Speech audiometry findings from HIV+ and HIV- adults in the MACS and WIHS longitudinal cohort studies

Abstract

The purpose of this study was to compare various speech audiometry measures between HIV+ and HIV- adults and to further evaluate the association between speech audiometry and HIV disease variables in HIV+ adults only. Three hundred ninety-six adults from the Multicenter AIDS Cohort Study (MACS) and Women's Interagency HIV Study (WIHS) completed speech audiometry testing. There were 262 men, of whom 117 (44.7%) were HIV+, and 134 women, of whom 105 (78.4%) were HIV+. Speech audiometry was conducted as part of the standard clinical audiological evaluation that included otoscopy, tympanometry, and pure-tone air- and bone-conduction thresholds. Specific speech audiometry measures included speech recognition thresholds (SRT) and word recognition scores in quiet presented at 40dB sensation level (SL) in reference to the SRT. SRT data were categorized in 5-dB steps from 0 to 25dB hearing level (HL) with one category as ≥30dB HL while word recognition scores were categorized as <90%, 90-99%, and 100%. A generalized estimating equations model was used to evaluate the association between HIV status and both ordinal outcomes. The SRT distributions across HIV+ and HIV- adults were similar. HIV+ and HIV- adults had a similar percentages of word recognition scores <90%, a lower percentage of HIV- adults had 90-99%, but HIV- adults had a higher percentage of 100%. After adjusting for covariables, HIV+ adults were borderline significantly more likely to have a higher SRT than HIV- adults (odds ratio [OR]=1.45, p=0.06). Among HIV+ adults, HIV-related variables (i.e., CD4+ T-cell counts, HIV viral load, and ever history of clinical AIDS) were not significantly associated with either SRT or word recognition score data. There was, however, a ceiling effect for word recognition scores, probably the result of obtaining this measure in quiet with a relatively high presentation level. A more complex listening task, such as speech-in-noise testing, may be a more clinically informative test to evaluate the effects of HIV on speech communication.

Keywords: Adults; HIV; Speech audiometry.

Figure 1

The distributions of speech recognition threshold scores stratified by HIV status are shown. White lines indicate median data, boxes are the 25^th and 75^th percentiles, whiskers are values outside of the middle 50% of the distribution (the interquartile range), and plus signs represent extreme values or “outliers”.

Figure 2

a. The distributions of WRS for HIV+ (blue) and HIV- (red) participants are shown. b. The percentages for HIV+ (blue) and HIV- (red) participants for the three word recognition score categories, <90%, 90–99%, and 100%, are shown.

Figure 2

a. The distributions of WRS for HIV+ (blue) and HIV- (red) participants are shown. b. The percentages for HIV+ (blue) and HIV- (red) participants for the three word recognition score categories, <90%, 90–99%, and 100%, are shown.