Minimal Clinically Important Differences May Help Alzheimer Disease Clinical Trials

Researchers identified triangulated minimal clinically important differences (MCID) in cognitive test scores for Alzheimer disease (AD). The study findings were published in Neurology.

Clinicians can make better decisions and researchers can design appropriate clinical trials with MCIDs. In 2018, the US Food and Drug Administration (FDA) stated that drugs for AD should show clinically meaningful improvement on cognitive tests, however, the MCID of many clinical cognitive assessments remains unknown. Few studies have evaluated which tests best gauge clinically relevant worsening in AD.

This is the first study to assess which combination of cognitive tests best estimates clinically relevant yet minimal worsening in Clinical Dementia Rating-Sum of Boxes (CDR-SB), according to the researchers. For the study, the researchers established MCIDs in test scores for 8 cognitive tests for AD.

They included 451 individuals (16% amyloid positive) who were cognitively unimpaired from the Swedish BioFINDER. Of that group, 90 individuals had subjective cognitive decline and 292 individuals had mild cognitive impairment (MCI).

The researchers developed MCIDs linked with a change of ≥0.5 or ≥1.0 on CDR-SB for cognitively unimpaired, MCI, and amyloid positive cognitively unimpaired tests and analyzed receiver operating characteristic (AUC) curve for predicting this change.

Reliable change index was larger among participants with MCI compared with those cognitively unimpaired in Stroop, Letter S, Symbol Digit, and Trailmaking Test (TMT) A and B. They were similar among both cognitively unimpaired and MCI groups for Mini Mental State Examination (MMSE), Alzheimer’s Disease Assessment Scale (ADAS), and Animal Fluency.

The researchers identified significant, yet weak to moderate, correlations between CDR-SB and cognitive tests.

The best predictive univariate test for cognitively unimpaired was ADAS delayed recall (AUC, 0.75) while Animal Fluency (AUC, 0.54) was worst.

MMSE was best (AUC, 0.75) and TMT A was worst (AUC, 0.61) for MCI. Animal Fluency, Letter S, and TMT B lacked sufficient follow-up data.

When tests were combined, the best model for identifying clinical change for in individuals who were cognitively unimpaired was age and changes in ADAS delayed recall, MMSE, and TMT B (AUC, 0.79; 95% CI, 0.72-0.86). The best predicting model (AUC, 0.82; 95% CI, 0.76-0.88) for participants with MCI included changes in Stroop, MMSE, and age.

Among amyloid positive participants who were cognitively unimpaired, changes in ADAS delayed recall, Stroop, Symbol Digit, and TMT B, including gender in the model, comprised the best predictive cognitive measure. Removing Stroop, the researchers found ADAS delayed recall, MMSE, symbol digit modalities test, TMT B, and gender (Akaike Information Criterion, 130.5; AUC, 0.87; 95% CI, 0.79-0.94; sensitivity 75%; specificity 88%) was the best predicting composite cognitive measure.

“Our MCIDs may be applied in clinical practice or clinical trials for identifying if

a clinically relevant change has occurred,” the researchers said. “The composite measure can be useful as a clinically relevant cognitive test outcome in preclinical AD trials.”

Study limitations included possible fluctuation in cognition in individuals cognitively unimpaired and reversion of participants with MCI participants to being cognitively unimpaired.

The researchers said tests should be validated in independent, more diverse populations with wider age range and education level and that Stroop violates the expectation of ordered effect size.

Disclosure: Some study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.