Table 3.
Association of HIV status, sex, age at vaccination, nutritional status, and maternal CD4 T-cell count with log measles IgG concentration
Infant characteristic | Log IgG concn (log mIU/ml) (mean ± SE) | Univariate difference in log IgG concn (95% CI)a | P | Multivariatea,b difference in log IgG concn (95% CI) | P |
---|---|---|---|---|---|
All infants (n = 236) | |||||
HIV status | |||||
HIV exposed | 5.67 ± 0.84 | Ref. | Ref. | ||
HIV infected | 5.46 ± 1.02 | −0.21 (−0.56 to 0.14) | 0.248 | −0.28 (−0.59 to 0.04) | 0.083 |
HIV-exposed infants (n = 201) | |||||
Sex | |||||
Female | 5.67 ± 0.85 | Ref. | Ref. | ||
Male | 5.67 ± 0.83 | 0.00 (−0.23 to 0.23) | 0.985 | −0.04 (−0.27 to 0.20) | 0.763 |
Age at vaccination (mo) | |||||
8.5–10 | 5.61 ± 0.79 | Ref. | Ref. | ||
10–11 | 5.80 ± 0.81 | 0.19 (−0.05 to 0.44) | 0.962 | 0.23 (−0.02 to 0.48) | 0.076 |
11–12 | 5.62 ± 1.13 | 0.01 (−0.46 to 0.49) | 0.120 | 0.08 (−0.38 to 0.53) | 0.742 |
Any breastfeeding at 8 mo | |||||
No | 5.67 ± 0.81 | Ref. | Ref. | ||
Yes | 5.70 ± 1.22 | 0.04 (−0.59 to 0.67) | 0.908 | -0.02 (−0.64 to 0.60) | 0.954 |
Stunting at vaccination | |||||
No | 5.70 ± 0.87 | Ref. | Ref. | ||
Yes | 5.44 ± 0.60 | −0.26 (−0.52 to 0.00) | 0.047 | −0.34 (−0.60 to −0.07) | 0.012 |
Wasting at vaccination | |||||
No | 5.63 ± 0.85 | Ref. | Ref. | ||
Yes | 5.94 ± 0.76 | 0.31 (0.01 to 0.61) | 0.046 | 0.29 (−0.01 to 0.60) | 0.062 |
Underweight at vaccination | |||||
No | 5.63 ± 0.81 | Ref. | Ref. | ||
Yes | 5.86 ± 0.95 | 0.23 (−0.10 to 0.57) | 0.175 | 0.23 (−0.11 to 0.56) | 0.184 |
Maternal CD4 T-cell count at randomization (cells/μl) | |||||
<200 | 5.78 ± 0.37 | 0.08 (−0.14 to 0.31) | 0.470 | 0.04 (−0.21 to 0.27) | 0.756 |
200–350 | 5.67 ± 0.91 | −0.03 (−0.38 to 0.32) | 0.876 | −0.01 (−0.33 to 0.32) | 0.975 |
>350 | 5.70 ± 0.89 | Ref. | Ref. | ||
Mother received HAART during pregnancy | |||||
Yes | 5.67 ± 0.85 | 0.00 (−0.28 to 0.28) | 0.995 | 0.00 (−0.27 to 0.27) | 0.986 |
No | 5.67 ± 0.82 | Ref. | Ref. |
Differences in log measles IgG concentration were estimated from linear regression models with robust empirical variance. Ref. indicates the reference group.
Multivariate models for HIV status among all infants included sex, age at vaccination, any breastfeeding at 8 months, maternal CD4 T-cell count, and maternal receipt of HAART as covariates. Multivariate models among HIV-exposed infants included sex, age at vaccination, any breastfeeding at 8 months, maternal CD4 T-cell count, maternal receipt of HAART, and a single indicator of nutritional status (stunting, wasting, or underweight). Each nutritional status indicator was entered into a separate multivariate model.