The Solanezumab Benefit for Alzheimer’s: Oh, So Small, But Probably Real

Prospects for Alzheimer’s immunotherapy appear to have brightened. Yesterday at the 137th annual meeting of the American Neurological Association in Boston, Massachusetts, researchers presented analyses from Phase 3 trials for two therapeutic antibodies, solanezumab and bapineuzumab. Most of the new data were on solanezumab, Eli Lilly and Company’s humanized monoclonal antibody against the mid-region of monomeric Aβ, and they appeared to deliver the long-awaited signal the amyloid field needed to restore its confidence after a string of clinical setbacks.

Rachelle Doody, Baylor College of Medicine, Houston, Texas, reported a first analysis of the solanezumab Phase 3 raw data that Lilly had given to the Alzheimer’s Disease Cooperative Study (ADCS) for independent study under a contract that may well be unique in AD therapy development. “We analyze the data without any constraints imposed by the company,” Doody said. Speaking on behalf of the ADCS, Doody presented a data analysis based on the cooperative’s own modeling, not on Lilly’s internal statistical analysis plan that is on file with regulatory agencies.The ADCS analysis encompassed combined data from two solanezumab trials (EXPEDITION1 and EXPEDITION2). Both trials administered 400 milligrams of antibody once every four weeks for 18 months. One trial took place primarily in the Americas, one mostly in Western Europe and Australia; both trials used sites in Japan and other countries as well. The ADCS analysis came to a more positive result than what Lilly had disclosed in a press release on 24 August 2012. Specifically, Doody described a cognitive benefit in both trials overall and in patients with mild Alzheimer’s disease, as well as a functional benefit on independent activities of daily living (iADL) in the latter group.

Researchers’ reactions at the meeting ranged from enthusiasm, to cautious optimism, to quibbling with the statistics. Regardless of whether solanezumab itself, or possibly a more effective follow-up immunotherapy, will eventually become an FDA-approved treatment for clinical use, many scientists felt that yesterday’s result validated the amyloid hypothesis and could prove to be a turning point for the immunotherapy field at large.

“This is the first evidence for a clinical benefit with an anti-amyloid treatment. The field should be encouraged,” said Susanne Ostrowitzki of Hoffmann-La Roche in Basel, Switzerland. “We think the effect is real. It is a start,” Paul Aisen of the University of California, San Diego, who heads the ADCS, told Alzforum. John Morris, Washington University, St. Louis, Missouri, found the signal on functional benefit particularly noteworthy. “It’s been a discouraging 12 years since the field started looking for disease-modifying drugs. Today we got the first hint that we may have a small effect, even a functional effect. We have a long way to go toward a truly effective medicine, but today we got a real encouragement.”

Last August, Lilly reported that solanezumab missed both cognitive and functional endpoints (see ARF related news story). Pre-specified subgroup analysis, however, indicated the drug slowed cognition in patients with mild AD. The ADCS analysis shows a trend toward a cognitive effect across mild to moderate AD, i.e., people who entered the study with an MMSE of 16 to 26. Pooling data from both trials, Doody found that cognition declined more slowly in patients who received the immunotherapy. The effect turned up in two versions of the ADAS-Cog (ADAS-Cog11 and 14) and reached higher statistical significance in patients with mild disease. This fits with a theme that resonated through the meeting, namely that treating earlier will more likely bring success.

Going beyond cognition, Doody reported that by the end of the EXPEDITION2 trial, solanezumab had slowed decline on activities of daily living, though only in people with mild AD. This effect, on both the ADCS-ADL measures and the more challenging measure of iADL, also showed up in analyses of pooled data from the two trials, again only in patients with mild AD.

Doody emphasized that, while statistically significant, the solanezumab effect is small. After 81 weeks, the difference in the ADAS-Cog14 score between treated and untreated patients averaged 1.41 points, while scores overall worsened by about eight points during that time. Indeed, some scientists pointed out that the effect is so small—less than that of the approved cholinesterase inhibitor drugs—as to be imperceptible to patients. Nevertheless, attendees at the meeting saw the glass half-full. “I am very excited by what we heard today,” said William Mobley, University of California, San Diego, in a press briefing following the scientific presentations. “This little signal is big. It increases the evidence that Aβ is a target, which is quite significant,” he added. Reisa Sperling, Brigham and Women’s Hospital, Boston, agreed. “This is a shot in the arm that we might be going in the right direction with Aβ therapy,” she said.

Other scientists said they would have preferred to see the data presented as standardized effect sizes. Both effect size and standard deviation contribute to p value. The analysis incorporated a large number of covariates, which can achieve statistical significance by way of reducing deviation, even if the effect is very small, some trialists noted. Even they, however, said that because the results consistently trend in the same direction across groups and measures, they were believable. “We did a sensitivity analysis, which showed that the effects are small but real,” Doody said.

Also yesterday, Lilly issued a press release announcing its own detailed analyses of the same dataset, but Lilly did not present its own analysis at ANA. These are model-based analyses, and they generate somewhat different, analysis-dependent results. For example, the Lilly release claims various percentage reductions of decline at 18 months in mild AD for the two trials. Short of being able to compare both complete analyses side by side, scientists found it difficult to interpret the data available thus far. It was not possible to confirm with Lilly whether its scientists will formally present data from their own analyses, and when.

Presenting data on bapineuzumab, Sperling offered pooled analysis of her own. In this case, two Phase 3 clinical trials of bapineuzumab were conducted in patients who were positive or negative for ApoE4. At ANA, Sperling added new information to presentations last month at the European Federation of Neurological Societies (EFNS) annual meeting in Stockholm, Sweden (see ARF related news story). There, Sperling and Stephen Salloway, Brown University, Providence, Rhode Island, showed that in both trials the immunotherapy affected biomarkers, but not cognitive or functional decline. In Boston, Sperling reported that even combining the data from the ApoE4-positive and -negative trials did not change that outcome. “It was disappointing that we didn’t see an effect with a larger sample size,” Sperling told Alzforum.

The pooled study also showed that the brain volumes of people with mild Alzheimer’s disease who took the antibody shrank more than did the brains of people on placebo. This finding prompted some head-scratching after Sperling’s talk, much like a similar finding of increased shrinkage in responders of the AN1792 trial had eight years ago (see ARF related news story).

With biomarker effects from the bapineuzumab trials, and cognitive and functional effects from the solanezumab trials, researchers are now eager to see if the latter shifted biomarkers as well. These data will be presented at the Clinical Trials in Alzheimer’s Disease (CTAD) conference in Monaco later this month. Sperling said solanezumab could be a candidate for the A4 prevention trial (see ARF related news story), which she leads with Paul Aisen at UCSD. “We’ll be very interested to see that biomarker data,” she said.

Both Phase 3 programs await numerous further analyses. For example, the ApoE4 non-carrier bapineuzumab trial, startlingly, turned out to have included 36.1 percent of participants who were amyloid-negative on PIB PET. Was this a technical error with PET or a clinical misdiagnosis? This needs to be studied closely, Sperling said. She will also look at how the patients who did, indeed, have brain amyloid fared when analyzed as their own, more homogeneous group. Other ongoing analyses include a look at how the achieved antibody concentration in particular people relates to their outcomes, as well as a deeper analysis of how white matter side effects called ARIA, which were common in the bapineuzumab trials, relate to the study’s outcome measures.

In the future, trials of anti-amyloid drugs should ensure that participants indeed have brain amyloid, either by spinal tap or PET, said Salloway.

“The take-home message today is that there is light at the end of the tunnel,” said Randall Bateman from WashU. “We are going to be moving forward to get treatments with larger clinical and cognitive benefit.” For her part, Carole Ho of Genentech, South San Francisco, noted that all therapeutic AD antibodies are not equal. Each has unique binding properties and effector properties on microglial cells. Much of the outcome hinges on antibody engineering, and antibodies can differ from one another quite dramatically just as small-molecule drugs do.

Sperling put it this way, “I am encouraged that we see any clinical response to any antibody at this point. I think we should test many agents in the safest possible way.”

Citation