FDA ad­vi­sors back Lil­ly’s Alzheimer’s drug, set­ting stage for ap­proval af­ter de­lay

An FDA ad­vi­so­ry com­mit­tee gave its strong back­ing to Eli Lil­ly’s Alzheimer’s drug do­nanemab, giv­ing fur­ther sup­port for the agency’s ex­pect­ed ap­proval af­ter a sur­prise de­lay ear­li­er this year.

The com­mit­tee of out­side ex­perts vot­ed 11-0 that do­nanemab’s ben­e­fits out­weigh its risks. It al­so vot­ed 11-0 that the da­ta pre­sent­ed by In­di­anapo­lis-based Lil­ly showed the drug was ef­fec­tive.

While the FDA doesn’t have to fol­low the com­mit­tee’s rec­om­men­da­tions, it of­ten does, and a re­jec­tion of the treat­ment would be a sur­prise de­spite a de­ci­sion by the agency in March to de­lay an ap­proval de­ci­sion. At that time, Lil­ly said the FDA had asked for more da­ta on safe­ty as well as its plans to let pa­tients stop treat­ment once a key bio­mark­er fell sig­nif­i­cant­ly.

Dur­ing Mon­day’s de­bate on the drug, many of the pan­el’s ex­perts not­ed that while do­nanemab came with risks, they thought those could be man­aged and should not stop the agency from mak­ing it avail­able to pa­tients.

“It seems like there are some very clear pa­ra­me­ters we can put in place,” said Cyn­thia Carls­son, an ad­comm mem­ber and pro­fes­sor of med­i­cine at the Uni­ver­si­ty of Wis­con­sin. “We all know that our pa­tients have dif­fer­ent val­ues they at­tach to dif­fer­ent ben­e­fits”

If ap­proved by the agency, po­ten­tial­ly in the com­ing weeks, Lil­ly’s drug would go head-to-head with Ei­sai and Bio­gen’s Leqem­bi. While hailed as the first drug to con­clu­sive­ly show it can slow the symp­toms of Alzheimer’s, Leqem­bi’s sales have got­ten off to a slow start.

Safe­ty con­sid­er­a­tions

Dur­ing its pre­sen­ta­tion on Mon­day, the FDA out­lined sev­er­al lin­ger­ing ques­tions about do­nanemab’s safe­ty. But it ap­pears most like­ly to take the path of em­pha­siz­ing the drug’s risks in a la­bel rather than throw­ing up a mean­ing­ful road­block.

“Safe­ty is a sig­nif­i­cant con­cern for these ther­a­pies,” Tere­sa Bu­rac­chio, di­rec­tor of the of­fice of neu­ro­science at the FDA’s Cen­ter for Drug Eval­u­a­tion and Re­search, said in open­ing re­marks.

A ma­jor fo­cus is a symp­tom called ARIA, which can be a sign of brain bleeds or swelling. In its pre­sen­ta­tion, the FDA not­ed that ARIA symp­toms can mim­ic a stroke, which can set the stage for a dan­ger­ous sce­nario.

One 70-year-old pa­tient in Lil­ly’s study who de­vel­oped stroke-like symp­toms and was treat­ed with the clot-bust­ing drug tenecteplase died four days lat­er. The case echoed a sim­i­lar anec­dote about death fol­low­ing the use of blood thin­ners in a pa­tient who had a stroke that was tak­ing Leqem­bi, which works in a sim­i­lar way.

Na­tal­ie Brana­gan, a clin­i­cal safe­ty re­view­er at the FDA, said that if do­nanemab is ap­proved, pa­tients should car­ry a med­ical in­for­ma­tion card to in­di­cate that they are be­ing treat­ed with do­nanemab. But she said the agency would be un­like­ly to bar pa­tients from re­ceiv­ing blood thin­ners, and would leave that tough de­ci­sion to care­givers and doc­tors.

“You are be­tween a rock and a hard place,” Brana­gan said. “You are go­ing to have a dif­fi­cult time as that pre­scriber fig­ur­ing out the right thing to do for that pa­tient.”

The com­mit­tee asked the FDA how it would han­dle known risk fac­tors for ARIA, such as the num­ber of mi­cro­he­m­or­rhages re­vealed on a pa­tient’s brain scan be­fore start­ing treat­ment. Brana­gan said that some pa­tients did well on the treat­ment even as they de­vel­oped new mi­cro­he­m­or­rhages dur­ing Lil­ly’s study, so the agency would “have a hard time say­ing this would be an ab­solute con­traindi­ca­tion.”

Two de­lays

The FDA’s de­ci­sion on the drug has al­ready been de­layed twice. Al­though the ad­comm’s vote is non­bind­ing, it’s a piv­otal mo­ment in Lil­ly’s thir­ty-year ef­fort to de­vel­op a drug that slows de­men­tia.

Near the be­gin­ning of the meet­ing, Lil­ly chief med­ical of­fi­cer David Hy­man said that the com­pa­ny be­lieves there is “sig­nif­i­cant align­ment” with the FDA on in­ter­pret­ing do­nanemab’s da­ta.

One ques­tion the FDA will have to de­cide is whether treat­ment with Lil­ly’s drug can stop once its shown an ef­fect. In Lil­ly’s tri­al, pa­tients stopped re­ceiv­ing do­nanemab once amy­loid plaques, thought by some to cause the dis­ease, were large­ly cleared from their brains. Bio­gen and Ei­sai, in con­trast, sug­gest that their amy­loid-bust­ing drug Leqem­bi be giv­en in­def­i­nite­ly.

But Kevin Krudys, a clin­i­cal ef­fi­ca­cy re­view­er at the FDA, said there is “con­sid­er­able un­cer­tain­ty” about how much amy­loid needs to be re­duced for ben­e­fit. Al­though Lil­ly said that amy­loid was slow to reac­cu­mu­late when pa­tients stopped tak­ing do­nanemab, Krudys not­ed that da­ta from a short pe­ri­od of the com­pa­ny’s study, and that Lil­ly did not in­clude an ap­pro­pri­ate com­par­i­son group.

Re­mov­ing amy­loid ear­ly in the course of the dis­ease ap­pears to mod­est­ly slow cog­ni­tive de­cline, but it can al­so cause signs of brain bleed­ing and swelling known as amy­loid-re­lat­ed imag­ing ab­nor­mal­i­ties, or ARIA. Reisa Sper­ling, a neu­rol­o­gist and clin­i­cal in­ves­ti­ga­tor at Mass­a­chu­setts Gen­er­al Hos­pi­tal, said that ARIA seems to be a side ef­fect in­her­ent to the en­tire drug class.

“It’s un­like­ly that we’re go­ing to be able to com­plete­ly avoid ARIA and still achieve the amy­loid re­duc­tion that ap­pears to be nec­es­sary for clin­i­cal ben­e­fit,” Sper­ling said.

Stag­ing pa­tients

Lil­ly al­so com­pared the ef­fi­ca­cy of its drug in pa­tients with slight­ly ear­li­er or more ad­vanced dis­ease, de­ter­mined by how much of the pro­tein tau had built up in their brain. It was more ef­fec­tive in pa­tients with less tau and less ad­vanced dis­ease. One of Lil­ly’s great­est con­cerns is that the drug would be re­strict­ed to pa­tients with low-to-medi­um tau, and the com­pa­ny ar­gued that do­nanemab ben­e­fit­ed both groups.

Bu­rac­chio, the FDA’s neu­ro­science di­rec­tor, said that the agency doesn’t ex­pect the dif­fer­ences in the tau sub­groups would change its de­ci­sion to ap­prove or re­ject the drug, but it could fac­tor in­to the word­ing on the drug’s la­bel­ing.

Rather than fo­cus­ing on the low­er ef­fi­ca­cy seen in pa­tients with high tau lev­els, some ad­comm mem­bers ex­pressed con­cern about ex­trap­o­lat­ing to pa­tients with no-to-low tau, a group that was ex­clud­ed from Lil­ly’ study, even though the com­pa­ny and some sci­en­tists ex­pect these pa­tients may re­ceive even greater ben­e­fit be­cause they are in an even ear­li­er stages of dis­ease.

“It could be that it’s bet­ter to wait a while to get the biggest ben­e­fit of the drug as op­posed to start­ing it as ear­ly as you can,” said Dean Foll­mann, an ad­comm mem­ber and bio­sta­tis­ti­cian at the Na­tion­al In­sti­tutes of Health. “Fur­ther study of this to un­der­stand can we keep giv­ing it or is there an op­ti­mal time will be im­por­tant to do.”

There was al­so de­bate about how to ad­dress pa­tients with the APOE4 gene, which in­creas­es the odds of de­vel­op­ing Alzheimer’s. Hav­ing the gene al­so dra­mat­i­cal­ly rais­es the risk of ARIA, es­pe­cial­ly if pa­tients have two copies of the gene. Brana­gan said the FDA would strong­ly rec­om­mend get­ting test­ed for the gene, but wouldn’t re­quire it, part­ly due to the many un­easy ques­tions it can raise for a pa­tient and their rel­a­tives.