Maximizing Health Benefits

What would be the preferred HHS recommendation if the goal were to maximize benefits to WLHIV? The author reasons that WLHIV who have undetectable viral loads pose negligible risk to their newborns through breastfeeding. It has been well known that, ever since the advent of anti-retroviral treatment, viral suppression prevents transmission. Recent studies show that individuals who are adherent to effective treatment pose very low risks of viral transmission. If viral loads are undetectable, no mode of transmission presents a significant risk, including sexual, needle-borne, or perinatal. The conclusion, therefore, that WLHIV with undetectable viral loads who breast-feed pose negligible risk to their newborns is fully warranted.

Dr. Gross argues for what we will call the “ideal scenario,” that is a WLHIV who is fully adherent to treatment and has a zero detectable viral load. If this ideal scenario could be reliably achieved and sustainably maintained, it would provide a clear justification for changing extant public health recommendations. Yet in the real world, it is uncertain whether WLHIV can achieve and sustain undetectable viral loads. There are multiple reasons why a WLHIV may not achieve or maintain a state of undetectable viral load. While today's antiretroviral drugs are far more effective and durable than in the past, some WLHIV still struggle with drug resistance and incomplete suppression while adherent to treatment.

Helpfully, the period of breastfeeding is usually a relatively short 6-12 months during which adherence is critical. Nonetheless, some women may not be adherent during the full period during which they breast-feed. We know that in real life, medical care for people living with HIV is imperfect and that WLHIV cannot always sustain taking their medication. Dr. Gross notes studies showing a minority of WLHIV in the US were not retained in care and sustainably suppressed through delivery and post-partum period. Meanwhile, although pre-exposure prophylaxis (PrEP) is not a clear analogy, the data show that individuals often do not consistently take their medication to prevent transmission. Self-reported data, moreover, are unreliable, as viral testing reveals cases when patients incorrectly report adherence. Thus, even as WLHIV understand the importance of maintaining adherence, there will be a proportion of cases in which the infant is at risk of contracting HIV from breast milk.

Dr. Gross has accounted for this variable by suggesting ongoing clinical supervision, which is essential. She rightly notes the concern that current guidelines unintentionally encourage surreptitious breastfeeding without the benefit of honest clinical management. We would add to her proposal the imperative of regular use of viral load testing during the full period during which she breastfeeds. Indeed, one reason to consider changing HHS guidelines is to enable a conversation between clinicians and WLHIV about when and how frequently viral load tests should be done.

One critical problem would occur if, having started breastfeeding, the WLHIV becomes lost to reliable follow up. We suspect that current HHS recommendations are designed to avoid any possibility of viral transmission to infants. This is a highly cautious position that takes a population-based perspective. That is, while undoubtedly many WLHIV could safely breastfeed their babies, a small proportion will present a risk. Having a clear “no exceptions” recommendation leaves little room for tolerating risk.

Dr. Gross reasons that this small population-based risk is outweighed by the benefits of breastfeeding including reduction of SIDS and other childhood risks. Certainly, the beneficial effects of breastfeeding are well established, but more so in low-and middle-income countries than in high-income countries like the US.

The core question is this: How do we think about the loss of benefits from breastfeeding among a large portion of virally suppressed women who do not breastfeed, compared with the risk for some women of being not virally suppressed. Undoubtedly, if HHS guidelines expressly gave WLHIV realistic options, many women would breastfeed who would not otherwise do so under current guidelines. Increased risk compared to the status quo comes if unsuppressed WLHIV who do not currently breastfeed under restrictive HHS policies might do so if guidance were loosened. Under that calculus, it would be vital to better understand the value of breastfeeding compared with the small risk of transmission. The author has undertaken this balancing with care and rigor.

Patient Autonomy and Informed Consent

The author rightly believes that WLHIV have the right to make informed choices whether to breastfeed. While this proposition has wide support, the case for autonomy in this context is not entirely on solid footing. True, competent adults have near absolute rights to make informed choices affecting their own health and wellbeing. In this case, however, the decision directly affects the health and safety of others — namely, infants. Again, in the ideal scenario, there are few, if any, tradeoffs. A WLHIV with fully contained viral loads gains emotional wellbeing from breast-feeding, while her baby benefits from improved overall childhood outcomes without risking contraction of HIV. Yet, if the WLHIV is not able to maintain that ideal viral state (whether intentional or not), there is a risk to her infant. Arguably, a woman would not have a right to autonomy if it posed a meaningful risk to others, including her own baby.

There is another point of importance. A recommendation not to breastfeed does not coerce a woman to make any given decision. That is, simply because HHS advises a WLHIV not to breastfeed, does not mean the agency is requiring the woman to follow that advice. Of course, public health recommendations, and accompanying counseling, must make clear that the woman is still free to choose. A key problem with current guidelines is that WLHIV may feel pressured into bottle-feeding their infant, fearing that disregarding HHS advice could open them up to allegations of child abuse or neglect. Might social workers or neighbors contact child protective services if a WLHIV breastfeeds her child? Even if HHS clearly discloses the risk of breast-feeding, it must advise the woman that she can make her own choice. Given evidence on viral suppression, the guidelines would be on far firmer footing to note that for some women (who are virally suppressed) there is genuine equipoise about the right choice.

Consider an analogy, which admittedly is not a perfect one. When Zika virus disease (ZVD) was endemic in the Americas and present in the US, the World Health Organization advised women potentially exposed to ZVD to talk to their doctors about becoming pregnant. One of us (LOG) opposed this view because it did not clearly and objectively inform the woman of the real risk to her newborn. (The World Health Organization [WHO] subsequently changed its recommendation.) At the same time, the Centers for Disease Control and Prevention (CDC) from the outset recommended that women potentially exposed to ZVD postpone pregnancy, while acknowledging the need for case-by-case decision-making based on based on individual circumstances, clinical judgment, and an individualized assessment of risks and expected outcomes. Neither the WHO nor the CDC recommendation required a woman to make a particular reproductive choice. But the CDC recommendation more clearly offered sound public health advice.

Recall, that many women, especially if they are poor and lack affordable access to care, will not consult their physicians. For poor WLHIV, it is certainly conceivable that they would sometimes fail to undergo ongoing clinical testing and supervision. Public health agencies, therefore, have a responsibility to provide advice based on the best scientific evidence.

The question remains: what is the best way to maximize informed choice for WLHIV? In the ideal scenario with sustained undetectable viral loads, the public health advice should expressly include the option of breastfeeding. But in the real-world, a “hard” recommendation not to breastfeed could be viewed as erring on the side of safety.

A Justice Perspective

Dr. Gross makes a strong case that current HHS recommendations are unjust. She points to data showing that minority women and their infants suffer disproportionately from diseases breastfeeding may prevent, such as diabetes, cardiovascular disease, prematurity, and SIDS. Discouraging WLHIV from breastfeeding, she argues, could cause poorer health outcomes among disadvantaged classes, thereby “systematically reinforcing” health disparities based on race, poverty, and education. The author's equity argument is strong, but again it is fact-dependent. In the ideal scenario, the benefit-risk ratio from breastfeeding is clear. Given a large population of poor women of color who will be treatment adherent during the full course of pregnancy, denying them the opportunity to breastfeed would be detrimental to the WLHIV and their infants.

But how should we view this more likely scenario? That is, a high proportion of WLHIV sustain undetectable viral loads while breastfeeding, but a few cannot maintain viral suppression and could transmit the infection to their infants. Certainly, more people would benefit by adopting the author's recommendation, but for the few who are not virally suppressed, the harm to their newborns is high. Assuming that changing HHS recommendations could result in a small number of new HIV infections among children, the personal, social, and economic ramifications are steep. (If our proposals below were adopted, however, there is a reasonable chance that we could reduce the risk of HIV transmission through breastfeeding to near zero.)

Reforming DHHS Recommendations: A Proposal

Drawing from the available scientific evidence and ethical analysis, the most appropriate additional HHS guideline following the broad statement that breastfeeding is not recommended could be framed as follows:

By focusing on distinguishing the low risks experienced by WLHIV who are durably virally suppressed, the guidance could maximize the opportunities for autonomy and benefit/risk balancing while continuing the simple, safety-oriented guidance for WLHIV who are not, or are not sure, to avoid more high-risk breast feeding than occur under the status quo guidelines.

A full package of public health reforms would provide for the following:

1. Health professionals should be trained to identify WLHIV who can safely breastfeed based on their viral load status at birth and other clinical and social factors, several of which are identified in Dr. Gross's paper.

2. A risk profile and clinical tools should be developed to identify people who are good candidates for breastfeeding.

3. WLHIV who choose to breastfeed should receive increased adherence support and clinical monitoring, including consideration of more frequent viral load testing to ensure durable suppression.

4. WLHIV should be provided with counseling, clinical and social support, and home services that could enable them to safely breastfeed their children.

Dr. Gross's article should open a major public health and ethical debate about the efficacy and fairness of HHS breastfeeding recommendations. If HHS reforms its recommendation, as it should, then there should be ongoing research to better understand the risks and benefits of breastfeeding. Armed with empirical data, policy makers could fine tune public health recommendations and accompanying services to maximize the health, well-being, and rights of WLHIV.