Table 3. Multivariate analysis of risk factors for perinatal HIV transmission, western Kenya, 1996–2001.

	Adjusted relative risks (ARR) for perinatal HIV transmissiona
	All womenb, N = 454		Placental malaria–negative, n = 348		Placental malaria–positiveb, n = 107
	ARR (95% CI)	p	ARR (95% CI)	p	ARR (95% CI)	p
Log10 viral load	1.8 (1.6 to 2.1)	<0.001	1.7 (1.4 to 2.0)	<0.001	3.5 (2.5 to 4.8)	<0.001
Episiotomy or perineal tear	1.6 (1.2 to 2.1)	0.004	–		4.8 (2.3 to 9.7)	<0.001
Low birth weight	–		1.9 (1.1 to 3.2)	0.03	–
Gravidity <3 versus >3	–		1.8 (1.2 to 2.8)	0.003	–
Placental malaria status
Negative	Referencec		N/A
<10,000 parasites/μL	0.4 (0.2 to 0.6)b	<0.001	N/A		Reference
>10,000 parasites/μL	0.7 (0.3 to 21.5)b	NS	N/A		2.0 (1.1 to 3.9)	0.04

^a–, factor was not retained in the final model; CI, confidence interval; N/A, not applicable; NS, not significant.
^bA significant interaction was found between viral load and placental malaria density (p = 0.02) in these analyses. The effect of this interaction on the relative risk for placental malaria is shown in Figures 1 and 2. All relative risks given in this table do not include this interaction but give a weighted average of the placental malaria effect at various levels of viral load.
^cAn alternative model, in which placental malaria was fit as a binary variable (positive or negative), showed that placental malaria was protective for perinatal HIV transmission (relative risk 0.4, 95% CI 0.3 to 0.7, p < 0.001). In that model, log₁₀ viral load and episiotomy or perineal tear remain independent risk factors.