The effect of diabetes on measures of audition and cognition among veterans

Abstract

Type 2 Diabetes Mellitus (DM) is a metabolic disease with complications that can impair auditory and cognitive function. This thesis describes cognitive function based on two cognitive behavioral measures (Digit Symbol Substitution [DSS] and Letter Number Sequencing subtests of the WAIS-III) and the electrophysiological P300 event-related potential and audiometric function based on pure-tone thresholds (0.25kHz threshold and pure-tone averages developed for low/mid- [0.5, 1, 2kHz], high- [3, 4, 6, 8kHz] and ultra-high frequencies [10, 12.5, 14kHz]), speech understanding (QuickSIN and time-compressed speech). Groups were constructed based on indices related to glycated hemoglobin (HbA1c) levels (control, prediabetic, all DM), or whether insulin was required to manage the DM (control, non-insulin-dependent [NIDDM], insulin-dependent [IDDM]). Following initial univariate statistical analyses, separate generalized linear regression models were fit to response measures. After adjusting for age, elevated pure-tone thresholds at 250Hz and low/mid-frequency PTA were associated with IDDM. Among diabetics, poorer thresholds were observed in IDDM compared with NIDDM groups. Additionally, IDDM group had, on average, the poorest cognitive and audiometric measures across groups. Overall results indicate DM-induced changes in cognition and audition, trending towards statistical significance, and generally associate better DM control with better outcome measures. Notably, effect sizes were small and speech frequencies were, on average, clinically normal in our cohort; therefore, not surprisingly, speech understanding was not significantly affected. This is the baseline report of a longitudinal study.