Impact of Maternal Viral Suppression on Growth Patterns for HIV-Exposed Uninfected Infants in Kenya

Background of the Study

With the success of prevention of vertical transmission (PVT) programs, the global population of children who are human immunodeficiency virus (HIV)-exposed but uninfected (HEU) has grown to nearly 16 million.^[1,2] While infants who are HEU do not acquire HIV, they are at risk for worse health outcomes compared to their unexposed peers.^[3,4] Studies have shown higher rates of comorbid conditions and mortality for children who are HEU compared to unexposed peers.^[3,5] A myriad of risk factors may contribute to these differences, including perinatal HIV exposure, exposure to antiretroviral therapy (ART), the challenges of being a member of an HIV-affected household, or a combination of these factors.^[3] One such risk factor, higher maternal viral loads, has been found to increase the risk for worse morbidity and mortality in children who are HEU.^[6]

Malnutrition is more commonly found in children who are HEU compared to those who are unexposed to HIV.^[7–10] Optimal nutrition in early childhood creates the critical foundation necessary to minimize mortality and morbidity over the lifespan.^[11,12] Globally, the deaths of children under five years of age are linked to undernutrition among 45% of the population,^[12] and children with malnutrition have worse cognitive development, independent of other potentially confounding factors.^[13] While the data are clear that children who are HEU have worse nutritional outcomes compared to their unexposed peers, the impact of maternal viral suppression on growth patterns for children who are HEU is not well understood.^[9,14]

Objectives and Hypothesis of the Study

In this study, we examined the impact of perinatal maternal viral suppression on the growth patterns of infants who are HEU and, secondarily, determined the rates of malnutrition in this population. We hypothesized that infants who are HEU and born to mothers who are not virally suppressed would have worse growth patterns and malnutrition rates compared to those born to mothers who are virally suppressed.

Study Design

We conducted a retrospective cohort study using electronic medical record (EMR) data from infants who were HEU and their mothers. This study was performed as a secondary analysis of data originally obtained for pregnant and postpartum women living with HIV,^[15] who were identified at the time of their first clinic visit, or previously, as having a positive test for HIV. The Institutional Research and Ethics Committee (IREC) approval at Moi University in Kenya was obtained. The study received a waiver for patient-level consent from the Moi IREC and was exempted from Indiana University Institutional Review Board because de-identified, routinely collected data were utilized for the analysis.

Study Setting

Infants and mothers jointly received integrated HIV care through the PVT services and associated maternal-child health clinics at Moi Teaching and Referral Hospital (MTRH), Eldoret, Kenya, from September 2015 to March 2019. For infants, this preventative treatment included HIV prophylaxis, growth monitoring, vaccinations, infectious disease prevention/treatment, and nutritional advice.^[16] Further, information about typical infant visits at MTRH during a similar time frame is detailed in Deathe et al., 2022.^[17] The HIV care and research infrastructure are a product of the Academic Model Providing Access to Healthcare (AMPATH) program, an over 30-year partnership between Moi University School of Medicine, MTRH, and a consortium of North American academic medical centers led by Indiana University School of Medicine.^[18] AMPATH contributes data to the East Africa international epidemiology databases to Evaluate AIDS consortium.^[19] All AMPATH facilities, including MTRH, provide standard-of-care HIV treatment services based on national guidelines.^[20]

Standards of Care at Time of Study

As per the World Health Organization (WHO) and Kenyan HIV treatment guidelines at the time of our study, first-line ART was efavirenz-based (typically tenofovir disoproxil fumarate + lamivudine + efavirenz) for women 15 years and older of childbearing age, including during pregnancy.^[21-23] Second-line ART was based on protease inhibitors (PIs) (atazanavir or lopinavir) plus ritonavir regimen. National guidelines recommended obtaining a viral load at the first antenatal visit (if the woman was already on ART), and then every six months thereafter (or every six months after ART initiation) while pregnant or breastfeeding, as long as the viral load remained <1,000 copies/mL. If the viral load was ≥1,000 copies/mL, additional counseling and support were provided to ensure excellent ART adherence,^[22] and the viral load was recommended to be repeated after three months. If the viral load remained ≥1,000 copies/mL at three months, a second-line regimen was initiated. HIV testing for infants was conducted within the first six weeks of life per the guidelines. If the first HIV DNA polymerase chain reaction testing was positive, the infant was to be presumed HIV-infected. If it was negative, the infant was considered HIV-exposed with repeat testing conducted at six months and 12 months of age; at 18 months of age, HIV antibody testing was completed according to the guidelines.^[22] The recommended ART prophylaxis initiated at birth was six weeks of zidovudine and nevirapine, with the latter recommended to continue until six weeks after cessation of breastfeeding.^[22]

Definitions

To determine malnutrition status, we converted child age, height, and weight into z-scores for weight-for-height (WFH), weight-for-age (WFA), and height-for-age (HFA), according to WHO’s reference means.^[24] Moderate-to-severe underweight, stunting, and wasting status were defined as z-scores ≤ −2 standard deviations (SD) for WFA, HFA, and WFH using WHO calculations.^[24,25]

Perinatal maternal viral suppression, henceforth referred to as maternal viral suppression, was defined as <1,000 copies/mL. If the mother had a viral load of ≥1,000 copies/mL at any point between 6 months before her first prenatal visit or 18 months after delivery, then she was considered not suppressed. Maternal ART adherence was determined by clinicians at each visit and categorized as good or poor.

Data Management

Clinical data from routine care visits were collected on paper-based forms and then entered by data clerks into the AMPATH EMR system. Recording errors were minimized by excluding infants with weight changes of >3 kg/month or height changes of >10 cm/month.^[26] Infants with heights that were less than previous recordings by ≥2 cm were also excluded from the study. Z-scores for WFA, HFA, or WFH that changed ≥2 units/month or had absolute z-scores of >6 units were also excluded from the study. These criteria were determined based on WHO growth charts and clinical judgment.

Statistical Analysis

Study characteristics were calculated as means and SD or percentages. t-tests were utilized to compare suppressed and not suppressed mother and infant characteristics. Percentages for underweight status, stunting, and wasting at 6, 12, and 18 months of age (with a range of 2 months allowed for each time point) were calculated, and we utilized Chi-square and Fisher’s exact tests. When we compared the percentages for malnutrition at these three time points, there was no adjustment made when using Chi-squared and Fisher’s exact tests. Graphic representations of WFA, HFA, and WFH z-scores from the first clinic visit to the last visit available were generated. We completed area under the curve analysis for WFA, HFA, and WFH to compare male and female sexes, as well as not suppressed and suppressed maternal viral load status. We also performed multivariate logistic regression to determine the adjusted odds ratios for moderate-to-severe underweight status, stunting, and wasting based on perinatal maternal viral suppression. Since infants were enrolled at different ages and due to extensive missing data for prematurity, we controlled for infant WFA z-score at first clinic visit, sex, maternal age at time of delivery, infant age at each time point, and ART before PVT enrollment. All analyses were completed in SAS Institute Inc. 2013. Base SAS® 9.4 Procedures Guide. Cary, NC: SAS Institute Inc. and R v 4.04. Statistical significance was set to α < 0.05.^[27]

Sociodemographic Characteristics

Infants were selected based on their mothers receiving integrated HIV care from an MTRH clinic between August 2015 and September 2017. Mothers were included in this cohort if they were living with HIV and had a measured viral load within six months before or after their delivery date. Infants were included if they were born at least 18 months (at the timing of confirmatory HIV testing) before database closure and had both height and weight measurements present in the EMR. Of the 845 potentially eligible women, 114 were not linked to an infant and were therefore excluded [Supplemental Figure 1]. This resulted in 735 infants linked to 731 mothers, including four sets of twins. As this study focuses on infants who were HIV-exposed but not infected, nine infants were excluded due to HIV infection. Data were not available on HIV status for 24 infants and perinatal viral suppression data were not available for 28 mothers. The final sample included 674 infants and 670 mothers.

Close

Figure 1:

Z-scores with age in years (a) for all HIV-exposed but uninfected infants, (b) by male and female sex, and (c) by maternal viral suppression. The shaded region represents the 95% confidence interval. WFA: Weight-for-age, HFA: Height-for-age, WFH: Weight-for-heightMagerko, et al.: HIV-Exposed Uninfected Growth Patterns in KenyaMagerko, et al.: HIV-Exposed Uninfected Growth Patterns in Kenya

Export to PPT

There were 674 infants who were HEU included in the final sample, with approximately half male (49%) and with a median age of 1.5 months (interquartile range [IQR] 1.4, 1.8) at first clinic visit and 16.4 months (IQR 15.1, 17.3) at last clinic visit [Table 1]. The mothers had a median maternal age of 32.2 (IQR 28.1, 36.3) at delivery; 87% were on ART before prevention of mother-to-child transmission (PMTCT) enrollment; and 98% were considered to have good adherence to ART. Mothers who were virally suppressed made up 85% of the population of dyads studied.

Main Variable Results

Significant differences were found between virally not suppressed and suppressed mothers for maternal ART regimen (P < 0.01). Mothers who were not suppressed were more likely to be on a PI-based regimen (i.e., lopinavir or atazanavir, 23% vs. 10%) and to have switched ART regimens during pregnancy (8% vs. 3%). Women who were not suppressed were also more likely to have initiated ART before PVT enrollment. Infant initial WFA, HFA, and WFH did not differ between not suppressed and suppressed maternal viral load groups, nor did the length of follow-up or mean number of clinic visits per month.

When comparing rates of moderate-to-severe malnutrition in infants at 6, 12, and 18 months of age-stratified by maternal viral suppression status, statistically significant differences were found at 6 months indicating underweight status (P < 0.05) for infants of non-suppressed mothers [Table 2]. The rates of underweight status at 12 months, stunting at 6 months, and wasting at 6 months suggested a trend toward higher rates of malnutrition among infants born to mothers who were not virally suppressed (P < 0.20).

Z-score plots are seen in Figure 1a for all infants who are HEU for malnutrition indicators including WFA, HFA, and WFH. The plots were further stratified by male and female sex [Figure 1b], and the test statistics when comparing the two groups were as follows: WFA was 0.03 (P = 0.01), HFA was 0.06 (P < 0.01), and WFH was 0.01 (P = 0.24). In Figure 1c, z-scores were examined by maternal viral suppression with significant differences seen in the area under the curve for WFA (test statistic = 0.07, P = 0.04), HFA (test statistic = 0.11, P = 0.01), and WFH (test statistic 0.06, P = 0.02).

Covariates Results

Infants had decreased odds of being stunted, underweight, and wasted if they had higher WFA z-scores at their first clinic visit (every unit increase in z-score was associated with 30%, 43%, and 61% lower odds of underweight, stunted status, and wasted status, respectively). Infants were also more likely to be underweight or stunted with each additional month of age. They had increased odds of being stunted if they were male [Table 3]. Maternal viral suppression and ART initiation Antenatal HIV Care enrollment were not associated with increased odds of moderate-severe malnutrition.

Strengths and Limitations of the Study

With this study’s observational nature, there are some limitations to note. Prematurity was not well-documented within the data set. WFA z-score at the initial visit, was utilized to account for possible differences due to prematurity, but a true measure of prematurity would have been preferred. In addition, our large sample had relatively few women who were virally non-suppressed (<15%), despite the broad criteria (viral load ≥1,000 copies/mL at any point between 6 months before pregnancy and 18 months post-partum), making it difficult to find meaningful patterns and associations within the data. However, this is a welcomed challenge and a testament to the success of their clinical care. Other important factors such as the specific infant antiretroviral prophylaxis, feeding type, number of siblings, spacing, geographic location, and socioeconomic status were unable to be assessed given the EMR data collection method. However, this large data set yielded important results to guide us in understanding the growth trajectories of children who are HEU.

Key Messages

• Infants born to non-virally suppressed mothers living with HIV are at increased risk for being underweight at six months.

• Male infants were more likely to be wasted or stunted across both groups.

• With each passing month, increase in age, infants had a higher odds of malnutrition.