Table 3. Associations between cytomegalovirus (CMV) disease and mannan-binding lectin (MBL)/ficolin levels.
Effect | Odds ratio | 95% CI |
A | ||
MBL, log10 ng/mL | 1.03 | 0.73–1.45 |
H-ficolin, log10 ng/mL | 0.89 | 0.26–3.03 |
L-ficolin, log10 ng/mL | 2.34 | 0.44–12.36 |
M-ficolin, log10 ng/mL | 0.66 | 0.28–1.52 |
B | ||
MBL, log10 ng/mL | 0.85 | 0.55–1.30 |
Age1, years | 1.03 | 0.67–1.59 |
Female gender | 2.13 | 0.84–5.41 |
HIV RNA, log10 copies/mL | 1.53 | 1.08–2.16 |
Duration of HIV-infection, years | 0.91 | 0.84–0.98 |
C | ||
Effect of MBL per log10 ng/mL | ||
CD4 T-cell count ≤20 cells/mm3 | 0.92 | 0.60–1.41 |
CD4 T-cell count >20 cells/mm3 | 1.26 | 0.69–2.28 |
D | ||
Effect of MBL per 100 ng/mL | ||
≤100 ng/mL | 0.71 | 0.12–4.15 |
100–500 ng/mL | 1.17 | 0.83–1.65 |
>500 ng/mL | 0.99 | 0.96–1.01 |
Abbreviations: MBL, mannan-binding lectin; CI, confidence interval.
Per 10 years increase.
A) Univariable conditional logistic regression models for MBL, H-ficolin (3), L-ficolin (2) and M-ficolin (1). B) Multivariable conditional logistic regression model for MBL adjusted for the effects of potentially confounding risk factors for CMV disease based on 89 case-control pairs without missing data. C) Conditional logistic regression model for MBL, investigating its relationship with CD4 T-cell count. D) Conditional logistic regression model with a linear spline for MBL.