Clinical trial results released today by Eli Lilly and Co. indicate its antibody donanemab clearly, if perhaps modestly, slows the progression of Alzheimer’s disease. Following on the heels of comparable results for a similar antibody, lecanemab, the data bolster the long-held but contested hypothesis that preventing the accumulation of a protein called beta amyloid in the brain could help the many millions of people who develop the fatal neurodegenerative disorder.

“We are extremely pleased that donanemab yielded positive clinical results with compelling statistical significance,” Daniel Skovronsky, Eli Lilly’s chief scientific and medical officer, and president of Lilly Research Laboratories, said in a press release that noted the antibody had slowed the cognitive and functional decline of Alzheimer’s patients by 35% compared with a placebo.

“This is monumental and aligns very closely to lecanemab as far as the amount of benefit,” Donna Wilcock, an Alzheimer’s disease scientist at the University of Kentucky College of Medicine, says. “In fact, it looks like donanemab may have greater benefit, although the trial designs do not allow direct comparison.”

But Eli Lilly’s preliminary donanemab results also reveal a sobering risk of brain swelling and hemorrhaging, side effects that the company disclosed may be linked to two—perhaps three—deaths in the clinical trial and that echo hazards seen with lecanemab, which is being marketed by Eisai and Biogen. With the U.S. Food and Drug Administration considering full approval for lecanemab next month and Eli Lilly vowing to quickly submit donanemab to the agency for review, many physicians, Alzheimer’s patients, and their caregivers may soon face difficult conversations about whether to risk immediate harm to take these therapies, the first to be clinically proven to somewhat thwart a slow but inexorable destroyer of the brain.

Donanemab is “not a miracle drug,” cautions Sorbonne University neurologist Nicolas Villain, who helped clinically test lecanemab and worries that Eli Lilly’s antibody may carry even higher risks.

Both donanemab and lecanemab bind to various forms of beta amyloid in the brain and are intended to promote their clearance from the extracellular deposits knowns as plaques that many scientists suspect cause neurons to malfunction and die. Brain scans taken during Eli Lilly’s phase 3 trial, which involved more than 1700 people with early signs of Alzheimer’s disease, showed that donanemab cleared the protein effectively. More than half of participants receiving the antibody via monthly infusions had their amyloid plaque fall to a level that trial designers had designated in advance as a criterion for stopping treatment and switching to the placebo group—Eli Lilly designed this aspect of the trial to help see whether amyloid would then resume accruing or hold at the lower level.

But more important, the preliminary results from the trial found that treated individuals declined more slowly in two different assessments of cognitive and physical functions over 18 months than those given a placebo. That echoes the groundbreaking result for lecanemab reported more than a year ago.

Eli Lilly only presented topline numbers from the trial today, saying it plans to reveal more data at an Alzheimer’s disease meeting in July and in a forthcoming publication. For now, the company says the drug slowed the disease’s expected progression by 35% based on a scale it developed and by 36% on the more widely used Clinical Dementia Rating-Sum of Boxes, or CDR-SB, scale, which gathers clinical and caregiver assessment to rank dementia severity on an 18-point scale. Eli Lilly also reported that 47% of the treated trial participants had no worsening of disease severity based on the CDR-SB, compared with 29% among the placebo group. The treated individuals also, on average, maintained a greater ability to perform key daily activities such as using the toilet independently and dressing themselves, the company said.

The Eli Lilly trial also looked at another brain protein, tau, that has attracted the interest of many researchers. Unlike beta amyloid, which forms plaques outside brain cells, tau malfunctions inside neurons, becoming enmeshed in clumps known as tangles. Their presence tends to indicate more advanced disease and Eli Lilly’s team predicted that those with higher levels of tau would be less responsive to the antibody treatment. So, participants in the trial were stratified based on tau levels in their brain to check for responsiveness to treatment, an aspect of the trial that U.S. National Institute on Aging neurologist Madhav Thambisetty praised as “innovative and beautifully designed,” as accounting for tau in an amyloid-targeting trial is not standard.

Those with high levels of tau also seemed to also benefit from donanemab, if perhaps less, the company reports. That raises hopes the drug could be used in later stages of the disease, although whether regulatory agencies will be convinced of that is an open question given it’s a new way to identify patient populations who will best benefit from the drug.

Outside researchers will want to scrutinize Eli Lilly’s analysis to assess how meaningful the clinical benefit from the drug is given its risk of serious side effects. For one, the company’s own Alzheimer’s disease scale has not been validated as extensively as CDR-SB. For another, the absolute difference between treatment and placebo groups on CDR-SB was just under 0.7; some Alzheimer’s researchers argue it takes at least a full point on that scale to represent a clear difference in disease severity, but others, including patients and caregivers, say smaller differences can signify a meaningful retention of cognitive and physical skills.

Beyond that debate, Eli Lilly’s results are likely to fuel concerns about the dangers of amyloid-targeting antibodies. Lecanemab has been linked to several deaths and cases of severe brain damage, as well as to less severe brain swelling and hemorrhaging. Although Eli Lilly did not disclose any participant data related to genotyping or use of blood thinners, outside scientists are eager to see such data to evaluate the drug’s risks. People on these medicines or who have two copies of the APOE4 gene variant, which raises the risk of Alzheimer’s disease, may be particularly susceptible to harm from amyloid-targeting antibodies, according to evidence from several trials.

Eli Lilly did report that a form of brain swelling known as ARIA-E occurred in 24% of people who received its antibody, with about 6% reporting symptoms. A more severe condition involving microhemorrhages, and known as ARIA-H, affected 31.4% of antibody-treated people, versus 13.6% of the placebo groups. (ARIA-E and ARIA-H are known symptoms of Alzheimer’s disease.) Eli Lilly says most ARIA cases were “mild to moderate” and stabilized after “appropriate management,” but the company acknowledges the incidence of “serious” ARIA was 1.6%. It did not disclose any more details on the deaths of three trial participants.

“Looks as dangerous as lecanemab,” Robert Howard, a psychiatrist at University College London who has tested treatments for dementia, tweeted about donanemab. Villain even suggests that Eli Lilly’s topline data on ARIA may indicate its antibody is more harmful than lecanemab.

Still, many scientists who have backed the so-called amyloid hypothesis of Alzheimer’s are likely to see vindication in today’s results, as they offer further evidence that beta amyloid is a significant, if not the main, driver of the disease. And even if the overall risk-benefit ratio of the antibodies proves poor, their partial success may inspire better, safer interventions to remove or prevent amyloid buildup.

Many other questions, such as whether donanemab or lecanemab offers the bigger advantage and how much treatments will cost, remain to be answered. Although some scientists say the “treatment era” of Alzheimer’s is finally about to start, others urge caution. Ultimately, donanemab’s “clinical meaningfulness, safety & accessibility remain to be established & can only be assessed by full, transparent and timely disclosure of all data especially those related to adverse events and safety concerns,” Thambisetty tweeted this morning.