乌鸦传媒

WASHINGTON -- An extension study in Eli Lilly's EXPEDITION program for solanezumab found that the drug may offer some benefits in cognition and function when started earlier rather than later in disease progression, researchers reported here.

Patients who started solanezumab in the open-label extension didn't have the same preservation of cognition and function as those who'd been on the drug in the initial randomized, controlled trials -- meaning they never "caught up" to those who were treated earlier, according to , of the University of Southern California.

This was an indication of a genuine disease-modifying effect, the trial investigators indicated; if the drug was simply providing a burst of symptomatic relief, the delayed-start group would reach the same level of function as those first put on the drug.

But all patients continued to worsen, losing an average of 4 points on the 30-point Mini-Mental State Examination scale during the two-year extension study, and researchers not involved in the study noted that actual clinical effects of the drug remain to be seen.

"It has a number of effects on certain biological aspects of the Alzheimer's process, but whether it benefits patients yet is unknown," of Washington University in St. Louis, who was not involved in the study, told 乌鸦传媒. "It gives you some idea that you are targeting the disease mechanism. Hopefully it translates to clinical benefit."

Morris added that the study is "an innovative way to look at the solanezumab data. It's appropriate and gives justification for continuing to evaluate solanezumab and other drugs that demonstrate benefits."

, of the Mayo Clinic, said that while the researchers "have taken the statistical analysis to a higher level than had been done" in previous delayed-start analyses, the corollary is that "these differences are really small."

"They certainly could be clinically important eventually, but they may not be evident on a patient-by-patient basis," he said. "If people are worsening in both groups, and one group is worsening less, you may not be able to tell."

Delayed-Start Analysis

EXPEDITION and EXPEDITION2 were 18-month phase III placebo-controlled studies looking at patients with mild-to-moderate Alzheimer's disease. Neither met its primary endpoint -- but researchers did see a significant benefit on pooled analyses for the subgroup of patients with mild disease, in terms of both cognition and function.

So they decided to move forward with a delayed-start analysis by taking the patients who'd been on placebo in the first two trials and giving them solanezumab, while those in the original active-drug arms continued to receive it -- hence, the ongoing open-label EXPEDITION-EXT study.

Since all of the patients in EXPEDITION-EXT were receiving treatment, none of them could be blinded (although they remained unaware of their original assignments) and there was no inactive control. But the researchers noted that all patients and investigators were blinded for the initial trials and they weren't told about which group they were initially assigned to.

The goal of a delayed-start analysis, the researchers said, is to determine whether a drug has disease-modifying effects. It posits that if patients who start the treatment late never "catch up" to those who started early, then the drug actually impacts the course of the disease.

A few other trials have used delayed-start methodology for trying to prove an impact on disease progression. Some studies of the acetylcholinesterase inhibitors donepezil (Aricept) and tacrine (Cognex) had used the design, which was developed by Paul Leber of FDA in the late 1990s.

It was also memorably used in of rasagiline (Azilect), which sought to prove disease modification in Parkinson's. An FDA advisory committee unanimously voted against approval of the indication citing methodological issues.

But Aisen and colleagues said they developed a new method for delayed-start analyses to address some of those methodological issues, including the use of non-inferiority tests and better control for potential bias from unbalanced treatment groups.

Evidence of Benefit?

Ultimately, 491 of the original placebo group patients ended up going on solanezumab, and by that time the active group had 484 patients. A total of 286 and 295 patients, respectively, completed two years in the EXPEDITION-EXT trial. Deaths and adverse events accounted for relatively few of the discontinuations -- unspecified caregiver, physician, and participant decisions were the most common reasons reported.

The dose remained 400 mg every four weeks, the same dose used in the active groups in the original trials.

The researchers selected 28 weeks into the extension phase for the primary analysis time point -- that's 108 weeks total, which includes the 18 months of the initial studies plus the 28 weeks of the extension.

Ultimately, they found that the delayed-start group didn't "catch up" to the early-starters at 28 weeks.

Treatment differences were significant and non-inferior at that time point for both the ADAS-Cog14 (P=0.01) and the ADCS-IADL (P=0.043), they reported. Those differences remained significant through most of the two years of the extension study, but they did not always meet non-inferiority criteria.

And they became non-significant by the end of the study, Knopman noted.

"How long do you have to go that the groups remain separated before you can declare victory?" Knopman said, noting that the need to show benefit needs to be balanced by the fact that patients are sick and progressing and will be dropping out, diminishing the power of the study.

Differences in Mini-Mental State Exam (MMSE) scores were significant at most time points but they, too, never met non-inferiority criteria. And differences in the Clinical Dementia Rating scale-Sum of Boxes (CDR-SB) score rarely met significance in the extension even though they often hit non-inferiority.

Aisen and colleagues wrote that these findings "demonstrated that starting solanezumab early had benefits that persisted after placebo-treated patients began solanezumab 18 months later."

But those benefits were subtle, as patients on the drug showed substantial worsening on the four functional measures tracked in the study. In addition to the average four-point drop in MMSE scores seen during the extension, both patient groups showed increases of about 10 points on the ADAS-Cog14 scale and increases of about three points in the CDR-SB. Mean scores on the ADCS-IADL declined by about nine points during the two-year extension.

A safety analysis presented as a poster also found that the drug was well-tolerated. ARIA-E was seen in eight patients in the delayed-start group and 11 patients in the early-start group, but these were asymptomatic, the researchers said. Also, there were no differences in ARIA-H, a measurement of hemorrhage. But since there was no placebo control during the extension, picking out treatment-related effects was difficult. The investigators indicated that nothing new was seen in the safety data.

Zaven Khachaturian, PhD, the editor-in-chief of Alzheimer's & Dementia, said the study is "a good start" and is important because it's "the first to show effects in patients starting early."

Knopman noted that the improved controls for non-inferiority and balance were good, although questions remain about the actual margin of non-inferiority, and that the researchers successfully built on the lessons of ADAGIO.

The company is continuing the evaluate the safety and efficacy of solanezumab in patients with mild Alzheimer's disease and confirmed amyloid pathology in the ongoing EXPEDITION3 trial. It will also include a delayed-start analysis.

FDA Perspective

Several regulators joined a panel discussion after Aisen presented the solanezumab data, including officials from FDA and the European Medicines Agency.

The overwhelming message from those regulators was that researchers shouldn't become so concerned with secondary analyses that they forget that the goal is a large treatment effect, adding that disease modification is certainly not a requirement for approval.

"Obviously this is something we want," William Dunn, MD, director of CDER's neurology division at the FDA, said of disease modification, "but we shouldn't take our eyes off the prize, and that prize is most clearly a large treatment effect."

Lisa Lavange, PhD, director of CDER's biostatistics office, concurred that officials want disease-modifying drugs that have a huge public health impact, and that statistical methods are a clever way to show that. However, those analyses must have clinical relevance or other important outcomes, she said.

"It doesn't make sense to prove non-inferiority to [a measure] that we don't care about," she said. "Trials have shown drugs to be non-inferior to a drug that's proven to be active, and that works. But if it's non-inferior to something that doesn't work, that's not good."

Comment