Lack of HIV RNA test result is a barrier to breastfeeding among women living with HIV in Botswana

Feeding method in the first 6 months of life plays an integral role in a child’s health and development trajectory [1]. Breastfeeding is one of the foundations of child health and development and is associated with substantially lower child morbidity and mortality, including among HIV-exposed children [2, 3]. The 2010 and 2016 World Health Organization (WHO) guidelines on HIV and infant feeding define a primary goal of maximizing infant HIV-free survival [2‐4] and support breastfeeding in the setting of maternal antiretroviral therapy (ART) or infant antiviral prophylaxis. In low-resource settings where diarrhoeal diseases, pneumonia and malnutrition are common causes of infant morbidity and mortality, WHO recommends exclusive breastfeeding for 6 months and continued breastfeeding up to 24 months with maternal or infant ARV protection in order to give infants the best chance for HIV-free survival [2, 5].

Despite outstanding ART coverage in pregnancy (95%), with 64.5% of WLWHIV conceiving on ART and very low (< 1.5%) mother-to-child HIV transmission (MTCT) [6, 7], Botswana has been slow to adopt a policy of universal breastfeeding for HIV-exposed infants. With over 20% of infants in the country HIV-exposed [8, 9] the lack of breastfeeding for these children contributes substantially to Botswana’s high under-five mortality (41.6 deaths per 1000 live births) compared to the Sustainable Development Goal target to reduce under-five mortality to at least as low as 25 deaths per 1000 live births [10, 11]. In recognition of this problem, Botswana chose to align its guidelines with those of the WHO in May 2016 to encourage greater breastfeeding for women living with HIV (WLHIV) on ART [12]. However, because Botswana can perform HIV RNA (viral load) testing as part of its ART program, the 2016 Botswana guidelines also specified that there should be documented HIV RNA suppression within the last 3 months of pregnancy to support a recommendation to breastfeed at the time of delivery. The updated guidelines also retain an individualised approach to feeding, recommending counselling during pregnancy to ensure each woman makes the best decision for her situation.

After decades of an MTCT prevention programme that relied on the use of formula, and with the additional requirement of documented viral suppression in the third trimester of pregnancy to support breastfeeding, it is currently unknown how the change in 2016 guidelines impacted actual feeding choices in Botswana. Therefore, we set out to describe feeding choices and potential barriers to breastfeeding among WLHIV in Botswana after the new guidelines were implemented.

We evaluated maternal feeding choice among WLHIV and identified ongoing low levels of breastfeeding following the 2016 guideline change in Botswana, and evidence that the requirement for a suppressed HIV RNA result during pregnancy may be a barrier to breastfeeding. To our knowledge, this is the first published data on feeding choices in Botswana after the new guidelines were implemented and in the ‘treat all’ ART era.

Breastfeeding among WLHIV in Botswana was low overall despite the guideline shift in 2016, with only 39.7% of women breastfeeding upon discharge from the maternity ward and an increase in breastfeeding of only 1.3% from 2016/2017 to 2018/2019. While this level of breastfeeding is higher than previous reports from Botswana before 2016 where fewer than 20% of WLHIV chose to breastfeed [14, 15], it is significantly lower than in other African settings where studies have reported choice to exclusively breastfeed by 80–90% of WLHIV [16, 17]. Most African countries with the highest HIV prevalence have adopted a public health approach rather than an individualized approach to support breastfeeding for WLHIV [18, 19]. In 2011, Botswana’s neighbouring country, South Africa, committed to promoting exclusive breastfeeding for WLHIV following the WHO 2010 guidelines on HIV and infant feeding and ended the universal formula programme for WLHIV. In contrast, Botswana continues to provide infant formula free-of-charge for all WLHIV who choose to formula feed.

Our hypothesis that the requirement for documented HIV RNA suppression in pregnancy served as a barrier to breastfeeding was confirmed. Only 25.7% of women had an HIV RNA (viral load) result documented in the obstetric record overall, and the presence of a result (95% of which were virally suppressed) did correlate with the choice to breastfeed. This low availability was only partially reflective of actual HIV RNA testing in pregnancy, as our electronic record verification identified that 62% of women with no documented HIV RNA result on the obstetric record did have an HIV RNA test performed during pregnancy. Thus, one implication of our findings is that improved documentation of HIV RNA results in the obstetric record (or improved access of maternity nurses to the electronic records) could improve breastfeeding uptake. However, we also noted an increase in available HIV RNA results over time that was larger than the increase in breastfeeding, suggesting that other factors may have also played a role in the reluctance to choose to breastfeed. Feeding decisions may have been influenced by earlier guidance in Botswana which emphasized formula feeding, fear of MTCT [20], ambiguous/confusing messages regarding infant feeding [20, 21] and lack of social supports for breastfeeding among WLHIV [22]. In addition, while most HIV RNA results were undetectable (< 400 copies/mL), some women without a result or with unsuppressed HIV RNA still chose to breastfeed, indicating that factors beyond MTCT risk were likely involved in feeding decisions.

We identified additional factors associated with lack of breastfeeding in our multivariable analysis, most of which were consistent with prior studies [20, 23]. Lack of prior antenatal visits was a strong risk factor, which may have been related to lack of a possible HIV RNA result in pregnancy or differences in counselling practices for these women. Women who delivered in a tertiary hospital were less likely to breastfeed compared with women who delivered in a primary/district hospital, but it is not known whether this was related to differences in patient populations at these tertiary centres, counselling differences, or other factors associated with urban locations. We also found age, education, nationality, and ART status during pregnancy to be associated with feeding choice.

Our study had several limitations. We were only able to evaluate feeding method at the time of maternity ward discharge. The decision to breastfeed at the time of discharge from maternity does not equate with exclusive breastfeeding for the WHO recommended 6 months. A prospective study would be needed to examine breastfeeding patterns throughout infancy. While there is a high percentage of viral suppression in Botswana overall (85%) [6], the 95% suppression in this report may be artificially high if women who tested were those most motivated by PMTCT concerns. However, multiple prior studies in Botswana support this high level of suppression in pregnancy, and we believe this indicates a true low MTCT risk for most breastfeeding women [5, 14]. Strategies that evaluate time on ART and adherence to ART that can be performed by maternity nurses throughout ANC and at the time when infant feeding decisions occur may be a useful alternative to reliance on HIV RNA documentation.

Despite a guideline shift in 2016 designed to support breastfeeding among WLHIV, fewer than 40% of WLHIV in Botswana from 2016 to 2019 chose to breastfeed their infants. This low rate of breastfeeding underscores the need to address barriers affecting feeding choices, including the requirement for documented HIV RNA suppression. While additional barriers to breastfeeding exist and should be explored, in a setting of long-standing use of suppressive ART with almost two-thirds of WLHIV conceiving on ART and high overall suppression, the requirement for documented HIV RNA suppression is unlikely to prevent a substantial amount of HIV transmission during breastfeeding but likely contributes to low rates of breastfeeding.