Solanezumab does not appear to slow cognitive decline in patients with mild Alzheimer’s disease, according to the newly-published results of the EXPEDITION-3 trial (Honig et al. N Engl J Med 2018;378:321-330). Prior studies also failed to show a benefit.

Solanezumab is a monoclonal antibody targeting soluble beta-amyloid, promoting efflux of beta-amyloid from the brain to the plasma (sink effect). In the EXPEDITION-1 and -2 trials in mild-to-moderate Alzheimer’s disease (n=2052), solanezumab 400 mg IV q4 weeks for 18 months failed to show significant improvement versus placebo in change from baseline to week 80 in the 11-item ADAS-cog scale score; the treatment effect was -0.8 points and -1.3 points in the two trials (Doody et al. N Engl J Med 2014;370:311-321).

However, a prespecified secondary analysis of pooled data from the EXPEDITION studies showed a reduction in cognitive decline the subgroup of patients with mild AD (MMSE score 20-26) (Siemers et al. Alzheimers Dement 2016;12:110-120). This prompted the hypothesis that anti-amyloid immunotherapy may be beneficial earlier in the disease course, before irreversible damage has occurred.

This was the rationale for the EXPEDITION-3 trial in patients with mild AD (Honig 2018). Subjects (n=2129) were randomized to placebo or solanezumab 400 mg q4weeks for 76 weeks. The primary outcome measure was the same as for EXPEDITION-1 and -2; the result in the EXPEDITION-3 cohort of mild AD patients was also the same. The change from baseline to week 80 in ADAS-cog14 score was 6.65 with solanezumab versus 7.44 with placebo (treatment difference -0.8). The change from baseline in MMSE score was -3.17 with solanezumab and -3.66 with placebo.

A separate delayed-start analysis to identify a putative disease-modifying effect was inconclusive (Lin-Seifert et al. J Prev Alzheimers Dis 2018;5:8-14).

Why did anti-amyloid immunotherapy fail? One hypothesis, based on mathematical modelling, is that shorter beta-amyloid isoforms (e.g. Aβ40) have beneficial effects on glutamate neurotransmission and cognition, which partially offset the adverse effects of longer forms (e.g. Aβ42). Thus, indiscriminately reducing all beta-amyloid forms may worsen cognition in patients with a low baseline beta-amyloid load (Geerts et al. Alzheimers Res Ther 2018;10:14; free full text at www.ncbi.nlm.nih.gov/pmc/articles/PMC5797372/pdf/13195_2018_Article_343.pdf). As such, the authors offer the counterintuitive suggestion that amyloid-lowering agents may actually worsen cognition outcomes in subjects earlier in the disease course with a lower beta-amyloid load.