Gilead investors should by now be used to the group’s antiviral remdesivir generating confusing and contradictory datasets. So it proved again today, as a study in moderately ill Covid-19 patients delivered a supposed win for a five-day course of the drug, but perversely a 10-day regimen did not show a significant benefit.
The extent to which this matters is a separate point, of course, as remdesivir is already available under US emergency use authorisation measures to tackle the coronavirus pandemic. But the borderline efficacy suggested by the latest dataset has clearly come up short of expectations: Gilead, which had sat on a 19% share price gain year to date, opened off 3%.
The moderate Covid-19 results came a month after Gilead’s trial in severely ill subjects suggested that a five-day course of remdesivir might be just as helpful as 10-day treatment. The big caveat about this study was that it had no control cohort (The winds change again for Gilead and remdesivir, April 29, 2020).
Also in the mix is a clearly negative remdesivir study from China – this had a robust, controlled design but had been terminated early – anecdotal data from emergency access, and a positive trial run by the NIAID. The last is by far the biggest remdesivir dataset, but is perhaps limited by its adaptive design.
|Cross-trial comparisons of remdesivir's four datastes|
|Gilead moderate trial||Gilead severe trial||China trial||NIAID trial|
|Enrolment||584 (target 1,600)||397 (target 6,000)||Halted at 237 (target 453)*||1,063|
|Covid-19 severity||Hospitalised, not on ventilation, ≤4 days since PCR confirmation of disease||Hospitalised, severe, ≤4 days since PCR confirmation of disease**||Hospitalised, confirmed lung involvement, ≤12 days since illness onset||Hospitalised, ≤72 hours (some exceptions) since PCR confirmation of disease|
|Design||Open-label, 5-day or 10-day course, vs SoC||Open-label, uncontrolled, 2-cohort (5-day/10-day)||Quadruple-blinded, placebo-controlled||Double-blinded, placebo-controlled|
|Primary endpoint||Odds ratio for improvement at day 11***||Odds ratio for improvement at day 14^||Time to clinical improvement at day 28||Time to recovery^^|
|Result||31% improvement for 10-day (not stat sig); 65% improvement for 5-day (p=0.017)||54-65% had ≥2-point improvement||21 days vs 23 days (HR=1.23, not stat sig)||11 days vs 15 days (p<0.001)|
|Mortality result||1% for 10-day, 0% for 5-day, 2% for SoC||8-11%||14% vs 13% (not stat sig)||8% vs 12% (not stat sig)|
|*Terminated early because, China's Covid-19 epidemic having been brought under control, no further eligible patients could be enrolled; **cohort of mechanically ventilated subjects was added in Apr; ***changed from hospital discharge at day 14; ^changed from normalisation of fever and oxygen saturation at day 14; ^^changed from disease severity improvement at day 15.|
What do the latest findings, from Gilead’s study in moderately ill patients, add to the understanding? Probably that remdesivir has some activity, but that overall it is not a game changer.
The trial compared remdesivir dosed over five consecutive days, and over 10 consecutive days, against standard of care, and measured efficacy at day 11 as primary endpoint. The efficacy measure was change of at least one point on a seven-point scale where 1 signifies death and 7 represents discharge from hospital.
Gilead said across the course of the study the five-day cohort had a 65% increase in chances of day 11 clinical improvement versus control; this yielded a p value of 0.017, but Gilead did not say what statistical bar it had set, or whether this had been met. The 10-day group showed a 31% increase, and this did not hit statistical significance.
Apart from the apparent paradox of such a finding, it is the absolute numbers at day 11 that lay bare remdesivir’s borderline efficacy. 70% of 10-day patients saw at least a one-point improvement, versus 76% for those on a five-day course and a still respectable 66% for standard of care.
Clearly no in-depth analysis is possible until the data are published in a peer-reviewed publication. Possible confounding factors include baseline imbalances between the three cohorts, and the fact that only 584 subjects had been enrolled into the study at this point.
The moderate trial actually has an enrolment target of 1,600 patients. Similarly, Gilead’s severe trial yielded data after recruiting only 397 out of a target 6,000 Covid-19 subjects.
The US FDA’s May 1 decision to grant remdesivir emergency use authorisation related to both the five and 10-day courses of the drug, but only for patients hospitalised with severe Covid-19. EUAs are not the equivalent of an FDA approval, and can be withdrawn quickly; perhaps this is fuelling this morning’s Gilead selloff.
That said, at this stage of the coronavirus pandemic a drug with even borderline efficacy might be enough, and remdesivir likely still meets this low bar.