Reasons for Optimism and Patience—A COVID-19 Update

A data leak in Gilead Science’s clinical trial of the antiviral remdesivir as a therapy for the virus causing COVID-19 is generating some excitement. Our view is the data are directionally encouraging, but without further information, that is as far as we can go for now.

As background, there are currently three major trials for this disease: China trials, NIH trials and Gilead’s trials. There are a few others, but they are less important at this point.

The China trials were proper, placebo-controlled blinded trials. Unfortunately, they were terminated/suspended last week. We do not know at this point if that was due to a lack of efficacy. The given explanation is China is doing well with disease containment and ran out of both severe and mild-moderate patients to recruit. There may be some truth to this, but the likelier explanation is there is major competition for patients due to the sheer number of trials being conducted. There has been some suggestion the data from these terminated trials would be released, which would certainly be helpful.

The US National Institute of Allergy and Infectious Diseases (US NIAID—part of the NIH) is running a multi-year, comprehensive adaptive trial that will test multiple drugs. The trial’s readout will be remdesivir versus a placebo in a blinded study. The study will measure clinical improvement on an ordinal scale, ranging from death to discharge. We expect this trial to readout sometime in May, and this will likely be the best quality information the FDA will use should it grant approval.

The Gilead trials have gone through a bit of an evolution. After it was clear the drug worked in a petri dish, Gilead gained compassionate-use approval from the FDA to initiate trials. Gilead began recruiting a few hundred patients, both moderate and severe, in two separate trials. Initially, these were both open-label studies with no control arms. However, the protocol was amended to include a comparator arm (standard of care) and added a couple thousand patients.

Over the April 11-12 weekend, a data release showed promising efficacy with remdesivir use in a “compassionate trial”—Gilead’s exploratory trial in a small cohort of 53 patients, which is unrelated to the two large trials just described (i.e., it did not have a comparator arm). These were hospitalized patients with clinical needs ranging from breathing ambient air to noninvasive oxygen supply to invasive intubation to blood oxygenation. Of these, 65% could be classified as critical—57% received intubation and 8% needed extracorporeal membrane oxygenation (ECMO, or blood oxygenation). At the trial’s end, 68% of patients showed improvement (greater than or equal to 2 points on the ordinal scale described previously), 58% of intubated patients were extubated, and 47% of all participants were discharged.

If clinical improvement is measured by the 2-point improvement or discharge rate, 84% of all patients saw benefit. In addition, extubation is a good sign that the lungs are recovering, and the patient will survive. All patients were followed for up to 28 days. The key takeaway was approximately 25% of invasive intubation patients were discharged. In addition, the mortality rate was 18% in patients who were intubated and only 5.3% in those who just received non-invasive oxygen. The 18% number is very encouraging because some reported mortality rates are multiples of this.

The leak causing excitement is from Gilead’s trial of 113 severe patients. The source described rapid suppression of fever in patients. Some patients came off the ventilator in a day, and most were discharged—some in just 6 days out of the designed treatment course of 10 days. The mortality rate was the most astonishing aspect—just two deaths for approximately 1.5%. According to the source, the data is locked, which means it goes straight to analysis and should not take much time if done in an unbiased way.

However, I’m approaching these data with caution. To start, this particular trial still lacks a proper control, as the comparator arm was added later. Also, in comparing this latest information to the compassionate trial, it’s important to note the discharge and mortality rates. The discharge rate from the leaked data suggested most patients have been discharged. This is very encouraging but far from definitive. What does “most” mean? Is it 51% or 95%? And how does this compare to the 47% reported in the compassionate trial? Also, how many of these patients were intubated invasively at baseline? On clinicaltrials.org, it clearly states the inclusion criteria is oxygen saturation less than 94% (which means, they need respiratory help and are classified as severe) but not mechanically intubated. So, these patients were in bad shape but did not require intubation and blood oxygenation at the start of the trial. Therefore, it is not that much of a surprise that discharge rates were better and mortality was much lower—it appears the patients were relatively less severe at baseline.

Compelling additional data should include more patients, stratification, additional clinical endpoints, comparator arms or placebo controls. As of yet, we don’t have that data, but I believe there is reason for cautious optimism—and some patience. If a therapy can keep patients from eventually needing an invasive ventilator, this will definitely help bend the curve. Ultimately, we need a solution to prevent prolonged critical care. This will reduce public anxiety and will help open our economy quicker.