Remdesivir for COVID-19: Why Not Dose Higher?

LETTER

A recent article by Xu et al. (1) examined the on- and off-target toxicity of remdesivir (RDV) and its parent nucleoside, GS-441524. Notably, primary human hepatocytes (PHHs) and HepG2 cells (liver cancer cell line) were exceptionally sensitive to RDV treatment. Still, the authors conclude, “In clinical settings of COVID-19 [coronavirus disease 2019] treatment in hospitalized patients, the risk associated with possible RDV-related liver enzyme elevations is substantially lower compared to its established benefits in hospitalized COVID-19 patients.” Whereas we commend the authors’ rigorous study, we deplore the omission of citations of key clinical studies demonstrating no clear benefit with RDV (2, 3). The clinical efficacy of RDV is contentious; major clinical trials conducted with RDV yield mixed results (Table 1). The claim of “established benefits in hospitalized COVID-19 patients” rests on favorable results of a single double-blind, randomized controlled trial (RCT) (4) and at the omission of unfavorable interim results from the WHO Solidarity trial (3) and a double-blinded RCT by Wang et al. (2). By undermining these trial results (5, 6), Gilead implicitly acknowledges that the clinical benefits of RDV are modest, requiring the most stringent trial design to extract a favorable, statistically significant result (4). Ironically, Gilead dismisses the interim results from Solidarity on the basis of potential heterogeneity in controls and its open-label nature (6) yet touts their self-sponsored, open-label trials (7, 8) lacking a control group (7). Beyond its questionable clinical efficacy, it is unclear whether the current RDV regimen effectively reduces viral loads in patients’ lungs (2, 9). Wang et al. (2) found no difference in viral reduction in the upper respiratory tracts of RDV-treated versus placebo groups. Given RDV’s limited clinical and antiviral efficacy, we ask, why not dose higher?

TABLE 1

Clinical efficacy of RDV, by major clinical trial

Study (reference)	Comparison	n	Double blind?	Controlled?	Primary endpoint	Outcome
Goldman et al. (7)	5- vs 10-day RDV in severe COVID-19	397	No	No	Clinical status at day 14 by 7-point ordinal scale	No statistically significant difference between treatment groups
Wang et al. (2)	RDV vs placebo in severe COVID-19	237	Yes	Yes (placebo)	Time to clinical improvement up to day 28	No statistically significant difference between placebo and RDV groups
Beigel et al. (4)	RDV vs placebo in severe COVID-19	1,062	Yes	Yes (placebo)	Time to recovery, discharge from hospital vs. hospitalization	Shortened median time to recovery in RDV vs placebo groups (10 vs 15 days)
Spinner et al. (8)	5- vs 10-day RDV vs SOCa in moderate COVID-19	596	No	Yes (SOC)	Clinical status at day 11 by 7-point ordinal scale	5-Day RDV had statistically significant higher odds of better clinical status
WHO Solidarity Trial Consortium (3)	One of trial drug regimens (including RDV) vs local SOC in severe COVID-19	2,750	No	Yes (local SOC)	In-hospital mortality	No effect on mortality for patients hospitalized with COVID-19

SOC, standard of care.

In a phase 1 trial with RDV in healthy volunteers, graded transaminase elevations were observed in 25% of participants in the 7-day multiple-ascending-dose (MAD) cohort (150 mg daily, 1,050 mg cumulative dose) and in 75% of participants in the 14-day MAD cohort (150 mg daily, 2,100 mg cumulative dose) (10), which concurs with the unique sensitivity of PHHs to RDV in vitro (1, 11). Comparing the magnitude of hepatotoxicity in healthy participants ties transaminase elevations to total dose exposure (Table 2); hepatotoxicity was not observed in the 225-mg single-dose cohort (10). For reference, the recommended dosage (200-mg loading dose, 100-mg maintenance) results in total doses of 600 mg (5 days) and 1,100 mg (10 days), which fall below the threshold for hepatotoxicity (1,050 to 2,100 mg). If viral suppression is a Cmax (maximum concentration)-driven effect and the degree of hepatotoxicity relates to cumulative exposure, then it may be possible to compress the dosing schedule to enable higher dosing while maintaining the same cumulative dose. For instance, a 300-mg loading dose with 200-mg maintenance for 5 days yields a cumulative dose of 1,100 mg. Although we foresaw these shortcomings with RDV some time ago and have advocated for clinical investigation of GS-441524 in regard to safety (12–16), investigating dose modifications with RDV may benefit patients more readily (17), and we urge Gilead to do so.

TABLE 2

Dose-dependent hepatotoxicity of RDV in healthy volunteers

Cohort	n	Dosea (mg i.v.)	Total dose (mg)	Duration (days)	ALT/AST elevationsc
SAD 5	4	150 (solution)	150	1	0/4
SAD 8	4	150 (lyophilized powder)	150	1	0/4
SAD 6b	4	225 (solution)	225	1	0/4
MAD 1	8	150 QD, 7 days	1,050	7	2/8
MAD 2	8	150 QD, 14 days	2,100	14	6/8

Doses of RDV trialed in the SAD and MAD arms of the phase 1 trial for a 50- to 70-kg human (10). Long-term repeated dosing at 150 mg yields transaminase elevations.

No participants experienced transaminase elevations in the 225-mg SAD cohort.

Boldface indicates that transaminase elevations are dose dependent and emergent even in healthy human volunteers.