Atabecestat Associated With Cognitive Worsening in Alzheimer Disease

Reisa A. Sperling, MD

A recent randomized, phase 2b/3 clinical EARLY trial found that atabecestat for the treatment of Alzheimer disease (AD) was associated with dose-related cognitive worsening as early as 3 months and presence of neuropsychiatric treatment-emergent adverse events (TEAEs). The TEAEs and cognitive status returned to baseline after 6 months off treatment. The trial was ended early due to hepatic-related TEAEs.

Patients treated with 25 mg of atabecestat showed cognitive worsening on the Preclinical Alzheimer Cognitive Composite (PACC) compared to placebo at month 6 (least-square mean difference [LSMD], –1.09; 95% CI, –1.66 to –0.53; P <.001) and month 12 (LSMD, –1.62; 95% CI, –2.49 to –0.76; P <.001). Similar results were seen at 3 months on Repeatable Battery for the Assessment of Neuropsychological Status (RBANS; LSMD, –3.70; 95% CI, –5.76 to –1.63; P <.001).

Reisa A. Sperling, MD, professor of neurology, Harvard Medical School, memory disorders unit and department of neurology, Brigham Women’s Hospital and Massachusetts General Hospital, and colleagues wrote that “there is abundant genetic evidence that overproduction of Aβ42 leads to AD and interference with production of Aβ42 may be protective. Decreasing Aβ production remains attractive as a potential disease-modifying strategy for AD. However, for β-secretase-1 inhibitor (BACEi) to be pursued safely researchers will need to explore lower doses with more modest enzyme inhibition, with careful safety/cognitive monitoring.”

The EARLY trial enrolled 557 amyloid positive and cognitively normal patients from 143 centers across 14 countries. Women made up 61.2% (n = 341) of patients and the mean age was 70.4 years (standard deviation [SD], 5.56). Patients were randomized to 10 mg or 25 mg of atabecestat, or placebo. The 10 mg dose was reduced to 5 mg, however, owing to TEAE concerns. Patient follow-up was performed after the trial’s early halt.

READ MORE: Troriluzole Not Effective for Symptomatic Treatment of Alzheimer Disease

LMSDs of 25 mg versus placebo are reported above. The LMSD of 5 mg versus placebo on PACC was −0.38 (95% CI; −0.92 to 0.17; P = .17) at 6 months and −0.79 (95% CI; −1.63 to 0.05; P = .06) at 12 months. Sensitivity analyses confirmed that neither the hepatic-related nor neuropsychiatric-related AEs fully accounted for cognitive worsening. Patients on atabecestat performed significantly worse than placebo on all components of PACC except the Mini-Mental State Examination. 

Patients on 5 mg of atabecestat performed insignificantly worse than placebo on RBANS at 3 months (LMSD, −1.97; 95% CI, −3.99 to 0.05; P = .06) and no significant differences were seen between either dose and placebo at 9 and 15 months. No significant differences were seen in cognitive function index (CFI) between atabecestat doses and placebo at 12 months, but nonsignificant subjective worsening was observed in the 25-mg dose group.

Sperling and colleagues found that the mean change in PACC from baseline to last received treatment was –1.23 (SD, 1.7; 95% CI, –1.72 to –0.74). The last on-treatment to last off-treatment assessment improved by 1.0 point (SD, 1.92; 95% CI, 0.47–1.54) for the 25-mg group, indicative of a return to cognitive baseline after cessation of treatment. Similarly, the RBANS mean change was –1.6 (SD, 8.29; 95% CI, –4.01 to 0.8) and last on-treatment to last off-treatment scores improved by 1.08 (8.81; 95% CI, –1.45 to 3.61) points in the 25-mg group.

Some patients treated with atbecestat experienced TE increases in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) of more than 3 times of the upper limit of normal (ULN). This was seen in 6.9% (n = 13) of patients in the 5-mg groups, 14.8% (n = 27) in the 25-mg group, and 0.5% (n = 1) patients on placebo. These increases were asymptomatic and resolved after drug discontinuation.

Sperling and colleagues also found small dose and duration-related decreases in whole-brain volume in atabcestat groups compared to placebo from baseline to months 6 and 12. No treatment group differences were seen in cerebrospinal fluid (CSF) neurofilament light or total tau. CSF p181tau decreased by 2.2 pg/mL (SD, 7.26) in the 5-mg group, decreased by 2.1 pg/mL (SD, 8.52) in the 25 mg group, and increased by 1.8 pg/mL (5.60) in the placebo group. No correlations were found with these markers and cognition or brain volume.

“There is abundant genetic evidence that overproduction of Aβ42 leads to AD and interference with production of Aβ42 may be protective. Decreasing Aβ production remains attractive as a potential disease-modifying strategy for AD. However, for BACE [inhibitors] to be pursued safely researchers will need to explore lower doses with more modest enzyme inhibition, with careful safety/cognitive monitoring,” Sperling and colleagues concluded.

REFERENCE

Sperling R, Henley D, Aisen PS, et al. Findings of efficacy, safety, and biomarker outcomes of atabecestat in preclinical Alzheimer disease: A truncated randomized phase 2b/3 clinical trial. JAMA Neurol. Published online January 19, 2021. doi:10.1001/jamaneurol.2020.4857