key: cord-1052942-rjc0hiaf authors: Grenet, G.; Mekhaldi, S.; MAINBOURG, S.; AUFFRET, M.; Cornu, C.; Cracowski, J.-L.; Gueyffier, F.; Lega, J.-C.; CUCHERAT, M. title: DPP4 inhibitors and respiratory infection, a systematic review and meta-analysis of the CardioVascular Outcomes Trials conducted before the pandemic and implications for the management of diabetes during COVID-19 date: 2020-07-29 journal: nan DOI: 10.1101/2020.07.28.20163386 sha: a9f9f5fc1a2c1decd5f1c135e177ec552dc04a4d doc_id: 1052942 cord_uid: rjc0hiaf Background Association between DPP4 inhibitors and respiratory infection remains unclear. CardioVascular Outcomes Trials (CVOTs) conducted before the COVID-19 pandemic are available. We aimed to estimate the effect of DPP4 inhibitors on the risk of respiratory infections. Methods We updated a previous systematic review and meta-analysis, searching for CVOTs assessing a DPP4 inhibitor in patients with type 2 diabetes mellitus. We focused on placebo-controlled CVOTs. Our primary outcome was 'any respiratory infection'. We added a sensitivity analysis integrating non-CVOTs and active-controlled CVOTs. Findings We included 47 714 patients in five placebo-controlled CVOTs. Median follow-up ranged from 1.5 years to 3 years. 4 369 events of overall respiratory infection were reported (rate of 9.2%). DPP4 inhibitors were not associated with a different risk compared to placebo (RR = 0.99 [95%CI: 0.93; 1.04]). The sensitivity analysis integrating the non-CVOTs studies and the active-controlled CVOT reached 11 349 events among 82 644 participants (rate of 13.7%). DPP4 inhibitors were not associated with a different risk of overall respiratory infection (RR = 1.00 [95% CI: 0.97; 1.03]). Interpretation Our up-dated meta-analysis provides the most powerful and least biased estimation of the association of DPP4 inhibitors and the risk of overall (non COVID-19) respiratory infection. We did not find any effect of the DPP4 inhibitors on the risk of respiratory infection. Our results support the recently published practical recommendations for the management of diabetes in patients with COVID-19, suggesting that DPP4 inhibitors should not be discontinued regarding the COVID-19 pandemic. Evidence before this study From before the COVID19 pandemic, respiratory infections are considered potential adverse effects of DPP4 inhibitors. Randomized trials assessing DPP4 inhibitors in patients with type 2 diabetes (T2D), their meta-analyses and pharmacovigilance studies reported conflicting results. Since the last meta-analyses assessing the risk of infections with DPP4 inhibitors, powerful cardiovascular outcomes randomized trials (CVOTs) became available. Recent practical recommendations for the management of diabetes during COVID-19 suggested that DPP4 inhibitors could be continued. We updated our previous meta-analysis of CVOTs and focused to the overall risk of respiratory infection associated with DPP4 inhibitors. We searched for published and unpublished CVOTs in Medline, the Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov, up to January 27, 2020, using key word as "diabetes mellitus", "hypoglycemic agents", "glucose control", "randomized controlled trial", "cardiovascular diseases". We included CVOTs comparing a DPP4 inhibitor versus placebo, in people with T2D, and analysed the risk of respiratory infection with DPP4 inhibitors. We focused on placebocontrolled CVOTs to avoid the pitfalls of small study effect and heterogeneous comparators. We added a sensitivity analysis integrating non-CVOTs and non-placebo CVOTs to challenge our results and to increase the statistical power. Our meta-analysis provides the most powerful and least biased estimation of the association of DPP4 inhibitors and the risk of overall (non COVID-19) respiratory infection. Our analyses integrated 11 349 events of any respiratory infections through 82 644 patients from randomized trials. Our results did not find any association between DPP4 inhibitors use and risk of non-COVID respiratory infections. The current COVID-19 pandemic has raised some questions about pros and cons of certain cardiovascular drugs. Our results support the recent practical recommendations for the management of diabetes in patients with COVID-19, suggesting that DPP4 inhibitors should not be discontinued regarding the COVID-19 pandemic. Background Association between DPP4 inhibitors and respiratory infection remains unclear. CardioVascular Outcomes Trials (CVOTs) conducted before the COVID-19 pandemic are available. We aimed to estimate the effect of DPP4 inhibitors on the risk of respiratory infections. We updated a previous systematic review and meta-analysis, searching for CVOTs assessing a DPP4 inhibitor in patients with type 2 diabetes mellitus. We focused on placebo-controlled CVOTs. Our primary outcome was 'any respiratory infection'. We added a sensitivity analysis integrating non-CVOTs and active-controlled CVOTs. We included 47 714 patients in five placebo-controlled CVOTs. Median follow-up ranged from 1·5 years to 3 years. 4 369 events of overall respiratory infection were reported (rate of 9·2%). DPP4 inhibitors were not associated with a different risk compared to placebo (RR = 0·99 [95% CI: 0·93; 1·04]). The sensitivity analysis integrating the non-CVOTs studies and the activecontrolled CVOT reached 11 349 events among 82 644 participants (rate of 13·7%). DPP4 inhibitors were not associated with a different risk of overall respiratory infection (RR = 1·00 [95% CI: 0·97; 1·03]). Our up-dated meta-analysis provides the most powerful and least biased estimation of the association of DPP4 inhibitors and the risk of overall (non COVID-19) respiratory infection. We did not find any effect of the DPP4 inhibitors on the risk of respiratory infection. Our results support the recently published practical recommendations for the management of diabetes in patients with COVID-19, suggesting that DPP4 inhibitors should not be discontinued regarding the COVID-19 pandemic. No source of funding Keywords: Type 2 diabetes; DPP4 inhibitors; respiratory infections; meta-analysis . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) From before the COVID19 pandemic, respiratory infections are considered to be potential adverse effect of dipeptidyl peptidase-4 (DPP-4) inhibitors. 1,2 Some randomized clinical trials (RCT) suggested an increased risk of respiratory (and urinary) infections with sitagliptin; 3 inhibitors. 5 They did not find an increased risk of infection, despite including 74 studies. However, they included only one cardiovascular outcomes trials (CVOT) -SAVOR TIMI 53-, 11 which weighted 30% of the meta-analysis of overall infection. SAVOR TIMI 53 did not appear in the meta-analysis for respiratory infection. 5 Since then, at least three other CVOTs assessing a gliptin versus placebo with a large sample size became available. [12] [13] [14] The last metaanalyses of CVOTs did not report such outcomes (respiratory infections) for DPP4 inhibitors. 15, 16 We did not find on-going meta-analysis specific to this question in PROSPERO (April 29, 2020). Pharmacovigilance studies also showed some discrepancies. A disproportionality analysis in the World Health Organization VigiBase reported an increased reporting of upper respiratory tract infections for users of DPP4 inhibitors, 17 not confirmed in other pharmacoepidemiological studies. [18] [19] [20] [21] Up to now, the potential effect of DPP4 inhibitors on overall respiratory infections remain unclear. In France, adverse events in patients with COVID-19 are already reported as suspected DPP4 inhibitor's adverse effect, whenever the causality remains unclear. 22 Recently published practical recommendations for the management of diabetes in patients with COVID-19 suggested that DPP4 inhibitors should not be discontinued. 23 On another hand, the DPP-4 enzyme could be a potential target for treating coronavirus infections. 24 We hypothesized that the meta-analysis of the powerful CVOTs would lead to a significantly increase in the statistical power of the analysis. Indeed, the sample sizes of the SAVOR TIMI 53 11 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint controlled CVOTs should provide an estimate at lower risk of bias compared to pharmacovigilance studies. We aimed to provide a powerful and low bias update estimate of the effect of DPP4 inhibitors on the overall risk of respiratory infections, allowing testing the recent recommendations for the management of diabetes during COVID-19 pandemic. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint We focused on respiratory infections in placebo-controlled CVOT assessing a DPP4 inhibitor. We updated a previous systematic review and meta-analysis. 25, 26 The search strategy was previously published. 26 We updated the searche for CVOTs assessing a DPP4 inhibitor versus a placebo, in type 2 diabetes mellitus, in PubMed and CENTRAL, up to January 27, 2020. We excluded active control as other gliptin, and other hypoglycemic drug, to avoid heterogeneity of control groups and inconsistency issues of indirect comparisons. We excluded specific populations as children and pregnant women. 25 Two reviewers independently review the bibliographic references (SMe and GG). consensus. In line with our previous meta-analysis, 26 we used the Cochrane Collaboration's tool for assessing risk of bias in RCTs. 28 Our primary outcome was the treatment effect on the risk of 'any respiratory infection'. Our primary analysis was the meta-analysis of the placebo-controlled CVOTs only, to avoid the risk of "small study effects" 29 and the potential heterogeneity of different comparators. We added a sensitivity analysis integrating i) the combined non-CVOTs from the previous meta-. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint analysis of Yang et al, 5 and ii) active-controlled CVOT (assessing a gliptin versus other hypoglycemic drug, if applicable). In this way, we were able to combine the information of smaller and shorter RCTs already summarized by Yang et al, and the last powerful CVOTs. Such analysis using smaller studies combined and larger trials apart has been successfully used before. 30 Treatment effect was estimated using risk ratio (RR) and its 95% confidence intervals (95% CI). P-value <0·05 was considered as significant. Analyses were performed using R 3.6.2 (Package [meta], version 4.11-0). 31 The heterogeneity between studies was assessed using I² test. We used fixed-effect model when I² was <50%, random-effects model otherwise. Reporting of the study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. 32 Publication bias was searched for using funnel plot. The protocol of the present review has not been registered, but the hypothesis, the bibliographic search, the outcome definition and the statistical plan were defined a priori, notably using our previous meta-analysis (which was registered in PROSPERO), 26 and the previous meta-analysis of Yang et al, specifically designed for assessing the risk of infections with DPP4 inhibitors. 5 Role of the funding source There was no funding source for this study. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint The results refers to the primary objective (placebo-controlled CVOTs), except for the sensitivity analysis of the primary outcome's section. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist is available in appendix (Appendix p2-3) . The update bibliographic search retrieved 762 references. The selection process of the update is presented in figure 1. Five trials with 47 714 patients were included. [11] [12] [13] [14] 33 Complementary to the previous systematic review, this update identified the final publication of one trial, 14 and one early-terminated trial "MK-3102-018". 33 The CAROLINA CVOT was not included in the primary analysis but integrated in the sensitivity analysis, as it compared a DPP4 inhibitor versus a sulfonylurea. 34 Baseline characteristics of included trials are presented in Table 1 . Included trials were published between 2013 and 2019. Median follow-up ranged from 1·5 years to 3 years. Percentage of males ranged from 62·9 to 70·7%, percentage of patients i) with high blood pressure or receiving antihypertensive drugs from 78·4 to 95·4%, ii) with dyslipidemia or receiving statins treatment from 71·2 to 90·4%, iii) receiving antiplatelet treatment from 68·3 to 97·2%, and percentage of current smokers at inclusion from 10·2 to 14·4%. Mean age ranged from 61 to 65·8+/-9·1 years, mean duration of diabetes from around 7·2 to 14·7+/-9·5 years, mean HbA1c at inclusion from 7·2+/-0·5 to 8·01+/-0·87%, mean body mass index (BMI) at inclusion from 28·7 to 31+/-5·3 kg.m -2 . The risk of bias assessments is presented in appendix (Appendix p4). All studies were randomized double-blinded placebo controlled trials. The risk of reporting bias of respiratory infections was unclear for all the included study. Only sparse data regarding respiratory infection were available for the TECOS trial. 13 Contacted, the sponsor explained that the safety data were unfortunately not available for safety meta-analysis. Definitions of the events for each trial are presented in appendix (Appendix p5). Five studies contributed to this analysis, including 4 369 events among 47 714 patients (rate of 9·2%). DPP4 inhibitors were not associated with a different risk of overall respiratory infection compared to placebo (RR = 0·99 [95% CI: 0·93; 1·04]). The heterogeneity was low (0%). Forest plot is shown in figure 2. Visual analysis of the funnel plot did not suggest a publication bias (see Appendix page 6). . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint Our up-dated meta-analysis provides the most powerful and least biased estimation of the association of DPP4 inhibitors and the risk of overall (non COVID-19) respiratory infection. We did not find any effect of the DPP4 inhibitors on the risk of respiratory infection. Our results support the recently published practical recommendations for the management of diabetes in patients with COVID-19, suggesting that DPP4 inhibitors should not be discontinued regarding the COVID-19 pandemic. 23 Our study has strengths and limitations. To the best of our knowledge, our study is the most comprehensive meta-analysis of large CVOTs assessing the risk of respiratory infection with DPP4 inhibitors, in patients with type 2 diabetes, before the COVID-19 pandemic. We even found one early-terminated CVOT not included in the last meta-analysis of glucose lowering drugs. 16 We summarized the information from five placebo-controlled CVOTs, including 47 714 patients. We firstly focused on placebo-controlled CVOTs, to avoid both the pitfalls of small study effect and of heterogeneous comparators. This could lead to a loss of information. However, the non-CVOTs have a relatively small sample size and time of follow up compared to the CVOTs. Thus, we expected the CVOTs to drive the main number of events. A previous meta-analysis of non-CVOTs DPP4 inhibitors studies provided treatment effect estimates on cardiovascular outcomes eventually disproved by CVOT. 35 To limit the risk of a loss of information with our methodological choice, we added a sensitivity analysis to verify the treatment effect estimation using both active-controlled CVOTs and non-CVOTs studies, which showed the robustness of our primary analysis. The placebo-controlled CVOTs reported more events (4 369) than the previous meta-analysis of non-CVOTs studies (3 904). 5 Integrating the placebo-controlled CVOTs, the active-controlled CVOT and the non-CVOTs studies, we provide a very precise estimation using an overall sample size of 11 349 events and 82 644 patients. Finally, we provide an estimation with a lower risk of bias and a greater statistical power compared to the previous pharmacovigilance studies. Indeed, they reported a number of events in patients exposed to DPP4 inhibitors ranging from 25 hospitalizations for communityacquired pneumonia, 18 87 pneumonia, 20 242 reports of any infections, 17 537 acute respiratory infections (compared to sulfonylureas), 19 and to 2440 infections overall. 21 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint We considered the risk of reporting bias as unclear for all the included study in our metaanalysis. Indeed, our outcomes of interest were not prespecified outcomes of interest of the included trials. Moreover, the reporting was heterogeneous between trials. This is especially the case for the TECOS trial, which surprisingly reported a very low rate of respiratory infections, despite a large sample size. However, the risk of reporting bias within each trial (i.e. between groups in each trial) was not considered as high, as most of the adverse events were "collected by systematic assessment" (see their report on the Clinical Trial website), in a double-blinded fashion, and coded using the Medical Dictionary for Regulatory Activities (MedDRA) definitions. Finally, as only one placebo-controlled CVOT was available per DPP4 inhibitor, we cannot exclude a molecule effect. It should be noted that the present review focused on the risk of respiratory infections and DPP4 inhibitors. The absence of demonstrated cardiovascular benefits of DPP4 inhibitors use in patients with type 2 diabetes, contrary to glucagon-like peptide-1 (GLP1) agonist receptors and Sodium-glucose Cotransporter-2 (SGLT2) inhibitors, has been documented before. 15, 26 The question of the role of the DPP4 inhibitors in type 2 diabetes management in general is beyond the scope of this paper. The current COVID-19 pandemic has raised many questions, especially regarding type 2 diabetes management. 23 We did not find any association between DPP4 inhibitors use and the risk of non-COVID respiratory infection. On-going randomized trial assessing sitagliptin in COVID-19 positive patients will help further understand the relationship between DPP4inhibitors and COVID-19 infections. 36 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 29, 2020. . for age, diabetes duration, HbA1c and body mass index at baseline and time of follow up unless otherwise indicated ("*": median and interquartile range if available). . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 29, 2020. . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.28.20163386 doi: medRxiv preprint Sitagliptin Study 024 Group. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, sitagliptin, compared with the sulfonylurea, glipizide, in patients with type 2 diabetes inadequately controlled on metformin alone: a randomized, double-blind, non-inferiority trial Efficacy and safety of alogliptin added to metformin in Japanese patients with type 2 diabetes: a randomized, double-blind, placebocontrolled trial with an open-label, long-term extension study DPP-4 inhibitors and risk of infections: a meta-analysis of randomized controlled trials Initial combination of linagliptin and metformin improves glycaemic control in type 2 diabetes: a randomized, doubleblind, placebo-controlled study Dipeptidyl peptidase-4 (DPP-4) inhibitors for type 2 diabetes mellitus Dipeptydil peptidase-4 inhibitors in type 2 diabetes: a meta-analysis of randomized clinical trials Longer term safety of dipeptidyl peptidase-4 inhibitors in patients with type 2 diabetes mellitus: systematic review and meta-analysis The efficacy and safety of dipeptidyl peptidase-4 inhibitors for type 2 diabetes: a Bayesian network meta-analysis of 58 randomized controlled trials Saxagliptin and cardiovascular outcomes in patients with type 2 diabetes mellitus Alogliptin after acute coronary syndrome in patients with type 2 diabetes Effect of Sitagliptin on Cardiovascular Outcomes in Type 2 Diabetes Effect of Linagliptin vs Placebo on Major Cardiovascular Events in Adults With Type 2 Diabetes and High Cardiovascular and Renal Risk: The CARMELINA Randomized Clinical Trial Association Between Use of Sodium-Glucose Cotransporter 2 Inhibitors, Glucagon-like Peptide 1 Agonists, and Dipeptidyl Peptidase 4 Inhibitors With All-Cause Mortality in Patients With Type 2 Diabetes: A Systematic Review and Meta-analysis Glucoselowering drugs or strategies, atherosclerotic cardiovascular events, and heart failure in people with or at risk of type 2 diabetes: an updated systematic review and meta-analysis of randomised cardiovascular outcome trials Use of dipeptidyl peptidase-4 inhibitors and the reporting of infections: a disproportionality analysis in the World Health Organization VigiBase. Diabetes Care Dipeptidyl peptidase-4 inhibitors and the risk of community-acquired pneumonia in patients with type 2 diabetes Comparative Safety of Dipeptidyl Peptidase-4 Inhibitors Versus Sulfonylureas and Other Glucose-lowering Therapies for Three Acute Outcomes Use of Dipeptidyl-Peptidase-4 Inhibitors and the Risk of Pneumonia: A Population-Based Cohort Study Postauthorization safety study of the DPP-4 inhibitor saxagliptin: a large-scale multinational family of cohort studies of five outcomes Diabetes and COVID-19 Practical recommendations for the management of diabetes in patients with COVID-19 COVID-19 and diabetes: Can DPP4 inhibition play a role? Protocol of GLUcose COntrol Safety and Efficacy in type 2 DIabetes, a NETwork meta-analysis: GLUCOSE DINET protocol-Rational and design GLUcose COntrol Safety & Efficacy in type 2 DIabetes, a systematic review and NETwork meta-analysis The Cochrane Collaboration's tool for assessing risk of bias in randomised trials Investigating and dealing with publication and other biases in meta-analysis Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes R: A Language for Data Analysis and Graphics Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement A randomized, placebocontrolled study of the cardiovascular safety of the once-weekly DPP-4 inhibitor omarigliptin in patients with type 2 diabetes mellitus Effect of Linagliptin vs Glimepiride on Major Adverse Cardiovascular Outcomes in Patients With Type 2 Diabetes: The CAROLINA Randomized Clinical Trial