Table 4.
pAO2 Mean | pAO2 SD | pAO2 %Negative | |||||
---|---|---|---|---|---|---|---|
Exposure | N | β (95% CI) | p | β (95% CI) | p | β (95% CI) | p |
%emphysema −950 HU | |||||||
Unadjusted | 54 | −0.0007 (−0.002, 0.0005) | 0.26 | 0.002 (0.0009, 0.003) | 0.0002 * | 0.57 (0.29, 0.86) | 0.0002 * |
Model 1 | 54 | −0.0003 (−0.002, 0.0009) | 0.62 | 0.001 (0.0003, 0.002) | 0.005 * | 0.37 (0.11, 0.63) | 0.007 * |
Model 2 | 53 | −0.0001 (−0.001, 0.001) | 0.85 | 0.001 (0.0003, 0.002) | 0.006 * | 0.32 (0.05, 0.60) | 0.02 * |
Visual emphysema severity (%) 1 | |||||||
Unadjusted | 50 | −0.0002 (−0.001, 0.0006) | 0.64 | 0.0008 (0.0002, 0.001) | 0.01 * | 0.29 (0.10, 0.48) | 0.004 * |
Model 1 | 50 | −0.00005 (−0.0008, 0.0007) | 0.90 | 0.0005 (0.0001, 0.001) | 0.02 * | 0.22 (0.05, 0.38) | 0.01 * |
Model 2 | 49 | 0.00006 (−0.0008, 0.0009) | 0.89 | 0.0005 (0.00003, 0.001) | 0.04 * | 0.19 (0.01, 0.36) | 0.04 * |
ADC 2 | |||||||
Unadjusted | 50 | −0.0005 (−0.001, 0.0002) | 0.17 | 0.0008 (0.0003, 0.001) | 0.003 * | 0.36 (0.12, 0.61) | 0.004 * |
Model 1 | 50 | −0.0004 (−0.001, 0.0004) | 0.33 | 0.0007 (0.0002, 0.001) | 0.005 * | 0.30 (0.08, 0.53) | 0.01 * |
Model 2 | 49 | −0.0002 (−0.001, 0.0006) | 0.63 | 0.0006 (0.0002, 0.001) | 0.01 * | 0.24 (0.02, 0.47) | 0.04 * |
DLCO | |||||||
Unadjusted | 34 | −0.0003 (−0.002, 0.001) | 0.65 | 0.001 (0.0003, 0.002) | 0.007 * | 0.23 (0.05, 0.41) | 0.01 * |
Model 1 | 34 | −0.0001 (−0.002, 0.002) | 0.93 | −0.001 (−0.002, −0.00003) | 0.04 * | −0.22 (−0.46, 0.01) | 0.06 |
Model 2 | 34 | −0.0001 (−0.002, 0.002) | 0.91 | −0.001 (−0.002, −0.00004) | 0.04 * | −0.22 (−0.45, 0.01) | 0.06 |
1 Calculated as the sum of the severity scores for centrilobular, panlobular, and paraseptal emphysema. 2 Per 0.01 change. * p-value < 0.05. ADC, apparent diffusion coefficient; HU, Hounsfield units; pAO2, partial pressure of oxygen; SD, standard deviation. Unadjusted: (weighted). Model 1: adjusted for age, sex, race/ethnicity, cigarette smoking status (weighted). Model 2: Model 1 + %-predicted FEV1 (weighted). Generalized linear models were used, and participants were weighted on the inverse ratio of probability of selection.