The high prevalence of serodiscordant relationships coupled with the high risk of
transmission within those relationships across sub-Saharan African countries makes these couples
a target for HIV prevention efforts. The development of appropriate and effective interventions
for HIV prevention in serodiscordant couples must be based on accurate assessments of sexual risk
behaviors, substance use, and fertility desires. My dissertation evaluates HIV acquisition and
transmission in heterosexual HIV serodiscordant couples in Lusaka, Zambia.
My first study analyzed the role that alcohol use plays on sexual behavior and HIV
acquisition in Zambian serodiscordant couples. Men who reported heavy drinking was associated
with increased odds of having =>1 outside sexual partnership (adjusted odds ratio [aOR]=2.02;
95% CI=1.58, 2.57), and condomless sex with main partner in the past 3-months (aOR=1.61; 95%
CI=1.28, 2.04). Women’s heavy drinking was associated with increased odds of having =>1
iii
outside partner (aOR=1.89, 95% CI=1.35, 2.64), and condomless sex with main partner in past 3-
months (aOR=1.54; 95% CI=1.31, 1.82). Women who reported being drunk daily or almost daily
was associated with increased HIV acquisition (aHR=3.71; 95% CI=0.90, 15.25). Men who
reported being drunk weekly or daily/almost daily was associated with increased HIV acquisition
(aHR=1.72; 95% CI= 1.01, 2.93), and transmission (aHR=1.42; 95% CI=0.97, 2.10).
My second study evaluated the incidence and predictors of HIV acquisition from an outside
partner in serodiscordant couples in Zambia. Forty-five unlinked HIV infections occurred among
women (1.85 per 100-couple-years [CY]) and 55 unlinked HIV infections occurred among men
(1.82 per 100 CY). Risk of female unlinked infection was associated with baseline female alcohol
consumption (aHR=5.44; 95% CI: 1.03, 28.73), recent genital ulcers and/or genital inflammation
(aHR=6.09; 95% CI: 2.72, 13.64 and aHR=11.92; 95% CI: 5.60, 25.37, respectively). Risk of male
unlinked HIV infection was associated with reporting being drunk weekly or daily/almost daily at
baseline (aHR=3.52; 95% CI=1.19, 10.46), recent genital inflammation (aHR=8.52; 95% CI: 3.82,
19.03), genital ulceration (aHR=4.27; 95% CI: 2.05, 8.89), self-reporting =>1 outside partner
(aHR=3.36; 95% CI: 1.53, 7.37).
My third study evaluated the effect of fertility desires on HIV acquisition among HIV
serodiscordant couples in Zambia. Among a sub-set of 1,029 serodiscordant couples, 311 agreed
that they wanted a child in the future (30.4%), and 368 agreed they did not want a child or did not
know (36.0%), and 344 couples disagreed about having a child (33.6%), of which in 212 couples
(61.6% of disagreeing couples) the man wanted a child but the woman did not, and in 132 couples
(38.3% of disagreeing couples) the woman wanted a child but the man did not. The adjusted risk
ratio for woman’s HIV acquisition was 2.06 (95% CI=1.40, 3.03) among women who wanted a
child, 1.75 (95% CI=1.07, 2.87) for men who wanted a child in the next 12-months, and 2.55 (95%
iv
CI=1.32, 4.93) among couples who agreed that they wanted a child compared to couples who
agreed they did not want a child.
In conclusion, heavy alcohol use plays an important role in linked and unlinked HIV
acquisition and transmission in serodiscordant couples. The strongest predictors of unlinked
infection in serodiscordant relationships were alcohol use, genital inflammation and ulceration
prior to infection. Further, women were particularly vulnerable when they or their partner wanted
to have a child. These findings will help improve the implementation of HIV prevention and safer
conception services among heterosexual serodiscordant couples in Southern Africa and beyond.