+Model DIABET-688; No. of Pages 8
ARTICLE IN PRESS Available online at
ScienceDirect www.sciencedirect.com Diabetes & Metabolism xxx (2015) xxx–xxx
Original article
Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study P. Valensi a , G. de Pouvourville b , N. Benard c , C. Chanut-Vogel c , C. Kempf d , E. Eymard c,∗ , C. Moisan c , J. Dallongeville e a
Department of Endocrinology–Diabetology–Nutrition, Jean-Verdier Hospital, AP–HP, Paris Nord University, CRNH-IdF, CINFO, Bondy, France b ESSEC, Cergy-Pontoise, France c MSD France, Courbevoie, France d CSD, Boulogne, France e INSERM U744, Institut Pasteur, Lille, France Received 11 February 2015; received in revised form 17 March 2015; accepted 22 March 2015
Abstract Aim. – The study compared the duration of maintenance of treatment in patients with type 2 diabetes (T2D) using dual therapy with either metformin and sitagliptin (M-Sita) or metformin and a sulphonylurea (M-SU). Materials and methods. – This observational study included adult patients with T2D who had responded inadequately to metformin monotherapy and therefore had started de-novo treatment with Met-Sita or Met-SU within the previous eight weeks. Patient follow-up and changes to treatment were performed according to their general practitioner’s usual clinical practice. The primary outcome was time to change in treatment for whatever cause. HbA1c and symptomatic hypoglycaemia were also documented. Results. – The median treatment duration for patients in the M-Sita group (43.2 months) was significantly longer (P < 0.0001) than in the M-SU group (20.2 months). This difference persisted after adjusting for baseline differences and confounders. A similar reduction in HbA1c was noted in both arms (–0.6%), and the incidence of hypoglycaemia prior to treatment modification was lower with M-Sita (9.7%) than with M-SU (21.0%). Adverse events potentially related to treatment were reported in 2.8% (n = 52) and 2.7% (n = 20) of patients in the M-Sita and M-SU arms, respectively. Conclusion. – Under everyday conditions of primary diabetes care, dual therapy with M-Sita can be maintained for longer than M-SU. In addition, while efficacy, as measured by changes in HbA1c , was similar between treatments, the incidence of hypoglycaemia was lower in patients taking M-Sita. © 2015 Elsevier Masson SAS. All rights reserved. Keywords: DPP-4 inhibitor; Observational study; Type 2 diabetes
1. Introduction Type 2 diabetes (T2D) is a major public-health challenge in Europe and in most other regions of the world [1]. In the absence of effective treatment, T2D is associated with significant morbidity and mortality, which are mostly attributable to long-term
∗ Corresponding author. 34, avenue Léonard-de-Vinci, 92400 Courbevoie, France. Tel.: +33 1 80 46 44 76. E-mail address: [email protected] (E. Eymard).
cardiovascular complications of the disease. Nevertheless, the risk of long-term complications can be significantly reduced by improving glycaemic control [2]. To this end, although several studies have shown residual dysglycaemia of 6.5% to 6.9% [3], the treatment goal in T2D is to restore and maintain glycaemia at as near to normal levels as possible, defined as an HbA1c level ≤ 7%. Since the end of the 1990s, available choices for the treatment of T2D have expanded considerably. This has allowed individualization of treatment in response to the specific clinical features and history of each patient. As T2D is a chronic disease with
http://dx.doi.org/10.1016/j.diabet.2015.03.007 1262-3636/© 2015 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 2
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
onset usually in older middle-aged people, the challenge for the physician is to choose drug therapies that ensure durable glycaemic control while minimizing the need for intensification of treatment. This will retard disease progression during the course of therapy and delay the exhaustion of effective therapeutic options. Although the efficacy of new oral antidiabetic drugs (OADs) has been convincingly demonstrated in phase-III clinical trials, their comparative effectiveness in everyday clinical practice remains poorly characterized. In this respect, data from observational studies can make an important contribution to the evidence base [4]. For this reason, this observational study (ODYSSEE) was undertaken to compare dual therapy with metformin and sitagliptin [5], a dipeptidyl peptidase (DPP)-4 inhibitor, with dual therapy with metformin and a sulphonylurea in patients failing to respond adequately to metformin monotherapy. As a comprehensive measure of efficacy, the study evaluated treatment maintenance duration, defined as the time from starting treatment until the physician chose to change the treatment. Treatment maintenance should be distinguished from treatment persistence, which refers to the length of time a patient takes medication as prescribed. Treatment maintenance duration is a simple and reliable variable to document outcome in observational studies in real-life treatment settings. It covers all circumstances relating to changes in treatment, including inadequate efficacy, unacceptable occurrence of hypoglycaemia or other adverse events, and poor patient acceptability. The primary objective of the present study was to assess whether metformin plus sitagliptin dual therapy was superior to metformin plus sulphonylurea dual therapy in terms of treatment maintenance duration in patients with T2D. Secondary objectives were to describe treatment regimens using sitagliptin in standard clinical practice and the profiles of such treated patients, and to document the evolution of the clinical status of such patients during treatment. Information on the treatment regimens and patient profiles are to be presented elsewhere. 2. Methods This was a multicentre longitudinal, prospective, observational study conducted in primary care in France. Patients were enrolled between July 2009 and December 2010. The planned follow-up period for each patient was three years, and the last follow-up visit took place in August 2013. 2.1. Participating investigators A group of general practitioners (GPs) in metropolitan France were randomly identified from an extensive national listing of practising physicians by specialty (Cegedim OneKey registries; Cegedim, Boulogne-Billancourt, France). From this group, physicians who provided care to large numbers of patients with T2D were invited to participate. Of the 13,521 primary care practitioners contacted, 1569 agreed to take part in the study. These participating GPs were compared with all GPs in France using data from the National Health Insurance database.
2.2. Patients Participating physicians were expected to consecutively include all patients fulfilling the eligibility criteria who had consulted them during an enrolment period of nine months. Adult (> 18 years) patients with T2D who had initiated de-novo dual therapy with either metformin and sitagliptin (M-Sita group) or metformin and sulphonylurea (M-SU group) within the previous eight weeks were eligible for the study. Also, only patients who, in the physician’s judgment, were equally eligible for a prescription with either sitagliptin or a sulphonylurea were to be included. Allocation of patients to a given medical strategy was not predefined, but based on the GPs’ standard practices, and the prescribing decision was clearly separate from the decision to include the patient in the study. Characteristics of the study patients were compared with those of patients documented in the Cegedim longitudinal patient data (LPD) primary care database, a large panel of patients consulting one of 1200 French primary care physicians considered representative of all primary care practitioners in terms of age, gender and place of practice. 2.3. Treatment Treatments were prescribed at the discretion of the physician. The possible treatment regimens required for enrolment in the study were sitagliptin 100 mg (either as dual therapy with metformin or as part of another treatment regimen) or dual therapy with metformin and sulphonylurea, irrespective of dose. To limit as much as possible any influence of study participation on treatment prescription, physicians were asked to include all patients prescribed sitagliptin, regardless of the treatment regimen. However, only patients prescribed the M-Sita dual therapy were intended for inclusion in the primary analysis. 2.4. Data collection and patient follow-up At the time of study inclusion, the patients’ demographic data and medical characteristics were documented, including comorbidities, recent laboratory tests, previous medications, prescribed medicinal products and reasons for prescribing sitagliptin or sulphonylurea. Patient follow-up and changes to treatment were performed according to the physician’s everyday clinical practice and were not specified by the study protocol. However, the protocol did stipulate that the physician was to document the patient at the time of the patient’s usual quarterly follow-up visits (as recommended by French national guidelines for the monitoring of patients with T2D [6]) over a 3-year period. The information collected included any changes to the treatment regimen and the reasons for change, patient-reported health and dietary habits, laboratory measures of HbA1c , microalbuminuria or urinary albumin excretion rate (UAER), proteinuria and serum creatinine. The onset and number of symptomatic hypoglycaemic episodes were to be documented by the physicians; blood measures were also collected if available and patients could complete a diary. In this diary, patients could report on body weight, occurrence and severity of hypoglycaemic episodes, glycaemia
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
values and any interruption of treatment. If information was collected by the physician by telephone instead of during a visit, this was also documented. 2.5. Evaluation criteria The primary study endpoint was the duration of maintenance of dual therapy expressed as the number of days from treatment initiation to treatment change, defined as the addition, replacement or withdrawal of any agent used in the initial dual therapy. Changes in dose of the initially prescribed dual therapy were not considered treatment changes. Secondary endpoints included the nature of changes in treatment regimen, changes in HbA1c , the proportion of patients achieving the HbA1c target of < 7%, the number of episodes of symptomatic and asymptomatic hypoglycaemia, microalbuminuria, and changes in proteinuria and serum creatinine. 2.6. Statistical analysis 2.6.1. Sample size calculation The target sample size for the study was determined a priori by power calculations to detect a hazard ratio (HR) of 0.85 for the difference in treatment duration in the M-Sita and M-SU groups with a power of 90% and an ␣ risk of 0.05. This HR corresponds to the difference in treatment duration observed in a randomized interventional phase-III trial comparing patients treated with MSita and those treated with metformin and glipizide [7]. To this end, it was necessary to recruit two equivalent groups of 1400 patients, each group receiving either M-Sita or M-SU. 2.6.2. Descriptive statistics These statistics included frequencies (n) and percentages for categorical variables, and means ± SD for continuous variables. Missing data were not replaced but excluded from the descriptive analyses. Patient characteristics at baseline were compared using Student’s t-test for continuous variables, and 2 or Fisher’s exact test for categorical variables. Changes in clinical variables over the study follow-up period were compared between groups using analysis of variance (ANOVA), with repeated measures or a mixed model with repeated measures. All statistical tests were two-sided, and a probability threshold of 0.05 was taken as statistically significant. 2.6.3. Primary endpoint analysis The primary endpoint (treatment maintenance duration) was compared between the M-Sita and M-SU groups using Kaplan–Meier survival analysis and a log-rank test; all premature follow-up discontinuations without documented treatment modification were censored. As the two groups to be compared were not randomized, a propensity score was calculated to describe each patient at inclusion. This score was generated using multiple logistic regression analyses of outcome as a function of age, gender, duration of diabetes, comorbidities, previous treatments, HbA1c at inclusion and physician’s characteristics. The primary endpoint analysis was then weighted by the propensity score to limit any bias due to the imbalance between the
3
two groups. Again, all statistical tests were two-sided and a probability threshold of 0.05 considered statistically significant. Three sensitivity analyses of the primary endpoint were performed to evaluate the robustness of the data and are provided as supplementary material (Table S1) online. 2.7. Ethical considerations This study was performed in accordance with Good Epidemiological Practice guidelines, and other relevant international and national guidelines. All patients were informed of the goals of the study and invited to participate. However, because participation in the study did not influence the care received, patients’ written informed consent was not required. All patient data were rendered anonymous before being included in the study database. The study protocol was submitted to the Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la Santé (CCTIRS; Advisory Committee on Information Processing of Research in the Field of Health) and Commission Nationale de l’Informatique et des Libertés (CNIL; National Commission for Information Technology and Civil Liberties), which ensures that all medical information is kept confidential and anonymous. 3. Results 3.1. Participating physicians Of the 1569 physicians who agreed to take part in the study, 705 included at least one eligible patient. These 705 GPs are representative of all GPs in France in terms of age, gender and geographical distribution (data not shown). 3.2. Patients Overall, 4031 patients were enrolled in the study, of whom 578 failed to fulfil the eligibility criteria, mainly due to inconsistencies in the data for baseline medication (369 patients) or due to missing information on treatment in the case report form (184 patients), leaving 3453 eligible patients. Of these, 846 patients were prescribed sitagliptin as part of another treatment regimen. The remaining patients comprised the analysis population (Fig. 1). Of the 2607 eligible patients analyzed here, 821 (31.5%) discontinued the study before the end of the 36-month follow-up period stipulated by the protocol, most frequently within the first 6 months of the study. This proportion did not differ between the two treatment groups (31.9% for the M-Sita group and 30.4% for the M-SU group). Characteristics of the study population are presented in Table 1. A number of small but significant differences were observed between the M-Sita and M-SU groups. The former were, on average, younger (P < 0.001) with shorter disease duration (P = 0.049) and had, on average, higher body mass index (BMI) scores (P = 0.004). Age at the time of inclusion and gender of the patients in each treatment group analyzed did not
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 4
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
Enrolled paents n = 4031 Inconsistent treatment informaon (n = 369) Missing treatment informaon (n = 184) Major protocol deviaons (n = 25) Eligible paents n = 3453 (85.7%) Other sitaglipn regimens (n = 846) Analysis populaon n = 2607 (75.5%)
M-Sit group N = 1874 (71.9%)
M-SU group n = 733 (28.1%)
Fig. 1. Patient flow and distribution between treatment groups. Percentages were calculated in relation to the previous box. M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin.
differ significantly from those of all patients starting the same treatment, as documented by the LPD database (data not shown). In the M-SU group, fixed-dose combinations were prescribed to 59 patients (8.0%). The most frequently prescribed sulphonylureas were gliclazide (n = 395; 53.9%), glibenclamide (n = 176; 24.0%) and glimepiride (n = 158; 21.6%). 3.3. Treatment maintenance Over the follow-up period, 341 patients (46.5%) in the M-SU group and 621 patients (33.1%) in the M-Sita group underwent a change in initial dual therapy (Table 2). In both patient groups, the change most frequently meant the addition of another OAD (Table 2). In the M-SU group, the sulphonylurea was frequently
replaced by another agent. None of the other possible changes involved more than 10% of patients. In the primary effectiveness analysis, the median treatment maintenance duration was 20.2 months [95% confidence interval (CI): 17.0–25.1 months] in the M-SU group and 43.2 months (95% CI: 41.4–non-estimable months) in the M-Sita group. This difference was statistically significant (log-rank test: P < 0.0001). Kaplan–Meier survival curves for treatment maintenance are presented in Fig. 2. In the multivariate analyses adjusted by the propensity score, the HR for undergoing a change in treatment regimen was 0.65 (95% CI: 0.57–0.73). Three sensitivity analyses were performed and support the longer treatment maintenance duration in the M-Sita group (Table S1; see supplementary material associated with this article online).
Table 1 Patients’ characteristics at baseline. M-SU group (n = 733)
M-Sita group (n = 1874)
P
Age (years)
64.2 ± 11.5 [n = 733]
62.4 ± 10.8 [n = 1866]
< 0.001
Gender (male)
422 (57.6%) [n = 733]
1108 (59.4%) [n = 1866]
0.4
29.6 ± 5.4 [n = 686]
30.3 ± 5.2 [n = 1738]
0.004
BMI
(kg/m2 )
Time since diagnosis (years)
7.0 ± 5.6 [n = 644]
6.4 ± 5.9 [n = 1702]
0.049
HbA1c (%) < 6.5% 6.5–7% > 7–8% > 8–9% > 9% Fasting blood glucose (g/L)
7.6 ± 1.0 [n = 678] 74 (10.9%) 139 (20.5%) 280 (41.3%) 124 (18.3%) 61 (9.0%) 1.53 ± 0.40 [n = 512]
7.5 ± 1.0 [n = 1735] 180 (10.4%) 396 (22.8%) 736 (42.4%) 282 (16.3%) 141 (8.1%) 1.55 ± 0.38 [n = 1348]
0.092
Previous therapy, n (%) None OAD monotherapy OAD dual therapy OAD triple therapy OAD and GLP-1 analogue Insulin therapy or other
48 (6.7%) 415 (57.6%) 235 (32.6%) 8 (1.1%) 8 (1.1%) 6 (0.8%)
100 (5.4%) 1210 (65.3%) 464 (25.0%) 30 (1.6%) 38 (2.1%) 11 (0.6%)
0.218
Data are presented as means ± SD (continuous data) or as number and percentages; M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin; BMI: body mass index; OAD: oral antidiabetic drug; GLP-1: glucagon-like peptide-1
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
5
Fig. 2. Treatment maintenance duration up to a change in initial dual therapy. Shaded areas represent the 95% Hall–Wellner bands. Figures along the horizontal axis correspond to numbers of patients at risk at each time point. Red: metformin and sulphonylurea group; blue: metformin and sitagliptin group.
The principal reason for changing treatment in both groups was insufficient efficacy, accounting for approximately twothirds of documented cases (Table 2). The proportion of patients whose treatment was changed due to hypoglycaemia was higher in the M-SU group (13.5%) than in the M-Sita group (4.2%). 3.4. Clinical outcomes In both study arms, a reduction in HbA1c level of approximately –0.6% was observed during the first 6 months of treatment, a reduction that was maintained up to the end of the observation period (Fig. 3). The proportion of patients achieving an HbA1c < 7% at least once during the follow-up period, and before a change in treatment, was higher in the M-Sita group than in the M-SU group (Table 2). The time to reach an HbA1c < 7% was somewhat shorter in the M-Sita group than in the M-SU group (P = 0.045; log-rank test). No difference was observed in the time to reach HbA1c < 6.5% (P = 0.34). The proportion of patients reporting at least one symptomatic hypoglycaemia episode during the follow-up period was higher with M-SU than with M-Sita (Table 2). The proportion of patients reporting hypoglycaemia in their diaries was low. Variations in body weight during follow-up were minor and of a similar scale between the two treatment groups (P = 0.985; repeated measures ANOVA). 3.5. Safety Adverse events were reported by 7.9% of patients in the M-SU group (58/733) and by 6.9% of those in the M-Sita group (130/1874). The most frequent events were those affecting the gastrointestinal (GI) system, notably diarrhoea, nausea and abdominal pain. There was no obvious difference in the
nature or frequency of adverse events between the two treatment groups. Serious adverse events were reported in 4.8% of patients (n = 35) in the M-SU group and by 3.6% (n = 68) in the M-Sita group. Only a minority of events were considered treatmentrelated. These were observed in 2.7% of patients (n = 20) in the M-SU group and in 2.8% (n = 52) in the M-Sita group. Of these, 1.6% (16 events in 12 patients) in the M-SU group and 2.3% (50 events in 43 patients) in the M-Sita group were related to the GI system. Three patients in the M-SU group and seven patients in the M-Sita group experienced adverse events that led to treatment discontinuation. Six serious adverse events were considered related to treatment: these included one case of myalgia, two cases of hypoglycaemia and one case of unstable diabetes in the M-SU group; and one case of renal failure and one case of hypoglycaemia in the M-Sita group. No deaths related to treatment were reported during the study. 4. Discussion The primary objective of the present study was to compare, in everyday clinical practice, treatment maintenance duration in patients with T2D, and inadequate glycaemic control with metformin monotherapy, who were then treated with dual therapy with either M-Sita or M-SU. Treatment maintenance duration is considered an indirect measure of the level of treatment satisfaction and acceptability by both physicians and patients. Our present study observed a highly significant difference in this comprehensive outcome measure, which corresponded to an approximately twofold increase in the median treatment maintenance duration of more than 1 year. During the follow-up period, 46.5% of patients in the M-SU group and 33.1% in the M-Sita group discontinued their prescribed dual therapy, the principal reason in both groups being insufficient efficacy. The proportion
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 6
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
Table 2 Primary and secondary endpoints. M-SU group (n = 733)
M-Sita group (n = 1874)
No treatment change
392 (53.5%)
1253 (66.9%)
Any treatment change Withdrawal of Sita or SU Withdrawal of metformin Addition of agent(s) Replacement of Sita or SU Replacement of metformin Replacement of all treatment Reasons for treatment changea Insufficient efficacy Poor tolerability Hypoglycaemia Other treatment event Patient decision Other
341 (46.5%) 33 (4.5%)
621 (33.1%) 73 (3.9%)
13 (1.8%)
50 (2.7%)
136 (18.6%)
333 (17.8%)
129 (17.6%)
114 (6.1%)
16 (2.2%)
31 (1.7%)
14 (1.9%)
20 (1.1%)
n = 215
n = 433
138 (64.2%) 25 (11.6%) 29 (13.5%) 6 (2.8%)
301 (69.5%) 57 (13.2%) 18 (4.2%) 7 (1.6%)
10 (4.7%) 38 (17.7%)
19 (4.4%) 54 (12.5%)
419 (58.8%)
1185 (64.8%)
250 (35.1%)
734 (40.1%)
127 (21.0%)
152 (9.7%)
10.20 ± 30.19
7.16 ± 17.45
–1.6 ± 7.0 73 (58.9%) 9 (7.3%) 42 (33.9%)
–2.5 ± 6.3 281 (62.4%) 56 (12.4%) 113 (25.1%)
HbA1c Patients achieving < 7% at least once [n (%)] Patients achieving < 6.5% at least once [n (%)] Hypoglycaemia At least one symptomatic episode [n (%)] Frequency of episodes (patient-years; mean ± SD) Weight change between inclusion and 36 months (kg) Mean ± SE Decreases Stable Increases
a Multiple responses were possible; M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin.
of patients who changed treatment due to hypoglycaemia was significantly higher with M-SU than with M-Sita. As differences in patient characteristics at inclusion were taken into account by the propensity score, these data can be taken to indicate that M-Sita dual therapy is associated with longer treatment maintenance independently of important confounding factors such as diabetes duration, comorbidities and HbA1c levels. Another French observational study (VILDA) evaluated treatment maintenance with a different DPP-4 inhibitor, vildagliptin [8], with maintenance defined less stringently as
discontinuation of vildagliptin. In that study, 11% of patients had discontinued vildagliptin after 2 years, a somewhat higher rate than that observed in our present study. However, it should be noted that VILDA was not restricted to patients receiving gliptin/metformin dual therapy. Glycaemic control until a change in treatment was a secondary outcome measure. The mean change in HbA1c levels until treatment change was approximately –0.6% in both treatment groups, and this reduction was maintained over the study duration. Approximately two-thirds of patients receiving M-Sita achieved an HbA1c level < 7% at least once during the course of the study. However, data on HbA1c were missing for a significant minority of patients, which means that these data should be interpreted with caution. The efficacy of sitagliptin observed in our study is similar to the efficacy reported in previous interventional trials [7,9], and is also very similar to that reported in the real-life VILDA study [8]. As in randomized clinical trials of sitagliptin [7,9], our present study found a significantly lower frequency of symptomatic hypoglycaemic episodes reported in patients taking M-Sita compared with M-SU. In addition, the proportion of patients who stopped treatment due to unacceptable hypoglycaemia was three times lower in the M-Sita group. The observed rates of hypoglycaemia were close to those reported in another recent observational study [10], where 6.4% of elderly patients prescribed a DPP-4 inhibitor and 26% of those prescribed an insulin secretagogue experienced at least one symptomatic hypoglycaemic event. In addition to the safety benefits obtained, patients may also be more adherent to medications carrying a lower risk of hypoglycaemia [11] and, thus, achieve a greater level of glycaemic control. The present study was not designed to collect exhaustive information on safety. However, the adverse events documented reflected the known safety profile of sitagliptin, and serious treatment-related adverse events occurred in < 1% of patients. However, the unanticipated challenge faced in implementing our study was the recruitment of comparable numbers of patients to the two treatment arms. In the end, twice as many patients were included in the M-Sita arm as in the M-SU arm. In part, this may reflect changes in prescribing practice as, over the last few years, DPP-4 inhibitors have replaced sulphonylureas as the OAD class most frequently associated with metformin in dual therapy in France [12]. However, an unconscious or voluntary inclusion bias on the part of the participating GPs, who may have been more assiduous in enrolling patients treated with sitagliptin, cannot be excluded. Nevertheless, any potential bias resulting from this should have been limited to some extent after adjusting the primary outcome with the propensity score. Our study has a number of limitations. As with all studies, the participation of physicians and patients was voluntary, and some resultant selection bias cannot be excluded. However, on comparing our included patients with all patients in the French LPD database with T2D prescribed the study treatments, no clinically relevant differences were observed. In addition, as with all observational studies, the absence of randomization may have led to an inclusion bias; however, to minimize such a bias, the primary
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model
ARTICLE IN PRESS
DIABET-688; No. of Pages 8
P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
7
7.8 meormin+sulphonylurea
7.7
meormin+sitaglipn 7.6
Group: Time: Interacon:
7.5
p=0,024 p
ARTICLE IN PRESS Available online at
ScienceDirect www.sciencedirect.com Diabetes & Metabolism xxx (2015) xxx–xxx
Original article
Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study P. Valensi a , G. de Pouvourville b , N. Benard c , C. Chanut-Vogel c , C. Kempf d , E. Eymard c,∗ , C. Moisan c , J. Dallongeville e a
Department of Endocrinology–Diabetology–Nutrition, Jean-Verdier Hospital, AP–HP, Paris Nord University, CRNH-IdF, CINFO, Bondy, France b ESSEC, Cergy-Pontoise, France c MSD France, Courbevoie, France d CSD, Boulogne, France e INSERM U744, Institut Pasteur, Lille, France Received 11 February 2015; received in revised form 17 March 2015; accepted 22 March 2015
Abstract Aim. – The study compared the duration of maintenance of treatment in patients with type 2 diabetes (T2D) using dual therapy with either metformin and sitagliptin (M-Sita) or metformin and a sulphonylurea (M-SU). Materials and methods. – This observational study included adult patients with T2D who had responded inadequately to metformin monotherapy and therefore had started de-novo treatment with Met-Sita or Met-SU within the previous eight weeks. Patient follow-up and changes to treatment were performed according to their general practitioner’s usual clinical practice. The primary outcome was time to change in treatment for whatever cause. HbA1c and symptomatic hypoglycaemia were also documented. Results. – The median treatment duration for patients in the M-Sita group (43.2 months) was significantly longer (P < 0.0001) than in the M-SU group (20.2 months). This difference persisted after adjusting for baseline differences and confounders. A similar reduction in HbA1c was noted in both arms (–0.6%), and the incidence of hypoglycaemia prior to treatment modification was lower with M-Sita (9.7%) than with M-SU (21.0%). Adverse events potentially related to treatment were reported in 2.8% (n = 52) and 2.7% (n = 20) of patients in the M-Sita and M-SU arms, respectively. Conclusion. – Under everyday conditions of primary diabetes care, dual therapy with M-Sita can be maintained for longer than M-SU. In addition, while efficacy, as measured by changes in HbA1c , was similar between treatments, the incidence of hypoglycaemia was lower in patients taking M-Sita. © 2015 Elsevier Masson SAS. All rights reserved. Keywords: DPP-4 inhibitor; Observational study; Type 2 diabetes
1. Introduction Type 2 diabetes (T2D) is a major public-health challenge in Europe and in most other regions of the world [1]. In the absence of effective treatment, T2D is associated with significant morbidity and mortality, which are mostly attributable to long-term
∗ Corresponding author. 34, avenue Léonard-de-Vinci, 92400 Courbevoie, France. Tel.: +33 1 80 46 44 76. E-mail address: [email protected] (E. Eymard).
cardiovascular complications of the disease. Nevertheless, the risk of long-term complications can be significantly reduced by improving glycaemic control [2]. To this end, although several studies have shown residual dysglycaemia of 6.5% to 6.9% [3], the treatment goal in T2D is to restore and maintain glycaemia at as near to normal levels as possible, defined as an HbA1c level ≤ 7%. Since the end of the 1990s, available choices for the treatment of T2D have expanded considerably. This has allowed individualization of treatment in response to the specific clinical features and history of each patient. As T2D is a chronic disease with
http://dx.doi.org/10.1016/j.diabet.2015.03.007 1262-3636/© 2015 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 2
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
onset usually in older middle-aged people, the challenge for the physician is to choose drug therapies that ensure durable glycaemic control while minimizing the need for intensification of treatment. This will retard disease progression during the course of therapy and delay the exhaustion of effective therapeutic options. Although the efficacy of new oral antidiabetic drugs (OADs) has been convincingly demonstrated in phase-III clinical trials, their comparative effectiveness in everyday clinical practice remains poorly characterized. In this respect, data from observational studies can make an important contribution to the evidence base [4]. For this reason, this observational study (ODYSSEE) was undertaken to compare dual therapy with metformin and sitagliptin [5], a dipeptidyl peptidase (DPP)-4 inhibitor, with dual therapy with metformin and a sulphonylurea in patients failing to respond adequately to metformin monotherapy. As a comprehensive measure of efficacy, the study evaluated treatment maintenance duration, defined as the time from starting treatment until the physician chose to change the treatment. Treatment maintenance should be distinguished from treatment persistence, which refers to the length of time a patient takes medication as prescribed. Treatment maintenance duration is a simple and reliable variable to document outcome in observational studies in real-life treatment settings. It covers all circumstances relating to changes in treatment, including inadequate efficacy, unacceptable occurrence of hypoglycaemia or other adverse events, and poor patient acceptability. The primary objective of the present study was to assess whether metformin plus sitagliptin dual therapy was superior to metformin plus sulphonylurea dual therapy in terms of treatment maintenance duration in patients with T2D. Secondary objectives were to describe treatment regimens using sitagliptin in standard clinical practice and the profiles of such treated patients, and to document the evolution of the clinical status of such patients during treatment. Information on the treatment regimens and patient profiles are to be presented elsewhere. 2. Methods This was a multicentre longitudinal, prospective, observational study conducted in primary care in France. Patients were enrolled between July 2009 and December 2010. The planned follow-up period for each patient was three years, and the last follow-up visit took place in August 2013. 2.1. Participating investigators A group of general practitioners (GPs) in metropolitan France were randomly identified from an extensive national listing of practising physicians by specialty (Cegedim OneKey registries; Cegedim, Boulogne-Billancourt, France). From this group, physicians who provided care to large numbers of patients with T2D were invited to participate. Of the 13,521 primary care practitioners contacted, 1569 agreed to take part in the study. These participating GPs were compared with all GPs in France using data from the National Health Insurance database.
2.2. Patients Participating physicians were expected to consecutively include all patients fulfilling the eligibility criteria who had consulted them during an enrolment period of nine months. Adult (> 18 years) patients with T2D who had initiated de-novo dual therapy with either metformin and sitagliptin (M-Sita group) or metformin and sulphonylurea (M-SU group) within the previous eight weeks were eligible for the study. Also, only patients who, in the physician’s judgment, were equally eligible for a prescription with either sitagliptin or a sulphonylurea were to be included. Allocation of patients to a given medical strategy was not predefined, but based on the GPs’ standard practices, and the prescribing decision was clearly separate from the decision to include the patient in the study. Characteristics of the study patients were compared with those of patients documented in the Cegedim longitudinal patient data (LPD) primary care database, a large panel of patients consulting one of 1200 French primary care physicians considered representative of all primary care practitioners in terms of age, gender and place of practice. 2.3. Treatment Treatments were prescribed at the discretion of the physician. The possible treatment regimens required for enrolment in the study were sitagliptin 100 mg (either as dual therapy with metformin or as part of another treatment regimen) or dual therapy with metformin and sulphonylurea, irrespective of dose. To limit as much as possible any influence of study participation on treatment prescription, physicians were asked to include all patients prescribed sitagliptin, regardless of the treatment regimen. However, only patients prescribed the M-Sita dual therapy were intended for inclusion in the primary analysis. 2.4. Data collection and patient follow-up At the time of study inclusion, the patients’ demographic data and medical characteristics were documented, including comorbidities, recent laboratory tests, previous medications, prescribed medicinal products and reasons for prescribing sitagliptin or sulphonylurea. Patient follow-up and changes to treatment were performed according to the physician’s everyday clinical practice and were not specified by the study protocol. However, the protocol did stipulate that the physician was to document the patient at the time of the patient’s usual quarterly follow-up visits (as recommended by French national guidelines for the monitoring of patients with T2D [6]) over a 3-year period. The information collected included any changes to the treatment regimen and the reasons for change, patient-reported health and dietary habits, laboratory measures of HbA1c , microalbuminuria or urinary albumin excretion rate (UAER), proteinuria and serum creatinine. The onset and number of symptomatic hypoglycaemic episodes were to be documented by the physicians; blood measures were also collected if available and patients could complete a diary. In this diary, patients could report on body weight, occurrence and severity of hypoglycaemic episodes, glycaemia
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
values and any interruption of treatment. If information was collected by the physician by telephone instead of during a visit, this was also documented. 2.5. Evaluation criteria The primary study endpoint was the duration of maintenance of dual therapy expressed as the number of days from treatment initiation to treatment change, defined as the addition, replacement or withdrawal of any agent used in the initial dual therapy. Changes in dose of the initially prescribed dual therapy were not considered treatment changes. Secondary endpoints included the nature of changes in treatment regimen, changes in HbA1c , the proportion of patients achieving the HbA1c target of < 7%, the number of episodes of symptomatic and asymptomatic hypoglycaemia, microalbuminuria, and changes in proteinuria and serum creatinine. 2.6. Statistical analysis 2.6.1. Sample size calculation The target sample size for the study was determined a priori by power calculations to detect a hazard ratio (HR) of 0.85 for the difference in treatment duration in the M-Sita and M-SU groups with a power of 90% and an ␣ risk of 0.05. This HR corresponds to the difference in treatment duration observed in a randomized interventional phase-III trial comparing patients treated with MSita and those treated with metformin and glipizide [7]. To this end, it was necessary to recruit two equivalent groups of 1400 patients, each group receiving either M-Sita or M-SU. 2.6.2. Descriptive statistics These statistics included frequencies (n) and percentages for categorical variables, and means ± SD for continuous variables. Missing data were not replaced but excluded from the descriptive analyses. Patient characteristics at baseline were compared using Student’s t-test for continuous variables, and 2 or Fisher’s exact test for categorical variables. Changes in clinical variables over the study follow-up period were compared between groups using analysis of variance (ANOVA), with repeated measures or a mixed model with repeated measures. All statistical tests were two-sided, and a probability threshold of 0.05 was taken as statistically significant. 2.6.3. Primary endpoint analysis The primary endpoint (treatment maintenance duration) was compared between the M-Sita and M-SU groups using Kaplan–Meier survival analysis and a log-rank test; all premature follow-up discontinuations without documented treatment modification were censored. As the two groups to be compared were not randomized, a propensity score was calculated to describe each patient at inclusion. This score was generated using multiple logistic regression analyses of outcome as a function of age, gender, duration of diabetes, comorbidities, previous treatments, HbA1c at inclusion and physician’s characteristics. The primary endpoint analysis was then weighted by the propensity score to limit any bias due to the imbalance between the
3
two groups. Again, all statistical tests were two-sided and a probability threshold of 0.05 considered statistically significant. Three sensitivity analyses of the primary endpoint were performed to evaluate the robustness of the data and are provided as supplementary material (Table S1) online. 2.7. Ethical considerations This study was performed in accordance with Good Epidemiological Practice guidelines, and other relevant international and national guidelines. All patients were informed of the goals of the study and invited to participate. However, because participation in the study did not influence the care received, patients’ written informed consent was not required. All patient data were rendered anonymous before being included in the study database. The study protocol was submitted to the Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la Santé (CCTIRS; Advisory Committee on Information Processing of Research in the Field of Health) and Commission Nationale de l’Informatique et des Libertés (CNIL; National Commission for Information Technology and Civil Liberties), which ensures that all medical information is kept confidential and anonymous. 3. Results 3.1. Participating physicians Of the 1569 physicians who agreed to take part in the study, 705 included at least one eligible patient. These 705 GPs are representative of all GPs in France in terms of age, gender and geographical distribution (data not shown). 3.2. Patients Overall, 4031 patients were enrolled in the study, of whom 578 failed to fulfil the eligibility criteria, mainly due to inconsistencies in the data for baseline medication (369 patients) or due to missing information on treatment in the case report form (184 patients), leaving 3453 eligible patients. Of these, 846 patients were prescribed sitagliptin as part of another treatment regimen. The remaining patients comprised the analysis population (Fig. 1). Of the 2607 eligible patients analyzed here, 821 (31.5%) discontinued the study before the end of the 36-month follow-up period stipulated by the protocol, most frequently within the first 6 months of the study. This proportion did not differ between the two treatment groups (31.9% for the M-Sita group and 30.4% for the M-SU group). Characteristics of the study population are presented in Table 1. A number of small but significant differences were observed between the M-Sita and M-SU groups. The former were, on average, younger (P < 0.001) with shorter disease duration (P = 0.049) and had, on average, higher body mass index (BMI) scores (P = 0.004). Age at the time of inclusion and gender of the patients in each treatment group analyzed did not
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 4
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
Enrolled paents n = 4031 Inconsistent treatment informaon (n = 369) Missing treatment informaon (n = 184) Major protocol deviaons (n = 25) Eligible paents n = 3453 (85.7%) Other sitaglipn regimens (n = 846) Analysis populaon n = 2607 (75.5%)
M-Sit group N = 1874 (71.9%)
M-SU group n = 733 (28.1%)
Fig. 1. Patient flow and distribution between treatment groups. Percentages were calculated in relation to the previous box. M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin.
differ significantly from those of all patients starting the same treatment, as documented by the LPD database (data not shown). In the M-SU group, fixed-dose combinations were prescribed to 59 patients (8.0%). The most frequently prescribed sulphonylureas were gliclazide (n = 395; 53.9%), glibenclamide (n = 176; 24.0%) and glimepiride (n = 158; 21.6%). 3.3. Treatment maintenance Over the follow-up period, 341 patients (46.5%) in the M-SU group and 621 patients (33.1%) in the M-Sita group underwent a change in initial dual therapy (Table 2). In both patient groups, the change most frequently meant the addition of another OAD (Table 2). In the M-SU group, the sulphonylurea was frequently
replaced by another agent. None of the other possible changes involved more than 10% of patients. In the primary effectiveness analysis, the median treatment maintenance duration was 20.2 months [95% confidence interval (CI): 17.0–25.1 months] in the M-SU group and 43.2 months (95% CI: 41.4–non-estimable months) in the M-Sita group. This difference was statistically significant (log-rank test: P < 0.0001). Kaplan–Meier survival curves for treatment maintenance are presented in Fig. 2. In the multivariate analyses adjusted by the propensity score, the HR for undergoing a change in treatment regimen was 0.65 (95% CI: 0.57–0.73). Three sensitivity analyses were performed and support the longer treatment maintenance duration in the M-Sita group (Table S1; see supplementary material associated with this article online).
Table 1 Patients’ characteristics at baseline. M-SU group (n = 733)
M-Sita group (n = 1874)
P
Age (years)
64.2 ± 11.5 [n = 733]
62.4 ± 10.8 [n = 1866]
< 0.001
Gender (male)
422 (57.6%) [n = 733]
1108 (59.4%) [n = 1866]
0.4
29.6 ± 5.4 [n = 686]
30.3 ± 5.2 [n = 1738]
0.004
BMI
(kg/m2 )
Time since diagnosis (years)
7.0 ± 5.6 [n = 644]
6.4 ± 5.9 [n = 1702]
0.049
HbA1c (%) < 6.5% 6.5–7% > 7–8% > 8–9% > 9% Fasting blood glucose (g/L)
7.6 ± 1.0 [n = 678] 74 (10.9%) 139 (20.5%) 280 (41.3%) 124 (18.3%) 61 (9.0%) 1.53 ± 0.40 [n = 512]
7.5 ± 1.0 [n = 1735] 180 (10.4%) 396 (22.8%) 736 (42.4%) 282 (16.3%) 141 (8.1%) 1.55 ± 0.38 [n = 1348]
0.092
Previous therapy, n (%) None OAD monotherapy OAD dual therapy OAD triple therapy OAD and GLP-1 analogue Insulin therapy or other
48 (6.7%) 415 (57.6%) 235 (32.6%) 8 (1.1%) 8 (1.1%) 6 (0.8%)
100 (5.4%) 1210 (65.3%) 464 (25.0%) 30 (1.6%) 38 (2.1%) 11 (0.6%)
0.218
Data are presented as means ± SD (continuous data) or as number and percentages; M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin; BMI: body mass index; OAD: oral antidiabetic drug; GLP-1: glucagon-like peptide-1
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
5
Fig. 2. Treatment maintenance duration up to a change in initial dual therapy. Shaded areas represent the 95% Hall–Wellner bands. Figures along the horizontal axis correspond to numbers of patients at risk at each time point. Red: metformin and sulphonylurea group; blue: metformin and sitagliptin group.
The principal reason for changing treatment in both groups was insufficient efficacy, accounting for approximately twothirds of documented cases (Table 2). The proportion of patients whose treatment was changed due to hypoglycaemia was higher in the M-SU group (13.5%) than in the M-Sita group (4.2%). 3.4. Clinical outcomes In both study arms, a reduction in HbA1c level of approximately –0.6% was observed during the first 6 months of treatment, a reduction that was maintained up to the end of the observation period (Fig. 3). The proportion of patients achieving an HbA1c < 7% at least once during the follow-up period, and before a change in treatment, was higher in the M-Sita group than in the M-SU group (Table 2). The time to reach an HbA1c < 7% was somewhat shorter in the M-Sita group than in the M-SU group (P = 0.045; log-rank test). No difference was observed in the time to reach HbA1c < 6.5% (P = 0.34). The proportion of patients reporting at least one symptomatic hypoglycaemia episode during the follow-up period was higher with M-SU than with M-Sita (Table 2). The proportion of patients reporting hypoglycaemia in their diaries was low. Variations in body weight during follow-up were minor and of a similar scale between the two treatment groups (P = 0.985; repeated measures ANOVA). 3.5. Safety Adverse events were reported by 7.9% of patients in the M-SU group (58/733) and by 6.9% of those in the M-Sita group (130/1874). The most frequent events were those affecting the gastrointestinal (GI) system, notably diarrhoea, nausea and abdominal pain. There was no obvious difference in the
nature or frequency of adverse events between the two treatment groups. Serious adverse events were reported in 4.8% of patients (n = 35) in the M-SU group and by 3.6% (n = 68) in the M-Sita group. Only a minority of events were considered treatmentrelated. These were observed in 2.7% of patients (n = 20) in the M-SU group and in 2.8% (n = 52) in the M-Sita group. Of these, 1.6% (16 events in 12 patients) in the M-SU group and 2.3% (50 events in 43 patients) in the M-Sita group were related to the GI system. Three patients in the M-SU group and seven patients in the M-Sita group experienced adverse events that led to treatment discontinuation. Six serious adverse events were considered related to treatment: these included one case of myalgia, two cases of hypoglycaemia and one case of unstable diabetes in the M-SU group; and one case of renal failure and one case of hypoglycaemia in the M-Sita group. No deaths related to treatment were reported during the study. 4. Discussion The primary objective of the present study was to compare, in everyday clinical practice, treatment maintenance duration in patients with T2D, and inadequate glycaemic control with metformin monotherapy, who were then treated with dual therapy with either M-Sita or M-SU. Treatment maintenance duration is considered an indirect measure of the level of treatment satisfaction and acceptability by both physicians and patients. Our present study observed a highly significant difference in this comprehensive outcome measure, which corresponded to an approximately twofold increase in the median treatment maintenance duration of more than 1 year. During the follow-up period, 46.5% of patients in the M-SU group and 33.1% in the M-Sita group discontinued their prescribed dual therapy, the principal reason in both groups being insufficient efficacy. The proportion
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model DIABET-688; No. of Pages 8 6
ARTICLE IN PRESS P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
Table 2 Primary and secondary endpoints. M-SU group (n = 733)
M-Sita group (n = 1874)
No treatment change
392 (53.5%)
1253 (66.9%)
Any treatment change Withdrawal of Sita or SU Withdrawal of metformin Addition of agent(s) Replacement of Sita or SU Replacement of metformin Replacement of all treatment Reasons for treatment changea Insufficient efficacy Poor tolerability Hypoglycaemia Other treatment event Patient decision Other
341 (46.5%) 33 (4.5%)
621 (33.1%) 73 (3.9%)
13 (1.8%)
50 (2.7%)
136 (18.6%)
333 (17.8%)
129 (17.6%)
114 (6.1%)
16 (2.2%)
31 (1.7%)
14 (1.9%)
20 (1.1%)
n = 215
n = 433
138 (64.2%) 25 (11.6%) 29 (13.5%) 6 (2.8%)
301 (69.5%) 57 (13.2%) 18 (4.2%) 7 (1.6%)
10 (4.7%) 38 (17.7%)
19 (4.4%) 54 (12.5%)
419 (58.8%)
1185 (64.8%)
250 (35.1%)
734 (40.1%)
127 (21.0%)
152 (9.7%)
10.20 ± 30.19
7.16 ± 17.45
–1.6 ± 7.0 73 (58.9%) 9 (7.3%) 42 (33.9%)
–2.5 ± 6.3 281 (62.4%) 56 (12.4%) 113 (25.1%)
HbA1c Patients achieving < 7% at least once [n (%)] Patients achieving < 6.5% at least once [n (%)] Hypoglycaemia At least one symptomatic episode [n (%)] Frequency of episodes (patient-years; mean ± SD) Weight change between inclusion and 36 months (kg) Mean ± SE Decreases Stable Increases
a Multiple responses were possible; M-SU: metformin and sulphonylurea; M-Sita: metformin and sitagliptin.
of patients who changed treatment due to hypoglycaemia was significantly higher with M-SU than with M-Sita. As differences in patient characteristics at inclusion were taken into account by the propensity score, these data can be taken to indicate that M-Sita dual therapy is associated with longer treatment maintenance independently of important confounding factors such as diabetes duration, comorbidities and HbA1c levels. Another French observational study (VILDA) evaluated treatment maintenance with a different DPP-4 inhibitor, vildagliptin [8], with maintenance defined less stringently as
discontinuation of vildagliptin. In that study, 11% of patients had discontinued vildagliptin after 2 years, a somewhat higher rate than that observed in our present study. However, it should be noted that VILDA was not restricted to patients receiving gliptin/metformin dual therapy. Glycaemic control until a change in treatment was a secondary outcome measure. The mean change in HbA1c levels until treatment change was approximately –0.6% in both treatment groups, and this reduction was maintained over the study duration. Approximately two-thirds of patients receiving M-Sita achieved an HbA1c level < 7% at least once during the course of the study. However, data on HbA1c were missing for a significant minority of patients, which means that these data should be interpreted with caution. The efficacy of sitagliptin observed in our study is similar to the efficacy reported in previous interventional trials [7,9], and is also very similar to that reported in the real-life VILDA study [8]. As in randomized clinical trials of sitagliptin [7,9], our present study found a significantly lower frequency of symptomatic hypoglycaemic episodes reported in patients taking M-Sita compared with M-SU. In addition, the proportion of patients who stopped treatment due to unacceptable hypoglycaemia was three times lower in the M-Sita group. The observed rates of hypoglycaemia were close to those reported in another recent observational study [10], where 6.4% of elderly patients prescribed a DPP-4 inhibitor and 26% of those prescribed an insulin secretagogue experienced at least one symptomatic hypoglycaemic event. In addition to the safety benefits obtained, patients may also be more adherent to medications carrying a lower risk of hypoglycaemia [11] and, thus, achieve a greater level of glycaemic control. The present study was not designed to collect exhaustive information on safety. However, the adverse events documented reflected the known safety profile of sitagliptin, and serious treatment-related adverse events occurred in < 1% of patients. However, the unanticipated challenge faced in implementing our study was the recruitment of comparable numbers of patients to the two treatment arms. In the end, twice as many patients were included in the M-Sita arm as in the M-SU arm. In part, this may reflect changes in prescribing practice as, over the last few years, DPP-4 inhibitors have replaced sulphonylureas as the OAD class most frequently associated with metformin in dual therapy in France [12]. However, an unconscious or voluntary inclusion bias on the part of the participating GPs, who may have been more assiduous in enrolling patients treated with sitagliptin, cannot be excluded. Nevertheless, any potential bias resulting from this should have been limited to some extent after adjusting the primary outcome with the propensity score. Our study has a number of limitations. As with all studies, the participation of physicians and patients was voluntary, and some resultant selection bias cannot be excluded. However, on comparing our included patients with all patients in the French LPD database with T2D prescribed the study treatments, no clinically relevant differences were observed. In addition, as with all observational studies, the absence of randomization may have led to an inclusion bias; however, to minimize such a bias, the primary
Please cite this article in press as: Valensi P, et al. Treatment maintenance duration of dual therapy with metformin and sitagliptin in type 2 diabetes: The ODYSSEE observational study. Diabetes Metab (2015), http://dx.doi.org/10.1016/j.diabet.2015.03.007
+Model
ARTICLE IN PRESS
DIABET-688; No. of Pages 8
P. Valensi et al. / Diabetes & Metabolism xxx (2015) xxx–xxx
7
7.8 meormin+sulphonylurea
7.7
meormin+sitaglipn 7.6
Group: Time: Interacon:
7.5
p=0,024 p
