Systematic review of the efficacy and safety of antiretroviral drugs against SARS, MERS or COVID-19: initial assessment.

Abstract Introduction Several antiretroviral drugs are being considered for the treatment of COVID‐19, the disease caused by a newly identified coronavirus, (SARS‐CoV‐2). We systematically reviewed the clinical outcomes of using antiretroviral drugs for the prevention and treatment of coronaviruses and planned clinical trials. Methods Three databases were screened from inception to 30 March 2020 for studies reporting clinical outcomes of patients with SARS, MERS or COVID‐19 treated with antiretrovirals. Results From an initial screen of 433 titles, two randomized trials and 24 observational studies provided clinical outcome data on the use of antiretroviral drugs; most studies reported outcomes using LPV/r as treatment. Of the 21 observational studies reporting treatment outcomes, there were three studies among patients with SARS, six studies among patients with MERS and 12 studies among patients with COVID‐19. In one randomized trial 99 patients with severe COVID‐19 illness were randomized to receive LPV/r (400/100 mg twice a day) and 100 patients to standard of care for 14 days: LPV/r was not associated with a statistically significant difference in time to clinical improvement, although LPV/r given within 12 days of symptoms was associated with shorter time to clinical improvement; 28 day mortality was numerically lower in the LPV/r group (14/99) compared to the control group (25/100), but this difference was not statistically significant. The second trial found no benefit. The certainty of the evidence for the randomized trials was low. In the observational studies 3 out of 361 patients who received LPV/r died; the certainty of evidence was very low. Three studies reported a possible protective effect of LPV/r as post‐exposure prophylaxis. Again, the certainty of the evidence was very low due to uncertainty due to limited sample size. Conclusions On the basis of the available evidence it is uncertain whether LPV/r and other antiretrovirals improve clinical outcomes or prevent infection among patients at high risk of acquiring COVID‐19.


| INTRODUCTION
Several antiretroviral drugs are being considered for use in the treatment of COVID-19, the disease caused by a newly identified coronavirus, (SARS-CoV-2). Protease inhibitors have been considered as candidate therapy because they inhibit enzymes that activate envelope glycoproteins as part of the process of viral entry into cells [1]. The use of lopinavir/ritonavir (LPV/r) has been supported by data from in vitro studies, animal models and positive clinical outcomes when LPV/r was given to patients infected with severe acute respiratory syndrome (SARS) and Middle East Respiratory Syndrome (MERS) diseases also caused by coronaviruses [2][3][4][5]. Other antiretrovirals have been proposed based on virtual screening and in vitro studies, and several clinical trials are planned. Lopinavir/ritonavir (LPV/r) is included in rapid guidance issued by researchers from Wuhan University based on clinical use during prior epidemics of severe acute respiratory syndrome (SARS) and MERS coronavirus (CoV) infections [6].
This systematic review summarizes the clinical outcomes of using antiretroviral drugs for the prevention and treatment of coronaviruses and planned clinical trials.
Three databases -Medline via PubMed, EMBASE and the Cochrane Librarywere screened from inception to 30 March 2020 for studies reporting clinical outcomes of patients with SARS, MERS or COVID-19 treated with antiretrovirals; studies using antiretrovirals for the prevention of these infections were also sought. The WHO database of publications on COVID-19 was also searched https://www. who.int/emergencies/diseases/novel-coronavirus-2019/globalresearch-on-novel-coronavirus-2019-ncov.
Any study design that reported clinical outcome data was included, and there were no language restrictions. Clinicaitrials.gov and Chictr.org.cn were searched for ongoing and completed trials. Data are summarized per study, but not pooled in meta-analysis due to the limited number of studies reporting outcomes for each disease. The review was conducted by a single reviewer (NF), with data extraction validated by a second reviewer (AR). The quality (or certainty) of the evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach [12].

| Antiretroviral drugs for treatment
From an initial screen of 433 titles, two randomized controlled trials and 21 observational studies provided clinical outcome data on the use of antiretroviral drugs for treatment, and 3 studies reported outcomes for prevention. Three studies were excluded: one because cause of infection was unclear [13], one because the original study was retracted during the conduct of this systematic review [14] and one because lamivudine was given to control chronic hepatitis B infection and its use could not be linked to SARS outcomes [15]. Among the included studies, the majority reported outcomes using LPV/r as treatment; two two studies reported outcomes among HIVpositive individuals who were on a combination antiretroviral drugs for management of HIV [16,17].
Characteristics of included studies and patient outcomes are summarized in Table 1.

| SARS
Two observational studies and one case report among patients with SARS [2,17,18] reported outcomes of patients who were given antiretrovirals. A study from China reported a reduction in mortality in patients receiving LPV/r of 2.3% (95% CI 0% to 6.8%) compared to matched controls (15.6%, 9.8% to 22.8%) [2]. A second study from China reported that none of the 41 patients given LPV/r died compared with seven of 111 patients in the control group [18]. The third study, also from China, was a case report of a 30-year-old HIV-positive man who recovered; he was receiving abacavir, efavirenz, tenofovir and LPV/r as antiretroviral therapy [17]. All patients also received ribavirin and steroids of varying dose and duration.

| MERS
Six observational studies, including two retrospective observational studies [3,21] and four case reports [16,19,20,22]one was from Greece, one from Austria, two from Saudi Arabia and two from the Republic of Koreaprovided data on patients diagnosed with MERS. There were 42 deaths among 165 patients who were given LPV/r together with other interventions including ribavirin and pegylated interferon.

| COVID-19
One randomized, controlled open-label study reported on the efficacy and safety of LPV/r for treating hospitalized adults with severe COVID-19 [23]. In this trial 99 patients received LPV/r (400/100 mg twice a day; median time between symptom onset and randomization 13 days) and 100 patients received standard care for 14 days. LPV/r was not associated with a statistically significant difference in time to clinical improvement; 28 day mortality was numerically lower in the LPV/r group (14/99) compared to the control group (25/100), but this difference was not statistically significant in the intention-to-treat analysis. Accelerated clinical recovery and reduced mortality were observed in those treated within 12 days of symptom onset, but not in those treated later. Almost half of patients in the LPV/r group (46 patients, 48.4%) and control group (49 patients, 46.7%) reported one or more adverse events: gastrointestinal-related complaints including nausea, vomiting and diarrhoea were more common in the lopinavir/ritonavir group. A second randomized trial assessed patients admitted to hospital with mild/moderate COVID-19, and compared outcomes of 21 patients given LPV/r (200mg/50mg twice a day) with 16 patients given ardibol and 7 patients who were not given any antiviral therapy [24]. In this trial, LPV/r did not show any benefit in terms of time to viral clearance (PCR negativity) or progression to severe disease. For both trials, certainty of the evidence was low due to risk of bias (investigators not blinded to the intervention, and imprecision.
In the observational studies, three case reports [25][26][27]. three case series [28][29][30], and six observational studies [31][32][33][34][35][36] reported outcomes of patients with COVID-19 who received LPV/r; nine studies were from China, one was from Singapore and two from the Republic of Korea. Among the 361 patients in the nine studies where outcomes could be associated with receipt of LPV/r, three patients died. One study reported that 53 of 56 patients received LPV/r and three patients died; however, it was unclear how many of the patients who died had received LPV/r [31].
LPV/r is recommended by WHO as part of second-line antiretroviral therapy [37]. Among people living with HIV receiving LPV/r diarrhoea, nausea and vomiting are commonly reported side effects at start of treatment [22]. These side effects were reported by four out of five individuals who received LPV/r for the treatment of COVID-19 in Singapore, and only one individual completed the 14-day treatment course as a result of adverse events [29].
The certainty of the evidence for outcomes across these three diseases is very low. The sample size was small and only two studies provided comparative outcomes (one using historical controls) and none used a randomized design to be able to assess the comparative effectiveness of different interventions. Timing, duration and dose of treatment varied, and in the majority of studies patients were provided with other

| Antiretroviral drugs as post-exposure prophylaxis
Three studies reported a possible protective effect of LPV/r against coronavirus infection [38][39][40]. The first, a retrospective observational study from China, noted that 0 out of 19 patients hospitalized on same floor as SARS patients contracted the disease. Of the 19 patients, 11 were on differing regimens of antiretroviral therapy; none received LPV/r [38]. The second study, from South Korea, retrospectively enrolled health care workers considered at high risk of MERS infection. Of 22 healthcare workers given post-exposure prophylaxis (PEP) comprising ribavirin and LPV/r, none were infected; this compared to 9 of 21 healthcare workers not given PEP who became infected [39]. The third study, from China, compared characteristics of 8 HIV-positive individuals on different antiretroviral regimens who had contracted COVID-19 infection with 1166 patients who had not been infected [40]. No statistically significant relationship was found between type of antiretroviral regimen and infection status. The certainty of the evidence across outcomes was again very low due to uncertainty due to limited sample size, lack of uniformity of regimens being used to treat patients, and lack of information regarding intensity of exposure (Appendix S2).

| Registered clinical trials
Of 85 titles screened, 25 registered trials were identified that plan to assess the safety and efficacy of antiretrovirals -20 assessing LPV/r (including 1 for the treatment of MERS and one for SARS, the rest for COVID-19), two ritonavir, two darunavir and cobicistat and one tenofovir alafenamide fumarate. Estimated completion dates are from March 2020 to January 2022 (Appendix S3).

| CONCLUSIONS
This systematic review identified two randomized trials and 21 observational studies provided clinical outcome data on the use of LPV/r for the treatment of COVID-19, SARS and MERS. The randomized trials showed no clinical benefit, the observational studies were inconclusive, and the certainty of the body of evidence across all important outcomes was low or very low. Based on available evidence it is uncertain whether LPV/r and other antiretrovirals improve clinical outcomes in severe symptomatic disease or prevent infection among patients at high risk of acquiring COVID-19. Any differences in potential therapeutic effect of LPV/r between SARS, MERS and COVID-19 may partly be due to different clinical presentations; many of the patients had complicated courses including stays in intensive care units and were on multiple concurrent, unproven treatments. Several randomized trials are planned to assess the safety and efficacy of antiretroviral drugs, including LPV/r, for the treatment of COVID-19, MERS-CoV and SARS-CoV. While the conduct of such trials is challenging [41], high quality evidence is needed to improve clinical and programmatic decisions to use antiretroviral drugs for current and future coronavirus outbreaks.
The procurement and use of LPV/r or other antiretroviral drugs to treat or prevent COVID-19 infection should take into consideration the need to ensure continued availability for people living with HIV who need LPV/r as part of their antiretroviral therapy. Overuse of LPV/r for COVID-19 in the current epidemic runs a risk of shortage of a drug that is currently used as a second line treatment for people living with HIV.
WHO plans to update this review at least monthly throughout 2020, and longer as needed, to update the evidence as new studies are completed.