Breastfeeding Among People With Human Immunodeficiency Virus in North America: A Multisite Study

Abstract

Graphical Abstract

This graphical abstract is also available at Tidbit: https://tidbitapp.io/tidbits/breastfeeding-among-people-with-hiv-in-north-america-a-multisite-study.

Open in new tabDownload slide

The majority of evidence regarding the risks of human immunodeficiency virus (HIV) transmission via breastfeeding comes from resource-limited countries, where exclusive breastfeeding is recommended because of higher infant mortality rates due to diarrheal disease and malnutrition when formula is used [1]. Prior to the availability of antiretroviral therapy (ART), transmission risk via breastfeeding was estimated to be 16% [2]. With maternal ART, transmission risk decreased to 1%–5%, though maternal viral loads during breastfeeding were often unknown [3]. In the Promoting Maternal and Infant Survival Everywhere (PROMISE) study, in which postpartum maternal ART was compared to infant nevirapine [4, 5], 0.3% transmission via breastfeeding was documented in the first 6 months and 0.6% transmission at 12 months for women on ART. Two of the 8 transmissions in the maternal ART arm of the PROMISE study occurred in people with undetectable viral loads at the time of infant diagnosis, but these mothers had only become virally suppressed late in pregnancy. Partly in response to the evidence of a lower incidence in breastfeeding transmission than previously estimated, more people with HIV in resource-rich countries are expressing a desire to breastfeed and discussing the risks and benefits with their healthcare provider.

In North American countries, national guidelines have historically recommended replacement feeding (with formula or pasteurized donor breastmilk), due to the concern for lactational transmission of HIV. The rationale was that in areas of the world such as the United States (US), Canada, and Europe where formula feeding is available and is perceived as acceptable, feasible, affordable, sustainable, and safe, replacement feeding eliminates the risk of lactational HIV transmission [6]. While formula may be acceptable to many parents, others prefer breastmilk feeding due to concern about inadvertent disclosure of HIV status by going against community or family expectations, or due to perceived benefits for maternal and infant health and desire to bond with their baby via breastfeeding. Some have had the opportunity to breastfeed children born prior to the current pregnancy, including in settings where breastfeeding is encouraged for all persons regardless of HIV status, and wish to continue this practice upon the birth of their current infant [7–9]. The affordability, sustainability, and safety of formula has also been challenged recently with issues in the US due to lack of clean drinking water (eg, in Flint, Michigan [10] and Jackson, Mississippi [11]) and significant nationwide formula shortages that disproportionately affected low-income families [12].

Data regarding breastfeeding experiences among people with HIV in high-resource settings are lacking. In response to feedback from clinicians across the US and Canada who work with individuals with HIV who choose to breastfeed, our objectives were to characterize people who breastfeed, including their motivations, challenges, and facilitators, and duration of breastfeeding, and to describe institutional practices surrounding breastfeeding for patients with HIV, including counseling, infant prophylaxis, and infant and maternal monitoring.

In this document, the term “breastfeeding” is used for simplicity to describe feeding a child the parent's own milk (either direct feeding or with expressed milk). We recognize that other terminology, such as “chestfeeding,” may be preferred by some individuals. Gender-inclusive language is used throughout, except when reviewing data, where results are presented using the same terms used in the original publications, such as “pregnant women.”

METHODS

After obtaining approval from the Baylor College of Medicine Institutional Review Board (H-43415), sites were recruited via announcements at HIV conferences, the ReproID listserve sponsored by the University of California, San Francisco (a multidisciplinary group of 600 individuals interested in the clinical management of HIV in pregnancy), and word of mouth. Investigators from participating sites in the US and Canada obtained approval from their respective institutional review boards. Any person with HIV, of any age, who breastfed between 2014 and 2022 in the US or Canada and received their care at a participating institution was eligible for inclusion in the study. People were excluded from the study if no maternal or infant data were available.

Institutional policies were reported for the site by the senior site investigator. Retrospective chart review was used to collect maternal and neonatal demographic and HIV information, peripartum viral suppression data, postpartum engagement and retention in care, and breastfeeding characteristics and challenges. Research teams were provided with predefined lists of motivations and challenges and given the opportunity to select 1 or more choices and also add free text to comment on motivations and challenges to breastfeeding. De-identified data were entered directly into a secure electronic Research Electronic Data Capture (REDCap) database.

For this study, undetectable viral load is defined as either reported as “undetectable” or ≤40 copies/mL. We recognize that specific laboratory cutoffs differed among locations and over time during the study interval. Index pregnancy refers to the most recent pregnancy reviewed for this study. Mixed feeding is defined as giving the infant any other liquid (including formula) or solids in addition to breastmilk.

De-identified data were exported from the database and all analyses were performed using Excel. Number and percentages were used to describe the cohort; mean (standard deviation) and median (interquartile range) were used for normally and abnormally distributed data, respectively.

RESULTS

A total of 72 cases were collected from 11 sites in the US (8 sites, 44 cases) and Canada (3 sites, 28 cases). The largest number of cases reported from 1 site was 26 cases (Canada) and the fewest was 1 case (US).

Sixty-two percent of participants (n = 45) who breastfed were born in African countries, and all but 2 were planning on living in North America after delivery. Ninety-two percent of participants (n = 66) had been diagnosed with HIV prior to the index pregnancy, seventy-two percent (n = 52) had disclosed their HIV status to a partner, eighty-six percent (n = 62) were on ART prior to pregnancy, and ninety percent (n = 65) had an undetectable viral load at delivery (Table 1). Most were multiparous (n = 58 [81%]), had entered prenatal care in the first trimester (n = 37 [51%]), and attended at least 1 postpartum visit with both obstetrics (n = 56 [78%]) and an HIV primary care provider (n = 35 [49%]) in the first year postpartum (Table 2).

The most commonly reported reasons for choosing to breastfeed were health benefits to the baby, family/community expectations and fear of disclosure of HIV status, and parent–child bonding. Total duration of breastfeeding ranged from 1 day to 72 weeks, with a median of 24 weeks. Most people stated they exclusively breastfed. Almost a third of the cohort reported breastfeeding challenges, the most common of which was inadequate breastmilk supply. Other reported difficulties included nipple pain/damage, mastitis, and difficulty with infant latching (Table 2). Providers reported concerns about how to manage these challenges, especially how to manage elevations in maternal viral load while breastfeeding, when and how to wean, and how to care for infants born prematurely or with other medical conditions that require care away from the birthparent.

A variety of neonatal antiretroviral regimens were used in this cohort. In order of simplest to most complex, 12 infants received zidovudine alone for 4–6 weeks (similar to recommendations for non-breastfed infants), 7 received nevirapine monotherapy for 6 weeks or until weaning, 12 received zidovudine with or followed by nevirapine for 6 weeks, until weaning, or after weaning, 10 received full triple therapy for 6 weeks and then nevirapine monotherapy until full weaning, and 22 received triple therapy until after cessation of breastfeeding.

Maternal and infant monitoring also differed significantly by site. Institutions reported that maternal viral loads were generally tested every 1–2 months while breastfeeding; however, we lacked data to show how often testing was actually performed. For infants, some were tested at birth; most were tested at 2 weeks, 2 months, 4 months, and then every 1–2 months while being breastfed; some were then tested 1, 4, and 6 months after weaning. Sixty-eight infants (94%) had documented negative HIV polymerase chain reaction (PCR) results ≥6 weeks after complete cessation of breastfeeding; the remaining 4 neonates were lost to follow-up.

Seven of 11 sites had breastfeeding policies in place for patients with HIV. Most had involved pediatric infectious diseases specialists, obstetric providers, and/or HIV infectious disease specialists in the development of the recommendations. Five sites also included the legal and/or ethics departments in their policy development (Table 3).

DISCUSSION

This is the largest study to date of people with HIV in high-resource countries who opted to breastfeed. Two reports of breastfeeding among women with HIV in European countries have been published: 13 cases from Italy [13] and 30 from Germany [14]. Of note, since the start of data collection for this work, 3 sites that submitted data to this study have published descriptive information from 3 [15], 10 [16], and 8 [17] included participants, respectively. Data collected for this cohort were more extensive than that collected for previous case series, so the decision was made to include all cases in final analyses. Fifty-one of the 72 cases have not been included in any previous publications. Country of origin, duration of known HIV diagnosis prior to pregnancy, partner disclosure, motivations to breastfeed, viral loads on entry in care and at delivery, presence/absence of institutional policies, and obstetric and HIV primary care follow-up in the year after delivery have not been included in most of the previous publications.

Motivations to Breastfeed

In the current study, the most common reason for choosing to breastfeed was health benefits for the child, followed by family/community expectations, concern about HIV status disclosure, and parent bonding. More than half of people who chose to breastfeed came from African countries. Although almost all were living in the US or Canada during the index pregnancy, many lived within closely knit ethnic communities where the expectation is that they will breastfeed. Not breastfeeding is frequently perceived by community members as a proxy for having HIV, which remains deeply stigmatizing—as has been documented in other countries [9, 18].

Besides individual motivations, the potential health benefits can also be considered. Breastfeeding reduces infants’ risk of diabetes, asthma, and obesity and the mothers’ risk of diabetes, obesity, hypertension, and breast/ovarian cancers—all of which are more prevalent in low-income communities that are disproportionately affected by HIV [19]. Discouraging breastfeeding in people with HIV may contribute to the inequities of health so common in resource-rich settings.

Challenges to Breastfeeding Counseling

Exclusive Breastfeeding or Not?

Twenty-two percent (16 patients) reported mixed feeding to their providers. While we do not have specific information for the timing or quantity of mixed feeding during breastfeeding, we do know that nearly one-fifth of reported breastfeeding difficulties were related to inadequate breastmilk supply. A study from the 1990s (prior to the availability of ART) demonstrated a higher incidence of transmission with mixed feeding [20]. As a result, clinicians have been hesitant to allow any formula supplementation if a person with HIV wants to breastfeed. However, there is no evidence that mixed feeding increases transmission risk when the lactating parent is virally suppressed on ART. At least 1 case in this cohort with inadequate milk supply had an infant who was given formula and subsequently told they should no longer breastfeed. Whether, and how, to include mixed feeding in infant feeding discussions has been the subject of ongoing debate among experts in both countries.

Expanded Infant Prophylaxis or Not?

Another enigma for pediatricians has been what kind of prophylaxis (if any) beyond the standard 2 weeks of zidovudine should be administered to infants of breastfeeding parents and for how long. Based on studies from resource-limited settings, either maternal ART or infant prophylaxis with nevirapine has been recommended [4, 5, 21, 22]. Most global guidelines recommend maternal lifelong ART including periods during pregnancy and postpartum. However, in resource-rich countries pediatricians have been unsure whether and/or what kind of infant antiretroviral prophylaxis to advise, and management differs by country and by institution. Because of concerns about maternal adherence to ART postpartum [23, 24], some clinicians prefer to prescribe infant prophylaxis for the duration of breastfeeding. We saw similar variation in this North American cohort. No perinatal HIV transmissions have been reported in any currently published cohort in high-resource settings, but the question about whether and what protection infant prophylaxis provides remains unanswered.

Maternal and Infant Follow-up

Regarding maternal and infant follow-up after delivery, the majority of people in the study had at least 1 postpartum visit. Thirty-five (49%) were documented to have had at least 1 visit to an HIV primary care provider in the first year after delivery, which was likely an underestimate if the individual sought primary care outside of the reporting healthcare system. However, lack of maternal follow-up in the first year postpartum remains an issue [23, 25, 26] and strategies to encourage people to stay in care after delivery are especially important for those who choose to breastfeed. Enhanced parental and infant assessment, support, and counseling are crucial for maintaining safety during breastfeeding, and also for identifying issues that might increase transmission risk (eg, the development of cracked nipples or mastitis). The frequency of infant HIV testing and postpartum birthing parent viral load testing was variable. While the best model for follow-up is unclear, coordination between the infant and the breastfeeding parent's healthcare providers is essential.

Need for Guidance

The lack of evidence-based feeding recommendations results in many local programs developing individual approaches to support people with HIV who choose to breastfeed [27]. This is especially complicated because infant feeding requires collaboration among a large group of diverse specialties (prenatal care providers; pediatric care providers; adult and pediatric infectious disease specialists; labor and delivery, postpartum, and pediatric nurses; lactation consultants; doulas; social workers; and case management teams). Moreover, some institutions have concerns about liability and legal implications, which results in varied system support for such practices. Unfortunately, the lack of guidelines means that people with HIV are unlikely to receive standardized evidence-based support in achieving their infant feeding goals depending on where they receive care, and this can increase existing health disparities [19]. Our hope is that by collecting and presenting these data, and encouraging national and international discussion, standardized recommendations may help expand infant feeding options to all locations where parents with HIV and their infants receive care.

Strengths/Limitations

This is the largest observational study of breastfeeding among people with HIV in North America. It considers policies and practices from a wide variety of geographic locations in North America, unlike previous case series that have reported experiences from a single institution. The main limitation of this study is that it is retrospective and relies on information obtained from record review. Consequently, this cohort is limited to individuals who disclosed their feeding status, and to the circumscribed description of motivations, barriers, and facilitators described in the medical record. The rate of retention in care of birthing parents and infants may have been underestimated since some obtained postpartum follow-up outside of the participating institutions. No neonatal transmissions occurred among the 94% of infants for whom results were available ≥6 weeks after weaning. Throughout the study, birthing people and their infants were closely monitored by their HIV healthcare provider and by pediatric HIV specialists. These circumstances are not possible in all clinical settings, reducing generalizability. Finally, some of these cases have been published as case series since the start of data collection. Data presented here are more extensive, but care should be taken not to consider information from all published reports in an aggregate manner since there are overlapping cases [15–17].

CONCLUSIONS

Advocacy to support the infant feeding choices of people with HIV is progressively increasing, as is the number of people with HIV in high-resource countries who are choosing to breastfeed. People with HIV and their clinicians have started to rethink the risks of HIV transmission (low in virally suppressed people) and the benefits of breastfeeding (potentially high). We describe the current clinical practices surrounding breastfeeding and HIV in North America. At the institutional level, we found variability in breastfeeding policies, involved medical specialties, infant prophylaxis regimen and duration, and infant and maternal HIV testing schedules. For people with HIV, we highlight the challenges faced in weighing the potential risks of perinatal transmission with multiple family and community expectations, infant and maternal health concerns, and concerns about bonding with their children. Finally, this study highlights the relatively small numbers of patients with HIV who chose to breastfeed at any 1 location, and the need for further multisite studies to answer the many remaining questions regarding infant feeding choices among people with HIV.

Notes

Acknowledgments. The authors thank the individuals who took the time to obtain institutional approval to participate, find the requested survey information, and enter data for this project: Barbara Baker, Leah R. Harris, Latia Hickerson, Alison Livingston, Leah Harrison, Arifa Rahman, Klaudia Szczech, Suzanne Taillefer, Silvie Valois, and Kylie Wilson. The authors also acknowledge the patients and advocates who have bravely demanded evidence-based information and guidelines around feeding choices.

Financial support. I. B. was supported in part by Quebec Health Ministry, Fonds de recherche du Québec-Santé (FRQS). L. R. was supported by the National Institutes of Health (NIH) (grant numbers R01CA250850, R01 HD101352, and R01 AI34363) and the National Cancer Institute (grant numbers HHSN2612012000311 and HHSN2612012000351).

References

Author notes