Cardiovascular safety of combination therapies with incretin-based drugs and metformin compared with a combination of metformin and sulphonylurea in type 2 diabetes mellitus – a retrospective nationwide study U. M. Mogensen1 , C. Andersson2 , E. L. Fosbøl1 , T. K. Schramm3 , A. Vaag4 , N. M. Scheller 7 , C. Torp-Pedersen7 , G. Gislason2,5,6 & L. Køber1 1 Department of Cardiology, University Hospital Rigshospitalet, Copenhagen, Denmark 2 Department of Cardiology, University Hospital Gentofte, Copenhagen, Denmark 3 Department of Cardiology, University Hospital Frederiksberg, Copenhagen, Denmark 4 Department of Endocrinology, University Hospital Rigshospitalet, Copenhagen, Denmark 5 Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark 6 National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark 7 Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
Aim: Dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide-1 (GLP-1) agonists are widely used in combinations with metformin in the treatment of type 2 diabetes; however, data on long-term safety compared with conventional combination therapies are limited. Methods: Danish individuals without prior myocardial infarction or stroke that initiated combinations of metformin with sulphonylurea (SU), DPP-4 inhibitors, GLP-1 agonists or insulin between 9 May 2007 and 31 December 2011 were followed up for the risk of all-cause mortality, cardiovascular (CV) mortality or a combined end point of myocardial infarction, stroke and CV mortality. Rate ratios (RR) were calculated using time-dependent multivariable Poisson regression analysis. Results: A total of 40 028 patients (59% men, mean age 60 ± 13 years) used metformin with SU (n = 25 092), DPP-4 inhibitor (n = 11 138), GLP-1 agonist (n = 4345) or insulin (n = 6858). Crude incidence rates per 1000 patient years for the combined end point were 18 (SU), 10 (DPP-4 inhibitor), 8 (GLP-1 agonist) and 21 (insulin). In adjusted analyses with metformin + SU as reference, metformin + DPP-4 inhibitor was associated with an RR of 0.65 (0.54–0.80) for mortality, an RR of 0.57 (0.40–0.80) for CV mortality and an RR of 0.70 (0.57–0.85) for the combined end point. For metformin + GLP-1 agonist, the RR for mortality was 0.77 (0.51–1.17), for CV mortality 0.89 (0.47–1.68), and for the combined end point 0.82 (0.55–1.21). Conclusion: Incretin-based drugs combined with metformin were safe compared with conventional combinations of glucose-lowering therapy. Use of incretin-based therapy may be target for strategies to lower CV risk in type 2 diabetes, although it should be recognized that the multivariable analysis may not have fully accounted for important baseline differences. Keywords: cardiovascular disease, DPP-IV inhibitor, GLP-1 analogue, incretin therapy, type 2 diabetes Date submitted 6 December 2013; date of first decision 5 January 2014; date of final acceptance 2 May 2014
Introduction Dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide 1 (GLP-1) agonists have recently been introduced in the treatment of type 2 diabetes mellitus (T2DM). These incretinbased drugs are widely used in combination with metformin, but data on their cardiovascular disease (CVD) risk compared with conventional combination therapy regimens [metformin with sulphonylurea (SU) or insulin] remain limited.
Correspondence to: Ulrik Madvig Mogensen, MD, Department of Cardiology, The Heart Centre, University Hospital Rigshospitalet, 9441, Blegdamsvej 9, 2100 Copenhagen, Denmark. E-mail: [email protected]
Metformin is generally recommended as first-line therapy , but the majority of T2DM patients eventually need more than one glucose-lowering therapy (GLT) to maintain glycaemic control . SUs have been the most widely used add-on therapy for several years, but controversies exist with respect to their individual or general CVD risk when administered alone or in combination with metformin [3–7]. Thus, the optimal strategy for glucose-lowering treatment as add-on to metformin as second-line therapy is not clear [1,8]. Compared with SU and insulin, incretin-based therapies have a low risk of hypoglycaemia and favourable or neutral effects on body weight . Furthermore, incretin-based therapies have been associated with positive cardiovascular effects beyond glucose lowering [9–12]. Two recent randomized
trials reported neutral composite CVD outcome results of adding a DPP-4 inhibitor to standard therapy in T2D patients at high risk for CV events [13,14]. However, data for direct comparisons of the most commonly used combinations with metformin treatment are lacking. In this study, we aimed to compare mortality and CVD risk associated with combinations of incretin-based therapy and metformin versus the two other most widely used secondline combination therapies in low-risk T2DM patients without previous CV events.
Methods Registry Data Sources All residents in Denmark are assigned a unique civil registration number at birth enabling linkage of nationwide administrative registries at an individual level. For this study, we used four different registries. We obtained information on hospitalizations from The Danish National Patient Registry, which holds information on all admissions in Denmark since 1978 with discharge diagnoses coded according to the International Classification of Diseases (ICD) system, ICD-8 before and ICD-10 after 1994. All prescriptions dispensed from Danish Pharmacies have been consistently registered according to the Anatomical Therapeutic Chemical (ATC) system in the Danish Registry of Medicinal Product Statistics since 1995. We used information on amount and strength of dispensed tablets as well as date of dispensing, which has been shown to be accurate, to define treatment exposures . We obtained information on vital status from The Danish National Population Registry, where all deaths of Danish residents are registered within 2 weeks of occurrence. Causes of death were obtained from the National Causes of Death Register, which holds information on primary and contributing causes of death according to ICD-10.
Patient Population All individuals ≥18 years who initiated a combination of GLT after 9 May 2007 were identified. Users of GLT before 1 January 1997 with unknown treatment duration, and patients with prior myocardial infarction (ICD-10: I21-I22, ICD-8: 410) or stroke (ICD-10: I61-I64, ICD-8: 431–434) were excluded. Patients using metformin in combination with SU, DPP-4 inhibitors, GLP-1 agonists or insulin were included at date of first claimed prescription of combined therapy (Figure 1). Danish physicians generally comply with guidelines issued from EASD/ADA [1,16].
Outcomes All-cause mortality, CV mortality (ICD-10: I00-I99) and a combined end point of myocardial infarction (I21-I22), stroke (I61-I64) or CV mortality were assessed.
Figure 1. Flow chart of patient population. MI, myocardial infarction.
tolbutamide (A10BB03) and glipizide (A10BB07)], DPP-4 inhibitors [sitagliptin (A10BH01 and A10BD07), vildagliptin (A10BH02 and A10BD08) and saxagliptin (A10BH03)] and GLP1-analogues [exenatide (A10BX04) and liraglutide (A10BX07)]. Intervals of treatment with each type of GLT were calculated by dividing the amount of claimed medication by time elapsed between prescription claims. A continuous treatment period was assumed if this was compatible with at least the minimal dose when considering up to four consecutively claimed prescriptions (in a retrospective manner to avoid conditioning on the future) [17–19]. All intervals were extended with 30 days to avoid artificial breaks in the calculated treatment durations. A sensitivity analysis without this extension was performed.
Co-morbidity, Concomitant CV Therapy and Income Co-morbidities were included as Charlson co-morbidity score based on discharge diagnoses 10 year prior to baseline [20,21]. An additional analysis was conducted using the following CV risk factors: peripheral vascular disease, ischaemic heart disease, cerebrovascular disease, renal disease atrial fibrillation and congestive heart failure. Concomitant CV therapy was included in the models as time-dependent dichotomized variables indicating use of the respective drugs within the last year. These drugs included statins (C10AA), angiotensin-converting enzyme inhibitors/angiotensin receptor inhibitors (renin angiotensin system inhibitor, RASi) (C09), β-blockers (C07), calcium channel blockers (C08), other antihypertensive drugs (C02), nitrates (C01D), aspirin (B01AC06), vitamin K antagonists (B01AA), digoxin (C01AA) and spironolactone (C03D). Household income among all included patients 5 years prior to baseline was graded in quintiles and included in the analysis to account for socio-economic status.
Statistics Exposure Classiﬁcation All types of GLTs were classified according to drug type including metformin (A10B), insulin (A10A), SUs [glibenclamide (A10BB01), gliclazide (A10BB09), glimepiride (A10BB12),
2 Mogensen et al.
sas version 9.2 (SAS Institute, Cary, NC, USA) was used for statistical analysis. A p-value of <0.05 was considered statistically significant for all analyses. A multivariable timedependent Poisson regression model was used to compare rate
DIABETES, OBESITY AND METABOLISM
ratios (RR) associated with individual combination therapies, using SU + metformin as reference. Follow-up started at the date of a first claimed prescription of a combination of metformin with either SU, insulin, DPP-4 inhibitor or GLP1 agonist and ended when reaching a relevant end point or at 31 December 2011. During exposure to a relevant combination therapy, the patients contributed with risk time to this particular combination. If changing therapy, patients contributed risk time to the new therapy accordingly. Outcome events were attributed to the therapy the patient was receiving at the time of the event. All models included age, sex, duration of GLT (from first claimed prescription of any glucoselowering drug), calendar year, Charlson score, concomitant CV pharmacotherapy and income. Assumptions of the Poisson model were tested and found valid. As a mean to balance baseline differences, matched groups based on the propensity to receive SU + metformin were created. In the propensity-matched models, only first-time users of the individual dual combination regimens were included, and the analyses were performed as intentionto-treat. Matching was performed using the greedy match macro (Mayo Clinic College of Medicine, http://ndc.mayo. edu/mayo/research/biostat/upload/gmatch.sas, last accessed November 2012), using a calliper of 0.02 on the probability scale. All variables in Table 1 were used in the logistic regression model. The C-statistics for propensity score models was 0.69 for DPP-4 + metformin, 0.71 for insulin + metformin and 0.86 for insulin + metformin.
Ethics The study was approved by the Danish Data Protection Agency (J. no. 2007-58-015I; suite no. 00916 GEH-2010-001). In Denmark, retrospective register studies do not need ethical approval.
Results Baseline Characteristics A total of 40 028 patients were included in the study (flow chart, Figure 1). The study population had a mean age of 60.6 ± 13 years, 59% were male and the median interquartile range (IQR) duration of monotherapy prior to initiating combination therapy was 2.0 years (0.4–4.5 years). Baseline characteristics according to combination therapy are presented in Table 1. Patients initiating SU + metformin were older, whereas GLP1 + metformin generally was used among younger patients. Prior to initiating dual combination therapy, patients using DPP-4 inhibitors and GLP agonists had more often used metformin as the initial GLT, and a larger proportion had exclusively used metformin compared with patients initiating SU or insulin in combination with metformin. Users of insulin + metformin were younger than average, had more co-morbidity and had the shortest duration of prior GLT when initiating metformin combination therapy.
Prescription Patterns During the Study Period There was an increase in use of both DPP-4 inhibitors and GLP1 agonists during the study period, and they comprised 25
and 12%, respectively, of dual GLTs used in 2011. At the same time, the use of SU + metformin therapy declined from 68 to 48% of the chosen combination therapies (Figure 2). Both DPP-4 inhibitors and GLP-1 agonists were used primarily as add-on therapy; 95.0 and 96.4%, respectively, of the cumulative exposure time was in combination with one or more glucoselowering drugs. Among patients using incretins in combinations with metformin, the most widely used DPP-4 inhibitor was sitagliptin (n = 7116) compared with vildagliptin (n = 3787) and saxagliptin (n = 586). The most widely used GLP-1 agonist was liraglutide (n = 4137).
Mortality and CV Risk Patients were followed up for a median of 2.1 years (IQR 1.0–3.3 years), 2468 (6.2%) died and 1645 (4.1%) reached the combined end point of myocardial infarction (MI), stroke or CV death. Crude incidence rates for all-cause mortality, CV death and the combined end point are presented in Table 2. In combinations with metformin, incretin-based therapies were associated with overall lower event rates compared with insulin and SU. RRs from multivariable analyses are presented in Figure 3. Compared with SU, DPP-4 inhibitor was associated with a significantly lower risk of all end points when used in combinations with metformin. GLP-1 agonist in combination with metformin was associated with lower risk estimates, but this was not significant in adjusted analyses. The difference in outcome between DPP-4 inhibitor and GLP-1 agonist-based therapy was not significant; p value for difference was 0.47 for mortality, 0.21 for CV mortality and 0.46 for the combined end point.
Additional Analyses A number of sensitivity analyses were performed. Including only users of the initial combination therapy (censoring patients when changing therapy) and using metformin + SU as reference, metformin + DPP-4 inhibitor was associated with an RR of 0.64 (0.51–0.80) for mortality, an RR of 0.55 (0.37–0.81) for CV death and an RR of 0.70 (0.56–0.87) for the combined end point. GLP-1 agonist + metformin was associated with an RR of 0.84 (0.49–1.47) for mortality, an RR of 1.21 (0.58–2.58) for CV death and an RR of 0.70 (0.40–1.25) for the combined end point. Insulin + metformin was associated with an RR of 2.15 (1.83–2.51) for mortality, an RR of 1.74 (1.32–2.31) for CV death and an RR of 1.24 (1.0–1.55) for the combined end point. Including only patients who had exclusively used metformin prior to initiation of combination therapy, using metformin + SU as reference, the RRs for mortality were DPP-4 inhibitors [RR = 0.69 (0.54–0.88)], GLP-1 agonists [RR = 0.89 (0.57–1.42)] and insulin [RR = 2.53 (2.09–3.05)]; the RR for CV death were DPP-4 inhibitors [RR = 0.61 (0.40–0.93)], GLP-1 agonists [RR = 1.15 (0.6–2.24)] and insulin [RR = 1.85 (1.31–2.62)]; the RRs for the combined end points were DPP-4 inhibitors [RR = 0.74 (0.58–0.94)], GLP1 agonists [RR = 1.02 (0.68–1.53)] and insulin [RR = 1.23 (0.94–1.60)].
DIABETES, OBESITY AND METABOLISM
Table 1. Baseline characteristics. Metformin in combination with Population (n) Men, n (%) Age, mean ± s.d. Treatment duration in years, median (p25–p75) Co-morbiditites Congestive heart failure, n (%) Ischaemic heart disease, n (%) Atrial fibrillation, n (%) Renal disease, n (%) Peripheral vascular disease, n (%) COPD, n (%) Peptic ulcer, n (%) Cancer, n (%) Charlson score 0, n (%) 1–2, n (%) ≥3, n (%) Concomitant therapy Statin, n (%) RASi, n (%) β-Blockers, n (%) Aspirin, n (%) Thiazid, n (%) Calcium antagonist, n (%) Other antihypertensive, n (%) Nitrates, n (%) Vitamin K antagonist, n (%) Digoxin, n (%) Spironolactone, n (%) Inclusion year 2007, n (%) 2008, n (%) 2009, n (%) 2010, n (%) 2011, n (%) Gross income, quintile 1, n (%) 2, n (%) 3, n (%) 4, n (%) 5, n (%) GLT prior to combination therapy GLT at incident diabetes, n (%) Metformin monotherapy, n (%) Sulphonylurea monotherapy, n (%) Insulin monotherapy, n(%) Dual combination therapy, n(%) Any prior use of metformin, n (%) Prior exclusive use of metformin, n (%) Any prior use of SU, n (%) Prior exclusive use of SU, n (%) Prior use of a combination of GLT, n (%)
As some patients changed therapy during follow-up, the number of patients at baseline in each treatment adds up to more than the total. Gross income was categorized in quintiles for all diabetes patients in the population. A total of 235 patients (0.59%) did not have information on prior income at the time of inclusion. SU, sulphonylurea; DPP-4, dipeptidyl peptidase-4 inhibitors; GLP-1, glucagon-like peptide-1 agonists; s.d., standard deviation; COPD, chronic obstructive pulmonary disease; RASi, renin angiotensin system inhibitors; GLT, glucose-lowering therapy.
Considering breaks in the calculated treatment durations of <30 days as real discontinuations in treatment, using metformin + SU as reference, the RRs for mortality were DPP-4 inhibitors [RR = 0.64 (0.52–0.79)], GLP-1 agonists [RR = 0.70 (0.45–1.10)] and insulin [RR = 1.79 (1.54–2.09)]; the RR for
4 Mogensen et al.
CV death were DPP-4 inhibitors [RR = 0.57 (0.40–0.87)], GLP-1 agonists [RR = 0.88 (0.47–1.68)] and insulin [RR = 1.49 (1.14–1.94)]; the RRs for the combined end points were DPP-4 inhibitors [RR = 0.66 (0.53–0.81)], GLP-1 agonists [RR = 0.79 (0.54–1.17)] and insulin [1.12 (0.92–1.37)].
DIABETES, OBESITY AND METABOLISM
Figure 2. Incident use of dual glucose-lowering therapies in Denmark 2007–2011. Incident use is defined as the combination therapy used among patients initiating a combination of glucose-lowering therapy (GLT) for the first time. SU, sulphonylureas; TZD, thiazolidinediones (94% rosiglitazone, 6% pioglitazone); DPP-4, dipeptidyl peptidase-4 inhibitors; GLP-1, glucagon-like peptide 1 agonists.
Table 2. Crude incidence rates for all-cause mortality, cardiovascular (CV) death and a combined end point of myocardial infarction, stroke and CV death. Combination of metformin with
All-cause mortality Events Person years Crude incidence rates CV mortality Events Person years Crude incidence rates Combined end point Events Person years Crude incidence rates
644 31 497 20 (19–22)
118 11 619 10 (8–12)
23 3246 7 (4–11)
304 6916 44 (39–49)
256 31 497 8 (7–9)
38 11 619 3 (2–5)
10 3246 3 (2–6)
91 6916 13 (11–16)
570 31 189 18 (17–20)
118 11 552 10 (9–12)
27 3230 8 (6–12)
146 6828 21 (18–25)
Crude incidence rates represent numbers of events per 1000 person years (95% confidence interval). SU, sulphonylurea; GLP-1, glucagon-like peptide 1 agonist; DPP-4, dipeptidyl peptidase 4 inhibitor.
Propensity-Score-Matched Analysis Baseline characteristics of the propensity-score-matched groups are presented in Table S1, Supporting Information. Including only these matched groups of patients in the multivariable model analysed as intention-to-treat compared with SU + metformin, RRs associated with other combinations with metformin were as follows: for mortality, DPP-4 inhibitors [RR = 0.85 (0.71–1.02), p = 0.08], GLP-1 agonists [RR = 0.70 (0.40–1.23), p = 0.70] and insulin [RR = 1.93 (1.65–2.25), p < 0.0001]; and for the combined end point, DPP-4 inhibitors [RR = 0.89 (0.73–1.09), p = 0.27], GLP-1 agonists [RR = 0.64 (0.34–1.20), p = 0.17] and insulin [1.28 (1.03–1.59), p = 0.03].
Figure 3. Rate ratios (RR) and 95% confidence intervals (95% CI) associated with the use of metformin in combinations with DPP-4 inhibitors, GLP-1 agonists or insulin compared with a combination of metformin and SU for mortality, cardiovascular (CV) mortality and combined end point of acute myocardial infarction, stroke or CV mortality. DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide 1; SU, sulphonylurea.
Discussion In this nationwide study, we examined mortality and CV events associated with the use of metformin combined with DPP-4 inhibitors or GLP1-receptor agonists versus the more conventional combinations of metformin with SU or insulin. Overall, we found no signs of increased mortality or CV events associated with the use of metformin combined with incretinbased combination therapies. In adjusted analyses, DPP-4 inhibitors were associated with lower risks of both mortality and CV end points. Results were similar in a number of sensitivity analyses, but the lower risk associated with DPP-4 inhibitors did not reach significant levels in propensity-score-matched intention-to-treat analysis. Our findings of no increased risk of CV events associated with DPP-4 inhibitors are consistent with two previous studies in a ‘real world setting’ [22,23] and two recent large clinical trials [13,14]. The Cardiovascular Outcomes Study of Alogliptin in Subjects With Type 2 Diabetes and Acute Coronary Syndrome (EXAMINE)  found that the addition
original article of the DPP-4 inhibitor alogliptin to standard care did not result in an increase in CV risk among 5380 patients with a recent acute coronary syndrome followed up for a median of 18 months. In the randomized trial, Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus (SAVOR TIMI-53) , the DPP-4 inhibitor saxagliptin had neutral effects on CV events when added to standard care in 16 492 high-risk patients followed up for a median of 2.1 years. In contrast, saxagliptin was associated with a small but significantly increased risk of heart failure. In this study, we focused on the risk of CV events and mortality and did not include heart failure as an end point, as the sensitivity of the heart failure diagnose has been found low in the registries . The observed trend of lower risks associated with the use of DPP-4 inhibitors combined with metformin in this study might be explained by the inclusion of patients at low CV risk with short diabetes treatment duration and the use of SU combined with metformin as reference, instead of investigating the effect of adding DPP-4 inhibitors to standard GLT as in SAVOR TIMI-53  and EXAMINE . Prior to the recent randomized trials, positive effects on mortality and CV end points associated with the use of DPP-4 inhibitor were reported in pooled analyses of short- and medium-term clinical trials [25,26]. In a meta-analysis including 70 trials with a mean follow-up of 44 weeks comparing DPP-4 inhibitors with placebo or other drugs, DPP-4 inhibitors were associated with a lower risk of major CV event [odds ratio (OR) = 0.71 (0.59–0.86)], myocardial infarction [OR = 0.64 (0.44–0.84)] and mortality [OR = 0.60 (0.41–0.88)] . Only few other studies have investigated mortality and CV outcomes associated with the use of DPP-4 inhibitors in combination with metformin compared with other combinations of GLT. In a retrospective population-based study, combination of metformin with the DPP-4 inhibitor sitagliptin was associated with a lower risk of a composite end point of allcause hospital admission and all-cause mortality [HR = 0.82 (0.72–0.93)] when compared with metformin + SU . In a randomized trial comparing addition of metformin to a DPP-4 inhibitor (linagliptin) against an SU (glimepiride), during 2 years of follow-up, metformin + DPP-4 inhibitor was associated with significantly fewer CV events [12 vs. 26; relative risk 0.46 (95% CI 0.23-0.91), p = 0.0213] . This is consistent with our results, but whether these findings could be the result of adverse effects associated with the combination of SU + metformin, as suggested by the UK Prospective Diabetes Study , or favourable effects associated with the metformin + DPP-4 inhibitor compared with other combinations deserves further study. During our study period, the major DPP-4 inhibitor used was sitagliptin, while fewer used vildagliptin and saxagliptin. Previous studies have not suggested important differences in efficacy between DPP-4 inhibitors including sitagliptin and saxagliptin  and conclusions on whether sitagliptin might be associated with improved outcomes awaits results from an ongoing clinical trial . Due to less numbers of people using especially saxagliptin, we did not include tests for differences between individual DPP-4 inhibitors in this study.
6 Mogensen et al.
DIABETES, OBESITY AND METABOLISM
DPP-4 inhibitor tended to be associated with lower risk estimates compared with GLP-1 agonist, but the observed differences were not statistically significant. GLP-1 agonist have proven more effective than DPP-4 inhibitors in terms of reducing HbA1c and weight [30–32], but current data do not support a superiority of GLP-1 agonists over DPP-4 inhibitors in terms of long-term outcomes . It can be speculated whether the insignificant differences in the risk estimates between GLP-1 agonists and DPP-4 inhibitors might be due to differences in obesity among patients using each treatment, as we could not adjust for body mass index, or that GLP-1 agonists were used at more advanced stages of disease after failure of oral therapy. Very limited data exist on mortality associated with the use of GLP-1 agonists. In a meta-analysis on trials with a duration of at least 6 months, the difference in the incidence of major CV events between GLP agonists and comparators did not reach statistical significance [OR = 0.78, [0.54–1.13)] . In a retrospective analysis on 39 275 patients, exenatide was associated with a lower risk of CV events (hazard ratio 0.81; 95% CI 0.68–0.95, p = 0.01) compared with other GLT. This was in spite of the fact that patients using exenatide were found to have more severe diabetes and a more adverse CV risk profile at baseline . Whether incretin-based therapies are associated with a differential prognosis compared with other GLT will be further elucidated by means of additional ongoing prospective CV outcome studies on GLP-1 [36,37] and DPP-4 inhibitors [29,38].
Strengths and Limitations A major strength of this study is the completeness of the nationwide registers avoiding selection bias with a follow-up of several years. However, there are limitations inherent to the retrospective observational nature of the study, and we did not have information on a variety of clinical variables including HbA1c, hypertension, heart rate (which has been found to be raised by GLP-1), dyslipidaemia, smoking and body mass index. Instead, we adjusted for use of several concomitant CV therapies, co-morbidity and socio-economy. As with all observational studies, we were unable to provide firm evidence for direct treatment cause–effect relationships, and the risks of false-positive or -negative findings due to residual confounding, even after the most stringent corrections, may not be trivial. Hence, in propensityscore-matched intention-to-treat analysis, the lower risks associated with metformin + DPP-4 inhibitor compared with metformin + SU did not reach significant levels. On the other hand, the propensity-score-matched analysis had less statistical power compared with the time-dependent analysis based on the full cohort, and there remained a trend towards a lower risk associated with incretin-based therapy. DPP-4 inhibitors and GLP-1 agonists were used for shorter time periods, and this could potentially render their associated risk more favourable when compared with SU in combinations with metformin. However, all analyses were based on time-dependent analyses, taking into account the exposure time for each drug combination, and the combination of insulin with metformin had a
DIABETES, OBESITY AND METABOLISM
similar shorter time of use, but was associated with increased risk when compared with SU in combination with metformin. Finally, the possibility of non-random prescribing pattern leading to confounding by indication is difficult to overcome and cannot be excluded: patients with a more favourable clinical presentation of type 2 diabetes might be more likely to be treated with metformin + DPP-4 inhibitor, whereas patients with a higher burden of disease might be more likely to be treated with SU or insulin and indeed, patients using metformin + insulin tended to have more advanced disease.
Conclusion In this retrospective observational study, combination of incretin-based therapies with metformin was not associated with increased CV or mortality risk when compared with other dual combinations of GLTs. Randomized prospective studies are needed to establish whether incretin-based combination therapies may be associated with lower long-term risks compared with older classes of GLTs in low-risk patients.
Acknowledgements This study was supported by a grant from the Interreg IVA programme, a part of the European Union. Dr Gislason is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation.
Conﬂict of Interest U. M. M., C. A., E. L. F., T. K. S., N. M. S., C. T. P. and L. K. declared no conflicts of interest. A. V. owns stocks and has received honoraria for lectures as well as for consultancy services from Novo Nordisk. G. G. owns stocks in Novo Nordisk.
Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Baseline characteristics, propensity-score-matched groups.
References 1. Inzucchi SE, Bergenstal RM, Buse JB et al. Management of hyperglycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia 2012; 55: 1577–1596.
original article 5. 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: e1001204. 6. Hemmingsen B, Schroll JB, Lund SS et al. Sulphonylurea monotherapy for patients with type 2 diabetes mellitus. Cochrane Database Syst Rev 2013; 4: CD009008. 7. Meinert CL, Knatterud GL, Prout TE, Klimt CR. A study of the effects of hypoglycemic agents on vascular complications in patients with adultonset diabetes. II. Mortality results. Diabetes 1970; 19(Suppl): 789–830. 8. Bennett WL, Maruthur NM, Singh S et al. Comparative effectiveness and safety of medications for type 2 diabetes: an update including new drugs and 2-drug combinations. Ann Intern Med 2011; 154: 602–613. 9. Boland CL, Degeeter M, Nuzum DS, Tzefos M. Evaluating second-line treatment options for type 2 diabetes: focus on secondary effects of GLP-1 agonists and DPP-4 inhibitors. Ann Pharmacother 2013; 47: 490–505. 10. Ussher JR, Drucker DJ. Cardiovascular biology of the incretin system. Endocr Rev 2012; 33: 187–215. 11. Mannucci E, Dicembrini I. Incretin-based therapies and cardiovascular risk. Curr Med Res Opin 2012; 28: 715–721. 12. Scheen AJ. Cardiovascular effects of gliptins. Nat Rev Cardiol 2013; 10: 73–84. 13. Scirica BM, Bhatt DL, Braunwald E et al. Saxagliptin and cardiovascular outcomes in patients with type 2 diabetes mellitus. N Engl J Med 2013; 369: 1317–1326. 14. White WB, Cannon CP, Heller SR et al. Alogliptin after acute coronary syndrome in patients with type 2 diabetes. N Engl J Med 2013; 369: 1327–1335. 15. Gaist D, Sorensen HT, Hallas J. The Danish prescription registries. Dan Med Bull 1997; 44: 445–448. 16. Snorgaard O, Drivsholm TB, Breum L, et al. Guidelines for Type 2-Diabetes. Available from URL: http://www.endocrinology.dk/PDF/DiabetesFolder. pdf. Accessed 2 December 2013. 17. Fosbol EL, Gislason GH, Jacobsen S et al. The pattern of use of non-steroidal anti-inﬂammatory drugs (NSAIDs) from 1997 to 2005: a nationwide study on 4.6 million people. Pharmacoepidemiol Drug Saf 2008; 17: 822–833. 18. Olsen AM, Fosbol EL, Lindhardsen J et al. Cause-speciﬁc cardiovascular risk associated with nonsteroidal anti-inﬂammatory drugs among myocardial infarction patients – a nationwide study. PLoS One 2013; 8: e54309. 19. Andersson C, Vaag A, Selmer C et al. Risk of cancer in patients using glucose-lowering agents: a nationwide cohort study of 3.6 million people. BMJ Open 2012; 2: e000433. 20. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383. 21. Thygesen SK, Christiansen CF, Christensen S, Lash TL, Sorensen HT. The predictive value of ICD-10 diagnostic coding used to assess Charlson comorbidity index conditions in the population-based Danish National Registry of Patients. BMC Med Res Methodol 2011; 11: 83.
2. Turner RC, Cull CA, Frighi V, Holman RR. Glycemic control with diet, sulfonylurea, metformin, or insulin in patients with type 2 diabetes mellitus: progressive requirement for multiple therapies (UKPDS 49). UK Prospective Diabetes Study (UKPDS) Group. JAMA 1999; 281: 2005–2012.
22. Eurich DT, Simpson S, Senthilselvan A, Asche CV, Sandhu-Minhas JK, McAlister FA. Comparative safety and effectiveness of sitagliptin in patients with type 2 diabetes: retrospective population based cohort study. BMJ 2013; 346: f2267.
3. UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive bloodglucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet 1998; 352: 854–865.
23. Scheller NM, Mogensen UM, Andersson C, Vaag A, Torp-Pedersen C. All-cause mortality and cardiovascular effects associated with the DPPIVinhibitor sitagliptin compared with metformin, a retrospective cohort study on the Danish population. Diabetes Obes Metab 2013; 16: 231–236.
4. Lamanna C, Monami M, Marchionni N, Mannucci E. Effect of metformin on cardiovascular events and mortality: a meta-analysis of randomized clinical trials. Diabetes Obes Metab 2011; 13: 221–228.
24. Kumler T, Gislason GH, Kirk V et al. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail 2008; 10: 658–660.
DIABETES, OBESITY AND METABOLISM
25. Patil HR, Al Badarin FJ, Al Shami HA et al. Meta-analysis of effect of dipeptidyl peptidase-4 inhibitors on cardiovascular risk in type 2 diabetes mellitus. Am J Cardiol 2012; 110: 826–833.
33. Scheen AJ. Dipeptidylpeptidase-4 (DPP-4) inhibitors are favourable to glucagon-like peptide-1 (GLP-1) receptor agonists: yes. Eur J Intern Med 2012; 23: 126–131.
26. Monami M, Ahren B, Dicembrini I, Mannucci E. Dipeptidyl peptidase-4 inhibitors and cardiovascular risk: a meta-analysis of randomized clinical trials. Diabetes Obes Metab 2013; 15: 112–120.
34. Monami M, Dicembrini I, Nardini C, Fiordelli I, Mannucci E. Effects of glucagon-like peptide-1 receptor agonists on cardiovascular risk: a metanalysis of randomised clinical trials. Diabetes Obes Metab 2013; 16: 38–47.
27. Gallwitz B, Rosenstock J, Rauch T et al. 2-year efﬁcacy and safety of linagliptin compared with glimepiride in patients with type 2 diabetes inadequately controlled on metformin: a randomised, double-blind, noninferiority trial. Lancet 2012; 380: 475–483. 28. Scheen AJ, Charpentier G, Ostgren CJ, Hellqvist A, Gause-Nilsson I. Efﬁcacy and safety of saxagliptin in combination with metformin compared with sitagliptin in combination with metformin in adult patients with type 2 diabetes mellitus. Diabetes Metab Res Rev 2010; 26: 540–549. 29. Sitagliptin Cardiovascular Outcome Study (MK-0431-082) (TECOS), 2013. Available from URL: http://www.clinicaltrials.gov/ct2/show/NCT 00790205?term=tecos&rank=1. Accessed 8 August 2013. 30. Pratley RE, Nauck M, Bailey T et al. Liraglutide versus sitagliptin for patients with type 2 diabetes who did not have adequate glycaemic control with metformin: a 26-week, randomised, parallel-group, open-label trial. Lancet 2010; 375: 1447–1456. 31. Bergenstal RM, Wysham C, Macconell L et al. Efﬁcacy and safety of exenatide once weekly versus sitagliptin or pioglitazone as an adjunct to metformin for treatment of type 2 diabetes (DURATION-2): a randomised trial. Lancet 2010; 376: 431–439.
35. Best JH, Hoogwerf BJ, Herman WH et al. Risk of cardiovascular disease events in patients with type 2 diabetes prescribed the glucagon-like peptide 1 (GLP-1) receptor agonist exenatide twice daily or other glucoselowering therapies: a retrospective analysis of the LifeLink database. Diabetes Care 2011; 34: 90–95. 36. Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results – A Long Term Evaluation (LEADER). Available from URL: http://www.clinicaltrials.gov/ct2/show/NCT01179048?term=leader+lir aglutide&rank=1. Accessed 8 August 2013. 37. Exenatide Study of Cardiovascular Event Lowering Trial (EXSCEL): A Trial To Evaluate Cardiovascular Outcomes After Treatment With Exenatide Once Weekly In Patients With Type 2 Diabetes Mellitus. Available from URL: http://www.clinicaltrials.gov/ct2/show/NCT01144338?term=exscel&ra nk=1. Accessed 8 August 2013. 38. Rosenstock J, Marx N, Kahn SE et al. Cardiovascular outcome trials in type 2 diabetes and the sulphonylurea controversy: rationale for the active-comparator CAROLINA trial. Diab Vasc Dis Res 2013; 10: 289–301.
32. Scheen AJ. DPP-4 inhibitors in the management of type 2 diabetes: a critical review of head-to-head trials. Diabetes Metab 2012; 38: 89–101.
Cardiovascular safety of combination therapies with incretin-based drugs and metformin compared with a combination of metformin and sulphonylurea in type 2 diabetes mellitus--a retrospective nationwide study.
Dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide-1 (GLP-1) agonists are widely used in combinations with metformin in the treatment...
To evaluate the efficacy and safety of canagliflozin, a sodium glucose co-transporter 2 inhibitor, in Asian patients with type 2 diabetes mellitus (T2DM) inadequately controlled by metformin or metformin in combination with sulphonylurea.
Metformin is the first-line treatment for most patients with type 2 diabetes but many patients need additional treatment with insulin secretagogues (IS) to achieve glycemic control. We aimed to compare mortality and cardiovascular risk among users of
Long-term and high-dose treatment with metformin is known to be associated with vitamin B12 deficiency in patients with type 2 diabetes. We investigated whether the prevalence of B12 deficiency was different in patients treated with different combina
This study assessed the health costs resulting from the combination of metformin/dipeptidyl peptidase-4 (DPP-4) inhibitors compared with metformin/oral antidiabetes drugs in patients with type 2 diabetes and metabolic syndrome (MS).
Alogliptin is a selective dipeptidyl peptidase-4 inhibitor recently marketed for once-daily administration in the treatment of type 2 diabetes mellitus (T2DM). Fixed-dose combinations of alogliptin with both metformin and pioglitazone are also commer
Metformin is typically the first pharmacologic treatment recommended for type 2 diabetes mellitus (T2DM), but many patients do not achieve glycemic control with metformin alone and eventually require combination therapy with other agents. Canaglifloz
Type 2 diabetes (T2D) is a chronic and multifactorial metabolic disease, which brings great threats to public health. The morbidity of T2D keeps growing, and it is estimated that the population with T2D will rise to 552 million throughout the world b
Type 2 diabetes mellitus (T2DM) is a progressive condition requiring long-term treatment. Most patients with T2DM are unable to maintain normoglycemia using metformin alone; thus, combination therapy is a pivotal part of disease management. Addition
Metformin is a preferred drug for starting treatment in type 2 diabetes mellitus. But, eventually most of the patients need additional drug to control blood sugar level. The choice of drug depends upon several factors including patient specific crite
To evaluate the cost-effectiveness of vildagliptin plus metformin vs generic sulphonylurea plus metformin in patients with type 2 diabetes mellitus, not controlled with metformin, from a Portuguese healthcare system perspective.