From randomized clinical trials to registry analysis: where is the evidence?






Mykola D. KHALANGOT
MD, PhD, DrMedSc
Associate Professor
Department of Endocrinology
P. L. Shupyk National Medical Academy of Postgraduate
Education and Senior Research Fellow
Department of Epidemiology
V. P. Komissarenko Institute of Endocrinology and Metabolism
National Academy of Medical Sciences of Ukraine
Kiev, UKRAINE

From randomized clinical trials to registry analysis: where is the evidence?


by M. D. Khalangot, Ukraine



Recent results from large randomized clinical trials (RCTs) started a new wave of discussion on approaches to glucose-lowering treatment of type 2 diabetes (T2D). Differing results prompted experts to discuss underlying reasons; one particular focus was the increased cardiovascular mortality in the ACCORD trial (Action to Control CardiOvascular Risk in Diabetes), which remains unexplained. Such concerns justify a more in-depth search for explanations behind the controversial results. One possible factor is the sulfonylurea (SU) used: gliclazide only, as in ADVANCE (Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation), or glimepiride only, as in ACCORD and the VADT (Veterans Affairs Diabetes Trial). Nowadays, it is believed that high-quality observational studies may extend evidence over a wider population and are likely the best choice for identifying harms, or when RCTs are unethical or impractical. We should acknowledge a small number of RCTs that directly evaluate the effects of various SUs—in particular for direct comparisons of gliclazide and glibenclamide; nevertheless, the relevance of observational studies is increasing due to issues arising from comparative analysis of the latest megatrials. Observational studies, in particular those based on constantly growing primary care–based territorial registers, provide valuable information about the results of T2D treatment that can complement or clarify the results of RCTs. Pharmacoepidemiological analysis of T2D treatment results requires differential evaluation of individual SUs, rather than an overall generation-based assessment.

Medicographia. 2013;35:15-20 (see French abstract on page 20)



Recent results from large randomized clinical trials (RCTs),1-3 as well as observational epidemiological studies,4-12 have started a new wave of discussion on various approaches to glucose-lowering treatment of type 2 diabetes (T2D).13-19 Several assessments based on these trials led to the banning of one of the two oral glucose-lowering drugs (OGLDs) from the thiazolidinedione group, and to special clarifications by the European Medicines Authority Safety Review concerning use of the other OGLD (pioglitazone) from this drug group.20,21 Such firm conclusions have not been drawn concerning other OGLDs; however, certain concerns of the medical community, reinforced by differing results from several RCTs and some observational trials, justify a more in-depth investigation into possible explanations behind these differences. Publication of main RCT results in 2008-20091-3 led to recommendations by several experts calling for complete abandonment of OGLDs (with a single exception, made for metformin)17 or at least a ban on the use of all sulfonylureas (SUs),16 proving the high level of distrust in the current model of T2D treatment, which has been used for the last decade. SU critics argue that out of five large RCTs—the UGDP (University Group Diabetes Program), the UKPDS (United Kingdom Prospective Diabetes Study), ACCORD (Action to Control Cardiovascular Risk in Diabetes), ADVANCE (Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation), and the VADT (Veteran Affairs Diabetes Trial)—two trials (the UGDPandACCORD) showed an increase in cardiovascular disease (CVD) events and mortality, and the others showed no reduction in CVD events and mortality in patients receiving OGLDs, except for a group of overweight T2D patients treated with metformin versus diet only in the framework of the UKPDS.

Several experts discussing all possible reasons for such differing results in the main trials came to no conclusion explaining the increased cardiovascular mortality in ACCORD22 despite a multitude of theories. However, we believe that there are only a few theories to explain differing results for T2D glucose- lowering treatment intensification: for example, with the UGDP, it may be insufficient initial randomization that led to differences in the occurrence of adverse outcomes in tolbutamide and placebo groups,23 and perhaps tolbutamide, a first generation SU, was not the best choice, as second-generation SUs were shortly thereafter developed by the pharmaceutical industry.

ACCORD and ADVANCE trials, when compared, indicate quantitative differences in the use of insulin and thiazolidinediones.22 Indeed, differences in the frequency of severe hypoglycemia requiring medical assistance or with an impaired level of consciousness (%/year) (3.1% versus 1.0% and 0.7% versus 0.4%, in the intensive and standard treatment groups of ACCORD and ADVANCE, respectively) as well as corresponding changes in weight (kg) (+3.5 versus +0.4 and –0.1 versus –1.0 in ACCORD and ADVANCE, respectively) mentioned in these studies smoothly conform with a more frequent insulin use (77% and 41% in intensive groups of ACCORD and ADVANCE, respectively). However, in the VADT trial, even though insulin was used more often (90% versus 74% in the intensive treatment group and the standard treatment group, respectively), there was no increase in mortality in the intensive group of this trial. One explanation for increased mortality in the intensive group of ACCORD could be the faster lowering of hyperglycemia: in ACCORD, the target glycated hemoglobin level was achieved after 3 months; this took over 6 months in ADVANCE and VADT.





Surprisingly, the retrospective epidemiological analysis of the ACCORD study showed that among participants who experienced an episode of severe hypoglycemia, the relative risk of death was lower in those in the intensive glycemia treatment arm than in those in the standard treatment arm and the increased relative risk of mortality observed in the intensive treatment group in the ACCORD trial cannot be explained by severe hypoglycemia as it was measured in the study.12 We may suppose that unrecognized hypoglycemic episodes are one of many possible explanations for the increased mortality in ACCORD.

Is there a specific connection between mortality risk in T2D patients and various OGLDs, including second-generation SUs?

An apparent difference, regularly omitted by experts,22 between discussed large trials is in the SU used: gliclazide only (ADVANCE) or glimepiride only (ACCORD and the VADT). Both of these drugs are second-generation SUs that essentially differ pharmacodynamically and in their molecular mechanisms from tolbutamide (a first-generation SU which fell under suspicion due to the UGDP trial) and the still popular glibenclamide (glyburide), a second-generation SU. Thus, it seems that it is not the generation of the drug, but its qualities, including specificity for pancreatic β-cell, but not myocardial or vascular smooth muscle cell receptors, as well as noninterference in adaptation to ischemia (ischemic preconditioning) and extent of receptor bonding strength (reversibility) that determine the safety profile of a specific SU. Gliclazide and glimepiride are on equal footing in regard to the first two factors; however, when it comes to reversibility, gliclazide has an advantage.24-28 Gliclazide has additional beneficial qualities, such as antioxidant properties, which restore endothelial function, reduce platelet reactivity, and exert an anti-inflammatory effect.29-34

In the UKPDS, there was no difference in diabetes-related death or all-cause mortality between chlorpropamide (a first generation SU) and glibenclamide (a second-generation SU),35 which identifies a need for further comparisons of safety among second-generation SUs. According to the hierarchy of clinical studies carried out within the last decade, RCTs and, especially, systematic reviews of several of these trials are traditionally considered gold standards for judging the benefits of treatment, mainly because it is conceptually easier to attribute any observed effect to the treatments being compared.36,37 However, Pocock and Elbourne38 indicated that the search for corresponding RCTs in only the most prestigious journals can select a small, potentially atypical subgroup of available trials, which influences the quality of systematic reviews of these trials. A good example of this is the conclusion of a Canadian meta-analysis that glibenclamide indeed leads to more frequent hypoglycemia, compared with other sulfonylurea derivatives, but does not cause higher risk of cardiovascular events andmortality.39 Thismeta-analysis has been widely promoted in post-Soviet states, perhaps due to the fact that glibenclamide is the most frequently prescribed OGLD in these countries. It should be noted that the authors used the results of 21 RCTs for this meta-analysis, and only 335,40,41 of those were used for evaluating cardiovascular events and mortality. Only 2 RCTs35,40 out of 3 that were the basis for the above meta-analysis compared the effect of glibenclamide with other sulfonylurea derivatives, which in fact did not include gliclazide. Consequently, the conclusions about glibenclamide’s safety39 are based on only 1 RCT (457 patients, one-year follow- up) that proved glimepiride is similar to glibenclamide in terms of metabolic control and tolerance.40

Barton,37 in his comments on some current studies, states that “they do not justify a major revision of the hierarchy of evidence, but they do support a flexible approach in which [RCTs] and observational studies have complementary roles.” He believes that “high-quality observational studies may extend evidence over a wider population and are likely to be dominant in the identification of harms and when [RCTs] would be unethical or impractical.”37 Thus, observational studies are relevant, and their relevance increases with issues arising from comparative analysis of recent megatrials. This does not take away from the importance of a small number of RCTs that directly evaluate the effects of various SUs. This is particularly true for direct comparisons of gliclazide and glibenclamide that are otherwise not likely to be found. This situation could be explained by ethical considerations.

OGLD-related mortality risk in recent observational studies with some explanations for differing results

Recent retrospective studies4-9 and one prospective11 cohort observational study compared CVD and/or mortality risk in T2D patients treated with different SU monotherapies5,7,11 or assessed mortality risk with different SU monotherapies or thiazolidinediones versus metformin.4,8,9 In retrospective observational cohort studies of a primary care–based diabetes register carried out in Ukraine, we evaluated risk of total and CVD mortality in a cohort of T2D patients that were treated with either glibenclamide (n=50 341), glimepiride (n=2479), or gliclazide (n=11 368).6 Total mortality was lower for gliclazide and glimepiride versus the glibenclamide cohort: hazard ratios (HRs) were 0.33 (95% confidence interval [CI], 0.26-0.41; P<0.001) and 0.605 (95% CI, 0.413-0.886; P<0.01), respectively. CVD mortality risk reduction versus glibenclamide was significant only in the gliclazide cohort: 0.29 (95% CI, 0.21- 0.38; P<0.001). Thus, glibenclamide treatment of T2D is associated with greater risk of all-cause mortality versus gliclazide or glimepiride treatment, and CVD mortality versus gliclazide treatment. Total mortality risk for those treated with glimepiride was higher than in those treated with gliclazide: HR=1.85 (95% CI, 1.19-2.90; P=0.006).6

This observational, epidemiological primary care–based study revealed an adverse influence of glibenclamide in terms of higher general and cardiovascular mortality, compared with gliclazide, in the treatment of T2D patients, consistent with findings in studies by Johnsen et al42 andMonami et al.43 These latter studies, however, either touched upon the risk of myocardial infarction and related mortality,42 or were based on a much smaller cohort of patients.43 It is quite possible that a smaller patient cohort (568 patients), compared with the one used in our study (64 288 patients) allowed Monami et al to find a glibenclamide-associated age- and gender-adjusted mortality risk increase only for all-cause mortality, and not for CVD mortality.43 Glimepiride treatment is also linked to reduction in all-cause mortality, compared with glibenclamide treatment. However, we found no proof of CVD mortality reduction. Furthermore, the risk of all-cause mortality was significantly higher for glimepiride-treated patients, compared with gliclazide- treated ones.

Investigators from the French Registry of Acute ST-Elevation and Non-ST-Elevation Myocardial Infarction analyzed the outcomes of 1310 diabetic patients: among SU-treated patients, in-hospital mortality was lower in patients receiving gliclazide or glimepiride (2.7%), compared with glibenclamide (7.5%; P=0.019). The lower risk in patients receiving gliclazide/ glimepiride versus glibenclamide persisted after multivariate adjustment (odds ratio 0.15; 95% CI, 0.04-0.56).11 These results seemed to be a convincing reason for Matthew C. Riddle to say goodbye to glibenclamide.17

As I wrote earlier,44 in an article by Pantalone et al,7 the authors did not identify an increased total mortality risk among individual SUs—glibenclamide versus glimepiride or glipizide— but did suggest that glimepiride may be the preferred SU in patients with a history of coronary artery disease (CAD). The authors found that in a retrospective cohort of patients with CAD, the HR for mortality in the subgroup of glibenclamide versus glimepiride was 1.36 (95% CI, 0.96-1.91; P=0.081).

Pantalone et al7 refer to our assessment of total and cardiovascular mortality HRs in patients treated with gliclazide versus glibenclamide,6 considering them based on incorrect adjustment for variables. These authors estimated the HR for glipizide versus glyburide (glibenclamide), because glipizide and glimepiride are SUR1-specific (pancreatic-specific) sulfonylureas, available and commonly used in the United States (gliclazide is not available in the United States). Meanwhile, the interaction of these molecules with the SU receptors is different; for example, their half-maximal inhibitory concentration on channel activity differs by more than 10 times, whereas the corresponding differences between glibenclamide, glipizide, and glimepiride could be significantly lower.45 According to a recent nationwide register–based study in Denmark, monotherapy with glibenclamide, glimepiride, or glipizide, but not with gliclazide, is associated with higher mortality and CAD risk compared with metformin.9 It is even speculated that differences between SUs may underpin the different outcomes observed in the ACCORD and ADVANCE trials.1,2

In our study of 119 570 patients who had originally been assigned to monotherapy with glibenclamide, glimepiride, or gliclazide, the received treatment was confirmed to be unchanged in only 64 288 cases after a minimum of two followup checks.6 We managed to avoid bias in risk assessments that have arisen due to changes in treatment. This is why we have obtained gliclazide versus glibenclamide total and CVD mortality HRs that were so high that they have sustained adjustments for seven variables. The HR for glimepiride versus glibenclamide was not as high and was significant only for total mortality, without adjustment. We were unable to consider the influence of socioeconomic differences on the risk of SU-related mortality, but in the case of the glimepiride versus gliclazide comparison (HR 1.8; 95% CI, 1.2-2.9; P=0.006), this factor was not significant, as the cost for these drugs in Ukraine is the same.

It seems that Pantalone et al7 did not verify whether treatment remained unchanged during the observation period and did not mention the unadjusted HRs. The authors only provided HRs simultaneously adjusted for 22 variables, which greatly complicates the impact assessment of each variable. Furthermore, these authors indicated the need for prospective studies to assess the risk of individual SUs,7 but if gliclazide will not be included in such assessments, the truth will remain unrevealed. In reply to our comment,44 Pantalone et al46 confirmed that approximately 70% of the cohort remained on a single drug (baseline medication) throughout their time in the cohort, and believe that restricting an analysis to patients who continue the baseline drug throughout their duration in the cohort could create substantial bias.

This statement is partially true, as with increasing of the observation period, the selection factor may have a stronger influence on the results. On the other hand, only 70% (according to Ukrainian diabetic register data, only 50%) of the cohort remained on a single drug (baseline medication) throughout their time in the cohort, clearly lowering assessment accuracy. Retrospective analyses of large registers, that were based on primary care records and carried out in the United Kingdom4 and Denmark9 have avoided the lowering of assessment accuracy that can emerge due to changes in treatment. When creating regression models, these authors used an approach based on estimating the number of pharmacological “intervals,” in other words, periods of time during which there were no changes in treatment,4 or fixed 3-month periods.9

Such an approach allowed comparing the risks of adverse treatment outcomes that would be definitely associated with the use of a particular OGLD. Metformin was considered the reference treatment in both trials. Unfortunately, in the British trial, SUs were categorized only as first- and second-generation drugs. This prevented comparison ofmortality risks among second-generation SUs. Such comparison in the Danish study revealed significantly higher mortality risks in patients treated with glibenclamide, glimepiride, and glipizide, compared with metformin-treated ones, whereas there was no difference in mortality risk between gliclazide- and metformin-treated patients.9 Epidemiologists that analyzed the UK General Practice Research Database4 revealed an equal increase in total mortality risk for first- and second-generation SUs (versus metformin) and a decrease in such risk for thiazolidinediones. These data seem quite unexpected in light of a recent ban of rosiglitazone, and limitations laid down for pioglitazone (the latter is due to observational data from a French pharmacological register47 about an increase in bladder cancer in patients treated with pioglitazone); however, they are in agreement with confirmed data from the PROactive trial ([PROspective pioglitAzone Clinical Trial In macroVascular Events]; histologic diagnosis in one of the cases was reconsidered,48 leading to a statistical significance in differences between pioglitazone and control groups49), and have been recently confirmed once again with retrospective analysis of the UK General Practice Research Database.50 Evaluation of the balance between increasing bladder cancer risk and decreasing macrovascular events risk51 associated with the use of pioglitazone in the treatment of T2D should be continued; however, today we must note that for pioglitazone we have data from the experimental trial PROactive and from observational retrospective assessments received from medical registers that are in agreement and are mutually complementary.

Conclusion

Observational studies, particularly those based on constantly growing primary care–based territorial registers provide valuable information on the results of T2D treatment that can complement or clarify the results of RCTs. Pharmacoepidemiological analysis of T2D treatment results requires a differential evaluation of individual SUs rather than an overall generation- based assessment. Such individual evaluation provides convincing arguments about the safe use of gliclazide, an SU extensively verified in RCTs. _


References
1. Action to Control Cardiovascular Risk in Diabetes Study Group; Gerstein HC, Miller ME, Byington RP, et al. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358(24):2545-2559.
2. ADVANCE Collaborative Group; Patel A, MacMahon S, Chalmers J, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358(24):2560-2572.
3. Duckworth W, Abraira C, Moritz T, et al; VADT Investigators. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med. 2009;360(2):129-139.
4. Tzoulaki I, Molokhia M, Curcin V, et al. Risk of cardiovascular disease and all cause mortality among patients with type 2 diabetes prescribed oral antidiabetes drugs: retrospective cohort study using UK general practice research database. BMJ. 2009;339:b4731.
5. Gallagher AM, Smeeth L, Seabroke S, Leufkens HG, van Staa TP. Risk of death and cardiovascular outcomes with thiazolidinediones: a study with the general practice research database and secondary care data. PLoS One. 2011;6(12): e28157.
6. Khalangot M, Tronko M, Kravchenko V, Kovtun V. Glibenclamide-related excess in total and cardiovascular mortality risks: data from large Ukrainian observational cohort study. Diabetes Res Clin Pract. 2009;86:247-253.
7. Pantalone KM, Kattan MW, Yu C, et al. The risk of overall mortality in patients with type 2 diabetes receiving glipizide, glyburide, or glimepiride monotherapy: a retrospective analysis. Diabetes Care. 2010;33:1224-1229.
8. Pantalone KM, Kattan MW, Yu C, et al. Increase in overall mortality risk in patients with type 2 diabetes receiving glipizide, glyburide, or glimepiride monotherapy vs. metformin: a retrospective analysis. Diabetes Obes Metab. 2012; 14(9):803-809.
9. Schramm TK, Gislason GH, Vaag A, et al. Mortality and cardiovascular risk associated with different insulin secretagogues compared with metformin in type 2 diabetes, with or without a previous myocardial infarction: a nationwide study. Eur Heart J. 2011;32:1900-1908.
10. Ferrara A, Lewis JD, Quesenberry CP Jr, et al. Cohort study of pioglitazone and cancer incidence in patients with diabetes. Diabetes Care. 2011;34(4):923-929.
11. Zeller M, Danchin N, Simon D, et al; French Registry of AcuteST-Elevation and Non-ST-Elevation Myocardial Infarction Investigators. Impact of type of preadmission sulfonylureas on mortality and cardiovascular outcomes in diabetic patients with acute myocardial infarction. J Clin Endocrinol Metab. 2010;95(11): 4993-5002.
12. Bonds DE, Miller ME, Bergenstal RM, et al. The association between symptomatic, severe hypoglycaemia and mortality in type 2 diabetes: retrospective epidemiological analysis of the ACCORD study. BMJ. 2010;340:b4909.
13. Schernthaner G, Barnett AH, Betteridge DJ, et al. Is the ADA/EASD algorithm for the management of type 2 diabetes (January 2009) based on evidence or opinion? A critical analysis. Diabetologia. 2010;53(7):1258-1269.
14. Terry T, Raravikar K, Chokrungvaranon N, Reaven PD. Does aggressive glycemic control benefit macrovascular and microvascular disease in type 2 diabetes? Insights from ACCORD, ADVANCE, and VADT. Curr Cardiol Rep. 2012;14(1): 79-88.
15. Hamet P.What matters in ADVANCE and ADVANCE-ON. Diabetes Obes Metab. 2012;14(suppl 1):20-29.
16. Carnevale Schianca GP, Sola D, Rossi L, Fra GP, Bartoli E. The management of type 2 diabetic patients with hypoglycemic agents. ISRN Endocrinol. 2012; 2012:601380.
17. Havas S. The ACCORD trial and control of blood glucose level in type 2 diabetes mellitus: time to challenge conventional wisdom. Arch Intern Med. 2009; 169(2):150-154.
18. Boussageon R, Supper I, Bejan-Angoulvant T, et al. Reappraisal of metformin efficacy in the treatment of type 2 diabetes: a meta-analysis of randomised controlled trials. PLoS Med. 2012;9(4):e1001204.
19. Riddle MC. More reasons to say goodbye to glyburide. J Clin Endocrinol Metab. 2010;95(11):4867-4870.
20. European Medicines Agency recommends suspension of Avandia, Avandamet and Avaglim: anti-diabetes medication to be taken off the market. European Medicines Agency. http://www.ema.europa.eu/ema/index.jsp?curl=pages/ news_and_events/news/2010/09/news_detail_001119.jsp&mid=WC0b01a c058004d5c1.
21. Conclusions of European Medicines Authority Safety Review of Pioglitazone Welcomed by Association of British Clinical Diabetologists. https://www. diabetologists.org.uk/Shared_Documents/position_papers/ABCD_pioglitazone _press_release_23.7.11.pdf.
22. Hamnvik OP, McMahon GT. Glycemic targets for patients with type 2 diabetes mellitus. Mt Sinai J Med. 2009;76:227-233.
23. Leibel B. An analysis of the University Group Diabetes Study Program: data results and conclusions. Can Med Assoc J. 1971;105:292-294.
24. Hosker JP, Rudenski AS, Burnett MA, Matthews DR, Turner RC. Similar reduction of first- and second-phase B-cell responses at three different glucose levels in type 2 diabetes and the effect of gliclazide therapy. Metabolism. 1989; 38:767-772.
25. Gregorio F, Ambrosi F, Cristallini S, Pedetti M, Filipponi P, Santeusanio F. Therapeutical concentrations of tolbutamide, glibenclamide, gliclazide and gliquidone at different glucose levels: in vitro effects on pancreatic A- and B-cell function. Diabetes Res Clin Pract. 1992;18:197-206.
26. Ashcroft FM, Gribble F. Tissue specific effects of sulfonylureas. Lessons from studies of cloned KATP channels. J Diabetes Complications. 2000;14:192-196.
27. Lawrence CL, Proks P, Rodrigo GC, et al. Gliclazide produces high-affinity block of KATP channels in mouse isolated pancreatic beta cells but not rat heart or arterial smooth muscle cells. Diabetologia. 2001;44:1019-1025.
28. Gribble FM, Ashkroft FM. Differential sensitivity of beta-cells and extra pancreatic KATP channel to gliclazide. Diabetologia. 1999;42:845-848.
29. O’Brien RC, Luo M, Balazs N, Mercuri J. In vitro and in vivo antioxidant properties of gliclazide. J Diabetes Complications. 2000;14:201-206.
30. Fava D, Cassone-Faldetta M, Laurenti O, et al. Gliclazide improves anti-oxidant status and nitric oxide-mediated vasodilation in type 2 diabetes. Diabetic Med. 2002;19:752-757.
31. Okouchi M, Okayama N, Omi H, et al. The antidiabetic agent, gliclazide, reduces high insulin-enhanced neutrophil-transendothelial migration through direct effects on the endothelium. Diabetes Metab Res Rev. 2004;20:232-238.
32. Drzewoski J, Zurawska-Klis M. Effect of gliclazide modified release on adiponectin, interleukin 6, and tumor necrosis factor-plasma levels in individuals with type 2 diabetes. Curr Med Res Opin. 2006;22:1921-1926.
33. Rakel A, Renier G, Roussin A, Buithieu J, Mamputu JC, Serri O. Beneficial effects of gliclazide modified release compared with glibenclamide on endothelial activation and low-grade inflammation in patients with type 2 diabetes. Diabetes Obes Metab. 2007;9:127-129.
34. Konya H, Hasegawa Y, Hamaguchi T, et al. Effects of gliclazide on platelet aggregation and the plasminogen activator inhibitor type 1 level in patients with type 2 diabetes mellitus. Metabolism. 2010;59:1294-1299.
35. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet. 1998; 352(9131):837-853.
36. Sackett DI, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn’t. BMJ. 1996;312:7172.
37. Barton S. Which clinical studies provide the best evidence? The best RCT still trumps the best observational study. BMJ. 2000;321:255-256.
38. Pocock SJ, Elbourne DR. Randomized trials or observational tribulations? N Engl J Med. 2000;342:1907-1909.
39. Gangji AS, Cukierman T, Gerstein HC, Goldsmith CH, Clase CM. A systematic review and meta-analysis of hypoglycemia and cardiovascular events: a comparison of glyburide with other secretagogues and with insulin. Diabetes Care. 2007;30(2):389-394.
40. Draeger KE, Wernicke-Panten K, Lomp HJ, Schuler E, Rosskamp R. Longterm treatment of type 2 diabetic patients with the new oral antidiabetic agent glimepiride (Amaryl): a double-blind comparison with glibenclamide. Horm Metab Res. 1996;28(9):419-425.
41. Marbury T, Huang WC, Strange P, Lebovitz H. Repaglinide versus glyburide: a one-year comparison trial. Diabetes Res Clin Pract. 1999;43(3):155-166.
42. Johnsen SP, Monster TB, Olsen ML, et al. Risk and short-term prognosis of myocardial infarction among users of antidiabetic drugs. Am J Ther. 2006;13(2): 134-140.
43. Monami M, Balzi D, Lamanna C, et al. Are sulphonylureas all the same? A cohort study on cardiovascular and cancer-related mortality. Diabetes Metab Res Rev. 2007;23(6):479-484.
44. Khalangot MD. Diabetes Care. 2011;34(8):e138; author reply 139. Comment on: Pantalone et al. The risk of overall mortality in patients with type 2 diabetes receiving glipizide, glyburide, or glimepiride monotherapy: a retrospective analysis. Diabetes Care. 2010;33:1224-1229.
45. Proks P, Reimann F, Green N, Gribble F, Ashcroft F. Sulfonylurea stimulation of insulin secretion. Diabetes. 2002;51(suppl 3):S368-S376.
46. Pantalone KM, Kattan MW, Wells BJ, Zimmerman RS. Response to comment on: Pantalone et al. The risk of overall mortality in patients with type 2 diabetes receiving glipizide, glyburide, or glimepiride monotherapy: a retrospective analysis. Diabetes Care. 2010;33:1224-1229. Diabetes Care. 2011;34(8):e139.
47. Berthet S, Olivier P, Montastruc JL, Lapeyre-Mestre M. Drug safety of rosigli tazone and pioglitazone in France: a study using the French PharmacoVigilance database. BMC Clin Pharmacol. 2011;11:5.
48. Dormandy J, Bhattacharya M, van Troostenburg de Bruyn AR. Safety and tolerability of pioglitazone in high-risk patients with type 2 diabetes: an overview of data from PROactive. Drug Saf. 2009;32:187-202.
49. Hillaire-Buys D, Faillie JL, Montastruc JL. Pioglitazone and bladder cancer. Lancet. 2011;378:1543-1544.
50. Azoulay L, Yin H, Filion KB, et al. The use of pioglitazone and the risk of bladder cancer in people with type 2 diabetes: nested case-control study. BMJ. 2012;344:e3645.
51. Dormandy JA, Charbonnel B, Eckland DJ, et al. Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial. Lancet. 2005;366:1279-1289.


Keywords: observational study; oral glucose-lowering drug; randomized clinical trial; type 2 diabetes