Table 1.
Summary statistics for auditory testing/imaging biomarkers and unadjusted relationships between demographic and performance on pure tone audiometry AND associations between pure tone audiometry performance and neuropsychological testing in cognitively normal older adults
| Binary demographics | Unadjusted association with PTA | ||
| Median (IQR) PTA | P value | ||
| Sex | Males: 23.5 (12.5), n=187 Females: 21.25 (12.5), n=181 |
0.19* | |
| Educational attainment | Non-advanced: 23.75 (15), n=169 Advanced: 21.25 (10), n=199 |
0.09* | |
| Childhood socioeconomic position | Non-manual: 22.55 (11.25), n=218 Manual: 21.25 (15), n=150 |
0.95* | |
| Adulthood socioeconomic position | Non-manual: 21.25 (11.25), n=313 Manual: 23.75 (16.25), n=55 |
0.17* | |
| Tinnitus |
No tinnitus: 21.25 (12.5), n=
288
Tinnitus: 26.25 (13.75), n= 80 |
0.0004* | |
| Hearing aid use |
Non-user: 21.25 (10), n=
313
User: 38.75 (16.25), n= 55 |
<0.0001* | |
| APOE genotype | Non-ε4 carrier: 22.5 (12.5), n=258 ε4 carrier: 21.25 (13.25), n=110 |
0.089* | |
| Aβ positivity, n (%) | Aβ-negative: 21.88 (12.5), n=306 Aβ-positive: 23.75 (12.5), n=62 |
0.25* | |
| Continuous demographics | Summary statistic | Unadjusted association with PTA | |
| Age, years, median (IQR) | 70.6 (1.1), n=368 | Rho=0.07, p=0.2† | |
| Childhood cognition, z-score, median (IQR) | 0.48 (0.84), n=368 | Rho=−0.02, p=0.72† | |
| Cognitive tests | Summary statistic | Association with PTA | |
| Univariate | Multivariate β-coefficient‡ (95% CI) | ||
| Logical memory delayed recall (Wechsler Memory Scale-Revised—out of 25), mean (SD) | 11.6 (3.6), n=368 | r=−0.065, p = 0.21§ |
−0.009 (−0.044 to 0.026)¶ |
| Digit-symbol substitution (Wechsler Adult Intelligence Scale-Revised—out of 93), mean (SD) | 48.8 (10.0), n=368 | r=−0.047, p = 0.36§ |
−0.005 (−0.098 to 0.087)¶ |
| 12-item Face-Name (out of 96), mean (SD) | 65.6 (18.2), n=364 | r=0.099, p = 0.85§ |
−0.01 (−0.20 to 0.18)¶ |
| Matrix Reasoning (Wechsler Abbreviated Scale of Intelligence—out of 32), median (IQR) | 25 (4), n=368 | Rho=−0.11, p=0.04§ | −0.037 (−0.084 to 0.007)** |
| MMSE (out of 30), median (IQR) | 30 (1), n=368 | Rho=−0.13, p=0.013§ | −0.009
(−0.019 to −0.001)** |
| MMSE without ‘No ifs, ands or buts’ (out of 29), median (IQR) | 29 (1), n=368 | Rho=−0.07, p=0.21§ | −0.005 (−0.015 to 0.003)** |
| “No ifs, ands or buts” repetition (out of 1), proportion who answered Item correctly (%) | 306/368, (83.2%) | p=0.0017a | −0.031
(−0.056 to − 0.006)†† |
Analyses that attained statistical significance at the 5% level are highlighted in bold.
*Wilcoxon rank sum test.
†Spearman correlation.
‡Coefficient represents increase in cognitive score per dB HL increase in PTA (covariates were: age at scan; sex; APOEε4 genotype; WMHV; educational attainment; childhood cognitive ability; socioeconomic position; and Aβ status).
§Pearson’s correlation.
¶Linear regression models with robust SE.
**Bootstrapping.
††Logistic regression.
Aβ, β-amyloid; AD, Alzheimer’s disease; MMSE, mini-mental-state-examination; PET, positron emission tomography; PTA, pure tone average in best hearing ear; r, Pearson’s r; Rho, Spearman’s Rho; WMHV, white matter hyperintensity volume.