The effect of selective serotonin and norepinephrine reuptake inhibitors on clinical outcome of COVID‐19 patients: A systematic review and meta‐analysis

Abstract Background and Aim Due to the high social and economic burden and also mortality and morbidity caused by coronavirus disease 2019 (COVID‐19) in the past few years, researchers have aimed at finding solutions to suppressing the severity of infection. Recently, selective serotonin and serotonin‐norepinephrine reuptake inhibitors (SSRI/SNRI) have been investigated as an adjuvant treatment for COVID‐19. The aim of the current study was to investigate the impact of SSRI/SNRIs on outcomes of COVID‐19 patients. Methods In this systematic review and meta‐analysis, a comprehensive search strategy consisting of relevant words was performed by two researchers in PubMed, Scopus and EMBASE libraries. Studies reporting the effect of SSRI and/or SNRI use in COVID‐19 patients' outcome were included. Hospitalization, mortality, hospitalization event, and length of hospital stay were considered as main outcomes of this study. Analysis was carried out using Comprehensive Meta‐Analysis (CMA‐version 2) and final data were reported as odds ratio (OR) and 95% confidence interval (CI). Results Our search led to the final selection of 9 articles including 15,287 patients. The effect of fluvoxamine, fluoxetine, and the overall effect of SSRI/SNRI use on mortality of COVID‐19 patients were investigated in 3, 2, and 7 articles, respectively. The results of our analyses showed that these medications could significantly decrease mortality of COVID‐19 patients (OR and 95% [CI]: 0.595 [0.467–0.758], 0.620 [0.469–0.821], and 0.596 [0.437–0.813]). The effect of SSRI/SNRIs on hospitalization events of COVID‐19 patients was not significant (OR: 0.240% and 95% CI: 0.041–1.4). Also, length of hospital stay was longer in patients who administrated SSRIs. Conclusion According to this study's results, SSRI/SNRIs may be effective in reducing mortality of COVID‐19 patients, suggesting the superiority of fluvoxamine to fluoxetine. The safety profile and affordable cost of SSRI/SNRIs for a short‐term use may be other reasons to propose them as beneficial medications in preventing mortality in COVID‐19.


| INTRODUCTION
From the beginning of the coronavirus disease 2019 (COVID- 19) pandemic in early 2020, a global endeavor has been started to find the best treatment, and an outstanding progress has been accomplished to minimize the transmission and complications of this infectious disease. 1 Although the basic mechanism and origin of this disease are still under investigation, male sex, obesity, higher age, immunosuppression, and comorbidities such as diabetes and chronic kidney disease are considered risk factors that are related to the severity of COVID-19. 2,3 Transmission of COVID-19 could occur by droplets spreading through human-to-human contacts or touching contaminated surfaces, and the infection can be considered as an airborne contagion. 4 It has been shown in a short-time study that climate does not influence outbreaks of COVID-19. 5,6 The transmission dynamics of COVID-19 was found to be related to air pollution as well. 7 Remdesivir, favipiravir, hydroxychloroquine, azithromycin, and corticosteroids were all a part of a world-wide effort to treat this disease. [8][9][10][11][12] For the first time, a randomized clinical trial was performed on fluvoxamine, as a selective serotonin reuptake inhibitor (SSRI), to determine the effect of this drug on the clinical deterioration of symptomatic COVID-19 patients and showed a promising impact on preventing the progress of the disease. 13 After this study, other SSRIs and specifically fluoxetine were studied for their effect on the course of COVID-19 infection. 14,15 The exact mechanism of SSRI/serotonin-norepinephrine reuptake inhibitor (SNRI) on coronavirus is not well known but there are some theories and explanations justifying its probable efficacy. 16 SSRIs are a part of functional inhibitors of acid sphingomyelinase and Coronavirus has been proven to use acid sphingomyelinase to enter the cells and as a result, these drugs may inhibit the entrance of the virus into the respiratory and other human cells. 17,18 This mechanism leads to decrease in cytokine and interleukin release, 19,20 conveying a positive effect on interference with host cells. 21 Showing anti-inflammatory effects 22 decrease in the aggregation of platelets and modulation of the degranulation of mast cells 16,23,24 are all of the many mechanisms described for their protective effects. Although many treatments strategies and vaccines have been developed and introduced during the past 2 years that have reduced complications and most importantly mortality rates, there still exists major concerns regarding the pathogenicity of probable new variants and future outbreaks due to the inefficient available data regarding the pathophysiology of the disease and ways to best control its complications. 25 A remarkable decrease in mortality and improvement of patients' clinical outcomes have been reported with SSRI/SNRIs in COVID-19 infection but the actual effect of these drugs remains questionable. 26 Considering that antidepressants such as SSRI/SNRIs may be used widely in the general population with their limited serious adverse effects and their manageable cost for short-term use, the benefits or drawbacks of these medications may be worthy of attention for use in COVID-19 infected patients.
In this systematic review and meta-analysis, we aimed to assess the impact of SSRI/SNRI use on mortality, hospitalization, and length of hospital stay in COVID-19 patients. Also, we are going to review eligible studies and investigate the accurate findings of each study.

| METHODS
This study was carried out according to Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guideline. 27

| Eligibility criteria
Studies that revised the effect of SSRI and/or SNRIs as repurposing drugs on PCR confirmed COVID-19 patients (hospitalized and nonhospitalized) and reported the drugs' effect on the outcome of the patients (mortality, hospitalization, and length of hospital stay), were found and selected to be included in this study. Selected study designs were randomized clinical trials, cohorts, case-control, and cross-sectional studies. We excluded the studies which were in vitro, cellular, molecular, hypothesis, non-English, conference papers, case series, case reports, reviews, editorials, unpublished or ongoing studies, or abstracts.

| Information source
The search process was performed on April 2022. Two separate authors have searched three electronic databases, including MED-LINE/PubMed, EMBASE, and Scopus, at the same time. For covering the possible missed studies, references of the final selected articles have also been evaluated for relevance to this study.

| Search strategy
We prepared a synonym Pandemic, COVID 19 Pandemic" that were combined together by using "OR" and then combined by using "AND" with the relevant words to SSRI/SNRIs which were "SSRI, Serotonin Uptake Inhibitors,

| Selection and data collection process
The selection and data collection process were performed manually, not using any filter or automatic tools, by two separate authors working parallel to each other. In the cases of noticing any conflicts between researchers, consultation with the third author was mandatory. All the search results from each database were exported to an endnote library, then duplications were removed. In the first step of the selection process, two authors have screened the articles' titles and abstracts, selected articles from this step entered the second step which was reading the full text of the articles and finding the matched articles with the deliberated eligibility criteria. The included articles after step 2, were discoursed in the group again, the decision on which outcomes to be extracted was made and the accuracy of extrapolated data was checked in between all authors.

| Data items
We attempted to extract the desirable demographic data and clinical outcomes from the selected studies. Study design, number of patients (case/control based on their design), study population, intervention/placebo, or exposure were our extracted demographic F I G U R E 1 PRISMA flowchart of included studies for determining the effect of SSRI/SNRI on clinical outcomes of COVID-19 patients. COVID-19, coronavirus disease 2019; PRISMA, Preferred Reporting Items for Systematic reviews and Meta-Analyses; SNRI, serotoninnorepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor data. Our main outcomes were mortality, hospitalization event, and length of hospital stay.

| Risk of bias assessment
The risk of bias and quality of each article were weighed by two separate authors, using National Institutes of Health (NIH) tools for evaluating the quality of studies. 28 The mentioned tool consists of questions assessing different parts of the studies such as randomization, allocation, blinding, selection, and selective reporting bias and validation for randomized clinical trials and defining population and exposure for observational or cohort studies. For each question, three options of "Yes" for being according to that index, "No" for not properly considered, and "N.A" or not applicable for not reported are considered.  infection had a higher risk of intubation/death. Also, the combination of antidepressants had no effect on the risk of intubation or death. 15 Confirming previous studies, Fei et al. 33 also found lower mortality rates in patients who were treated with SSRI and/or SNRI. This study pointed out that acute respiratory distress syndrome and endotracheal intubation were significantly lower in antidepressant-treated patients although respiratory failure was not significantly different.

| Effect measures and synthesis methods
The largest study with the largest sample size was a retrospective cohort that included 83,548 patients of which 3401 were exposed to SSRIs. A propensity score matching was performed between treated and control patients to diminish confounding factors. The study reported an 8% reduction in RR of mortality in patients who used any kind of SSRI medication, a 28% reduction in patients for whom fluoxetine was prescribed, and a 26% reduction in the group of patients using either fluoxetine or fluvoxamine. 32 The last study was a prospective cohort that congruent with other studies showed significantly lower mortality rate, ICU admission, and MV in fluvoxamine-treated patients. A remarkable finding of this study was that on day 14 of the infection, no COVID-19-related symptoms were detectable in the treatment group taking into account that 60% of non-fluvoxamine-treated patients presented with symptoms of infection 31 A very recent study reported findings that showed no benefit in using SSRIs contrary to the results of Hoertel et al. 15 In this retrospective study, 832 patients on SSRIs and 8212 patients without prior use of SSRIs were compared and no significant odds of reducing mortality was found. 35 This finding although at first glance may appear contrary to the results of this meta-analysis, but it is in line with our overall results and the results of Hoertel et al., 15 showing that long-term use of antidepressants probably in patients with a pre- Bonnet et al. 38 17 reported a significant hazard ratio for the association of some baseline characteristics of the patients and death/intubation. Obesity, male gender, older age, smoking, and underlying disease increase risk of severe COVID-19 and history of psychiatric disease were presented among the risk factors. Another study led to a conclusion that fluvoxamine's impact on mortality was more significant in women rather than men relevant to the fact that published articles on mortality of COVID-19 have also reported higher mortality rates among men. 30,40 This difference has been attributed to probable hormonal differences. Lower testosterone levels have been linked to increased inflammatory processes and estrogen has shown protective quality by showing anti-inflammatory effects. 40,41 This article described age, coronary artery disease, continuous renal replacement therapy, and fluvoxamine consumption as significant variables in determining the prognosis of COVID-19 infected patients. 30

| Possible mechanisms suggested for protective effects of SSRI/SNRIs on COVID-19 infection
Although the pure mechanism of the above-mentioned effects has not been fully discovered, several explanations have been reported in the literature. A meta-analysis and systematic review was conducted before the outbreak of COVID-19 that proposed the antiinflammatory effects of SSRIs. This study reported a significant decrease in the level of IL-6, tumor necrosis factor-alpha, IL-10, and C-C motif ligand 2 chemokine as inflammatory markers. 19 Another theory is the antisphingomyelinase effect of antidepressants.
Sphingomyelinase leads to ceramide production that is accounted as a receptor for viruses. The antisphingomyelinase property of antidepressants can therefore functionally inhibit entry of coronavirus into the human cells. 42,43 An in vitro experiment was performed on the effect of fluoxetine on the inhibition of coronavirus. This study reported a successful inhibitory effect of fluoxetine on the gene expression of this virus. 44 Fluvoxamine may also have an immunomodulatory effect by showing high affinity for sigma 1 receptor (S1R). 45 S1R has shown significant roles in modulating the immune system and inhibition of inflammatory cytokines in vitro. [46][47][48] Fluoxetine and fluvoxamine have shown to affect intravesicular pH thereby impairing the fusion of virus compartments and its internal spread. 49 Reduced platelet aggregation and degranulation of mast cells are other proposed mechanisms for fluvoxamine and its role in the reducing inflammation and mortality and improvement of patient outcomes in COVID-19 infection. 16,50 4.2 | The rationale for using SSRI/SNRI drugs for prevention of severe COVID-19 infection Although SSRI/SNRI drugs were first introduced as mood stabilizers, their indication now has been extended to the treatment of other disorders. [51][52][53][54] Considering the proposed mechanisms, it is somehow logical to consider these medications as an adjunct in the treatment of infectious diseases such as COVID-19. According to a report, a 1-month cost of an SSRI/SNRI will be an average of 75-320$ depending on the medication, considering it a manageable cost for a single duration of therapy. 55,56 Comparingly speaking a 5-day treatment course with remdesivir costs 2340$ which was recently introduced to prevent high-risk patients from developing severe infection. 56,57 Although SSRI/SNRIs have a range of adverse effects like all other medications, their adverse effects are variable between different populations and are mostly transient and leave no sequala for the patients in particular when used for a short duration of time. 58 Overall, SSRI/SNRI are safe, inexpensive, and effective drugs that may benefit COVID-19 patients before it is too late. Although these drugs are available in most parts of the world as psychiatric medications, providing extra dosage for treatment of COVID-19 needs multi-level and national-wide governance and cooperation with pharmaceutical companies to avoid shortages. 59 It is worth mentioning that although these drugs may reduce mortality of patients, they could not substitute vaccination as a preventive strategy. Regardless of the effect of vaccination on preventing COVID-19, all proposed preventive strategies are aimed at reducing the socioeconomic burden of this disease. 60

| Limitation
The major limitation of this study was the scarcity of available data on