key: cord-1003117-xpgzhns8
authors: Marzolini, C.; Stader, F.; Stoeckle, M.; Franzeck, F.; Egli, A.; Bassetti, S.; Hollinger, A.; Osthoff, M.; Weisser, M.; Gebhard, E. C.; Baettig, V.; Geenen, J.; Khanna, N.; Tschudin-Sutter, S.; Mueller, D.; Hirsch, H.; Battegay, M.; Sendi, P.
title: Effect of Systemic Inflammatory Response to SARS-CoV-2 on Lopinavir and Hydroxychloroquine Plasma Concentrations
date: 2020-07-07
journal: nan
DOI: 10.1101/2020.07.05.20146878
sha: 5e804ddf175d3263453f55ba419cf06809dc0857
doc_id: 1003117
cord_uid: xpgzhns8

Background: Coronavirus disease 2019 (COVID-19) leads to inflammatory cytokine release, which can downregulate the expression of metabolizing enzymes. This cascade affects drug concentrations in the plasma. We investigated the association between lopinavir (LPV) and hydroxychloroquine (HCQ) plasma concentrations and the values of acute phase inflammation marker C-reactive protein (CRP). Methods: LPV plasma concentrations were prospectively collected in 92 patients hospitalized at our institution. Lopinavir/ritonavir was administered 12 hourly, 800/200 mg on day 1, and 400/100 mg on day 2 until day 5 or 7. HCQ was given at 800 mg, followed by 400 mg after 6, 24 and 48 hours. Hematological, liver, kidney, and inflammation laboratory values were analyzed on the day of drug level determination. Results: The median age of study participants was 59 (range 24 up to 85) years, and 71% were male. The median duration from symptom onset to hospitalization and treatment initiation was 7 days (IQR 4;10) and 8 days (IQR 5;10), respectively. The median LPV trough concentration on day 3 of treatment was 26.5 ug/mL (IQR 18.9;31.5). LPV plasma concentrations positively correlated with CRP values (r=0.37, p<0.001), and were significantly lower when tocilizumab was preadministrated. No correlation was found between HCQ concentrations and CRP values. Conclusions: High LPV plasma concentrations were observed in COVID-19 patients. The ratio of calculated unbound drug fraction to published SARS-CoV2 EC50 values indicated insufficient LPV concentrations in the lung. CRP values significantly correlated with LPV but not HCQ plasma concentrations, implying inhibition of cytochrome P450 3A4 (CYP3A4) metabolism by inflammation.

The release of inflammatory cytokines such as interleukin-6 (IL-6) activates intracellular 80 signaling cascades, leading to the downregulation of cytochrome P450 enzymes (CYPs) 81 (7). The decrease in expression and activity of CYPs is explained by transcriptional 82 suppression of CYP mRNA, triggering a decrease in enzyme synthesis (5, 6). Systemic 83 inflammation affects CYPs differently with a more pronounced decrease in CYP3A4 84 expression followed by CYP2B6, CYP2C19, CYP2C9, CYP2D6 and CYP1A2 (5, 6, 8) . 85 Correlations have been reported between elevated C-reactive protein (CRP) values and 86 high plasma levels of antipsychotic drugs (9) and voriconazole (10). CRP production is 87 triggered by IL-6, and conversely, IL-6 suppression can be monitored with plasma CRP 88 levels (11). 89 The HIV drug lopinavir/ritonavir (LPV/r) has been repurposed for the treatment of SARS- CoV-2 (2). Recent brief reports of 8 (12), 12 (13), and 21 (14) COVID-19 patients, noted 91 considerably higher LPV plasma concentrations than those observed in HIV patients (15) . 92 Considering the inhibition of drug metabolism by cytokine release, and the administration 93 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 5 of LPV/r (metabolized by CYP3A4), we had the rationale to prospectively monitor LPV 94 plasma concentrations in our cohort of COVID-19 patients. 95 The objective of this study was to investigate the association between CRP values and LPV 96 plasma concentrations in COVID-19 patients. With this approach, we aimed to underscore 97 the hypothesis that high inflammatory markers in the blood correlate with high LVP 98 plasma concentrations. For comparison, we measured hydroxychloroquine (HCQ) 99 concentrations, because it is characterized by a different metabolism (16). We also 100 performed an age-stratified analysis to explore the combined effect of aging and 101 inflammation on drug plasma levels. (which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 6 were not administered, except in 3 individuals in whom a low-dose long-term treatment 119 with prednisone was continued (5 mg/d in 2 patients, 10 mg/d in 1 patient). Other drugs 120 affecting inflammation were not administered, with the exception of tocilizumab (TCZ).

Treatment concepts for COVID-19 and dosing rationale: Our institutional treatment 122 recommendations include the administration of LPV/r and HCQ for hospitalized patients.

To achieve rapidly high LPV/r plasma concentrations, we administered a double dose in 124 the first 24 hours. This approach in the early treatment phase was presumed necessary 125 to suppress the high SARS-CoV-2 viral load in the early stage of disease ("hit early and hit 126 hard"). The LVP/r treatment schedule included 800/200 mg twice daily on day 1, followed 127 by a maintenance dose of 400/100 mg every 12 hours for another 4 to 6 days. LVP/r 128 treatment was combined with HCQ for 2 days (i.e., 800 mg loading dose followed by 400 129 mg at 6 hours, 24 hours, and 48 hours). In patients with clinical signs and findings 130 suggestive for COVID-19-induced hyper-inflammation, the use of TCZ was considered at (which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 7 for liquid chromatography mass spectrometry methods (Recipe Chemicals + Instruments, 143 Munich, Germany). The lower limit of quantification was 0.1 μg/mL.

HCQ levels were measured from available plasma material obtained for LPV trough 145 determination. HCQ was quantified with a validated liquid chromatography mass 146 spectrometry method developed by the laboratory of clinical chemistry at the University 147 Hospital in Zurich, Switzerland. The lower limit of quantification was 10 ng/mL. Because age-related physiological changes can affect drug pharmacokinetics (18), we 156 categorized patients as <65 years or > 65 years. As indicated earlier, we used a tentative 157 CRP cutoff value of 75 mg/L to aid decision making for the administration of TCZ. This 158 CRP level was used as marker for inflammation for the analysis in the study (i.e., <75 vs > 159 75 mg/L).

In patients receiving TCZ prior to the measurement of LPV or HCQ plasma concentrations, (which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 8 measurement of LPV trough levels, the interval between the two time points was 167 considered to be too short for having an effect on LPV plasma concentrations.

Absolute numbers, percentages, medians, and interquartile ranges (IQRs) were used to 169 report demographic characteristics and laboratory results. The Mann-Whitney U test was (which was not certified by peer review) 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 July 7, 2020. Table 1 .

Combined effect of age and inflammation on LPV concentrations: Median LPV plasma 208 trough levels were insignificantly higher in patients who were >65 years (26.9 μg/m, n = 209 33) than in those who were <65 years (24.5 μg/mL, n = 59) ( Figure 2A) . Accordingly, (which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. Inflammation has been shown to have the greatest impact on CYP3A4 expression (7). This 262 increase may, in turn, impact the magnitude of DDIs, because LPV/r inhibits CYP3A4 in a 263 concentration-dependent manner (26). Co-administered CYP3A4 substrates can -per se 264 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 12 -be also affected by inflammation, and can further increase the magnitude of DDIs. This 265 interaction is illustrated by a case series of 12 patients who were followed up for direct 266 oral anticoagulant treatment (DOAC) before and after being infected with SARS-CoV-2. (which was not certified by peer review) 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 July 7, 2020. 

HCQ has been used historically for malaria and immune diseases. Its ability to inhibit 299 SARS-CoV-2 is thought to be due to an increase in endosomal pH, thereby impairing the 300 entry of the virus into the cell. HCQ also interferes with the glycosylation of cellular 301 receptors for SARS-CoV-2, resulting in reduced virus-cell binding. Finally, HCQ has 302 immunomodulatory activity that may suppress the cytokine storm (16).

The median HCQ concentrations observed in our study (i.e., 171 ng/mL, IQR 128-207) is 304 comparable to those reported in another study with COVID-19 patients (220 + 110 305 ng/mL) (38), and to steady-state trough levels observed in patients with lupus 306 erythematosus (i.e., 103-130 ng/mL)(39). Thus, the HCQ plasma concentrations in 307 COVID-19 patients, in contrast to reported LPV plasma concentrations, were not higher 308 than those previously observed in studies with other indications. Furthermore, no 309 correlation was observed with CRP values. This difference may possibly be explained by 310 the different metabolic pathways of HCQ and LPV/r, as inflammation affects CYPs 311 differently (7). Furthermore, HCQ is known to have higher concentrations in tissue than 312 in plasma (approximately 200-to 700-fold higher), resulting in a large distribution 313 volume and a long half-life (33). Therefore, HCQ plasma concentrations from COVID-19 314 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint 14 patients might not be suitable to reflect the effect of inflammation given that HCQ does 315 not achieve steady-state concentrations during the short treatment course. Similar to 316 LPV/r, HCQ was shown to have a low RLTEC (i.e., 0.11-0.34), indicating that HCQ levels 317 achieved in vivo do not result in adequate clinical activity against SARS-CoV-2 (33). These Acknowledgements 338 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. 

Age, years, median (range)

Variables are median and interquartile range (IQR) unless stated otherwise

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. (78) All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. The medial Lopinavir plasma concentration was 26.5 (IQR 18.9-31.5) μg/mL. The median 520 Hydroxychloroquine plasma concentration was 171 (IQR, 128-207) ng/mL. The dashed 521 line represents the historical lopinavir trough level observed in HIV-infected individuals 522 treated with lopinavir/ritonavir 400/100 mg twice daily (i.e., 7.1 μg/mL) (15).

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. (which was not certified by peer review) 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 July 7, 2020. The left bar includes LPV plasma levels from COVID-19 patients with no TCZ 539 administration (n = 57) or TCZ administration < 12 hours prior to LPV measurement (n = 540 19), (median 28.8 µg/mL). The right bar represents LPV samples from COVID-19 patients 541 with TCZ administration > 12 hours prior to LPV measurement (median 18.7 µg/mL).

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) 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 July 7, 2020. . https://doi.org/10.1101/2020.07.05.20146878 doi: medRxiv preprint