Accepted Manuscript Title: Metformin in combination with various insulin secretagogues in type 2 diabetes and associated risk of cardiovascular morbidity and mortality – a retrospective nationwide study Author: Ulrik Madvig Mogensen MD Charlotte Andersson MD PhD Emil Loldrup Fosbøl MD PhD Tina Ken Schramm MD PhD Allan Vaag MD DMSci Nikolai Madrid Scheller MB Christian Torp-Pedersen MD DMSci Gunnar Gislason MD PhD Lars Køber MD DMSci PII: DOI: Reference:
S0168-8227(14)00447-1 http://dx.doi.org/doi:10.1016/j.diabres.2014.09.047 DIAB 6202
To appear in:
Diabetes Research and Clinical Practice
Received date: Revised date: Accepted date:
23-4-2014 21-8-2014 15-9-2014
Please cite this article as: U.M. Mogensen, C. Andersson, E.L. Fosbol, T.K. Schramm, A. Vaag, N.M. Scheller, C. Torp-Pedersen, G. Gislason, L. Kober, Metformin in combination with various insulin secretagogues in type 2 diabetes and associated risk of cardiovascular morbidity and mortality ndash a retrospective nationwide study, Diabetes Research and Clinical Practice (2014), http://dx.doi.org/10.1016/j.diabres.2014.09.047 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Metformin in combination with various insulin secretagogues in type 2 diabetes and associated risk of cardiovascular morbidity
5
ip t
and mortality – a retrospective nationwide study Ulrik Madvig Mogensen1, MD; Charlotte Andersson2, MD PhD; Emil Loldrup Fosbøl1, MD PhD; Tina Ken Schramm3, MD PhD; Allan Vaag4, MD DMSci; Nikolai Madrid Scheller7, MB; Christian
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, Denmark
6
National Institute of Public Health, University of Southern Denmark
7
Department of Health Science and Technology, Aalborg University, Denmark
d
M
an
Department of Cardiology, University Hospital Rigshospitalet, Copenhagen, Denmark
te
15
1
Ac ce p
10
us
cr
Torp-Pedersen7, MD DMSci; Gunnar Gislason2,5,6, MD PhD; Lars Køber1, MD DMSci.
Corresponding author:
Ulrik Madvig Mogensen, MD
The Heart Centre, Department of Cardiology, University Hospital Rigshospitalet, 9441 20
Blegdamsvej 9, 2100 Copenhagen
Email:
[email protected] Word count: Total: 7600 Abstract: 250 25
Main text: 3511
Page 1 of 28
Tables/Figures: 2/3
Ac ce p
te
d
M
an
us
cr
ip t
Supplementary tables: 4
Page 2 of 28
Abstract Aims: Metformin is the first-line treatment for most patients with type 2 diabetes but many patients
5
ip t
need additional treatment with insulin secretagogues (IS) to achieve glycemic control. We aimed to compare mortality and cardiovascular risk among users of metformin in combination with
cr
pharmacologically different ISs.
us
Methods: Using nationwide administrative Danish registries, we followed all individuals without prior stroke or myocardial infarction who initiated metformin and an IS from 1997 through 2009.
10
an
Rate ratios (RR) of all-cause mortality, cardiovascular death, and a composite of myocardial infarction, stroke, or cardiovascular death were compared between user groups using time-
M
dependent multivariable Poisson regression models. The most common combination, glimepiride+metformin, was used as reference.
d
Results: A total of 56,827 patients were included, 56% male, the mean age was 6112.5 years, and
rates of mortality for combinations of ISs with metformin were; 15.4 (repaglinide), 28.1 (glipizide),
Ac ce p
15
te
median duration of prior monotherapy was 2.2 (inter quartile range 0.5-4.5) years. Crude incidence
23.7 (glibenclamide), 21.1 (gliclazide), 20.7 (glimepiride), 27.7 (tolbutamide) deaths per 1000 person years. In adjusted analysis, the associated mortality risk was similar for users of gliclazide+metformin (RR=1.01 [0.88-1.15]), repaglinide+metformin (RR=0.81 [0.62-1.05]), glibenclamide+metformin (RR=0.98 [0.87-1.10]), and tolbutamide+metformin (RR=1.04 [0.8520
1.28]). Users of glipizide+metformin was associated with increased all-cause mortality (RR=1.16 [1.02-1.32], p=0.02), cardiovascular death (RR=1.21 [1.01-1.46], p=0.04), and the combined endpoint (RR=1.20 [1.06-1.36, p=0.005). Conclusion: Most ISs in combination with metformin were associated with similar mortality and cardiovascular risk.Whether glipizide is associated with increased risk compared with other ISs
Page 3 of 28
Ac ce p
te
d
M
an
us
cr
ip t
when used in combinations with metformin warrants further study.
Page 4 of 28
Introduction The majority of type 2 diabetes patients eventually need more than one glucose lowering drug and the combination of metformin and sulfonylurea (SU) is widely used.(1) Concerns about the safety of this combination were raised when the UK Prospective Diabetes Study (UKPDS) reported an increased risk of diabetes related death and all-cause mortality in patients receiving a combination
ip t
5
cr
of metformin and SU when compared with SU alone(2).
A meta-analysis of observational studies found no effect on cardiovascular or all-cause mortality,
us
yet the study did find an increased risk of a combined endpoint of cardiovascular hospitalization or
10
an
all-cause mortality associated with the combination of SU and metformin(3). Furthermore, two recent meta-analyses on randomized trials have suggested that the combination of metformin and
M
SU is associated with increased mortality (4,5).
Whether pharmacological differences between individual SUs translate into differences in
d
their associated cardiovascular- and mortality risk is not clear(6-9). Compared with other ISs,
hypoglycemia(13,14). On the contrary, gliclazide was suggested to have a more favorable
Ac ce p
15
te
glibenclamide has been associated with an increased risk of mortality(10-12) and
profile(15-17). Accordingly, in a recent study on all individual SUs and repaglinide (altogether: ISs), gliclazide and repaglinide appeared to be associated with lower all-cause and cardiovascular death compared with other ISs when used as monotherapy(12). However, data on combination therapy are very limited. 20
Considering the cardiovascular risk burden that is carried by type 2 diabetes, it is
important to put increased focus on the cardiovascular safety of glucose lowering treatments. Unfortunately, randomized clinical trials are unlikely to be conducted in this setting. Given their widespread use and potential clinically relevant differences, we performed a nationwide study comparing cardiovascular- and mortality risk related to users of different combinations of ISs and 25
metformin.
Page 5 of 28
Methods
5
ip t
Data sources All residents in Denmark have a unique civil registration number. This allows linkage of data from
cr
different nationwide administrative registers on an individual level. The Danish National Patient registry has kept records of all hospital admissions since 1978. Each admission is registered with
us
one primary diagnosis defined as the main reason for admission, while any contributing causes are registered as secondary diagnoses, according to the International Classification of Diseases, 8th and 10th before and after 1994, respectively. The Danish Registry of Medicinal Product Statistics has
an
10
recorded all prescriptions dispensed from pharmacies in Denmark since 1995. All drugs are
M
registered by the Anatomical Therapeutic Classification (ATC) with information on date of dispensing and amount dispensed available. Information on vital status was obtained from the
from the National Causes of Death Register, in which all deaths and their primary and contributing
te
15
d
Centralized Civil Registration System through Statistics Denmark. Cause of death was obtained
Ac ce p
causes are registered according to ICD10 within 2 weeks after occurrence.
Study population and outcomes measures All Danish individuals ≥ 18 years initiating glucose-lowering therapy with a dual combination of 20
metformin and IS between January 1, 1997 and December 31, 2009 were identified. Individuals with a prior myocardial infarction or stroke were excluded to study time to first event in homogenous population without prior cardiovascular disease. A flowchart for the study population is presented in figure 1.
Page 6 of 28
All-cause mortality, cardiovascular death, and a combined endpoint of nonfatal myocardial infarction (ICD-10 codes: I21-I22), stroke (I61-I64), or cardiovascular death (I00-I99) were used as
ip t
endpoints in separate analyses.
cr
5
us
Oral glucose lowering therapy
an
All types of glucose-lowering treatments were classified according to drug type including metformin (A10B), glimepiride (A10BB12), glibenclamide (A10BB01), gliclazide (A10BB09), glipizide (A10BB07), tolbutamide (A10BB03) and repaglinide (A10BX02). Durations of all
M
10
treatments were calculated using dates of claimed prescriptions, number of tablets and strength of
d
the glucose lowering drug: Typical daily doses were defined for all drugs according to tablet
te
strength based on the recommended treatment doses of the individual drugs. By considering up to four consecutive claimed prescriptions, a continuous treatment period was assumed if this was compatible with at least the minimal dose of the respective glucose lowering drug(18,19).
Ac ce p
15
Treatment periods were extended for 30 days as to avoid potential inappropriate breaks in the calculated treatment periods. Sensitivity analysis without this adjustment was performed. Combination therapy was defined as all intervals where patients were claiming more than one glucose lowering drug within the same periods of time. 20
Comorbidity, concomitant cardiovascular therapy, and income Based on discharge diagnoses 10 year prior to the index date, a modified Charlson comorbidity index score(20) (not including diabetes and myocardial infarction) was calculated for each individual. Concomitant cardiovascular therapy was included as time-dependent variables (use within the last year) and included use of statins (C10AA), aspirin (B01AC06), beta-blockers (C07),
Page 7 of 28
calcium channel blockers (C08), angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (RASi) (C09), other anti-hypertensive drugs (C02), loop diuretics (C03C), nitrates (C01D), vitamin K antagonists (B01AA), digoxin (C01AA), spironolactone (C03D). To account for socioeconomic status, personal gross income 5 years prior to the index date was graded in quintiles and included in the analyses.
us
cr
ip t
5
SAS statistical software version 9.2 (SAS institute, NC, Cary) were used for all statistical analysis.
M
10
an
Statistical analysis
A p-value < 0.05 was considered statistically significant in all analyses.
d
Descriptive data was reported as meanstandard deviation (SD), median with interquartile range
te
(IQR) or frequencies in percentages as appropriate. Rate ratios (RR) of reaching the relevant
15
Ac ce p
outcomes dependent on use of different combinations of ISs and metformin were investigated using time dependent, multivariable Poisson regression models. Follow-up began on the day of a first claimed prescription of a dual combination of metformin and an IS. Patients were considered at risk in the individual treatment groups during exposure to the treatment. It was possible for all patients to change combination therapy or convert to multi therapy or monotherapy during follow-up dependent on the type and number of drugs they were filling a prescription for. If patients changed 20
to another treatment during follow-up, they were registered as receiving e.g. ‘multitherapy’, ‘insulin therapy’, or ‘other therapy’. Outcome events were attributed to the drugs the patient was receiving at the time of the event. Sensitivity analyses were performed to assess potential influence of exposure prior to an event. Follow-up ended on December 31, 2009 or the date of death or, in analyses of the combined end point, the date of myocardial infarction, stroke or cardiovascular
Page 8 of 28
death, whichever came first. Models were adjusted for age, sex, treatment duration, actual calendar year and year of initiation of combination therapy, concomitant cardiovascular treatment, and Charlson comorbidity index. Calendar year and treatment duration were included as 1 year intervals, age was included as 5 year intervals. A variable indicating initial use of an IS, metformin, or combination therapy at initiation of glucose lowering therapy was included in the model to adjust
ip t
5
for monotherapy treatment regimens prior to inclusion.
cr
Glimepiride was the most widely used IS in combination with metformin and was used as the
us
referent.
an
10
M
Sensitivity analyses
Sensitivity analyses were performed to test the robustness of the results: an analysis a) including
d
only patients who were naive to the combination treatment, b) excluding patients with insulin use
15
te
prior to initiation of combination therapy as well as censoring patients when initiating multiple (≥3 drugs) or insulin therapy during follow-up, c) using a grace period of 30 days to account for the
Ac ce p
possibility of a delayed effect of the treatment, d) adjusting for updated Charlson score throughout follow-up, e) including only patients who had exclusively used IS prior to combination therapy, f) without time-dependent adjustment for concomitant cardiovascular therapy (instead defining cardiovascular therapy as use at or prior to baseline), g) including only patients with a Charlson 20
score of 0.
Propensity matched analysis Matched samples of combination therapy with metformin and the individual IS’s were constructed based on the propensity to receive metformin and glimepiride in combination. Only first time users of the individual ISs were included. The individual propensity scores were calculated by logistic 25
regression, based on all covariates in the baseline table. Hence, the capability of the models for Page 9 of 28
predicting who would be treated with the different ISs ranged from moderate (C-statistics 0.55) to high (C-statistics 0.76). Baseline characteristics between matched groups are presented in a supplementary table. 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).
ip t
5
cr
Ethics
The study was approved by the Danish Data Protection Agency (no. 2007-41-1667). Ethical
an
us
approval is not needed for retrospective register studies in Denmark.
M
10
Results
d
A total of 204,321 inhabitants in Denmark aged 18 years or above initiated glucose lowering
te
therapy and of those 87,630 (43%) needed two or more glucose lowering therapies during the study
15
Ac ce p
period. Dual combination of metformin and IS was initiated by 64,816 (74%) and of these, 56,827 did not have a prior myocardial infarction or stroke and were included in the primary analyses (Figure1). Mean follow-up time (from initiation of first dual therapy of metformin and an IS to end of study or death) was 4.1 years (IQR 1.9-6.6 years) and 7863 (13.8%) died. A large proportion of patients changed therapy during follow-up as 35% received the same dual combination of IS and metformin at end of follow-up compared with index treatment; others had 20
changed to multitherapy (16%), insulin+metformin (18%), insulin monotherapy (13%), metformin monotherapy (17%), no therapy (10%) or other forms of glucose lowering therapies. Considering risk time associated with the use of a combination of IS and metformin only, there were 2665 deaths during 122,613 patient years of follow-up and a median follow up of 2.16 years.
Page 10 of 28
When initiating an IS+metformin combination the mean age was 61 ( standard deviation 12.5) years, 56% were male and median duration of monotherapy before initiating the combination therapy was 2.2 (inter quartile range 0.5-4.5) years. The largest proportion of patients had initiated glucose lowering therapy with an IS and used metformin as add-on. This changed during the study period and in the last 2 years of the study (2007-2009), the majority of patients had
cr
initiated glucose lowering therapy with metformin (Figure 2).
ip t
5
The baseline characteristics according to use of individual ISs are presented in Table 1. Since 5,091
us
patients (9.0%) changed combination therapy during follow-up, the number of patients in each treatment group adds up to more than the total number of patients in the study. In Supplementary Table 1, baseline characteristics are presented for each individual patient at time of inclusion in the
an
10
d
M
study (counting each patient only once).
receiving glipizide+metformin or tolbutamide+metformin were older than average and had the
Ac ce p
15
te
As demonstrated in Table 1, exposure groups were relatively well balanced. However, patients
highest proportion of patients using IS before metformin. Gliclazide+metformin and glimepiride+metformin had the highest proportion of male users, had longer treatment duration when initiating combination therapy and had more often used combinations of metformin with another IS prior to inclusion. Furthermore, they were more frequently prescribed statins and 20
antihypertensive drugs and were included later in the study period. Users of repaglinide+metformin were younger and tended to have lower comorbidity scores. Incidence rates of mortality, cardiovascular death, and the combined end point associated with use of metformin and the individual ISs are presented in table 2. Users of combinations of metformin with glibenclamide, tolbutamide, and glipizide had the highest
25
incidence rates for all endpoints, combinations with gliclazide and glimepiride were associated with Page 11 of 28
intermediate incidence rates, whereas users of metformin combined with repaglinide had the lowest incidence rates for all endpoints. Adjusted rate ratios for mortality, cardiovascular death, and the combined end point are presented in Figure 3 using glimepiride in combination with metformin as reference. Differences between users of individual IS’s were modest, only glipizide appeared to be associated with increased RR for
ip t
5
mortality. If accounting for multiple comparisons with simple Bonferroni adjusted significance
cr
level of 0.01, glipizide+metformin was significantly associated with the composite end point
an
us
(p=0.004), but not with all-cause (p=0.024) or cardiovascular death (p=0.049).
15
Ac ce p
te
d
M
10
Additional analyses
Sensitivity analyses defined in the method section gave similar results (Supplementary Table 2). Propensity score matched groups using glimepiride+metformin as reference is presented in Supplementary Table 3. Results from models including only matched patients are presented in Supplementary Table 4. In these analyses, the risk among users of glipizide+metformin was only 20
significantly increased for all-cause mortality (RR=1.19 [1.01-1.41], p=0.043), but not for cardiovascular death (RR=1.24 [0.97-1.58], p=0.08) or the combined end point (RR=1.13 [0.961.34], p=0.14).
Page 12 of 28
Discussion
5
ip t
In this nationwide retrospective study, we compared the cardiovascular- and mortality risk among users of individual ISs in combination with metformin and report three major findings: First, in
cr
combinations with metformin, users of gliclazide and repaglinide was associated with similar
us
cardiovascular risks as other widely used ISs. Second, users of glibenclamide with metformin was not associated with an increased risk as suggested by studies on IS monotherapy.(11,12,14) Third,
10
an
users of glipizide was associated with an increased risk of all endpoints compared with glimepiride when used in combinations with metformin. However, this difference did not reach significant
M
levels for all-cause mortality and cardiovascular death using Bonferroni correction to adjust for the number of combinations examined.
d
The majority of type 2 diabetes patients need more than one glucose lowering drug
this study, 50% used a combination of glucose lowering therapies 2.2 years after initiating use of
Ac ce p
15
te
during their life with diabetes(21,22) and combinations of IS and metformin are widely used(1). In
their first glucose lowering drug and 74% of patients used IS and metformin, making this the most widely used combination. Hence, any potential differences in risk among these combinations would be important to clarify.
It has been speculated that differences in outcome in recent clinical trials may have
20
been underpinned by differential effects of ISs (12,23). Glibenclamide was widely used in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study(24) which reported a signal of increased mortality in the intensive therapy arm, whereas gliclazide was prescribed by study protocol in the Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Relase Controlled Evaluation (ADVANCE)(25), with no signal of increased mortality. On the other hand,
Page 13 of 28
in the UKPDS and A Diabetes Outcome Progression Trial (ADOPT), glibenclamide was not
5
ip t
associated with increased risk (26).
We did not observe an increased risk associated with glibenclamide compared with other ISs when
cr
used in combinations with metformin. Very few other studies have investigated long-term outcomes
us
among patients using different combinations of ISs with metformin (27-29), and to the best of our knowledge, only one study has compared individual combinations directly(29). In a study by
10
an
Monami et al. on 696 type 2 diabetes patients, use of glibenclamide was associated with increased mortality (OR 2.09, 95% CI: 1.07-4.11) when compared with use of other IS (repaglinide, gliclazide
M
or glimepiride) in combinations with metformin(27). However, the analyses were based on small sample sizes in each combination group (range 17 to 469), with no descriptive statistics on whether
d
the groups were balanced in terms of baseline risk factors.
glimepiride in combinations with metformin is consistent with a recent study by Pantalone et al. on
Ac ce p
15
te
Our finding of a similar associated risk among users of glibenclamide compared with
7320 type 2 diabetes patients, reporting no difference in mortality risk among users of different combinations of IS and metformin(29). When comparing glimepiride with glibenclamide, the hazard ratio (HR) for overall mortality was 1.23 [0.98-1.94], p=0.07, tending to contrast the findings by Monami et al. as described above(27). Likewise, there was no difference between users 20
of glimepiride and glipizide in combinations with metformin (HR=1.03 [0.89-1.20]). In addition to the ISs investigated by Pantalone et al., we were able to include gliclazide, tolbutamide, and repaglinide and hence compared all individual ISs. These ISs did not differ significantly in terms of cardiovascular risk when used in combinations with metformin.
Page 14 of 28
Our finding of an increased risk among users of glipizide is not readily explained by its pharmacological properties. ISs stimulate insulin release by closing adenosine triphosphate dependent potassium (KATP) channels through binding to receptors on pancreatic beta cells, but similar receptors can be found in the myocardium and on vascular smooth muscle cells and ISs differs in their specificity for the pancreatic receptors(16). Hence, IS mediated inhibition of KATP
ip t
5
channels in cardiomyocytes could theoretically interfere with ischemic preconditioning, a protective
cr
adaption to prior ischemia(30). Glibenclamide(31-33), but neither gliclazide(34), glimepiride(35) or
us
importantly, glipizide(36) has been associated with impairment of ischemic preconditioning. Hence, glipizide is pancreas specific and is considered to have a risk profile comparable to that of glimepiride and gliclazide(1).
an
10
Glipizide has been associated with increased risk in some studies on
M
monotherapy(12,37,38). In a study by Schramm et al.(12), both glibenclamide, glipizide and glimepiride appeared to be associated with increased risk of cardiovascular adverse events
glimepiride(12). In a study by Johnson(39), increased risk of overall mortality was found when
te
15
d
compared with gliclazide and repaglinide, but there were no comparative data on glipizide versus
Ac ce p
comparing glipizide and glibenclamide, so-called “old sulfonylureas”, with gliclazide and glimepiride, but the isolated risk associated with glipizide was not reported. In a small study (37), glipizide was associated with increased risk of developing coronary artery disease when compared with gliclazide or glimepiride. 20
However, other studies have not found an increased risk associated with glipizide alone(8,9) or in treatment groups including glipizide(6). In a study by Andersson et al.(8), there was comparable mortality risk associated with individual ISs in diabetes patients with heart failure. Other studies have not found clinically relevant differences between ISs.(6,7,9,40). In propensity-score matched analyses, the risk among users of glipizide was weaker and not
25
significant for mortality and the composite end point. Taken together, our results suggest that most
Page 15 of 28
ISs in combination with metformin have similar mortality- and cardiovascular risk. The observed increased risk associated with glipizide must be interpreted with caution and further studies are
5
ip t
warranted.
Strength and limitations
cr
Strength of this study is the completeness of data with a nationwide unselected population of
us
patients using combinations of glucose lowering therapies. Furthermore, it was possible to adjust for changes in therapy during follow-up. We did not have information on smoking status, family
10
an
history of cardiovascular disease, HbA1C and body mass index. These limitations were partly addressed by using time-dependent adjustment for several potential confounders, including
M
concomitant cardiovascular therapy and by performing propensity matched analyses. However, residual confounding cannot be excluded and confounding as a result of perceived lower risk
d
associated with some ISs cannot be excluded. A large proportion of patients using glipizide were
15
te
from the first part of follow-up and might have received less optimal therapy, which might explain differences in outcome in spite of adjusting for index year in the analyses. In many western
Ac ce p
countries the risk of cardiovascular death has decreased during the study period, which coincided with changes in the use of different ISs. Hence, there was interaction between index year and treatment. However, parameter estimates did not change significantly when stratifying in 2 equal time periods (data not shown). 20
Conclusion
In a nationwide cohort of patients with type 2 diabetes needing combination therapy with metformin and IS users of gliclazide or repaglinide were not associated with a significantly better cardiovascular risk compared with users of other ISs, including glibenclamide, when used in combinations with metformin. Users of glipizide in combination with metformin appeared to be 25
associated with increased risk. However, this may represent a random finding and/or residual Page 16 of 28
confounding. Prospective, randomized placebo controlled trials are required to establish the cardiovascular risk associated with different glucose lowering therapy regimens.
UMM made the primary contributions, researched data and wrote the manuscript.
cr
5
ip t
Aknowledgements
CA, ELF and GG researched data and reviewed/edited the manuscript. NMS, CTP, AV, LK
us
edited/reviewed the manuscript. All authors contributed to interpretation of the results and revised
Funding
M
10
an
the manuscript for intellectual content. UMM is the guarantor.
This work was supported by a grant from the Interreg IVA program, a part of the European Union.
d
CA was supported by an independent research grant from the Danish agency for science,
te
technology and innovation (grant number FSS - 11-120873). GG is supported by an independent
15
Ac ce p
research scholarship from the Novo Nordisk Foundation. Conflicts of interest
UMM, CA, ELF, TKS, NMS, CTP, LK: None. AV owns stocks and has received honoraria for lectures as well as for consultancy services from Novo Nordisk. GG owns stocks in Novo Nordisk.
Page 17 of 28
Reference List
1. Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR, Sherwin R, Zinman B: Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 32:193-203, 2009
10
Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). UK Prospective Diabetes Study (UKPDS) Group. Lancet 352:854865, 1998
cr
2.
ip t
5
us
3. Rao AD, Kuhadiya N, Reynolds K, Fonseca VA: Is the combination of sulfonylureas and metformin associated with an increased risk of cardiovascular disease or all-cause mortality?: a meta-analysis of observational studies. Diabetes Care 31:1672-1678, 2008 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 13:221-228, 2011
an
15
M
5. Boussageon R, Supper I, Bejan-Angoulvant T, Kellou N, Cucherat M, Boissel JP, Kassai B, Moreau A, Gueyffier F, Cornu C: Reappraisal of metformin efficacy in the treatment of type 2 diabetes: a metaanalysis of randomised controlled trials. PLoS Med 9:e1001204, 2012
d
20
6. Evans JM, Ogston SA, Reimann F, Gribble FM, Morris AD, Pearson ER: No differences in mortality between users of pancreatic-specific and non-pancreatic-specific sulphonylureas: a cohort analysis. Diabetes Obes Metab 10:350-352, 2008
25
30
35
Ac ce p
te
7. Pantalone KM, Kattan MW, Yu C, Wells BJ, Arrigain S, Jain A, Atreja A, Zimmerman RS: The risk of overall mortality in patients with type 2 diabetes receiving glipizide, glyburide, or glimepiride monotherapy: a retrospective analysis. Diabetes Care 33:1224-1229, 2010 8. Andersson C, Gislason GH, Jorgensen CH, Hansen PR, Vaag A, Sorensen R, Merie C, Olesen JB, Weeke P, Schmiegelow M, Norgaard ML, Kober L, Torp-Pedersen C: Comparable long-term mortality risk associated with individual sulfonylureas in diabetes patients with heart failure. Diabetes Res Clin Pract 94:119-125, 2011 9. Horsdal HT, Johnsen SP, Sondergaard F, Jacobsen J, Thomsen RW, Schmitz O, Sorensen HT, Rungby J: Sulfonylureas and prognosis after myocardial infarction in patients with diabetes: a populationbased follow-up study. Diabetes Metab Res Rev 25:515-522, 2009 10. Zeller M, Danchin N, Simon D, Vahanian A, Lorgis L, Cottin Y, Berland J, Gueret P, Wyart P, Deturck R, Tabone X, Machecourt J, Leclercq F, Drouet E, Mulak G, Bataille V, Cambou JP, Ferrieres J, Simon T: Impact of type of preadmission sulfonylureas on mortality and cardiovascular outcomes in diabetic patients with acute myocardial infarction. J Clin Endocrinol Metab 95:4993-5002, 2010 11. 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 86:247-253, 2009 40
12. Schramm TK, Gislason GH, Vaag A, Rasmussen JN, Folke F, Hansen ML, Fosbol EL, Kober L, Norgaard ML, Madsen M, Hansen PR, Torp-Pedersen C: Mortality and cardiovascular risk associated with
Page 18 of 28
different insulin secretagogues compared with metformin in type 2 diabetes, with or without a previous myocardial infarction: a nationwide study. Eur Heart J 32:1900-1908, 2011
5
13. Holstein A, Plaschke A, Egberts EH: Lower incidence of severe hypoglycaemia in patients with type 2 diabetes treated with glimepiride versus glibenclamide. Diabetes Metab Res Rev 17:467-473, 2001
15. O'Brien RC, Luo M, Balazs N, Mercuri J: In vitro and in vivo antioxidant properties of gliclazide. J Diabetes Complications 14:201-206, 2000
cr
10
ip t
14. Gangji AS, Cukierman T, Gerstein HC, Goldsmith CH, Clase CM: A systematic review and metaanalysis of hypoglycemia and cardiovascular events: a comparison of glyburide with other secretagogues and with insulin. Diabetes Care 30:389-394, 2007
17. Kimoto K, Suzuki K, Kizaki T, Hitomi Y, Ishida H, Katsuta H, Itoh E, Ookawara T, Suzuki K, Honke K, Ohno H: Gliclazide protects pancreatic beta-cells from damage by hydrogen peroxide. Biochem Biophys Res Commun 303:112-119, 2003
an
15
us
16. Abdelmoneim AS, Hasenbank SE, Seubert JM, Brocks DR, Light PE, Simpson SH: Variations in tissue selectivity amongst insulin secretagogues: a systematic review. Diabetes Obes Metab 14:130-138, 2012
M
20
18. Fosbol EL, Gislason GH, Jacobsen S, Abildstrom SZ, Hansen ML, Schramm TK, Folke F, Sorensen R, Rasmussen JN, Kober L, Madsen M, Torp-Pedersen C: The pattern of use of non-steroidal antiinflammatory drugs (NSAIDs) from 1997 to 2005: a nationwide study on 4.6 million people. Pharmacoepidemiol Drug Saf 17:822-833, 2008
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 40:373-383, 1987
Ac ce p
25
te
d
19. Olsen AM, Fosbol EL, Lindhardsen J, Andersson C, Folke F, Nielsen MB, Kober L, Hansen PR, TorpPedersen C, Gislason GH: Cause-specific cardiovascular risk associated with nonsteroidal antiinflammatory drugs among myocardial infarction patients--a nationwide study. PLoS One 8:e54309, 2013
21. Wright A, Burden AC, Paisey RB, Cull CA, Holman RR: Sulfonylurea inadequacy: efficacy of addition of insulin over 6 years in patients with type 2 diabetes in the U.K. Prospective Diabetes Study (UKPDS 57). Diabetes Care 25:330-336, 2002 30
22. 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 281:2005-2012, 1999 23. Gore MO, McGuire DK: Resolving drug effects from class effects among drugs for type 2 diabetes mellitus: more support for cardiovascular outcome assessments. Eur Heart J 32:1832-1834, 2011
35
40
24. Gerstein HC, Miller ME, Byington RP, Goff DC, Jr., Bigger JT, Buse JB, Cushman WC, Genuth S, IsmailBeigi F, Grimm RH, Jr., Probstfield JL, Simons-Morton DG, Friedewald WT: Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med 358:2545-2559, 2008 25. Patel A, MacMahon S, Chalmers J, Neal B, Billot L, Woodward M, Marre M, Cooper M, Glasziou P, Grobbee D, Hamet P, Harrap S, Heller S, Liu L, Mancia G, Mogensen CE, Pan C, Poulter N, Rodgers A, Williams B, Bompoint S, de Galan BE, Joshi R, Travert F: Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med 358:2560-2572, 2008
Page 19 of 28
26. Kahn SE, Haffner SM, Heise MA, Herman WH, Holman RR, Jones NP, Kravitz BG, Lachin JM, O'Neill MC, Zinman B, Viberti G: Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy. N Engl J Med 355:2427-2443, 2006 5
27. Monami M, Luzzi C, Lamanna C, Chiasserini V, Addante F, Desideri CM, Masotti G, Marchionni N, Mannucci E: Three-year mortality in diabetic patients treated with different combinations of insulin secretagogues and metformin. Diabetes Metab Res Rev 22:477-482, 2006
29. Pantalone KM, Kattan MW, Yu C, Wells BJ, Arrigain S, Nutter B, Jain A, Atreja A, Zimmerman RS: The risk of overall mortality in patients with Type 2 diabetes receiving different combinations of sulfonylureas and metformin: a retrospective analysis. Diabet Med 29:1029-1035, 2012
cr
10
ip t
28. Monami M, Marchionni N, Masotti G, Mannucci E: Effect of combined secretagogue/biguanide treatment on mortality in type 2 diabetic patients with and without ischemic heart disease. Int J Cardiol 126:247-251, 2008
31. Tomai F, Crea F, Gaspardone A, Versaci F, De PR, Penta de PA, Chiariello L, Gioffre PA: Ischemic preconditioning during coronary angioplasty is prevented by glibenclamide, a selective ATPsensitive K+ channel blocker. Circulation 90:700-705, 1994
an
15
us
30. Gribble FM, Reimann F: Differential selectivity of insulin secretagogues: mechanisms, clinical implications, and drug interactions. J Diabetes Complications 17:11-15, 2003
33. Klepzig H, Kober G, Matter C, Luus H, Schneider H, Boedeker KH, Kiowski W, Amann FW, Gruber D, Harris S, Burger W: Sulfonylureas and ischaemic preconditioning; a double-blind, placebocontrolled evaluation of glimepiride and glibenclamide. Eur Heart J 20:439-446, 1999
d
20
M
32. Lee TM, Chou TF: Impairment of myocardial protection in type 2 diabetic patients. J Clin Endocrinol Metab 88:531-537, 2003
30
35. Mocanu MM, Maddock HL, Baxter GF, Lawrence CL, Standen NB, Yellon DM: Glimepiride, a novel sulfonylurea, does not abolish myocardial protection afforded by either ischemic preconditioning or diazoxide. Circulation 103:3111-3116, 2001
Ac ce p
25
te
34. Maddock HL, Siedlecka SM, Yellon DM: Myocardial protection from either ischaemic preconditioning or nicorandil is not blocked by gliclazide. Cardiovasc Drugs Ther 18:113-119, 2004
36. Flynn DM, Smith AH, Treadway JL, Levy CB, Soeller WC, Boettner WA, Wisniecki P, Plowchalk DR, Gernhardt SS, Tracey WR, Knight DR: The sulfonylurea glipizide does not inhibit ischemic preconditioning in anesthetized rabbits. Cardiovasc Drugs Ther 19:337-346, 2005 37. Sadikot SM, Mogensen CE: Risk of coronary artery disease associated with initial sulphonylurea treatment of patients with type 2 diabetes: a matched case-control study. Diabetes Res Clin Pract 82:391-395, 2008
35
40
38. Jorgensen CH, Gislason GH, Andersson C, Ahlehoff O, Charlot M, Schramm TK, Vaag A, Abildstrom SZ, Torp-Pedersen C, Hansen PR: Effects of oral glucose-lowering drugs on long term outcomes in patients with diabetes mellitus following myocardial infarction not treated with emergent percutaneous coronary intervention--a retrospective nationwide cohort study. Cardiovasc Diabetol 9:54, 2010 39. Johnsen SP, Monster TB, Olsen ML, Thisted H, McLaughlin JK, Sorensen HT, Lervang HH, Rungby J: Risk and short-term prognosis of myocardial infarction among users of antidiabetic drugs. Am J Ther 13:134-140, 2006
Page 20 of 28
40. Juurlink DN, Gomes T, Shah BR, Mamdani MM: Adverse cardiovascular events during treatment with glyburide (glibenclamide) or gliclazide in a high-risk population. Diabet Med 29:1524-1528, 2012
Ac ce p
te
d
M
an
us
cr
ip t
5
Page 21 of 28
Figure legends Figure 1. Flowchart. MI=myocardial infarction.
Proportions of initial drug use (insulin secretagogoues (IS), metformin or a combination of IS and
cr
5
ip t
Figure 2.
metformin at initiation of glucose lowering therapy) among included patients. There was no
us
interaction between initial drug use and the used combination treatment (p=0.71)
an
Figure 3.
Rate Ratios (RR) and 95% confidence intervals (95% CI) for mortality (a), cardiovascular mortality
M
(b) and a combined endpoint of cardiovascular death, myocardial infarction or stroke (c) when
te
d
comparing individual insulin secretagogoues+metformin using glimepiride+metformin as reference.
Ac ce p
10
Page 22 of 28
Figure
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 1
Page 23 of 28
Figure
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 2
Page 24 of 28
cr
ip t
Figure
Ac c
ep te
d
M
an
us
Figure 3
Page 25 of 28
Table 1. Baseline variables
Ac c
metfomin in combination with glimepiride glipizide 40026 4466 22643 (56.6) 2459 (55.1) 60.69 ± 12.5 62.29 ± 12.2 1.82 1.85 (0.4-3.9) (0.4-3.8)
an
us
gliclazide 5966 3401 (57) 61.18 ± 12.2 1.92 (0.5-4.0)
repaglinide tolbutamide 2118 2173 1157 (54.6) 1201 (55.3) 59.88 ± 11.8 63.12 ± 12.5 1.95 1.75 (0.5-3.8) (0.4-3.7)
192 (3.2) 253 (4.2) 32 (0.5) 57 (1) 164 (2.8) 140 (2.4) 231 (3.9)
1494 (3.7) 1787 (4.5) 275 (0.7) 482 (1.2) 1300 (3.3) 970 (2.4) 1549 (3.9)
157 (3.5) 206 (4.6) 22 (0.5) 59 (1.3) 134 (3) 114 (2.6) 177 (4)
55 (2.6) 68 (3.2) 11 (0.5) 21 (1) 51 (2.4) 31 (1.5) 84 (4)
76 (3.5) 102 (4.7) 5 (0.2) 33 (1.5) 67 (3.1) 54 (2.5) 70 (3.2)
6539 (88.1) 820 (11.1) 64 (0.9)
5281 (88.6) 632 (10.6) 47 (0.8)
35048 (87.7) 4526 (11.3) 313 (1)
3899 (87.4) 517 (11.6) 47 (1.1)
1886 (89.1) 213 (10.1) 19 (0.9)
1905 (87.7) 249 (11.5) 19 (0.9)
2207 (30.9) 2998 (43.9) 1394 (25.1) 1912 (29.2) 1642 (32.2) 1446 (24.7) 132 (2.8) 480 (6.46) 288 (3.88) 484 (6.52) 307 (4.13)
2659 (45.9) 3029 (53.5) 1322 (28.5) 1857 (34.9) 1295 (32.7) 1311 (27.9) 176 (4.7) 358 (6) 311 (5.21) 309 (5.18) 261 (4.37)
18352 (47.4) 19858 (52.7) 8632 (28.5) 12564 (35.2) 9543 (34.9) 9085 (28.6) 906 (3.7) 2264 (5.66) 2095 (5.23) 2270 (5.67) 1790 (4.47)
1496 (34.8) 1886 (45.6) 826 (24.8) 1187 (30.3) 1014 (33.7) 926 (26.8) 90 (3.3) 281 (6.29) 221 (4.95) 315 (7.05) 168 (3.76)
776 (37.9) 955 (48.1) 411 (26.2) 615 (32.6) 483 (33.1) 430 (25.7) 61 (4.4) 104 (4.91) 77 (3.64) 94 (4.44) 76 (3.59)
544 (26.2) 798 (40.4) 363 (23.5) 527 (28.4) 518 (34.2) 430 (26.2) 36 (2.7) 126 (5.8) 96 (4.42) 149 (6.86) 75 (3.45)
d
M
269 (3.6) 294 (4) 23 (0.3) 100 (1.4) 219 (3) 170 (2.3) 272 (3.7)
ep te
Numbers Men, n (%) Age, mean ± SD Treatment Duration in years, median (p25-p75) Comorbiditites Congestive Heart Failure, n (%) Atrial Fibrillation, n (%) Renal failure, n (%) Peripheral Vascular Disease, n (%) COPD, n (%) Peptic Ulcer, n (%) Cancer, n (%) Charlson score, n (%) 0 1-2 ≥3 Concomitant therapy Statin, n (%) RASi, n (%) Beta Blockers, n (%) Aspirin, n (%) Thiazid, n (%) Calcium antagonist, n (%) Other antihypertensive, n (%) Nitrates, n (%) Vitamin K antagonists, n (%) Digoxin, n (%) Spironolactone, n (%)
glibenclamide 7429 4028 (54.2) 61.78 ± 12.6 1.82 (0.4-3.8)
cr
ip t
Table
Page 26 of 28
cr
ip t 2255 (30.4) 4718 (63.5) 402 (5.4) 2225 (30) 4368 (58.8) 428 (5.8)
1798 (4.5) 8336 (20.8) 13693 (34.2) 16193 (40.5)
804 (18) 1330 (29.8) 1384 (31) 948 (21.2)
216 (10.2) 851 (40.2) 656 (31) 395 (18.7)
471 (21.7) 731 (33.6) 700 (32.2) 271 (12.5)
16994 (42.5) 20659 (51.6) 1869 (4.7) 16563 (41.4) 18747 (46.8) 3132 (7.8)
1357 (30.4) 2857 (64) 214 (4.8) 1330 (29.8) 2639 (59.1) 350 (7.8)
692 (32.7) 1315 (62.1) 91 (4.3) 712 (33.6) 481 (22.7) 424 (20)
509 (23.4) 1601 (73.7) 55 (2.5) 499 (23) 1464 (67.4) 132 (6.1)
us
319 (5.4) 1238 (20.8) 2294 (38.5) 2115 (35.5)
an
1360 (18.3) 2277 (30.7) 2081 (28) 1711 (23)
2366 (39.7) 3299 (55.3) 257 (4.3) 2287 (38.3) 2949 (49.4) 885 (14.8)
M
Inclusion period, n (%) January 1997 - March 2000 April 2000 - June 2003 July 2003 - September 2006 October 2006 - December 2009 GLT at incident diabetes Initiated GLT with metformin, n (%) Initiated GLT with IS, n (%) Initiated GLT with combination therapy Prior exclusive use of metformin, n (%) Prior exclusive use of IS, n (%) Received a prior combination therapy, n (%)
Ac c
ep te
d
Abbreviations; SD: Standard Deviation, P;percentile, COPD, Chronic Obstructive Pulmonary Disease; RASi, Renin Angiotensin System Inhibitors, DM, Diabetes Mellitus; GLT, Glucose Lowering Therapy; IS, Insulin Secretagogue
Page 27 of 28
Table
Table 2. Crude incidence rates for all-cause mortality, cardiovascular death and the combined end point according to combinations of insulin secretagogues and metformin.
Crude incidence rates Combined end point Events Person years Crude incidence rates
272 9770 28.1 (25.0-31.7)
164 14894 11.0 (9.5-12.8)
102 11308 9.0 (7.4-11.0)
761 79276 9.6 (8.9-10.3)
133 9770 13.7 (11.6-16.3)
387 14493 26.7 (24.2-29.5)
236 11047 21.4 (18.8-24.3)
1647 77409 21.3 (20.3-22.3)
60 3883 15.4 (12.0-19.9)
97 3482 27.7 (22.7-33.8)
28 3883 7.2 (4.9-10.4)
49 3482 14.0 (10.7-18.5)
72 3807 18.9 (15.0-23.8)
96 3406 28.2 (23.1-34.4)
cr
1645 79276 20.7 (19.7-21.7)
us
Cardiovascular death Events Person years
238 11308 21.1 (18.5-23.9)
an
Crude incidence rates
353 14894 23.7 (21.4-26.3)
M
All-cause mortality Events Person years
ip t
glibenclamide
Combination of metformin with gliclazide glimepiride glipizide repaglinide tolbutamide
287 9394 30.6 (27.2-34.3)
Ac
ce pt
ed
Crude incidence rates=events/1000 person years with 95% CI.
Page 28 of 28