapy in last 6 months	A1c ≤ 8.5% FPG ≤ 7 mmol/L (126 mg/dL)
Boehm, 2004⁴⁵ Boehm, 2002^9* United Kingdom, Germany, Ireland	Parallel-arms, randomized controlled trial	Intended duration: 104 weeks	Age < 18 years, A1c > 11%, no T2DM, BMI > 35 kg/m², duration of diabetes < 2 years	NR
Christiansen, 2003¹³ 9 countries	Parallel-arms, randomized controlled trial	Intended duration: 16 weeks	Age < 18 years, A1c > 11%, no T2DM, BMI > 35 kg/m², insulin doses ≥ 1.8 IU/kg/day, history of serious late diabetic complications or other serious disease	NR
Coscelli, 2003⁶⁷ Italy	Cross-over, randomized controlled trial, no washout period	Mean: 24 days Intended duration: 12 weeks	Age < 35 and > 70 years, any liver disease, any kidney disease, history of CVD, A1c > 9.5%, no T2DM, BMI < 27 and > 35 kg/m², not already taking twice daily premixed insulin (30/70) or NPH insulin therapy for at least 6 months, cancer, drug or alcohol abuse, insulin allergy, recurrent severe hypoglycemia, anemia, hemoglobinopathy, breastfeeding, pregnant, or intending to become pregnant, any treatment with OA agents, systemic glucocorticoids, or insulin doses > 2.0 IU/kg/day	FPG ≤ 7.8 mmol/L (140 mg/dL) 2-hr PPG < 10 mmol/L (180 mg/dL)
Cox, 2007⁷⁴ United States	Cross-over, randomized controlled trial, no washout period	Intended duration: 24 weeks	A1c < 7 and > 10%, no T2DM, have not used metformin, pregnant, breastfeeding, patients with a previous diagnosis of depression or treated with centrally acting medications (e.g., antidepressants or anxiolytics)	FPG < 6.7 mmol/L (121 mg/dL) 2-hr PPG < 8.0 mmol/L (144 mg/dL)
Hermansen, 2002⁵⁸ Denmark	Cross-over, randomized controlled trial, washout period: at least 5 days	Intended duration: 1 day	Age < 18 years, any liver disease, any kidney disease, history of CVD, neuropathy, retinopathy, A1c ≥ 11%, no T2DM, BMI > 32 kg/m², not insulin treated, insulin doses ≥ 1.4 U/kg/day, recurrent severe hypoglycemia, alcohol or drug abuse	NR
Herz, 2002⁷¹ Croatia	Cross-over, randomized controlled trial, no washout period	Intended duration: 4 weeks	Age < 38 and > 69 years, A1c ≥ 10%, no T2DM, BMI > 35 kg/m², not treated with a fixed mixture of human insulin twice daily for at least 1 month, not capable of exercising for 30 minutes on a cycle ergometer at a heart rate of 120 beats/minute during two exercise sessions separated by 30 minutes on rest, being treated with OA agents, systemic glucocorticoids, or insulin doses > 2.0 U/kg/day	FPG < 7.0 mmol/L (126 mg/dL) 2-hr PPG < 10.0 mmol/L (180 mg/dL)
Herz, 2002⁷² Czech Republic, Hungary, Slovenia, Croatia, Poland, Sweden, Australia and New Zealand	Parallel-arms, randomized controlled trial	Intended duration: 16 weeks	Age < 60 and > 80 years, any liver disease, history of CVD, retinopathy, A1c < 1.2 fold ULN at visit 1, FBG < 7.8 mmol/L (140 mg/dL) on at least 2 of 3 occasions during 4 week lead-in, no T2DM, BMI > 35 kg/m², insulin allergy, treatment with insulin in the last 6 months, taking OA agents other than SU or acarbose, not on maximum dose of SU for at least 1 month, duration of diabetes < 1 year, renal dialysis or renal transplant	FPG < 7 mmol/L (126 mg/dL) (encouraged by the study investigators but targets were at the discretion of the physician) 2 hour PPG < 10 mmol/L (180 mg/dL)
Herz, 2003¹² South Africa	Cross-over, randomized controlled trial, no washout period	Intended duration: 4 weeks	Age < 40 and > 70 years, A1c > 10%, no T2DM, BMI > 35 kg/m², not treated with human insulin 30/70 twice daily, have not practiced self-monitoring of BG for at least 3 months, usually injected human insulin 30-45 minutes before meals, being treated with OA agents, systemic glucocorticoids, or insulin doses > 2.0 U/kg/day	FPG < 7.0 mmol/L (126 mg/dL) 2-hr PPG < 10 mmol/L (180 mg/dL)
Hirao, 2008⁶¹ Japan	Parallel-arms, randomized controlled trial	Intended duration: 6 months	A1c < 8%, soft drink ketoacidosis	A1c < 7%
Holman, 2007³² United Kingdom, Ireland	Parallel-arms, randomized controlled trial	Mean: 52 weeks Median: 156 weeks, results reported at 52 weeks Intended duration: 156 weeks	Age < 18 years, any liver disease, any kidney disease, history of CVD, history of insulin treatment, retinopathy, A1c < 7 and > 10%, no T2DM, BMI > 40 kg/m², retinopathy, on less than maximally tolerated doses of metformin and SU for at least 4 months, unawareness of hypoglycemia, pregnant, duration of diabetes < 12 months, TZD treatment or triple OA agents within the previous 6 months, uncontrolled hypertension (SBP > 180 or DBP > 105 mmHg)	A1c 6.5% FPG 72 - 99 mg/dL PPG 90 - 126 mg/dL
Jacober, 2006⁶⁴ United States	Cross-over, randomized controlled trial, no washout period	Intended duration: 16 weeks	Age < 30 years, any liver disease, any kidney disease, history of CVD, history of insulin treatment, A1c < 1.2 - 2 times the ULN reference range as defined by the local laboratory, no T2DM, BMI > 40 kg/m², had adequate blood glucose control, as determined by the investigator, while receiving at least 2 OA agents of different classes used in combination for at least 2 months, undergoing treatment for a malignancy other than basal cell or squamous cell skin cancer, insulin allergy, pregnant or intending to become pregnant, history of severe hypoglycemia within 6 months, currently taking rosiglitazone, long term insulin therapy, chronic systemic glucocorticoid therapy, fibric acid derivatives, niacin or a bile acid sequestant to treat hypertriglyceridemia, chronic anemia	FPG < 120 mg/dL (6.7 mmol/L) For treatment with insulin lispro mixtures, the target 2-hr PPG was < 180 mg/dL (10.00 mmol/L)
Joshi, 2005⁵² India	Prospective study	Intended duration: 12 weeks	No T2DM	A1c < 7% but was up to the individual clinician to titrate
Kann, 2006⁵⁰ Austria, Czech Republic, Germany, Hungary, Poland, Slovakia, Slovenia	Parallel-arms, randomized controlled trial	Intended duration: 26 weeks	A1c ≤ 7 and > 12%, no T2DM, BMI > 40 kg/m², any kidney disease, history of CVD, duration of diabetes < 6 months, not receiving one of the following: SU (at least half maximum dose) with or without metformin, metformin (< 2 g/day), insulin therapy > 7days in last 6 months, alcohol or drug abuse, pregnant, breastfeeding, intending to become pregnant, taking medication interfering with glucose regulation	FPG 5 - 8mmol/L (90 – 144 mg/dL) for both groups 90-min PPG 5 - 10 mmol/L (90 – 180 mg/dL) for BIAsp group
Kapitza, 2004⁵⁶ Germany	Cross-over, randomized controlled trial, washout period: 3 -21 days	Intended duration: 5 hours	No T2DM, not treated with insulin for the past 6 months	BG < 10 mmol/L (180 mg/dL)
Kazda, 2006⁷⁶ Germany	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	Age < 30 or > 75 years, A1c < 6 or > 10.50%, no T2DM, BMI ≥ 40 kg/m², duration of diabetes < 1 and > 10 years, insulin treatment during last 3 months	FPG < 7 mmol/L (126 mg/dL) for insulin glargine 2-hr PPG < 10 mmol/L (180 mg/dL) for lispro groups
Kilo, 2003¹⁵ United States	Parallel-arms, randomized controlled trial	Intended duration: 12 weeks	Age < 18 years, any liver disease, any kidney disease, history of CVD, history of insulin treatment, A1c < 7.5%, FBG < 126 mg/dL, no T2DM, BMI > 40 kg/m², body weight > 100 kg, if significant cardiovascular, liver or kidney disease, NOT on metformin monotherapy or combination with SU or repaglinide for ≥ 3 months, controlled on metformin after 4 week run-in period	FPG 90 - 26 mg/dL
Kvapil, 2006⁵¹ Croatia, Czech Republic, Denmark, France, Greece, Hungary, Norway, Poland, Portugal, Russia, Spain	Parallel-arms, randomized controlled trial	Intended duration: 16 weeks	Any liver disease, any kidney disease, history of CVD, no T2DM, not on metformin, adequately controlled on metformin monotherapy, significant medical problems (proliferative retinopathy, impaired hepatic or renal function, recurrent severe hypoglycemia, cardiac disease, anemia), change in dose of medications known to interfere with metformin	Breakfast insulin aspart 70/30 dose adjusted to target post-breakfast and pre-dinner glucoses of 5 – 8 mmol/L (90 – 144 mg/dL); evening insulin aspart 70/30 dose adjusted to target post-dinner, nighttime, and pre-breakfast blood glucose of 5 – 8 mmol/L (90 – 144 mg/dL)
Malone, 2000⁴⁴ Malone, 2000¹⁴ Canada	Cross-over, randomized controlled trial, washout period: 3-11 days	Intended duration: 1 day	Age < 38 and > 74 years, A1c > 1.5 times ULN, no T2DM, BMI > 35 kg/m², not using a manufactured or self-prepared human insulin mixture in the morning, a short-acting insulin at dinner, and a second NPH insulin dose either at dinner or separately at bedtime, total daily insulin dose > 2.0 U/kg, using an OA agent or glucocorticoids within 2 weeks, using Ultralente insulin, pregnant, breastfeeding	NR
Malone, 2003⁶⁸ 14 countries	Parallel-arms, randomized controlled trial	Mean: 16 weeks	Age < 35 and > 75 years, A1c < 125% of ULN within 4 weeks, no T2DM, BMI > 40 kg/m², adequately controlled diabetes, not using a single OA agent, specifically metformin or SU, at a maximally clinically effective dose within last 3 months	FPG and pre-meal BG < 7 mmol/L (126 mg/dL) 2-hr PPG < 10 mmol/L (180 mg/dL)
Malone, 2004⁶⁵ United States	Cross-over, randomized controlled trial, no washout period	Intended duration: 16 weeks	Age < 30 and > 80 years, history of insulin treatment, A1c < 1.3 and > 2.0 times ULN while using ≥ 1 OA agents without insulin for 30 days before study start, no T2DM, BMI > 40 kg/m²	FPG 90 - 126 mg/dL 2-hr PPG 144 - 180 mg/dL
Malone, 2005⁶⁶ Spain and France	Cross-over, randomized controlled trial, no washout period	Intended duration: 16 weeks	Age < 30 and > 75 years, A1c < 1.3 and > 2.0 times ULN by a local laboratory within 30 days, no T2DM, used TZDs within 30 days, not using NPH once or twice daily, alone or in combination with an OA agent, or a once-daily human insulin mixture with an OA agent for at least 30 days	FPG 5 - 7 mmol/L (90 - 126 mg/dL); 8 - 10 mmol/L (144 - 180 mg/dL) for Humalog 75/25 only
Mattoo, 2003⁷⁰ India, Pakistan, Malaysia, Singapore, Egypt, Morocco, and South Africa	Cross-over, randomized controlled trial, no washout period	Intended duration: 2 weeks	Any liver disease, any kidney disease, history of CVD, retinopathy, no T2DM, BMI > 35 kg/m², not treated with conventional insulin therapy for at least 2 months, not complying with dietary and insulin treatment, not capable or willing to perform self-BG monitoring and use a patient diary, cancer, insulin allergy, drug or alcohol abuse, more than 1 unexplained episode of severe hypoglycemia within 6 months or a history of clinically significant hypoglycemia unawareness, treated with systemic glucocorticoids or insulin doses > 2.0 U/kg	NR
McNally, 2007⁴⁸ United Kingdom	Cross-over, randomized controlled trial, no washout period	Intended duration: 16 weeks	A1c ≥ 9.5%, no T2DM, BMI ≥ 40 kg/m², not treated with insulin for at least 6 months	FPG 5 - 7 mmol/L (90 – 126 mg/dL) Preprandial glucose 5 - 7 mmol/L (90 – 126 mg/dL)
McSorley, 2002¹¹ NR	Cross-over, randomized controlled trial, no washout period	Intended duration: 2 weeks	Age < 40 and > 75 years, no T2DM, history of type 2 diabetes < 1 year, not using twice-daily BHI 30 for at least 6 months	NR
Nauck, 2007⁴⁹ 13 countries	Parallel-arms, randomized controlled trial	Intended duration: 52 weeks	Age < 30 and > 75 years, A1c < 7 and > 11%, BMI < 25 and > 40 kg/m², not on "optimally effective" metformin and sulfonylurea treatment for at least 3 months, lack of stable body weight (> 10% variation in last 3 months), more than 3 episodes of severe hypoglycemia within 6 months prior to screening, use of a weight loss prescription drug in the last 3 months, treated with insulin, TZDs, alpha-glucosidase inhibitors, or meglitinides for > 2 weeks in last 3 months	FPG < 7 mmol/L (126 mg/dL) 2-hr PPG < 10 mmol/L (180 mg/dL)
Niskanen, 2004⁵⁵ United Kingdom, Finland, Norway, Sweden	Cross-over, randomized controlled trial, no washout period	Intended duration: 12 weeks	Age < 18 years, any liver disease, any kidney disease, history of CVD, A1c > 12%, no T2DM, BMI > 35 kg/m², did not require insulin for the past 6 months, insulin dose ≥ 1.8 IU/kg/day, not eligible for BID mixed insulin treatment, not willing or able to perform self monitoring of BG, previous treatment with insulin analogues or use of OA agents within the last 4 weeks, severe uncontrolled hypertension, known or suspected allergy to trial products, pregnant, alcohol or drug abuse	FPG 5.0 - 8.0 mmol/L (90 – 144 mg/dL) Postprandial BG (1-3 hours after a meal) 5.0 - 10.0 mmol/L (90 – 180 mg/dL)
Raskin, 2005³⁹ Raskin, 2007⁴⁰ Brod, 2007⁴¹ United States	Parallel-arms, randomized controlled trial	Intended duration: 28 weeks	Age < 18 and > 75 years, history of insulin treatment, A1c < 8%, no T2DM, BMI > 40 kg/m², body weight > 275 lbs, not on metformin > 1000 mg/day as a single agent or as part of combination therapy for at least 3 months, pregnant, breastfeeding, or not practicing contraception	FPG 80 - 110 mg/dL
Raskin, 2007⁶⁰ United States	Parallel-arms, randomized controlled trial	Intended duration: 34 weeks	Age < 18 years, history of insulin treatment, A1c < 7.5 and > 12, no T2DM, BMI > 42 kg/m², not treated with 2 OA agents for at least 6 months	FPG 4.4 - 6.1 mmol/L (79.2 – 109.8 mg/dL)
Raz, 2003⁵⁷ Israel	Parallel-arms, randomized controlled trial	Intended duration: 6 weeks	Age < 30 years, any liver disease, history of CVD, history of insulin treatment, A1c ≤ 8 and ≥ 13%, no T2DM, T1DM, BMI > 35 kg/m², alcohol or drug abuse, responding to glibenclamide therapy, not treated with glibenclamide as the only OA agent for at least 4 weeks	FPG 90 - 144 mg/dL PPG (1 – 3 hours after a meal) < 180 mg/dL
Raz, 2005⁵⁴ Canada, Israel, China, Australia, Croatia, Thailand, South Africa, Poland	Parallel-arms, randomized controlled trial	Intended duration: 18 weeks	Age < 18 years, any liver disease, history of CVD, A1c < 7.4 and > 14.7%, no T2DM, BMI > 40kg/m², no treatment with SU within last 3 months, alcohol or drug abuse, any serious disease, pregnant, likely to become pregnant or not using contraception	FPG, preprandial, and nighttime 5-8 mmol/L (90 – 180 mg/dL) for insulin aspart 70/30 PPG 5-8 mmol/L (90 -180 mg/dL) for insulin aspart 70/30
Roach, 1999⁷³ United Kingdom, Spain, South Africa	Cross-over, randomized controlled trial, no washout period	Intended duration: 13 weeks	Age < 18 and > 75 years, any liver disease, any kidney disease, history of CVD, history of OA agents, retinopathy, A1c > 9.2%, no T2DM, BMI > 35 kg/m², had not received insulin therapy using mixtures of short-acting or rapid-acting insulin and intermediate- or long-acting insulin twice daily for at least 30 days, cancer, anemia, hemoglobinopathy, alcohol or drug abuse, insulin allergy, recurrent severe hypoglycemia, breastfeeding, pregnant, or intending to become pregnant, treated with OA agents, systemic glucocorticoids, or insulin doses > 2.0 U/kg	NR
Roach, 1999¹⁰ United Kingdom, Germany, Hungary, the Netherlands, Switzerland	Cross-over, randomized controlled trial, no washout period	Intended duration: 12 weeks	Age < 18 and > 70 years, any liver disease, any kidney disease, history of CVD, A1c > 9.2%, no type 1 or type 2 diabetes, not treated with commercially available insulin for at least 120 days, cancer, drug or alcohol abuse, insulin allergy, recurrent severe hypoglycemia, anemia, or hemoglobinopathy, treated with OA agents, systemic glucocorticoids, or insulin doses > 2.0 U/kg	FPG ≤ 7.8 mmol/L (140 mg/dL) 2-hr PPG ≤ 10 mmol/L (180 mg/dL)
Roach, 2003⁶⁹ India	Cross-over, randomized controlled trial, no washout period	Intended duration: 8 weeks	Age < 25 and > 75 years, A1c > 12%, no T2DM, BMI > 35 kg/m², not taking twice daily insulin therapy with mixtures of short- or rapid-acting and intermediate- or long-acting insulin for at least 6 months, history of recurrent severe hypoglycemia, treated with OA agents, systemic glucocorticoids, or insulin doses > 2.0 U/kg	FPG ≤ 7.8 mmol/L (140 mg/dL) 2-hr PPG ≤ 10 mmol/L (180 mg/dL)
Roach, 2006⁶³ United States	Cross-over, randomized controlled trial, no washout period	Intended duration: 12 weeks	Age < 21 and > 80 years, any liver disease, any kidney disease, A1c < 7 and > 12%, no T2DM, inadequate glycemic control using single or multiple OA agents or once or twice-daily insulin or a combination of OA agents and insulin for at least 3 months, use of a TZD within 3 months, pregnant, evidence of major systemic illness or organ dysfunction	FPG < 6.0 mmol/L (108 mg/dL)
Robbins, 2007⁷⁹ United States, Australia, Greece, the Netherlands, Poland, and Puerto Rico	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	Age < 35 and > 75 years, any liver disease, any kidney disease, A1c < 6.5 and > 11, no T2DM, had a clinically unacceptable level of LDL cholesterol determined by investigator's opinion, not currently using metformin and/or a sulfonylurea with a stable dose of 0 to 2 daily insulin injections for at least 3 months, receiving continuous SC insulin infusions, > 3 daily insulin injections or a total daily insulin dose > 2.0 U/kg, change in the type or dose of lipid-altering medications or TZD use up to 3 months prior to study start, fasting triglyceride levels > 4.5 mmol/L (81 mg/dL), pregnant women or women not using an effective method of contraception	FPG < 120 mg/dL (6.7 mmol/L) 2-hr PPG < 144 mg/dL (8.0 mmol/L) for insulin lispro 50/50 group only
Rosenstock, 2008⁸⁰ United States, Puerto Rico	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	Age < 30 and > 75 years, any liver disease, any kidney disease, history of CVD, A1c < 7.5 and > 12%, no T2DM, BMI > 45 kg/m², not taking insulin glargine for at least 90 days in combination with OA agents as monotherapy, dual therapy, or triple therapy, history of scheduled mealtime insulin use or more than one episode of severe hypoglycemia within the prior 6 months, total daily insulin dose > 2.0 U/kg	FPG < 110 mg/dL (6.1 mmol/L)
Schernthaner, 2004⁷⁷ NR	Cross-over, randomized controlled trial, no washout period	Intended duration: 12 weeks	Diagnosed after 35 years of age, any liver disease, any kidney disease, history of CVD, A1c > 11%, no T2DM, BMI > 40 kg/m², no severe diabetic complications	NR
Schwartz, 2006⁶² United States	Cross-over, randomized controlled trial, washout period: clinic visits were at 3-11 day intervals; last dose of usual insulin taken at least 10 hours before test meal	Intended duration: 1-day 1-dose	Age < 30 years, A1c ≥ 2-fold ULN nondiabetic reference range of the local laboratory (4.3% - 6.1%) at screening, no T2DM, BMI > 40 kg/m², not using insulin, excluding insulin glargine, for at least 30 days prior to screening, known allergy to trial products, insulin doses > 2.0 U/kg, any condition interfering with the accurate assessment of the glucodynamic and pharmacokinetic properties of insulin, any condition that precluded a patient from following protocol, pregnant or not using contraception	NR
Sun, 2007⁷⁵ United States	Retrospective cohort study	Intended duration: 18 months	History of insulin treatment, no T2DM, no initiation or less than 3-month use of insulin therapy with either once-daily insulin glargine, twice-daily premixed insulin analogue, or twice-daily premixed human insulin, without switching to another insulin regimen during the observation period, multiple A1c records before and after insulin initiation, taking twice-daily premixed insulin aspart 70/30	NR
Tamemoto, 2007⁴⁷ Japan	Parallel-arms, randomized controlled trial	Intended duration: 24 weeks	Age < 40 or > 75 years, A1c < 7.5 or > 12%, FBG < 140 mg/dL, T1DM, duration of diabetes < 1 year, lack of treatment with OA agents (in particular, had to be on a SU--glibenclamide > 5 mg/d or glimepiride > 3 mg/d over 12 weeks), prior use of insulin in last 12 weeks, fasting C-peptide < 0.7 ng/mL	A1c < 7% FPG < 120 mg/dL
Tirgoviste, 2003⁴³ Roach, 2001⁴² Romania and Russia	Parallel-arms, randomized controlled trial	Intended duration: 16 weeks	Age < 30 years, any liver disease, any kidney disease, history of CVD, retinopathy, A1c ≤ 1.4 times ULN, no T2DM, BMI > 32 kg/m², not treated with a 15 mg dose of glibenclamide as their only medication for at least 3 months prior, FBG ≤ 7.8 mmol/L (140 mg/dL), PPG ≤ 10 mmol/L (180 mg/dL), adrenal insufficiency, insulin allergy, treated with systemic glucocorticoids, hemoglobinopathy	FPG < 7 mmol/L or 125 mg/dL 2-hr PPG < 10 mmol/L or <180 mg/dL
Ushakova, 2007⁵⁹ Russia	Parallel-arms, randomized controlled trial	Intended duration: 8 weeks titration; 8 weeks maintenance	Age < 40 and > 70 years, any liver disease, any kidney disease, history of CVD, history of insulin treatment, retinopathy, A1c < 8%, no T2DM, BMI > 35.0 kg/m², was not treated with at least 1 OA agents for at least 6 months, recurrent major hypoglycemia, using medication known to interfere with glucose metabolism, pregnant or breastfeeding women	FPG: 79.2 - 126 mg/dL (4.4 - 7.0 mmol/L) PPG: 79.2 - 162 mg/dL (4.4 - 9.0 mmol/L)
Yamada, 2007⁷⁸ Japan	Parallel-arms, randomized controlled trial	Intended duration: 4 months	Any liver disease, any kidney disease, history of CVD, retinopathy, A1c ≤ 6.5%, no T2DM, treatment with a twice-daily injection of 70/30 or 50/50 premixed human insulin for < 3 months, patients who were anti-GAD antibody positive, severe hypertension (SBP/DBP 180/100 mmHg)	Self-monitored FPG < 130 mg/dL Clinic-measured PPG < 180 mg/dL

Evidence Table 3. Population characteristics of the studies comparing a premixed insulin analogue to other diabetes treatments
Author, year	Group, N	Mean age (age range) in years	Male, n (%)	Race, n (%)	Mean BMI in kg/m² Mean weight in kg	Mean A1c in % Mean FBG in mg/dL#	Mean duration of diabetes in years	Previous treatment, n (%)
#All numbers have been converted from mmol/L to mg/dL. To convert from mg/dL to mmol/L, divide by 18. *Number has been imputed. †The study population for Boehm 2002⁹ was patients with either type 1 or type 2 diabetes. The type 2 diabetic population was the same study population used for Boehm 2004.⁴⁵ ‡Raskin 2007⁴⁰ was conducted among a subpopulation of Raskin 200539 who were not using thiazolidinediones. µmol/L = micromole per liter; A1c = hemoglobinA1c; AA = African American; BMI = body mass index; BID = twice daily; C = Caucasian; dL = deciliter; FPG = fasting blood glucose; H = Hispanic; IQR = interquartile range; kg = kilogram; kg/m² = kilogram per square meter; mg/dL = milligram per deciliter; NPH = neutral protamine Hagedorn; NR = not reported; OA = oral antidiabetic medication; TID = thrice daily
Abrahamian, 2005⁵³	Insulin aspart 70/30, 89	62.6	46* (52)	NR	BMI: 28	A1c: 9.8	12.7	Insulin naive: No
Abrahamian, 2005⁵³	NPH/regular 70/30, 88	62.3	31* (35)	NR	BMI: 28.3	A1c: 9.85	9.5	Insulin naive: No
Bebakar, 2007⁴⁶	Insulin aspart 70/30 + OA agents, 128	55	48 (92)	NR	BMI: 26.2	A1c: 8.6	4.4	Insulin naive: Yes OA agents: 128 (100)
Bebakar, 2007⁴⁶	OA agents, 63	52.7	41 (69)	NR	BMI: 25.4	A1c: 8.5	4.3	Insulin naive: Yes OA agents: 63 (100)
Boehm, 2004⁴⁵ Boehm, 200²⁹†	Insulin aspart 70/30, 58	62.8	32 (55*)	NR	BMI: 29.1	A1c: 8.11	15.5	Insulin naive: No
Boehm, 2004⁴⁵ Boehm, 200²⁹†	NPH/regular 70/30, 67	62.6	34 (51*)	NR	BMI: 27.2	A1c: 8.21	12.9	Insulin naive: No
Christiansen, 2003¹³	Insulin aspart 70/30, 201	59.3	94* (47)	NR	BMI: 28	A1c: 8.8	9.2	Insulin: 66 (33) OA agents: 78 (39) Insulin and OA agents: 55 (27)
Christiansen, 2003¹³	NPH insulin, 202	59.6	101* (50)	NR	BMI: 28.4	A1c: 8.8	10.5	Insulin: 66 (33) OA agents: 75 (37) Insulin and OA agents: 59 (29)
Coscelli, 2003⁶⁷	Insulin lispro 75/25, 18	59.1	7 (39)	NR	BMI: 29.5 Weight: 79	FBG: 154.2	14.9	Insulin naive: No Insulin: 18 (100)
	NPH/regular 70/30, 15	59.2	8 (53)	NR	BMI: 30.1 Weight: 80.2	FBG: 150.9	13.8	Insulin naive: No Insulin: 15 (100)
	Total, 33	59.1	15 (45)	C: 33 (100)	BMI: 29.8 Weight: 79.5	FBG: 152.5	14.4	Insulin naive: No Insulin: 33 (100)
Cox, 2007⁷⁴	Total, 45	52.6	NR	NR	BMI: 35.08	NR	11.9	Insulin naive: NR OA agents: 45 (100)
Hermansen, 2002⁵⁸	Total, 61	60.1	40 (66*)	NR	BMI: 27.3	A1c: 8.3	11.6	Insulin naive: No Insulin: 61* (100)
Herz, 2002⁷¹	Insulin lispro 75/25, 19	56.3	12 (63*)	NR	BMI: 27 Weight: 76	NR	8.9	Insulin naive: No Insulin: 19* (100)
Herz, 2002⁷¹	NPH/regular 70/30, 18	55.3	6 (33*)	NR	BMI: 26.3 Weight: 75.8	NR	7.5	Insulin naive: No Insulin: 18* (100)
Herz, 2002⁷²	Insulin lispro 75/25, 71	68.1	37 (52.1)	NR	BMI: 28	A1c: 9.82	11.4	Insulin naive: No
Herz, 2002⁷²	Glyburide, 72	67.7	32 (44.4)	NR	BMI: 27.8	A1c: 9.9	12.4	Insulin naive: No
Herz, 2003¹²	Insulin lispro 75/25, 13	54.8	10 (77*)	NR	BMI: 29.2	A1c: 7.81	NR	Insulin naive: No Insulin: 13* (100)
Herz, 2003¹²	NPH/regular 70/30, 12	53.6	7 (58*)	NR	BMI: 29.3	A1c: 7.6	NR	Insulin naive: No Insulin: 12* (100)
Hirao, 2008⁶¹	Insulin aspart 70/30, 80	58.5	47 (59*)	NR	BMI: 23.7 Weight: 62.5	A1c: 10.5	9.5	Insulin naive: No OA agents: 41 (51*)
Hirao, 2008⁶¹	Insulin aspart + NPH insulin, 80	57.9	49 (61*)	NR	BMI:23.7 Weight: 62.1	A1c: 10.7	12.2	Insulin naive: No OA agents: 39 (49*)
Holman, 2007³²	Insulin aspart 70/30+ usual care, 235	61.7	159 (67.7)	AA: 2 (0.9) C: 221 (94) Asian: 11 (4.7) Mixed: 1 (0.4) Other: 0 (0)	BMI: 30.2 Weight: 86.9	A1c: 8.6 FBG: 175	9 median (IQR: 6 - 12)	Insulin naive: Yes OA agents: 221 (94*)
	Insulin aspart + usual care, 239	61.6	152 (63.6)	AA: 5 (2.1) C: 214 (89.5) Asian: 15 (6.3) Mixed: 4 (1.7) Other: 1 (0.4)	BMI: 29.6 Weight: 84.9	A1c: 8.6 FBG: 173	9 median (IQR: 6 - 14)	Insulin naive: Yes OA agents: 227 (95*)
	Insulin detemir + usual care, 234	61.9	143 (61.9)	AA: 2 (0.9) C: 218 (93.2) Asian: 9 (3.8) Mixed: 2 (0.9) Other: 3 (1.3)	BMI: 29.7 Weight: 85.5	A1c: 8.4 FBG: 171	9 median (IQR: 6 - 12)	Insulin naive: Yes OA agents: 224 (96*)
	Total, 708	61.7	454 (64.1)	AA: 9 (1.3) C: 653 (92.2) Asian: 35 (4.9) Mixed: 7 (1) Other: 4 (0.6)	BMI: 29.8 Weight: 85.8	A1c: 8.5 FBG: 173	9 median (IQR: 6 - 13)	Insulin naive: Yes OA agents: 672 (95*)
Jacober, 2006⁶⁴	Total, 60	54.9	34 (56.7)	AA: 3 (5) C: 45 (75) Asian: 3 (5) H: 9 (15)	BMI: 32.9 Weight: 95.1	A1c: 9.21	8.4	Insulin naive: Yes OA agents: 60 (100)
Joshi, 2005⁵²	Insulin aspart 70/30, 114	52.41	76 (67*)	NR	Weight: 70.4	A1c: 8.79 FBG: 186.59	9.53	Insulin naive: NR Insulin: 62 (54.39) OA agents: 102 (89.47)
Joshi, 2005⁵²	Insulin aspart + insulin glargine, 31	51.1	24 (77*)	NR	Weight: 69.63	A1c: 8.53 FBG: 190.23	11.98	Insulin naive: NR Insulin: 21 (67.74) OA agents: 25 (80.65)
Kann, 2006⁵⁰	Insulin aspart 70/30 + metformin, 128	61.5	69 (54*)	NR	BMI: 29.9 Weight: 84.2	A1c: 9.21	10.3	Insulin naive: NR
Kann, 2006⁵⁰	Insulin glargine + glimepiride, 127	61	62 (49*)	NR	BMI: 30.6 Weight: 86.6	A1c: 8.9	10.2	Insulin naive: NR
Kapitza, 2004⁵⁶	Total, 31	57	21 (68*)	NR	BMI: 29	A1c: 8.7	12	Insulin naive: No Insulin: 31* (100)
Kazda, 2006⁷⁶	Insulin lispro 50/50, 54	58.7	32 (59*)	NR	BMI: 31	A1c: 8.1 FBG: 167.4	5.9	Insulin naive: No Insulin: 0 in last 3 months
	Insulin lispro, 52	60.4	32 (62*)	NR	BMI: 31.7	A1c: 8.2 FBG: 176.4	5.3	Insulin naive: No Insulin: 0 in last 3 months
	Insulin glargine, 53	59.1	23 (43*)	NR	BMI: 30.1	A1c: 8.1 FBG: 172.8	5.5	Insulin naive: No Insulin: 0 in last 3 months
Kilo, 2003¹⁵	Insulin aspart 70/30 + metformin, 46	57.2	25 (54)	AA: 4 (9) C: 33 (72) H: 0 (0) Other: 9 (20)	BMI: 30.4	A1c: 9.5 FBG: 241.8	10.4	Insulin naive: Yes Insulin: 0 (0) OA agents: 46 (100) Insulin and OA agents: 0 (0)
	NPH insulin + metformin, 47	55.1	19 (40)	AA: 9 (19) C: 30 (64) H: 1 (2) Other: 7 (15)	BMI: 30.4	A1c: 9.5 FBG: 242.7	10.7	Insulin naive: Yes Insulin: 0 (0) OA agents: 47 (100) Insulin and OA agents: 0 (0)
	NPH/regular 70/30 + metformin, 47	55.4	29 (52)	AA: 6 (13) C: 35 (74) H: 1 (2) Other: 5 (11)	BMI: 30.6	A1c: 9.3 FBG: 227.2	8.4	Insulin naive: Yes Insulin: 0 (0) OA agents: 47 (100) Insulin and OA agents: 0 (0)
Kvapil, 2006⁵¹	Insulin aspart 70/30, 107	55.2	50 (47*)	NR	BMI: 30.9 Weight: 87.3	A1c: 9.6	8.2	Insulin naive: NR
	Insulin aspart 70/30 + metformin, 108	56.4	53 (49*)	NR	BMI: 30.4 Weight: 85.1	A1c: 9.3	6.7	Insulin naive: NR
	Metformin + glibenclamide, 114	58.1	52 (46*)	NR	BMI: 30.5 Weight: 84	A1c: 9.4	8.1	Insulin naive: NR
Malone, 2000⁴⁴ Malone, 200014	Insulin lispro 75/25, 41	59.2	26 (63*)	NR	BMI: 29.1	NR	14	Insulin naive: No Insulin: 41* (100)
	NPH/regular 70/30, 43	60.5	27 (63*)	NR	BMI: 29.2	NR	16.2	Insulin naive: No Insulin: 43* (100)
	Total, 84	59.9	53 (63*)	NR	BMI: 29.2	NR	15.1	Insulin naive: No Insulin: 84* (100)
Malone, 2003⁶⁸	Insulin lispro 75/25 + metformin, 296	58	169 (57)	C: 263 (88.9) H: 22 (7.4) Other: 9 (3) African: 2 (0.7)	BMI: 29.8 Weight: 83	A1c: 9.17	8	Insulin naive: NR OA agents: 296 (100)
Malone, 2003⁶⁸	Glibenclamide + metformin, 301	59	146 (49)	C: 268 (89) H: 18 (6) Other: 12 (4) African: 3 (1)	BMI: 29.6 Weight: 81.7	A1c: 9.27	7.4	Insulin naive: NR OA agents: 301 (100)
Malone, 2004⁶⁵	Insulin lispro 75/25 + metformin, 52	54.5 (32.3 - 79.1)	33 (63.5)	NR	BMI: 30.1 Weight: 88.5	A1c: 8.7 FBG: 150.2	8.1	Insulin naive: Yes OA agents: 52 (100)
Malone, 2004⁶⁵	Insulin glargine + metformin, 53	55.3 (35.5 - 75.1)	33 (62.3)	NR	BMI: 31.7 Weight: 94.4	A1c: 8.7 FBG: 155.3	9.8	Insulin naive: Yes OA agents: 53 (100)
Malone, 2005⁶⁶	Insulin lispro 75/25 + metformin, 50	59.18	25 (50)	NR	BMI: 29.41 Weight: 77.82	A1c: 8.5 FBG: 155.34	13.52	Insulin naive: No Insulin: 50* (100) OA agents: 26 (52*)
Malone, 2005⁶⁶	Insulin glargine + metformin, 47	59.63	18 (38)	NR	BMI: 29.64 Weight: 77.21	A1c: 8.48 FBG: 147.78	11.9	Insulin naive: No Insulin: 47* (100) OA agents: 28 (60*)
Mattoo, 2003⁷⁰	Insulin lispro 75/25, 72	54 (30-72)	34 (47.2)	NR	BMI: 26.9 (17.8 - 34.6) Weight: 71	NR	13.2	Insulin naive: No Insulin: 72* (100)
	NPH/regular 70/30, 79	52 (32-72)	35 (44.3)	NR	BMI: 26.5 (17.1 - 34.5) Weight: 71	NR	11.8	Insulin naive: No Insulin: 79* (100)
	Total, 151	53 (30-72)	69 (45.7)	NR	BMI: 26.7 (17.1 - 34.6) Weight: 71	NR	12.5	Insulin naive: No Insulin: 151* (100)
McNally, 2007⁴⁸	Insulin aspart 70/30, 80	61.8	49 (61*)	NR	BMI: 29.7 Weight: 83.3	A1c: 7.5	11.5	Insulin naive: No Insulin: 80 (100)
	NPH/regular 70/30, 80	62.7	63 (79*)	NR	BMI: 30.5 Weight: 89.1	A1c: 7.5	12.1	Insulin naive: No Insulin: 80 (100)
	Total, 160	62.3	112 (70*)	NR	BMI: 30.1 Weight: 86.2	A1c: 7.5	11.8	Insulin naive: No Insulin: 160 (100)
McSorley, 2002¹¹	Total, 13	64	8 (62*)	NR	BMI: 28.1	A1c: 7.7	13	Insulin naive: No Insulin: 13* (100)
Nauck, 2007⁴⁹	Insulin aspart 70/30 + metformin + sulfonylurea, 248	58	126.5 (51)	NR	BMI: 30.2 Weight: 83.4	A1c: 8.6 FBG: 203.4	10	Insulin naive: NR OA agents: 248 (100)
Nauck, 2007⁴⁹	Exenatide + metformin + sulfonylurea, 253	59	118.9 (47)	NR	BMI: 30.6 Weight: 85.5	A1c: 8.6 FBG: 198	9.8	Insulin naive: NR OA agents: 253 (100)
Niskanen, 2004⁵⁵	Total, 133	62.3	79 (59*)	NR	BMI: 28.1	A1c: 8.5	12.1	Insulin naive: No Insulin: 133* (100)
Raskin, 2005³⁹ Brod, 2007⁴¹	Insulin aspart 70/30 + metformin, 117	52.6	62 (53)	AA: 18 (15) C: 64 (55) Asian: 2 (2) H: 32 (27) Other: 2 (2)	BMI: 31.5 Weight: 90.6	A1c: 9.7 FBG: 252 A1c > 8.5% at baseline, n (%): 10.2 (89)	9.5	Insulin naive: Yes OA agents:117 (100)
Raskin, 2005³⁹ Brod, 2007⁴¹	Insulin glargine + metformin, 116	52.3	65 (56)	AA: 20 (17) C: 60 (52) Asian: 5 (4) H: 30 (26) Other: 1 (1)	BMI: 31.4 Weight: 89.9	A1c: 9.8 FBG: 243 A1c > 8.5% at baseline, n (%): 10.1 (99)	8.9	Insulin naive: Yes OA agents: 116 (100)
Raskin, 2007⁴⁰‡	Insulin aspart 70/30 + metformin, 79	52	41 (51.9)	AA: 10.3 (13) C: 41.1 (52) Asian: 2.4 (3) H: 25.3 (32) Other: 0.78 (1)	BMI: 31.2 Weight: 88.7	A1c: 9.9 FBG: 255.6	NR	Insulin naive: Yes Insulin: 0 (0) OA agents: 79 (100)
Raskin, 2007⁴⁰‡	Insulin glargine + metformin, 78	51.7	42 (53.8)	AA: 11.7 (15) C: 36.7 (47) Asian: 3.1 (4) H: 25 (32)	BMI: 30.8 Weight: 86.2	A1c: 9.9 FBG: 239.4	NR	Insulin naive: Yes Insulin: 0 (0) OA agents: 78 (100)
Raskin, 2007⁶⁰	Insulin aspart 70/30, 102	53.4	46 (45*)	AA: 12* (12) C: 53* (52) Asian: 3* (3) H: 34* (33) Other: 0*(0)	BMI: 32.4	A1c: 8.1	9.2	Insulin naive: Yes Insulin: 0 (0) OA agents: 102 (100)
Raskin, 2007⁶⁰	Metformin and pioglitazone, 98	54.2	38 (39*)	AA: 10* (10) C: 43* (44) Asian: 4* (4) H: 36* (37) Other: 5*(5)	BMI: 33.4	A1c: 8.1	8.3	Insulin naive: Yes Insulin: 0 (0) OA agents: 98 (100)
Raz, 2003⁵⁷	Insulin aspart 70/30 + rosiglitazone, 26	60.3 (43–77)	19 (73.1)	C: 22 (84.6) Asian: 1 (3.8) Other: 3 (11.5)	BMI: 27.7	A1c: 9.9 FBG: 259.8 Serum fructosamine: 398 µmol/L	10.9	Insulin naive: NR OA agents: 26 (100%)
Raz, 2003⁵⁷	Glibenclamide + rosiglitazone, 23	57.8 (43–71)	13 (56.5)	C: 19 (82.6) Asian: 2 (8.7) Other: 2 (8.7)	BMI: 27.6	A1c: 10.3 FBG: 265.2 Serum fructosamine: 409.2 µmol/L	10.3	Insulin naive: NR OA agents: 23 (100%)
Raz, 2005⁵⁴	Insulin aspart 70/30, 97	55.2	63 (65)	NR	BMI: 29.5	A1c: 9.5	10	Insulin naive: NR OA agents: 97* (100)
	Insulin aspart 70/30 + pioglitazone, 93	56.7	49 (53)	NR	BMI: 29.4	A1c: 9.6	9.2	Insulin naive: NR OA agents: 93* (100)
	Glibenclamide + pioglitazone, 91	55.8	56 (62)	NR	BMI: 29.5	A1c: 9.4	9.9	Insulin naive: NR OA agents: 91* (100)
Roach, 1999⁷³	Insulin lispro 75/25, 44	56.5	23 (52*)	NR	BMI: 28.3	NR	12.8	Insulin naive: No Insulin: 44* (100)
Roach, 1999⁷³	NPH/regular 70/30, 45	57.4	19 (42*)	NR	BMI: 29.4	NR	11.5	Insulin naive: No Insulin: 45* (100)
Roach, 1999¹⁰	Insulin lispro 50/50 + insulin lispro 75/25, 34	58	18 (53*)	NR	BMI: 28.4	NR	12.2	Insulin naive: No Insulin: 34* (100)
Roach, 1999¹⁰	NPH/regular 50/50 + NPH/ regular 70/30, 29	60.2	12 (41*)	NR	BMI: 28.4	NR	13.1	Insulin naive: No Insulin: 29* (100)
Roach, 2003⁶⁹	Insulin lispro 75/25, 57	53.9	21 (40)	Asian: 52 (100)	Weight: 62.8	NR	12.4	Insulin naive: No Insulin: 57* (100)
Roach, 2003⁶⁹	Insulin lispro 50/50 + insulin lispro 75/25, 58	54.2	22 (40)	Asian: 55 (100)	Weight: 65.1	NR	13.1	Insulin naive: No Insulin: 58* (100)
Roach, 2006⁶³	Total, 20	53.5	10 (50)	AA: 4* (20) C: 16* (80)	BMI: 36.7 Weight: 108	A1c: 8.4	NR	Insulin naive: No
Robbins, 2007⁷⁹	Insulin lispro 50/50 + metformin, 157	57.4	79 (50.3)	AA: 9 (5.7) C: 102 (65) Asian: 22 (14) H: 24 (15.3)	BMI:32.1 Weight: 89.1	A1c: 7.8	11.3	Insulin naive: No Insulin: 125 (79.6)
Robbins, 2007⁷⁹	Insulin glargine + metformin, 158	58.1	78 (49.4)	AA: 9 (5.7) C: 100 (63.3) Asian: 23 (14.6) H: 26 (16.4)	BMI: 32 Weight: 88.1	A1c: 7.8	12.5	Insulin naive: No Insulin: 123 (77.8)
Rosenstock, 2008⁸⁰	Insulin lispro 50/50, 187	55.4	99 (53)	AA: 25 (13.4) C: 103 (55.1) H: 49 (26.2) Other: 10 (5.3)	BMI:34.1 Weight: 99.1	A1c: 8.83 FBG: 171.81	10.9	Insulin naive: No Insulin: 187 (100) Insulin and OA agents: 185 (98.9)
Rosenstock, 2008⁸⁰	Insulin glargine + insulin lispro, 187	54	98 (52)	AA: 18 (9.6) C: 102 (54.6) H: 53 (28.3) Other: 14 (7.5)	BMI: 34.8 Weight: 99.8	A1c: 8.89 FBG: 181.48	11.2	Insulin naive: No Insulin: 187 (100) Insulin and OA agents: 184 (98.5)
Schernthaner, 2004⁷⁷	Insulin lispro 50/50, 18	66.1	3 (17*)	NR	BMI: 29.5	A1c: 8.3	16.2	Insulin naive: No Insulin: 18* (100)
	NPH/regular 70/30, 17	67.8	5 (29*)	NR	BMI: 28.8	A1c: 8.5	14.2	Insulin naive: No Insulin: 17* (100)
	Total, 35	67	8 (23*)	NR	BMI: 29.2	NR	15.3	Insulin naive: No Insulin: 35* (100)
Schwartz, 2006⁶²	Insulin lispro 75/25, 8	NR	NR	NR	NR	NR	NR	Insulin naive: No
	Insulin lispro 50/50, 7	NR	NR	NR	NR	NR	NR	Insulin naive: No
	NPH/regular 70/30, 8	NR	NR	NR	NR	NR	NR	Insulin naive: No
	Total, 23	61.3	17 (73.9)	AA: 2 (8.7) C: 13 (56.5) H: 8 (34.8)	BMI: 33 Weight: 98.5	A1c: 8.1 FBG: 158.7	NR	Insulin naive: No Insulin: 23 (100)
Sun, 2007⁷⁵	Insulin lispro 75/25, 895	62.8	439* (49.1)	AA: 161* (17.5) C: 268* (30.1) H: 63* (6.8)	Weight: 93.9	A1c: 8.6	20.5	Insulin naive: Yes
	Insulin glargine, 3624	58.4	1740* (48.5)	AA: 362* (10.3) C: 906* (25.2) H: 72* (2.4)	Weight: 93.3	A1c: 8.6	24.7	Insulin naive: Yes
	NPH/regular 70/30, 3647	65.7	1641* (44.7)	AA: 584* (16.5) C: 1204* (32.6) H: 73* (1.6)	Weight: 92.3	A1c: 8.4	18.3	Insulin naive: Yes
Tamemoto, 2007⁴⁷	Insulin aspart 70/30, 14	55.9	6 (54)	NR	BMI: 23.9	A1c: 9.13 FBG: 183.3	9.8	Insulin naive: NR OA agents: 14 (100)
Tamemoto, 2007⁴⁷	Insulin glargine, 20	61.7	13 (68)	NR	BMI: 25.5	A1c: 8.45 FBG: 184.1	10.4	Insulin naive: NR OA agents: 19 (100)
Tirgoviste, 2003⁴³ Roach, 2001⁴²	Insulin lispro 75/25, 85	58.7	30 (35*)	NR	BMI: 26.8 Weight: 74.1	A1c: 9.85 FBG: 208.8	10.3	Insulin naive: Yes OA agents: 85 (100)
	Glibenclamide, 87	60.3	31 (36*)	NR	BMI: 27.6 Weight: 75.8	A1c: 10.07 FBG: 219.6	10.2	Insulin naive: Yes OA agents: 87 (100)
	Total, 172	59.5	61 (35*)	NR	Weight: 75	NR	10.2	Insulin naive: Yes OA agents: 172 (100)
Ushakova, 2007⁵⁹	Insulin aspart 70/30 TID, 104	58	17 (16.3)	NR	BMI: 29.8 Weight: 79.3	A1c: 10.4	9.9	Insulin naive: Yes OA agents: 104 (100)
	Insulin aspart 70/30 BID + metformin, 100	58.4	27 (27)	NR	BMI: 29.2 Weight: 78.4	A1c: 10.4	8.4	Insulin naive: Yes OA agents: 100 (100)
	OA agents, 104	58.4	21 (20.2)	NR	BMI: 29.3 Weight: 78	A1c: 10.1	8.3	Insulin naive: Yes OA agents: 104 (100)
Yamada, 2007⁷⁸	Insulin lispro 50/50, 15	66	12 (80*)	NR	BMI: 27	A1c: 7.59 FBG: 130.3	13.7	Insulin naive: No Insulin: 15 (100)
Yamada, 2007⁷⁸	NPH/regular 70/30 + NPH/regular 50/50, 15	66.3	11 (73*)	NR	BMI: 23.8	A1c: 7.33 FBG: 141.8	15.9	Insulin naive: No Insulin: 15 (100)

Evidence Table 4. Comparative effectiveness of premixed insulin analogues and other diabetes treatments on intermediate outcomes
Author, year	Intervention	Fasting blood glucose in mg/dL#	2-hr PPG in mg/dL after breakfast#	Pre-dinner glucose in mg/dL#	2-hr PPG in mg/dL after dinner#	A1c in %#	Quality of life#
# Numbers are mean (SD) unless otherwise specified. * Number has been imputed. † Number has been estimated from a figure. ◊ Among those who were not using thiazolidinediones. ‡ One-hundred and four (36%) of the 291 participants of this trial are patients with type 1 diabetes. The remaining population has type 2 diabetes and is the same study population as Boehm 2004.⁴⁵ Only data for the Boehm 2004 study is presented because it has the longest followup. ¶ Dosing during the outpatient phase. § Dosing during the inpatient phase. µg = microgram; A1c = Hemoglobin A1c; B = baseline; BDI-II = Beck Depression Inventory – Revised; B-F = mean difference from baseline; bid = twice daily; CI = confidence interval; D = duration; dl = deciliter; DM = diabetes mellitus; EQ-5D = EuroQol-5D; F = final; FBG = fasting blood glucose; fix = fixed dosing; GP = group; GP1-GP2 = mean difference between the difference from baseline; hr = hour; IU = international unit; kg = kilogram; l = liter; mg = milligrams; min = minutes; ml = milliliter; mmol = millimole; NA = not applicable; NPH = neutral protamine Hagedorn; NR = not reported; NS = not significant; OA = oral antidiabetic; p = p-value; per = period; PPG = postprandial glucose; qd = once daily; ref = reference group; SE = standard error; SEM = standard error of the mean; T = time of day when insulin taken; tid = thrice daily; U = units; v = dose varied; WHO-DTSQ = World Health Organization-Diabetes Treatment Satisfaction Questionnaire
Insulin aspart 70/30 vs. long-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin detemir (v) Start: 16 median IU/day Range: 10 – 24 IU/day T: Bedtime, twice if required D: 1 year Usual care D: 1 year	GP1 F-B: -45 (56) p: <0.001 GP2 F-B: -59 (52) GP1-GP2: 14*	PPG (time not specified) (mg/dL) GP1 F-B: -68 (63) p: <0.001 GP2 F-B: 47 (54) GP1-GP2: -115*	GP1 F: 113.04 Median GP2 F: 115.56 Median		Total glycated hemoglobin GP1 B: 8.6 (0.8) F: 7.3 (0.9) p: <0.001 vs. GP2 F-B: -1.3 (1.1) GP2 B: 8.4 (0.8) F: 7.6 (1) F-B: -0.8 (1) GP1-GP2: 0* Glycated hemoglobin ≤ 7.0%, n (%) GP1 98 (41.7) p: <0.001 vs. GP2 GP2 65 (27.8)	EQ-5D GP1 F: 0.76 (95% CI: 0.73 – 0.8) p: overall 0.48 GP2 F: 0.78 (95% CI: 0.75 – 0.81)
Kann, 2006⁵⁰	GP1: Insulin aspart 70/30 (v) Start: 0.1 U/kg bid Mean: 0.4 U/kg T: Breakfast, dinner D: 26 weeks Metformin (v) Start: 500 mg bid or current dose T: Breakfast, dinner D: 26 weeks GP2: Insulin glargine (v) Start: 0.2 U/kg qd Mean: 0.39 U/kg T: Preferred time (constant through study) D: 26 weeks Glimepiride (v) Start: 1 mg daily or current dose T: Breakfast D: 26 weeks	GP1 B: 187.2† F: 136.8 (95% CI: 131.58 – 143.46) F-B: -50.4* GP2 B: 190.8† F: 136.8 (95% CI: 129.6 – 145.26) F-B: -54* GP1-GP2: 3.6* Fasting plasma glucose (time not specified) (mg/dL) GP1 F-B: -46.8 (4.32) p: 0.23 vs. GP2 GP2 F-B: -39.6 (4.5) GP1-GP2: -7*	90-min PPG - after breakfast (mg/dL) GP1 B: 248.4† F: 158.4† F-B: -90* GP2 B: 241.2† F: 187.2† F-B: -54* GP1-GP2: -36*	GP1 B: 187.2† F: 172.8† p: NS F-B: -14.4* GP2 B: 190.8† F: 156.6† F-B: -34.2* GP1-GP2: 19.8	90-min PPG - after dinner (mg/dL) GP1 B: 221.4† F: 156.6† F-B: -66.6* GP2 B: 223.2† F: 183.6† F-B: -39.6* GP1-GP2: -27*	GP1 F: 7.5 (1.1) p: 0.01 GP2 F: 7.9 (1.3) GP1-GP2: -0.5 (95% CI: -0.8 – -0.2) p: 0.0002 A1c < 7%, n (%) GP1 42* (33.1) p: 0.2711 GP2 33* (26.2)
Raskin, 2005^{39 40}	GP1: Insulin aspart 70/30 (v) Start: 10 or 12 U T: Breakfast, dinner D: Unclear Metformin (v) Range: 1500 – 2550 mg/day T: NR D: Unclear GP2: Insulin glargine (v) Start: 10-12 U T: Bedtime D: Unclear Metformin (v) Range: 1500 - 2550 mg/day T: NR D: Unclear	GP1 F: 118.75† p: <0.05 GP2 F: 112.5† GP1◊ F: 122.4† p: NS GP2◊ F: 117† Fasting plasma glucose (time not specified) (mg/dL) GP1 B: 252 (67.4) F: 127 (40.6) p: NS F-B: 125 (72.9) GP2 B: 243 (68.8) F: 117 (44.3) F-B: 125 (74.4) GP1-GP2: 0* GP1◊ F-B: -128.88 (75.06) p: 0.11 GP2◊ F-B: -126 (79.02) GP1-GP2: -3*	90-min PPG - after breakfast (mg/dL) GP1 F: 153.125† p: NS GP2 F: 168.75† GP1◊ F: 154.8† p: NS GP2◊ F: 172.8†	GP1 F: 120.31† p: <0.05 GP2 F: 134.38† GP1◊ F: 129.6† p: NS GP2◊ F: 131.4†	90-min PPG - after dinner (mg/dL) GP1 F: 135.5† p: <0.05 GP2 F: 171.88† GP1◊ F: 127.62 (40.68) p: 0.0008 GP2◊ F: 176.22 (59.04) Dinner postprandial glucose increment (mg/dL) GP1-GP2◊: 19.386† p: 0.003	GP1 B: 10* F: 6.91 (1.17) p: <0.01 F-B: -2.79 (0.11) p: <0.01 GP2 B: 9* F: 7.41 (1.24) F-B: -2.36 (0.11) GP1-GP2: -1* GP1◊ F: 7 (1.3) p: 0.035 F-B: -2.89 (1.6) p: 0.035 GP2◊ F: 7.4 (1.3) F-B: -2.46 (1.6) A1c < 7.0%, n (%) GP1 71.3* (66†) p: <0.001 GP2 45.6* (40†) GP1◊ (65) p: 0.003 GP2◊ (41) A1c ≤ 6.5%, n (%) GP1 43.2* (40†) p: 0.036 GP2 31.9* (28†) GP1◊ (40) p: 0.17 GP2◊ (29)
Tamemoto, 2007⁴⁷	GP1: Insulin aspart 70/30 (v) Start: 10-16 U/day Mean: 26.7 U T: Breakfast, dinner D: 6 months Continued OA agents (NR) T: NR D: 6 months GP2: Insulin glargine (v) Start: 6-8 U/day T: NR D: 6 months Continued OA agents (NR) T: NR D: 6 months	GP1 B: 183.3 (54.6) p: 0.90 F: 141.4 (59.8) p: 0.79 vs. GP2; <0.01 vs. baseline F-B: -41.9* GP2 B: 184.1 (42.1) F: 136.0 (40.3) F-B: -48.1* GP1-GP2: 6.2				GP1 F-B: -1.2 (1.06) p: 0.49 GP2 F-B: -0.95 (0.84) GP1-GP2: 0* A1c < 7%, n (%) GP1 1 (9.1) p: NS GP2 6 (31.6)
Insulin aspart 70/30 vs. rapid-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin aspart (v) Start: 18 median IU/day Range: 9 – 24 IU/day T: Breakfast, lunch, dinner D: 1 year Usual care D: 1 year	GP1 F-B: -45 (56) p: <0.001 vs. GP2 GP2 F-B: -23 (49) GP1-GP2: -22*	PPG (time not specified) (mg/dL) GP1 F-B: -68 (63) p: <0.001 vs. GP2 GP2 F-B: -83 (54) GP1-GP2: 15*	GP1 F: 113.04 Median GP2 F: 128.52 Median		Total glycated hemoglobin GP1 B: 8.6 (0.8) F: 7.3 (0.9) p: 0.08 vs. GP2 F-B: -1.3 (1.1) GP2 B: 8.6 (0.8) F: 7.2 (0.9) F-B: -1.4 (1) GP1-GP2: 0* Glycated hemoglobin ≤ 7.0% GP1 98 (41.7) p: 0.08 vs. GP2 GP2 116 (48.7)	EQ-5D GP1 F: 0.76 (95% CI: 0.73 – 0.8) p: overall 0.48 GP2 F: 0.76 (95% CI: 0.73 – 0.79)
Insulin aspart 70/30 vs. rapid-acting with long-acting insulin analogues
Joshi, 2005⁵²	GP1: Insulin aspart 70/30 (v) Mean: 40.19 U/day T: bid D: 12 weeks GP2: Insulin aspart (v) Mean: 28.26 U/day T: Before every meal D: 12 weeks Insulin glargine (v) Mean: 24.52 U/day T: Bedtime D: 12 weeks	GP1 B: 186.59 (47.35) F: 114.83 (18.68) F-B: -72* p: <0.0001 GP2 B: 190.23 (55.63) F: 110.61 (16.79) F-B: -79* p: <0.0001 GP1-GP2: 7*	PPG (time not specified) (mg/dL) GP1 B: 287.29 (58.4) F: 171.54 (28.75) F-B: -115* p: <0.0001 GP2 B: 281.42 (68.76) F: 177.52 (24.72) F-B: -103* p: <0.0001 GP1-GP2: -12*			GP1 B: 8.79 (1.13) F: 7.2 (0.83) F-B: -1.58 p: <0.0001 GP2 B: 8.53 (1.22) F: 7.37 (0.83) F-B: -1.16 p: <0.0001 GP1-GP2: -1* p: <0.05 A1c < 7%, n (%) GP1 52* (45.61) GP2 10* (32.26)
Insulin aspart 70/30 vs. premixed human insulins
Abraha-mian, 2005⁵³	GP1: Insulin aspart 70/30 (v) Mean: 0.49 U/kg (start), 0.61 U/kg (end) T: Breakfast, lunch, dinner D: 24 weeks GP2: NPH/regular 70/30 (v) Mean: 0.46 U/kg (start), 0.59 U/kg (end) T: Breakfast, dinner D: 24 weeks	GP1 F: 151† (SEM 4†) GP2 F: 143† (SEM 4†)	90-min PPG - after breakfast (mg/dL) GP1 F: 175† (SEM 10†) GP2 F: 189† (SEM 20†) 90-min PPG increment - after breakfast (mg/dL) F: p: 0.0572 vs. GP2 (favoring GP1)	GP1 F: 142 (SEM 7†) p: 0.0069 vs GP2 GP2 F: 166 (SEM 15†)	90-min PPG - after dinner (mg/dL) GP1 F: 154 (SEM 15†) p: 0.0022 vs GP2 GP2 F: 182 (SEM 7†) 90-min PPG increment - after dinner (mg/dL) F: p: 0.4096	GP1 B: 9.8 (1.55) F: 7.6 (1.1) F-B: -2* p: <0.0001 GP2 B: 9.85 (1.55) F: 7.7 (1.1) F-B: -2* p: <0.0001 GP1-GP2: 0* p: 0.641 vs GP2
Boehm, 200^{49 45}‡	GP1: Insulin aspart 70/30 (v) Start: 0.57 U/kg T: Breakfast, dinner D: 24 months GP2: NPH/regular 70/30 (v) Start: 0.57 U/Kg T: Breakfast, dinner D: 24 months	GP1 B: 151.2 F: 160.56 (SE 5.04) F-B: 10* GP2 B: 149.4 F: 148.32 (SE 4.86) F-B: -1* GP1-GP2: 12.06 (95% CI: -0.9 – 25.2) p: NS	90-min PPG - after breakfast (mg/dL) GP1 B: 212.4 F: 187.2 (SE 6.66) F-B: -25* GP2 B: 212.4 F: 205.2 (SE 6.48) F-B: -7* GP1-GP2: -18.18 (95% CI: -35.46 – -0.9) p: <0.05		90-min PPG - after dinner (mg/dL) GP1 B: 181.8 F: 165.96 (SE 5.94) F-B: -16* GP2 B: 187.2 F: 183.6 (SE 5.76) F-B: -3* GP1-GP2: -18.54 (95% CI: -34.02 – -3.24) p: <0.02	GP1 F: 8.35 (0.2) GP2 F: 8.13 (0.16) GP1-GP2: 0.03 (90% CI: -0.29 – 0.34) p: 0.89
Herman-sen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day		GP1 F: 13.9† GP2 F: 15.0† 2-hr PPG excursion GP1 F: 7.7 (2.7) p: <0.01 Ratio between treatments = 0.81 (95% CI: 0.71 - 0.93) p: <0.01
Kapitza, 2004⁵⁶	GP1: Insulin aspart 70/30 (NA) T: Breakfast (15 min after) D: 1 day GP2: NPH/regular 70/30 (NA) T: Breakfast (15 min before) D: 1 day		2-hr PPG increment - after breakfast (mg/dL) GP1 F: 52.2† GP2 F: 91.8†
Kapitza, 2004⁵⁶	GP1: Insulin aspart 70/30 (NA) T: Breakfast (15 min after) D: 1 day GP2: NPH/regular 70/30 (NA) T: Breakfast (right before) D: 1 day		2-hr PPG increment - after breakfast (mg/dL) GP1 F: 52.2† GP2 F: 81†
Kapitza, 2004⁵⁶	GP1: Insulin aspart 70/30 (NA) T: Breakfast (right before) D: 1 day GP2: NPH/regular 70/30 (NA) T: Breakfast (15 min before) D: 1 day		2-hr PPG increment - after breakfast (mg/dL) GP1 F: 81† GP2 F: 91.8†
Kapitza, 2004⁵⁶	GP1: Insulin aspart 70/30 (NA) T: Breakfast (right before) D: 1 day GP2: NPH/regular 70/30 (NA) T: Breakfast (right before) D: 1 day		2-hr PPG increment - after breakfast (mg/dL) GP1 F: 81† GP2 F: 81†
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks GP2: NPH/regular 70/30 (v) Start: 0.16 U/day Mean: 29 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks	GP1 F-B: -75 (72.3) GP2 F-B: -63 (86.2) GP1-GP2: -12*	GP1 B: 265† (±SE 5-10†) F: 190† (±SE 5-10†) F-B: -75 GP2 B: 266† (±SE 5-10†) F: 180† (±SE 5-10†) F-B: -86 GP1-GP2: 11		GP1 B: 250† (±SE 5-10†) F: 165† (±SE 5-10†) F-B: -85 GP2 B: 235† (±SE 5-10†) F: 168† (±SE 5-10†) F-B: -67 GP1-GP2: -18	GP1 F-B: -1.3 (SE 0.2†) GP2 F-B: -1.1 (SE 0.2†) GP1-GP2: 0*
McNally, 2007⁴⁸	GP1: Insulin aspart 70/30 (v) Start: 100 U/mL Mean: 68.8 U Range: 6 - 238.7 U T: Breakfast, dinner D: 16 weeks GP2: NPH/regular 70/30 (v) Start: 100 U/mL Mean: 66.6 U Range: 11.3 – 240 U T: Breakfast, dinner D: 16 weeks					GP1 F: 7.28 GP2 F: 7.22 GP1-GP2: 0.06 (95% CI: -0.04 – 0.17) p: 0.21	WHO-DTSQ GP1 F: 30.6 (5.84) GP2 F: 30.95 (5.01) GP1-GP2: -0.46 p: 0.25
Insulin aspart 70/30 vs. intermediate-acting human insulins
Christ-iansen, 2003¹³	GP1: Insulin aspart 70/30 (v) Start: insulin naïve: 8 - 16 U/day; taking NPH prior to trial: pretrial dose T: Breakfast, dinner D: 16 weeks GP2: NPH insulin (v) Start: insulin naïve: 8 - 16 U/day; taking NPH prior to trial: pretrial dose T: Breakfast, dinner D: 16 weeks	GP1 F-B: -25.2 GP2 F-B: -27 GP1-GP2: 2*				GP1 F-B: 0.67 p: <0.0001 vs. baseline GP2 F-B: 0.61 p: <0.0001 vs. baseline GP1-GP2: 0*
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks GP2: NPH insulin (v) Start: 0.16 U/day Mean: 28 U/day T: Bedtime D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks	GP1 F-B: -75 (72.3) GP2 F-B: -91 (72) GP1-GP2: 16*	GP1 B: 265† (±SE 5-10†) F: 190† (±SE 5-10†) F-B: -75 GP2 B: 266† (±SE 5-10†) F: 180† (±SE 5-10†) F-B: -86 GP1-GP2: 11		GP1 B: 250† (±SE 5-10†) F: 165† (±SE 5-10†) F-B: -85 GP2 B: 240† (±SE 5-10†) F: 190† (±SE 5-10†) F-B: -50 GP1-GP2: -35	GP1 F-B: -1.3 (SE 0.2†) GP2 F-B: -1.2 (SE 0.2†) GP1-GP2: 0*
Insulin aspart 70/30 vs. oral antidiabetic agents
Bebakar, 2007⁴⁶	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg Range: 0.16 U/kg (qd group) - 0.43 U/kg (bid group) T: once or twice daily D: 24 weeks GP2: OA agents (v) T: NR D: 24 weeks	GP1 F-B: -39.6 (54) p: <0.005 vs. GP2 GP2 F-B: -9 (48.24) GP1-GP2: -31* Fasting plasma glucose (time not specified) (mg/dL) GP1 F-B: -34.38 (39.96) p: <0.05 vs. GP2 GP2 F-B: -18.18 (39.6) GP1-GP2: -16*	90-min PPG - after breakfast (mg/dL) GP1 F-B: -43.38 (84.24) p: <0.05 vs. GP2 GP2 F-B: -14.04 (71.46) GP1-GP2: -18*	GP1 F-B: -36.72 (69.66) p: <0.005 vs. GP2 GP2 F-B: 1.44 (61.92) GP1-GP2: -38*	90-min PPG - after dinner (mg/dL) GP1 F-B: -68.22 (80.64) p: <0.005 vs. GP2 GP2 F-B: -9.36 (75.24) GP1-GP2: -59*	GP1 F-B: -1.16 (1.01) p: <0.005 vs. GP2 GP2 F-B: -0.58 (0.95) GP1-GP2: 0* A1c < 7%, n (%) GP1 32* (25) GP2 8* (12)
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.51 U/kg T: Breakfast, dinner D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg daily (end) T: once or twice daily D: 16 weeks	GP1-GP2: -0.18 (SE 4.86) p: NS	90-min PPG - after breakfast (mg/dL) GP1-GP2: -5.22 (SE 7.2) p: NS	GP1-GP2: 10.26 (SE 6.12) p: NS	90-min PPG - after dinner (mg/dL) GP1-GP2: 2.7 (SE 6.66) p: NS	GP1-GP2: 0.2 (SE 0.15) p: NS
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.3 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg T: NR D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg daily (end) T: once or twice daily D: 16 weeks	GP1-GP2: -1.26 (SE 4.86) p: NS	90-min PPG - after breakfast (mg/dL) GP1-GP2: -5.22 (SE 7.2) p: NS	GP1-GP2: 9.18 (SE 6.12) p: NS	90-min PPG - after dinner (mg/dL) GP1-GP2: -0.36 (SE 6.66) p: NS	GP1-GP2: -0.20 (SE 0.15) p: NS
Raskin, 2007⁶⁰	GP1: Insulin aspart 70/30 (v) Start: 6 U bid Mean: 0.6 U/kg/day T: Breakfast, dinner D: 34 weeks Metformin (fix) Mean: 2446 mg T: NR D: 34 weeks Pioglitazone (fix) Mean: 32.5 mg T: NR D: 34 weeks GP2: Metformin (fix) Mean: 2439 mg T: NR D: Unclear Pioglitazone (fix) Mean: 31.7 T: NR D: Unclear	GP1 B: 173.16 (39.78) p: NS F: 129.78 (50.04) p: < 0.001 F-B: -44.1 (49.86) p: < 0.001 GP2 B: 163.26 (35.46) F: 162.18 (40.86) F-B: 1.08 (43.56) GP1-GP2: -45* Met target FBG values of 79.2 – 109.8 mg/dL (4.4 - 6.1 mmol/L ) GP1 34* (37) GP2 2* (2)		GP1 B: 167.4† (SE 7.2†) F: 124.2 † (SE 3.6†) p: < 0.05 F-B: 43 GP2 B: 178.2† (SE 7.2 †) F: 156.6 † (SE 7.2 †) F-B: 22 GP1-GP2: 21*		GP1 B: 8* F: 6.5 (1) p: < 0.0001 F-B: -1.5 (1.1) p: <0.05 vs. baseline; < 0.0001 vs. GP2 GP2 B: 8* F: 7.8 (1.2) F-B: -0.2 (0.9) p: <0.05 vs. baseline GP1-GP2: -2* A1c < 7.0%, n (%) GP1 71* (76) p: < 0.001 GP2 21* (24) A1c < 6.5%, n (%) GP155 (59†) p: < 0.001 GP2 10 (11†)
Raz, 2003⁵⁷	GP1: Insulin aspart 70/30 (v) Start: 6 - 8 U bid T: Breakfast, dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks GP2: Glibenclamide (fix) Range: 7.5 – 15 mg T: Dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks	GP1 F-B: 58 p: NS vs. GP2 GP2 F-B: 34.2 GP1-GP2: 24*	PPG (time not specified) (mg/dL) GP1 F-B: 80.6 GP2 F-B: 52.9 GP1-GP2: 28*	GP1 F-B: 36.2 p: NS vs. GP2 GP2 F-B: 43.3 GP1-GP2: -7*	GP1 F-B: 72.8 p: NS vs. GP2 GP2 F-B: 47 GP1-GP2: 26*	GP1 B: 9.9 F: 9.4 F-B: 0.7 p: NS vs. GP2 GP2 B: 10.3 F: 10.1 F-B: 0.2 GP1-GP2: 1*
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.7 U/kg T: Breakfast, dinner D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 B: 178* F: 162† F-B: -16 p: NS GP2 B: 171* F: 169 (65) F-B: -2 p: NS GP1-GP2: -14*	90-min PPG - after breakfast (mg/dL) GP1 F: 196.2† GP2 F: 223.2†		90-min PPG - after dinner (mg/dL) GP1 F: 199.8† GP2 F: 212.4† 90-min PPG increment - after dinner (mg/dL) GP1-GP2: -8.1 (8.46) p: NS	GP1 B: 9.5 (1.3) F: 9 (1.3) F-B: -0.5 p: NS GP2 B: 9.4 (1.4) F: 9 (2.1) F-B: -0.4 p: NS GP1-GP2: -0.9*
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.5 U/kg T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 B: 184* F: 153 (45) p: 0.012 vs GP2 F-B: -31 p: NS GP2 B: 171* F: 169 (65) F-B: -2 p: NS GP1-GP2: -29*	90-min PPG - after breakfast (mg/dL) GP1 F: 178.2† GP2 F: 223.2†		90-min PPG increment - after dinner (mg/dL) F: 178.2† GP2 F: 212.4† GP1-GP2: -12.96 (8.64) p: NS	GP1 B: 9.6 (1.3) F: 8.4 (1.2) F-B: -1.2 p: NS GP2 B: 9.4 (1.4) F: 9 (2.1) F-B: -0.4 p: NS GP1-GP2: -0.64 (0.23) p: 0.005
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Continuation of OA agents (v) T: NR D: 16 weeks	GP1 B: 235.8† F: 163.8† F-B: 72 GP2 B: 225† F: 171† F-B: 54 GP1-GP2: 18*		GP1 B: 261† F: 156.6† F-B: 105 GP2 B: 243† F: 178.2† F-B: 65 GP1-GP2: 40*		GP1 F-B: -2.9 (1.5) p: < 0.001 vs. GP2 GP2 F-B: -2.1 (1.4) GP1-GP2: -0.65 (95% CI: -0.958 – -0.337) p: <0.001 vs. GP2 A1c ≤ 7.0%, n (%) GP1 42 (42) p: 0.012 GP2 27 (26.2) A1c after 3-month extension GP1 F: 7.2 (1.2) A1c ≤ 7.0% after 3-month extension, n (%) GP1 22 (51.2)	Diabetes Health Profile GP1 F-B: p: <0.001 vs. baseline GP2 F-B: p; <0.001 vs. baseline
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 44.8 U T: Breakfast, dinner D: 16 weeks metformin (varied) Start: 500 mg qd or bid or 850 mg qd T: NR D: 14 weeks (started after 2 weeks) GP2: Continuation of OA agents (v) T: NR D: 16 weeks	GP1 B: 234† F: 160.2† F-B: 74 GP2 B: 225† F: 171† F-B: 54 GP1-GP2: 20*		GP1 B: 252† F: 172.8† F-B: 80 GP2 B: 243† F: 178.2† F-B: 65 GP1-GP2: 15*		GP1 F-B: -3 (1.6) p: < 0.001 vs. GP2 GP2 F-B: -2.1 (1.4) GP1-GP2: -0.85 vs (95% CI: -1.163 – -0.537) p: <0.001 A1c ≤ 7.0%, n (%) GP1 45 (45) p: 0.002 GP2 27 (26.2) A1c after 3-month extension GP1 F: 7.2 (1.4) A1c ≤ 7.0% after 3-month extension, n (%) GP1 23 (54.8)	Diabetes Health Profile GP1 F-B: p: <0.001 vs. baseline GP2 F-B: p; <0.001 vs. baseline
Insulin aspart 70/30 vs. exenatide
Nauck, 2007⁴⁹	GP1: Insulin aspart 70/30 (v) Start: 15.7 U/day Mean: 24.4 U/day T: Breakfast, dinner D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks GP2: Exenatide (v) Start: 5 µg bid Range: 5 - 10 µg bid T: Breakfast D: 52 weeks 'Optimally' effective metformin and sulfonylurea (v) T: NR D: 52 weeks	GP1 B: 177.12† (SE 3.006†) F: 147.06† (SE 1.512†) p: 0.037 F-B: -30.06* p: <0.001 GP2 B: 173.34† (SE 2.16†) F: 153† (SE 2.16†) F-B: -20.34* p: <0.001 Fasting plasma glucose (time not specified) (mg/dL) GP1 F-B: -30.6 p: <0.001 GP2 F-B: -32.4 p: <0.001 GP1-GP2: 1.8 (95% CI: -7.2 – 10.8) p: 0.689	GP1 B: 229.5† (SE 3.6†) F: 171† (SE 3.06†) F-B: -58.5* p: <0.001 GP2 B: 222.84† (SE 3.06†) F: 153† (SE 2.16†) p: <0.001 F-B: -69.84* p: <0.001 GP1-GP2: 11.34* PPG excursion - after breakfast (NA) GP1 p: <0.001	GP1 B: 171.72† (SE 3.42†) F: 141.12† (SE 3.06†) p: <0.001 vs. baseline F-B: -30.6* GP2 B: 168.84† (SE 3.78†) F: 147.24† (SE 3.06†) p: <0.001 vs. baseline F-B: -21.6* GP1-GP2: -9*	GP1 B: 210.06† (SE 3.78†) F: 165.06† (SE 3.06†) p: <0.001 F-B: -45 GP2 B: 203.94† (SE 3.06†) F: 147.06† (SE 3.78†) p: <0.001 F-B: -57.6 p: <0.001 GP1-GP2: 13* PPG excursion after dinner (NA) GP1 p: <0.001	GP1 F-B: -0.89 p: <0.001 GP2 F-B: -1.04 p: <0.001 GP1-GP2: 0.15 (95% CI: -0.01 – 0.32) p: 0.067 A1c ≤ 7.0%, n (%) GP1 57 (24) p: 0.038 GP2 72 (32)
Insulin aspart 70/30 vs. insulin lispro 75/25
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day		GP1 F: 13.9† GP2 F: 14.5† 2-hr PPG excursion GP1 F: 7.7 (2.7) p: NS vs. GP2 Ratio between treatments = 0.97 (95% CI: 0.85 - 1.11) p: NS
Niskanen, 2004⁵⁵	GP1: Insulin aspart 70/30 (v) Mean: 0.65 to 0.67 U/kg T: Breakfast, dinner D: 12 weeks GP2: Insulin lispro 75/25 (v) Mean: 0.67 to 0.71 U/kg T: Breakfast, dinner D: 12 weeks	GP1 F: 136.8 GP2 F: 135 GP1-GP2: 3.6 (95% CI: -0.54 – 10.8) p: 0.422	90-min PPG after breakfast (mg/dL) GP1 F: 171 GP2 F: 174.6 GP1-GP2: -3.6 (95% CI: -18 – 9) p: 0.524	GP1 F: 8.7 GP2 F: 8.6 GP1-GP2: 0.1 (95% CI: -0.5 – 0.7) p: 0.824	90-min PPG after dinner (mg/dL) GP1 F: 172.8 GP2 F: 180 GP1-GP2: -7.2 (95% CI: -19.8 – 3.6) p: 0.186	GP1 F: 8.15 GP2 F: 8.01 GP1-GP2: 0.14 (90% CI: 0.008 – 0.275) p: 0.082
Insulin aspart 70/30 vs. insulin aspart 70/30 + oral antidiabetic agents
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.51 U/kg T: Breakfast, dinner D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.3 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg T: NR D: 16 weeks	GP1-GP2: 0.9 (SE 4.86) p: NS	90-min PPG - after breakfast (mg/dL) GP1-GP2: 0 (SE 7.38) p: NS	GP1-GP2: 1.08 (SE 6.3) p: NS	90-min PPG - after dinner (mg/dL) GP1-GP2: 3.06 (SE 6.66) p: NS	GP1-GP2: 0.39 (SE 0.15) p: <0.01
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.7 U/kg T: Breakfast, dinner D: 18 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.5 U/kg T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 B: 178* F: 162† F-B: -16 p: NS GP2 B: 184* F: 153 (45) F-B: -31 p: NS GP1-GP2: 15*	90-min PPG - after breakfast (mg/dL) GP1 F: 196.2† GP2 F: 178.2†		90-min PPG - after dinner (mg/dL) GP1 F: 199.8† GP2 F: 178.2† 90-min PPG increment - after dinner (mg/dL) GP1-GP2: 4.86 (8.46) p: NS	GP1 B: 9.5 (1.3) F: 9 (1.3) F-B: -0.5 p: NS GP2 B: 9.6 (1.3) F: 8.4 (1.2) F-B: -1.2 p: NS GP1-GP2: 0.60 (0.22) p: 0.008
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 44.8 U T: Breakfast, dinner D: 16 weeks Metformin (v) Start: 500 mg qd or bid or 850 mg qd T: NR D: 14 weeks (started after 2 weeks)	GP1 B: 235.8 † F: 163.8 † F-B: 72 GP2 B: 234 † F: 160.2 † F-B: 74 GP1-GP2: -2*		GP1 B: 261 † F: 156.6 † F-B: 105 GP2 B: 252 † F: 172.8 † F-B: 80 GP1-GP2: 25*		GP1 F-B: -2.9 (1.5) GP2 F-B: -3 (1.6) GP1-GP2: 0.20 (95% CI: -0.108– 0.514 ) A1c ≤ 7.0%, n (%) GP1 42 (42) p: 0.012 GP2 45 (45) p: 0.002 A1c after 3-month extension GP1 F: 7.2 (1.2) GP2 F: 7.2 (1.4) A1c ≤ 7.0% after 3-month extension, n (%) GP1 22 (51.2) GP2 23 (54.8)	Diabetes Health Profile GP1 F-B: p: <0.001 vs. baseline GP2 F-B: p; <0.001 vs. baseline
Insulin aspart 70/30 vs. rapid-acting insulin analogues with intermediate-acting human insulin
Hirao, 2008⁶¹	GP1: Insulin aspart 70/30 (NR) T: BID D: 6 months GP2: Insulin aspart (NR) T: TID D: 6 months NPH insulin (unclear) T: Optional multiple daily injections D: 6 months					GP1 B: 10.2 (2.1) F: 7.6 (1.3) p: NS F-B: -2.5 GP2 B: 10.4 (2) F: 7.8 (1.8) F-B: -2.5 GP1-GP2: 0* A1c < 7.0%, n (%) GP1 (32.1) p: NS GP2 (32.8) A1c < 6.5%, n (%) GP1 (17.9) GP2 (16.4)
Insulin lispro 75/25 vs. long-acting insulin analogues
Cox, 2007⁷⁴	GP1: Insulin lispro 75/25 (v) T: Breakfast, dinner D: 12 weeks Metformin (NR) T: NR D: 12 weeks GP2: Insulin glargine (v) T: Bedtime D: 12 weeks Metformin (NR) T: NR D: 12 weeks	GP1 F: 8.5 (1.5) p: 0.056 GP2 F: 7.8 (2)	GP1 F: 11 (1.9) p: 0.642 GP2 F: 10.9 (2.1) GP1-GP2: 2.2 (0.7) p: NS	GP1 F: 176.4 (45) p: 0.076 GP2 F: 192.6 (54)	GP1 F: 198 (41.4) p: 0.001 GP2 F: 221.4 (52.2) GP1-GP2: 55.8 (23.4) p: NS		BDI-II GP1 B: 8.2 (6) p: NS F: 5.5 (3.8) p: 0.115 F-B: -2* p: 0.018 GP2 B: 8.2 (6) F: 6.8 (5.9) F-B: -1* p: NS GP1-GP2: -1*
Jacober, 2006⁶⁴	GP1: Insulin lispro 50/50 (v) Mean: 0.353 IU/kg; 36.73 IU T: Breakfast, lunch D: 4 months Insulin lispro 75/25 (v) T: Dinner D: 4 months Existing OA agents (NR) T: NR D: 4 months GP2: Insulin glargine (v) Mean: 0.276 IU/kg; 27.98 IU T: Bedtime D: 4 months Existing OA agents (NR) T: NR D: 4 months	GP1 F: 130† (25†) p: NS GP2 F: 125† (15†)	GP1 F: 153.5 (35.6) p: 0.0034 GP2 F: 172.1 (35)	GP1 F: 123.1 (36.1) p: 0.0205 GP2 F: 139 (41.9)	GP1 F: 145.4 (38.2) p: 0.0066 GP2 F: 161.9 (42.3)	Overall results GP1 B: 8* F: 7.08 (0.11) p: 0.003 F-B: -1.01 (0.1) p: 0.0068 vs. GP2 GP2 B: 8* F: 7.34 (0.11) F-B: -0.75 (0.1) GP1-GP2: 0* 1st per. results GP1 F: 6.97 (0.62†) GP2 F: 7.32 (0.93†) 2nd per. results GP1 F: 7.22 (0.77†) GP2 F: 7.33 (0.92†) A1c ≤ 7%, n (%) GP1 26* (44) p: 0.1026 GP2 18* (31)
Malone, 2004⁶⁵	GP1: Insulin lispro 75/25 (v) Mean: 0.62 U/kg T: Breakfast, dinner D: 16 weeks Metformin (NR) Mean: 1945 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.57 U/kg T: Bedtime D: 16 weeks Metformin (NR) Mean: 1997 mg Range: 1500 - 2550 mg T: NR D: 16 weeks	GP1 B: 150* F: 139.3 (36.6) p: <0.001 F-B: -11.3 (44.5) p: 0.001 vs. GP2 GP2 B: 153* F: 123.9 (34.9) F-B: -29 (47.4) GP1-GP2: 18* Met target FBG of 90 to 126 mg/dL, n (%) GP1 31 (45) p: 0.019 GP2 44 (65)	GP1 F: 156.4 (43.6) p: 0.012 GP2 F: 171.1 (44.9) Met target 2-hr PPG of 144 to 180 mg/dL, n (%) GP1 55 (80) p: 0.036 GP2 43 (63)		GP1 F: 164.8 (42.5) p: <0.001 GP2 F: 193.8 (51) Met target 2-hr PPG of 144 to 180 mg/dL, n (%) GP1 50 (72) p: <0.001 GP2 29 (43)	GP1 B: 8.7 (1.3) F: 7.4 (1.1) p: 0.002 F-B: -1.32 (1.01) p: 0.003 vs. GP2; <0.001 vs. baseline GP2 B: 8.7 (1.3) F: 7.8 (1.1) F-B: -0.93 (0.89) p: <0.001 vs. baseline GP1-GP2: 0* A1c ≤ 7.0%, n (%) GP1 30 (42) p: <0.001 GP2 13 (18)
Malone, 2005⁶⁶	GP1: Insulin lispro 75/25 (v) Mean: 0.42 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 2128 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.36 U/kg T: Bedtime D: 16 weeks Metformin (fix) Mean: 2146 mg Range: 1500 - 2550 mg T: NR D: 16 weeks	GP1 F: 142.2 (34.56) p: 0.007 GP2 F: 133.02 (35.28) Met target FBG of < 126 mg/dL (7.0 mmol/L), n (%) GP1 33* (34) p: 0.01 GP2 49* (51)	GP1 F: 169.92 (46.08) p: <0.001 GP2 F: 194.94 (49.32) Met target 2-hr PPG of < 180 mg/dL (10 mmol/L), n (%) GP1 64* (66) p: <0.001 GP2 41* (42)		GP1 F: 172.62 (45) p: <0.001 GP2 F: 200.7 (45.36) Met target 2-hr PPG of < 180 mg/dL (10 mmol/L), n (%) GP1 62* (64) p: <0.001 GP2 39* (40)	GP1 B: 9* F: 7.54 (0.87) p: <0.001 F-B: -1 (0.85) p: <0.001 vs. GP2 GP2 B: 8* F: 8.14 (1.03) F-B: -0.42 (0.92) GP1-GP2: -1* A1c ≤ 7.0%, n (%) GP1 (30) p: 0.002 GP2 (12) A1c ≤ 6.5%, n (%) GP1 p: 0.1
Roach, 2006⁶³	GP1: Insulin lispro 75/25 (v) Mean: 23 U (morning) and 37 U (evening) Range: 0 – 72 U (morning); 11 – 88 U (evening) T: Breakfast, dinner D: 12 weeks OA agents (NR) Start: Current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks GP2: Insulin glargine (v) Mean: 44 U Range: 14 - 100 U T: Breakfast D: 12 weeks OA agents (NR) Start: Current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks	GP1 F: 104.4 (20.16) p: 0.649 GP2 F: 99 (38.52)	GP1 F: 187.2 (43.2) p: 0.551 GP2 F: 180 (37.8)	GP1 F: 91.8 (17.1) p: 0.141 GP2 F: 100.8 (25.38)	GP1 F: 144 (39.24) p: 0.005 GP2 F: 176.4 (36)	GP1 F: 6.9 (0.52) p: 0.035 GP2 F: 7.3 (0.81)
Insulin lispro 75/25 vs. premixed human insulins
Coscelli, 2003⁶⁷	GP1: Insulin lispro 75/25 (v) Mean: 38.1 Range: 12 - 72 T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days GP2: NPH/regular 70/30 (v) Mean: 37.3 Range: 10 - 72 T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days		GP1 F: 157 (43.2) p: <0.05 GP2 F: 180 (43.2) 2-hr PPG excursion GP1 F: 2.4 (48.9) p: 0.08 GP2 F: 17.9 (41.43)		2-hr PPG excursion GP1 F: 12.2 (48.01) p: <0.05 GP2 F: 35.5 (36.92)
Hermansen, 200258	GP1: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day		GP1 F: 14.5† GP2 F: 15.0† 2-hr PPG excursion GP1 F: 8.5 (3.3) Ratio between treatments = 0.81 (95% CI: 0.72 – 0.94) p: <0.01 GP2 F: 9.4 (2.7) Ratio between treatments = ref
Herz, 2002⁷¹	GP1: Insulin lispro 75/25 (v) Mean: 26.1 U T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 26.2 U T: Breakfast, dinner D: 4 weeks		GP1 F: 189 (SE 7.2) p: 0.016 GP2 F: 208.8 (SE 7.2)
Herz, 2003¹²	GP1: Insulin lispro 75/25 (v) Mean: 31.6¶ (morning) and 26.8¶ U (evening) and 32.4§ (morning) and 27.6§ U (evening) T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 32.3¶ (morning), 26.4¶ U (evening) and 33.3§ (morning), 27.5§ U (evening) T: Breakfast, dinner D: 4 weeks	GP1 F: 117† GP2 F: 117†	GP1 F: 223.2† GP2 F: 259.2† 2-hr PPG excursion GP1 F: 99 (SE 6.12) p: 0.002 GP2 F: 129.6 (SE 6.12)	GP1 F: 135† GP2 F: 135†	GP1 F: 181.8† GP2 F: 201.6† 2-hr PPG excursion GP1 F: 43.2 (SE 4.86) p: 0.018 GP2 F: 61.2 (SE 4.86)
Malone, 2000⁴⁴	GP1: Insulin lispro 75/25 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days GP2: NPH/regular 70/30 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days		GP1 F: 221.4 (52.2) p: 0.066 GP2 F: 230.4 (54) 2-hr PPG excursion GP1 F: 60.3 (41.04) p: <0.001 GP2 F: 74.34 (40.68)
Mattoo, 2003⁷⁰	GP1: Insulin lispro 75/25 (NR) Mean: 20 U (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks GP2: NPH/regular 70/30 (NR) Mean: 20 U (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks	GP1 F: 160.2 (54) p: 0.393 GP2 F: 163.8 (57.6)	GP1 F: 208.8 (66.6) p: 0.104 GP2 F: 216 (64.8) 2-hr PPG excursion GP1 F: 48.6 (57.6) p: 0.397 GP2 F: 54 (55.8)	GP1 F: 127.8 (39.6) p: 0.034 GP2 F: 135 (46.8)	GP1 F: 189 (57.6) p: 0.001 GP2 F: 208.8 (61.2) 2-hr PPG excursion GP1 F: 61.2 (52.2) p: 0.007 GP2 F: 72 (57.6)
Roach, 1999⁷³	GP1: Insulin lispro 75/25 (v) Mean: 0.37 (morning), 0.28 U/kg (evening) T: Breakfast, dinner D: 13 weeks GP2: NPH/regular 70/30 (v) Mean: 0.36 (morning), 0.27 (evening) T: Breakfast, dinner D: 13 weeks	GP1 F: 154.8† p: NS GP2 F: 157.5†	GP1 F: 161.1 (39.06) p: 0.017 GP2 F: 180 (41.04)	GP1 F: 170.1† p: NS GP2 F: 169.2†	GP1 F: 167.04 (45.18) p: 0.014 GP2 F: 184.86 (49.68)	GP1 F: 7.8 p: 0.408 GP2 F: 8.1
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day		GP1 F: 198 (67.5) p: <0.05 GP2 F: 213 (47) p: <0.05
Sun, 2007⁷⁵	GP1: Insulin lispro 75/25 (NR) T: bid D: 18 months GP2: NPH/regular 70/30 (unclear) T: bid D: 18 months					GP1 B: 8.6 (3.7) F: 8* F-B: -0.87 GP2 B: 8.4 (3.9) F: 7* F-B: -0.75 GP1-GP2: -0.16 p: <0.05
Insulin lispro 75/25 vs. oral antidiabetic agents
Herz, 2002⁷²	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg Mean: 0.46 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glyburide (fix) Start: 15 mg/day T: Breakfast, dinner D: 16 weeks	GP1 B: 199.44 (SE 6.3) p: 0.139 vs. GP2 F: 147.06 (SE 4.14) p: <0.001 vs. GP2 F-B: -52.74 (SE 5.94) p: <0.001 vs. GP2 GP2 B: 187.74 (SE 4.68) F: 176.76 (SE 5.22) F-B: -8.82 (SE 5.04) GP1-GP2: -44*	GP1 B: 255.6 (SE 9) p: 0.621 vs. GP2 F: 174.96 (SE 6.66) p: <0.001 vs. GP2 F-B: -80.82 (SE 9) p: <0.001 vs. GP2 GP2 B: 261.18 (SE 7.02) F: 236.52 (SE 7.02) F-B: -22.5 (SE 7.02) GP1-GP2: -59*	GP1 B: 222.48 (SE 8.82) p: 0.216 vs. GP2 F: 175.68 (SE 5.94) p: 0.120 vs. GP2 F-B: -47.34 (SE 7.92) p: 0.002 vs. GP2 GP2 B: 207.18 (SE 8.64) F: 189.9 (SE 7.02) F-B: -14.76 (SE 6.48) GP1-GP2: -32*	GP1 B: 241.2 (SE 9.54) p: 0.711 vs. GP2 F: 181.98 (SE 6.84) p: <0.001 vs. GP2 F-B: -58.86 (SE 8.82) p: <0.001 vs. GP2 GP2 B: 245.88 (SE 8.28) F: 227.52 (SE 7.56) F-B: -14.94 (SE 7.56) GP1-GP2: -44*	GP1 B: 9.82 (1.51) F: 8.64 (SE 0.17) p: <0.001 vs. GP2 F-B: -1.14 (SE 0.18) p: 0.001 vs. GP2 GP2 B: 9.9 (1.3) F: 9.45 (SE 0.16) F-B: -0.36 (SE 0.15) GP1-GP2: -1*	Treatment acceptance questionnaire (satisfaction rated from 1 (very low) to 5 (very high) GP1 F: 4.35 p: 0.014 vs. GP2 GP2 F: 3.98 Willingness to continue treatment GP1 (92) p: 0.041 GP2 (79)
Malone, 2003⁶⁸	GP1: Insulin lispro 75/25 (v) Mean: 0.19 (morning), 0.14 U/kg (evening) T: Breakfast, dinner D: 16 weeks Metformin (v) Mean: 1813 mg/day Range: 1500 – 2550 mg/day T: 2 to 3 times/day D: 16 weeks GP2: Metformin (v) Mean: 1968 mg/day Range: 1500 - 2550 mg/day T: 2 to 3 times/day D: 16 weeks Glibenclamide (v) Mean: 14.2 mg/day T: NR D: 16 weeks	GP1 B: 239.4 (68.22) F: 156.06 Median (60.48) F-B: -83* GP2 B: 233.82 (68.04) F: 169.74 Median (61.02) F-B: -64* GP1-GP2: -19* p: 0.173	GP1 B: 252 (+/- SE 246.6 - 257.4) F: 147.6 (+/- SE 145.8 - 151.2) F-B: -104* GP2 B: 259.2 (+/- SE 252 - 273.6) F: 190.8 (+/- SE 185.4 - 199.8) F-B: -68* GP1-GP2: -36* 2-hr PPG (time not specified) (mg/dL) GP1 F-B: -124.02 (84.42) p: 0.007 GP2 F-B: -68.94 (84.96) GP1-GP2: -55* 2-hr PPG excursion (time not specified) (mg/dL) GP1 F-B: -40.86 (69.66) p: 0.009 GP2 F-B: -3.96 (35.82) GP1-GP2: -37*			Test meal patients GP1 B: 9.64 (1.6) F: 7.29 (1.12) p: 0.192 vs. GP2 F-B: -3* GP2 B: 9.78 (1.83) F: 7.53 (1.27) F-B: -2* GP1-GP2: -1* All patients GP1 B: 9.17 (1.5) F: 7.29 (1.00) p: 0.661 vs. GP2 F-B: -1.87 (1.35) p: <0.001 GP2 B: 9.27 (1.55) F: 7.33 (1.14) F-B: -1.98 (1.28) p: <0.001 GP1-GP2: 0* A1c < 7.0%, (%) GP1 (40) GP2 (41) A1c < 6.5%, (%) GP1 (18) GP2 (19)
Tirgoviste, 2003⁴³	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glibenclamide (v) Start: 15 mg T: Breakfast, dinner D: 16 weeks	GP1 B: 221* F: 171 F-B: -50.4 p: <0.01 GP2 B: 209* F: 189 F-B: -19.8 GP1-GP2: -30*	GP1 B: 279* F: 187.2 F-B: -91.8 p: <0.001 GP2 B: 265* F: 234 F-B: -30.6 GP1-GP2: -61*	GP1 B: 233* F: 192.6 F-B: -39.6 p: <0.05 GP2 B: 219* F: 205.2 F-B: -14.4 GP1-GP2: -26*	GP1 B: 272* F: 192.6 F-B: -79.2 p: <0.001 GP2 B: 261* F: 234 F-B: -27 GP1-GP2: -52*	GP1 B: 9* F: 8.5 (1.3) p: 0.001 F-B: -1.4 p: 0.004 GP2 B: 10* F: 9.4 (1.8) F-B: -0.7 GP1-GP2: 0*
Insulin lispro 75/25 vs. insulin lispro 50/50
Roach, 2003⁶⁹	GP1: Insulin lispro 75/25 (v) Mean: 31.3 (morning), 27.6 U (evening) T: Breakfast, dinner D: 8 weeks GP2: Insulin lispro 50/50 (v) Mean: 31.5 U T: Breakfast D: 8 weeks Insulin lispro 75/25 (v) Mean: 27.9 U T: Dinner D: 8 weeks	GP1 F: 160.2 (SE 5.4) p: 0.129 GP2 F: 171 (SE 5.4)	GP1 F: 223.2 (SE 5.94) p: 0.0012 GP2 F: 196.2 (SE 5.04) 2-hr PPG excursion GP1 F: 63 (SE 5.04) p: <0.001 GP2 F: 25.2 (SE 5.04)			GP1 F: 8.14 (SE 1.07) p: 0.919 GP2 F: 8.14 (SE 1.14)
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day		GP1 F: 198 (67.5) p: <0.05 GP2 F: 159 (52.3) p: <0.05
Insulin lispro 50/50 vs. long-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg Mean: 0.59 IU/kg T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin glargine (v) Start: 0.16 IU/kg Mean: 0.43 IU/kg T: Bedtime D: 24 weeks	GP1 B: 167.4 (37.8) F: 151* F-B: -16.2 (32.4) p: <0.001 vs. GP2 GP2 B: 172.8 (43.2) F: 126* F-B: -46.8 (43.2) GP1-GP2: 31*	GP1 B: 214.2 (50.4) F: 164* F-B: -50.4 (52.2) p: 0.43 vs. GP2 GP2 B: 219.6 (55.8) F: 173* F-B: -46.8 (59.4) GP1-GP2: -3* 2-hr PPG excursion GP1 B: 48.6 (32.4) F: 17* F-B: -32.4 (43.2) p: <0.001 vs. GP2 GP2 B: 45 (39.6) F: 43* F-B: -1.8 (39.6) GP1-GP2: -30*	GP1 B: 166.5† (SE 5.4†) F: 144† (SE 7.56†) F-B: -22.5* GP2 B: 174.06† (SE 5.4†) F: 159.12† (SE 7.56†) F-B: -14.94* GP1-GP2: -7.56*	GP1 B: 198† (SE 5.94†) F: 149.94† (SE 5.94†) F-B: -48.06* GP2 B: 208.44† (SE 7.38†) F: 207† (SE 7.38†) F-B: -1.44* GP1-GP2: -46.62*	GP1 B: 8.1 (1.2) F: 7* F-B: -1.2 (1.1) p: <0.001 vs. GP2 GP2 B: 8.1 (1.3) F: 8* F-B: -0.3 (1.1) GP1-GP2: -1* A1c < 7%, n (%) GP1 29* (59.3) GP2 12* (24.5)	Willing to continue current treatment at end of study GP1 F: 83.3% GP2 F: 77.4% Overall satisfaction based on 5-point Likert scale (non validated): proportion with high or very high treatment satisfaction GP1 B: 18.5% F: 63% GP2 B: 26.4% F: 50.9%
Robbins, 2007⁷⁹	GP1: Insulin lispro 50/50 (v) Mean: 0.7 U/kg T: Breakfast, lunch, dinner D: 24 weeks Metformin (fix) Mean: 1641 mg T: bid D: 24 weeks GP2: Insulin glargine (v) Mean: 0.6 U/kg T: Bedtime D: 24 weeks Metformin (fix) Mean: 1636 mg T: bid D: 24 weeks	GP1 B: 152† (95% CI: 146† – 158†) F: 146 (33) p: < 0.001 vs GP2 F-B: 6 GP2 B: 148† (95% CI: 143† – 153†) F: 118 (29) F-B: 30 GP1-GP2: -24* FBG < 120 mg/dL (6.7 mmol/L), n (%) GP1 37 (24.2) p: < 0.001 vs GP2 GP2 91 (61.1)	GP1 B: 183† (95% CI: 177† – 191†) F: 156 (39) p: 0.03 vs. GP2 F-B: 27 GP2 B: 180† (95% CI: 174† – 186†) F: 166 (46) F-B: 14 GP1-GP2: 13*	GP1 B: 168† (95% CI: 161† – 174†) F: 149 (36) p: 0.04 vs GP2 F-B: 19 GP2 B: 170† (95% CI: 163† – 177†) F: 160 (50) F-B: 10 GP1-GP2: 9*	GP1 B: 183† (95% CI: 174† – 190†) F: 157 (40) p: <0.001 vs GP2 F-B: 26 GP2 B: 183† (95% CI: 176† – 191†) F: 193 (57) F-B: -10 GP1-GP2: 36* 2-hr PPG excursion GP1 F-B: -8 (50) p: < 0.001 vs. GP2 GP2 F-B: 23 (54) GP1-GP2: -31*	GP1 B: 7.8 (0.9) p: F: 7.1 (0.9) p: <0.001 vs. GP2 F-B: -0.7 (0.9) p: <0.001 vs GP2 GP2 B: 7.8 (1) F: 7.5 (1) F-B: -0.4 (0.9) GP1-GP2: -1* A1c ≤ 7.0%, n (%) GP1 85 (56.3) p: 0.005 vs GP2 GP2 58 (39.7) A1c ≤ 6.5%, n (%) GP1 46 (30.5) p: 0.001 vs GP2 GP2 21 (14.4)
Insulin lispro 50/50 vs. rapid-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg mean Mean: 0.59 IU/kg T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin lispro (v) Start: 0.25 IU/kg mean Mean: 0.50 IU/kg T: Breakfast, lunch, dinner D: 24 weeks	GP1 B: 167.4 (37.8) F: 151* F-B: -16.2 (32.4) GP2 B: 176.4 (50.4) F: 160* F-B: -16.2 (39.6) GP1-GP2: 0*	GP1 B: 214.2 (50.4) F: 164* F-B: -50.4 (52.2) GP2 B: 205.2 (61.2) F: 151* F-B: -54 (63) GP1-GP2: 4* 2-hr PPG excursion GP1 B: 48.6 (32.4) F: 17* F-B: -32.4 (43.2) GP2 B: 28.8 (43.2) F: -9* F-B: -37.8 (52.2) GP1-GP2: 6*	GP1 B: 166.5† (SE 5.4†) F: 144† (SE 7.56†) F-B: -22.5* GP2 B: 169.38† (SE 5.4†) F: 145.44† (SE 7.56†) F-B: -23.94* GP1-GP2: -1.44	GP1 B: 198† (SE 5.94†) F: 149.94† (SE 5.94†) F-B: -48.06* GP2 B: 205.38† (SE 4.5†) F: 141.12† (SE 4.5†) F-B: -64.26* GP1-GP2: 16.2	GP1 B: 8.1 (1.2) F: 7* F-B: -1.2 (1.1) GP2 B: 8.2 (1.2) F: 7* F-B: -1.1 (1.1) GP1-GP2: 0* A1c < 7%, n (%) GP1 29* (59.3) GP2 20* (40.4)	Willing to continue current treatment at end of study GP1 F: 83.3% GP2 F: 88.5% Overall satisfaction based on 5-point Likert scale (nonvalidated): proportion with high or very high treatment satisfaction GP1 B: 18.5% F: 63% GP2 B: 21.2% F: 65.4%
Insulin lispro 50/50 vs. rapid-acting with long-acting insulin analogues
Rosenstock, 2008⁸⁰	GP1: Insulin lispro 50/50 (v) Start: Insulin glargine dose at entry (52.5 U) Mean: 123 U T: Breakfast, lunch, dinner D: 24 weeks Insulin lispro 75/25 (v) Start: Allowed to switch evening dose to insulin lispro 75/25 T: Dinner D: Unclear GP2: Insulin glargine (v) Start: 50% of insulin glargine dose at entry (54.9 U) Mean: 70 U T: Bedtime D: 24 weeks Insulin lispro (v) Start: 50% of insulin glargine dose at entry divided in 3 equal doses (54.9 U) Mean: 76 U T: Breakfast, Lunch, dinner D: 24 weeks	GP1 F: 159 (55) p: 0.013 GP2 F: 147 (43)	GP1 F: 174 (56) p: 0.002 GP2 F: 155 (53)	GP1 B: 208† F: 144† p: > 0.05 vs. GP2 F-B: 64 GP2 B: 212† F: 150† F-B: 62 GP1-GP2: 2*		GP1 B: 8.8 (1) p: 0.598 F: 6.95 F-B: -1.87 p: 0.021 GP2 B: 8.9 (1.1) F: 6.78 F-B: -2.09 GP1-GP2: 0.22 (90% CI: 0.07 – 0.38) A1c < 7.5%, n (%) GP1 (82) GP2 (83) A1c < 7.0%, n (%) GP1 81 (54) p: < 0.05 GP2 101 (69) A1c < 6.5%, n (%) GP1 53 (35) p: < 0.05 GP2 74 (50) A1c < 6.0%, n (%) GP1 (12.5†) GP2 (12.5†)
Insulin lispro 50/50 vs. premixed human insulins
Roach, 1999¹⁰	GP1: Insulin lispro 50/50 (v) Mean: 0.31 U/kg T: Breakfast D: 3 months Insulin lispro 75/25 (v) Mean: 0.26 U/kg T: Dinner D: 3 months GP2: NPH/regular 50/50 (v) Mean: 0.32 U/kg T: Breakfast D: 3 months NPH/regular 70/30 (v) Mean: 0.26 U/kg T: Dinner D: 3 months	GP1 F: 160.38 p: NS GP2 F: 162.18	GP1 F: 150.3 p: <0.001 GP2 F: 182.16 2-hr PPG excursion GP1 F: -10.44 p: <0.001 GP2 F: 21.42	GP1 F: 171 p: 0.01 GP2 F: 166.68	GP1 F: 179.28 p: NS GP2 F: 188.64 2-hr PPG excursion GP1 F: 6.48 p: NS GP2 F: 21.96	GP1 F: 7.73 p: 0.371 GP2 F: 7.66
Schern-thaner, 2004⁷⁷	GP1: Insulin lispro 50/50 (v) Mean: 64.6 IU T: Breakfast, lunch, dinner D: 12 weeks Diet/exercise D: 12 weeks GP2: NPH/regular 70/30 (v) Mean: 61.8 IU T: Breakfast, dinner D: 12 weeks Diet/exercise D: 12 weeks	GP1 B: 155* F: 177.7 (SE 9.6) F-B: 23.3 (SE 7.8) p: 0.005 vs. baseline GP2 B: 154* F: 147.4 (SE 6.3) F-B: -7 (SE 8) p: 0.387 vs. baseline GP1-GP2: 30* p: <0.001	GP1 B: 198* F: 189.8 (SE 10.2) F-B: -8.3 (SE 11) p: 0.456 vs. baseline GP2 B: 198* F: 191.3 (SE 10.5) F-B: -6.9 (SE 7.8) p: 0.384 vs. baseline GP1-GP2: -1* p: 0.836 2-hr PPG increment GP1 F-B: -32.3 (SE 9.7) p: 0.002 vs. baseline GP2 F-B: 1 (SE 7.3) p: NS vs. baseline GP1-GP2: -33* p: <0.001	GP1 B: 192* F: 174.8 (SE 7.3) F-B: -17.3 (SE 9.6) p: 0.079 vs. baseline GP2 B: 192* F: 187.8 (SE 9.5) F-B: -4.3 (SE 8.5) p: 0.614 vs. baseline GP1-GP2: -13* p: 0.064	GP1 B: 209* F: 166.3 (SE 7.2) F-B: -42.8 (SE 10) p: <0.001 vs. baseline GP2 B: 209* F: 198.2 (SE 10) F-B: -10.9 (SE 9.7) p: 0.268 vs. baseline GP1-GP2: -32* p: <0.001 2-hr PPG increment GP1 F-B: -21 (SE 9.7) p: 0.037 vs. baseline GP2 F-B: -4.6 (SE 8.2) p: NS vs. baseline GP1-GP2: -16* p: 0.055	GP1 B: 9* F: 7.6 (SE 1.1) F-B: -0.8 (SE 1.1) p: <0.001 vs. baseline GP2 B: 8* F: 8.1 (SE 1.4) F-B: -0.3 (SE 1.1) p: 0.034 vs. baseline GP1-GP2: -1* p: 0.021
Schwartz, 2006⁶²	GP1: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day		GP1 F: 159 (52.3) p: <0.05 GP2 F: 213 (47) p: <0.05
Yamada, 2007⁷⁸	GP1: Insulin lispro 50/50 (v) Start: Current dose Mean: 0.37 (start), 0.38 U/kg (end) T: bid D: 4 months GP2: NPH/regular 70/30 (v) Start: current dose Mean: 0.34 (start), 0.37 U/kg (end) T: bid D: 4 months NPH/regular 50/50 (v) Start: current dose Mean: 0.34 (start), 0.37 U/kg (end) T: bid D: 4 months	GP1 B: 130.3 (50.7) F: 158.5 (63.4) F-B: 28* p: NS vs. baseline GP2 B: 141.8 (51.9) F: 136.4 (47.2) F-B: -6* p: NS vs. baseline GP1-GP2: 34* p: NS				GP1 B: 7.59 (0.44) F: 7.24 (0.49) F-B: -1* p: <0.05 vs. baseline GP2 B: 7.33 (0.58) F: 7.29 (0.65) F-B: 0* p: NS vs. baseline GP1-GP2: -1* p: <0.05

Evidence Table 5. Comparative safety of premixed insulin analogues and other diabetes treatments on hypoglycemia
Author, year	Intervention	Mild hypoglycemia, n (%)	Moderate hypoglycemia, n (%)	Serious hypoglycemia, n (%)	Daytime hypoglycemia, n (%)	Nighttime hypoglycemia, n (%)	Other hypoglycemia, n (%)
* Number has been imputed. ◊ Among those not using thiazolidinediones. ‡ One-hundred and four (36%) of the 291 participants of this trial are patients with type 1 diabetes. The remaining population has type 2 diabetes and is the same study population as Boehm 2004.⁴⁵ Only data for the Boehm 2004 study is presented because it has the longest followup. ¶ Results occurring during the outpatient phase. § Results occurring during the inpatient phase. µg = microgram; B = baseline; BG = blood glucose; bid = twice daily ; BS = blood sugar; CGMS = Continuous Glucose Monitoring System; CI = confidence interval; CNS = central nervous system; D = duration; dl = deciliter; F = final; F-B = mean difference from baseline; fix = fixed dosing; GP = group; GP1-GP2 = mean difference between the difference from baseline; h = hour; IQR = interquartile range; IU = international unit; kg = kilogram; L = liter; mg = milligram; ml = milliliter; mmol = millimole; NPH = neutral protamine Hagedorn; NR = not reported; NS = not significant; OA = oral antidiabetic; p = p-value; PG = plasma glucose; qd = once daily; RR = relative risk; SD = standard deviation; SE = standard error; T = time of day when insulin taken; tid = thrice daily; U = unit; v = dosing varied
Insulin aspart 70/30 vs. long-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin detemir (v) Start: 16 median IU/day Range: 10 – 24 IU/day T: Bedtime, twice if required D: 1 year Usual care D: 1 year		Grade 2: symptoms and BG < 56 mg/dL GP1 Median number of events/patient-year: 3.9 (IQR 1.0 - 9.0) p: 0.01 GP2 Median number of events/patient-year: 0 (IQR 0 - 2.0)	Grade 3: third party assistance required GP1 11 (4.7) p: overall 0.20 Median number of events/patient-year: 0 p: overall 0.10 GP2 4 (1.7) Median number of events/patient-year: 0			Grades 1, 2, or 3 GP1 216 (91.9) p: overall < 0.001 GP2 173 (73.9)
Kann, 2006⁵⁰	GP1: Insulin aspart 70/30 (v) Start: 0.1 U/kg bid Mean: 0.4 U/kg T: Breakfast, dinner D: 26 weeks Metformin (v) Start: 500 mg bid or current dose T: Breakfast, dinner D: 26 weeks GP2: Insulin glargine (v) Start: 0.2 U/kg qd Mean: 0.3 U/kg T: preferred time (constant through study) D: Glimepiride (v) Start: 1 mg daily or current dose T: Breakfast D: 26 weeks	Treat self, PG < 55.8 mg/dL (3.1 mmol/L) GP1 26* (20.3) p: 0.0124 GP2 11* (9)		Unable to treat self GP1 1 (1) GP2 1 (1) Hypoglycemic coma GP1 2 (1.6) GP2 0 (0)	% mild episodes that occurred in daytime GP1 number (%) of events: 61 (77) GP2 number (%) of events: 25 (71)		Symptoms only GP1 14* (10.6) GP2 9* (6.6)
Raskin, 2005^{39 40}	GP1: Insulin aspart 70/30 (v) Start: 10 or 12 U/day T: Breakfast, dinner D: Unclear Metformin (v) Range: 1500 – 2550 mg/day T: NR D: Unclear GP2: Insulin glargine (v) Start: 10-12 U/day T: Bedtime D: Unclear Metformin (v) Range: 1500 – 2550 mg/day T: NR D: Unclear	PG < 56 mg/dL with or without symptoms, self-treated GP1 46.4* (43) event rate: 3.4/patient-year p: < 0.05 GP2 18.2* (16) event rate: 0.7/patient-year GP1◊ 33 (41.8) number of events: 121 GP2 ◊ 11 (14.1) number of events: 23		Neurological symptoms, required assistance, PG < 56 mg/dL or reversal with treatment GP1 0 (0) GP2 1 (0.88) GP1◊ 0 (0) GP2◊ 0 (0)		Mild or serious between 11pm and 8am GP1◊ 19.8* (25) p: 0.021 GP2◊ 7.8* (10)	Symptoms but PG ≥ 56 mg/dL GP1 event rate: 9.8/patient-year p: < 0.05 GP2 event rate: 4.7/patient-year Reported hypoglycemic event GP1◊ 54 (68) p: 0.0013 GP2◊ 33 (42)
Tamemoto, 2007⁴⁷	GP1: Insulin aspart 70/30 (v) Start: 10 - 16 U/day Mean: 26.7 U T: Breakfast, dinner D: 6 months Continued OA agents (NR) T: NR D: 6 months GP2: Insulin glargine (v) Start: 6 - 8 U/day T: NR D: 6 months Continued OA agents (NR) T: NR D: 6 months						From self-monitored blood glucose data, < 70 mg/dL GP1 2 (50) number of events: 11 GP2 4 (57) number of events: 43 Self-reported events GP1 4 (80) GP2 6 (55)
Insulin aspart 70/30 vs. rapid-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin aspart (v) Start: 18 median IU/day Range: 9 – 24 IU/day T: Breakfast, lunch, dinner, D: 1 year Usual care D: 1 year		Grade 2: symptoms and BG < 56 mg/dL GP1 Median number of events/patient-year: 3.9 (IQR 1.0-9.0) p: 0.002 GP2 Median number of events/patient-year: 8.0 (IQR 2.9-17.7)	Grade 3: third party assistance required GP1 11 (4.7) p: overall 0.20 Median number of events/patient-year: 0 p: overall 0.10 GP2 16 (6.7) Median number of events/patient-year: 0			Grades 1, 2, or 3 GP1 216 (91.9) p: 0.08 GP2 229 (96.2)
Insulin aspart 70/30 vs. rapid-acting with long-acting insulin analogues
Joshi, 2005⁵²	GP1: Insulin aspart 70/30 (v) Mean: 40.19 U/day T: bid D: 12 weeks GP2: Insulin aspart (v) Mean: 28.26 U/day at 12 weeks T: tid D: 12 weeks Insulin glargine (v) Mean: 24.52 U/day T: Bedtime D: 12 weeks	BS < 50 mg/dL but self managed GP1 19* (16.7) p: < 0.05 vs GP2 GP2 18* (58.06)		Requiring 3rd party assistance GP1 0 (0) GP2 0 (0)
Insulin aspart 70/30 vs. premixed human insulin
Abra-hamian, 2005⁵³	GP1: Insulin aspart 70/30 (v) Mean: 0.49 (start), 0.61 U/kg (end) T: Breakfast, lunch, dinner D: 24 weeks GP2: NPH/regular 70/30 (v) Mean: 0.46 (start), 0.59 U/kg (end) T: Breakfast, dinner D: 24 weeks	Not defined GP1 number of events: 130 GP2 number of events: 185		Major GP1 number of events: 2 GP2 number of events: 0		Not defined p: NS
Boehm, 2004⁴⁵ Boehm, 2002⁹‡	GP1: Insulin aspart 70/30 (v) Start: 0.57 U/kg T: Breakfast, dinner D: 24 months GP2: NPH/regular 70/30 (v) Start: 0.57 U/Kg T: Breakfast, dinner D: 24 months	GP1 35 (63) number of events: 398 p: 1 GP2 41 (63) number of events: 555		Major hypoglycemia GP1 3 (5) number of events: 3 p: 0.14 GP2 9 (14) number of events: 19
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Requiring third-party assistance GP1 number of events: 2 GP2 number of events: 2			Overall hypoglycemia rates (not specified) GP1 number of events: 23 GP2 number of events: 11
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks GP2: NPH/regular 70/30 (v) Start: 0.16 U/day Mean: 29 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks	Symptoms with BS < 50 mg/dL but not requiring third party assistance GP1 11 (24) GP2 9 (19)		BS < 50 mg/dL with severe CNS symptoms and required third party assistance GP1 0 (0) GP2 0 (0)		Between midnight and 6 am GP1 7 (15) GP2 11 (23)	Symptoms only GP1 13 (28) GP2 11 (23) Any (reported symptoms or BS < 50 mg/dL) GP1 20 (43) p: overall 0.245 GP2 15 (32)
McNally, 2007⁴⁸	GP1: Insulin aspart 70/30 (v) Start: 100 U/mL Mean: 68.8 U Range: 6 - 238.7 U T: Breakfast, dinner D: 16 weeks GP2: NPH/regular 70/30 (v) Start: 100 U/mL Mean: 66.6 U Range: 11.3 - 240 U T: Breakfast, dinner D: 16 weeks	Self reported minor hypoglycemia (patient able to self-treat and blood glucose < 50.4 mg/dL (2.8 mmol/L)) GP1 63* (90) GP2 65* (84)		Patients unable to self-treat GP1 2 (3) number of events: 2 GP2 5 (6) number of events: 7	< 45 mg/dL (2.5 mmol/L) recorded by CGMS between 0600 - 0000 h GP1 29* (41) p: 0.1 GP2 31* (41) < 63 mg/dL (3.5 mmol/L) recorded by CGMS between 0600 - 0000 h GP1 51* (73) p: 0.6 event rate: 2.58/patient-week p: 0.32 GP2 52* (70) event rate: 2.36/patient-week Daytime self-reported rates p: NS	< 45 mg/dL (2.5 mmol/L) recorded by CGMS between 0000 - 0600 h GP1 18* (25) p: 0.039 GP2 28* (37) < 63 mg/dL (3.5 mmol/L) recorded by CGMS between 0000 - 0600 h GP1 36* (51) p: 0.015 event rate: 1.18/patient-week p: 0.011 GP2 50* (66) event rate: 1.62/patient-week Nighttime self-reported rates GP1 event rate: 1.5/patient-year (SD = 4.54) p: 0.002 GP2 event rate: 3.8/patient-year (SD = 8)	< 45 mg/dL (2.5 mmol/L) recorded by CGMS at any time GP1 32* (46) p: 0.28 GP2 40* (54) < 63 mg/dL (3.5 mmol/L) recorded by CGMS at any time GP1 57* (82) p: 1 event rate: 3.76/patient-week p: 0.62 GP2 62* (82) event rate: 3.93/patient-week Total self-reported rates p: NS
McSorley, 2002¹¹	GP1: Insulin aspart 70/30 (NR) T: Breakfast, dinner D: 2 weeks GP2: NPH/regular 70/30 (NR) T: Breakfast, dinner D: 2 weeks			Required third party assistance or injection of glucose or glucagon GP1 0 (0) GP2 0 (0)			Experienced symptoms, but did not require assistance GP1 4 (31) number of events: 7 GP2 3 (23) number of events: 5
Insulin aspart 70/30 vs. intermediate-acting human insulins
Christiansen, 2003¹³	GP1: Insulin aspart 70/30 (v) Start: insulin naive = 8 - 16 U/day; taking NPH prior to trial = started at pretrial dose T: Breakfast, dinner D: 16 weeks GP2: NPH insulin (v) Start: insulin naive = 8 - 16 U/day; taking NPH prior to trial = started at pretrial dose T: Breakfast, dinner D: 16 weeks	Not requiring third party assistance or glucagon injection GP1 77 (38) number of events: 341 RR = 1.21 (95% CI: 0.77 - 1.9) p: 0.4 GP2 68 (34) number of events: 285		Requiring third party assistance or use of glucagon GP1 NR (<2) GP2 NR (<2)		Minor (not requiring assistance) and nocturnal (midnight to 6 am) GP1 22* (10.9) p: NS GP2 22* (11.4) p: NS
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks GP2: NPH insulin (v) Start: 0.16 U/day Mean: 28 U/day T: Bedtime D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 mg - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks	Symptoms with BS < 50 mg/dL but not requiring third party assistance GP1 11 (24) GP2 6 (13)		BS < 50 mg/dL with severe CNS symptoms and required third party assistance GP1 0 (0) GP2 0 (0)		Between midnight and 6 am GP1 7 (15) GP2 11 (23)	Symptoms only GP1 13 (28) GP2 10 (21) Any (reported symptoms or BS < 50 mg/dL) GP1 20 (43) p: overall 0.245 GP2 13 (28)
Insulin aspart 70/30 vs. oral antidiabetic agents
Bebakar, 2007⁴⁶	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Range: 0.16 U/kg (qd) - 0.43 U/kg (bid) T: once or twice daily D: 24 weeks GP2: OA agents (v) T: NR D: 24 weeks	Symptoms and PG < 56 mg/dL and handled by self or PG < 56 mg/dL GP1 number of events: 177 GP2 number of events: 45		Severe CNS symptoms and unable to treat self + PG < 56 mg/dL or reversal of symptoms with treatment GP1 number of events: 1 GP2 number of events: 1			Mild and severe GP1 178 (54) p: < 0.005 GP2 46 (30)
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.51 U/kg T: Breakfast, dinner D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg (end) T: once or twice daily	Symptoms confirmed by BG < 50.4 mg/dL (2.8 mmol/l), handled by patient; asymptomatic BG < 50.4 mg/dL GP1 10 (9) number of events: 20 GP2 9 (8) number of events: 28		Required assistance, BG < 50.4 mg/dL (2.8 mmol/)l, need for food or IV glucose GP1 0 (0) GP2 0 (0)			Total hypoglycemic events (includes minor and symptomatic only) GP1 event rate: 0.037/patient-week GP2 event rate: 0.04/patient-week Symptoms without confirmatory BG GP1 22 (21) number of events: 44 GP2 23 (20) number of events: 43
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.3 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg daily (end) T: once or twice daily D: 16 weeks	Symptoms confirmed by BG < 50.4 mg/dL (2.8 mmol/l), handled by patient; asymptomatic BG < 50.4 mg/dL GP1 13 (12) number of events: 23 GP2 9 (8) number of events: 28		Required assistance, BG < 50.4 mg/dL (2.8 mmol/l), need for food or IV glucose GP1 0 (0) GP2 0 (0)			Symptoms without confirmatory BG GP1 22 (20) number of events: 44 GP2 23 (20) number of events: 43 Total hypoglycemic events (includes minor and symptomatic only) GP1 event rate: 0.039/patient-week GP2 event rate: 0.04/patient-week
Raz, 2003⁵⁷	GP1: Insulin aspart 70/30 (v) Start: 6-8 U bid T: Breakfast, dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks GP2: Glibenclamide (fix) Range: 7.5 – 15 mg T: Dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks	BG < 50 mg/dL handled by self GP1 event rate: 1.8/year p: 0.03 GP2 event rate: 0/year					Minor episodes with symptoms but no blood sugars GP1 event rate: 5.3/year p: <0.01 vs. GP2 GP2 event rate: 0/year
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.7 U/kg T: Breakfast, dinner D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	BG < 50 mg/dL but did not require third party assistance GP1 15 (15) number of events: 47 GP2 3 (3) number of events: 3		BG < 50 mg/dL or requiring third party assistance GP1 0 (0) GP2 0 (0)		Midnight to 6 am GP1 number of events: 8 GP2 number of events: 0	All hypoglycemic episodes - symptoms or BG < 50 mg/dL GP1 event rate: 0.132/patient-week GP2 event rate: 0.032/patient-week
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.5 U/kg T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	BG < 50 mg/dL but did not require third party assistance GP1 11 (12) number of events: 15 GP2 3 (3) number of events: 3		BG < 50 mg/dL or requiring third party assistance GP1 0 (0) GP2 0 (0)		Midnight to 6 am GP1 number of events: 0 GP2 number of events: 0	All hypoglycemic episodes - symptoms or BG < 50 mg/dL GP1 event rate: 0.083/patient-week GP2 event rate: 0.032/patient-week
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Continuation of OA agents (v) T: NR D: 16 weeks	BG < 55.8 mg/dL (3.1 mmol/L), with or without symptoms, and handled by patient GP1 4 (4) GP2 1 (1)		BG < 55.8 mg/dL (3.1 mmol/L) and required 3rd party help or symptoms reversed after intake of food, glucagon, or IV glucose GP1 0 (0) GP2 0 (0)			Symptoms only GP1 28 (27.5) GP2 4 (3.8) Overall hypoglycemia GP1 Event rate: 0.73/ person-year GP2 Event rate: 0.69/person-year
Insulin aspart 70/30 vs. exenatide
Nauck, 2007⁴⁹	GP1: Insulin aspart 70/30 (v) Start: 15.7 U/day Mean: 24.4 U/day T: Breakfast, dinner D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks GP2: exenatide (v) Start: 5 µg bid Range: 5 - 10 µg bid T: Breakfast D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks			Severe, not further defined GP1 0 (0) GP2 0 (0)	Not further defined GP1 Event rate: 4.4/patient-year p: NS GP2 Event rate: 4.1/patient-year	Nocturnal, not further defined GP1 25 (62) event rate: 1.1/patient-year p: NS GP2 44 (17) Event rate: 0.6/patient-year	Symptoms or PG < 61.2 mg/dL (3.4 mmol/L) GP1 Event rate: 5.6/patient-year p: NS GP2 Event rate: 4.7/patient-year
Insulin aspart 70/30 vs. insulin lispro 75/25
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Requiring third-party assistance GP1 number of events: 2 GP2 number of events: 5			Overall hypoglycemia rates (not specified) GP1 number of events: 23 GP2 number of events: 19
Niskanen, 2004⁵⁵	GP1: Insulin aspart 70/30 (v) Mean: 0.65 to 0.67 U/kg T: Breakfast, dinner D: 12 weeks GP2: Insulin lispro 75/25 (v) Mean: 0.67 to 0.71 U/kg T: Breakfast, dinner D: 12 weeks	BG < 50.4 mg/dL (2.8 mmol/L) with or without symptoms or symptoms not confirmed by BG reading GP1 57 (43) number of events: 269 GP2 53 (40) number of events: 233		Required third party assistance GP1 1 (1) GP2 1 (1)
Insulin aspart 70/30 vs. insulin aspart 70/30 + oral antidiabetic agents
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.51 U/kg T: Breakfast, dinner D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.3 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks	Symptoms confirmed by BG < 50.4 mg/dL (2.8 mmol/l), handled by patient; asymptomatic BG < 50.4 mg/dL GP1 10 (9) number of events: 20 GP2 13 (12) number of events: 23		Required assistance, BG < 50.4 mg/dL (2.8 mmol/l), need for food or IV glucose GP1 0 (0) GP2 0 (0)			Symptoms without confirmatory BG GP1 22 (21) number of events: 44 GP2 22 (20) number of events: 44 Total hypoglycemic events (includes minor and symptomatic only) GP1 Event rate: 0.037/patient-week GP2 Event rate: 0.039/patient-week
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg Mean: 0.7 U/kg T: Breakfast, dinner D: 18 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg Mean: 0.5 U/kg T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	BG < 50 mg/dL but did not require third party assistance GP1 15 (15) number of events: 47 GP2 11 (12) number of events: 15		BG < 50 mg/dL or requiring third party assistance GP1 0 (0) GP2 0 (0)		Midnight to 6 am GP1 number of events: 8 GP2 number of events: 0	All hypoglycemic episodes - symptoms or BG < 50 mg/dL GP1 Event rate: 0.132/patient-week GP2 Event rate: 0.083/patient-week
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 44.8 U T: Breakfast, dinner D: 16 weeks Metformin (v) Start: 500 mg qd or bid or 850 mg qd T: NR D: 14 weeks (started after 2 weeks)	BG < 55.8 mg/dL (3.1 mmol/L), with or without symptoms, and handled by patient GP1 4 (4) GP2 9 (9)		BG < 55.8 mg/dL (3.1 mmol/L) and required 3rd party help or symptoms reversed after intake of food, glucagon, or IV glucose GP1 0 (0) GP2 0 (0)			Symptoms only GP1 28 (27.5) p: NS vs. GP2 GP2 28 (28) Overall hypoglycemia GP1 event rate: 0.73/ person-year p: NS vs. GP2 GP2 event rate: 0.69/ person-year
Insulin aspart 70/30 vs. rapid-acting insulin analogues with intermediate-acting human insulin
Hirao, 2008⁶¹	GP1: Insulin aspart 70/30 (NR) T: bid D: 6 months GP2: Insulin aspart (NR) T: tid D: 6 months NPH insulin (NR) T: Optional multiple daily injections D: 6 months			Major hypoglycemia, not further defined GP1 0 (0) GP2 0 (0)
Insulin lispro 75/25 vs. long-acting insulin analogues
Cox, 2007⁷⁴	GP1: Insulin lispro 75/25 (v) T: Breakfast, dinner D: 12 weeks Metformin (NR) T: NR D: 12 weeks GP2: Insulin glargine (v) T: Bedtime D: 12 weeks Metformin (NR) T: NR D: 12 weeks			Severe, not defined GP1 0 (0) GP2 0 (0)			Symptoms or BG < 63 md/dL (3.5 mmol/L) GP1 p: NS
Jacober, 2006⁶⁴	GP1: Insulin lispro 50/50 (v) Mean: 0.353 IU/kg; 36.73 IU T: Breakfast, lunch D: 4 months Insulin lispro 75/25 (v) T: Dinner D: 4 months Existing oral therapy (NR) T: NR D: 4 months GP2: Insulin glargine (v) Mean: 0.276 IU/kg; 27.98 IU T: Bedtime D: 4 months Existing oral therapy (NR) T: NR D: 4 months			Self reported GP1 0 (0) GP2 0 (0)		Self reported symptoms or PG ≤ 72 mg/dL GP1 Event rate: 0.8/patient/30 days (SD: 2.12) p: 0.3604 GP2 Event rate: 1.05/patient/30 days (SD: 1.59)	Self reported symptoms or PG ≤ 72 mg/dL GP1 42* (72.2) p: 0.033 Event rate: 3.98/patient/30 days (SD: 4.74) p: 0.0013 GP2 56* (94.8) Event rate: 2.57/patient/30 days (SD: 3.22)
Malone, 2004⁶⁵	GP1: Insulin lispro 75/25 (v) Mean: 0.62 U/kg T: Breakfast, dinner D: 16 weeks Metformin (NR) Mean: 1945 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.57 U/kg T: Bedtime D: 16 weeks Metformin (NR) Mean: 1997 mg Range: 1500 - 2550 mg T: NR D: 16 weeks			Requiring third-party assistance due to disabling hypoglycemia GP1 0 (0) GP2 0 (0)		BG < 63 mg/dL or symptoms occurring between bedtime and before breakfast GP1 30 (30) number of events: 39 GP2 28 (28) number of events: 63	BG < 63 mg/dL or symptoms GP1 57 (57) number of events: 181 Event rate: 0.68/patient/30 days (SD: 1.38) p: 0.041 GP2 40 (40) number of events: 87 Event rate: 0.39/patient/30 days (SD: 1.24)
Malone, 2005⁶⁶	GP1: Insulin lispro 75/25 (v) Mean: 0.42 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 2128 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.36 U/kg T: Bedtime D: 16 weeks Metformin (fix) Mean: 2146 mg Range: 1500 - 2550 mg T: NR D: 16 weeks			Not defined GP1 0 (0) GP2 0 (0)	BG < 63 mg/dL (3.5 mmol/L) or symptoms GP1 Event rate: 0.46/patient/30 days (SD: 1.28) p: 0.003 GP2 Event rate: 0.1/patient/30 days (SD: 0.51)	BG < 63 mg/dL (3.5 mmol/L) or symptoms occurring between bedtime and breakfast for the patient GP1 Event rate: 0.14//patient/30 days (SD: 0.49) p: 0.002 GP2 Event rate: 0.34/patient/30 days (SD: 0.85)	Overall rate of BG < 63 mg/dL (3.5 mmol/L) or symptoms GP1 Event rate: 0.61/patient/30 days (SD: 1.41) p: 0.477 GP2 Event rate: 0.44/patient/30 days (SD: 1.07)
Roach, 2006⁶³	GP1: Insulin lispro 75/25 (v) Mean: 23 U (morning) and 37 U (evening) Range: 0 – 72 U (morning); 11 – 88 U (evening) T: Breakfast, dinner D: 12 weeks OA agents (NR) Start: current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks GP2: Insulin glargine (v) Mean: 44 U Range: 14 U - 100 U T: Breakfast D: 12 weeks OA agents (NR) Start: current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks	Self reported BG < 63 mg/dL (3.5 mmol/L) or symptoms GP1 3 (15) GP2 2 (10)		Not defined GP1 0 (0) GP2 0 (0)	PG < 63 mg/dL (3.5 mmol/L) GP1 0 (0) GP2 1 (5)	PG < 63 mg/dL (3.5 mmol/L) GP1 8 (40) GP2 2 (10)	PG < 63 mg/dL (3.5 mmol/L) GP1 8 (40) GP2 3 (15)
Insulin lispro 75/25 vs. premixed human insulins
Coscelli, 2003⁶⁷	GP1: Insulin lispro 75/25 (v) Mean: 38.1 Range: 12 - 72 T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days GP2: NPH/regular 70/30 (v) Mean: 37.3 Range: 10 - 72 T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days						Not defined GP1 p: NS vs. GP2
Hermansen, 2002⁵⁸	GP1: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Requiring third-party assistance GP1 number of events: 5 GP2 number of events: 2			Overall hypoglycemia (not specified) GP1 number of events: 19 GP2 number of events: 11
Herz, 2002⁷¹	GP1: Insulin lispro 75/25 (v) Mean: 26.1 U T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 26.2 U T: Breakfast, dinner D: 4 weeks						Symptoms or BG < 54 mg/dL (3.0 mmol/L)¶ GP1 Event rate: 0.7/patient/30 days (SE = 0.2) p: 0.042 GP2 Event rate: 1.2/patient/30 days (SE = 0.3) Symptoms or BG < 54 mg/dL (3.0 mmol/L)§ GP1 Event rate: 0.9/patient/30 days (SE = 0.2) p: 0.569 GP2 Event rate: 0.9/patient/30 days (SE = 0.1)
Herz, 2003¹²	GP1: Insulin lispro 75/25 (v) Mean: 31.6 (morning)¶ and 26.8 U (evening)¶ and 32.4 (morning)§ and 27.6 U (evening)§ T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 32.3 (morning)¶ and 26.4 U (evening)¶ and 33.3 (morning)§ and 27.5 U (evening)§ T: Breakfast, dinner D: 4 weeks						Symptoms or any spontaneous BG < 54 mg/dL (3.0 mmol/L)¶ GP1 Event rate: 0.049/patient/30 days (SE = 0.018) p: 0.586 GP2 Event rate: 0.1/patient/30 days (SE = 0.018) Symptoms or any spontaneous BG < 54 mg/dL (3.0 mmol/L)§ GP1 Event rate: 0.241/patient/30 days (SE = 0.053) p: 0.524 GP2 Event rate: 0.222/patient/30 days (SE = 0.053)
Malone, 2000⁴⁴	GP1: Insulin lispro 75/25 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days GP2: NPH/regular 70/30 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days						BG < 63 mg/dL (3.5 mmol/L) or symptoms occurring between lunch and dinner GP1 number of events: 3 GP2 number of events: 5 BG < 63 mg/dL (3.5 mmol/L) or symptoms GP1 number of events: 7 GP2 number of events: 10 BG < 63 mg/dL (3.5 mmol/L) or symptoms occurring within 4 hours of test meal GP1 number of events: 5 GP2 number of events: 8
Mattoo, 2003⁷⁰	GP1: Insulin lispro 75/25 (NR) Mean: 20 (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks GP2: NPH/regular 70/30 (NR) Mean: 20 (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks						Symptoms or BG < 63 mg/dL (3.5 mmol/L) GP1 event rate: 0.4/patient/14 days (SD = 0.9) p: 0.725 GP2 event rate: 0.4/patient/14 days (SD = 0.8)
Roach, 1999⁷³	GP1: Insulin lispro 75/25 (v) Mean: 0.37 (morning), 0.28 (evening) U/kg T: Breakfast, dinner D: 13 weeks GP2: NPH/regular 70/30 (v) Mean: 0.36 (morning), 0.27 (evening) U/kg T: Breakfast, dinner D: 13 weeks			Required third party assistance GP1 1 (1) GP2 1 (1)		Symptoms or BG < 54 mg/dL (3.0 mmol/L) occurring between mean reported bedtime and mean reported breakfast time for each country GP1 13 (15) p: 0.266 GP2 8 (9)	Symptoms or BG < 54 mg/dL (3.0 mmol/L) GP1 34* (42) p: 0.398 GP2 28* (35)
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day	GP1 0 (0) GP2 1 (5)
Insulin lispro 75/25 vs. oral antidiabetic agents
Herz, 2002⁷²	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg Mean: 0.46 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glyburide (fix) Start: 15 mg/day T: Breakfast, dinner D: 16 weeks			Requiring assistance of third party GP1 0 (0) GP2 0 (0)			Any (BG < 54 mg/dL (3 mmol/L) or symptoms) GP1 B: 0.14 episodes/ patient/30 days (SE 0.14) p: 0.361 vs GP2 F: 0.31 episodes/ patient/30 days (SE 0.21) p: 0.028 vs GP2 F-B: 0.17 episodes/ patient/30 days (SE 0.02) p: 0.077 vs GP2
Malone, 2003⁶⁸	GP1: Insulin lispro 75/25 (v) Mean: 0.19 U/kg (morning) and 0.14 U/kg (evening) T: Breakfast, dinner D: 16 weeks Metformin (v) Mean: 1813 mg/day Range: 1500 – 2550 mg/day T: 2 to 3 times/day D: 16 weeks GP2: Metformin (v) Mean: 1968 mg/day Range: 1500 - 2550 mg/day T: 2 to 3 times/day D: 16 weeks Glibenclamide (v) Mean: 14.2 mg/day T: NR D: 16 weeks			Unable to treat self or BG < 36 mg/dL (2.0 mmol/L) (events/patient/30 days) GP1 B: 0.01 Median (0.09) F: 0.01 Median (0.11) F-B: 0* (1) GP2 B: 0 Median (0) F: 0.02 Median (0.15) F-B: 0* (1.3) GP1-GP2: 0*		Symptoms or BG < 63 mg/dL (3.5 mmol/L) occurring after bedtime (events/patient/30 days) GP1 B: 0.03 (0.23) F: 0.01 ( 0.11) F-B: 0* GP2 B: 0 (0) F: 0.08 ( 0.4) F-B: 0* GP1-GP2: 0* Symptoms or BG < 63 mg/dL (3.5 mmol/L) occurring after bedtime GP1 (1) GP2 (5)	Overall events/ patient/30 days GP1 B: 0.08 (0.59) F: 0.31 (1.07) F-B: 0* GP2 B: 0.07 (0.57) F: 0.48 (1.17) F-B: 0* GP1-GP2: 0*
Tirgoviste, 2003⁴³	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glibenclamide (v) Start: 15 mg T: Breakfast, dinner D: 16 weeks						Symptoms and/or BG < 54 mg/dL (3.0 mmol/L) GP1 38 (44.7) p: 0.001 GP2 9 (10.3)
Insulin lispro 75/25 vs. insulin lispro 50/50
Roach, 2003⁶⁹	GP1: Insulin lispro 75/25 (v) Mean: 31.3 (morning), 27.6 U (evening) T: Breakfast, dinner D: 8 weeks GP2: Insulin lispro 50/50 (v) Mean: 31.5 U T: Breakfast D: 8 weeks Insulin lispro 75/25 (v) Mean: 27.9 U T: Dinner D: 8 weeks			Required third party assistance GP1 0* (0) GP2 0* (0)			Symptoms GP1 28* (26.1) p: 0.078 number of events: 65 p: 0.681 GP2 34* (32.4) number of events: 68
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day	GP1 0 (0) GP2 0 (0)
Insulin lispro 50/50 vs. long-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg mean Mean: 0.59 IU/kg T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin glargine (v) Start: 0.16 IU/kg mean Mean: 0.43 IU/kg T: Bedtime D: 24 weeks			Not defined GP1 0 (0) GP2 0 (0)			Symptoms or PG < 54 mg/dL (3.0 mmol/L) GP1 24* (44.4) Event rate: 1.5/ 100 patient-days GP2 17* (32.1) Event rate: 1/ 100 patient-days
Robbins, 2007⁷⁹	GP1: Insulin lispro 50/50 (v) Mean: 0.7 U/kg T: Breakfast, lunch, dinner D: 24 weeks Metformin (fix) Mean : 1641 mg T: bid D: 24 weeks GP2: Insulin glargine (v) Mean : 0.6 U/kg T: Bedtime D: 24 weeks Metformin (fix) Mean: 1636 mg T: bid D: 24 weeks			Required 3rd party assistance and BG < 50 mg/dL (2.9 mmol/L) or prompt recovery GP1 3 (1.9) number of events: 8 p: NS GP2 2 (1.3) number of events: 4		Episodes occurring after bedtime and before awakening GP1 Event rate: 0.2/30 patient-days (SD = 0.7) p: 0.30 vs GP2 GP2 Event rate: 0.3/30 patient-days (SD = 0.6) s	Overall (signs or symptoms or BG < 63 mg/dL (3.5 mmol/L)) GP1 81 (51.9) Event rate: 0.8/30 patient-days (SD = 1.4) p: 0.07 for event rate vs GP2 GP2 77 (48.4) Event rate: 0.5/30 patient-days (SD = 1.0)
Insulin lispro 50/50 vs. rapid-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg mean Mean: 0.59 IU/kg T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin lispro (v) Start: 0.25 IU/kg mean Mean: 0.50 IU/kg T: Breakfast, lunch, dinner D: 24 weeks			Not defined GP1 0 (0) GP2 0 (0)			Symptoms or PG < 54 mg/dL (3.0 mmol/L) GP1 24* (44.4) Event rate: 1.5/ 100 patient-days GP2 28* (53.8) Event rate: 1.4/100 patient-days
Insulin lispro 50/50 vs. rapid-acting with long-acting insulin analogues
Rosen-stock, 2008⁸⁰	GP1: Insulin lispro 50/50 (v) Start: Insulin glargine dose at entry (52.5 U) Mean: 123 U T: Breakfast, lunch, dinner D: 24 weeks Insulin lispro 75/25 (v) Start: allowed to switch evening dose to insulin lispro 75/25 T: Dinner D: Unclear GP2: Insulin glargine (v) Start: 50% of insulin glargine dose at entry (54.9 U) Mean: 70 U T: Bedtime D: 24 weeks Insulin lispro (v) Start: 50% of insulin glargine dose at entry divided in 3 equal doses (54.9 U) Mean: 76 U T: Breakfast, lunch, dinner D: 24 weeks			Required 3rd party assistance GP1 6 (3.21) p: 0.751 Event rate: 0.1/patient-year (SD = 0.65) p: 0.266 GP2 4 (2.14) Event rate: 0.05/patient-year (SD = 0.31)		Nocturnal (not further defined) GP1 109 (58.29) p: 1 Event rate: 4.78/ patient-year (SD = 7.15) p: 0.139 GP2 110 (58.82) Event rate: 6.17/ patient-year (SD = 10.68)	BG < 72 mg/dL (4.0 mmol/L) GP1 165 (88.24) p: 1 Event rate: 46.5/ patient-year (SD = 48) p: 0.747 GP2 165 (88.24) Event rate: 44.95/ patient-year (SD = 46.8) BG < 60 mg/dL (3.3 mmol/L) GP1 148 (79.14) p: 0.898 Event rate: 20.75/ patient-year (SD = 26.86) p: 0.574 GP2 150 (80.21) Event rate: 19.26/ patient-year (SD = 24.51) BG < 50 mg/dL (2.8 mmol/L) GP1 104 (55.61) p: 0.294 Event rate: 7.34/ patient-year (SD = 12.88) p: 0.23 GP2 115 (61.5) Event rate: 5.93/ patient-year (SD = 9.92)
Insulin lispro 50/50 vs. premixed human insulin
Roach, 1999¹⁰	GP1: Insulin lispro 50/50 (v) Mean: 0.31 U/kg T: Breakfast D: 3 months Insulin lispro 75/25 (v) Mean: 0.26 U/kg T: Dinner D: 3 months GP2: NPH/regular 50/50 (v) Mean: 0.32 U/kg T: Breakfast D: 3 months NPH/regular 70/30 (v) Mean: 0.26 U/kg T: Dinner D: 3 months			Occurrence of coma or requirement for intravenous glucose, glucagon, or both GP1 0 (0) GP2 0 (0)		Symptoms or BG < 54 mg/dL (3.0 mmol/L) occurring between median bedtime (10:30pm) and median breakfast (7:45am) GP1 Event rate: 0.3/3 patient-months (SD: 1.0) p: 0.199 GP2 Event rate: 0.6/3 patient-months (SD: 1.4)	Symptoms or BG < 54 mg/dL (3.0 mmol/L) GP1 25* (40) p: NS GP2 23* (37)
Schernthaner, 2004⁷⁷	GP1: Insulin lispro 50/50 (v) Mean: 64.6 IU T: Breakfast, lunch, dinner D: 12 weeks Diet/exercise D: 12 weeks GP2: NPH/regular 70/30 (v) Mean: 61.8 IU T: Breakfast, dinner D: 12 weeks Diet/exercise D: 12 weeks			BG < 36 mg/dL, coma, or treatment with glucagon or intravenous glucose GP1 0 GP2 1			BG < 65 mg/dL or symptoms GP1 14 (41.2) p: NS GP2 10 (29.4)
Schwartz, 2006⁶²	GP1: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day	GP1 0 (0) GP2 1 (5)
Yamada, 2007⁷⁸	GP1: Insulin lispro 50/50 (v) Start: current dose Mean: 0.37 (start), 0.38 U/kg (end) T: twice daily D: 4 months GP2: NPH/regular 70/30 (v) Start: current dose Mean: 0.34 (start), 0.37 U/kg (end) T: twice daily D: 4 months NPH/regular 50/50 (v) Start: current dose Mean: 0.34 U/kg (start), 0.37 U/kg (end) T: twice daily D: 4 months			Requiring third party assistance GP1 0 (0) GP2 0 (0)

Evidence Table 6. Comparative effectiveness of premixed insulin analogues and other diabetes treatments on adverse events
Author, year	Intervention	Weight in kg, mean (SD)	Injection site reaction, n (%)	Total serious adverse events, n (%)	Withdrawn due to adverse events, n (%)	Other serious adverse events, n (%)
* Number has been imputed. ◊ Among those who were not using thiazolidinediones. ‡ One-hundred and four (36%) of the 291 participants of this trial are patients with type 1 diabetes. The remaining population has type 2 diabetes and is the same study population as Boehm 2004.⁴⁵ Only data for the Boehm 2004 study is presented because it has the longest followup. ¶ Dosing during the outpatient phase. § Dosing during the inpatient phase. µg = microgram; B = baseline; bid = twice daily; BMI = body mass index; Cl = confidence interval; D = duration; F = final; F-B = mean difference from baseline; GP = group; GP1-GP2 = mean difference between the difference from baseline; IU = international unit; kg = kilogram; kg/m² = kilogram per square meter; mg = milligram; ml = milliliter; NA = not applicable; NPH = neutral protamine Hagedorn; NR = not reported; NS = not significant; OA = oral antidiabetic; p = p-value; qd = once daily; SD = standard deviation; SE = standard error; T = time of day when insulin taken; tid = thrice daily; U = unit; U/l = units per liter; v = dosing varied
Insulin aspart 70/30 vs. long-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin detemir (v) Start: 16 median IU/day Range: 10 – 24 IU/day T: Bedtime, twice if required D: 1 year Usual care D: 1 year	GP1 F-B: 4.7 (4) p: < 0.001 GP2 F-B: 1.9 (4.2) GP1-GP2: 3*		GP1 41 (17.4) p: overall 0.25 GP2 30 (12.8)	GP1 2 (1) GP2 4 (2)	Gastrointestinal and abdominal pain GP1 3 (1.3) p: overall 0.21 GP2 2 (0.9) Lower respiratory tract and lung infection GP1 4 (1.7) p: overall 0.02 GP2 0 (0)
Kann, 2006⁵⁰	GP1: Insulin aspart 70/30 (v) Start: 0.1 U/kg bid Mean: 0.4 U/kg T: Breakfast, dinner D: 26 weeks Metformin (v) Start: 500 mg bid or current dose T: Breakfast, dinner D: 26 weeks GP2: Insulin glargine (v) Start: 0.2 U/kg qd Mean: 0.39 U/kg T: Preferred time (constant through study) D: 26 weeks Glimepiride (v) Start: 1 mg daily or current dose T: Breakfast D: 26 weeks	GP1 B: 84* F: 84.8 (17.2) F-B: 0.7 p: NS vs. baseline GP2 B: 86* F: 88.1 (14.6) F-B: 1.5 (95% CI: 0.84 – 2.19) p: < 0.0001 vs. baseline GP1-GP2: -1*		GP1 10 (7.8) GP2 11 (8.7)	GP1 5 (4) GP2 2 (2)
Raskin, 2005³⁹ Raskin, 2007⁴⁰	GP1: Insulin aspart 70/30 (v) Start: 10 or 12 U T: Breakfast, dinner D: Unclear Metformin (v) Range: 1500 – 2550 mg/day T: NR D: Unclear GP2: Insulin glargine (v) Start: 10-12 U/day T: Bedtime D: Unclear Metformin (v) Range: 1500 - 2550 mg/day T: NR D: Unclear	GP1 F-B: 5.4 (4.8) p: < 0.01 GP2 F-B: 3.5 (4.5) GP1-GP2: 1* GP1◊ F-B: 5.6 (4.6) p: 0.0004 GP2◊ F-B: 3 (4.3) GP1-GP2: 3*		GP1◊ 4 (5) GP2◊ 5 (6)	GP1 4 (3) GP2 1 (1) GP1◊ 3 (4) GP2◊ 0 (0)
Tamemoto, 2007⁴⁷	GP1: Insulin aspart 70/30 (v) Start: 10 - 16 U/day Mean: 26.7 U T: Breakfast, dinner D: 6 months Continued OA agents (NR) T: NR D: 6 months GP2: Insulin glargine (v) Start: 6 - 8 U/day T: NR D: 6 (expected) months Continued OA agents (NR) T: NR D: 6 months	GP1 F-B: 0.42 p: NS GP2 F-B: 0.51 GP1-GP2: -1*			GP1 0 (0) GP2 0 (0)
Insulin aspart 70/30 vs. rapid-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin aspart (v) Start: 18 median IU/day Range: 9 – 24 IU/day T: Breakfast, lunch, dinner D: 1 year Usual care D: 1 year	GP1 F-B: 4.7 (4) p: 0.005 vs. GP2 GP2 F-B: 5.7 (4.6) GP1-GP2: -1*		GP1 41 (17.4) p: overall 0.25 GP2 30 (12.6)	GP1 2 (1) GP2 0 (0)	Gastrointestinal and abdominal pain GP1 3 (1.3) p: overall 0.21 GP2 0 (0) Lower respiratory tract and lung infection GP1 4 (1.7) p: overall 0.02 GP2 0 (0)
Insulin aspart 70/30 vs. rapid-acting with long-acting insulin analogues
Joshi, 2005⁵²	GP1: Insulin aspart 70/30 (v) Mean: 40.19 U/day T: bid D: 12 weeks GP2: Insulin aspart (v) Mean: 28.26 U/day T: before every meal D: 12 weeks Insulin glargine (v) Mean: 24.52 U/day T: Bedtime D: 12 weeks	GP1 B: 70.4 (12.18) F: 70.61 (11.23) F-B: 1* p: NS vs. baseline GP2 B: 69.63 (10.31) F: 69.68 (9.58) F-B: 0* p: NS vs. baseline GP1-GP2: 1*		GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 0 (0)
Insulin aspart 70/30 vs. premixed human insulins
Abrahamian, 2005⁵³	GP1: Insulin aspart 70/30 (v) Mean: 0.49 (start), 0.61 U/kg (end) T: Breakfast, lunch, dinner D: 24 weeks GP2: NPH/regular 70/30 (v) Mean: 0.46 (start), 0.59 U/kg (end) T: Breakfast, dinner D: 24 weeks			GP1 number of events: 16 GP2 number of events: 15	GP1 3 (3) GP2 0 (0)
Boehm, 2004⁴⁵ Boehm, 20029‡	GP1: Insulin aspart 70/30 (v) Start: 0.57 U/kg T: Breakfast, dinner D: 24 months GP2: NPH/regular 70/30 (v) Start: 0.57 U/kg T: Breakfast, dinner D: 24 months	GP1 F-B: 0.05 (SE 0.81) p: 0.07 vs. GP2 GP2 F-B: 2 (SE 0.69) GP1-GP2: -2*			GP1 5 (6) GP2 6 (6)
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day				GP1 1 (2) GP2 0 (0)	GP1 number of events: 1 GP2 number of events: 0
Kapitza, 2004⁵⁶	GP1: Insulin aspart 70/30 (NA) T: Breakfast D: 1 day GP2: NPH/regular 70/30 (NA) T: Breakfast D: 1 day				GP1 0 (0) GP2 0 (0)
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 Weeks run-in, then 12 weeks GP2: NPH/regular 70/30 (v) Start: 0.16 U/day Mean: 29 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 Weeks run-in, then 12 weeks	GP1 F-B: 0.7 p: 0.251 vs. GP2 GP2 F-B: 1 GP1-GP2: 0*			GP1 2 (4) GP2 0 (0)	Blurred vision and pain in the extremities GP1 1 (2) GP2 0 (0)
McNally, 2007⁴⁸	GP1: Insulin aspart 70/30 (v) Start: 100 U/mL Mean: 68.8 U Range: 6 - 238.7 T: Breakfast, dinner D: 16 weeks GP2: NPH/regular 70/30 (v) Start: 100 U/mL Mean: 66.6 U Range: 11.3 - 240 U T: Breakfast, dinner D: 16 weeks			Resulted in death, was life-threatening or caused (or prolonged) hospitalization GP1 3* (4) GP2 5* (6)	GP1 2 (1) GP2 1 (1)
McSorley, 2002¹¹	GP1: Insulin aspart 70/30 (NR) T: Breakfast, dinner D: 2 weeks GP2: NPH/regular 70/30 (NR) T: Breakfast, dinner D: 2 weeks				GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 0 (0)
Insulin aspart 70/30 vs. intermediate-acting human insulin
Christiansen, 2003¹³	GP1: Insulin aspart 70/30 (v) Start: insulin naïve: 8 - 16 U/day; taking NPH prior to trial: pretrial dose T: Breakfast, dinner D: 16 weeks GP2: NPH insulin (v) Start: insulin naïve: 8 - 16 U/day; taking NPH prior to trial: pretrial dose T: Breakfast, dinner D: 16 weeks	GP1 1 (0) GP2 1 (0)		GP1 5 (2) number of events: 5 GP2 7 (3) number of events: 8	GP1 2 (1) GP2 2 (1)	Allergic reaction to protamine GP1 1 (0) GP2 0 (0)
Kilo, 2003¹⁵	GP1: Insulin aspart 70/30 (v) Start: 0.16 U/day Mean: 26 U/day T: Dinner D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks GP2: NPH insulin (v) Start: 0.16 U/day Mean: 28 U/day T: Bedtime D: 12 weeks Metformin (fix) Mean: about 2200 mg Range: 500 - 2550 mg T: 1-3 times/day D: 4 weeks run-in, then 12 weeks	GP1 F-B: 0.7 p: 0.251 overall GP2 F-B: 0.1 GP1-GP2: 1*			GP1 2 (4) GP2 0 (0)	Blurred vision and pain in the extremities GP1 1 (2) GP2 0 (0)
Insulin aspart 70/30 vs. oral antidiabetic agents
Bebakar, 2007⁴⁶	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Range: 0.16 U/kg (qd) - 0.43 U/kg (bid) T: Once or twice daily D: 24 weeks GP2: OA agents (v) T: NR D: 24 weeks	GP1 F-B: 0.98 p: < 0.005 vs. GP2 GP2 F-B: 0 GP1-GP2: 1*		GP1 number of events: 5 GP2 number of events: 0	GP1 6 (5) GP2 0 (0)
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.51 U/kg/day T: Breakfast, dinner D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg (end) T: once or twice daily D: 16 weeks	GP1 F-B: 1.6 GP2 F-B: 0.1 GP1-GP2: 1.46 (SE 00.41) p: < 0.001		GP1 total events for all groups: 5 GP2 total events for all groups: 5	GP1 1 (1) GP2 0 (0)
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.3 U/kg/day T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg (end) T: once or twice daily D: 16 weeks	GP1 F-B: 0.8 GP2 F-B: 0.1 GP1-GP2: 0.66 (SE 0.41) p: NS		GP1 total events for all groups: 5 GP2 total events for all groups: 5	GP1 2 (2) GP2 0 (0)
Raz, 2003⁵⁷	GP1: Insulin aspart 70/30 (v) Start: 6-8 U bid T: Breakfast, dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks GP2: Glibenclamide (fix) Range: 7.5 – 15 mg T: Dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks	GP1 F-B: 0.23 p: NS vs. GP2 GP2 F-B: 0.03 GP1-GP2: 0*			GP1 0 (0) GP2 0 (0)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.7 U/kg/day T: Breakfast, dinner D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 F-B: 2.2 GP2 F-B: 2.2 GP1-GP2: 0* Experienced weight gain GP1 3* (3) p: < 0.05 overall GP2 2* (2)		GP1 2 (2) GP2 0 (0)	GP1 3 (3) GP2 2 (2)	Cellulitis GP1 1 (1) GP2 0 (0) Peripheral edema GP1 0* (0) GP2 1* (1)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.5 U/kg/day T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 F-B: 4 GP2 F-B: 2.2 GP1-GP2: 2* Experienced weight gain GP1 7* (8) p: < 0.05 overall GP2 2* (2)		GP1 0 (0) GP2 0 (0)	GP1 1 (1) GP2 2 (2)	Cellulitis GP1 0 (0) GP2 0 (0) Peripheral edema GP1 6* (6) GP2 1* (1)
Raskin, 2007⁶⁰	GP1: Insulin aspart 70/30 (v) Start: 6 U bid Mean: 0.6 U/kg/day T: Breakfast, dinner D: 34 weeks Metformin (fixed) Mean: 2446 mg T: NR D: 34 weeks Pioglitazone (fixed) Mean: 32.5 mg T: NR D: 34 weeks GP2: Metformin (fixed) Mean: 2439 mg T: NR D: Unclear Pioglitazone (fixed) Mean: 31.7 T: NR D: Unclear				GP1 3 (2.9) GP2 4 (4.1)	Peripheral edema GP1 8* (9) GP2 11* (12)
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Continuation of OA agents (v) T: NR D: 16 weeks			GP1 0 (0) GP2 0 (0)	GP1 1 (1) GP2 1 (1)
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 44.8 U T: Breakfast, dinner D: 16 weeks Metformin (v) Start: 500 mg qd or bid or 850 mg qd T: NR D: 14 weeks (started after 2 weeks) GP2: Continuation of OA agents (v) T: NR D: 16 weeks			GP1 0 (0) GP2 0 (0)	GP1 1 (1) GP2 1 (1)
Insulin aspart 70/30 vs. exenatide
Nauck, 2007⁴⁹	GP1: Insulin aspart 70/30 (v) Start: 15.7 U/day Mean: 24.4 U/day T: Breakfast, dinner D: 52 weeks 'Optimally' effective metformin and sulfonylurea (v) T: NR D: 52 weeks GP2: Exenatide (v) Start: 5 µg bid Range: 5 - 10 µg bid T: Breakfast D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks	GP1 F-B: 2.9 p: < 0.001 GP2 F-B: -2.5 p: < 0.001 GP1-GP2: 5.4 (95% CI: 5 – 5.9) p: < 0.001		GP1 11 (4.4) GP2 19 (7.5)	GP1 0 (0) GP2 20 (8)
Insulin aspart 70/30 vs. insulin lispro 75/25
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day				GP1 1 (2) GP2 0 (0)	GP1 number of events: 1 GP2 number of events: 0
Niskanen, 2004⁵⁵	GP1: Insulin aspart 70/30 (v) Mean: 0.65 to 0.67 U/kg T: Breakfast, dinner D: 12 weeks GP2: Insulin lispro 75/25 (v) Mean: 0.67 to 0.71 U/kg T: Breakfast, dinner D: 12 weeks		GP1 1 (1) GP2 2 (2)		GP1 1 (1) GP2 1 (1)	Resulted in death, life-threatening experience, inpatient hospitalization, persistent or significant disability/ incapacity, or congenital anomaly/birth defect GP1 11 (8) GP2 3 (2)
Insulin aspart 70/30 vs. insulin aspart 70/30 + oral antidiabetic agents
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.51 U/kg/day T: Breakfast, dinner D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.3 U/kg/day T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks	GP1 F-B: 1.6 GP2 F-B: 0.8 GP1-GP2: 0.8 (SE 0.41) p: NS vs. GP2		GP1 total events for all groups: 5 GP2 total events for all groups: 5	GP1 1 (1) GP2 2 (2)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.7 U/kg/day T: Breakfast, dinner D: 18 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.5 U/kg/day T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks	GP1 F-B: 2.2 GP2 F-B: 4 GP1-GP2: -2* Experienced weight gain GP1 3* (3) p: < 0.05 overall GP2 7* (8)		GP1 2 (2) GP2 0 (0)	GP1 3 (3) GP2 1 (1)	Cellulitis GP1 1 (1) GP2 0 (0) Peripheral edema GP1 0* (0) GP2 6* (6)
Ushakova, 2007⁵⁹	GP1: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 55.5 U T: bid for 2 weeks, then tid D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.3 - 0.5 U/kg Mean: 44.8 U T: Breakfast, dinner D: 16 weeks Metformin (varied) Start: 500 mg qd or bid or 850 mg qd T: NR D: 14 weeks (started after 2 weeks)			GP1 0 (0) GP2 0 (0)	GP1 1 (1) GP2 1 (1)
Insulin lispro 70/30 vs. rapid-acting insulin analogue with intermediate-acting human insulin
Hirao, 2008⁶¹	GP1: Insulin aspart 70/30 (NR) T: bid D: 6 months GP2: Insulin aspart (NR) T: tid D: 6 months NPH insulin (NR) T: Optional multiple daily injections D: 6 months	GP1 B: 23.8 (4.1) p: NS F: 25.2 (4) p: < 0.0001 vs. baseline; NS vs. GP2 F-B: 1.47 (1.82) p: 0.013 vs. GP2 GP2 B: 24 (4.2) F: 24.8 (4.5) p: < 0.0001 vs. baseline F-B: 0.69 (1.04) GP1-GP2: 0*
Insulin lispro 75/25 vs. long-acting insulin analogues
Cox, 2007⁷⁴	GP1: Insulin lispro 75/25 (v) T: Breakfast, dinner D: 12 weeks Metformin (NR) T: NR D: 12 weeks GP2: Insulin glargine (v) T: Bedtime D: 12 weeks Metformin (NR) T: NR D: 12 weeks				GP1 0 (0) GP2 0 (0)
Jacober, 2006⁶⁴	GP1: Insulin lispro 75/25 (v) Mean: 0.353 IU/kg; 36.73 IU T: Breakfast, lunch D: 4 months Insulin lispro 75/25 (v) T: Dinner D: 4 months Existing OA agents (NR) T: NR D: 4 months GP2: Insulin glargine (v) Mean: 0.276 IU/kg; 27.98 IU T: Bedtime D: 4 months Existing OA agents (NR) T: NR D: 4 months	GP1 B: 98* F: 99.7 (18.6) p: 0.9106 F-B: 1.98 (0.44) p: < 0.0001 vs. baseline GP2 B: 97* F: 99 (19.1) F-B: 1.52 (0.46) p: 0.0015 vs. baseline GP1-GP2: 0* p: 0.457 vs. GP2			GP1 0 (0) GP2 0 (0)
Malone, 2004⁶⁵	GP1: Insulin lispro 75/25 (v) Mean: 0.62 U/kg T: Breakfast, dinner D: 16 weeks Metformin (NR) Mean: 1945 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.57 U/kg T: Bedtime D: 16 weeks Metformin (NR) Mean: 1997 mg Range: 1500 - 2550 mg T: NR D: 16 weeks	GP1 B: 91* F: 93 (18.8) p: 0.006 F-B: 2.3 (4) p: 0.006 GP2 B: 91* F: 93.1 (19.3) F-B: 1.6 (4) GP1-GP2: 0*			GP1 0 (0) GP2 1 (1)	Required hospitalization GP1 4 (4) GP2 1 (1)
Malone, 2005⁶⁶	GP1: Insulin lispro 75/25 (v) Mean: 0.42 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 2128 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.36 U/kg T: Bedtime D: 16 weeks Metformin (fix) Mean: 2146 mg Range: 1500 - 2550 mg T: NR D: 16 weeks	GP1 B: 77* F: 78.31 (15.13) p: 0.001 F-B: 0.82 (2.56) p: 0.001 vs. GP2 GP2 B: 77* F: 77.05 (14.38) F-B: 0.06 (2.49) GP1-GP2: 1*			GP1 1 (1) GP2 0 (0)	GP1 3 GP2 3
Roach, 2006⁶³	GP1: Insulin lispro 75/25 (v) Mean: 23 U (morning) and 37 U (evening) Range: 0 - 72 U (morning); 11 - 88 U (evening) T: Breakfast, dinner D: 12 weeks OA agents (NR) Start: Current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks GP2: Insulin glargine (v) Mean: 44 U Range: 14 U - 100 U T: Breakfast D: 12 weeks OA agents (NR) Start: Current dose T: NR D: 12 weeks Metformin (v) Start: 500 mg qd T: NR D: 12 weeks	GP1 F: 103.9 (17.8) p: 0.068 GP2 F: 102.5 (17.9)			GP1 1 (3) GP2 0 (0)
Insulin lispro 75/25 vs. premixed human insulins
Coscelli, 2003⁶⁷	GP1: Insulin lispro 75/25 (v) Mean: 38.1 U/l Range: 12 – 72 U/l T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days GP2: NPH/regular 70/30 (v) Mean: 37.3 U/l Range: 10 – 72 U/l T: Breakfast, dinner D: 12 days Diet/exercise D: 12 days	GP1 B: 79 (13.1) F: 79.4 (12.9) p: NS vs. baseline F-B: 0* GP2 B: 80.2 (11.8) F: 80.4 (12.8) p: NS vs. baseline F-B: 0* GP1-GP2: 0*			GP1 0 (0) GP2 0 (0)	GP1 1 GP2 2
Hermansen, 2002⁵⁸	GP1: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day				GP1 0 (0) GP2 0 (0)	GP1 number of events: 0 GP2 number of events: 0
Herz, 2002⁷¹	GP1: Insulin lispro 75/25 (v) Mean: 26.1 U T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 26.2 U T: Breakfast, dinner D: 4 weeks				GP1 0 (0) GP2 0 (0)
Herz, 2003¹²	GP1: Insulin lispro 75/25 (v) Mean: Mean: 31.6¶ (morning) and 26.8¶ U (evening) and 32.4§ (morning) and 27.6§ U (evening) T: Breakfast, dinner D: 4 weeks GP2: NPH/regular 70/30 (v) Mean: 32.3¶ (morning) and 26.4¶ U (evening) and 33.3§ (morning) and 27.5§ U (evening) T: Breakfast, dinner D: 4 weeks				GP1 0 (0) GP2 0 (0)
Malone, 2000⁴⁴	GP1: Insulin lispro 75/25 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days GP2: NPH/regular 70/30 (fix) Mean: 35.4 U (0.43 U/kg) T: Breakfast D: 2 days				GP1 0 (0) GP2 0 (0)
Mattoo, 2003⁷⁰	GP1: Insulin lispro 75/25 (NR) Mean: 20 U (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks GP2: NPH/regular 70/30 (NR) Mean: 20 U (morning), 32 U (evening) T: Breakfast, dinner D: 2 weeks	GP1 p: NS vs. baseline for all patients			GP1 0 (0) GP2 0 (0)
Roach, 1999⁷³	GP1: Insulin lispro 75/25 (v) Mean: 0.37 (morning), 0.28 U/kg (evening) T: Breakfast, dinner D: 13 weeks GP2: NPH/regular 70/30 (v) Mean: 0.36 (morning), 0.27 U/kg (evening) T: Breakfast, dinner D: 13 weeks	GP1 p: NS vs. GP2			GP1 0 (0) GP2 0 (0)
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day		GP1 1 (5) GP2 0 (0)	GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 0 (0)
Insulin lispro 75/25 vs. oral antidiabetic agents
Herz, 2002⁷²	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg Mean: 0.46 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glyburide (fix) Start: 15 mg/day T: Breakfast, dinner D: 16 weeks	GP1 B: 78.65 (SE 1.36) p: 0.519 vs GP2 F: 79.7 (SE 1.47) p: 0.151 vs GP2 F-B: 1.02 (SE 0.35) p: < 0.001 vs GP2 GP2 B: 77.34 (SE 1.53) F: 76.61 (SE 1.55) F-B: -0.85 (SE 0.18) GP1-GP2: 2*			GP1 2 (3) GP2 1 (1)	Liver carcinoma GP1 1 (1) GP2 0 (0)
Malone, 2003⁶⁸	GP1: Insulin lispro 75/25 (v) Mean: 0.19 (morning), 0.14 U/kg (evening) T: Breakfast, dinner D: 16 weeks Metformin (v) Mean: 1813 mg/day Range: 1500 – 2550 mg/day T: 2 to 3 times/day D: 16 weeks GP2: Metformin (v) Mean: 1968 mg/day Range: 1500 - 2550 mg/day T: 2 to 3 times/day D: 16 weeks Glibenclamide (v) Mean: 14.2 mg/day T: NR D: 16 weeks	GP1 B: 83 (15.2) F: 84 (15.1) F-B: 1* GP2 B: 81.7 (15.7) F: 82.2 (15.4) F-B: 0* GP1-GP2: 1* p: 0.33			GP1 1 (0) GP2 2 (1)	Treatment-emergent adverse events GP1 number of events: 7 GP2 number of events: 5
Tirgoviste, 2003⁴³	GP1: Insulin lispro 75/25 (v) Start: 0.3 - 0.5 U/kg T: Breakfast, dinner D: 16 weeks GP2: Glibenclamide (v) Start: 15 mg T: Breakfast, dinner D: 16 weeks	GP1 F-B: 1.32 (2.4) p: < 0.001 GP2 F-B: -0.7 (2.6) GP1-GP2: 2*
Insulin lispro 75/25 vs. insulin lispro 50/50
Schwartz, 2006⁶²	GP1: Insulin lispro 75/25 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day GP2: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day		GP1 1 (5) GP2 1 (4)	NR GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 0 (0)
Insulin lispro 50/50 vs. long-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg/day--mean Mean: 0.59 IU/kg/day T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin glargine (v) Start: 0.16 IU/kg/day--mean Mean: 0.43 IU/kg/day T: Bedtime D: 24 weeks	GP1 F-B: 1.8 (3.4) GP2 F-B: 0.7 (3.8) GP1-GP2: 1* BMI (in kg/m²) GP1 F-B: 0.6 (1.1) p: 0.19 vs GP2 GP2 F-B: 0.2 (1.3) GP1-GP2: 1*			GP1 0 (0) GP2 0 (0)
Robbins, 2007⁷⁹	GP1: Insulin lispro 50/50 (v) Mean: 0.7 U/kg T: Breakfast, lunch, dinner D: 24 weeks Metformin (fixed) Mean: 1641 mg T: bid D: 24 weeks GP2: Insulin glargine (v) Mean: 0.6 U/kg T: Bedtime D: 24 weeks Metformin (fixed) Mean: 1636 mg T: bid D: 24 weeks	GP1 B: 89.1 (20.4) F: 90 (20.5) p: < 0.001 vs. GP2 F-B: 1.2 (3.2) p: < 0.001 vs. GP2 GP2 B: 88.1 (19) F: 87.6 (19.3) F-B: -0.5 (2.8) GP1-GP2: 1*		GP1 N and % of events: 11 (7%) p: NS GP2 N and % of events: 5 (3.2%)	GP1 5 (3.2) GP2 1 (0.6)
Insulin lispro 50/50 vs. rapid-acting insulin analogues
Kazda, 2006⁷⁶	GP1: Insulin lispro 50/50 (v) Start: 0.30 IU/kg/day--mean Mean: 0.59 IU/kg/day T: Breakfast, lunch, dinner D: 24 weeks GP2: Insulin lispro (v) Start: 0.25 IU/kg/day-mean Mean: 0.50 IU/kg/day T: Breakfast, lunch, dinner D: 24 weeks	GP1 F-B: 1.8 (3.4) GP2 F-B: 2.3 (4.3) GP1-GP2: 0* BMI (in kg/m²) GP1 F-B: 0.6 (1.1) GP2 F-B: 0.9 (1.5) GP1-GP2: 0*			GP1 0 (0) GP2 0 (0)
Insulin lispro 50/50 vs. rapid-acting with long-acting insulin analogues
Rosenstock, 2008⁸⁰	GP1: Insulin lispro 50/50 (v) Start: Insulin glargine dose at entry (52.5 U) Mean: 123 U T: Breakfast, lunch, dinner D: 24 weeks Insulin lispro 75/25 (v) Start: Allowed to switch evening dose to insulin lispro 75/25 T: Dinner D: Unclear GP2: Insulin glargine (v) Start: 50% of insulin glargine dose at entry (54.9 U) Mean: 70 U T: Bedtime D: 24 weeks Insulin lispro (v) Start: 50% of insulin glargine dose at entry divided in 3 equal doses (54.9 U) Mean: 76 U T: Breakfast, Lunch, dinner D: 24 weeks			GP1 6 (3.2) number of events: 9 p: 0.600 for incidence GP2 9 (4.8) number of events: 13
Insulin lispro 50/50 vs. premixed human insulins
Roach, 1999¹⁰	GP1: Insulin lispro 50/50 (v) Mean: 0.31 U/kg T: Breakfast D: 3 months Insulin lispro 75/25 (v) Mean: 0.26 U/kg T: Dinner D: 3 months GP2: NPH/regular 50/50 (v) Mean: 0.32 U/kg T: Breakfast D: 3 months NPH/regular 70/30 (v) Mean: 0.26 U/kg T: Dinner D: 3 months	GP1 p: NS vs. GP2			GP1 0 (0) GP2 0 (0)
Schernthaner, 2004⁷⁷	GP1: Insulin lispro 50/50 (v) Mean: 64.6 IU T: Breakfast, lunch, dinner D: 12 weeks Diet/exercise D: 12 weeks GP2: NPH/regular 70/30 (v) Mean: 61.8 IU T: Breakfast, dinner D: 12 weeks Diet/exercise D: 12 weeks				GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 5 (12)
Schwartz, 2006⁶²	GP1: Insulin lispro 50/50 (fix) Start: 2/3 of usual daily dose Mean: 43.8 U T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 2/3 of usual daily dose Mean: 44.1 U T: Breakfast D: 1 day		GP1 1 (4) GP2 0 (0)	NR GP1 0 (0) GP2 0 (0)	GP1 0 (0) GP2 0 (0)
Yamada, 2007⁷⁸	GP1: Insulin lispro 50/50 (v) Start: current dose Mean: 0.37 (start), 0.38 U/kg (end) T: bid D: 4 months GP2: NPH/regular 70/30 (v) Start: current dose Mean: 0.34 (start), 0.37 U/kg (end) T: bid D: 4 months NPH/regular 50/50 (v) Start: current dose Mean: 0.34 (start), 0.37 U/kg (end) T: bid D: 4 months	BMI (in kg/m²) GP1 B: 27 (5.8) F: 27.3 (5.9) F-B: 0* p: NS vs. baseline GP2 B: 23.8 (3.4) F: 23.6 (3.6) F-B: 0* p: NS vs. baseline GP1-GP2: 0* p: NS vs. baseline			GP1 0 (0) GP2 0 (0)

Evidence Table 7. Quality of studies comparing a premixed insulin analogue to other diabetes treatments
Author, year	Clear quest	Rand# / Rand app#	Comp gp* / Exp asc* / Out not present*	Blind	Out assess	FU long enough	Lost to FU / Desc of WD	Conc	Fund / COI	Overall quality†
# Questions only rated for trials. * Questions only rated for non-randomized studies. † Overall quality ratings were good, fair, or poor, which were defined as: Good (low risk of bias). These studies have the least bias and results are considered valid. A study that adheres mostly to the commonly held concepts of high quality including the following: a formal randomized controlled study; clear description of the population, setting, interventions, and comparison groups; appropriate measurement of outcomes; appropriate statistical and analytic methods and reporting; no reporting errors; low dropout rate; and clear reporting of dropouts. Fair. These studies are susceptible to some bias, but it is not sufficient to invalidate the results. They do not meet all the criteria required for a rating of good quality because they have some deficiencies, but no flaw is likely to cause major bias. The study may be missing information, making it difficult to assess limitations and potential problems. Poor (high risk of bias). These studies have significant flaws that imply biases of various types that may invalidate the results. They have serious errors in design, analysis or reporting; large amounts of missing information; or discrepancies in reporting. Blind = blinding; Clear quest = clearly stated study questions or objectives; COI = conflict of interest; Comp gp = selection of comparison group; Conc = conclusions reflective of results; Desc of WD = description of withdrawals; Exp asc = ascertainment of exposure; FU = followup; Fund = funding source; Indep blind = independent blind assessment; NA = not applicable; NR = not reported; Out assess = outcome assessment; Out not present = demonstration that outcome was not present at study start; Rand = randomized; Rand app = randomization scheme appropriate
Abrahamian, 2005⁵³	Yes	Yes / NR	NA	No	Indep blind	Yes	< 10% / No	Partially	Pharmaceutical / Yes	Fair
Bebakar, 2007⁴⁶	Yes	Yes / NR	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Boehm, 2004⁴⁵ Boehm, 2002⁹	Yes	Yes / Yes	NA	Outcome assessors	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Christiansen, 2003¹³	No	Yes / NR	NA	Patients, providers	Indep blind	Yes	< 10% / Yes	Partially	Pharmaceutical / Yes	Fair
Coscelli, 2003⁶⁷	Yes	Yes / NR	NA	No	Indep blind	Yes	NR / No	Yes	Pharmaceutical / NR	Fair
Cox, 2007⁷⁴	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	> 10% / No	Yes	Pharmaceutical / NR	Fair
Hermansen, 2002⁵⁸	Yes	Yes / Yes	NA	No	Indep blind	No	< 10% / Yes	Yes	Pharmaceutical / NR	Fair
Herz, 2002⁷¹	Yes	Yes / NR	NA	No	Indep blind	Yes	< 10% / Yes	Yes	NR / NR	Fair
Herz, 2002⁷²	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / Yes	Partially	Pharmaceutical / Yes	Poor
Herz, 2003¹²	Yes	Yes / NR	NA	No	Indep blind	Yes	> 10% / No	Yes	Pharmaceutical / NR	Fair
Hirao, 2008⁶¹	Yes	Yes / NR	NA	No	Indep blind	Yes	> 10% / Yes	Yes	Non-pharmaceutical / No	Fair
Holman, 2007³²	Yes	Yes / Yes	NA	Outcome assessors	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Jacober, 2006⁶⁴	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	> 10% / Yes	Partially	Pharmaceutical / NR	Fair
Joshi, 2005⁵²	No	NA	Drawn from same community / other / NA	No	Indep blind	Yes	Complete FU / Yes	Partially	Pharmaceutical / Yes	Poor
Kann, 2006⁵⁰	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Kapitza, 2004⁵⁶	Yes	Yes / NR	NA	No	Indep blind	No	Complete FU / No	Yes	Pharmaceutical / NR	Fair
Kazda, 2006⁷⁶	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Kilo, 2003¹⁵	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Kvapil, 2006⁵¹	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Ligthelm, 2006⁸³	Yes	Yes / Yes	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Malone, 2000⁴⁴ Malone, 2000¹⁴	Yes	Yes / NR	NA	Patients, providers	Indep blind	No	< 10% / Yes	Yes	NR / NR	Fair
Malone, 2003⁶⁸	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	Complete FU / Yes	Yes	Pharmaceutical / NR	Good
Malone, 2004⁶⁵	Yes	Yes / Yes	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	NR / NR	Poor
Malone, 2005⁶⁶	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	NR / Yes	Fair
Mattoo, 2003⁷⁰	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / No	Yes	Pharmaceutical / NR	Fair
McNally, 2007⁴⁸	Yes	Yes / Yes	NA	Patients, providers	Indep blind, self report	Yes	< 10% / Yes	Yes	NR / Yes	Good
McSorley, 2002¹¹	Yes	Yes / NR	NA	Patients, providers	Indep blind	Yes	NR / No	Yes	Pharmaceutical / NR	Fair
Nauck, 2007⁴⁹	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Niskanen, 2004⁵⁵	Yes	Yes / Yes	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Raskin, 2005³⁹ Raskin, 2007⁴⁰, Brod, 2007⁴¹	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Raskin, 2007⁶⁰	No	Yes / NR	NA	No	Indep blind	Yes	< 10% / Yes	Yes	NR / No	Fair
Raz, 2003⁵⁷	Yes	Yes / Yes	NA	No	Indep blind, medical record review, self report	No	< 10% / Yes	Yes	Pharmaceutical / NR	Fair
Raz, 2005⁵⁴	Yes	Yes / NR	NA	No	Indep blind	Yes	> 10% / Yes	Yes	Pharmaceutical / Yes	Fair
Roach, 1999⁷³	Yes	Yes / NR	NA	No	Indep blind, self report	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Fair
Roach, 1999¹⁰	Yes	Yes / NR	NA	No	Indep blind	Yes	Complete FU / Yes	Yes	Pharmaceutical / NR	Fair
Roach, 2003⁶⁹	Yes	Yes / NR	NA	Patients, providers	Indep blind	No	< 10% / Yes	Yes	Pharmaceutical / NR	Good
Roach, 2006⁶³	Yes	Yes / NR	NA	No	Indep blind	Yes	> 10% / Yes	Yes	Pharmaceutical / Yes	Fair
Robbins, 2007⁷⁹	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / Yes	Good
Schernthaner, 2004⁷⁷	Yes	Yes / NR	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / NR	Fair
Rosenstock, 2008⁸⁰	Yes	Yes / Yes	NA	No	Indep blind	Yes	> 10% / Yes	Yes	Pharmaceutical / Yes	Good
Schwartz, 2006⁶²	Yes	Yes / NR	NA	Patients, providers	Indep blind	No	< 10% / Yes	Yes	Pharmaceutical / Yes	Fair
Tamemoto, 2007⁴⁷	Yes	Yes / No	NA	No	No description	Yes	< 10% / Yes	Yes	NR / NR	Poor
Tirgoviste, 2003⁴³, Roach, 2001⁴²	Yes	Yes / Yes	NA	No	Indep blind, self report	Yes	Complete FU / Yes	Yes	NR / NR	Good
Sun, 2007⁷⁵	Yes	NA	Drawn from the same community / secure record / NA	No	Medical record review	Yes	> 10% / No	Yes	Pharmaceutical / Yes	Fair
Ushakova, 2007⁵⁹	Yes	Yes / Yes	NA	No	Indep blind	Yes	< 10% / Yes	Yes	Pharmaceutical / No	Good
Yamada, 2007⁷⁸	Yes	Yes / Yes	NA	No	Indep blind	Yes	NR / No	Partially	NR / NR	Fair

Evidence Table 8. Applicability of studies comparing a premixed insulin analogue to other diabetes treatments
Author, year	Pop source / % enrolled / Run-in periods excluding >10%	Demographically representative	Illness severity representative	Dose, schedule, or route of administration reflective of current practice	Interventions or monitoring reflective of current practice	Comparison best alternative / Adequate dose, interval, schedule	Clinical outcomes measured / adverse events reported	Standards of care different from US
kg = kilogram; NA = not applicable; NR = not reported; OA = oral antidiabetic; Pop source = population source; U = units; US = United States
Abrahamian, 2005⁵³	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: No: assumed mostly Caucasian	Yes	Yes for only 2 of the 3	No: monitoring of blood glucose occurred 7 times/day	Yes / Yes	No / No	No
Bebakar, 2007⁴⁶	NR / ≥50% / Yes	Sex: Yes Age: NR Race/ethnicity: No: Western Pacific countries	No: insulin naive	Yes for all 3	Yes	Yes / NR	Yes / No	Yes
Boehm, 2004⁴⁵ Boehm, 2002⁹	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: NR	Yes	Yes for all 3	Yes	Yes / Yes	Yes / Yes	No
Christiansen, 2003¹³	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: different from US racial and ethnic make-up	Yes	Yes for all 3	No: monitoring too frequent	Yes / Yes	No / Yes	Yes
Coscelli, 2003⁶⁷	Subspecialty clinics / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: 100% Caucasian	No: excluded patients with diabetic complications; must have been taking insulin; average duration of diabetes was 14 years	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Cox, 2007⁷⁴	NR / ≥50% / NA	Sex: NR Age: Yes Race/ethnicity: No:	No: no early diabetics	Yes for all 3	NA	Yes / Yes	No / Adverse outcomes not reported	No
Hermansen, 2002⁵⁸	NR / ≥50% / NA	Sex: Yes Age: Yes Race/ethnicity: NR	No: subjects needed to have been on insulin and insulin dose < 1.4 U/kg, excluded those with diabetes complications	Yes for all 3	No: patients were given a single dose of insulin and a standard meal and then monitored for 5 hours afterwards	Yes / Yes	No / Yes	No
Herz, 2002⁷¹	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: all participants were currently taking insulin	Yes for all 3	No: patients were hospitalized for a few days while they performed an exercise test	Yes / Yes	No / Yes	No
Herz, 2002⁷²	Subspecialty clinics / ≥50% / No	Sex: Yes Age: No: subjects 60 to 80 years old Race/ethnicity: NR	No: excluded those with new diagnosis of type 2 diabetes	Yes for all 3	Yes	No: compared to glyburide when patients were already on maximum dose of glyburide / Yes	No / No	No
Herz, 2003¹²	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: all respondents have been taking insulins, none were currently taking an OA agent	Yes for all 3	No: investigators telephoned patients at least once weekly	Yes / Yes	No / Yes	No
Hirao, 2008⁶¹	Clinics and hospitals affiliated with JDDM / NR / NA	Sex: Yes Age: NR Race/ethnicity: No: Japanese study	Yes	Yes for all 3	Yes	Yes / Yes	Yes / Yes	Yes
Holman, 2007³²	Clinical centers / ≥50% / NA	Sex: Yes Age: Yes Race/ethnicity: No: over 90% Caucasian	No: patients were insulin naive	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Jacober, 2006⁶⁴	NR / ≥50% / NA	Sex: Yes Age: Yes Race/ethnicity: No: study contained more Caucasians and fewer African Americans and Mexican Americans	No: study likely excluded newly diagnosed and those with comorbidities	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Joshi, 2005⁵²	Outpatient clinics, subspecialty clinics / ≥50% / NA	Sex: No: 67 to 77% male per group Age: Yes Race/ethnicity: No: all from India	Yes	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Kann, 2006⁵⁰	NR / NR / No	Sex: Yes Age: NR Race/ethnicity: NR	No: male and female insulin-naive patients	Yes for all 3	Yes	Yes / Yes	Yes / Yes	Yes
Kapitza, 2004⁵⁶	NR / NR / NA	Sex: Yes Age: Yes Race/ethnicity: NR	No: participants had to have been on insulin for at least 6 months	Yes for only 1 of the 3	NA	Yes / Yes	No / Adverse outcomes not reported	No
Kazda, 2006⁷⁶	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: No: assumed mostly Caucasian	No: included those who have a longer duration of diabetes	Yes for all 3	Yes	No: would usually add glargine to OA agents as opposed to give it alone / Yes	No / Yes	No
Kilo, 2003¹⁵	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: Yes	Yes	Yes for all 3	No: 8-point glucose profile measurement is not used in clinical practice	Yes / Yes	No / Yes	No
Kvapil, 2006⁵¹	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: No: assumed mostly Caucasian	No: treatment naive patients not included	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Ligthelm, 2006⁸³	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: No: predominantly Caucasian with an Asian minority	No: only patients who previously used insulin	Yes for all 3	No: interventions and monitoring likely too frequent	No: better alternatives are available / Yes	No / Yes	No
Malone, 2000⁴⁴ Malone, 200014	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: all patients needed to be on insulin	Yes for all 3	No: patients were monitored in house and had frequent blood glucose measurements	Yes / Yes	No / Yes	No
Malone, 2003⁶⁸	NR / ≥50% / No	Sex: Yes Age: NR Race/ethnicity: No: 90% Caucasian, 2% African American, 7% Hispanic	No	Yes for all 3	No: there was intense titration of dosing and patient visits every 4 weeks for 16 weeks	Yes / Yes	No / Yes	No
Malone, 2004⁶⁵	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: patients were insulin naive and had to be poorly controlled on an OA agent for at least 30 days	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Malone, 2005⁶⁶	NR / NR / Yes	Sex: Yes Age: Yes Race/ethnicity: NR	No: all had previously taken insulin	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Mattoo, 2003⁷⁰	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: participants had to be taking insulin for at least 6 months	Yes for only 1 of the 3	Yes	Yes / Yes	No / Yes	No
McNally, 2007⁴⁸	NR / ≥50% / No	Sex: Yes Age: No: mean age of population is 62 with a standard deviation of 9 years. Study is unlikely capturing the younger (<44 years) diabetic population Race/ethnicity: NR	No: all respondents have been pretreated on insulin for at least 6 months	Yes for all 3	Yes	Yes / Yes	No / No	No
McSorley, 2002¹¹	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: participants had to be diagnosed with diabetes for at least 1 year	Yes for all 3	No: there was a sampling period where standard meals were provided for the participants	Yes / Yes	No / Yes	No
Nauck, 2007⁴⁹	Outpatient clinics / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: suboptimal blood sugar control	Yes for all 3	Yes	No: comparator is a new drug that is not being used often / Yes	Yes / Yes	Yes
Niskanen, 2004⁵⁵	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: included only patients who had been receiving insulin	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Raskin, 2005³⁹ Raskin, 2007⁴⁰ Brod, 2007⁴¹	NR / NR / Yes	Sex: Yes Age: NR Race/ethnicity: Yes	No: insulin naive patients	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Raskin, 2007⁶⁰	NR / >=50% / Yes	Sex: YesAge: Yes Race/ethnicity: Yes	No: insulin naive patients were enrolled	Yes for all 3	Yes	No: comaprison is placebo, technically / No: comparison is technically placebo	No / Yes	No
Raz, 2003⁵⁷	Outpatient clinics / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: 82% Caucasian	No: insulin naive patients	Yes for all 3	Yes	No / No: insulin dose was adjusted while gliben-clamide and rosiglitazone doses were not adjusted	No / Yes	No
Raz, 2005⁵⁴	NR / ≥50% / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: excluded those with serious complications or disease	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Roach, 1999⁷³	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: excluded those not taking insulin and those with diabetic complications	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Roach, 1999¹⁰	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: NR	No: participants must have been on insulin and could not have had any diabetes complications	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Roach, 2003⁶⁹	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: 100% Western Asian (Indian)	No: all had to have been taking insulin; excluded respondents taking OA agents	Yes for all 3	Yes	Yes / Yes	No / Yes	Yes
Roach, 2006⁶³	NR / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: 80% Caucasian and 20% African American; no Hispanics were included	No: needed to be on an OA agent or insulin for at least 3 months	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Robbins, 2007⁷⁹	NR / >=50% / No	Sex: Yes Age: Yes Race/ethnicity: No: only 6% African American and 22% were Asian but rest were representative (i.e. Caucasian and Hispanic numbers representative)	No: did not include those with renal or liver complications	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Rosenstock, 2008⁸⁰	NR / >=50% / NA	Sex: YesAge: NR Race/ethnicity: Yes	No: only those already receiving insulin	Yes for all 3	No: blood glucose was checked much more often: patients met with physicians much more often	Yes / Yes	No / Yes	No
Schernthaner, 2004⁷⁷	NR / NR / No	Sex: No: there were fewer males (23%) enrolled in the study Age: No: average age is 67 with standard deviation of 8.4 years. Unlikely capturing younger diabetics (e.g., <50 years of age) Race/ethnicity: NR	No: excluded respondents with severe diabetic complications; average time on insulin was over 5 years	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Schwartz, 2006⁶²	Subspecialty clinics / ≥50% / NA	Sex: No: 74% of the population was male Age: No: mean age of the population was 61 with a standard deviation of 10; study unlikely captured younger diabetics Race/ethnicity: No: the study population had fewer blacks and more Hispanics	No: only diabetics already on insulin were enrolled	No for all three	NA	Yes / Yes	No / No	Yes
Sun, 2007⁷⁵	NR / NR / NA	Sex: Yes Age: Yes Race/ethnicity: Yes	No: insulin naive	Yes for all 3	Yes	Yes / Yes	No / Adverse outcomes not reported	No
Tamemoto, 2007⁴⁷	Outpatient clinics / NR / No	Sex: Yes Age: Yes Race/ethnicity: No: assumed 100% Japanese	No: had diabetes for at least 1 year	Yes for all 3	Yes	Yes / Yes	No / Yes	No
Tirgoviste, 2003⁴³ Roach, 2001⁴²	NR / NR / No	Sex: Yes Age: NR Race/ethnicity: NR	No: only patients needing 1 OA agent	Yes for all 3	No: there were 5 visits in 12 weeks. Dose adjustments for insulin were made every 2-3 days.	No: OA agent dose could not increase / No: could not increase the OA agent dose	No / Yes	Yes
Ushakova, 2007⁵⁹	NR / >=50% / NR	Sex: No: majority of respondents were female Age: Yes Race/ethnicity: No: likely majority Caucasian	No: those treated with insulin or with diabetes-related complications were excluded	Yes for all 3	Yes	Yes / Yes	No / Yes	Yes
Yamada, 2007⁷⁸	NR / NR / NA	Sex: No: mostly male Age: Yes Race/ethnicity: No: assumed mostly Japanese	No: excluded insulin naive patients and those with severe comorbidity	Yes for all 3	Yes	Yes / Yes	No / Yes	No

Evidence Table 9. Comparative effectiveness of premixed insulin analogues and other diabetes treatments on clinical outcomes
Author, year	Intervention	Overall mortality, n (%)	CVD mortality, n (%)	CVD morbidity, n (%)	Nephropathy, n (%)
µg = microgram; bid = twice daily; CVD = cardiovascular disease; D = duration; F-B = mean difference from baseline; fix = fixed dosing; GP = group; GP1-GP2 = mean difference between the difference from baseline; IQR = interquartile range; IU = international unit; kg = kilogram; mg = milligram; NPH = neutral protamine Hagedorn; NR = not reported; p = p-value; qd = once daily; T = time of day when insulin taken; tid = thrice daily; U = unit; v = dose varied
Insulin aspart 70/30 vs. long-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin detemir (v) Start: 16 median IU/day Range: 10 – 24 IU/day T: Bedtime, twice if required D: 1 years Usual care D: 1 year	GP1 3 (1) GP2 0 (0)	Myocardial infarction GP1 3 (1) GP2 0 (0)		Change in plasma creatinine, mean (SD) GP1 F-B: 0.05 (0.09) p: 0.008 vs. GP2 GP2 F-B: 0.02 (0.11) GP1-GP2: 0* Change in ratio of albumin to creatinine, median (IQR) GP1 F-B: -0.9 (-8 – 9.7) p: overall 0.07 GP2 F-B: -1.8 (-10.6 – 2.7) GP1-GP2: 1*
Kann, 2006⁵⁰	GP1: Insulin aspart 70/30 (v) Start: 0.1 U/kg bid Mean: 0.4 U/kg T: Breakfast, dinner D: 26 weeks Metformin (v) Start: 500 mg bid or current dose T: Breakfast, dinner D: 26 weeks GP2: Insulin glargine (v) Start: 0.2 U/kg qday Mean: 0.39 U/kg T: preferred time D: 26 weeks Glimepiride (v) Start: 1 mg daily or current dose T: Breakfast D: 26 weeks			Peripheral vascular disorder GP1 1 (0.8) GP2 0 (0) Cardiac failure GP1 0 (0) GP2 1 (0.8)
Insulin aspart 70/30 vs. rapid-acting insulin analogues
Holman, 2007³²	GP1: Insulin aspart 70/30 (v) Start: 16 median IU/day Range: 10 – 26 IU/day T: bid D: 1 year Usual care D: 1 year GP2: Insulin aspart (v) Start: 18 median IU/day Range: 9 – 24 IU/day T: Breakfast, lunch, dinner D: 1 year Usual care D: 1 year	GP1 3 (1) GP2 1 (0)	Myocardial infarction GP1 3 (1) GP2 1 (0)		Change in plasma creatinine, mean (SD) GP1 F-B: 0.05 (0.09) p: 0.62 vs. GP2 GP2 F-B: 0.05 (0.12) GP1-GP2: 0* Change in ratio of albumin to creatinine, median (IQR) GP1 F-B: -0.9 (-8 – 9.7) p: overall 0.07 GP2 F-B: -0.9 (-12.4 – 6.2) GP1-GP2: 0*
Insulin aspart 70/30 vs. premixed human insulins
Boehm, 2004⁴⁵	GP1: Insulin aspart 70/30 (v) Start: 0.57 U/kg T: Breakfast, dinner D: 24 months GP2: NPH/regular 70/30 (v) Start: 0.57 U/Kg T: Breakfast, dinner D: 24 months	GP1 3 (5) GP2 1 (2)	Cardiac failure GP1 1 (2) GP2 0 (0)	Cardiovascular adverse events GP1 15 (26) events: 19 GP2 17 (25) events: 19
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Transient ischemic attack GP1 1 (2*) GP2 0 (0)
Insulin aspart 70/30 vs. oral antidiabetic agents
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.51 U/kg/day T: Breakfast, dinner D: 16 weeks GP2: Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg (end) T: once or twice daily D: 16 weeks	GP1 0 (0) GP2 0 (0)	Myocardial infarction GP1 0 (0) GP2 0 (0)
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.3 U/kg/day T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks GP2: metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks Glibenclamide (v) Start: 1.75 mg Mean: 2.33 (start), 6.58 mg (end) T: once or twice daily D: 16 weeks	GP1 1 (1) GP2 0 (0)	Myocardial infarction GP1 1 (1) GP2 0 (0)
Raskin, 2007⁶⁰	GP1: Insulin aspart 70/30 (v) Start: 6 U bid Mean: 0.6 U/kg/day T: Breakfast, dinner D: 34 weeks Metformin (fix) Mean: 2446 mg T: NR D: 34 weeks Pioglitazone (fix) Mean: 32.5 mg T: NR D: 34 weeks GP2: metformin (fix) Mean: 2439 mg T: NR D: Unclear Pioglitazone (fix) Mean: 31.7 T: NR D: Unclear			Withdrawn due to arrhythmia, angina pectoris, and coronary artery disease GP1 1 (1) GP2 2 (2)	Withdrawn due to increased blood creatinine GP1 1 (1) GP2 0 (0)
Raz, 2003⁵⁷	GP1: Insulin aspart 70/30 (v) Start: 6 - 8 U bid T: Breakfast, dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks GP2: Glibenclamide (fix) Range: 7.5 – 15 mg T: Dinner D: 6 weeks Rosiglitazone (fix) Start: 4 mg T: Breakfast D: 6 weeks			Non-fatal myocardial infarction GP1 1 (4) GP2 0 (0)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.7 U/kg/day T: Breakfast, dinner D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks			Non-fatal myocardial infarction GP1 1 (1) GP2 0 (0)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.5 U/kg/day T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks GP2: Glibenclamide (v) Start: 5 to 10 mg Mean: 14 mg T: Breakfast D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks			Non-fatal myocardial infarction GP1 0 (0) GP2 0 (0)
Insulin aspart 70/30 vs. exenatide
Nauck, 2007⁴⁹	GP1: Insulin aspart 70/30 (v) Start: 15.7 U/day Mean: 24.4 U/day T: Breakfast, dinner D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks GP2: Exenatide (v) Start: 5 µg bid Range: 5 - 10 µg bid T: Breakfast D: 52 weeks 'Optimally' effective metformin and sulfonylurea therapy (v) T: NR D: 52 weeks	GP1 1 (0.4) GP2 2 (0.8)		Unspecified cardiac disorder adverse events GP1 5 (2) GP2 10 (4)
Insulin aspart 70/30 vs. insulin lispro 75/25
Hermansen, 2002⁵⁸	GP1: Insulin aspart 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Transient ischemic attack GP1 1 (2*) GP2 0 (0)
Niskanen, 2004⁵⁵	GP1: Insulin aspart 70/30 (v) Mean: 0.65 U/kg to 0.67 U/kg T: Breakfast, dinner D: 12 weeks GP2: Insulin lispro 75/25 (v) Mean: 0.67 U/kg to 0.71 U/kg T: Breakfast, dinner D: 12 weeks	GP1 0 (0) GP2 1 (1)	Myocardial infarction GP1 0 (0) GP2 1 (1)
Insulin aspart 70/30 vs. insulin aspart 70/30 + oral antidiabetic agents
Kvapil, 2006⁵¹	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.51 U/kg/day T: Breakfast, dinner D: 16 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.3 U/kg/day T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 1660 mg daily Range: 500 - 3000 mg qd T: NR D: 16 weeks	GP1 0 (0) GP2 1 (1)	Myocardial infarction GP1 0 (0) GP2 1 (1)
Raz, 2005⁵⁴	GP1: Insulin aspart 70/30 (v) Start: 0.3 U/kg/day Mean: 0.7 U/kg/day T: Breakfast, dinner D: 18 weeks GP2: Insulin aspart 70/30 (v) Start: 0.2 U/kg/day Mean: 0.5 U/kg/day T: Breakfast, dinner D: 18 weeks Pioglitazone (fix) Start: 30 mg Mean: 30 mg T: Breakfast D: 18 weeks			Non-fatal myocardial infarction GP1 1 (1) GP2 0 (0)
Insulin aspart 70/30 vs. rapid-acting insulin analogue with intermediate-acting human insulin
Hirao, 2008⁶¹	GP1: Insulin aspart 70/30 (NR) T: bid D: 6 months GP2: Insulin aspart (NR) T: tid D: 6 months NPH insulin (NR) T: Optional multiple daily injections D: 6 months	GP1 1 (1) GP2 0 (0)
Insulin lispro 75/25 vs. long-acting insulin analogues
Malone, 2004⁶⁵	GP1: Insulin lispro 75/25 (v) Mean: 0.62 U/kg T: Breakfast, dinner D: 16 weeks Metformin (NR) Mean: 1945 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.57 U/kg T: Bedtime D: 16 weeks Metformin (NR) Mean: 1997 mg Range: 1500 - 2550 mg T: NR D: 16 weeks			Congestive heart failure GP1 1 (1) GP2 0 (0)
Malone, 2005⁶⁶	GP1: Insulin lispro 75/25 (v) Mean: 0.42 U/kg T: Breakfast, dinner D: 16 weeks Metformin (fix) Mean: 2128 mg Range: 1500 - 2550 mg T: NR D: 16 weeks GP2: Insulin glargine (v) Mean: 0.36 U/kg T: Bedtime D: 16 weeks Metformin (fix) Mean: 2146 mg Range: 1500 - 2550 mg T: NR D: 16 weeks	GP1 1 (1) GP2 1 (1)	Myocardial infarction GP1 1 (1) GP2 0 (0)
Insulin lispro 75/25 vs. premixed human insulins
Hermansen, 2002⁵⁸	GP1: Insulin lispro 75/25 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day GP2: NPH/regular 70/30 (fix) Start: 0.4 U/kg T: Breakfast D: 1 day			Transient ischemic attack GP1 0 (0) GP2 0 (0)
Insulin lispro 75/25 vs. oral antidiabetic agents
Malone, 2003⁶⁸	GP1: Insulin lispro 75/25 (v) Mean: 0.19 (morning), 0.14 U/kg (evening) T: Breakfast, dinner D: 16 weeks Metformin (v) Mean: 1813 mg/day Range: 1500 – 2550 mg/day T: 2 to 3 divided doses with meals D: 16 weeks GP2: Metformin (v) Mean: 1968 mg/day Range: 1500 – 2550 mg/day T: 2 to 3 divided doses with meals D: 16 weeks Glibenclamide (v) Mean: 14.2 mg/day T: NR D: 16 weeks	GP1 1 (0) GP2 0 (0)
Insulin lispro 50/50 vs. premixed human insulins
Schernthaner, 2004⁷⁷	GP1: Insulin lispro 50/50 (v) Mean: 64.6 IU T: Breakfast, lunch, dinner D: 12 weeks Diet/exercise D: 12 weeks GP2: NPH/regular 70/30 (v) Mean: 61.8 IU T: Breakfast, dinner D: 12 weeks Diet/exercise D: 12 weeks	GP1 0 (0) GP2 1 (2)
Insulin lispro 50/50 vs. long-acting insulin analogues
Robbins, 2007⁷⁹	GP1: Insulin lispro 50/50 (v) Mean: 0.7 U/kg T: Breakfast, lunch, dinner D: 24 weeks Metformin (fixed) Mean : 1641 mg T: bid D: 24 weeks GP2: Insulin glargine (v) Mean: 0.6 U/kg T: Bedtime D: 24 weeks Metformin (fixed) Mean: 1636 mg T: bid D: 24 weeks		&nb