key: cord-0816716-8siopd2k
authors: Chouchana, L.; Boujaafar, S.; Gana, I.; Preta, L.-H.; Regard, L.; Legendre, P.; Azoulay, C.; Canouï, E.; Zerbit, J.; Carlier, N.; Terrier, B.; Kerneis, S.; Batista, R.; Treluyer, J.-M.; Zheng, Y.; Benaboud, S.
title: Plasma concentration and safety of lopinavir/ritonavir in patients with Covid-19: a retrospective cohort study
date: 2020-05-22
journal: nan
DOI: 10.1101/2020.05.18.20105650
sha: 9a17bdcd3c3e268ce2a27a487b2c4be1ac40118b
doc_id: 816716
cord_uid: 8siopd2k

Background: There is an urgent need of active treatment for coronavirus disease 2019 (Covid-19). Although efficacy have not been proven, lopinavir/ritonavir 400 mg/100 mg twice daily has been proposed as a treatment of moderate to severe Covid-19. Previously published cohorts showed Covid-19 is associated with major inflammation. To date, no data are available regarding lopinavir/ritonavir plasma concentration and its safety in Covid-19 patients. Methods: Real-world Covid-19 experience based on a retrospective cohort study Results: On the cohort of 31 patients treated by lopinavir/ritonavir for Covid-19, we observed very high lopinavir plasma concentrations, increased of 4.6-fold (IQR 2.9-6.4), with regards to average plasma concentrations in HIV treatment. All except two patients were above the upper limit of the concentration ranges of HIV treatment. In this cohort, about one over four to five patients prematurely stopped lopinavir/ritonavir therapy due to a moderate adverse drug reaction, mainly hepatic and gastrointestinal disorders. Conclusion: Patients with Covid-19 pneumonitis treated with lopinavir/ritonavir have plasma concentrations dramatically higher than expected. Owing to that high plasma concentration may be required for antiviral activity against SARS-CoV-2, it appears that lopinavir dosage should not be reduced in the absence of adverse effect. About 80% of the patients well tolerated lopinavir/ritonavir therapy under these plasma concentrations. However, cautious is necessary as drug repurposing can be associated with a new drug safety profile.

Since early December 2019, a pandemic infectious disease due to a novel coronavirus, called 61 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is spreading all over the world. 62

To date, no specific therapeutic agent has proven its clinical efficacy against this outbreak. 63

However, due to the urgent need for a treatment, several antiviral drugs are being repurposed and 64 used off-label to treat Covid-19, including lopinavir/ritonavir (LPV/r) (1). The latter is a human 65 immunodeficiency virus (HIV) protease inhibitor approved since years. LPV is prescribed in 66 patients with HIV in association with another protease inhibitor, ritonavir (RTV), used as a potent 67 P450 3A4 cytochrome (CYP) inhibitor in order to dramatically increase LPV plasma exposure. 68 LPV/r has been proposed in previous coronavirus outbreaks in 2003 and 2012 due to SARS-69

CoV-1 and to Middle East respiratory syndrome coronavirus (MERS-CoV), respectively (2, 3). 70

In vitro activity and data from rodent models showed LPV has an antiviral activity against 71 MERS-CoV(4, 5). It has also proven its potency to inhibit SARS-CoV-2 replication in vitro (6). 72

Furthermore, in an open-label study comparing the addition of LPV/r to ribavirin in patients with 73 SARS-CoV, the authors showed a reduced risk of acute respiratory distress syndrome (ARDS) or 74 death by comparison with a historical control group treated by ribavirin alone (2). Viral load was 75 also reduced in the LPV/r treated patients. 76

Based on these results, this drug has been considered to be potentially useful in patients with 77 6 first symptoms. Furthermore, no therapeutic drug monitoring was performed for LPV/r to assess 83 an ideal drug exposure. Nevertheless, due to the absence of specific treatment for Covid-19, 84 LPV/r is still a therapeutic option used for patients with moderate to severe Covid-19 pneumonia. 85

To date, more than 40 clinical trials assessing the efficacy of LPV/r against Covid-19 are ongoing 86 around the world (8). 87

In order to assess LPV/r plasma concentration and its safety in Covid-19 patients, we conducted a 88 retrospective cohort study as part of routine care in adult patients hospitalized for a Covid-19 89 infection and requiring oxygen. 90 91 92

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) patients did not receive any antibiotic at admission or before being hospitalized. LPV/r was 103 started within the first 48 hours after hospitalization and median time to starting LPV/r after 104 symptom onset was 8 days (IQR, 7-10 days). No patient had drug-drug interaction involving 105 RTV and 13 (42%) patients did not received concomitant therapy. LPV/r treatment duration was 106 7 (IDR, 3-8) days. At the end of the LPV/r course, five patients had recovered, eight patients 107 were transferred to intensive care unit or died, 17 were still hospitalized with oxygen 108 dependency. Clinical outcome is unknown for one patient that has been transferred to another 109 hospital during treatment. 110 respectively. In all patients, considering either trough or peak for LPV and RTV, plasma 117 concentrations were above the average concentrations observed during HIV treatment (4,000 118 ng/mL and 9,000 ng/mL for trough and peak concentrations, respectively) ( Figure 1A ). All 119 except two patients had plasma concentration over the upper limit of the concentration ranges 120 observed in HIV. With regards with these plasma concentration ranges, LPV concentrations in 121

Covid-19 patients were increased of about 4.6-fold (IQR 2.9-6.4) ( Figure 1B) . Extent of LPV 122 plasma concentration increase was not associated with CRP or IL-6 levels (data not shown). 123 124 Safety 125 LPV/r was discontinued before the end of scheduled course for 14 (45%) patients (Table 2) . 126

Reasons for early ending LPV/r therapy were the occurrence of adverse drug reaction, therapeutic 127 limitation or patient deceased and poor efficacy in seven (22%), four (13%) and three (10%) 128 patients, respectively. Suspected adverse drug reaction were assessed as possibly related to LPV/r 129 therapy. They consisted in four cases of liver injuries (three cases moderate cytolytic hepatitis 130 between three and six times above the upper limit of normal range and one case of isolated 131 hyperbilirubinemia), two cases of gastrointestinal disorders (nausea/vomiting and diarrhea) and 132 one case of psychiatric disorders (agitation/anxiety). These adverse effects were mild and all 133 patients recovered after drug withdrawn. In patients with cytolytic hepatitis, LPV plasma 134 concentrations were 6.1-fold above average plasma concentrations in HIV patients. 135 136 137 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 22, 2020. . https://doi.org/10.1101/2020.05.18.20105650 doi: medRxiv preprint plasma exposure with regards to the therapeutic levels in HIV. Although LPV/r regimen was 140 400/100 mg bid as in HIV treatment, LPV plasma concentrations were in median about 4.6-fold 141 higher, and up to 8-fold higher in some patients. All patients of this retrospective cohort had a 142 LPV plasma concentration over the average concentrations observed in HIV patients. As LPV is 143 extensively metabolized by the hepatic cytochrome P450 system, almost exclusively by the 144 CYP3A isozyme, LPV/r therapy was not started in the event of drug-drug interaction. None was 145 detected during therapy and no patient had liver cirrhosis, which could have impaired LPV/r 146 clearance. Only two patients had a medical history of hepatitis, including one patient with an 147 active hepatitis B without hepatic insufficiency. 148

Some clinical evidence show that patients with moderate to severe Covid-19 have a major 149 inflammation. Elevated levels of blood IL-6 are commonly seen in severe Covid-19 illness and 150 could led to a cytokine storm syndrome (9). Previous studies in macaques inoculated with SARS-151

CoV found that defects in T-cell and B-cell function and the excess production of type 2 152 cytokines could lead to a deficiency in control of viral replication and prolonged proinflammatory 153 responses (10). In Covid-19 patients, a persistent blood virus load and an exacerbated 154 inflammatory response has been shown to be partially driven by the transcriptional factor NFκB 155 (11) . These results are in line with our cohort where patients had very high CRP and IL-6 levels 156 at starting therapy. Besides, it is long-standing known that inflammatory responses and infections 157 decrease drug metabolism capacity in human and experimental animals (12). In cultured human 158 hepatocytes, direct treatments with tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), 159 interleukin-6 (IL-6), interferon-γ (IFN-γ) and transforming growth factor-β (TGF-β) can reduce 160 the expression of CYP1A2, CYP2C8 and CYP3A4 (13, 14). In animals and humans, several 161 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 22, 2020. . https://doi.org/10.1101/2020.05.18.20105650 doi: medRxiv preprint studies or reports have shown the role of inflammation in reducing drug metabolism and CYP450 162 inhibition (15, 16). CYP3A4 expression is transcriptionally regulated by pregnane X receptor 163 (PXR), which is a ligand-dependent transcription factor (17). NF-κB activation by tumor necrosis 164 factor-α, in a setting of inflammatory response, has been shown to suppress PXR activation and 165 PXR-mediated gene expression (17, 18) . This results in a reduced CYP450 expression and a 166 decreased drug metabolism, especially CYP450-mediated metabolism. Other genes regulated by 167 PXR, including ABCB1, coding for P-glycoprotein transporter, could also have their expression 168 decreased, contributing to decrease non-renal drug clearance, such as LPV/r. Owing to its low 169 oral bioavailability and extensive metabolism by CYP3A4 isoenzyme, LPV needs to be co-170 administered with low doses of RTV, a potent CYP3A inhibitor, to achieve drug concentrations 171 high enough to inhibit viral replication (19). An added reduced expression of CYP3A4 could 172 dramatically increase LPV plasma concentrations in Covid-19 patients. 173

Interestingly, nevertheless very high plasma concentrations in all assayed patients, LPV/r was 174 overall well tolerated. The main adverse effects were moderate hepatobiliary disorders that have 175 been attributed to LPV/r therapy. In HIV, moderate-to-severe elevations in serum 176 aminotransferase levels (>5 times the upper limit of normal) are found in 3% to 10% of patients, 177 although rates may be higher in patients with HIV-HCV coinfection (20). These elevations are 178 usually asymptomatic and self-limited and can resolve even with continuation of the medication. 179

Drug causality could be hardly assessable as about 20% of the patients with Covid-19 have been 180 reported to have increased transaminases (9). However, in our cohort, transaminases quickly 181 decreased after stopping lopinavir, making drug causality probable. Overall, adverse effects 182 reported in this Covid-19 cohort are in line with LPV/r safety profile (21). Adverse effects led to 183 stop LPV/r treatment in seven (22%) patients, which could appear elevated, even though, due to 184 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 22, 2020. . https://doi.org/10.1101/2020.05.18.20105650 doi: medRxiv preprint considering whether a dose is appropriate for SARS-CoV-2, it is critical to consider free drug 210 concentrations. For standard HIV-1, Cmin after 75 ng/mL that is substantially below EC 50 211 reported for coronaviruses (27). Therefore, targeting SARS-CoV-2 antiviral activity in patients is 212

challenging. It appears that LPV/r dosage should not be reduced on the basis of therapeutic drug 213 monitoring in the absence of adverse drug reaction. 214

Our study has several limitations. First, due to the retrospective design using data from routine 215 care, plasma assays were drawn at different times that prevent from an accurate pharmacokinetic 216 estimation. Second, according to different practices in wards, Covid-19 severity at baseline was 217 heterogeneous between patients. However, most importantly, in all except three due to drawing 218 issue, LPV plasma concentrations regardless assays timing were unexpectedly high comparing 219 HIV experience. Finally, unbound concentrations, which are the active ones, were not estimated. 220

In conclusion, in our single center Covid-19 experience, we found that LPV/r treated patients had 222 unexpected very high plasma concentrations, well above those observed in HIV patients. 223

However, LPV dosage should not be reduced in the absence of adverse effect, owing that to high 224 plasma concentration may be required for an in vivo antiviral activity as suggested by previous 225 studies. We observed that about one over four to five patients stopped LPV/r therapy in relation 226 with moderate adverse drug reactions attributed to LPV therapy. Cautious is needed in this 227 context of drug repurposing, which can be associated with a new drug safety profile. These early 228 data need to be confirmed with prospective studies. 229 230 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 22, 2020. The calibration range was from 160 (limit of quantification, LOQ) to 40,000 ng/mL for LPV and 268 10 (LOQ) to 2,500 ng/mL for ritonavir. The inter-and intra-assay bias and CV were inferior to 269 15% for LPV and ritonavir. No significant matrix effect on quantification of LPV and ritonavir 270 was found in this method. LPV plasma concentrations were considered as peak level (Cmax) and 271 as through level (Cmin) for assays drawn between one hour to five hours, and from eight hours 272 after tablet intake, respectively. In HIV treatment, observed Cmin and Cmax ranges are 1,000-273 8,000 ng/mL (target > 4,000 ng/mL for non-naïve patients) and 7,000-11,000 ng/mL, respectively 274 (31, 32). 275 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 22, 2020. . https://doi.org/10.1101/2020.05.18.20105650 doi: medRxiv preprint LPV/r treatment safety 277 LPV/r treatment safety was assessed in a routine care setting. Cases of suspected adverse drug 278 reaction were spontaneously notified to the Pharmacovigilance regional center. After case-by-279 case review assessment and review by a senior pharmacologist, cases were reported to the French . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 22, 2020. . https://doi.org/10.1101/2020.05.18.20105650 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 22, 2020. Lopinavir/ritonavir therapy characteristics Time from symptom onset to starting lopinavir/ritonavirdays 8 (7-10) Type of specific anti-COVID-19 drug associated with lopinavir/ritonavir -None -Cephalosporin or penicillin* -Cephalosporin* and macrolide* -Macrolide* -Corticosteroids -Corticosteroids and antibiotics -Sarilumab (anti-IL-6) 13 (42%) 6 (19%) 3 (10%) 3 (10%) 2 (6%) 2 (6%) 2 (6%) Duration of lopinavir/ritonavir therapydays 7 (3-8) Lopinavir plasma concentration-ng/mL -at trough -at peak 20,153 (IQR, 16,633-26,505) 22,550 (IQR 12,720-33,612) Lopinavir plasma concentrationfold increase** 4.6 (IQR 2.9-6.4) Reasons for therapy termination -Scheduled end of treatment -Adverse drug reaction -Therapeutic limitation or deceased -Poor efficacy 17 (52%) 7 (22%) 4 (13%) 3 (10%) Types of adverse drug reaction -Cytolytic hepatitis -Isolated hyperbilirubinemia -Nausea and vomiting -Diarrhea -Agitation/anxiety 

Impaired type I interferon activity and exacerbated inflammatory responses in severe 354

Impact of infectious and inflammatory disease on cytochrome P450-356 mediated drug metabolism and pharmacokinetics

Gene-specific effects of inflammatory cytokines on 358

cytochrome P450 2C, 2B6 and 3A4 mRNA levels in human hepatocytes

IL-6) and an anti-IL-6 monoclonal antibody on drug-metabolizing enzymes in human 362 hepatocyte culture

Effects of Escherichia coli 364 lipopolysaccharide on telithromycin pharmacokinetics in rats: inhibition of metabolism via 365 CYP3A

Role of NF-kappaB in regulation of PXR-mediated gene expression: a mechanism for the 373 suppression of cytochrome P-450 3A4 by proinflammatory agents

Possible involvement of the 376 competition for the transcriptional coactivator glucocorticoid receptor-interacting protein 1 377 in the inflammatory signal-dependent suppression of PXR-mediated CYP3A induction 378 in vitro

ABT-378, a highly potent inhibitor 383 of the human immunodeficiency virus protease

National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda (MD)

Dosing will be a key success 412 factor in repurposing antivirals for COVID-19

Small 415 molecules targeting severe acute respiratory syndrome human coronavirus

Radiological 418 findings from 81 patients with COVID-19 pneumonia in Wuhan

ARDS Definition Task Force

Acute respiratory distress syndrome

Ageyear 63 (51-78) Sexmale:female sex 22:9 Comorbidities Hypertension Diabetes Cardiovascular diseases (others) Malignancy or immunosuppression Chronic respiratory disease (including asthma) Hepatitis or liver cirrhosis (Child-Pugh B or more) Rheumatic disease Chronic kidney failure None