In January, Eli Lilly revealed results from a mid-stage study showing an Alzheimer’s medicine it’s developing slowed cognitive and functional decline in people with early signs of the disease. The findings — one of the few times such an outcome has been reported in clinical testing — sparked cautious optimism in a field accustomed to the failure of drug after drug.

Detailed data from the Phase 2 trial, published Saturday in the New England Journal of Medicine, now offer more context. While a rare success, they could also temper enthusiasm for a medicine that may still be some ways away from definitive proof of its effectiveness.

Called TRAILBLAZER-ALZ, Lilly’s study enrolled people with early symptomatic Alzheimer’s who had the disease’s characteristic plaques in their brains. Researchers measured study participants’ symptoms on a composite scale that combines measures of cognition, such as recalling words, and functions essential for independent living, such as preparing meals.

Over a year and a half, the 131 volunteers who received monthly infusions of Lilly’s drug, called donanemab, declined by 6.9 points on the 144-point scale. By comparison, the 126 people who received a placebo in the study saw their scores drop by 10.1 points, a statistically significant difference of 3.2 points.

Trial participants had an average score of 106 points at the beginning of the study.

Alzheimer’s specialists polled by analysts at Mizuho Securities USA expected a difference of at least three points, suggesting the study’s primary result may be well received among physicians and researchers. The strength of the statistical difference — measured by a number known as the p-value — fell short of some analysts’ expectations, however.

Expressed in percent terms, treatment appeared to slow declines from disease by 32%. “When we talk about slowing [Alzheimer’s progression] … what we get excited about is that over an 18 month trial, that means we’ve essentially pushed back the disease six months,” Mark Mintun, head of Lilly’s Alzheimer’s disease development unit, said in an interview.

Along with the data’s publication in the New England Journal of Medicine, Lilly is presenting its findings at the International Conference on Alzheimer’s and Parkinson’s Diseases on Saturday.

Lilly had previously announced the 32% figure in a Jan. 11 press release. Newly disclosed, however, is a detailed accounting of how the drug performed on several important secondary measures, on several of which donanemab didn’t show a clear-cut difference.

Using a scale known as clinical dementia rating-sum of boxes or CDR-SB, for instance, volunteers treated with donanemab did not do significantly better than those on placebo when participants were assessed at the trial’s end at 76 weeks.

There was a larger difference in CDR-SB between groups at 36 and 52 weeks, according to study researchers.

CDR-SB, which tests patients’ memory and asks caregivers about their symptoms, is considered an important measure by the Food and Drug Administration.

Biogen, which is waiting to see whether the agency will approve its Alzheimer’s medicine aducanumab, used CDR-SB as the primary measure in its pivotal trials. And Lilly is using it as the main goal of its second Phase 2 study testing donanemab, results from which are expected in 2023.

Donanemab treatment in TRAILBLAZER-ALZ also didn’t lead to clearly better results on the functional assessment used as part of the primary measurement, possibly indicating the drug’s main effect within the composite scale was on cognition.

The mixed findings could invite skepticism of the drug’s true benefit, although Lilly highlighted the consistent slowing of decline across measures. According to Lilly’s Mintun, TRAILBLAZER-ALZ also wasn’t large enough to detect a benefit on each secondary endpoint separately.

“We are now discussing with regulators what changes are necessary to optimize the usefulness of the next trial to replicate [this] data,” Mintun said.

Like many other experimental Alzheimer’s medicines that have failed in testing, donanemab was designed to clear out built-up plaques of a toxic protein known as amyloid. Finding new ways to target these plaques has been the central focus of Alzheimer’s research for decades, but time and time again drugs that reduce plaque haven’t proven beneficial clinically.

Across Lilly’s study, donanemab did both, sharply reducing plaques as well as slowing disease progression. On an individual level, however, researchers were not able to show an association between the plaque reductions and clinical outcomes.

Lilly developed donanemab to target a specific form of amyloid plaques, which the company argues makes the medicine different than drugs that were tested previously.

Donanemab does come with some safety concerns, most notably a type of brain swelling that’s a common side effect of amyloid-targeting drugs. Roughly a quarter of trial volunteers who got donanemab had this side effect, compared to 1% of those receiving placebo. This swelling led to symptoms in 6% of donanemab patients. Seven study participants stopped taking the drug as a result and two were hospitalized.

All told, roughly 30% of people given donanemab discontinued treatment at some point in the study due to side effects, versus 7% of the placebo group.