From before the coronavirus disease 2019 (COVID-19) pandemic, the effect of dipeptidyl peptidase 4 inhibitors (DPP-4i) on respiratory infections (RI) remains unclear. The meta-analysis of Yang et al. (1) specifically assessed the risk of infection with DPP-4i. However, cardiovascular outcomes trials (CVOTs) did not appear in their meta-analysis for RI (1). In addition, the meta-analyses of CVOTs did not report assessment of RI for DPP-4i. Meta-analyses and pharmacovigilance studies showed some discrepancies (2). Recently published practical recommendations did not suggest discontinuation of DPP-4i because of the COVID-19 pandemic (3,4). The dipeptidyl peptidase 4 enzyme could be a target for treating COVID-19 (2). We aimed to provide a powerful and less biased estimate of the effect of DPP-4i on the overall risk of RI and thus verify the validity of the recent recommendations for the management of diabetes during the COVID-19 pandemic (3,4).
We updated a previous systematic review and meta-analysis, focusing our primary analysis on RI in placebo-controlled CVOTs assessing a DPP-4i in patients with type 2 diabetes, using PubMed and Cochrane Central Register of Controlled Trials (CENTRAL), up to 27 January 2020. In line with the previous meta-analysis of Yang et al. (1), our primary outcome was “any respiratory infection” (defined as influenza, nasopharyngitis, sinusitis, pharyngotonsillitis, pharyngitis, bronchitis, respiratory tract infection [RTI], upper RTI, lower RTI, or pneumonia). Secondary outcomes were any upper RI, any lower RI, and each of their components. Our primary analysis included placebo-controlled CVOTs only to avoid the risk of “small study effects” and the potential heterogeneity of different comparators. We added a sensitivity analysis integrating 1) the combined non-CVOTs from the previous meta-analysis of Yang et al. (1) and 2) an active-control CVOT identified during the systematic review. We estimated treatment effect using risk ratio and 95% CI. A P value <0.05 was considered significant. Analyses were performed with R. We used a fixed-effects model when I2 was <50% and random-effects model otherwise. The full report is available on medRxiv (5).
Five trials with 47,714 patients were included. At baseline, mean ± SD age ranged from 61 to 65.8 ± 9.1 years, HbA1c from 7.2 ± 0.5 to 8.0 ± 0.9%, and BMI from 28.7 to 31.3 ± 5.3 kg ⋅ m−2; 62.9–70.7% were male; 78.4–95.4% had arterial hypertension; and 10.2–14.4% were current smokers. The risk of bias regarding the reporting of RI was unclear for all the included studies.
The primary analysis included 47,714 patients; 4,369 participants were affected by an event of the primary outcome (rate: 9.2%). DPP-4i were not associated with either an increased or a decreased risk of overall RI compared with placebo (risk ratio 0.99 [95% CI 0.93; 1.04], I2 = 0%) (Fig. 1). The sensitivity analysis integrating non-CVOT studies and the active-controlled CVOT reached 82,644 participants and 11,349 events (rate: 13.7%). The risk of overall RI with DPP-4i was neither increased nor decreased (1.00 [0.97; 1.03]) (Fig. 1). Regarding secondary outcomes, DPP-4i were not associated with a different risk of 1) any upper RI (1.00 [0.93; 1.08]), 2) any lower RI (0.96 [0.87; 1.06]), or 3) any components themselves (influenza 1.08 [0.91; 1.28], nasopharyngitis 0.96 [0.86; 1.08], sinusitis 0.84 [0.17; 4.09], pharyngotonsillitis not estimable, pharyngitis 2.98 [0.12; 73.01], bronchitis 0.93 [0.66; 1.30], RTI 0.62 [0.24; 1.59], upper RTI 1.01 [0.89; 1.13], lower RTI 0.90 [0.41; 2.01], and pneumonia 1.05 [0.90; 1.23]).
![Figure 1. Forest plot of the risk of any RI (ARI); up, primary analysis; down, sensitivity analysis (SA), integrating 1) the placebo-controlled CVOTs, CArdiovascular safety and Renal Microvascular outcomE with LINAgliptin in patients with Type 2 Diabetes mellitus at high vascular risk (CARMELINA) (clinical trial reg. no. NCT01897532, ClinicalTrials.gov [linagliptin]), Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trial (NCT00968708 [alogliptin]), Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 (SAVOR-TIMI 53) (NCT01107886 [saxagliptin]), Trial Evaluating Cardiovascular Outcomes With Sitagliptin (TECOS) (NCT00790205 [sitagliptin]), and A Study to Assess Cardiovascular Outcomes Following Treatment With Omarigliptin (MK-3102) in Participants With Type 2 Diabetes Mellitus (MK-3102-018) (NCT01703208 [omarigliptin]); 2) the active-controlled CVOT, CARdiovascular Outcome study of LINAgliptin versus glimepiride in patients with type 2 diabetes (CAROLINA) (NCT01243424 [linagliptin vs. glimepiride]); and 3) the non-CVOTs combined from the meta-analysis of Yang et al. (1).](/view-large/figure/4024776/dc202018f1.jpeg)
Forest plot of the risk of any RI (ARI); up, primary analysis; down, sensitivity analysis (SA), integrating 1) the placebo-controlled CVOTs, CArdiovascular safety and Renal Microvascular outcomE with LINAgliptin in patients with Type 2 Diabetes mellitus at high vascular risk (CARMELINA) (clinical trial reg. no. NCT01897532, ClinicalTrials.gov [linagliptin]), Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trial (NCT00968708 [alogliptin]), Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 (SAVOR-TIMI 53) (NCT01107886 [saxagliptin]), Trial Evaluating Cardiovascular Outcomes With Sitagliptin (TECOS) (NCT00790205 [sitagliptin]), and A Study to Assess Cardiovascular Outcomes Following Treatment With Omarigliptin (MK-3102) in Participants With Type 2 Diabetes Mellitus (MK-3102-018) (NCT01703208 [omarigliptin]); 2) the active-controlled CVOT, CARdiovascular Outcome study of LINAgliptin versus glimepiride in patients with type 2 diabetes (CAROLINA) (NCT01243424 [linagliptin vs. glimepiride]); and 3) the non-CVOTs combined from the meta-analysis of Yang et al. (1).
Forest plot of the risk of any RI (ARI); up, primary analysis; down, sensitivity analysis (SA), integrating 1) the placebo-controlled CVOTs, CArdiovascular safety and Renal Microvascular outcomE with LINAgliptin in patients with Type 2 Diabetes mellitus at high vascular risk (CARMELINA) (clinical trial reg. no. NCT01897532, ClinicalTrials.gov [linagliptin]), Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trial (NCT00968708 [alogliptin]), Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction 53 (SAVOR-TIMI 53) (NCT01107886 [saxagliptin]), Trial Evaluating Cardiovascular Outcomes With Sitagliptin (TECOS) (NCT00790205 [sitagliptin]), and A Study to Assess Cardiovascular Outcomes Following Treatment With Omarigliptin (MK-3102) in Participants With Type 2 Diabetes Mellitus (MK-3102-018) (NCT01703208 [omarigliptin]); 2) the active-controlled CVOT, CARdiovascular Outcome study of LINAgliptin versus glimepiride in patients with type 2 diabetes (CAROLINA) (NCT01243424 [linagliptin vs. glimepiride]); and 3) the non-CVOTs combined from the meta-analysis of Yang et al. (1).
Our updated meta-analysis provides the most powerful and least biased estimation of the association of DPP-4i and the risk of overall (non-COVID-19) RI. We did not find any effect of DPP-4i on the risk of RI. Our results support the practical recommendations, suggesting that DPP-4i should not be discontinued with regard to the COVID-19 pandemic (3,4). Placebo-controlled CVOTs reported more events than the previous meta-analysis of non-CVOT studies (1). The sensitivity analysis integrating active-controlled CVOTs and non-CVOT studies shows the robustness of the primary analysis and provided a very precise estimation with an overall sample size of 82,644 patients. However, the risk of reporting bias was unclear. Since we pooled all DPP-4i and all RI, we cannot exclude that some DPP-4i could have different effects and that DPP-4i could have a different effect on specific RI. The extrapolation of those results to COVID-19 respiratory disease may be limited. The role of the DPP-4i for diabetes management in general is beyond the scope of this article.
We did not find any relation between DPP-4i and non-COVID RI. Ongoing trials will help with understanding the effect of DPP-4i in COVID-19.
Article Information
Duality of Interest. J.-L.C. reports investigator-driven grants from Bioprojet Pharma, Pfizer, and United Therapeutics outside the submitted work. J.-C.L. has received speaking fees and honoraria from Roche. M.C. has received consulting fees from Boehringer Ingelheim, Sanofi, and AstraZeneca and speaker honoraria from Sanofi. In the last 5 years, F.G. received for his institution fees from Portola Pharmaceuticals for central reading of ultrasound records, Neurochlore for Data and Safety Monitoring Board coordination, EryTech Pharma for modeling projects from RCTs, and Stève Consultants for exploring the French social security database. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. G.G. and S.Me. performed the literature search. G.G. and S.Ma. performed data extraction. G.G. conducted the analyses under the supervision of M.C. G.G. wrote the first draft of the manuscript. S.Ma., M.A., C.C., J.-L.C., F.G., J.-C.L., and M.C. contributed to data interpretation of the findings and provided critical input for important intellectual content. G.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
