Table 5.
Adjusted effect estimates and standard errors in hearing thresholds with one interquartile range increment in tibia lead (15 µg/g) from linear mixed effects models (No=448; total observations=2264).
| Effect estimate [beta (SE)], dB HL | |||
|---|---|---|---|
| Frequency | Time | Tibia lead | Time×tibia lead |
| 0.25 kHz | 0.263 (0.019)§ | −0.129 (0.319) | 0.026 (0.018) |
| 0.5 kHz | 0.157 (0.019)§ | −0.067 (0.364) | 0.026 (0.018) |
| 1 kHz | 0.244 (0.017)§ | 0.429 (0.336) | 0.038 (0.017)† |
| 2 kHz | 0.468 (0.027)§ | 0.740 (0.543) | 0.082 (0.026)‡ |
| 3 kHz | 0.727 (0.031)§ | 1.250 (0.954) | 0.045 (0.030) |
| 4 kHz | 0.748 (0.028)§ | 1.767 (0.980) | 0.049 (0.026) |
| 6 kHz | 0.868 (0.034)§ | 0.913 (1.139) | 0.058 (0.033) |
| 8 kHz | 1.204 (0.036)§ | 0.452 (1.103) | 0.090 (0.035)† |
| PTA | 0.406 (0.017)§ | 0.756 (0.431) | 0.050 (0.017)‡ |
All linear mixed effects models included an interaction between time elapsed from the baseline and tibia lead along with the main effects, adjusting for baseline age, race (white or other), education (>12 yr of education or not), occupational noise (low, medium, high), time varying covariates of body mass index (kg/m2), pack-years of cigarettes (0, <30, 30+), noise notch (yes or no), diabetes and hypertension, and random intercepts and random slopes for time. The spatial exponential correlation structure was considered the covariance matrix.
p<0.05,
p<0.01,
p<0.001.