Clinical outcomes of COVID-19 amongst HIV patients: a systematic literature review

OBJECTIVES The global pandemic coronavirus disease 2019 (COVID-19) emerged in the city of Wuhan, China around December 2019. Since then, the virus has caused severe morbidity and mortality worldwide and has put pressure on the global medical system. Still, there are limited data regarding the clinical impact of COVID-19 on people living with human immunodeficiency virus (HIV). The primary aim of this study was, therefore, to systematically review up-to-date studies reporting the clinical outcomes of COVID-19 amongst HIV patients. METHODS A thorough literature search was carried out using MEDLINE, Embase, Scopus, and the Cochrane Library Databases in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. RESULTS A total of 22 studies were identified. Amongst 730 HIV/COVID-19-coinfected patients, 79.4% were males, the median age was 51.5 years, and the number of reported patients receiving antiretroviral drugs was 708 (97.2%). Most coinfected patients had mild to moderate symptoms, including cough (37.7%), fever (37.5%), and dyspnoea (24.7%). Among pre-existing comorbidities, hypertension (26.3%) was the most prevalent in HIV/COVID-19 coinfected patients, and 87% of coinfected patients recovered. CONCLUSIONS Based on the existing data in this systematic literature review, HIV patients with pre-existing comorbidities, obesity, and older age should be considered as a high-risk group for COVID-19. Furthermore, coinfected patients appear to have marginally comparable clinical outcomes with the general population. The study’s findings highlight the need for further investigation to elucidate the impact of COVID-19 infection on HIV patients.

INTRODUCTION munodeficiency virus (HIV) infection results in a reduced number of CD4 cells and abnormal immune responses, leading to a weakened immune system and vulnerability to various pathogens and opportunistic infections [6], concerns about the outcomes of COVID-19 in HIV patients were immediately raised and carefully considered.
Although some scholars have speculated that antiretroviral drugs may favour HIV patients due to their activity against SARS-CoV-2 and other coronaviruses [7], there has been no evidence that HIV patients receiving certain antiretroviral drugs have an altered risk of COVID-19 infection and severity [8]. A recent study on the effectiveness of antiviral drugs has now directed focus on a drug known as tenofovir, which has been extensively used for HIV treatment and as pre-exposure prophylaxis for HIV prevention. The study found that tenofovir can bind to the RNAdependent RNA polymerase (RdRp) of SARS-CoV-2 and may therefore impair its function [9]. While this promising finding may be useful in future research, no treatment for COVID-19 has been found. Accordingly, researchers have been trying to provide a clearer insight into various aspects of COVID-19 outcomes in HIV patients. A study found that older age, late diagnosis, low CD4 cell count, and treatment-naive status were potential determinants of COVID-19 incidence amongst HIV patients [10]. Similarly, a study in China confirmed that advanced age and preexisting comorbidities, such as hypertension and diabetes, are associated with unfavourable outcomes and increased mortality from COVID-19 [11].
Despite the current urgency to obtain a clear understanding of COVID-19 outcomes in HIV patients, large-scale observational studies on disease severity, symptoms, multimorbidity, complications, and mortality of HIV and COVID-19 coinfected patients have not yet been conducted. Given the limited data on this subject, in July 2020, the Centers for Disease Control and Prevention asserted that older HIV patients and those who have pre-existing comorbidities might be at increased risk for severe illness [12]. In view of the uncertainties relating to COVID-19 in people living with HIV and the unexpected nature of the virus, special alertness towards HIV and COVID-19 coinfection is needed. This concern is consistent with the annual rise in HIV incidence, which increases the likelihood of patients becoming coinfected with COVID-19. Despite the limited data, this systematic literature review assembled existing data on the earliest and current reported cases to provide the basis for what is known and to present relevant findings. Therefore, this study identified and quantified different aspects of COVID-19 outcomes amongst HIV patients, including the proportion of various symptoms, severity, pre-existing comorbidities, recovery, death, and the most commonly prescribed antiretroviral drugs in HIV patients with COVID-19 coinfection.

Literature search
The methods and design used in this study comply with the Cochrane Collaboration Handbook, MOOSE (Meta-Analyses of Observational Studies in Epidemiology) and are reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [13] (Supplementary Material 1). The inclusion and exclusion criteria are summarized in Table 1.

Information sources
A thorough literature search was carried out using the following databases: Embase, MEDLINE, Scopus, and selected Internet sites (e.g., PubMed, Cochrane Collaborative Review Group on HIV Infection and acquired immunodeficiency syndrome [AIDS], and the Science Citation Index).

Search strategy
Certain keywords and terms obtained from a scoping search and acquired knowledge in the field were used in the phase of literature search. Two categories of the main keywords were used for the search, COVID-19 and HIV, and related study keywords and terms were searched accordingly.

Eligibility criteria
Inclusion and exclusion criteria were used for the search ( Table 1). Studies that did not meet the inclusion criteria or discussed different topics that did not contribute to the research were excluded.

Data extraction
Data were extracted using a standardized form and the assessment was carried out by classifying studies as eligible, non-eligible, or possibly eligible according to the eligibility and inclusion criteria [14]. A study was assessed for eligibility if its title and abstract were relevant and could not be excluded, and the process was based on the approval of the supervisor (Romero, R). In order to examine the data for consistency and clarity or in case of disagreement on the inclusion of studies, a third reviewer could be appointed to assess eligibility independently. The following data were extracted: sample size, study design, population characteristics (age, sex, CD4 cell count, viral load), used antiretroviral drugs, outcome (recovery, death, still in hospital), severity (mild, moderate, severe/critical) and related comorbidities (e.g., diabetes mellitus, hypertension, cardiovascular diseases, respiratory diseases, chronic kidney disease, obesity). Where necessary, authors were asked to provide the raw data or additional information.

Quality assessment
The quality and risk of bias assessment was carried out using the Newcastle-Ottawa Quality Scale [15], which includes 8 customized assessment sheet criteria that are categorised into 3 groups: selection, comparability, and outcome. The assessment sheet evaluated the representativeness of cases, research methods and the outcomes of each study. Quality assessment was conducted using different criteria based on various study designs. A score between 1 and 3 out of 8 implies a higher risk of bias due to inadequacy of reporting, the use of invalid and unreliable measures to define conditions and exposures, mis-representativeness of cases, unclear population characteristics, unblinded assessors, and insufficient follow-up of cases. Medium-risk studies scored 4-5 out of 8 and lower-risk studies scored 6-8 out of 8 (Supplementary Material 2).

Data synthesis
Data from the included studies were tabulated and categorized based on the study sample, comorbidities, recovery, severity, antiretroviral drug use, and other factors influencing the total number of patients. Accordingly, a descriptive analysis was used to report findings, and bar charts were used to display key findings. Stata version 16.0 (Stata Corp., College Station, TX, USA) was used for the analysis.
Unfortunately, due to a lack of data and variability in the comparison groups amongst the included studies, a meta-analysis for comparison between groups could not be performed. However,  when data permitted, a meta-analysis of proportions was conducted by calculating the pooled estimate of the proportion of recovery amongst patients with HIV/COVID-19 coinfection, using a random-effects model [16]. The meta-analysis of proportions was performed using the "metafor" package in R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). Data were weighted prior to combination to account for the various samples from different studies and to avoid pooling the data as if they were derived from a single sample.
A random-effects analysis was used, with a binomial distribution modelling within-study variability and parameters estimated using a maximum likelihood procedure. To calculate the confidence intervals (CIs) for between-study variance, the Jackson [17] method was used. The overall proportions were shown, along with 95% CIs. The I 2 statistic was used to report between-study heterogeneity, and the p-value confirmed the result.

Publication bias assessment
Publication bias was assessed using funnel plots, and the degree of asymmetry was tested using Egger's regression test [18]. Both analyses were performed in R using the "metafor" package.

Ethics statement
As the present study was a systematic review, no ethics statement was needed.

Study selection
A total of 632 articles were found. Following the PRISMA guidelines, 247 duplicate articles were excluded and after screening titles and abstracts, 339 additional articles were excluded. Only 22 of the studies were included for analysis after full-text screening of 44 articles.
All 22 included articles were observational studies, of which there were 9 cohort studies (including 3 retrospective studies), 9 case series, 2 cross-sectional studies, 1 case-control study, and 1 case report. Figure 1 shows a PRISMA flow chart that explains the inclusion and exclusion process for this systematic literature review.

Study characteristics
The finally selected studies were listed according to the first author, publication year, country, study design, sample, antiretroviral drugs, findings, and outcomes. The study characteristics are detailed in Table 2 [10,. All of the selected studies were observational studies that provided data on the outcomes of COV-ID-19 amongst HIV patients, clinical outcomes amongst antiretroviral therapy (ART) users, demographic data, and severity determinants in coinfected cases. Six studies were conducted in the United States, 6 in China, 3 in Spain, 2 in Italy, and the remaining 5 studies were conducted in France, Germany, United Kingdom, Turkey, and Iran. All of the studies were conducted between April 15, 2020 and November 17, 2020. An overview of the included studies is provided in Table 3 [10,.

Quality and risk of bias within studies
The quality of studies and risk of bias were reported using the Newcastle-Ottawa Scale following approval from the supervisor (Romero, R). Although all of the included studies are observational, different criteria were customised to assess the quality and bias within studies based on the study design. The assessment was carried out using a rating scale ranging from 1 to 8 out of 8. A higher score implies a lower risk of bias due to adequacy of reporting, the use of valid and reliable measures to define conditions and exposures, representativeness of cases, presence of nonexposed in cohorts, clarity of characteristics, blinding assessors, sufficient follow-up of cases, and clear outcomes. None of the included studies scored lower than 4, and therefore none were classified as high-risk studies. Where certain data were missing and no additional information could be received from the authors, the ratings were classified as "unknown" (Supplementary Material 2).

Synthesis of results
The synthesis of results from different observational study designs across outcomes of COVID-19 amongst HIV patients is provided in Table 4. Overall, 730 HIV patients with COVID-19 coinfection were reported in all studies. The patients were predominantly male (79.4%) and the mean age of all reported cases was 49.11 years. A low CD4 cell count ( < 200 cells/mm 3 ) was reported in 87 of 470 patients (18.5%) and the viral load was higher than 50 copies/mL in 41 of 393 patients (10.4%). The number of   Figure 3 and Table 4). When data allowed, the meta-analysis of proportions included all 22 studies to estimate the recovery proportion of HIV/COV-ID-19-coinfected patients. None of the included studies were excluded from the analysis due to the low-medium risk of bias (scored 4-8). The pooled recovery proportion of HIV/COVID-19 coinfection was 87% (95% CI, 83 to 91), as shown in Figure 4. The weight-  ed mean of the estimates from each included study revealed differences in sample sizes between studies. However, using a random-effects model, the results showed an insignificant degree of heterogeneity between studies, as confirmed by the p-value (I 2 = 27%, p = 0.12, π 2 = 0.1335). Although low, a considerable amount of the variability between studies (I 2 = 27%) could be attributed to clinical and methodological differences; for example, some studies included patients with multiple morbidities and those of older age, both of which are risk factors for a severe course of COV-ID-19, resulting in a lower recovery ratio. Furthermore, this systematic literature review included various observational study designs that may have contributed to a relatively high degree of variability.

Publication bias
The funnel plot showed a symmetrical scattering of points throughout the funnel, indicating that publication bias was unlikely. Furthermore, the Egger regression test revealed no significance in the asymmetry degree of the funnel plot (p = 0.85) (Supplementary Material 3).

DISCUSSION
By December 2020, more than 73 million confirmed cases of COVID-19 and more than 1.6 million worldwide deaths had been reported by the WHO [2]. In line with the aim of this study, 730 patients coinfected with COVID-19 and HIV were investigated.
The majority of patients were male and most of them received antiretroviral drugs. Furthermore, 81.9% of coinfected cases had mild to moderate clinical symptoms, primarily including cough, fever, and dyspnoea [20,35]. Due to the impaired immune defences due to the underlying disease, including in patients receiving treatment, HIV patients were thought to be at higher risk of developing severe forms of COVID-19 [40]. Although comparative tests were not possible due to the limited data and differences in comparative groups between studies, some of the included studies revealed a similar representation of people living with HIV across the population. An earlier cohort from New York City, United States found no differences in adverse outcomes between HIV/ COVID-19-coinfected patients and a similar comparison group [34]. In Wuhan, China, a cohort study indicated that case-severity rates of COVID-19 in HIV patients were comparable to those in the overall population [28], and in Italy, coinfected HIV patients with COVID-19 were found to be at no higher risk of severe illness than HIV-negative patients [24].
While further research is needed to provide evidence regarding the prognosis of COVID-19 in HIV patients and other groups, this study has shown that older age and the presence of comorbidities, including hypertension, obesity, diabetes mellitus, renal diseases, cardiovascular diseases, chronic respiratory disease, liver disease, and malignancy, are associated with a poor prognosis in COVID-19 patients, including the risk of death. This aligns with an earlier report finding that the most prevalent comorbidities among COVID-19 patients were hypertension, obesity, chronic lung disease, and diabetes mellitus [41]. Further evidence on the relationship between pre-existing comorbidities and adverse outcomes of COVID-19 was reported in a previous systematic review, which found that older age and the presence of one or more comorbidities increased COVID-19 severity [42], and these findings also align with those presented by Meyerowitz et al. [30] who reported that 83% of all HIV/COVID-19-coinfected patients had comorbidities associated with severe COVID-19 outcomes. Considering the age factor, a relatively large cohort study found that older patients (median age, 63 years) had a higher risk of developing severe disease [43]. Despite the lower median age in this systematic literature review (51.5 years), some of the included studies confirmed a similar association in coinfected patients with COV-ID-19 and HIV [23,29], taking into account that the median age difference may be due to the changing age distribution of COV-ID-19 between May 2020 and August 2020, which is thought to be due to behavioural and occupational factors as younger people are less likely to follow safety precautions [44]. A similar age shift was identified in Europe, where the median age of COVID-19 cases was 54 years in the first half of 2020 and then decreased to 39 years during June-July 2020 [45].
As of December 31, 2020, almost 68 million (97%) people had recovered from COVID-19 infection worldwide [2]. Despite the high recovery percentage globally, this systematic literature review showed a lower percentage of recovery in both a simple pooling of unweighted studies (90.6% recovery) and a meta-analysis of proportions following weighting of studies (87.0% recovery). These findings are consistent with the high proportion of death amongst HIV/COVID-19-coinfected patients (9.4%). However, different studies have shown conflicting findings regarding the coinfectionrelated deaths in patients with HIV and COVID-19. Huang et al. [28] reported a lower percentage of death related to coinfection (5.71%). A study in France [23], in contrast, reported a considerably lower percentage of deaths (1.8%). This significant variation of results is most likely due to selection bias, as many patients were admitted to hospital with severe symptoms. Another reason may be the presence of confounders, including certain comorbidities or determinants that increase the risk of death from COVID-19, such as hypertension, obesity, diabetes mellitus, and old age [42], which is consistent with the findings of the included studies in this systematic literature review [19,23,25,26,34,36].
Owing to the increase in the number of COVID-19 cases worldwide and the global demand for treatment, some HIV antiviral drugs have been under the spotlight. This particular interest was driven by a finding from the 2004 severe acute respiratory syndrome (SARS) outbreak, where it was observed that none of 19 HIV/AIDS patients receiving ART had contracted SARS despite being in close contact with SARS patients [46], which led to the hypothesis that the use of ART could prevent SARS from developing and could potentially reduce the severity and mortality of COVID-19. While the previous literature did not provide conclusive evidence on the effect of ART against SARS-CoV-2 infection [47], Elfiky [9] found that various antiviral drugs, including riba-virin, remdesivir, sofosbuvir, galidesivir, and tenofovir (NRTI) showed potential activity against SARS-CoV-2 by binding to its RdRp. This finding is congruous with previous studies suggesting that NRTIs inhibit the RdRp of COVID-19 and therefore may be effective against COVID-19 infection [48,49]. While this systematic literature review found that NRTIs were the most commonly used type of ART amongst HIV/COVID-19 patients, this finding cannot explain the relatively low severity of COVID-19 amongst HIV patients. Future research is therefore needed to further understand the role of antiretroviral drugs in reducing the severity of COVID-19 amongst HIV patients.

Strengths and limitations
In view of the rapidly changing evidence of the ongoing global pandemic, this study has been continuously updated over the course of 2020 and therefore includes relevant data. This research addresses an important and interesting concern that adds to the essential information needed to further understand the impact of COVID-19 on HIV patients. Considering the unstable and mutable nature of COVID-19, this study provides an up-to-date and in-depth analysis of the clinical outcomes of COVID-19 amongst HIV patients. Furthermore, the results highlight additional evidence on the clinical determinants of COVID-19 severity in HIV patients and their impact on patients' outcomes and quality of life. There are also some limitations to this study. First, most of the included studies have coinfected patients with a CD4 cell count greater than 200 cells/mm 3 and an HIV viral load lower than 50 copies/mL; thus, the results may not represent uncontrolled HIV patients. Second, while data on the antiretroviral drugs used in coinfected patients with HIV/COVID-19 were available, a comparison between the antiretroviral drugs received in coinfected cases of HIV/COVID-19 and HIV without COVID-19 was not possible due to lack of information. Some of the included studies lacked a comparison group and therefore provided limited information on the determinants associated with HIV/COVID-19 coinfection. For some of the results, the findings were influenced by a single large study [22]. Furthermore, most of the included studies had small samples, and various confounders were not included in the reporting of such data.

CONCLUSION
Based on the findings of this systematic literature review, coinfected patients with HIV/COVID-19 have marginally comparable clinical outcomes with the general population. HIV patients with pre-existing comorbidities should be considered as a high-risk COVID-19 group, along with those who are obese and older. Healthcare providers will therefore need to take particular care of HIV patients with multiple morbidities to prevent unnecessary poor outcomes. Given the mutable nature of COVID-19, HIV patients should still be advised to take additional precautions and protect themselves as new information is rapidly updated. Despite the lack of evidence on the effectiveness of antiretroviral drugs against