Gilead is surg­ing again off da­ta that sug­gest its an­tivi­ral remde­sivir might im­prove sur­vival.

The new da­ta come from an analy­sis Gilead con­duct­ed com­par­ing the death rate and re­cov­ery time of pa­tients in one of its remde­sivir tri­als to a group of 800 pa­tients “with sim­i­lar base­line char­ac­ter­is­tics and dis­ease sever­i­ty” who re­ceived on­ly stan­dard-of-care around the same time. The re­sult, they said, sug­gest­ed that pa­tients who re­ceived remde­sivir had a 62% bet­ter chance at sur­viv­ing than those who did not.

Ret­ro­spec­tive looks are more sus­pect than place­bo-con­trolled tri­als, so the da­ta are far from con­clu­sive. But it does add new ev­i­dence to the ques­tion of whether remde­sivir can im­prove mor­tal­i­ty. The NIH tri­al that led the FDA to au­tho­rize remde­sivir for use in Covid-19 pa­tients on­ly showed that remde­sivir could help pa­tients re­cov­er faster. Few­er pa­tients on remde­sivir died in that tri­al than on stan­dard-of-care — 8% vs 11% — but the re­sults were not sig­nif­i­cant.

For­mer FDA chief sci­en­tist Lu­ciana Bo­rio said that da­ta from ob­ser­va­tion­al stud­ies are dif­fi­cult to in­ter­pret, but that the bulk of ev­i­dence points to remde­sivir help­ing pa­tients sur­vive Covid-19. “It is very like­ly that Remde­sivir does in­deed af­ford pa­tients a mor­tal­i­ty ben­e­fit,” she said in an email. “The gold stan­dard RCT con­duct­ed by the NIH demon­strat­ing the clin­i­cal ben­e­fit of Remde­sivir in im­prov­ing time to re­cov­er wasn’t pow­ered to de­tect a dif­fer­ence in mor­tal­i­ty. It doesn’t mean that there wasn’t one”

In a tweet, for­mer FDA chief Scott Got­tlieb called the re­sults “en­cour­ag­ing” but urged that it need­ed to be con­firmed. The com­pa­ny ac­knowl­edged in their press re­lease that a prospec­tive study was need­ed.

This is very en­cour­ag­ing but needs to be con­firmed in a prospec­tive tri­al. It ap­pears to be a ret­ro­spec­tive analy­sis of the phase III da­ta us­ing his­tor­i­cal matched con­trols, sug­gest­ing a sur­vival ben­e­fit in se­vere Covid pa­tients. https://t.co/84dRM­sIn­TA

— Scott Got­tlieb, MD (@ScottGot­tlieb­MD) Ju­ly 10, 2020

Gilead was up 2% pre-mar­ket on the news, from $74.71 to $76.18.

Oth­er an­a­lysts and ex­perts, though, were more skep­ti­cal and in some cas­es en­tire­ly dis­mis­sive. Pe­ter Bach, the di­rec­tor for the cen­ter for health pol­i­cy and out­comes at Memo­r­i­al Sloan Ket­ter­ing Can­cer Cen­ter, said on Twit­ter that com­par­ing tri­al par­tic­i­pants to re­al-world pa­tients on­ly works when doc­tors have a good un­der­stand­ing of the dis­ease, some­thing that isn’t the case with Covid-19.

Get­ting ?’s on this. Pan­dem­ic has taught us how wrong in­fer­ences can be when drawn from com­par­isons of un­like co­horts as here. The first prin­ci­ples test for these analy­ses is that nat­ur­al his­to­ry and its pre­dic­tors both well un­der­stood. Not true in covid. https://t.co/J90gpX­HZ­KA

— Pe­ter B. Bach, MD (@pe­ter­bach­md) Ju­ly 10, 2020

Baird an­a­lyst Bri­an Sko­r­ney com­pared the study to work from Di­di­er Raoult, the French physi­cian who pub­lished in Feb­ru­ary one of the first tri­als on hy­drox­y­chloro­quine in Covid-19. That study, de­spite gin­ning up sig­nif­i­cant in­ter­est in the drug, was seen as rife with is­sues, in­clud­ing in “los­ing” 6 of the tri­al’s 26 pa­tients “in fol­low-up,” 4 of whom died.

Who did this analy­sis? Di­di­er Raoult? https://t.co/cmdzH5vg3R

— Bri­an Sko­r­ney (@Bri­an­Sko­r­ney) Ju­ly 10, 2020

In the study, 312 pa­tients from Gilead’s open-la­bel SIM­PLE tri­al were com­pared to 818 re­al-world pa­tients who had sim­i­lar char­ac­ter­is­tics. Pa­tients who re­ceived remde­sivir had a 7.6% chance of dy­ing, com­pared to 12.6% for those who re­ceived stan­dard-of-care. That trans­lates to 62% im­proved odds of sur­vival — or slight­ly more than dou­ble the 30% im­proved odds the NIH tri­al’s raw num­bers sug­gest­ed.

Sun­Trust an­a­lyst Robyn Kar­nauskas said the sur­vival rates in each tri­al were “com­pa­ra­ble,” while ac­knowl­edg­ing the study was “not an ap­ples-to-ap­ples com­par­i­son.”

The remde­sivir pa­tients in the analy­sis al­so had a 74.4% chance of re­cov­er­ing by day 14, ver­sus 59.9% for stan­dard-of-care.

Not every tri­al, though, has found remde­sivir has a ben­e­fit. On the same day as the NIH read­out, The Lancet pub­lished a study from Chi­na that found remde­sivir im­proved nei­ther mor­tal­i­ty nor oth­er pa­tient met­rics. That study, though, was end­ed ear­ly af­ter the first out­break in Chi­na end­ed.

For hos­pi­tals that can get their hands on it, remde­sivir so far has been one of on­ly two proven treat­ments for Covid-19, along with a gener­ic steroid called dex­am­etha­sone. More an­tivi­rals and neu­tral­iz­ing an­ti­bod­ies, though, are now be­ing stud­ied, hav­ing be­come a grow­ing fo­cus of the Trump ad­min­is­tra­tion. If more ther­a­pies prove ef­fec­tive, the new Gilead da­ta could help physi­cians choose the best op­tion.

“While not as vig­or­ous as a ran­dom­ized con­trolled tri­al, this analy­sis im­por­tant­ly draws from a re­al-world set­ting and serves as an im­por­tant ad­junct to clin­i­cal tri­al da­ta, adding to our col­lec­tive un­der­stand­ing of this virus and re­flect­ing the ex­tra­or­di­nary pace of the on­go­ing pan­dem­ic,” in­ves­ti­ga­tor Su­san Olen­der of Co­lum­bia Uni­ver­si­ty Irv­ing Med­ical Cen­ter said in a state­ment.

So­cial: AP Im­ages

For a look at all End­points News coro­n­avirus sto­ries, check out our spe­cial news chan­nel.

A volunteer in AstraZeneca’s Covid-19 vaccine trial in Brazil has died, Brazilian health authorities said Wednesday, triggering fresh alarms over the future of the Oxford program. But later reports noted that the death was in the placebo group and AstraZeneca issued word that there were no concerns about continuing the study.

The Brazilian health agency Anvisa said it had received data from an investigation into the issue, per a Reuters report. The report was then updated citing a Brazilian newspaper with unnamed sources saying that the volunteer was in the placebo arm of the trial.

Four years after a Phase I/II setback in spinal muscular atrophy (SMA), Novartis is hoping its drug branaplam will find more success in a new neurological indication: Huntington’s disease.

The decision was announced a year after the head of research, Jay Bradner, said he did not see a “big opportunity” in SMA, according to Reuters. Novartis says it has preclinical data showing that branaplam reduces levels of mutant huntingtin protein, and SMA data showing patients on the drug had reductions in huntingtin mRNA. The FDA gave branaplam their orphan drug designation, and Novartis plans to move forth with a Phase IIb trial next year.