Breastfeeding in HIV-positive mothers under optimized conditions: ‘real-life’ results from a well-resourced healthcare setting

Introduction

Mother to child transmission (MTCT) of HIV can occur during pregnancy, delivery, or through breastfeeding. Without preventive measures, the risk of such transmission ranges from 15 to 45 % [1]. Advancements in antiretroviral therapy (ART) have significantly reduced this risk to below 1 % [2]. In a recent study from our clinic, the MTCT rate between 2014 and 2020 was 0.2 % [3]. Breastfeeding among mothers living with HIV (MLWH) has long been a topic of intense debate in resource-rich settings, as it involves weighing the well-documented benefits of breastfeeding against the potential risk of MTCT of the virus. The benefits of breastfeeding are widely documented and universally acknowledged [4]. Breastfeeding not only fosters maternal-infant bonding but also provides essential nutrients and antibodies crucial for the infant’s health [4]. Additionally, it positively influences neurological and cognitive development while strengthening the immune system [4]. The general recommendation for the population without HIV is exclusive breastfeeding for the first six months of life, followed by complementary feeding up to two years or longer, depending on the preferences of the mother and child. Before delivery, healthcare providers should counsel patients with HIV on the risk of transmission through breastfeeding. This risk is influenced by maternal viral load (VL), ART adherence, infant postnatal prophylaxis (PNP), and feeding practices [5]. One pivotal study, the PROMISE study in Sub-Saharan Africa and India, reported an MTCT risk of 0.3 % after six months of exclusive breastfeeding under ART, increasing to 0.7 % after 12 and 24 months [6]. Breastfeeding recommendations for MLWH vary globally, reflecting healthcare resource availability. In low-resource settings, where safe formula access is limited, exclusive breastfeeding remains essential. WHO recommends six months of exclusive breastfeeding with maternal ART and short-term infant PNP, followed by continued breastfeeding with complementary feeding up to 24 months if maternal ART adherence is maintained [7], 8]. In contrast, high-resource settings have seen a paradigm shift. Earlier guidelines discouraged breastfeeding due to MTCT risk, but recent studies indicate that under optimal conditions characterized by shared decision-making, close medical monitoring, regular infant HIV PCR testing, maternal ART adherence, and sustained undetectable VL – the risk approaches zero [5], 9], 10]. The German-Austrian guidelines permit breastfeeding under strict conditions, including a fully suppressed maternal VL (<50 copies/ml) throughout pregnancy and breastfeeding, consistent ART adherence, and monthly VL monitoring of both mother and infant [11]. Despite these advances, challenges remain. Comprehensive counseling is crucial to support ART adherence and informed feeding decisions. Complications such as mastitis, nipple injury, abrupt cessation, or mixed feeding may increase MTCT risk [12], [13], [14]. Postpartum ART adherence is another concern, with only 53 % of women consistently taking prescribed medications [15]. Additionally, ART drug concentrations in breast milk vary, and while typically low, potential toxicity in the infant cannot be entirely excluded [16], [17], [18], [19], [20], [21]. PNP remains standard in many resource-rich countries, initiated within 6–12 h after birth, even with undetectable maternal VL. In some cases, prophylaxis is omitted under optimal maternal and infant conditions [11], [22], [23], [24]. This study evaluates breastfeeding outcomes among MLWH in a resource-rich setting under optimal conditions, focusing on MTCT rates and the role of multidisciplinary care.

Materials and methods

Results

Participant characteristics

Between January 2017 and December 2023, a total of 409 WLWH were seen at the Charité – Universitätsmedizin Berlin, who gave birth to 416 infants, including seven pairs of twins. Of these, 77 mothers (18.8 %) chose to breastfeed and were included in the analysis. A total of 365 women (89.2 %) met the eligibility criteria for breastfeeding, while 38 women (9.3 %) did not qualify due to a viral load exceeding 50 copies/mL. Data on VL were missing for six women (1.5 %). The proportion of WLWH meeting the eligibility criteria for breastfeeding remained consistently high across the years, ranging from 85.4 to 93.1 % in Figure 1. However, the percentage of women who actually breastfed showed a clear upward trend. In 2017, only 3.2 % of eligible women breastfed, increasing gradually to 13.1 % in 2019 and 29.3 % in 2021. This trend continued, reaching 37.5 % in 2022 and 40.4 % in 2023 in Figure 1. The median age of breastfeeding participants was 34.5 years (IQR 6.5). 69 (89.6 %) were diagnosed with HIV prior to pregnancy, with a median time since diagnosis of 8.2 years (IQR 6.01), while 8 (10.4 %) received their diagnosis during the present pregnancy. Most participants were from Sub-Saharan African (49/77, 63.6 %), followed by Western European origin (22/77, 28.6 %). Participants from Eastern Europe/Russia, Asia, and the Americas each accounted for 2.6 % (2/77).

Figure 1:

The total number of mothers seen per year (blue bars), the number who met eligibility criteria for breastfeeding (grey bars), and the proportion of mothers who initiated breastfeeding (red line) are shown. While the proportion of women eligible for breastfeeding remained consistently high throughout the years (ranging from 85.4 to 93.1 %), the percentage of eligible women who chose to breastfeed increased steadily from 3.2 % in 2017 to 40.4 % in 2023.

At the time of delivery, all participants had undetectable HI-VL (<50 copies/mL) within the last two months, and the median CD4 cell count was 580.5/μL (IQR 240.0). The majority (62.9 %, 44/70) had CD4 cell counts>500/μL, with only one participant (1.4 %) below 200/μL. Data on CD4 cell counts were not available for seven participants. Mode of delivery was reported for 76 participants. Spontaneous vaginal delivery was the most common mode (40/75, 52.6 %), followed by cesarean section (31/75, 40.8 %) and vacuum extraction (5/75, 6.6 %). These data are summarized in Table 1.

Table 1:

Descriptive analyses of the cohort.

Characteristics of the breastfeeding mothers	Number of observations	%
n=77
Age, years, median (IQR)	77, 34.5 (6.5)
<20	0	0
20–27	5	6.5
28–35	42	54.5
>35	30	39.0
Initial diagnosis of HIV infection prior to pregnancy	77
Yes	69	89.6
No	8	10.4
Region of origin	77
Sub-saharan Africa	49	63.6
Western Europe	22	28.6
Eastern Europe/Russia	2	2.6
Asia	2	2.6
America (north-, middle- and south)	2	2.6
HIV viral load, cop/mL
<50	77	100
CD4 cell, count/μL – median (IQR)	70, 580.5 (240.0)
<200	1	1.4
≥200–350	5	7.1
≥350–500	20	28.6
>500	44	62.9
ART regime	77
2 NRTI’s+PI	13	16.9
2 NRTI’s+INI	31	40.3
2 NRTI’s+NNRTI	27	35.1
Other	6	7.8
ART before pregnancy	76
Yes	69	90.8
No	7	9.2
Mode of delivery	76
Spontaneous vaginal delivery	40	52.6
Cesarean section	31	40.8
Vaginal extraction	5	6.6
Midwifery care	69
Yes	40	58
No	29	42
Brestfeeding experience	72
Yes	38	52.8
No	34	47.2
Breastfeeding duration, months, median (IQR)	72	7.3 (7.6) range: 0.1–33.9 monthsa

1. IQR, interquartile range; MW, mean value; SD, standard deviation; HIV, human immunodeficiency virus; CD4, cluster of differentiation 4; ART, antiretroviral therapy; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; INI, integrase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; SVD, spontaneous vaginal delivery; CS, cesarean section; VE, vaginal extraction. ^aOf the two children still being breastfed at the end of the follow-up, one had been breastfeeding for 39 months and the other for 24 months. These durations reflect the length of breastfeeding up to the last follow-up visit.

Breastfeeding practices

Of the 77 participants, 39 (50.6 %) practiced exclusive breastfeeding for at least three months, while 32 (41.6 %) engaged in mixed feeding. The median breastfeeding duration was 7.3 months (IQR 7.6), with a range of 0.1–33.9 months. Seven women (9.1 %) breastfed for less than one month. Of the two children still being breastfed at the end of the follow-up, one had been breastfeeding for 39 months and the other for 24 months. Exclusive breastfeeding in the first three months was associated with significantly longer overall breastfeeding durations (median 10.0 months; IQR 7.4) compared to mixed feeding (median 6.2 months; IQR 6.3; p=0.001) (see Figure 2). Maternal breastfeeding experience had no impact on the likelihood of exclusive breastfeeding (p=0.779).

Figure 2:

Newborns who were exclusively breastfed for at least the first three months and did not receive any formula during that time were breastfed for significantly longer durations compared to those who were not exclusively breastfed (p=0.001). The median breastfeeding duration was higher among exclusively breastfed newborns. The boxplot highlights a broader variability in breastfeeding durations among newborns who were not exclusively breastfed, with notable outliers in both groups.

Impact of mode of delivery and midwifery care

The duration of breastfeeding was significantly associated with the mode of delivery (p=0.005). Women who had a spontaneous vaginal delivery breastfed for a median duration of approximately 10 months, whereas those who underwent a cesarean section had a median duration of around 5 months. The shortest breastfeeding durations were observed in the vacuum extraction group, with a median of approximately 3–4 months. Variability within groups is displayed in Figure 3.

Figure 3:

The duration of breastfeeding was significantly associated with the mode of delivery (p=0.005). Women who had a spontaneous vaginal delivery breastfed for a median duration of approximately 10 months, whereas those who underwent a cesarean section had a median duration of around 5 months. The shortest breastfeeding durations were observed in the vaginal extraction group, with a median of approximately 3–4 months. The variability within groups is displayed in Figure 2, with an outlier identified in the spontaneous vaginal delivery group.

Midwifery care was provided to 40 participants (58 %), and exclusive breastfeeding was significantly associated with midwifery care (p=0.009). Women who received midwifery support were more likely to practice exclusive breastfeeding compared to those without such care, as shown in Figure 4.

Figure 4:

Distribution of exclusive vs. non-exclusive breastfeeding among mothers with and without midwifery care. Among those who received midwifery support (n=40), the majority practiced exclusive breastfeeding (n=26). In contrast, among mothers without midwifery care (n=29), exclusive breastfeeding was less common (n=11). The association between midwifery care and exclusive breastfeeding was statistically significant (p=0.009).

Discussion

The proportion of breastfeeding mothers in our cohort increased from 3.2 % in 2017 to 40.4 % in 2023, reflecting a significant shift in clinical practice. This trend underscores the importance of healthcare providers staying informed about evolving evidence to support informed decision-making for MLWH. Notably, there was a decline in breastfeeding rates in 2020 in Figure 1, coinciding with the COVID-19 pandemic. This decrease may have been influenced by disruptions in healthcare services, changes in hospital policies, and increased uncertainty regarding breastfeeding safety during the pandemic [25]. However, following this period, breastfeeding rates rebounded, suggesting a recovery in clinical support and confidence in breastfeeding practices. This recovery underscores the critical role of sustained viral suppression and close medical monitoring in ensuring safe breastfeeding practices. It aligns with guidelines from high-resource settings that permit breastfeeding under specific conditions, emphasizing the need for continuous clinical support [26], [27], [28]. The decision to breastfeed in these settings requires shared decision-making between mothers and healthcare providers to ensure individualized care and optimal outcomes. The main challenge remains optimizing breastfeeding practices for MLWH to maximize benefits while minimizing the risk of MTCT. Achieving this goal necessitates strict ART adherence, tailored interventions, and multidisciplinary support. Breastfeeding remains a significant contributor to infant HIV infections in resource-limited settings, with maternal immunosuppression, low CD4 cell counts, and breast complications such as mastitis increasing transmission risk [1], [29], [30], [31], [32], [33]. Maternal VL is the most critical determinant in preventing transmission [34], 35]. However, the concept of “undetectable equals untransmittable” (U=U) may not fully apply to breastfeeding [36]. A systematic review reported an HIV transmission rate of 2.93 % after 12 months of breastfeeding under ART [37]. Similarly, the PROMISE study observed two transmission cases in breastfeeding mothers with VLs below 40 copies/mL, suggesting that persistent HIV exposure, local breast inflammation, or other unknown factors may contribute to residual transmission risk [6].

The importance of feeding practices is further underscored by data from a large South African intervention cohort study, which demonstrated that exclusively breastfed infants had a significantly lower risk of HIV transmission compared to those receiving mixed feeding. [38]. However, it is important to note that the mothers in this study were not receiving ART, which limits the generalizability of these findings to current clinical contexts, where maternal ART is standard [38]. In such settings, the influence of feeding practices on transmission risk is likely reduced. Nevertheless, these earlier findings provide insight into potential biological mechanisms that may still be relevant. In our own setting, where all mothers were on ART, no HIV transmissions occurred, regardless of feeding type. This supports the notion that while exclusive breastfeeding may reduce transmission risk in the absence of ART, effective maternal ART is the key protective factor in contemporary settings. One possible explanation is that exclusive breastfeeding supports gut barrier integrity, while mixed feeding especially the introduction of solid foods may increase intestinal permeability to HIV and other pathogens. This mechanism is particularly relevant in low-resource settings with increased exposure to contaminated water and infections. In contrast, in high-resource settings with access to clean water, high hygiene standards, and formula alternatives, this protective effect may be less significant. Another key concern is the increased risk of maternal breast complications such as mastitis, which has been linked to higher HIV shedding in breast milk [30], 38]. While this risk exists across all settings, midwifery-led lactation support in high-resource healthcare systems can play a crucial role in prevention and early intervention. Access to trained midwives providing tailored counseling, monitoring for early signs of inflammation, and ensuring optimal infant positioning may help reduce mastitis and related complications. Notably, in our cohort, exclusive breastfeeding was significantly associated with midwifery care (p=0.009), emphasizing the need to integrate midwifery-led support into breastfeeding recommendations for MLWH.

European guidelines on breastfeeding among MLWH remain inconsistent. The INSURE survey found that while 52 % of national guidelines still discourage breastfeeding for MLWH, 48 % allow it under specific conditions. Nearly half of respondents reported an increasing trend in breastfeeding among MLWH, highlighting the need for harmonized, evidence-based guidance across Europe [39]. Our cohort largely followed the German–Austrian Guidelines for HIV Therapy in Pregnancy and HIV-Exposed Newborns [11], which initially opposed breastfeeding but, since 2020, have supported shared decision-making under strict conditions, including sustained viral suppression, ART adherence, and close monitoring. However, eight mothers in our study, initiated breastfeeding despite receiving a new HIV diagnosis during pregnancy, and one mother was treated with a dual-drug regimen instead of the recommended three-drug regimen.

A key aspect to consider is how our cohort compares to other studies regarding breastfeeding duration. In our study, the median breastfeeding duration was 7.3 months (IQR 7.6 months), with a wide range from 0.1 to 33.9 months. Two children were still breastfeeding at the end of follow-up, one for 39 months and the other for 24 months.

When comparing this to the Swiss study (2019–2021), where 25 out of 41 ART-treated mothers breastfed with a median duration of 6.3 months (range 0.7–25.7 months), our cohort had a slightly longer median duration and a broader range, particularly with some mothers breastfeeding beyond two years [10]. The UK study presents a notably different pattern. The median breastfeeding duration was only 56 days (IQR: 23–140 days), with a range from 1 day to 2.3 years. Strikingly, 93.3 % of breastfeeding mothers in the UK cohort breastfed for 6 months or less, and 18.7 % (28 of 150) stopped within the first 7 days. This suggests that, compared to both our cohort and the Swiss study, breastfeeding durations in the UK tend to be significantly shorter. The data indicate that only a small minority of mothers breastfed beyond 6 months, whereas in our cohort, a substantial proportion breastfed for extended periods, with some continuing beyond two or even three years [40]. This suggests a much shorter breastfeeding duration compared to our cohort and the Swiss study. The differences could be attributed to variations in national guidelines, cultural attitudes toward breastfeeding among MLWH, healthcare provider recommendations, or access to lactation support. The relatively short breastfeeding duration in the UK cohort may indicate that many mothers discontinued breastfeeding early, possibly due to lack of support, or personal concerns regarding MTCT risk. However, as there is no information on the availability or continuity of postnatal breastfeeding support in this context, such interpretations should be made with caution.

This raises important questions about the long-term safety of ART exposure in breastfed infants, potential health benefits, and the need for extended clinical monitoring of both mother and child.

PNP protocols for HIV-exposed infants (HEI) also vary widely across Europe. The Penta survey highlighted significant discrepancies in risk stratification, drug selection, and prophylaxis duration. For low-risk infants, some countries do not recommend PNP, while others administer up to four weeks of ZDV. High-risk infants often receive a three-drug regimen (ZDV, lamivudine, and nevirapine), but regimen duration and dosing protocols differ [24]. In our cohort, nearly all breastfed infants (98.7 %) received ZDV regimen, though two mothers opted out of PNP after counseling. 24.7 % of infants received ZDV for longer than the recommended 14 days, possibly reflecting maternal uncertainty or misunderstanding despite prior counseling.The ongoing debate over whether PNP is always necessary for breastfed infants highlights a critical gap in current guidelines.

Despite the overall positive outcomes in our cohort, guideline-recommended monthly maternal VL monitoring during breastfeeding was not consistently implemented due to logistical challenges. Pediatricians and obstetricians cannot bill for adult blood draws, requiring MLWH to schedule separate infectious disease appointments, often necessitating travel with an infant. This raises questions about the necessity of frequent VL monitoring in ART-adherent women with sustained suppression. Instead, our approach focused on fostering trust, ensuring ART adherence, and providing social work support via phone. These real-world insights reflect the practical strengths and challenges of managing MLWH in high-resource settings, where patient-centered care plays a critical role.

Exclusive breastfeeding and longer breastfeeding duration were significantly associated with vaginal delivery compared to cesarean section or vacuum extraction (p=0.005), suggesting that delivery mode may influence breastfeeding outcomes due to maternal recovery, early skin-to-skin contact, and neonatal feeding readiness. However, our cohort size (n=77) limits generalizability. Additionally, we did not analyze ART concentrations in breast milk, preventing conclusions on potential toxicities or long-term effects. A systematic review of 24 pharmacokinetic studies suggests that breastfed infants are exposed to considerable levels of NRTIs and NNRTIs, though methodological heterogeneity limits direct comparisons [41].

Our findings emphasize the importance of interdisciplinary care, particularly midwifery-led support, in preventing MTCT and promoting safe breastfeeding among MLWH. Future research should address long-term ART exposure in breastfed infants, refine risk-adapted care models, and validate these findings in larger cohorts to inform evidence-based guidelines.