Table 1.
Model parameter | Estimate | 95% CI | P | |
Viral load model | Intercepta | 0.093 | 0.086–0.10 | <0.001 |
Viral load at t−1 (per 1 log10 copies/ml higher) | −0.026 | −0.031 to −0.021 | <0.001 | |
CD4+ cell count at t0 (per 100 cells/μl higher) | −0.004 | −0.007 to −0.002 | <0.001 | |
Age at t0 (per 10 years higher) | 0.022 | 0.017–0.027 | <0.001 | |
Female | 0.004 | −0.006 to 0.014 | 0.42 | |
Black race | −0.019 | −0.04–0.002 | 0.07 | |
Transmission by IDU | −0.011 | −0.024 to 0.003 | 0.12 | |
CD4+ cell count model | Interceptb | −69.0 | −71.8 to −66.2 | <0.001 |
CD4+ cell count at t−1 (per 100 cells/μl higher) | −17.0 | −18.1 to −15.9 | <0.001 | |
Viral load at t0 (per 1 log10 copies/ml higher) | −37.6 | −39.6 to −35.7 | <0.001 | |
Age at t0 (per 10 years higher) | −1.7 | −3.3 to −0.03 | 0.046 | |
Female | −0.8 | −4.2 to 2.7 | 0.67 | |
Black race | 2.3 | −3.8 to 8.3 | 0.48 | |
Transmission by IDU | 0.7 | −4.6 to 6.0 | 0.80 |
Annualized change in viral load (copies/ml, log10 scale) and CD4+ cell count (cells/μl, absolute scale) was defined as [(measurement at t1 – measurement at t0) × 365]/(t1 – t0). All continuous variables were centred, see a,bbelow), estimate sizes and 95% confidence intervals (CI) from a multivariable linear regression model. IDU, injecting drug use.
aThe intercept in the viral load model is the annual change in viral load for an individual with viral load at t−1 = 4 log10 copies/ml, CD4+ cell count at t0 = 500 cells/μl, age at t0 = 35 years, male, white race and transmission mode other than by IDU.
bThe intercept in the CD4+ cell count model is the annual change in CD4+ cell count for an individual with viral load at t0 = 4 log10 copies/ml, CD4+ cell count at t-1 = 500 cells/μl, age at t0 = 35 years, male, white race and transmission mode other than by IDU.
Other comments: The reference race group is white race. All other people who were not classified as black or white race were also included in the model but in a separate category (not shown).