Table 2.
Association of clinical and laboratory factors with in utero HIV transmission (PEPI-Malawi trial, Malawi, 2004–2009)*
| Factor | Univariate models | Multivariate model | ||||
|---|---|---|---|---|---|---|
| OR | 95%CI | P | OR | 95%CI | P | |
| Identified as recently infected using the multi-assay algorithm (yes vs. no) | 3.03 | 1.63, 5.63 | 0.0005 | 2.49 | 1.30, 4.78 | 0.006 |
|
| ||||||
| Age (per 10 year increase) | 0.63 | 0.45–0.89 | 0.009 | 0.66 | 0.43–0.93 | 0.02 |
|
| ||||||
| Parity (per pregnancy) a | 0.91 | 0.82, 1.01 | 0.09 | -- | -- | -- |
|
| ||||||
| CD4 cell count (per 100 cell increase) | 0.94 | 0.88, 1.00 | 0.04 | 0.98 | 0.92, 1.06 | 0.66 |
|
| ||||||
| Log10 HIV viral load (per log increase) | 2.03 | 1.65, 2.51 | <0.0001 | 2.01 | 1.60, 2.51 | <0.0001 |
|
| ||||||
| Infant gender (female vs. male) | 1.30 | 0.96, 1.77 | 0.09 | 1.28 | 0.94, 1.74 | 0.12 |
|
| ||||||
| Presentation at delivery (late vs. early presenter) b | 1.66 | 0.21, 13.36 | 0.63 | 1.84 | 0.22, 15.25 | 0.57 |
|
| ||||||
| Infant prophylaxis regimen (control vs. extended prophylaxis) c | 1.22 | 0.88, 1.70 | 0.24 | 1.17 | 0.84, 1.64 | 0.35 |
Logistic regression analysis. Statistically significant results are shown in bold. OR: odds ratios; CI: confidence intervals. All univariate models included 2,561 women except for the model for presenter status; presenter status was only available for 2,554 women; women who presented early in labor delivery received single dose nevirapine (sdNVP) prophylaxis, while women who presented late in labor did not receive sdNVP (see Methods). The multivariate model included 2,554 women. In utero transmission was defined as a positive HIV DNA test at birth. The association of variables such as the availability of running water and electricity, maternal body mass index (BMI) and infant birth weight with identification of women as recently infected using the multi-assay algorithm was not statistically significant (see Table 1). Those variables were also not significantly associated with in utero transmission of HIV when included in the multivariate models shown in Table 2 (data not shown).
All of the variables included in the univariate models were also included in the multivariate logistic regression model, with one exception: parity was not included in the multivariable model because it was correlated with age.
Women who presented early in labor delivery received single dose nevirapine (sdNVP) prophylaxis, while women who presented late in labor did not receive sdNVP (see Methods).
In this analysis, results from the extended nevirapine and the extended nevirapine plus zidovudine study arms were combined.