Association of Metformin with Mortality in COVID-19 Patients: A Systematic Review and Meta-Analysis

Introduction

The novel coronavirus disease (COVID-19), now a worldwide public health concern, is associated with varied fatality.¹ Some of the recognized risk factors for poor prognosis of COVID-19 patients include age (>65 years), chronic obstructive pulmonary disease (COPD), hypertension, cardiovascular disease, and type 2 diabetes mellitus.² Although type 2 diabetes mellitus has been reported as a risk factor for COVID-19, the effect of pharmacologic agents used to treat type 2 diabetes mellitus on COVID-19 outcomes remains unclear.

Metformin, the first-line treatment for type 2 diabetes mellitus with adequate safety profile, is being suggested for treatment of COVID-19 as a host-directed therapy.³ A retrospective study highlighted metformin use to be associated with a higher risk of severe COVID-19.³ On the contrary, some retrospective cohort studies have revealed a significant reduction in mortality with metformin use in type 2 diabetes mellitus patients diagnosed with COVID-19.^1,4,5,6 The mechanism by which metformin seems to reduce the mortality risk of COVID-19 patients remains unclear. However, its anti-inflammatory actions are hypothesized to be one of the mechanisms.^7,8,9

Several studies have demonstrated higher prevalence of mortality in type 2 diabetes mellitus patients with COVID-19^1,4,5,6; however, the effects of antidiabetic pharmacotherapy on COVID-19 complications need further exploration. Studying the probable effects of pharmacologic agents on clinical prognosis may provide an important understanding to design further studies related to mitigating the risk factors.² Thus, in this systematic review and meta-analysis, we aimed to explore the association of metformin use and mortality in COVID-19 patients. We hypothesized that mortality would be lower in COVID-19 patients receiving metformin.

Methods

We followed the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines for systematic reviews¹⁰ and Meta-analyses Of Observational Studies in Epidemiology (MOOSE)¹¹ guidelines in designing, conducting, and reporting of this systematic review. The systematic review protocol was registered at PROSPERO (registration ID: CRD42020204384).

Results

Search Results

The systematic search yielded a total of 135 publications using the keyword “Metformin and COVID-19.” After removing duplicates, 86 articles were found to be potential publications for screening. A total of 76 articles (42 review articles and 14 not reporting the required data) were excluded after the application of predefined inclusion and exclusion criteria. Ten studies^{4,15,16,17,18,19,20,21,22,23} were included for systematic review, while 9^{4,15,16,17,18,20,21,22,23} were included in the meta-analysis (►Fig. 1). All included studies were retrospective cohort studies.

Close

Fig. 1

Prisma flowchart. n, number.

Export to PPT

Association between Metformin Use with Mortality and Hospitalization Days

The analysis of the nine retrospective cohort studies suggested that metformin is associated with lower mortality rate (RR: 0.63; 95% CI: 0.50–0.78; p < 0.001; I ²: 73.9%) with substantial heterogeneity in individual study estimates. The results highlight that mortality was 37% lower in patients receiving metformin (►Fig. 2), while Cheng et al¹⁹ reported no significant difference in the mortalities of individuals between the metformin and the no-metformin groups (adjusted hazard ratio [HR]: 0.87; 95%CI: 0.36–2.12; p = 0.757). Univariate analysis by Pérez-Belmonte et al¹⁸ showed significant association of metformin with mortality (RR: 0.73; CI: 0.58–0.90; p = 0.004). However, the multivariate analysis revealed no significant association (RR: 1.10; CI: 0.76–1.60; p = 0.616). Multiple logistic regression analysis by Crouse et al¹⁶ revealed that metformin use is an independent factor affecting mortality in COVID-19 patients. Lally et al²³ showed that the patients taking metformin were at significantly reduced hazard of death (adjusted HR: 0.48; 95%CI: 0.28–0.84). Multivariate analysis by Luo et al¹⁵ showed that the use of metformin was negatively correlated with the mortality of type 2 diabetes mellitus patients. Four studies qualified for quantitative analysis of hospitalization days. The analysis revealed no significant difference in hospitalization days between the two groups (MD: 1.07; 95%CI: –0.55 to 2.69; p = 0.197; I ²: 0.0; ►Table 2).

Close

Fig. 2

Association of metformin use and mortality in covid-19 patients.

Export to PPT

Table 2 Random-effects meta-analysis results between metformin and control groups for hospitalization and laboratory parameters

Outcomes	No. of studies	Sample size	MD	95% CI	p-value	I-squared (%)	Heterogeneity statistic (p-value)
Hospitalization (d)	4	1,646	1.07	−0.55 to 2.69	0.197	0.0	0.18 (0.758)
Hematologic parameters (109/L)
White blood cells	3	731	0.01	−0.29 to 0.31	0.953	0.0	1.79 (0.409)
Neutrophils	2	403	−0.11	−0.48 to 0.26	0.561	0.0	0.61 (0.436)
Lymphocytes	3	731	0.12	−0.02 to 0.28	0.085	39.2	3.29 (0.193)
Monocytes	2	403	−0.02	−0.08 to 0.05	0.650	51.7	2.07 (0.150)
Platelets	2	403	12.57	−12.46 to 37.57	0.324	45.8	1.84 (0.174)
Eosinophils	1	283	0.02	0.002 to 0.04	0.034	–	–
Biochemical parameters
ALT (U/L)	3	731	0.02	−2.27 to 2.31	0.986	0.0	0.41 (0.815)
AST (U/L)	3	731	−1.66	−6.11 to 2.78	0.464	78.3	9.20 (0.010)
Urea (mmol/L)	2	403	−0.65	−1.49 to 0.19	0.129	53.8	2.16 (0.141)
LDH (U/L)	1	328	−12.30	−28.79 to 4.19	0.144	–	–
Ferritin (μg/L)	1	328	78.47	−53.00 to 209.94	0.242	–	–
Albumin (g/L)	1	328	2.23	1.034 to 3.433	< 0.001	–	–
Triglyceride (mmol/L	1	131	0.12	−0.18 to 0.42	0.432	–	–
FBG (mmol/L)	2	611	1.00	−0.59 to 2.58	0.218	71.7	3.53 (0.060)
HbA1c (%)	1	328	0.70	0.27 to 1.13	0.001	–	–
HbA1c (mmol/L)	1	131	1.77	0.98 to 2.56	< 0.001	–	–
eGFR (mL/min)	1	1,213	3.43	0.55 to 6.32	0.020	–	–
γ-Glutamyltransferase	1	283	−0.23	−5.33 to 4.86	0.928	–	–

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HbA1c, hemoglobin A1c; LDH, lactic dehydrogenase; MD, weighted mean difference.

Discussion

Emerging evidence suggest that the presence of comorbidities increases the mortality risk in COVID-19 patients. Diabetes forms an important risk factor to predict the adverse outcomes.¹ Several studies have demonstrated high mortality rate in type 2 diabetes mellitus patients^1,5,6,24; however, the effect of antidiabetics on the mortality remains to be elucidated.²⁵ Thus, the present meta-analysis was conducted to assess the association of metformin use with mortality in COVID-19 patients. The present meta-analysis was based on data from nine studies on COVID-19 patients. The analysis demonstrated lower mortality in the group receiving metformin; however, no significant difference in hospitalization days was observed between the two groups. The analysis also revealed significant difference in the prevalence of obesity, hypertension, heart failure, cerebrovascular disease and COPD. No significant difference was observed in clinical severity between the two groups.

Several retrospective cohort studies have demonstrated low mortality in patients receiving metformin.^15,17,24,25 In line with our findings, other meta-analysis have revealed a significantly lower mortality in patients receiving metformin.^7,26 Retrospective cohort studies have shown metformin use to be an independent factor associated with mortality.^16,20 However, retrospective cohort studies have revealed no difference in hospitalization days.^20,21 Additionally, no difference in the clinical severity between the groups receiving and not receiving metformin was revealed in retrospective cohort studies.^15,21 Furthermore, retrospective cohort studies have reported a lower number of patients with obesity,¹⁷ hypertension,^17,22 heart failure,^19,22 cerebrovascular disease,¹⁹ and COPD²² in patients receiving metformin.

The potential mechanisms underlying the reduction of mortality by metformin in COVID-19 patients remain unclear.⁷ However, several mechanisms have been proposed for the same. Metformin is thought to increase the expression of angiotensin converting enzyme 2 (ACE2),^2,27,28,29 a receptor for severe acute respiratory syndrome coronavirus 2.² Metformin has been demonstrated to activate adenosine monophosphate-activated protein kinase (AMPK) in in vitro and in vivo experiments,^2,7 thus regulating ACE2 and protein stability.² Activation of AMPK further leads to the inhibition of nuclear factor kappa B (NF-kB), which is a transcription regulator involved in inflammation. Inhibition of NF-kB subsequently inhibits the production of proinflammatory cytokines, such as interleukins 8 and 1α.⁷ Metformin is also suggested to have antithrombotic effects. Evidence suggests excessive inflammatory responses as well as disseminated thromboembolic events are associated with increased mortality in COVID-19 patients.³⁰ Therefore, metformin is suggested to have beneficial effect by its antifibrinolytic activities³⁰ and inhibiting the cytokine storm in COVID-19 patients.⁷ Additionally, metformin induces autophagy, thus controlling inflammation and activating innate and adaptive immunity.³¹ Furthermore, metformin is suggested to act on Na⁺/H⁺ exchangers and vacuolar ATPase, subsequently inhibiting viral infection by raising the cellular pH leading to interference with the endocytic cycle.

This study had several limitations. A small number of studies have been included in the meta-analysis; retrospective cohort studies were included, and substantial heterogeneity was observed among the studies, which poses a significant limitation for the conclusions drawn. Although an extensive search was done, studies might have been missed inadvertently. Exclusion of studies in languages other than English might have resulted in missing of relevant studies.

Conclusion

To conclude, our findings suggest that metformin may contribute to reduced mortality in COVID-19 patients. However, randomized studies demonstrating the effects of metformin in COVID-19 are needed to support clinical recommendations of metformin in high-risk population. Further investigations are warranted to look into the likelihood of metformin improving prognosis of COVID-19.