A lot has happened since the last post on Lopinavir/Ritonavir (it was only March 4, but it seems like a year ago). Here is a quick reminder of where we left off:
- Lopinavir/ritonavir (tradename KALETRA) is a combination of protease inhibitors used in HIV, which might also be effective in COVID-19.
- Prior studies on SARS and MERS have suggested benefit from lopinavir/ritonavir… but most patients in these studies were also given ribavirin. So, the trials were really testing the triple-cocktail of lopinavir/ritonavir/ribavirin.
- Young et al. reported a small series of patients with COVID-19 treated with lopinavir/ritonavir. This didn’t seem to work – but it was unclear whether that might be due to lack of ribavirin, the low dose of lopinavir (they used half of the typical dose), or whether the drug is simply ineffective.
So that brings us up to the current trial…
Cao et al: A trial of lopinavir-ritonavir in adults hospitalized with severe COVID-19
This is an open-label RCT of the efficacy of lopinavir/ritonavir among hospitalized patients with COVID-19, pneumonia, and hypoxemia (defined as <94% saturation on room air, or PaO2/FiO2 < 300). 199 patients were included, with good matching between groups:
Patients treated with lopinavir/ritonavir received 400 mg/100 mg BID for two weeks (the standard dose, unlike the Young et al. study which used half of this lopinavir dose). Lopinavir/ritonavir was stopped early in 13 patients (14%) due to adverse events, mostly gastrointestinal intolerance. On average, therapy was started 13 days after illness onset – potentially too late to affect the clinical disease course:
The primary endpoint was time to clinical improvement, which was defined as time from randomization either to an improvement of two points on a seven-category ordinal scale or to discharge from the hospital (whichever came first). The ordinal scale was as follows:
- Not hospitalized, resumption of normal activities
- Not hospitalized, unable to resume normal activities
- Hospitalized, not on oxygen
- Hospitalized, on supplemental oxygen
- Hospitalized, on either high-flow nasal cannula or noninvasive ventilation
- Hospitalized, on invasive mechanical ventilation and/or ECMO
There wasn't any significant effect on this primary endpoint:
Likewise, other clinical endpoints were generally neutral (or very close to neutral):
In a post-hoc, subgroup analysis of patients treated within <12 days, there was a barely significant difference in time to clinical improvement. However, this remains thoroughly unimpressive (the Kaplan Meier plots largely overlap, as shown below)
A modified intention-to-treat analysis was performed, excluding three patients in the lopinavir/ritonavir group who died before they were able to receive treatment (figure below). It doesn't appear that patients who died in the control group were similarly removed, rendering the validity of this analysis dubious. A one-day improvement was found in patients receiving lopinavir-ritonavir, with borderline statistical significance (hazard ratio 1.39, confidence interval 1-1.91). As a secondary analysis lacking statistical robustness, this finding shouldn't hold much weight.
So, the clinical endpoints were really quite neutral. There are many potential explanations for this. Perhaps treatment was initiated too late to affect the clinical disease trajectory. Perhaps the clinical endpoints were too rough to capture small differences in clinical course. Perhaps lack of blinding somehow interfered with patient assessments. In this context, viral load may be particularly informative. Lopinavir/ritonavir functions solely as antiviral drug, so if it's doing anything beneficial then we would expect to see an accelerated drop in viral titers. In smaller studies with fuzzy clinical endpoints, viral load might be the most sensitive and objective sign that a drug is doing something.
And there was no difference in the viral load. Indeed, after two weeks of lopinavir/ritonavir therapy, there was actually a trend towards a higher viral load among patients in the lopinavir/ritonavir group. Virology data was also analyzed in various patient subgroups:
This analysis also showed no differences. Lack of any effect in the viral load or positivity may be the most damning evidence in this paper, which suggests that lopinavir/ritonavir simply doesn't work.
kudos to the investigators
This study started recruiting a week after initial sequencing of the SARS-CoV-2 genome, which is remarkable (most studies in the United States couldn't pass through an IRB review that rapidly). The study provides actionable evidence, which may help us adjust our treatments for these critically ill patients (omission of a potentially toxic treatment is extremely helpful). So, although the results are a bit disappointing, the authors deserve enormous credit for executing the study.
- Lopinavir/Ritonavir doesn’t appear to work for COVID-19. There are some limitations in this study, but lack of any impact on viral titers strongly suggests a lack of efficacy. Further study is reasonable, but lopinavir/ritonavir shouldn't be used outside of an RCT.
- Lopinavir/Ritonavir/Ribavirin might still work (as this synergistic combination of drugs previously showed some efficacy for SARS). This merits evaluation in further trials. Empiric use of the triple cocktail outside of RCTs is a bit dubious at this point, since this involves exposing the patient to three agents with a considerable potential for side-effects.
- From the clinician’s perspective, lopinavir/ritonavir does seem to be down and out. At least for now. As always, we will continue to follow the COVID-19 situation and maintain an open mind.
- Full manuscript of Cao et al. at NEJM.org.
- Preview of lopinavir/ritonavir efficacy in COVID-19? PulmCrit 3/4 (spoiler alert: the preview was correct)
- IBCC chapter section on lopinavir/ritonavir.
Image credit: Cashino NDT on UpSplash