Diabetes Care Volume 38, July 2015
e108
RESPONSE TO COMMENT ON THOMSEN ET AL.
Incretin-Based Therapy and Risk of Acute Pancreatitis: A Nationwide PopulationBased Case-Control Study. Diabetes Care 2015;38:1089–1098
Reimar Wernich Thomsen,1 Lars Pedersen,1 Niels Møller,2 Johnny Kahlert,1 Henning Beck-Nielsen,3 and Henrik Toft Sørensen1
e-LETTERS – COMMENTS AND RESPONSES
Diabetes Care 2015;38:e108–e109 | DOI: 10.2337/dc15-0675
We thank Smits et al. (1) for their interest and kind comments on our recent study in Diabetes Care on the risk of acute pancreatitis (AP) associated with use of incretin-based therapies and other glucose-lowering drugs (2). We generally agree with our Dutch colleagues’ thoughtful and cautious comments. We would rather not make any firm conclusionsdlet alone health care guidelinesdbased on a single study. Nonetheless, several large epidemiological studies have yielded remarkably similar results by now, finding no evidence for a specific association between incretin-based therapies and AP, with relative risks of AP close to 1.0 when comparing users of incretinbased and other glucose-lowering drugs (3). Remarkably, metformin, with its different mode of glucose-lowering action, conferred the same risk for AP as did incretin-based therapies in our study, pointing to a general diabetes effect on AP (2). Smits et al. (1) question whether the number of AP events in our study was sufficient to adequately control for confounding, without hampering statistical power. We were able to include 89 AP events in incretin-based
therapy patients, compared with 29 AP events in dipeptidyl peptidase-4 inhibitor therapy patients in the large Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 (SAVOR-TIMI 53) and Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trials together (4). Our statistical models proved robust with 11 different confounding factors. Statistical precision was as indicated by the CIs; i.e., the adjusted odds ratio for AP was 0.95 (95% CI 0.75–1.21) for incretin-based versus no glucose-lowering therapy and the adjusted odds ratio was 0.97 (95% CI 0.76–1.23) for incretin-based versus other glucose-lowering therapyda rather robust null result. Smits et al. (1) are generally concerned about the “residual confounding that is inherent to the nature of database studies,” but they do not comment on the likely nature and direction of such possible confounding in the present association. Did we include the right confounders, and were they appropriately measured? Important confounders, such as obesity, gallstones, and comorbidities, measured through medical registries clearly increased AP risk
in our analyses (2). These factors were also positively associated with incretinbased therapy. Thus, adjustment for these factors reduced the AP risk estimates associated with incretin-based therapy toward the null. Therefore, we find it unlikely that further adjustment for uncontrolled confounding would have moved the AP risk estimate in the opposite direction, away from the null, for incretin-based therapies. We believe we should embrace different methodologies to study adverse drug effects. Randomized clinical trials are not perfect, even in concept (5)din fact, they are subject to various errors, including underestimation of adverse effects due to short duration, nonadherence to treatment, undercounting of outcomes, and inclusion of healthier patients than those treated in real life. Population-based studies like ours have the advantage of including all people, including those who may have the largest risk of APdmultimorbid, frail individuals and those who abuse alcohol and drugs. We look forward to the pooled results from the ongoing cardiovascular safety trials. In the meantime, large and carefully designed observational studies based on multinational
1
Department of Clinical Epidemiology, Institute of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark Medical Department M (Endocrinology and Diabetes) and Institute of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark 3 The Danish Centre for Strategic Research in Type 2 Diabetes, Diabetes Research Centre, Department of Endocrinology, Odense University Hospital, Odense, Denmark 2
Corresponding author: Reimar Wernich Thomsen, [email protected]. © 2015 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.
care.diabetesjournals.org
existing databases could provide further safety evidence. Duality of Interest. H.B.-N. received research grants from Novo Nordisk. The Department of Clinical Epidemiology is involved in studies with funding from various companies as research grants to (and administered by) Aarhus University. The Department of Clinical Epidemiology is a member of The Danish Centre for Strategic Research in Type 2 Diabetes (DD2), supported by the Danish Agency for Science, Technology and Innovation (grants 09067009 and 09-075724). DD2 is also supported by the Danish Health and Medicines Authority, the
Thomsen and Associates
Danish Diabetes Association, and an unrestricted donation from Novo Nordisk A/S. Partners in the DD2 project are listed on the project website at www.DD2.nu. No other potential conflicts of interest relevant to this article were reported.
References 1. Smits MM, Muskiet MHA, Tonneijck L, van Raalte DH. Comment on Thomsen et al. Incretin-based therapy and risk of acute pancreatitis: a nationwide population-based case-control study. Diabetes Care 2015;38:1089–1098 (Letter). Diabetes Care 2015;38: e106–e107. DOI: 10.2337/dc15-0367 2. Thomsen RW, Pedersen L, Møller N, Kahlert J, Beck-Nielsen H, Sørensen HT. Incretin-based
therapy and risk of acute pancreatitis: a nationwide population-based case-control study. Diabetes Care 2015;38:1089–1098 3. Li L, Shen J, Bala MM, et al. Incretin treatment and risk of pancreatitis in patients with type 2 diabetes mellitus: systematic review and meta-analysis of randomised and nonrandomised studies. BMJ 2014;348:g2366 4. Schernthaner G, Sattar N. Lessons from SAVOR and EXAMINE: some important answers, but many open questions. J Diabetes Complications 2014;28:430–433 5. Rothman KJ. Six persistent research misconceptions. J Gen Intern Med 2014;29: 1060–1064
e109
e108
RESPONSE TO COMMENT ON THOMSEN ET AL.
Incretin-Based Therapy and Risk of Acute Pancreatitis: A Nationwide PopulationBased Case-Control Study. Diabetes Care 2015;38:1089–1098
Reimar Wernich Thomsen,1 Lars Pedersen,1 Niels Møller,2 Johnny Kahlert,1 Henning Beck-Nielsen,3 and Henrik Toft Sørensen1
e-LETTERS – COMMENTS AND RESPONSES
Diabetes Care 2015;38:e108–e109 | DOI: 10.2337/dc15-0675
We thank Smits et al. (1) for their interest and kind comments on our recent study in Diabetes Care on the risk of acute pancreatitis (AP) associated with use of incretin-based therapies and other glucose-lowering drugs (2). We generally agree with our Dutch colleagues’ thoughtful and cautious comments. We would rather not make any firm conclusionsdlet alone health care guidelinesdbased on a single study. Nonetheless, several large epidemiological studies have yielded remarkably similar results by now, finding no evidence for a specific association between incretin-based therapies and AP, with relative risks of AP close to 1.0 when comparing users of incretinbased and other glucose-lowering drugs (3). Remarkably, metformin, with its different mode of glucose-lowering action, conferred the same risk for AP as did incretin-based therapies in our study, pointing to a general diabetes effect on AP (2). Smits et al. (1) question whether the number of AP events in our study was sufficient to adequately control for confounding, without hampering statistical power. We were able to include 89 AP events in incretin-based
therapy patients, compared with 29 AP events in dipeptidyl peptidase-4 inhibitor therapy patients in the large Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 (SAVOR-TIMI 53) and Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trials together (4). Our statistical models proved robust with 11 different confounding factors. Statistical precision was as indicated by the CIs; i.e., the adjusted odds ratio for AP was 0.95 (95% CI 0.75–1.21) for incretin-based versus no glucose-lowering therapy and the adjusted odds ratio was 0.97 (95% CI 0.76–1.23) for incretin-based versus other glucose-lowering therapyda rather robust null result. Smits et al. (1) are generally concerned about the “residual confounding that is inherent to the nature of database studies,” but they do not comment on the likely nature and direction of such possible confounding in the present association. Did we include the right confounders, and were they appropriately measured? Important confounders, such as obesity, gallstones, and comorbidities, measured through medical registries clearly increased AP risk
in our analyses (2). These factors were also positively associated with incretinbased therapy. Thus, adjustment for these factors reduced the AP risk estimates associated with incretin-based therapy toward the null. Therefore, we find it unlikely that further adjustment for uncontrolled confounding would have moved the AP risk estimate in the opposite direction, away from the null, for incretin-based therapies. We believe we should embrace different methodologies to study adverse drug effects. Randomized clinical trials are not perfect, even in concept (5)din fact, they are subject to various errors, including underestimation of adverse effects due to short duration, nonadherence to treatment, undercounting of outcomes, and inclusion of healthier patients than those treated in real life. Population-based studies like ours have the advantage of including all people, including those who may have the largest risk of APdmultimorbid, frail individuals and those who abuse alcohol and drugs. We look forward to the pooled results from the ongoing cardiovascular safety trials. In the meantime, large and carefully designed observational studies based on multinational
1
Department of Clinical Epidemiology, Institute of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark Medical Department M (Endocrinology and Diabetes) and Institute of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark 3 The Danish Centre for Strategic Research in Type 2 Diabetes, Diabetes Research Centre, Department of Endocrinology, Odense University Hospital, Odense, Denmark 2
Corresponding author: Reimar Wernich Thomsen, [email protected]. © 2015 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.
care.diabetesjournals.org
existing databases could provide further safety evidence. Duality of Interest. H.B.-N. received research grants from Novo Nordisk. The Department of Clinical Epidemiology is involved in studies with funding from various companies as research grants to (and administered by) Aarhus University. The Department of Clinical Epidemiology is a member of The Danish Centre for Strategic Research in Type 2 Diabetes (DD2), supported by the Danish Agency for Science, Technology and Innovation (grants 09067009 and 09-075724). DD2 is also supported by the Danish Health and Medicines Authority, the
Thomsen and Associates
Danish Diabetes Association, and an unrestricted donation from Novo Nordisk A/S. Partners in the DD2 project are listed on the project website at www.DD2.nu. No other potential conflicts of interest relevant to this article were reported.
References 1. Smits MM, Muskiet MHA, Tonneijck L, van Raalte DH. Comment on Thomsen et al. Incretin-based therapy and risk of acute pancreatitis: a nationwide population-based case-control study. Diabetes Care 2015;38:1089–1098 (Letter). Diabetes Care 2015;38: e106–e107. DOI: 10.2337/dc15-0367 2. Thomsen RW, Pedersen L, Møller N, Kahlert J, Beck-Nielsen H, Sørensen HT. Incretin-based
therapy and risk of acute pancreatitis: a nationwide population-based case-control study. Diabetes Care 2015;38:1089–1098 3. Li L, Shen J, Bala MM, et al. Incretin treatment and risk of pancreatitis in patients with type 2 diabetes mellitus: systematic review and meta-analysis of randomised and nonrandomised studies. BMJ 2014;348:g2366 4. Schernthaner G, Sattar N. Lessons from SAVOR and EXAMINE: some important answers, but many open questions. J Diabetes Complications 2014;28:430–433 5. Rothman KJ. Six persistent research misconceptions. J Gen Intern Med 2014;29: 1060–1064
e109
