The Protease Inhibitor Lopinavir, Boosted with Ritonavir, as Treatment for COVID-19: A Rapid Review

Background The HIV protease inhibitor lopinavir, boosted with ritonavir, has been used off-label to treat COVID-19. We aimed to synthesize the clinical evidence for lopinavir/ ritonavir as a treatment for COVID-19. Methods We performed a rapid review by searching databases including PubMed, GoogleScholar, medRxiv, ClinicalTrials.gov and the Cochrane COVID-19 Study Register, for COVID-19 studies comparing outcomes between patients who did and did not receive lopinavir/ritonavir. The quality of evidence was assessed using the GRADE criteria. Results We identified five completed randomized controlled trials (RCTs) and 14 retrospective cohort studies. Two large RCTs of 5,040 and 2,771 hospitalized adults with COVID-19 found no evidence that lopinavir/ritonavir influenced the primary outcome of mortality, or secondary outcomes including progression to mechanical ventilation or time to discharge. Results remained similar in all subgroup analyses including by age, gender, baseline ventilation and time since symptom onset. The three smaller RCTs (n=86–199) also found no evidence of a benefit in the primary outcomes of time to clinical improvement or time to viral clearance. The 14 observational studies included between 50 and 415 participants, and were limited by a lack of adjustment for potential confounding variables. The majority of these studies found no evidence that lopinavir/ritonavir was associated with improved mortality or other clinical outcomes, although results regarding viral clearance were mixed. Conclusions Good evidence from large clinical trials does not support using lopinavir/ritonavir to treat COVID-19 amongst hospitalized patients.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes novel coronavirus disease 2019 (COVID- 19), is known to have infected over 51 million people globally at the time of writing, with over 1.2 million reported deaths [1]. Due to limited SARS-CoV-2 testing, the true numbers of cases are likely far higher. Currently, there are few effective treatments for COVID-19 that have been rigorously tested in randomized controlled trials. The rapid spread of the pandemic and urgent need for effective treatments has led to interest in re-purposing currently available drugs for immediate use [2]. The antiretroviral drug lopinavir is a protease inhibitor, which is widely used for the treatment of HIV and is a potential candidate for treatment of COVID-19. Lopinavir is formulated in combination with another protease inhibitor, ritonavir (lopinavir/ ritonavir, branded as Kaletra or Aluvia). Ritonavir inhibits the metabolizing enzyme cytochrome P450 3A and thereby increases the half-life of lopinavir [3].
There is preliminary clinical evidence of effectiveness of lopinavir/ritonavir against other coronaviruses such as severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV). During the SARS-CoV outbreak in 2003, an open-label, non-randomized study found reduced risk of severe hypoxia or death in 41 SARS-CoV patients who were treated with lopinavir/ ritonavir and ribavirin, compared with 111 historical controls treated with ribavirin alone [4]. However, there has been no evidence from randomized clinical trials to demonstrate the efficacy of lopinavir/ritonavir in treating SARS-CoV or MERS-CoV.

Introduction
Potential mechanism of action SARS-CoV-2 is a single-stranded RNA-enveloped betacoronavirus, similar to SARS-CoV and MERS-CoV. These viruses enter host cells and replicate, producing strands of the viral RNA which are translated by host cell ribosomes into polypeptides [2,5]. The enzyme 3-chymotrypsin-like protease (3CL pro ) cleaves these polypeptides into non-structural proteins. These then replicate and transcribe the viral genome, prior to assembly and release of new progeny viruses [5][6][7]. As lopinavir is a protease inhibitor, it may inhibit the action of 3CL pro , which would in turn disrupt the process of viral replication and release from host cells [6,7]. Lopinavir has antiviral activity against SARS-CoV [4] and MERS-CoV [8], and has in vitro activity against SARS-CoV-2 with a half-maximal effective concentration (EC 50 ) at 26.63 µM [9]. However, this indicates that SARS-CoV-2 is less susceptible to lopinavir than HIV, as the EC 50 for lopinavir against HIV is between 0.01-0.03 µM [3]. One reason for the reduced potency may be that coronavirus proteases like 3CL pro do not contain a C2-symmetric pocket, which is the target of HIV protease inhibitors [8,10]. Darunavir, another HIV protease inhibitor, was not active against SARS-CoV-2 in an unpublished in vitro study [11]. A recent study using in vitro and mouse models found stronger evidence for anti-MERS-CoV activity for another antiviral, the RNA-dependent RNA polymerase inhibitor remdesivir, compared with lopinavir/ritonavir [8].

Dosing, safety, side effect profile and drug interactions
The recommended dose of lopinavir/ritonavir for HIV treatment is 400 mg/100 mg twice daily [3]. For COVID-19, the same dose has commonly been used, generally for 14 days [2]. Lopinavir/ritonavir is contraindicated in porphyria and caution is advised in patients with haemophilia, cardiac conduction disorders, pancreatitis, structural heart disease and those with increased risk of cardiovascular disease [12]. Common side effects include gastrointestinal disturbance, in particular diarrhoea, which is often worse in the first few weeks and can be treated with loperamide. Dyslipidaemia, diabetes mellitus, pancreatitis and hepatic disorders have also been reported, but these complications are usually only experienced after several months of therapy [12,13]. Drug interactions with lopinavir/ritonavir are common due to their inhibition of cytochrome P450, which can lead to increased levels of co-administered drugs that are metabolized by this enzyme. Drugs that interact with lopinavir/ritonavir and are commonly used in primary care include the lipid lowering simvastatin, combined oral hormonal contraceptives, anti-epileptic drugs and the inhaled corticosteroid fluticasone [14].

Global use and price
Lopinavir/ritonavir is currently recommended by the World Health Organization as a second-line treatment of HIV, and is used by approximately 580,000 people globally for this indication [15]. With the current interest in repurposing its use for COVID-19, there have been concerns regarding ensuring adequate supply for people living with HIV. Currently, a 2-week course costs approximately £140 (USD 170) in the UK [12], and approximately £6.50 (USD 8) in certain low-and middle-income countries under pooled patent agreements [15].
Here, we used rapid review methods [16] to synthesise the evidence for the use of lopinavir/ritonavir as a treatment for COVID-19.

Methods
We adhered as closely as possible to the PRISMA guidelines for systematic reviews [17]. We searched for clinical trials or observational studies comparing the efficacy of lopinavir/ritonavir against standard care or other drugs, for the treatment of COVID-19 in humans. We excluded studies which had fewer than 15 participants in either the lopinavir/ritonavir or the comparator group, as these studies were unlikely to have statistical power to detect any difference in treatment outcomes. We did not use any language restrictions and included pre-prints. Systematic reviews were used as a point of reference.

Search strategy and selection criteria
We conducted electronic searches in PubMed, GoogleScholar, the Cochrane COVID-19 Study Register, ClinicalTrials.gov and medRxiv on the 13 October 2020. We excluded articles published before 2019. We used search terms including coronavirus*, COVID-19, SARS-CoV-2, lopinavir and Kaletra. Full search terms are listed in Additional file 1. We searched the reference lists of identified articles to find further relevant articles. After removing duplicates, one author conducted title, abstract and full text screening, with any uncertainties checked by a second author. We used the Population, Intervention, Comparator, Outcomes, Study design (PICOS) framework to extract relevant data from selected articles using a standardized form (Table 1) [18]. We assessed the quality of the evidence using the GRADE criteria [19].

Results
We identified a total of 879 records, including 386 through PubMed, 85 through medRxiv, 264 through the Cochrane COVID-19 Study Register, 87 through ClinicalTrials.gov and 57 through GoogleScholar. From these, 175 duplicates were identified, and a further 631 records were excluded through title and abstract screening. 54 studies were further excluded in the full-text screening, including 45 studies that did not compare clinical outcomes between patients receiving lopinavir/ritonavir and those not receiving lopinavir/ ritonavir, and 9 studies with sample sizes too small to allow meaningful comparisons. We finally included 19 studies, including 5 completed clinical trials and 14 observational studies ( Table 2). Two of the clinical trials provided good quality evidence, one provided moderate quality evidence, and the other included studies were assessed as being at high risk of bias, mainly due to small sample sizes and the lack of adjustment for potential confounders. We present our findings in narrative format as heterogeneity of outcomes and study designs made quantitative synthesis inappropriate.

Clinical trial findings
The RECOVERY trial is an open-label, randomized platform trial of patients hospitalized with clinically suspected or laboratory confirmed SARS-CoV-2 infection in 176 UK hospitals [20]. As part of this study, 5,040 patients were randomized to receive either usual care plus lopinavir/ritonavir 400 mg/100 mg twice daily for 10 days (n=1,616), or usual care alone (n=3,424). Median days from symptom onset to enrolment was similar in both arms (lopinavir/ritonavir 8 days [IQR [5][6][7][8][9][10][11][12], usual care 8 days [ ). There was no difference between the two groups in the primary outcome of mortality by 28 days (rate ratio 1.03, 95% CI 0.91, 1.17; P=0.60). Analyses of pre-specified sub-groups by age, sex, ethnicity, time since symptom onset, respiratory support at randomization and baseline predicted risk also found no evidence of benefit. Secondary outcomes of hospital discharge (risk ratio [RR] 0.98, 95% CI 0.91, 1.05; P=0.53) and a composite of mechanical ventilation or death by 28 days (RR 1.09, 95% CI 0.99, 1.20; P=0.092), were similar between groups. There was one serious adverse event of elevated serum alanine aminotransferase in the lopinavir/ritonavir group, although the authors note that they did not record non-serious adverse reactions.
The SOLIDARITY trial, led by the World Health Organization, is a large, international, randomized platform trial of adults admitted with COVID-19 to 405 hospitals in 30 countries [21]. 1,399 participants randomized to receive lopinavir/ritonavir 400 mg/200 mg twice daily for 14 days, were compared with 1,372 participants who were randomized to receive usual care. The primary outcome of in-hospital mortality was similar in both groups (148/1,399 versus 146/1,372, RR 1.00, 95% CI 0.79, 1.25; P=0.97), and there were no differences in sub-group analyses stratified by age or ventilation status at randomization. Proportions achieving the secondary outcome of progression to requiring ventilation were similar in the lopinavir/ritonavir group (9.6%) and the control group (9.5%). The proportion remaining in hospital at 7 and 14 days (68% versus 59%, 31% versus 22%, respectively) was higher in the lopinavir/ritonavir arm, but the authors hypothesize that this was due to patients being asked to remain in hospital to complete the lopinavir/ritonavir treatment course, and this resolved by 21 days (12% versus 11%). 95% CIs and P-values were not presented for secondary outcomes. Serious adverse events were not reported, but there were no deaths from hepatic failure in the lopinavir/ritonavir arm.
Cao et al. [22] conducted an open-label randomized controlled trial at a single hospital in Wuhan, China at the peak of the epidemic there. They enrolled 199 hospitalized adults with a positive SARS-CoV-2 reverse transcriptase polymerase chain reaction (RT-PCR) test, pneumonia and oxygen saturations ≤94% on ambient air, and randomized them to receive lopinavir/ritonavir 400 mg/100 mg twice a day for 14 days (n=99) or standard care (n=100). Median time from onset of illness to randomization was 13 days (interquartile range [IQR] [11][12][13][14][15][16]. Baseline characteristics were similar between the two groups. The median age was 58 years (IQR 49-68). Most patients (55%) were enrolled >12 days after symptom onset, and were severely unwell, requiring urgent clinical attention. After 28 days, intention to treat (ITT) analysis revealed no difference in the primary outcome of time to clinical improvement between the two arms (16 days in both groups; hazard ratio 1.31; 95% CI: 0.95, 1.85; P=0.09). Restricting the analysis to patients enrolled within 12 days of symptom onset did not alter results. When modified ITT analysis was conducted, in which three patients who died within 24 h of randomization and did not receive lopinavir/ritonavir were excluded, there was weak evidence of a small improvement in the time to clinical improvement with lopinavir/ritonavir (median of 15  proportion of patients with clinical improvement at 28 days, time from randomization to death, nor duration of oxygen therapy or mechanical ventilation. There was also no difference in viral clearance between groups. Gastrointestinal symptoms were more common in the lopinavir/ritonavir arm, and 13.8% of patients stopped treatment early due to adverse events. Overall, serious adverse events were higher in the usual care arm (32 versus 19 events), largely due to a higher frequency of acute respiratory distress syndrome (27 versus 12 events).
Li et al. [23] report findings from a single-blind randomised controlled trial in China. The investigators initially aimed to enrol 125 adults with laboratory confirmed SARS-CoV-2, but due to control of the epidemic, the trial was limited to 86 participants. Patients with mild or moderate clinical status (with or without signs of pneumonia) were suitable for inclusion. The mean age of was 49.4 years (range 19-79). 34 participants were randomized to receive lopinavir/ritonavir for 14 days, 35 to receive umifenovir (a fusion inhibitor [24]) and 17 to standard care with no antiviral. There was no difference in the primary outcome of mean time to negative pharyngeal SARS-CoV-2 PCR test between the lopinavir/ritonavir, umifenovir and control groups Huang et al. [25] conducted a single-site, openlabel, randomized controlled trial amongst 101 adults hospitalized with laboratory confirmed SARS-CoV-2 infection and mild to moderate COVID-19, defined as respiratory rate <30 breaths per min, oxygen saturations >93% and FiO2 >39.9 kPa. Participants were randomized to receive ribavirin plus interferon-a (n=33), or lopinavir/ritonavir plus interferon-a (n=36), or lopinavir/ritonavir plus ribavirin plus interferon-a (n=32). There was no evidence of a difference in the primary outcome of time to SARS-CoV-2 negativity between the groups (13 versus 12 versus 15 days respectively; P=0.23). Secondary outcomes of proportions who were SARS-CoV-2 negative by day 14 (51.5% versus 61.1% versus 46.9% respectively; P-value not given), and progression to severe disease (3.0% versus 5.6% versus 6.3%; P=0.58) were similar between the three groups.
Most studies found no association between lopinavir/ritonavir treatment and COVID-19 outcomes. Lopinavir/ritonavir use was not associated with a change in mortality outcomes in five studies [26,27,[30][31][32], or with preventing clinical deterioration in three studies [26,27,32], nor with clinical improvement in three studies [32,33,35]. One Chinese study found that amongst patients receiving lopinavir/ritonavir (n=28) a higher proportion developed unfavourable outcomes, including death or disease progression (53.6% versus 21.4%; P<0.001), compared with those not on lopinavir/ritonavir (n=295) [29]. However, patients with critical disease severity at baseline were more likely to receive lopinavir/ritonavir compared with those with non-severe and severe disease (P<0.001), meaning that the worse outcomes amongst those receiving lopinavir/ritonavir could be explained by this bias. Another Chinese study found an association between receiving lopinavir/ritonavir and faster resolution of lung involvement on computed tomography (CT) scan, but patients receiving lopinavir/ritonavir were also more likely to have influenza coinfection, and the scheduling of CT scans was not clear [34]. Regarding clearance of SARS-CoV-2 in RT-PCR testing, results were more mixed. Of the 11 studies, 6 found no association between lopinavir/ritonavir use and SARS-CoV-2 clearance [27,28,[31][32][33]36], while lopinavir/ritonavir was associated with better SARS-CoV-2 clearance in 3 studies [34,35,38] and worse clearance in 2 studies [37,39]. The two studies with shorter time since symptom onset did not find an association between lopinavir/ritonavir use and shorter time to viral clearance [28,35]. Three studies reported increased side effects (gastrointestinal disturbances) [32] and adverse events (liver and renal function derangement) [26,32,35] among patients receiving lopinavir/ritonavir.

Discussion
We identified five randomized clinical trials, including 8,197 hospitalized patients, which assessed lopinavir/ ritonavir as a treatment for COVID-19. Two of these studies provided good quality evidence due to their large sample size and objective primary outcome [20,21], although neither were placebo controlled and results from one of these trials had not yet been peer-reviewed. The remaining three trials had a higher risk of bias; none were blinded and two were under-powered. No randomized trial found a benefit from lopinavir/ritonavir with regard to their primary outcomes which included mortality, time to clinical improvement and negative pharyngeal SARS-CoV-2 PCR test. There was no clear benefit in secondary outcomes, nor in subanalyses broken down by patients who received treatment earlier in the course of the disease. Gastrointestinal side effects were more common in patients treated with lopinavir/ritonavir compared with controls. We also reviewed 14 observational studies including 2,549 patients, of whom 1,003 received lopinavir/ritonavir. There was no association between use of lopinavir/ritonavir and reduced mortality nor clinical improvement. Results regarding viral shedding were mixed, and in general were unadjusted for potential confounders such as disease severity or concomitant treatments.

Comparison with existing literature
We did not find any other published systematic reviews reporting results of recent, large trials of lopinavir/ritonavir for COVID-19. Our rapid review, containing five RCTs and 14 observational studies, therefore provides an important update.

Strengths and limitations
We used a broad search strategy, and rapid, pragmatic approach that allowed identification of emerging evidence in the evolving COVID-19 pandemic. We adhered as closely as possible to the PRISMA checklist for systematic reviews. We specifically searched for and included pre-prints, because much of the current evidence is being made available in this form in the interest of data sharing during this pandemic. We did not apply language restrictions in order to capture COVID-19-related research being conducted in China. Due to the rapid nature of the review, we did not attempt to contact authors, meaning we were unable to clarify or confirm data where we had queries.

Limitations of the current evidence
We did not find any randomized trials of lopinavir/ ritonavir use among patients with COVID-19 in the community, where earlier treatment may be more effective. However, amongst hospitalized patients, there was no evidence of a benefit in sub-group analyses among those treated with lopinavir/ritonavir earlier in the course of the disease. Our review included a large number of observational studies which, by virtue of their design, are more prone to bias. None of the included studies adjusted their analyses for concomitant medication use or COVID-19 severity, which are likely to have affected outcomes. Lack of adjustment for important co-variates (potential confounders) can increase risk of bias. We also excluded many cohort studies from this analysis as they did not specifically describe the number of patients treated with lopinavir/ritonavir for COVID-19, or did not describe a comparator group. This represents a missed opportunity to increase the evidence base to identify treatment strategies for COVID-19.

Implications for research and clinical practice
Ongoing platform randomized trials should continue to assess alternative treatments for COVID-19 in order to determine safety and efficacy [40][41][42]. Ideally, these would be double blinded, but the urgency and constraints of conducting research in pandemic settings means this has not always been possible [42][43][44]. These studies are mainly focused on hospital inpatients, and trials of treatments in primary care, where patients are likely to present earlier, should also be prioritized. Developing studies in community settings would allow assessment of whether earlier treatment with repurposed drugs such as lopinavir/ritonavir could prevent the development of serious complications of COVID-19 [45]. Treatments that are hypothesized to have an antiviral effect may be more likely to work earlier during the replicative phase of SARS-CoV-2 infection, rather than later during the inflammatory stage of COVID-19.
In conclusion, we found evidence that lopinavir/ritonavir should not be used to treat COVID-19 amongst hospitalized patients. Further trials of communitybased lopinavir/ritonavir treatment may be warranted.

Disclosure statement
JD and OG are funded by the Wellcome Trust PhD Programme for Primary Care Clinicians (216421/Z/19/Z and 203921/Z/16/Z, respectively). For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. TC is funded by the British Heart Foundation (BHF) PhD Studentship (FS/19/13/34235). GH is funded by an NIHR Advanced Fellowship. The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. The authors declare no competing interests.

Additional file
Additional file 1: A list of search terms can be found at https://www.intmedpress.com/uploads/documents/ AVT-20-OA-4692_Dorward_Addfile_1.pdf