Table 3.
Participant Characteristics | Univariable OR (95% CI) | Full multivariable OR (95% CI) | Final multivariable OR (95% CI) |
---|---|---|---|
Age, years | |||
≤ 29 | 1.78 (0.60–5.25) | 2.62 (0.57–12.01) | 2.55 (0.86–7.54) |
30–39 | 1.72 (0.87–3.39) | 3.08 (1.16–8.18)* | 2.32 (1.16–4.63)* |
40–49 | 1.59 (1.00–2.52)* | 2.07 (1.08–3.98)* | 1.66 (1.03–2.67)* |
≥ 50 | Referent | Referent | Referent |
Race/Ethnicity | |||
White, non-Hispanic/Latino | Referent | Referent | Referent |
Black, non-Hispanic/Latino | 0.48 (0.28–0.81)** | 0.39 (0.17–0.91)* | 0.51 (0.29–0.87)* |
Hispanic/Latino | 0.67 (0.34–1.34) | 0.64 (0.22–1.84) | 0.68 (0.34–1.37) |
Other/Unknown | 0.88 (0.29–2.63) | 0.79 (0.15–4.23) | 0.71 (0.23–2.24) |
Insurance | |||
Private | 1.30 (0.81–2.10) | 0.83 (0.38–1.83) | |
Public/None | Referent | Referent | |
CD4 cell count (cells/mm3) | |||
< 200 | 0.63 (0.29–1.39) | 0.53 (0.17–1.63) | |
200–349 | 0.82 (0.47–1.45) | 0.70 (0.32–1.54) | |
350–499 | 1.03 (0.62–1.72) | 1.09 (0.55–2.18) | |
≥ 500 | referent | referent | |
Unknown/Missing | 0.58 (0.18–1.90) | † | |
HIV RNA viral load (copies/mL) | |||
Undetectable | 0.95 (0.61–1.47) | 0.91 (0.50–1.68) | |
Detectable | referent | ||
Unknown/Missing | 0.65 (0.19–2.22) | † | |
Cocaine | 3.59 (1.21–10.63)* | 3.39 (1.12–10.31)* | 2.75 (1.27–5.96)* |
Poppers | 4.75 (2.15–10.50)** | 5.03 (1.63–15.53)* | 3.90 (1.72–8.84)* |
Injection drug use | 1.30 (0.38–4.52) | 1.09 (0.17–7.17) | |
Methamphetamine | 3.99 (1.10–14.53)* | 1.73 (0.20–15.25) | |
Club drugs | 8.24 (1.05–64.91)* | 1.12 (0.08–16.30) | |
Erectile dysfunction drugs | 4.47 (2.22–9.00)** | 3.66 (1.84–7.26)** | 2.71 (1.64–4.46)** |
Cannabis use | 1.76 (1.13–2.74)* | 1.69 (0.87–3.28) | |
Alcohol use | 1.57 (1.03–2.38)* | 1.33 (0.72–2.44) | |
Current cigarette smoker | 0.82 (0.53–1.26) | 0.68 (0.36–1.26) | |
Time since 3/15/2007, per year | 1.03 (0.94–1.12) | 1.03 (0.97–1.10) |
CI confidence interval; GLMM generalized linear mixed model analyses; OR odds ratio
Indicates P-value < 0.05 and
indicates P-value < 0.001
Participant numbers too small to calculate ORs and CIs in the full multivariable analyses