Efficacy and safety of adding fluoxetine to the treatment regimen of hospitalized patients with non-critical COVID-19 pneumonia: A double-blind randomized, placebo-controlled clinical trial

Abstract

Introduction

Selective serotonin reuptake inhibitors are considered the drugs, whose effectiveness in viral pandemics has been studied. The aim of this study was to evaluate of adding fluoxetine to the treatment regimen of patients with COVID-19 pneumonia.

Methods

This study was a double-blind randomized placebo controlled clinical trial .36 patients in the fluoxetine and 36 patients in the placebo group were enrolled. Patients in the intervention group were first treated with fluoxetine 10 mg for 4 days and then the dose of 20 mg was continued for 4 weeks. Data analysis was conducted using SPSS V. 22.0.

Results

There was no statistically significant difference between the two groups in terms of clinical symptoms at the beginning of the study and also the score of anxiety and depression, oxygen saturation at the time of hospitalization, mid-hospitalization and discharge periods. The need for mechanical ventilator support (p = 1.00), the need for admission in the intensive care unit (ICU) (p = 1.00), rate for mortality (p = 1.00), and discharge with relative recovery (p = 1.00) were not significantly different between the two groups. The distribution of CRP within the study groups showed a significant decrease during different time periods (p = 0.001), and although there was no statistically significant difference between the two groups on the first day (p = 1.00) and at discharge (p = 0.585), mid-hospital CRP showed a significant decrease in the fluoxetine group (p = 0.032).

Conclusion

Fluoxetine resulted in a faster reduction of patients' inflammation without association with depression and anxiety.

INTRODUCTION

The coronavirus disease 2019 (COVID-19) was primarily detected in the city of Wuhan, China, and then became an ongoing pandemic.^1,2 Clinical manifestations of coronavirus-acute respiratory syndrome-2 (SARS-COV-2) disease range from asymptomatic cases to severe pneumonia and acute respiratory distress syndrome that require intensive care.^3,4 Very limited treatment options for COVID-19 management have made vital the urgent need to identify safe, effective and globally available drugs to control the disease.^5,6 Although there is no definite therapeutic intervention for this disease, there are several drugs whose positive results are proven in some randomized controlled clinical trials.⁷ Among the drugs whose effectiveness in viral pandemics such as influenza, enteroviruses (coxsackievirus) have been studied in numerous clinical trials in human and animal species were psychotropic drugs such as selective serotonin reuptake inhibitors (SSRIs).⁵ Fluoxetine is a SSRI with low and tolerable side effects in patients with psychiatric disorders.^8,9

Previous studies have shown that treatment with antidepressants in patients with major depressive disorder (MDD) has been associated with a significant reduction in hyperactive inflammatory processes.^10,11 Based on the results of a meta-analysis, the use of antidepressants was associated with decreased plasma levels of several cytokines such as interleukin-6, tumor necrosis factor-alpha, and chemokine ligand 2, which may be associated with COVID-19 disease severity.¹¹ The mentioned effects of SSRIs which inhibit the production of alpha-microglial tumor necrosis factor and nitric oxide by regulating cAMP signaling have been shown to be more potent than serotonin-norepinephrine reuptake inhibitors. Short-term use of low-dose antidepressants, especially SSRIs, is well tolerated in the elderly group, who are prone to COVID-19.^12,13

According to the results of a study, treatment with fluoxetine inhibited the SARS-COV-2 infection up to 99% in a dose-dependent manner, leading to Vero and Calu-3 cell damage in COVID-19 infection, and increased the endolysosomal cholesterol resources.⁵ As the use of psychiatric medications may interfere with drugs used in COVID-19 treatment, and lead to worsening of these drugs adverse effects, identifying these interactions (mainly through two pharmacokinetic drug–drug interactions and also a combination of drug side effects) can be critical in improving the survival of patients with COVID-19.^7,10

Although several studies have been performed on the anti-inflammatory role of fluoxetine in viral diseases,^9,14 these studies were conducted in vitro, and on SARS-COV-2 and H1N1 influenza virus, and studies on human samples were in a descriptive design. Considering that achieving the best valid and reliable treatment options are through implementation of clinical trials, this study aimed to determine the efficacy and safety of adding fluoxetine to the treatment regimen of hospitalized patients with noncritical COVID-19 pneumonia.

MATERIALS AND METHODS Study design, sampling method and sample size

The present double-blind randomized placebo controlled clinical trial was conducted in Imam Khomeini general hospital, affiliated with Mazandaran University of Medical Sciences (MAZUMS), Sari, Mazandaran, Iran. The inclusion criteria consisted of (1) age of 16 to 65 years old, (2) dominant to the Persian language, (3) hospitalized patients with COVID-19 infection with positive PCR results or typical lung CT scan findings, (4) no history of MDD and other acute psychiatric disorders and no consumption of psychotropic drugs during the last 8 weeks, (5) no comorbid physical disorders such as cancer, (6) no substance and smoking related disorders, (7) not being admitted to the intensive care unit (ICU) at the beginning of the study, (8) no consumption of melatonin as a protocol treatment during hospitalization, (9) constant treatment regimen, (10) no pregnancy, (11) not having a critical stage of COVID-19 infection, (12) ability to take oral medications and (13) a score of less than 8 on hospital anxiety and depression scale (HADS).

The exclusion criteria included: (1) Occurrence of drug allergies or severe drug side effects, and (2) changing the treatment protocol.

Considering the effect size of 0.6, 95% confidence level, and 80% test power, and using the comparison of the two means formula and utilization of G-power software, sample size was calculated to equal to 72 individuals (36 patients in each group). The patients were included in the study by convenient sampling and then they were placed in two intervention groups (fluoxetine) and placebo (placebo) through random allocation.

Measurement instruments

The data collection tools were questionnaire of sociodemographic information and clinical characteristics, and fluoxetine side effects check list, and the HADS.

Demographic information and clinical features

The questionnaire which was designed according to the literature review included variables such as age, gender, the hospitalization ward, marital status, residency, education, occupation, body mass index (BMI), length of hospitalization, oxygen saturation, history of previous physical illnesses, methods of COVID-19 diagnosis, symptoms at admission, laboratory tests, patient's respiratory status, need for intubation, duration of intubation, need for hospitalization in the ICU during the intervention and treatment outcomes including recovery, partial recovery and death.

Fluoxetine side effects check list

This check list was prepared based on all the side effects listed for fluoxetine on the Food and Drug Administration website and included a list of all drug side effects regardless of their incidence.¹⁵

Hospital anxiety and depression scale (HADS)

Hospital anxiety and depression scale was developed by Sigmund et al in 1983¹⁶ and leads to the distinction between the psychological and the physical aspects of patients who have been referred to medical settings with physical symptoms.¹⁷ The scale is a 14-item tool designed to screen the depression and anxiety symptoms in patients over the past week. This tool has seven-part depression and seven-part anxiety subscales and is scored from zero to three on a four-point Likert scale. The scores on the scale range from zero to 21. For both subscales, scores from 0–7, 8–10, 11–14 and 15–21 are considered normal (no depression or anxiety), mild, moderate and severe respectively.¹⁸ The validity and reliability of this scale were measured frequently in different studies. Based on the results of a study in China, the reliability of this tool for subscales of anxiety and depression was confirmed with Cronbach's alpha 0.75 and 0.76, respectively.¹⁹ Also, the reliability of the anxiety and depression subscales of the Persian version of the tool was determined and confirmed with Cronbach's alpha of 0.78 and 0.86, respectively.¹⁸

Procedure

This research was approved by MAZUMS with the Ethics committee code of IR.MAZUMS.REC.1399.760 and Iranian Registry of Clinical Trials number of 20 200 904 048 616 N1. After obtaining the permission to conduct the research, necessary explanations were provided to the relevant officials in the COVID-19 wards, and the researchers were introduced to the head nurses of the wards to carry out this study. Before participating in the study, the objectives and design of the study were explained to each patient and written informed consent was obtained.

These patients were informed that they can withdraw from the study if they were not willing to cooperate at any stage. Of note, all the patients' personal information was remained confidential throughout the study and afterward.

Participants were evaluated in two stages in terms of inclusion criteria. At first, the information of the patients' records was checked and if they seemed eligible, additional information was obtained to complete the inclusion criteria. The interview was conducted by a last year psychiatric assistant to rule out psychiatric disorders according to the criteria of the fifth edition of Diagnostic and Statistical Manual of Mental Disorders-Text Revised.

Then, the randomization of eligible individuals was performed by block allocation randomization process. The blocking process was performed with random allocation software. Thus, 12 blocks of 6 were generated and the samples were assigned to each block based on the random numbers generated. This procedure was performed by one of the project colleagues who was not involved in examining the patients. The patients and the person evaluating the therapeutic effects were blinded to the intervention or placebo group. Questionnaires of sociodemographic information and clinical characteristics and HADS were completed before the intervention.

For eligible patients enrolled in the study, in addition to the recommended medication regimen according to the national COVID-19 treatment protocol which was similar in the intervention and placebo groups, treatment with methylprednisolone at a dose of 0.5 mg per weight daily, prophylaxis of venous thromboembolism with enoxaparin anticoagulant and supportive measures was also done.

Patients in the intervention group were first treated with fluoxetine 10 mg for 4 days²⁰ and then the dose of 20 mg was continued for 4 weeks. Simultaneously, patients in the placebo group received the placebo for 32 days. Even in cases that were discharged, the intervention continued and patients' medical status was monitored by telephone each week. The fluoxetine capsule used in this study was obtained from Abidi Pharmaceuticals Company which is available in green and yellow colors on the market.

The placebo was prepared by Abidi Company too. The fluoxetine and placebo of the patients were placed separately in cans with the patient's randomization code by a pharmacologist colleague and then was given to the researcher who was in contact with the patients. At the end of the intervention, HADS was completed by patients in both groups again.

It should be noted that during the intervention, patients' vital signs, inflammatory markers and the disease states were evaluated on a daily based during the admission period, and weekly electrocardiograms were taken for them. CRP levels were measured at the time of admission and, if necessary, on the third and fifth days, as well as middle time of hospitalization and at the time of discharge. In addition, oxygen saturation status in patients of both groups was checked daily until discharge. Side effects of fluoxetine were checked using the relevant checklist on the third and seventh days and at the end of hospitalization.

Data analysis

Data analysis was conducted using Statistical Package for the Social Sciences, version 22.0 (SPSS Inc). To describe quantitative variables, descriptive statistics such as mean and standard deviation and also to describe qualitative variables, frequency and percent were used. To evaluate the normality of values, the Shapiro–Wilk test was performed. The Fisher exact test was administered to compare the sociodemographic and clinical symptoms, as well as depression and anxiety at the time of admission between the study groups. The chi-square test was used to compare the disease outcomes and the Friedman and the Mann–Whitney tests were administered to compare the CRP distribution within and between study groups over different time periods. Also, to control the effect of interfering factors, the generalized estimating equations test was used. It should be noted that a p-value less than 0.05 was considered a significant level.

RESULTS

Out of 202 inpatients with COVID-19 hospitalized during the sampling period, 73 were not included in the study due to lack of inclusion criteria. Of 129 patients who were eligible for the study, 57 individuals did not initially consent to participate in this study. Thus, 36 patients in the fluoxetine and 36 patients in the placebo group were enrolled in the trial. In the fluoxetine group, 3 patients were excluded due to ICU admission for ventilator support (n = 1), bleeding diseases (n = 1) and death (n = 1), and in the placebo group, 2 patients were excluded due to ICU admission for ventilator support (n = 1) and death (n = 1). Finally, 33 patients in the intervention and 34 patients in the placebo group completed the study and were included in the final analysis (Figure 1).

FIGURE 1. The consort flow diagram.

Comparison of baseline demographic characteristics between the two groups

The demographic characteristics of patients in two groups are presented in Table 1. The results showed that the mean age of patients in this study was 52.62 ± 11.02 years (53.52 ± 9.81 years in the placebo group and 51.72 ± 12.19 years in the fluoxetine group). In the study, 37 (51.4%) participants were men, 67 (93.1%) were married and most of them (41.7%) had high school diplomas and associate degrees. Also 24 (33.3%) of them were homemakers and most of them (91.7%) had city residency. The results also indicated that according to the Chi-square test, there were no statistically significant differences between the two groups regarding the variables such as, gender (p = 0.479), marital status (p = 0.077), education (p = 0.495) and occupation (p = 0.255). According to Fisher's test, there was also no statistically significant difference between the two groups regarding the place of residency (p = 0.674) (Table 1).

TABLE 1 Demographic characteristics by study groups.

[Table omitted. See PDF]

^*Frequency (Percentage)

^**Inter-group comparison with Chi-Square Tests

^***Inter-group comparison with Fisher's exact Test

^****Frequency (Mean ± SD).

Comparison of basic clinical characteristics between the two groups

Basic clinical characteristics were not significantly different between the two groups as depicted in Table 2. Fisher's exact test on the clinical and diagnostic characteristics showed that that there were no statistically significant differences between the placebo and the fluoxetine group regarding BMI (p = 1.00), medical history of diabetes (p = 0.300), medical history of cardiovascular diseases (p = 0.115), history of anti-diabetic medications (p = 0.285), history of heart drugs (p = 0.115), history of corticosteroids (p = 0.493) and the history of other drugs (p = 0.260) consumption. However, variables such as history of dyslipidemia (p = 0.018), medical history of other diseases (p = 0.024), the history of anti-hypertensive agents (p = 0.045) and the history of anti-dyslipidemia agent's (p = 0.018) consumption were significantly different between the two groups.

TABLE 2 Clinical and diagnostic characteristics by study groups.

[Table omitted. See PDF]

Note: Bold indicates statistical significant value (p < 0.05).

Significant level ≤ 0.05.

^*Frequency (Percentage)

^**Inter-group comparison with Fisher's Test

^***Inter group comparison with Chi-Square Tests.

According to the Chi-square test, the method of COVID-19 diagnosis (p = 0.926) was not significantly different between the two groups. However, the history of antihypertensive medication (p = 0.025) was significantly higher in the placebo group (Table 2).

Comparison of clinical symptoms at admission between the two groups

According to Fisher test in Table 3, respiratory symptoms (p = 0.260), fever (p = 0.786), fatigue (p = 0.260), myalgia (p = 0.396), neurological symptoms (p = 0.541), psychiatric symptoms (p = 1.00) and gastrointestinal symptoms (p = 0.173) and lactate dehydrogenase (LDH) (p = 1.00) were not statistically different in the two groups. Furthermore, according to Chi-square test, the percentage of lung involvement in CT scan (p = 0.798), ferritin (p = 0.250), D dimer (p = 0.085) and interleukin-6 (p = 0.474) were not statistically different in the two groups.

TABLE 3 Clinical symptoms at admission by study groups.

[Table omitted. See PDF]

Comparison of depression and anxiety at the time of admission between the study groups

The HADS results showed that the two groups were asymptomatic and similar to each other at the time of admission in terms of anxiety and depression. According to the Fisher test at the end of the study, the anxiety (p = 1.00) and depression (p = 1.00) scores had no statistically significant differences between the two groups (Table 4).

TABLE 4 Comparison of depression and anxiety according to the hospital depression and anxiety scale at the time of admission between the study groups.

[Table omitted. See PDF]

Comparison of disease outcomes in hospitalized patients with COVID-19 by study groups

According to Fisher's test, in terms of disease outcomes in patients hospitalized with COVID-19, such as the need for mechanical ventilator support (p = 1.00), the need for ICU admission (p = 1.00), mortality (p = 1.00), and the disease outcome (p = 1.00) there were no statistically significant differences between the two groups. According to Chi-square test, oxygen saturation at hospitalization (p = 1.00), mid-hospitalization (p = 0.767) and at discharge (p = 0.243) were not statistically different between the two groups (Table 5).

TABLE 5 Comparison of disease outcomes in hospitalized patients with COVID-19 by study groups.

[Table omitted. See PDF]

Comparison of CRP distribution within and between study groups over different time periods

According to the Friedman test, the distribution of CRP within the study groups showed a significant decrease during different time periods, (p = 0.001) and also according to the Mann–Whitney test, there was no statistically significant difference between the two groups regarding CRP on the first day (p = 1.00) and at discharge (p = 0.585). However, mid-hospital CRP showed a significant decrease in the fluoxetine group (p = 0.032) (Table 6).

TABLE 6 Comparison of CRP distribution within and between study groups over different time periods.

[Table omitted. See PDF]

Note: Significant level ≤ 0.05.

^*Mean ± SD

^**Inter-group comparison with Mann-whitney U

^***Intra-group comparison with Friedman Test.

DISCUSSION

Given there are several studies showed the effectiveness of antiviral activity of antidepressants such as fluoxetine on SARS-CoV-2 in laboratory conditions, this study aimed to evaluate the efficacy and safety of adding fluoxetine to the treatment regimen of hospitalized patients with non-critical COVID-19 pneumonia in a double-blind, placebo-controlled randomized clinical trial. In this study patients in fluoxetine and placebo groups had similar demographic and basic clinical features. The results of this study indicated that the need for mechanical ventilator support, the need for admission in the ICU, rate of mortality, and treatment outcomes were not significantly different between the two group. Also, the results indicated that there was no statistically significant difference between the two groups regarding the level of CRP on the first day and at discharge. However, mid-hospital CRP levels showed a greater decrease in the fluoxetine group.

A retrospective case–control study conducted to assess the relationship between 20 mg/day fluoxetine consumption in 110 patients with COVID-19 pneumonia showed that mortality in the fluoxetine-treated group was one-third of the control group. In comparison with our study results, that study had a larger sample size and was conducted in the retrospective design.²¹

The results of a study examining the relationship between the use of some antidepressants and the reduction in the risk of intubation and death in patients with COVID-19 showed that a history of using fluoxetine, citalopram, paroxetine, venlafaxine and mirtazapine was associated with a reduced risk of intubation and death.¹²

Another study that examined the effect of SSRIs such as fluoxetine, citalopram and paroxetine on viral replication supported the antiviral in vitro effects of fluoxetine on COVID-19 disease. However, this effect had not been observed by paroxetine and citalopram. Also, it was concluded that fluoxetine inhibits the proliferation of SARS-COV-2 at a concentration of 0.8 μg/mL. Treatment with fluoxetine decreased the protein expression and showed that this drug acts as an upstream of gene expression.⁶ It was also indicated that fluoxetine has properties such as anti-inflammatory activity leading to cell survival, antiapoptotic properties and a strengthening role in the antioxidant defense system.²² Recent studies also have suggested that several antidepressants may inhibit sphingomyelin acid activity that may prevent epithelial cell infection with SARS-CoV-2, and fluoxetine may inhibit the antiviral effects on SARS-CoV-2 in an in vitro condition.^23,24

Another study was conducted to investigate the effect of early treatment with fluvoxamine on the risk of emergency care and hospitalization among patients with COVID-19. That study was conducted among high-risk symptomatic adults from 11 clinical sites in Brazil, who, in addition to being positive for SARS-CoV-2, had a known risk factor for developing severe disease. Patients were randomly divided into fluvoxamine (100 mg twice daily for 10 days) and placebo groups. The results showed that treatment with fluvoxamine reduced the need for hospitalization among high-risk outpatients with early diagnosis of COVID-19.²⁵ Also, the results of another study in the USA with a lower sample size which assessed the effect of fluvoxamine (100 mg 3 times per day) on the prevention of severe COVID-19 in outpatients, showed consistent results with the previously mentioned study.²⁶ The findings of these studies were not consistent with the current study due to differences in the selected drug. However, it should be considered that fluvoxamine and fluoxetine are both categorized in SSRIs group.

A retrospective cohort study, which examined the mortality risk among patients with COVID-19 who were prescribed SSRIs when compared with matched control pateints, showed that the mortality rate among patients treated with SSRIs was 14.6% and among patients in the control group, it was from 16.3% to 16.6%, which showed a reduction of 8%–10% in relative risk. Also, the mortality rate among patients treated with fluoxetine was 9.8%, and among the control group, it was between 13.3% and 13.4%. Overall, mortality was reduced by 8% among patients receiving each type of SSRIs and by 28% among patients receiving fluoxetine. That study was performed only on patients admitted to the ICU, and those patients experienced more severe symptoms of COVID-19 and also had a higher mean age. Also, the participants in that study were already treated with SSRIs and therefore had a constant blood level of the drug at the time of admission.²⁷

These findings have provided evidence in favor of the effectiveness of SSRIs on the outcomes of patients with COVID-19. However, larger and more precise clinical trials are needed to elucidate the effect of SSRIs in general or fluoxetine in particular on the severity of disease outcomes.

Study limitations

The double-blind nature of this study along with the severe symptoms and difficult physical condition of patients led to many of them being reluctant to participate in the study, which made the process of finding volunteers difficult. One of the limitations of the present study is the small sample size compared to other studies. Another limitation of this study was that the majority of hospitalized patients were elderly patients with multiple physical comorbidities, which may cause decreased generalization of the results to a younger population with lower comorbidities.

Also, another limitation of this study was that the drug was started at a dose of 10 mg and then increased to a dose of 20 mg to reduce the risk of side effects. This made the time to reach a stable serum level of the drug prolonged and in some cases the patient was discharged during this time. However, in this study, none of the patients reported the side effects of fluoxetine and the evidence of reduced inflammatory processes was observed in the fluoxetine group. It is suggested that in future studies, the sample size would be calculated with the higher power to examine whether stronger findings can be achieved with larger sample size. Also the dose of the drug could be increased more rapidly and, if possible, increasing the dose to higher than 20 mg per day, can lead to greater efficacy of the drug.

CONCLUSION

Although CRP levels were not significantly different between the two groups on the first day and at the time of discharge, CRP levels in the fluoxetine group were significantly lower versus that in the placebo group in the middle of the treatment course.

AUTHOR CONTRIBUTIONS

FE, FS, MZ, FAA, MM, HM, SA and RG conceived and designed the study. FS, FAA, FE, HM and RG collected the clinical data. MM performed the statistical analysis. FE, FS, MZ and FAA interpreted the clinical data. FS drafted the manuscript. FE, HM and MZ revised it critically for important intellectual content. All authors read and approve the final manuscript.

ACKNOWLEDGMENTS

The study was done with the financial support from the Vice Chancellor for Research of MAZUMS. We thank all those who helped us in this study. We also thank the patients for their cooperation in the fulfillment of this study and all hospital staff involved in patient treatment.

FUNDING INFORMATION

This study was supported by the Mazandaran University of Medical Sciences. (Grant number = 8354).

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflict of interest.

DATA AVAILABILITY STATEMENT

Since our data contain potentially sensitive personal information, it is forbidden to share these data with a third party without obtaining an additional written informed consent for information sharing. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

ETHICS APPROVAL AND CONSENT FOR PUBLICATION

This research project was approved by the Ethics Committee of MAZUMS (IR.MAZUMS.REC.1399.760), Sari, Iran.

PATIENT CONSENT STATEMENT

Informed consents were obtained from all participants and they were ensured that their identity would be kept anonymous throughout the study.