Utility of the Repeatable Batter for the Assessment of Neuropsychological Status (RBANS) as Endpoint in AD Trials

Abstract

AbstractBackgroundThe Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) is a brief neurocognitive battery that is commonly used in clinical settings to detect and track early symptomatic Alzheimer’s disease (AD) and other neurodegenerative conditions that involve neurocognitive compromise. The RBANS has also been used as an endpoint in trials involving presymptomatic and early symptomatic AD. As the RBANS generates scores that are scaled to population norms, it theoretically may be subject to floor effects in more impaired populations. The present study examined screening RBANS data from several multinational clinical trials of presymptomatic and early symptomatic AD as a function of CDR sum of box (CDR‐SB) stages.MethodAggregated data from six multinational clinical trials of presymptomatic and early symptomatic AD were analyzed. There were 19,126 subjects with screening CDR‐SB scores ranging from 0–7. RBANS total scale index score was evaluated as function CDR‐SB scores at each stage. Descriptive statistics, including skewness and kurtosis, were also summarized. Frequency distributions of RBANS total were plotted at each CDR‐SB stage.ResultThe overall mean for the RBANS total scale index score was 84.3 (SD = 17.3). The distributions of RBANS total scale index score remained reasonably normal across the range of CDR‐SB, but the largest magnitude of change on the RBANS total scale index score as a function of CDR‐SB was over the range of CDR‐SB 0–1.5 (CDR‐SB = 0: n = 4,154, mean = 100.7, SD = 12.35, skewness = 0.19, kurtosis = 0.37; CDR‐SB = 1.5: n = 1,575, mean = 78.7, SD = 12.91, skewness = 0.22, kurtosis = 0.24).ConclusionThe findings from this study suggest that the standard normative‐referenced index scoring of the RBANS may be most useful in preclinical trials, although alternative scoring approaches de‐linking the raw scores to population norms (e.g., baseline z‐score transformations) could provide greater sensitivity to change in samples with more significant neurocognitive impairment.