Beginning as early as 2003, the first cohorts to follow HEU children reported increased morbidity and mortality when compared to HU children (8). While the overall mortality rate in studies of HEU infants varies (ranging from 4.6 to 18.7% in the African setting, see Table 1) (7–16), the majority of studies have demonstrated increased mortality among HEU vs. HU infants across all settings, with mortality rates ranging from twice to fourfold above HU controls. Moreover, it appears that the cause of mortality, when investigated, is predominantly infectious. Specifically, studies in Botswana and Durban, South Africa, demonstrated higher rates of treatment failure in HEU infants diagnosed with pneumonia when compared to HU infants, with higher associated mortality (17, 18). HEU infants also suffered higher mortality from invasive pneumococcal disease (IPD) when compared to HU infants (33.7 vs. 22.4%) in a South African surveillance study (19) and increased mortality from lower respiratory tract infection (OR: 2.1, CI: 1.1–3.8) compared to HU infants (20). The emerging pattern is one of increased mortality from infectious diseases, and primarily from respiratory illness, among HEU infants.

In addition to increased overall mortality, recent studies have reported increased rates of all-cause hospitalization among HEU children when compared to HU children. Among 825 HEU children in the European Collaborative Study, 25% had been hospitalized in the first 2 years of life, with a reported rate of 0.5 per 5 child-years (22). A study of 736 HEU infants in India found that 35% of HEU infants had been hospitalized within the first year of life, with an overall rate in infancy of 906 per 1000 person-years (PY) (23). Again in that study, the majority (56%) of hospitalizations were due to infectious diseases (primary three causes included acute gastroenteritis 18.6%, sepsis/meningitis 11.5%, and pneumonia 6.6%). This pattern of high incidence of hospitalization has also been observed in resource-rich settings. In Belgium, the incidence of severe infections was estimated at 16.8% HEU infant years (24). In France, the risk of serious infections during the first year of life was estimated at 9.3% in HEU children (25). In a Canadian cohort, a higher rate of hospitalization was seen among infants born to mothers with detectable viral load at delivery (31 per 100 PY) compared to mothers with undetectable viral load (10 per 100 PY) (p = 0.02) (26). In addition to hospitalization, increased health-care utilization among HEU infants has been described. In one of the largest longitudinal studies of HEU children in Zimbabwe, sick clinic visits were 1.2 times more likely among HEU compared to HU children (27). Again, where documented, the overwhelming cause of health-care visits/hospitalizations was due to infectious diseases. In a small prospective South African cohort, HEU infants were 3.6 times more likely to develop a severe infection requiring hospitalization in their first year of life compared to HU controls (28). In a laboratory-based surveillance study in South Africa, HEU infants were found to have twice the rate of hospitalization from IPD compared to HU infants (19). A prospective South African cohort study of children undergoing surgery observed significantly higher rates of postsurgical complications among HEU compared to HU controls (23.7 vs. 5.7%); these were most often surgical site complications (impaired wound healing, infection, breakdown) or sepsis (29). In a large series from the Caribbean, HEU infants were found to have three times the incidence of neonatal infections compared to HU infants (26 per 1000 vs. 8.1 per 1000) (30). Taken together, these studies indicate an increased rate of health-care visits, hospitalization, and severe infections in the first 2 years of life among HEU infants across all settings.

Amidst the reports of increased morbidity, mortality, and hospitalization among HEU children, a number of studies have documented the causative pathogens, whose presence may suggest specific areas of immune dysfunction (Table 2).

Concurrent with decreased protection from maternal antibodies, differences in early-life innate immune development between HEU and HU infants may predispose them to increased infectious disease morbidity, most commonly from viral infections (76). Innate immunity orchestrates the initial, non-specific response to pathogens, while shaping future adaptive responses to prevent or clear infections (3) While there is very limited data on innate immune development among HEU infants, studies have shown that the first year of life is a period of rapid change in innate immune response, with differences in the developing immune profile of HEU vs. HU infants.

While well-known phenomena of neonatal innate immune development, such as deficiency in α and β defensins or complement pathways, may contribute to increased susceptibility to fungal infections in infancy, these may be magnified in HEU infants through additional mechanisms. Comparisons of early-life innate immunity in HEU and HU infants have demonstrated altered secretion of immune-mediating cytokines, with decreased IL-12 production in HEU vs. HU infants (77). Upregulation of cell surface receptors, such as increased MHCII expression, has been identified on unstimulated HEU APCs, relative to HU APCs (78). Functional comparison of NK cell activity at 1 month of age also demonstrated an increase of an intermediate NK phenotype for activation and perforin expression in HEU vs. HU, which was no longer significant by 1 year of age (79). Studies have also demonstrated a hyperinflammatory innate immune profile among HEUs and a proinflammatory innate immune response to stimulation with pathogen-associated molecular patterns (PAMPs) early in life. In a prospective study of innate immune development in the first year of life among HEUs, HEU APCs responded more strongly than HU in both the proportion of responder cells, and quantity of cytokine produced per cell, with a strong pattern of hyperresponsive polyfunctional monocytes and classical dendritic cells. The majority of differences were observed between 2 weeks and 6 months of age and were largely in response to bacterial PAMPs (80).

Different ethnic groups (with varied genetic backgrounds by extrapolation) can exhibit varied innate immune responses (81, 82). More specifically, TLR polymorphisms are associated with heterogeneity of innate responses (83). Therefore, in comparing HEU and HU immune development, it is essential to consider any differences in the ethnic distribution of populations where HIV is more vs. less prevalent. Future studies are needed to examine relative contributions of potential etiological factors to changes in innate immune ontogeny, and how those changes in immune development contribute to early-life susceptibility to infectious disease (75). For example, there has been recent focus on vaccine-induced non-specific innate immunity that exhibits adaptive-like characteristics, called “trained immunity” (84). BCG vaccination-induced trained immunity is shown to be associated with drastic reductions (17–37%) in LRTIs (85); however, the effects on innate immune cells appear to be transient. Epigenetic programing of monocytes by H3K4 trimethylation were detected very early in the first year but diminished by 12 months post-BCG exposure (86). Similar to BCG-induced transient changes in innate immunity, altered intra uterine or early-life exposures may transiently impact HEU innate immunity early in life. We do not know if the hyperinflammatory state in HEU infants early in life negatively impacts non-specific “trained immunity” or specific vaccine-induced immunity.

The cause of the immunological changes seen among the HEU is not clear, though is likely multifactorial (Figure 1). First, maternal health status and viral control is strongly suspected to play a role. Multiple studies have observed an association between maternal viral load and infant immune parameters. A prospective study of HEU infants in Mozambique reported that despite similar median levels of naive, memory, and activated CD4 and CD8 T cell counts between 1-month-old HEU and HU infants, HEU infants born to mothers with a high viral load had reduced numbers of naive CD8 and increased numbers of memory CD8 T cells compared to infants born to mothers with a lower viral load (87). Differences related to maternal viral load have also been observed at the functional level. CBMCs of HEU infants, whose mothers had a controlled viral load, produced higher levels of the anti-inflammatory cytokine IL-10 and were functionally more similar to the HU control group, while infants born to mothers with detectable viral load had reduced IL-10 and significantly higher levels of TNF-α (63). Overall, lower maternal immune status has been shown to be associated with adverse clinical outcomes in their infants. A birth cohort in Zambia found that infants of mothers with CD4 counts <350 cells/mm³ had a higher risk of hospitalization and mortality compared to infants of mothers with CD4 >350 cells/mm³ (59). A European cohort found that lower maternal CD4 count was associated with increased risk of serious bacterial, but not viral infections (35); however, increased risk of congenital and early postnatal CMV infection was seen in infants of mothers with CD4 counts <200 vs. >200 cells/mm³ in a US cohort (39). In the Zvitambo trial reporting on health outcomes among HEU, the incidence of oral candidiasis was significantly increased among HEU infants born to mothers with CD4 counts <200 vs. >800 cells/mm³ (OR: 3.91 vs. 1.91, p < 0.05) (34). Finally, in the two cases of HEU infants with PCP reported in Texas, the maternal CD4 counts at time of delivery were 27 and 47 cells/mm³, respectively (29).

Combination antiretroviral therapy (cART) during pregnancy, while essential and life-saving, may paradoxically be a contributing factor to immunological changes seen in the HEU infants. The recommended treatment for all HIV-infected pregnant women includes two nucleoside reverse transcriptase inhibitors (NRTIs) along with a protease inhibitor, for the duration of pregnancy (88). NRTI drugs induce to different degrees mitochondrial dysfunction, due to their affinity for mitochondrial gamma DNA polymerase. This affinity can interfere with mitochondrial replication, resulting in mitochondrial DNA (mt DNA) depletion and dysfunction. Aberrant histological morphology of mitochondria, mt DNA mutations, alterations in mitochondrial DNA levels in cord blood mononuclear cells, aneuploidy in cord blood cells, downregulation of select pattern recognition receptor genes, and altered secretion of soluble markers of inflammation (89–92) have all been described in both non-human primates and neonates exposed in utero to NRTIs (93). In the Woman and Infants Transmission (WITS) cohort study from North America, HEU infants who had had exposure to ARVs had lower total lymphocyte counts and absolute CD4 counts at 0–2 months of age, and again at 6–25 months of age, when compared to HEU who were not exposed to ARVS. Maternal CD4 counts were also significantly associated with lower infant CD4 counts in both groups (60). In the French Perinatal Cohort study, any ARV exposure was associated with lower CD4 and CD8 T cell counts in these infants, and again maternal CD4 counts <500 cells/mm³ were associated with lower infant CD4 counts (94). World Health Organization (WHO) guidelines recommend exclusive breast-feeding and infant ARV prophylaxis for the duration of breast-feeding (95), though little is known to date about effects of long-term ARV exposure (i.e., during the entire duration of breast-feeding) on the infant’s developing immune system.

The role of breast-feeding in HEU infant susceptibility to infectious diseases is not clear. On the one hand, studies have consistently shown a benefit to exclusive breast-feeding vs. formula or mixed feeding in resource-limited settings in preventing infant morbidity and mortality (96–98). This is likely due to increased transfer of maternal immune factors, including antibodies, and reduced risk of infectious exposure from potentially contaminated water sources associated with formula feeding (9, 21, 99). On the other hand, in the context of maternal HIV infection, breast-feeding presents an increased risk of HIV transmission from mother to infant (100–102). Finding the balance between reducing risk of HIV transmission, while preserving the benefits of breast-feeding, is a programmatic challenge (103). In response to the accumulated data on morbidity related to infant feeding practices, particularly in the context of ARV therapy, WHO guidelines on HIV and infant feeding practices were updated in 2010. They remain consistent with the guidelines of 2006, which recommended exclusive breast-feeding for HIV-infected mothers unless replacement feeding is acceptable, feasible, affordable, sustainable, and safe, in which case avoidance of all breast-feeding by HIV-infected mothers is recommended (104). This has resulted in different practices in resource-rich and resource-limited settings, complicating our understanding of how this may impact the developing infant’s immune system.

In resource-rich settings where exclusive formula feeding is recommended, infants may be at risk for infectious diseases in infancy due to reduced protective maternal antibodies (105, 106). However, this may be counterbalanced by expanded immunization coverage in these settings and, more importantly, increased access to health care and health services, which can mitigate this risk. In resource-limited settings, however, the recommendations for exclusive breast-feeding likely protect infants from infectious diseases-associated morbidity and mortality. Breast milk contains compounds that modulate PRR-mediated immune responses, including immunoglobulins, cytokines, antimicrobial proteins/peptides, nucleotides, and oligosaccharides (107–110). Breast milk immune composition is impacted by maternal nutritional status, exclusivity and duration of breast-feeding, and maternal or infant infections (110, 111), and maternal HIV-seropositivity does not appear to be an independent determinant of altered breast milk immune composition or quality (112). To counterbalance the increased risk of HIV acquisition during breast-feeding, the WHO recommends that all mothers remain on ARV therapy during the duration of breast-feeding (Option B+) and recommends daily antiretroviral prophylaxis for the infants in order to reduce the transmission of HIV. The protection provided through breast-feeding must be balanced against the risks associated with increased ARV exposure, and, most importantly, postnatal HIV acquisition.

Finally, an important factor in the shaping of the innate immune system is the environment (and associated pathogens) to which a person is exposed. HEU infants may also be exposed to a different variety of pathogens in the household, as immunosuppressed HIV-positive individual(s) in the home carry a heavier burden of disease, and potentially more opportunistic pathogens (13). HEU infants may therefore have a different innate imprint from HU infants in the same setting due to increased microbial exposure in early life. As an example, increased lipopolysaccharide has been identified in the serum of infants who lack access to running water in the household (113), which can lead to chronic inflammation in the gastrointestinal system, and a gut that is more prone to bacterial translocation. Increased exposure to lipopolysaccharide can lead to immune tolerance of endotoxin; indeed, lacking access to running water correlates positively with expression of IL-10 in childhood (114). Determining if the transient diminished capacity to contain infection is mediated at least in part by non-specific environmental exposure is an important next step toward understanding morbidity and mortality in HEU infants.

Across multiple cohort settings, HEU infants have been shown to have increased morbidity and mortality when compared to HU infants, and specifically higher rates of infectious diseases. While the studies reporting on microbiological diagnosis have been predominantly from South Africa, Belgium, and France, the overall reports of morbidity and mortality from infectious diseases have spanned different treatment eras (prior to and after widespread ARV availability) and geographical regions, suggesting that this is most likely related to infrastructure, research capacity, and expertise in these settings, rather than a specific geographical cluster of disease. The most commonly reported infections include invasive bacterial infections with encapsulated organisms, upper and lower respiratory tract infections, and diarrheal illnesses, most often within the first year of life, with additional case reports of PJP pneumonia. Of note, there have not been reports of increased prevalence of vaccine preventable diseases, including measles or disseminated TB, often common contributors to childhood mortality in resource-limited settings. Taken together, these findings suggest that there is likely an immunological basis for the increased susceptibility of HEU to infectious diseases in infancy.

The most consistently demonstrated difference in immune status between HEU and HU infants has been decreased maternal antibody transfer, which likely explains the increased susceptibility to encapsulated organisms in early life. However, there have been no direct studies correlating lower protective antibody levels and risk of, e.g., sepsis among HEU infants. At the same time, additional mechanisms that could explain this susceptibility to encapsulated organisms have not yet been studied, including complement deficiencies and splenic function, which may increase their susceptibility to these bacterial pathogens. Given that the susceptibility to these infections is predominantly seen during infancy and not later in life, this suggests that the humoral antibody response of HEU infants is preserved. The mechanisms that may underlie their susceptibility to opportunistic pathogens (PJP), chronic viral infections (CMV, HCV), and possibly acute viral infections (upper and lower respiratory tract infections, viral gastroenteritis), is less clear. There is some evidence of decreases in T lymphocyte number and function; however, these changes appear to be limited to select HEU infants, such as those exposed to high levels of maternal viral load or specific antiretroviral regimens. HEU infants have been shown to have a more experienced immune profile and innate immune responses, but it is not known how this may lead to increased susceptibility to disease.

In this review, we have attempted to correlate the reported clinical with the immunological findings among HEU infants. From this review, it becomes clear that there remains a significant knowledge gap in understanding the immunological basis for the increased susceptibility to infectious diseases in HEU vs. HU infants. While a number of clinical, epidemiological, and surveillance-based studies have reported on infectious outcomes, these same studies have not been able to extensively study the immune profile of affected infants, likely due to the challenges in obtaining specimens and conducting immunological assays in resource-limited settings. At the same time, while a number of detailed immunological assays on a small cohort of HEU children (predominantly in resource-rich settings) have been done, there are few correlates to clinical outcomes, given the relatively small number of HEUs that become severely ill in these settings. Future work needs to be directed at directly correlating infectious disease outcomes to immunological findings among HEU infants. This would best be achieved through prospective cohort studies where detailed microbiological and immunological data would be collected in a sufficiently large number of HEU infants, with concurrent collection of detailed data on maternal immune status, in utero and postnatal drug exposure, and feeding patterns, in order to identify potential confounders and the relative contribution of each of these factors.

With an estimated 1.3 million HEU children born annually, their increased susceptibility to infectious diseases and associated mortality in infancy suggests that where resources are limited, targeted interventions may be necessary for those at highest risk for adverse outcomes, such as those HEU infants born to mothers with advanced disease, highest immune suppression, and uncontrolled viremia. While standard of care in HEU follow-up varies across settings, specific health interventions could be envisioned. These include expanded vaccination coverage (including maternal immunization) to target the most identified causes of illness (pneumococcal vaccine, meningococcal vaccine, rotavirus vaccine), improved nutritional support and micronutrient supplementation, and potentially, antimicrobial prophylaxis during the first 6 months of life. A number of studies are currently underway to determine the efficacy of such interventions, including on the efficacy of antibiotic prophylaxis, the immunogenicity, and efficacy of additional vaccines in the HEU population.