Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2023 Feb 23;18(2):e0282210. doi: 10.1371/journal.pone.0282210

Is metformin use associated with low mortality in patients with type 2 diabetes mellitus hospitalized for COVID-19? a multivariable and propensity score-adjusted meta-analysis

Zhiyuan Ma 1, Mahesh Krishnamurthy 1,*
Editor: Jennifer A Hirst2
PMCID: PMC9949644  PMID: 36821577

Abstract

Background

Coronavirus disease 2019 (COVID-19) is a new pandemic that the entire world is facing since December of 2019. Increasing evidence has shown that metformin is linked to favorable outcomes in patients with COVID-19. The aim of this study was to address whether outpatient or inpatient metformin therapy for type 2 diabetes mellitus is associated with low in-hospital mortality in patients hospitalized for COVID-19.

Methods

We searched studies published in PubMed, Embase, Google Scholar and Cochrane Library up to November 1, 2022. Raw event data extracted from individual study were pooled using the Mantel-Haenszel approach. Odds ratio (OR) or hazard ratio (HR) adjusted for covariates that potentially confound the association using multivariable regression or propensity score matching was pooled by the inverse-variance method. Random effect models were applied for meta-analysis due to variance among studies.

Results

Twenty-two retrospective observational studies were selected. The pooled unadjusted OR for outpatient metformin therapy and in-hospital mortality was 0.48 (95% CI, 0.37–0.62) and the pooled OR adjusted with multivariable regression or propensity score matching was 0.71 (95% CI, 0.50–0.99). The pooled unadjusted OR for inpatient metformin therapy and in-hospital mortality was 0.18 (95% CI, 0.10–0.31), whereas the pooled adjusted HR was 1.10 (95% CI, 0.38–3.15).

Conclusions

Our results suggest that there is a significant association between the reduction of in-hospital mortality and outpatient metformin therapy for type 2 diabetes mellitus in patients hospitalized for COVID-19.

Introduction

Coronavirus disease 2019 (COVID-19), caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is a new pandemic, leading to high mortality [1–5]. One of risk factors linked to worse outcomes for COVID-19 is preexisting type 2 diabetes mellitus [6]. Patients with diabetes hospitalized with COVID-19 are two- to three-fold more likely to be admitted into intensive care units than that of non-diabetics and the mortality rate is at least doubled [7–10].

There seems two distinct but overlapping pathologic phases in response to SARS-CoV-2 infection: the initial viral response phase triggered by the virus itself and the subsequent inflammatory phase triggered by the host response [11]. Accumulating evidence has shown cytokine storm syndrome in patients with severe COVID-19. Hence, agents that alleviate hyper-inflammation in COVID-19 patients could be beneficial. Metformin has been shown to modulate the immune responses and restore immune homeostasis in immune cells via AMPK-dependent mechanisms [12]. Historically, metformin was used during the treatment of influenza outbreak, due to its host-directed anti-viral properties [13]. Metformin has been widely used as the first-line therapy for type 2 diabetes mellitus. In light of the pathogenesis of SARS-CoV-2, several possible beneficial effects of metformin therapy in COVID-19 patients with pre-existing type 2 diabetes mellitus have been speculated, such as anti-inflammatory effects [14], reduction in neutrophils [15], increasing the cellular pH to inhibit viral infection, interfering with the endocytic cycle, reversing established lung fibrosis [16], and unique protective effects on microvessels [17]. Conversely, metformin therapy has also been challenged for the potential risk for lactic acidosis, particularly in the cases of multi-organ failure and for promoting SARS-CoV-2 infection by possibly increasing in ACE2 availability in the respiratory tract [18].

Accumulating evidence from retrospective studies suggests that treating pre-existing type 2 diabetes mellitus patients with metformin in COVID-19 patients may not be harmful, but even endows a protective effect and offers a low mortality. Due to the inherent nature of retrospective studies, biases are likely to confound the association between metformin therapy and outcomes. In this study, to limit the confounding bias, we aimed to address whether outpatient (prior to admission) or inpatient metformin (during hospitalization) therapy for type 2 diabetes mellitus is associated with low in-hospital mortality in patients hospitalized for COVID-19 by multivariable regression and propensity score-adjusted meta-analysis.

Materials and methods

Strategy of literature search for meta-analysis

A systemic literature search of studies published in English was performed in PubMed, Embase, Google Scholar and Cochrane Library up to November 1, 2022 using “((SARS-CoV-2) OR (Covid-19)) AND ((Metformin) OR (Biguanides) OR (Fortamet) OR (Glumetza) OR (Glucophage) OR (Biomet)) AND ((Mortality) OR (Death))”. All abstracts were screened and the ‘related articles’ function was employed to broaden the search on appropriate abstracts including meta-analysis. Full texts were then retrieved and reviewed. The meta-analysis was in adherence to the principles outlined by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). The study was registered on the PROSPERO (CRD42021284113).

Study selection and risk of bias assessment for meta-analysis

Two reviewers independently extracted the following data from each study: first author, year of publication, study population characteristics (age and gender), study size, study design, country, definition of mortality, number of metformin users, effect size including odds ratios (ORs) or hazard ratios (HRs), confounding adjustment, and timing of metformin use. Studies included in the analysis met: 1) comparing the effect of metformin use prior to admission or during hospitalization on COVID-19 patients with type 2 diabetes; 2) the raw data on events can be extracted from the manuscript or supplementary materials or effect sizes (OR, HR) were reported; and 3) outcomes include in-hospital mortality. Studies were discarded, if: 1) duplicated studies from the same Registry or dataset; 2) studies without a control group or without in-hospital mortality reported. 3) total patients with type 2 diabetes in the study < 50; 4) studies that were not published in English. Discrepancies were resolved by discussion. The risk of bias in the included studies was assessed using the Newcastle-Ottawa Scale (NOS) [19].

Data syntheses and statistical analysis

To assess whether metformin use is associated with low in-hospital mortality by comparing and integrating the results of different studies, meta-analysis was carried out as described previously [20]. OR was used as the main summary statistics. The OR represents the odds of a death event occurring in the metformin group in comparison to the non-metformin group. The point of estimate of the OR is considered statistically significant at the P < 0.05 level if the 95% confidence interval does not include the value 1. Raw event data extracted from individual study were pooled using the Mantel-Haenszel approach and effect size (OR, HR) was pooled using the inverse-variance method. In our study, random effect models were employed due to variance among studies. Heterogeneity between studies was investigated by the standard chi-squared Q-test. Publication bias was assessed by Funnel plots and Egger’s test. A funnel plot is a type of scatter plot that resembles an inverted funnel (the 95% confidence interval), in which the treatment effects estimated from individual studies on the horizontal axis (OR), against a measure of study size on the vertical funnel (SE [log OR]). In addition, sensitivity analyses were performed by excluding one study at a time or subgroup to assess the robustness of the results. To better limit confounding bias due to baseline characteristics, several studies conducted analyses with propensity score matching approaches, in which propensity score was derived as the probability of being treated with metformin based on the participant’s observed covariates between metformin and non-metformin users. All analyses were conducted with R (version 4.1.2) and R package ‘meta’.

Results

Study selection and characteristics

After studies were carefully reviewed based on the inclusion and exclusion criteria, 22 retrospective observational studies were selected for further meta-analysis of metformin use and in-hospital mortality. The flowchart for the study selection is shown in Fig 1. Of these 22 studies, 17 studies reported on the association of outpatient metformin use and in-hospital mortality [21–37], 6 studies investigated the link between inpatient metformin use and in-hospital mortality [37–42], and 2 study probed the association of both outpatient and inpatient metformin use and in-hospital mortality [37,42]. There were 16 studies with extractable raw event data. According to adjustment for confounding bias, 16 studies performed multivariable regression or propensity scoring matching. The characteristics and quality assessment of the included studies are summarized in Table 1.

Fig 1. Flowchart illustration of the study selection.

Fig 1

Open in a new tab

Table 1. Characteristics of studies comparing mortality in metformin users vs non-users in patients with type 2 diabetes hospitalized for COVID-19.

Study Metformin (users/non-users) Country
 Endpoint Confounding adjustment Adjusted OR (95% CI)/
Note		 NOS
No. Mean age (years) Gender
(% male)
Outpatient Metformin Use
Studies with extractable raw event data
Abu-Jamous et at. 2020 [21] 23/
167		 NA NA UK In-hospital mortality Matching: age, sex, and number of admissions 0.19 (0.05–0.70) 7
Bramante & Buse et al. 2021 [22] 676/
8879		 60.4/
54.2
(Median)		 56.2/
46.6		 USA mortality (in-hospital and before- hospital) PS: age, race/ethnicity, gender, English‐speaking status, T2DM, BMI category, history of bariatric surgery, NAFLD)/NASH, CAD, heart failure, CKD, hypertension, hyper- or hypo-coagulable state, interstitial lung disease, tobacco use, and home medications 0.38 (0.16–0.91)/
MSA/
BMI > 25 kg/m2 		8
Bramante & Ingraham et al. 2021 [23] 2333/
3923		 73.0/
76.0		 51.6/
44.6		 USA In-hospital mortality Cox proportional hazards and PS: age, sex, comorbidities, alcohol abuse, HIV, asthma, inflammatory bowel disease, dementia, Charlson comorbidity index, and medications, and state Adjusted HR: 0.89 (0.78–1.01)
OR: 0.91 (0.78–1.07)/MSA		 8
Cariou et al. 2020 [24] 746/
571		 70
(both groups)		 65
(both groups)		 Franch In-hospital mortality Duplicated report from the CORONADO study as Lalau et al. 2020/ not included for analysis 7
Chen et al. 2020 [25] 43/
77		 62.0/
67.0		 NA China In-hospital mortality 6
Crouse et al. 2020 [26] 76/
144		 NA NA USA In-hospital mortality 6
Ghany et al. 2021 [27] 243/
350		 NA NA USA In-hospital mortality Cox proportional hazards: age, gender, Charlson score, diabetes, hypertension and ejection fraction Adjusted HR: 0.74 (0.53–0.98) 8
Lalau et al. 2020 [28] 1496/
953		 68.5/
74.6		 66.8/
59.6		 France In-hospital 28-day mortality IPTW: sex, age, BMI, arterial hypertension, history of disease, active cancer, treated OSA, use of any of antidiabetic drugs 0.71 (0.54–0.94)/
MSA		 8
Li et al. 2020 [29] 37/
94		 64.6/
67.7		 59.5/
55.3		 China In-hospital mortality 6
Luk et al. 2021 [35] 737/
254		 65.6/
68.9		 55.0/
51.6		 China In-hospital mortality Adjusted for age, sex, smoking, diabetes duration, HbA1c level, comorbidities (hypertension, CAD, heart failure, cerebrovascular disease, CKD, chronic liver disease, COPD), preadmission use of other glucose-lowering drugs, statins and RAAS inhibitors, and in-hospital use of other glucose-lowering drugs. Adjusted HR: 0.51 (0.27–0.97) 8
Ma et al. 2022 [30] 361/
995		 NA 60.4/
54.1		 USA In-hospital mortality and hospice Multivariable, PS and IPTW: age, height, weight, gender, race, comorbidities (CHF, CAD, asthma, COPD), clinical lab tests (BUN, creatinine, ALT, total bilirubin) 0.25 (0.06–0.74)
PS: 0.22 (0.04–0.81)
IPTW: 0.25 (0.07–0.89)/MSA		 8
Mirani et al. 2020 [31] 69/21 69/
75		 72.5/
71.4		 Italy In-hospital mortality Adjusted for age and sex Adjusted HR: 0.55 (0.27–1.11) 7
Ojeda-Fernandez et al. 2022 [36] 23327/
8639		 70.7/
75.0		 56.7/
60.9		 Italy In-hospital mortality PS: age, sex, duration of diabetes, DDCI and co-morbidities, and medications of interest PS: 0.74 (0.67–0.81) 8
Ong et al. 2021 [37] 109/
169		 NA/
63.9		 62.4/
52.7		 Philipp-ines In-hospital mortality Multivariable model: age, CKD, ACS, Tocilizumab, Systemic steroids, Convalescent plasma therapy, Hemoperfusion, In-hospital insulin use, HbA1c 0.43 (0.23–0.82)/ including inpatient metformin use 7
Pérez-Belmonte et al. 2020 [33] 825/
663		 74.8/
77.1		 65.7/
57.2		 Spain In-hospital mortality PS: age, gender, history of smoking, hypertension, dyslipidemia, CKD, cerebrovascular disease, COPD, atrial fibrillation, CAD, heart failure, obesity, dementia, Barthel Index score, and Charlson Comorbidity Index score, treatment with ACEi, ARB, anticoagulant, and statin, admission blood glucose, serum creatinine, and transaminase levels
1.16 (0.78–1.72)/
MSA		 8
Ramos-Rincón et al. 2020 [34] 420/
370		 NA NA Spain In-hospital mortality Multivariable analysis: demographics (age, sex, acquisition), comorbidities and dependence, symptoms (dyspnea), physical examination, laboratory findings and treatment 0.98 (0.64–1.79)/ patients ≥80 years 8
Tamura et al.
2021 [42]		 116/
72		 62.1/
68.6		 62.1/
63.9		 Brazil In-hospital mortality 7
Studies with only pre-calculated effect size (OR, RR or HR) data
Oh et al. 2021 [32] 2047
(total)		 NA NA South Korea In-hospital mortality Multivariable model: age, sex, place of residence, disability, income, the Charlson Comorbidity Index, other anti-diabetic medications 1.26 (0.81–1.95) 8
Inpatient Metformin Use
Studies with extractable raw event data
Jiang et al. 2021 [39] 100/
228		 64.0/
67.0		 49.0/
54.8		 China 30-day all- cause mortality PS: age, gender, weight, FBG, severity of COVID-19, Charlson comorbidity index, CHD, metformin therapy prior to hospitalization, DDI, creatinine and site Adjusted HR: 0.54 (0.13–2.26)/
MSA		 8
Li et al. 2022 [40] 37/
94		 NA NA China In-hospital mortality 6
Luo et al. 2020 [41] 104/
179		 63.0/
65.0		 51.0/
57.5		 China In-hospital mortality 6
Ong et al. 2021 [37] 40/
169		 NA/
63.9		 65.0/
52.7		 Philipp-ines In-hospital mortality 7
Tamura et al. 2021 [42] 115/
73		 63/
67		 63.5/
61.6		 Brazil In-hospital mortality Multivariable analysis: In-hospital metformin therapy, Pre-hospital metformin therapy, In-hospital insulin therapy, Dexamethasone/prednisolone 0.034 (0.002–0.58) 7
Studies with only pre-calculated effect size (OR, RR or HR) data
Cheng et al. 2020 [38] 678/
535		 62.0/
64.0
(Median)		 53.8/
49.9		 China In-hospital 28-day mortality Cox proportional hazards and PS: age, gender, comorbidities, blood glucose, C-reactive protein, estimated glomerular filtration, alanine aminotransferase, and creatinine Adjusted HR before PS: 0.87 (0.36–2.12)
Adjusted HR after PS: 1.65 (0.71–3.86)/MSA		 8
Open in a new tab

PS: Propensity score; IPTW: Inverse probability of treatment weighting; NA: Not accessed; MSA: Metformin-specific adjustment (The probability of a patient being treated with metformin as dependent variable was estimated using a logistic regression model); ACS, acute coronary syndrome; CHF, congestive heart failure; CAD, coronary artery disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DCCI, Drug-Derived Complexity Index; ALT, alanine aminotransferase; BUN, blood urea nitrogen; TBILI, total bilirubin; NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; NOS: The Newcastle-Ottawa Scale.

Outpatient metformin use and in-hospital mortality

Due to the inherent nature of retrospective observational studies, to better probe the in-hospital mortality benefits and risks for outpatient metformin use, we set out to perform the meta-analysis using two approaches. First, meta-analysis of 16 studies with extractable raw event data from 30,891 metformin and 25,770 non-metformin users revealed that outpatient metformin use was associated with a significant reduction of in-hospital death (Fig 2A) with a pooled unadjusted OR of 0.48 (95% CI, 0.37–0.62; I2 = 87%). Publication bias was assessed using a funnel plot (Fig 2B) which showed the 16 studies used in our meta-analysis. Egger’s test was performed to quantitively assess publication bias and revealed no significant publication bias (p = 0.86). Second, meta-analysis of 10 studies with adjusted OR using either multivariable regression or propensity score matching demonstrated a pooled OR of 0.71 (95% CI, 0.50–0.99; I2 = 68%), suggesting that there was a significant association between decreased in-hospital mortality and outpatient metformin use (Fig 2C). The funnel plot for the included 10 studies showed an approximate symmetric distribution (Fig 2D). Egger’s test revealed no statistically significant publication bias (p = 0.54).

Fig 2. The association of outpatient metformin use and in-hospital mortality in patients with COVID-19 and pre-existing T2DM.

Fig 2

Open in a new tab

(A) Meta-analysis of 16 selected studies with extractable raw event data using the Mantel-Haenszel approach. (B) Funnel plot test of 16 studies included in panel A. (C) Meta-analysis of 10 selected studies with multivariable or propensity score-adjusted OR using the inverse-variance method to pool effect sizes. (D) Funnel plot of 10 studies included in panel C. (E) Meta-analysis of OR for in-hospital mortality in 6 studies with propensity score matched for metformin use by the inverse-variance method. (F) Meta-analysis of 5 studies with adjusted HR for in-hospital mortality using the inverse-variance method.

In the subgroup analysis of propensity scored-matched studies, outpatient metformin use was not associated with a statistically significant but a trend toward a reduction of in-hospital mortality with a pooled OR of 0.75 (95% CI, 0.53–1.05; I2 = 66%) (Fig 2E). In contrast, subgroup analysis of 5 studies with adjusted HR revealed outpatient metformin use was associated with a statistically significant decrease of in-hospital mortality with a pooled HR of 0.78 (95% CI, 0.69–0.89; I2 = 43%) (Fig 2F). In the subgroup analysis of studies with extractable raw events, excluding studies with NOS < 7 or sample size < 100 revealed a similar significant association between outpatient metformin use and in-hospital mortality (OR 0.50; 95% CI, 0.37–0.68; I2 = 90%) (S1 Fig). In addition, to determine the impact of individual studies on pooled effects, sensitivity analysis was performed by excluding one study a time. The pooled ORs in studies with extractable raw event data showed constantly significant differences with the upper limit of 95% CI less than 1 (S1B Fig), whereas studies with adjusted OR demonstrated unstable results in the sensitivity analysis (S1C Fig). These results indicate that outpatient metformin use is associated with a significant reduction of in-hospital death in unadjusted raw data with stability of meta-analysis and that the association was still statistically significant after adjustments for confounding but with instability in the sensitivity analysis.

Inpatient metformin use and in-hospital mortality

There were a few studies mainly from China investigating the association between inpatient metformin use and in-hospital mortality in patients with type 2 diabetes hospitalized for COVID-19. We performed the similar analyses with extractable raw data and adjusted OR. Meta-analysis of 5 studies with extractable raw event data from 396 metformin and 743 non-metformin users showed that inpatient metformin use was associated with a significant reduction of in-hospital death (Fig 3A) with a pooled unadjusted OR of 0.18 (95% CI, 0.10–0.31; I2 = 0%). The funnel plot for the included 5 studies demonstrated a good symmetric distribution (Fig 3B). However, meta-analysis of 2 studies with adjusted HR demonstrated a pooled HR of 1.10 (95% CI, 0.38–3.15; I2 = 43%), suggesting there is no statistically significant association between inpatient metformin use and in-hospital mortality (Fig 3C). No sensitivity or subgroup studies were performed due to small number of studies available.

Fig 3. The association of inpatient metformin use and in-hospital mortality in patients with COVID-19 and pre-existing T2DM.

Fig 3

Open in a new tab

(A) Meta-analysis of 5 selected studies with extractable raw event data using the Mantel-Haenszel approach. (B) Funnel plot of 5 studies included in panel A. (C) Meta-analysis of 2 selected studies with multivariable and propensity score-adjusted HR using the inverse-variance method to pool effect sizes.

Discussion

Diabetes has been linked to poor outcomes in patients with COVID-19, so treatments that are effective, safe, and available immediately are urgently needed. Metformin is a the first-line therapeutic agent that is safe, effective, and inexpensive for type 2 diabetes mellitus. In the light of host-directed anti-viral properties and anti-inflammatory effects of metformin, many retrospective studies including our previous study [30] has explored whether metformin is beneficial in patients with COVID-19 and preexisting type 2 diabetes mellitus. To further compare and integrate the results of different studies and to increase statistical power, we set out to perform a meta-analysis. In this meta-analysis study, we found that outpatient metformin use was associated with a significantly lower in-hospital mortality in patients with type 2 diabetes hospitalized for COVID-19 in the raw and adjusted analyses, compared to non-metformin use. We also found that there was a significant association between decreased in-hospital mortality and inpatient metformin use in the unadjusted analysis, but this significant association was not observed after adjustments for confounding factors.

Prior to this meta-analysis, several meta-analyses [43–45] of the association between metformin therapy and severity and mortality were published, but they have issues. First, metformin therapy was not differentiated between outpatient and inpatient use, as oral hypoglycemic medications including metformin are usually held during hospitalization. Second, while study endpoints such as pre-defined mortality were somewhat different among studies, effect sizes (OR or HR) of different mortality outcomes were combined in previous meta-analysis studies. Third, due to nature of retrospective studies, various effect sizes with adjustments for baseline characteristics were reported in several studies. However, unadjusted and adjusted effect sizes were mixed in different meta-analyses. Fourth, using pre-calculated standard error instead of raw event data may not be a good estimator of the precision of the binary effect size (OR or HR). Nevertheless, only pre-calculated OR was used in the previous meta-analyses.

In this study, we specifically addressed these issues by investigating the association between outpatient or inpatient metformin therapy and in-hospital mortality in patients with diabetes hospitalized for COVID-19. We also used extractable raw data instead of pre-calculated ORs or HRs to determine the pooled OR for the unadjusted analysis. We found that there was a significant association between the reduction of in-hospital mortality and outpatient or inpatient metformin therapy for type 2 diabetes mellitus in COVID-19 patients in the unadjusted analysis, which is in line with previous meta-analyses [43–45]. Furthermore, evidence obtained from observational studies is an of great importance source for clinical practice where randomized clinical trials are unavailable or infeasible [46]. Unlike clinical trials through randomization to ensure patient characteristics are comparable across treatment and control groups, observational studies usually attempt to adjust for confounding by multivariable regression and propensity score matching. We took the advantage of these approaches and applied adjusted ORs or HRs into our meta-analysis. Similarly, we found that outpatient metformin therapy was associated with decreased in-hospital mortality in the meta-analysis of studies with adjusted OR or adjusted HR. However, when only the propensity score-matched studies [22,23,28,30,33,36] were considered for outpatient metformin use and in-hospital mortality in patient with diabetes hospitalized for COVD-19, there was no significant association with reduced in-hospital mortality. A similar result was reached for inpatient metformin use and in-hospital mortality between the unadjusted and adjusted analyses. These discordant findings could be due to confounding bias, between-study heterogeneity and a small number of studies in the subgroup analysis. Therefore, further randomized clinical trials are needed to investigate the clinical benefits of metformin therapy in patient with diabetes hospitalized for COVD-19.

Several limitations are present in our study. First, owing to lack of randomized clinical trials, all studies included for analysis were retrospective studies. Although efforts were made to balance and control for potential confounding factors by multiple variable adjustments and propensity score matching, due to the inherent nature of retrospective observational studies, residual confounders are likely to exist and could not be balanced. Therefore, even meta-analysis of studies with adjusted ORs or HRs could be misleading. Second, the size of the retrospective studies that were included in the meta-analysis varied considerably, resulting in moderate-to-high between-study heterogeneity, as revealed by high I2. Third, there were multiple sources of bias, resulting in high risk of bias in the meta-analysis. Confounding bias could be resulted from different patient characteristics, such as comorbidities, age, gender, and countries. Baseline of HbA1c, metformin dosage, duration of diabetes and treatment were often not reported in studies, contributing to heterogeneity and bias. Furthermore, different reported outcomes in studies could lead to selection and measurement bias. Fourth, there were very few studies investigating inpatient metformin use and in-hospital mortality. More studies are needed to perform a robust meta-analysis for inpatient metformin therapy and in-hospital mortality in patient with diabetes hospitalized for COVD-19.

In conclusion, our findings demonstrate that there is a significant association between the reduction of in-hospital mortality and outpatient metformin therapy for type 2 diabetes mellitus in patients hospitalized for COVID-19 and that inpatient metformin use is associated with a significant decrease of in-hospital mortality in the unadjusted analysis, but this mortality association does not retain after adjustments for confounding bias. Therefore, further randomized clinical trials are needed to provide clinical evidence regarding metformin therapy and in-hospital mortality in patient with diabetes hospitalized for COVD-19.

Supporting information

S1 Checklist. PRISMA 2020 checklist.

(DOCX)

Click here for additional data file. (31.7KB, docx)
S1 Fig. Sensitivity analysis of the studies included in meta-analysis in Fig 2A and 2C.

(TIF)

Click here for additional data file. (280.5KB, tif)

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

The author(s) received no specific funding for this work.

References

  • 1.Guan WJ, Ni ZY, Hu Y, Liang WH, Ou CQ, He JX, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. N Engl J Med. 2020;382(18):1708–20. Epub 2020/02/29. doi: 10.1056/NEJMoa2002032 ; PubMed Central PMCID: PMC7092819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Wu Z, McGoogan JM. Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention. JAMA. 2020. Epub 2020/02/25. doi: 10.1001/jama.2020.2648 . [DOI] [PubMed] [Google Scholar]
  • 3.Yang X, Yu Y, Xu J, Shu H, Xia J, Liu H, et al. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. Lancet Respir Med. 2020;8(5):475–81. Epub 2020/02/28. doi: 10.1016/S2213-2600(20)30079-5 ; PubMed Central PMCID: PMC7102538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Cummings MJ, Baldwin MR, Abrams D, Jacobson SD, Meyer BJ, Balough EM, et al. Epidemiology, clinical course, and outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort study. Lancet. 2020;395(10239):1763–70. Epub 2020/05/23. doi: 10.1016/S0140-6736(20)31189-2 ; PubMed Central PMCID: PMC7237188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Grasselli G, Zangrillo A, Zanella A, Antonelli M, Cabrini L, Castelli A, et al. Baseline Characteristics and Outcomes of 1591 Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region, Italy. JAMA. 2020. Epub 2020/04/07. doi: 10.1001/jama.2020.5394 ; PubMed Central PMCID: PMC7136855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Lighter J, Phillips M, Hochman S, Sterling S, Johnson D, Francois F, et al. Obesity in Patients Younger Than 60 Years Is a Risk Factor for COVID-19 Hospital Admission. Clin Infect Dis. 2020;71(15):896–7. Epub 2020/04/10. doi: 10.1093/cid/ciaa415 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zhou F, Yu T, Du R, Fan G, Liu Y, Liu Z, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet. 2020;395(10229):1054–62. Epub 2020/03/15. doi: 10.1016/S0140-6736(20)30566-3 ; PubMed Central PMCID: PMC7270627. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Guan WJ, Liang WH, Zhao Y, Liang HR, Chen ZS, Li YM, et al. Comorbidity and its impact on 1590 patients with COVID-19 in China: a nationwide analysis. Eur Respir J. 2020;55(5). Epub 2020/03/29. doi: . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Roncon L, Zuin M, Rigatelli G, Zuliani G. Diabetic patients with COVID-19 infection are at higher risk of ICU admission and poor short-term outcome. J Clin Virol. 2020;127:104354. Epub 2020/04/20. doi: 10.1016/j.jcv.2020.104354 ; PubMed Central PMCID: PMC7195018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Huang I, Lim MA, Pranata R. Diabetes mellitus is associated with increased mortality and severity of disease in COVID-19 pneumonia—A systematic review, meta-analysis, and meta-regression. Diabetes Metab Syndr. 2020;14(4):395–403. Epub 2020/04/26. doi: 10.1016/j.dsx.2020.04.018 ; PubMed Central PMCID: PMC7162793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Siddiqi HK, Mehra MR. COVID-19 illness in native and immunosuppressed states: A clinical-therapeutic staging proposal. J Heart Lung Transplant. 2020;39(5):405–7. Epub 2020/05/05. doi: 10.1016/j.healun.2020.03.012 ; PubMed Central PMCID: PMC7118652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Ursini F, Russo E, Pellino G, D’Angelo S, Chiaravalloti A, De Sarro G, et al. Metformin and Autoimmunity: A "New Deal" of an Old Drug. Front Immunol. 2018;9:1236. Epub 2018/06/20. doi: 10.3389/fimmu.2018.01236 ; PubMed Central PMCID: PMC5994909. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Garcia EY. Flumamine, a new synthetic analgesic and anti-flu drug. J Philipp Med Assoc. 1950;26(7):287–93. Epub 1950/07/01. . [PubMed] [Google Scholar]
  • 14.Cameron AR, Morrison VL, Levin D, Mohan M, Forteath C, Beall C, et al. Anti-Inflammatory Effects of Metformin Irrespective of Diabetes Status. Circ Res. 2016;119(5):652–65. Epub 2016/07/16. doi: 10.1161/CIRCRESAHA.116.308445 ; PubMed Central PMCID: PMC4990459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Dalan R. Metformin, neutrophils and COVID-19 infection. Diabetes Res Clin Pract. 2020;164:108230. Epub 2020/05/25. doi: 10.1016/j.diabres.2020.108230 ; PubMed Central PMCID: PMC7242188 interest. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Esam Z. A proposed mechanism for the possible therapeutic potential of Metformin in COVID-19. Diabetes Res Clin Pract. 2020:108282. Epub 2020/06/28. doi: 10.1016/j.diabres.2020.108282 ; PubMed Central PMCID: PMC7314682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Wiernsperger N, Al-Salameh A, Cariou B, Lalau JD. Protection by metformin against severe Covid-19: An in-depth mechanistic analysis. Diabetes Metab. 2022;48(4):101359. Epub 2022/06/07. doi: 10.1016/j.diabet.2022.101359 ; PubMed Central PMCID: PMC9154087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ursini F, Ciaffi J, Landini MP, Meliconi R. COVID-19 and diabetes: Is metformin a friend or foe? Diabetes Res Clin Pract. 2020;164:108167. Epub 2020/04/28. doi: 10.1016/j.diabres.2020.108167 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol. 2010;25(9):603–5. Epub 2010/07/24. doi: 10.1007/s10654-010-9491-z . [DOI] [PubMed] [Google Scholar]
  • 20.Ma Z, Dong A, Fan J, Cheng H. Does sleeve lobectomy concomitant with or without pulmonary artery reconstruction (double sleeve) have favorable results for non-small cell lung cancer compared with pneumonectomy? A meta-analysis. Eur J Cardiothorac Surg. 2007;32(1):20–8. Epub 2007/04/20. doi: 10.1016/j.ejcts.2007.03.018 . [DOI] [PubMed] [Google Scholar]
  • 21.Abu-Jamous B, Anisimovich A, Baxter J, Mackillop L, Vizcaychipi MP, McCarthy A, et al. Associations of comorbidities and medications with COVID-19 outcome: A retrospective analysis of real-world evidence data. medRxiv. 2020:2020.08.20.20174169. doi: 10.1101/2020.08.20.20174169 [DOI] [Google Scholar]
  • 22.Bramante CT, Buse J, Tamaritz L, Palacio A, Cohen K, Vojta D, et al. Outpatient metformin use is associated with reduced severity of COVID-19 disease in adults with overweight or obesity. J Med Virol. 2021. Epub 2021/02/14. doi: 10.1002/jmv.26873 ; PubMed Central PMCID: PMC8013587. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Bramante CT, Ingraham NE, Murray TA, Marmor S, Hovertsen S, Gronski J, et al. Metformin and risk of mortality in patients hospitalised with COVID-19: a retrospective cohort analysis. Lancet Healthy Longev. 2021;2(1):e34–e41. Epub 2021/02/02. doi: 10.1016/S2666-7568(20)30033-7 ; PubMed Central PMCID: PMC7832552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Cariou B, Hadjadj S, Wargny M, Pichelin M, Al-Salameh A, Allix I, et al. Phenotypic characteristics and prognosis of inpatients with COVID-19 and diabetes: the CORONADO study. Diabetologia. 2020;63(8):1500–15. Epub 2020/05/31. doi: 10.1007/s00125-020-05180-x ; PubMed Central PMCID: PMC7256180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Chen Y, Yang D, Cheng B, Chen J, Peng A, Yang C, et al. Clinical Characteristics and Outcomes of Patients With Diabetes and COVID-19 in Association With Glucose-Lowering Medication. Diabetes Care. 2020;43(7):1399–407. Epub 2020/05/16. doi: 10.2337/dc20-0660 . [DOI] [PubMed] [Google Scholar]
  • 26.Crouse AB, Grimes T, Li P, Might M, Ovalle F, Shalev A. Metformin Use Is Associated With Reduced Mortality in a Diverse Population With COVID-19 and Diabetes. Front Endocrinol (Lausanne). 2020;11:600439. Epub 2021/02/02. doi: 10.3389/fendo.2020.600439 ; PubMed Central PMCID: PMC7838490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Ghany R, Palacio A, Dawkins E, Chen G, McCarter D, Forbes E, et al. Metformin is associated with lower hospitalizations, mortality and severe coronavirus infection among elderly medicare minority patients in 8 states in USA. Diabetes Metab Syndr. 2021;15(2):513–8. Epub 2021/03/05. doi: 10.1016/j.dsx.2021.02.022 ; PubMed Central PMCID: PMC7891082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Lalau JD, Al-Salameh A, Hadjadj S, Goronflot T, Wiernsperger N, Pichelin M, et al. Metformin use is associated with a reduced risk of mortality in patients with diabetes hospitalised for COVID-19. Diabetes Metab. 2020;47(5):101216. Epub 2020/12/15. doi: 10.1016/j.diabet.2020.101216 ; PubMed Central PMCID: PMC7832745. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Li J, Wei Q, Li WX, McCowen KC, Xiong W, Liu J, et al. Metformin Use in Diabetes Prior to Hospitalization: Effects on Mortality in Covid-19. Endocr Pract. 2020;26(10):1166–72. Epub 2021/01/21. doi: 10.4158/EP-2020-0466 ; PubMed Central PMCID: PMC7834011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ma Z, Patel N, Vemparala P, Krishnamurthy M. Metformin is associated with favorable outcomes in patients with COVID-19 and type 2 diabetes mellitus. Sci Rep. 2022;12(1):5553. Epub 2022/04/03. doi: 10.1038/s41598-022-09639-2 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Mirani M, Favacchio G, Carrone F, Betella N, Biamonte E, Morenghi E, et al. Impact of Comorbidities and Glycemia at Admission and Dipeptidyl Peptidase 4 Inhibitors in Patients With Type 2 Diabetes With COVID-19: A Case Series From an Academic Hospital in Lombardy, Italy. Diabetes Care. 2020;43(12):3042–9. Epub 2020/10/08. doi: 10.2337/dc20-1340 . [DOI] [PubMed] [Google Scholar]
  • 32.Oh TK, Song IA. Metformin use and risk of COVID-19 among patients with type II diabetes mellitus: an NHIS-COVID-19 database cohort study. Acta Diabetol. 2021;58(6):771–8. Epub 2021/02/15. doi: 10.1007/s00592-020-01666-7 ; PubMed Central PMCID: PMC7882044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Perez-Belmonte LM, Torres-Pena JD, Lopez-Carmona MD, Ayala-Gutierrez MM, Fuentes-Jimenez F, Huerta LJ, et al. Mortality and other adverse outcomes in patients with type 2 diabetes mellitus admitted for COVID-19 in association with glucose-lowering drugs: a nationwide cohort study. BMC Med. 2020;18(1):359. Epub 2020/11/17. doi: 10.1186/s12916-020-01832-2 ; PubMed Central PMCID: PMC7666969. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ramos-Rincón JM, Pérez-Belmonte LM, Carrasco-Sánchez FJ, Jansen-Chaparro S, De-Sousa-Baena M, et al. Association between prior cardiometabolic therapy and in-hospital mortality in very old patients with type 2 diabetes mellitus hospitalized due to COVID-19. A nationwide observational study in Spain. Research Square. 2020. doi: 10.21203/rs.3.rs-133358/v1 [DOI] [Google Scholar]
  • 35.Luk AOY, Yip TCF, Zhang X, Kong APS, Wong VW, Ma RCW, et al. Glucose-lowering drugs and outcome from COVID-19 among patients with type 2 diabetes mellitus: a population-wide analysis in Hong Kong. BMJ Open. 2021;11(10):e052310. Epub 2021/10/22. doi: 10.1136/bmjopen-2021-052310 ; PubMed Central PMCID: PMC8529616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Ojeda-Fernandez L, Foresta A, Macaluso G, Colacioppo P, Tettamanti M, Zambon A, et al. Metformin use is associated with a decrease in the risk of hospitalization and mortality in COVID-19 patients with diabetes: A population-based study in Lombardy. Diabetes Obes Metab. 2022;24(5):891–8. Epub 2022/01/12. doi: 10.1111/dom.14648 . [DOI] [PubMed] [Google Scholar]
  • 37.Ong AN, Tan CC, Canete MT, Lim BA, Robles J. Association Between Metformin Use and Mortality among Patients with Type 2 Diabetes Mellitus Hospitalized for COVID-19 Infection. J ASEAN Fed Endocr Soc. 2021;36(2):133–41. Epub 2021/12/31. doi: 10.15605/jafes.036.02.20 ; PubMed Central PMCID: PMC8666492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Cheng X, Liu YM, Li H, Zhang X, Lei F, Qin JJ, et al. Metformin Is Associated with Higher Incidence of Acidosis, but Not Mortality, in Individuals with COVID-19 and Pre-existing Type 2 Diabetes. Cell Metab. 2020;32(4):537–47 e3. Epub 2020/08/31. doi: 10.1016/j.cmet.2020.08.013 ; PubMed Central PMCID: PMC7439986. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Jiang N, Chen Z, Liu L, Yin X, Yang H, Tan X, et al. Association of metformin with mortality or ARDS in patients with COVID-19 and type 2 diabetes: A retrospective cohort study. Diabetes Res Clin Pract. 2021;173:108619. Epub 2020/12/15. doi: 10.1016/j.diabres.2020.108619 ; PubMed Central PMCID: PMC7833191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Li J, Wei Q, McCowen KC, Xiong W, Liu J, Jiang W, et al. Inpatient use of metformin and acarbose is associated with reduced mortality of COVID-19 patients with type 2 diabetes mellitus. Endocrinol Diabetes Metab. 2022;5(1):e00301. Epub 2021/09/30. doi: 10.1002/edm2.301 ; PubMed Central PMCID: PMC8646242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Luo P, Qiu L, Liu Y, Liu XL, Zheng JL, Xue HY, et al. Metformin Treatment Was Associated with Decreased Mortality in COVID-19 Patients with Diabetes in a Retrospective Analysis. Am J Trop Med Hyg. 2020;103(1):69–72. Epub 2020/05/25. doi: 10.4269/ajtmh.20-0375 ; PubMed Central PMCID: PMC7356425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Tamura RE, Said SM, de Freitas LM, Rubio IGS. Outcome and death risk of diabetes patients with Covid-19 receiving pre-hospital and in-hospital metformin therapies. Diabetol Metab Syndr. 2021;13(1):76. Epub 2021/07/15. doi: 10.1186/s13098-021-00695-8 ; PubMed Central PMCID: PMC8275913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Li Y, Yang X, Yan P, Sun T, Zeng Z, Li S. Metformin in Patients With COVID-19: A Systematic Review and Meta-Analysis. Front Med (Lausanne). 2021;8:704666. Epub 2021/09/08. doi: 10.3389/fmed.2021.704666 ; PubMed Central PMCID: PMC8416892. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Oscanoa TJ, Amado J, Vidal X, Savarino A, Romero-Ortuno R. Metformin therapy, severity and mortality of SARS-CoV-2 infection: a meta-analysis. Clin Diabetol. 2021;10(4):317–29. [Google Scholar]
  • 45.Yang W, Sun X, Zhang J, Zhang K. The effect of metformin on mortality and severity in COVID-19 patients with diabetes mellitus. Diabetes Res Clin Pract. 2021;178:108977. Epub 2021/07/25. doi: 10.1016/j.diabres.2021.108977 ; PubMed Central PMCID: PMC8295100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Thomas LE, Li F, Pencina MJ. Overlap Weighting: A Propensity Score Method That Mimics Attributes of a Randomized Clinical Trial. JAMA. 2020;323(23):2417–8. Epub 2020/05/06. doi: 10.1001/jama.2020.7819 . [DOI] [PubMed] [Google Scholar]

Decision Letter 0

Jennifer A Hirst

11 Nov 2022

PONE-D-21-37827Does Metformin Decrease Mortality in Patients with Type 2 Diabetes Mellitus Hospitalized for COVID-19? A Multivariable and Propensity Score-adjusted Meta-analysisPLOS ONE

Dear Dr. Krishnamurthy,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please take particular consideration to overinterpretation of findings and use of language that suggests causation.

Please submit your revised manuscript by Nov 28 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Jennifer A. Hirst, DPhil

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Additional Editor Comments (if provided):

General comments – please add line numbers to help reviewers

The main finding of this paper is that although unadjusted analysis suggests a significant benefit of metformin, there is no effect of metformin on mortality after adjustment for confounders. This should be highlighted in the discussion, rather than the significant finding of the unadjusted analysis. Amend this in the abstract and main body of paper.

Methods.

The Prospero page needs updating as it indicates that the authors are still at the search stage.

Searches: These only identified a very small number of studies. Was a medical librarian consulted to assist with searches? Please provide full details of the search strategy, as the described searches would have missed articles. Consider including terms such as covid*, death, biguanide*, and metformin brand names.

Were terms for diabetes included in the search?

In the meta-analysis of adjusted data, please give details of methods used to allow odds ratio and hazard ratios to be combined. Were they assumed to be the same? If this is the case, then I would expect sensitivity analyses to be carried out to assess the impact of these approximations.

Results:

Outpatient metformin use and in-hospital mortality Page 6 – the following sentence needs some clarification: “There were 15 studies with or without attempt to adjust for confounding examining the association between outpatient metformin use and in-hospital mortality”

What is meant by with or without adjusting? Is this a repeat of the numbers in the previous paragraph?

The results section contains methods and interpretation – please ensure that methods are in the methods section and interpretation is moved to the discussion leaving results only to be reported in the results

Include details of quality assessment scores for each of the included studies and risk of bias.

Discussion

This needs to be structured, starting with the key findings.

Discuss risk of bias.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: What would be of particular interest is to be the first to draw some lessons from rather robust studies. This consists in fact of – to my knowledge – seventh meta-analysis.

Moreover, there are major drawbacks.

There are several errors:

- We are speaking about observational studies, not about interventional studies. Therefore the words “decrease” (mortality), as early as in the title, and “reduction”, appearing numerous times in the text, are not appropriate.

- The terms “outpatients” and “inpatients” are also inappropriate. The question should rather be phrased as whether or not the putative beneficial association between metformin treatment and COVID outcomes is due to metformin treatment prior to hospital admission, its continuation during the hospital stay, or both. Therefore the wording would be: “metformin discontinued”, “metformin continued”, “maintenance of metformin”, etc. (during a hospital stay).

- Title: what means “A Multivariable and Propensity Score-adjusted Meta-analysis?”

- The wording is often weak (one example, in the abstract: “Random effect models were applied for meta-analysis due to variation among studies”. In fact, the issue is not that of “variations” (between studies). Rather it is as following: the inclusion of a large number of covariates does not compensate for a lack of key variables (such as BMI, the estimated glomerular filtration rate, duration of diabetes, etc.). Given that, the studies with a low sample size, with a low number of covariates, and with a lack of the major covariates should not even be considered. It is already impossible to compare studies with either 172 or more than 2 million people with diabetes!

- There are just a few words about the putative protective effects for metformin; this is clearly not enough (what about effects on microvascular blood flow and pericytes, microvascular permeability, hemostasis, glycocalyx, hydrogen sulfide, ER stress, etc.?).

- The first lines of the Discussion do not deal with the main results.

- To put it simply, the compilation of short and/or weak studies cannot lead to a firm conclusion. And, as it happens, the by far largest study (more than 2 million people: Ojeda-Fernandez, Diabetes Obesity Metab 2022) is lacking.

- Given all these difficulties, the limitations subsection is far too weak.

Reviewer #2: Thank you for pursuing this question of metformin's role.

- You state that you assessed bias and refer to internal to analysis confounding issues as well as publication bias, but you do not characterize the latter. What were your systematic bias assessment results across the studies?

- Heterogeneity is quite high across pooled analyses. This should be discussed. For instance, (1) how was A1C and complicated diabetes accounted across studies; and, (2) in which studies was metformin already part of the patients' regimen versus added while in house or once diagnosis with SARS-CoV-2 was accomplished.

- The discussion and conclusion language should acknowledge more directly (depending upon your response to the previous questions) that metformin may indicate lower risk because being managed for diabetes... currently, your presented data and analyses does not support a direct effect and yet you close the article expressing only a caveat of pooled analyses statistical significance.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Jean-Daniel Lalau

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2023 Feb 23;18(2):e0282210. doi: 10.1371/journal.pone.0282210.r002

Author response to Decision Letter 0

27 Nov 2022

1. “Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.” “please add line numbers to help reviewers”

Response: We have revised our manuscript according to PLOS ONE’s style requirements and added the line numbers.

2. “The main finding of this paper is that although unadjusted analysis suggests a significant benefit of metformin, there is no effect of metformin on mortality after adjustment for confounders. This should be highlighted in the discussion, rather than the significant finding of the unadjusted analysis. Amend this in the abstract and main body of paper.”

Response: We updated search and have performed new analysis. The results have been included in the section of Results and discussion.

3. “The Prospero page needs updating as it indicates that the authors are still at the search stage.”

Response: We have updated the status in the PROSPERO.

4. “Searches: These only identified a very small number of studies. Was a medical librarian consulted to assist with searches? Please provide full details of the search strategy, as the described searches would have missed articles. Consider including terms such as covid*, death, biguanide*, and metformin brand names.

Were terms for diabetes included in the search?”

Response: We did not consult librarian for assisting with the search. We did not include the keyword diabetes in the search. We manually screened diabetic patients in the studies after identification. Since our search was more than 1 year ago, we performed updated search using the expression “((SARS-CoV-2) OR (Covid-19)) AND ((Metformin) OR (Biguanides) OR (Fortamet) OR (Glumetza) OR (Glucophage) OR (Biomet)) AND ((Mortality) OR (Death))” as suggested. Three new studies have been included for analyses.

5. “In the meta-analysis of adjusted data, please give details of methods used to allow odds ratio and hazard ratios to be combined. Were they assumed to be the same? If this is the case, then I would expect sensitivity analyses to be carried out to assess the impact of these approximations.”

Response: Raw event data extracted from individual study were pooled using the Mantel-Haenszel approach and effect size (OR, HR) was pooled using the inverse-variance method. This has been described in the methods. We also included sensitivity analyses in Figure 2 and 3.

6. “Outpatient metformin use and in-hospital mortality Page 6 – the following sentence needs some clarification: “There were 15 studies with or without attempt to adjust for confounding examining the association between outpatient metformin use and in-hospital mortality”

What is meant by with or without adjusting? Is this a repeat of the numbers in the previous paragraph?”

Response: We have revised the text.

7. “The results section contains methods and interpretation – please ensure that methods are in the methods section and interpretation is moved to the discussion leaving results only to be reported in the results”

Response: We have revised the results.

8. “Include details of quality assessment scores for each of the included studies and risk of bias.”

Response: the Newcastle-Ottawa Scale scores have been included in Table 1.

9. “This needs to be structured, starting with the key findings.

Discuss risk of bias.”

Response: We have revised the section of discussion and discussed more on bias.

Reviewer #1:

1. “What would be of particular interest is to be the first to draw some lessons from rather robust studies. This consists in fact of – to my knowledge – seventh meta-analysis.”

Response: Thank you for your comments. Of note, we performed this study and submitted it more than a year ago. In this study, we aimed to address whether outpatient (prior to admission) or inpatient metformin (during hospitalization) therapy offers low in-hospital mortality in patients with type 2 diabetes mellitus hospitalized for COVID-19 especially by multivariable and propensity score-adjusted effect sizes (OR and HR). We discussed the issues in the previously published meta-analysis in the second paragraph of the section of the discussion. We think our study is interesting.

2. “We are speaking about observational studies, not about interventional studies. Therefore the words “decrease” (mortality), as early as in the title, and “reduction”, appearing numerous times in the text, are not appropriate.”

Response: Thank you for your comments. We agree that we did a meta-analysis of observational studies. We have revised our title. But we believe metformin therapy is interventional and treatment, so “decrease” and “reduction is related to metformin therapy not to study types.

3. “The terms “outpatients” and “inpatients” are also inappropriate. The question should rather be phrased as whether or not the putative beneficial association between metformin treatment and COVID outcomes is due to metformin treatment prior to hospital admission, its continuation during the hospital stay, or both. Therefore the wording would be: “metformin discontinued”, “metformin continued”, “maintenance of metformin”, etc. (during a hospital stay).”

Response: Thank you for your suggestion. We have specifically clarified “outpatient” and “inpatient” metformin use in the introduction.

4. “what means “A Multivariable and Propensity Score-adjusted Meta-analysis?””

Response: Meta-analysis of studies with adjusted effect size derived from multivariable regression or propensity score matching.

5. “The wording is often weak (one example, in the abstract: “Random effect models were applied for meta-analysis due to variation among studies”. In fact, the issue is not that of “variations” (between studies). Rather it is as following: the inclusion of a large number of covariates does not compensate for a lack of key variables (such as BMI, the estimated glomerular filtration rate, duration of diabetes, etc.). Given that, the studies with a low sample size, with a low number of covariates, and with a lack of the major covariates should not even be considered. It is already impossible to compare studies with either 172 or more than 2 million people with diabetes!”

Response: Thank you for your comments. We have revised the text. The random-effects model is conventionally used to address between-study heterogeneity. We did unadjusted and adjusted analyses to address other bias. we excluded studies with the number of subjects less than 50.

6. “There are just a few words about the putative protective effects for metformin; this is clearly not enough (what about effects on microvascular blood flow and pericytes, microvascular permeability, hemostasis, glycocalyx, hydrogen sulfide, ER stress, etc.?).”

Response: We have revised the text and added the protected effects on microvessels.

7. “The first lines of the Discussion do not deal with the main results.”

Response: We have revised the section of discussion.

8. “To put it simply, the compilation of short and/or weak studies cannot lead to a firm conclusion. And, as it happens, the by far largest study (more than 2 million people: Ojeda-Fernandez, Diabetes Obesity Metab 2022) is lacking.”

Response: We completed our search before Oct 1st, 2021 and submitted our paper on Nov. 29th, 2021. Now We completed new search and have included the above study along with 2 more studies.

9. “Given all these difficulties, the limitations subsection is far too weak.”

Response: We have revised the limitation subsection.

Reviewer #2:

1. “You state that you assessed bias and refer to internal to analysis confounding issues as well as publication bias, but you do not characterize the latter. What were your systematic bias assessment results across the studies?”

Response: We have performed the analyses of publication bias using funnel plots and Egger’s tests. We also included the Newcastle-Ottawa Scale scores in Table 1.

2. “Heterogeneity is quite high across pooled analyses. This should be discussed. For instance, (1) how was A1C and complicated diabetes accounted across studies; and, (2) in which studies was metformin already part of the patients' regimen versus added while in house or once diagnosis with SARS-CoV-2 was accomplished.”

Response: We have revised the limitation subsection.

3. “The discussion and conclusion language should acknowledge more directly (depending upon your response to the previous questions) that metformin may indicate lower risk because being managed for diabetes... currently, your presented data and analyses does not support a direct effect and yet you close the article expressing only a caveat of pooled analyses statistical significance.”

Response: We have performed new analyses with more studies. We have revised the results and conclusions.

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file. (47KB, docx)

Decision Letter 1

Jennifer A Hirst

30 Jan 2023

PONE-D-21-37827R1Is Metformin Use Associated with Low Mortality in Patients with Type 2 Diabetes Mellitus Hospitalized for COVID-19? A Multivariable and Propensity Score-adjusted Meta-analysisPLOS ONE

Dear Dr. Krishnamurthy,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Mar 16 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Jennifer A. Hirst, DPhil

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I must admit that I am not very interested in meta-analyses but I would also highlight the great work done by the authors for improving their manuscript.

I agree with all modifications but one: as long as studies are observational, it is never possible to speak about “reduction” (of mortality – because of metformin) (lines 21, 158, 211, 266, 269…), and even less about a “trend toward a reduction” (L. 158: non-scientific).

Reviewer #2: Thank you for your work addressing editorial and reviewer comments. The manuscript is improved.

- Consider doing a quick sensitivity analysis where you remove all studies that have a 6 or lower Newcastle-Ottowa score (indicating high risk). Those studies are substantially smaller than the studies that you scored more highly and so a single study removal method for sensitivity analysis would likely be insufficient to discern the effect by class.

- The bigger issue for me is that the manuscript is written in a way that suggests that metformin was applied for the purpose of COVID-19 therapy. That seems unlikely, particularly in the one category where you observed a multi-variate statistically significant effect. This matters in how readers (even technically grounded ways) understand the role of metformin here. You also seem to avoid talking about DM management in your background as a reason for exploring the effect. This needs to be addressed through-out. The Limitations section additions relevant to my earlier comments are appropriate. What I am talking about now is how the question is framed and the discussion and conclusions are proffered. I suspect that you simply have taken for granted that readers will understand this difference, but you do not state it, and your language suggests direct therapeutic application for the purpose of COVID-19 in a milieu where it has been among several problematic narratives avoiding more rounded approaches to patient management. This can be stated succinctly in the intro, discussion, and conclusion sections and easily corrected.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Jean-Daniel Lalau

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2023 Feb 23;18(2):e0282210. doi: 10.1371/journal.pone.0282210.r004

Author response to Decision Letter 1

4 Feb 2023

Reviewer #1:

1. “I must admit that I am not very interested in meta-analyses but I would also highlight the great work done by the authors for improving their manuscript.

I agree with all modifications but one: as long as studies are observational, it is never possible to speak about “reduction” (of mortality – because of metformin) (lines 21, 158, 211, 266, 269…), and even less about a “trend toward a reduction” (L. 158: non-scientific)”

Response: Thank you for your comment. As you pointed out that this is a meta-analysis of observational studies, we have revised the text and made it clear throughout the study that there is an association between metformin use and hospital mortality, but not direct causation. The word “reduction” is used as a noun or adjective not as a verb which may imply direct causation.

Reviewer #2:

1. “Consider doing a quick sensitivity analysis where you remove all studies that have a 6 or lower Newcastle-Ottowa score (indicating high risk). Those studies are substantially smaller than the studies that you scored more highly and so a single study removal method for sensitivity analysis would likely be insufficient to discern the effect by class”

Response: We have added additional analysis regarding this (Supp fig. 1A).

2. “The bigger issue for me is that the manuscript is written in a way that suggests that metformin was applied for the purpose of COVID-19 therapy. That seems unlikely, particularly in the one category where you observed a multi-variate statistically significant effect. This matters in how readers (even technically grounded ways) understand the role of metformin here. You also seem to avoid talking about DM management in your background as a reason for exploring the effect. This needs to be addressed through-out. The Limitations section additions relevant to my earlier comments are appropriate. What I am talking about now is how the question is framed and the discussion and conclusions are proffered. I suspect that you simply have taken for granted that readers will understand this difference, but you do not state it, and your language suggests direct therapeutic application for the purpose of COVID-19 in a milieu where it has been among several problematic narratives avoiding more rounded approaches to patient management. This can be stated succinctly in the intro, discussion, and conclusion sections and easily corrected.”

Response: Thank you for your comment. We have revised the text.

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file. (41.4KB, docx)

Decision Letter 2

Jennifer A Hirst

10 Feb 2023

Is Metformin Use Associated with Low Mortality in Patients with Type 2 Diabetes Mellitus Hospitalized for COVID-19? A Multivariable and Propensity Score-adjusted Meta-analysis

PONE-D-21-37827R2

Dear Dr. Krishnamurthy,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Jennifer A. Hirst, DPhil

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Acceptance letter

Jennifer A Hirst

14 Feb 2023

PONE-D-21-37827R2

Is Metformin Use Associated with Low Mortality in Patients with Type 2 Diabetes Mellitus Hospitalized for COVID-19? A Multivariable and Propensity Score-adjusted Meta-analysis

Dear Dr. Krishnamurthy:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Jennifer A. Hirst

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Checklist. PRISMA 2020 checklist.

    (DOCX)

    Click here for additional data file. (31.7KB, docx)
    S1 Fig. Sensitivity analysis of the studies included in meta-analysis in Fig 2A and 2C.

    (TIF)

    Click here for additional data file. (280.5KB, tif)
    Attachment

    Submitted filename: Response to reviewers.docx

    Click here for additional data file. (47KB, docx)
    Attachment

    Submitted filename: Response to reviewers.docx

    Click here for additional data file. (41.4KB, docx)

    Data Availability Statement

    All relevant data are within the paper and its Supporting Information files.

    Articles from PLOS ONE are provided here courtesy of PLOS

    RESOURCES