Abstract
Objective:
Assess the risk of death for offspring of pregnant women living with HIV (PWLHIV) and the association with sociodemographic, pregnancy, HIV-related, and birth factors.
Design:
We conducted a prospective cohort study of PWLHIV on antiretroviral therapy (ART) and their offspring in urban Tanzania who were enrolled in a vitamin D trial conducted from June 2015 to October 2019.
Methods:
We described rates of fetal, neonatal, and infant death and assessed risk factors for these outcomes with generalized estimating equations. We also estimated population-attributable risk percentages for the contribution of prematurity and small-for-gestational age (SGA) to neonatal and infant mortality.
Results:
Among 2,299 PWLHIV, there were a total of 136 fetal deaths (5.6%) and the stillbirth rate was 42.0 per 1,000 total births. Among 2,167 livebirths, there were 57 neonatal deaths (26.3 per 1,000 livebirths) and 114 infant deaths (52.6 per 1,000 livebirths). Twin birth was associated with neonatal death, while maternal CD4 T-cell count <350 cells/μL in pregnancy was associated with infant death (p-values <0.05). As compared to term-appropriate-for-gestational age (AGA) births, the relative risks for neonatal mortality for term-SGA, preterm-AGA, and preterm-SGA infants were 2.07 (95% CI: 1.00–4.28), 2.87 (95% CI 1.54–5.35) and 7.15 (95% CI: 2.11–24.30), respectively. We estimated that 42.7% of neonatal and 29.4% of infant deaths were attributable to prematurity and SGA in the cohort.
Conclusions:
The risk of death is high for offspring of PWLHIV in Tanzania and the combination of prematurity and fetal growth restriction may account for nearly half of neonatal deaths.
Keywords: HIV, Fetal Death, Neonatal Death, Infant Death, Preterm Birth, Adverse Birth Outcomes
Introduction
Maternal HIV infection has added complex challenges to child survival since the beginning of the epidemic. In the absence of antiretroviral therapy (ART), there was a 15–30% risk of HIV transmission for infants born to pregnant women living with HIV (PWLHIV) as well as increased risk of multiple adverse birth outcomes including stillbirth, prematurity, and low birthweight (LBW)[1, 2]. Combination drug ART in pregnancy then markedly reduced the risk of vertical transmission of HIV, but the risk of adverse birth outcomes, including neonatal death, preterm birth, and small size at birth remains elevated in some settings[3, 4].
Individual, community, and health systems factors have been shown to increase the risk of fetal and infant death in HIV-uninfected populations[5–7]. In addition to these factors, PWLHIV can experience additional risks related to HIV infection and ART, comorbidities as well as social, economic, and environmental factors related to living in an HIV-affected household[1, 8]. Studies have identified that HIV-related factors like low CD4 T-cell count, high viral load, and advanced WHO HIV disease stage in pregnancy as well as selected ART regimens may increase the risk for some adverse birth outcomes[9–12]. There has also been growing global attention to the importance of “small and vulnerable newborns” (SVNs), infants born preterm (<37 weeks gestation), small-for-gestational-age (SGA; <10th percentile of birthweight-for-gestational age relative to a standard reference population), or LBW (<2500 grams), in child mortality[13]. However, few studies have examined the contribution of SVNs to neonatal and infant mortality among PWLHIV, particularly in the era of Option B+.
We conducted a prospective cohort study of PWLHIV in urban Tanzania to estimate rates of fetal death, stillbirth, neonatal, and infant mortality and assess the relationship with sociodemographic, pregnancy, and HIV-related factors and SVN phenotypes. We also estimated the percentage of neonatal and infant deaths attributable to the combination of preterm birth and SGA in the cohort.
Methods
We conducted a secondary analysis among pregnant women living with HIV and their infants who were enrolled in a randomized, double-blind placebo-controlled trial of vitamin D3 supplementation conducted in urban Dar es Salaam, Tanzania (clinicaltrials.gov: NCT02305927)[14, 15]. The study was approved by the Harvard T.H. Chan School of Public Health IRB, the Tanzanian National Health Research Ethics Sub-Committee, the Muhimbili University of Health and Allied Sciences IRB, and the Tanzania Food and Drugs Authority.
Pregnant women were recruited for the trial at five public facilities that provided antenatal care for pregnant women living with HIV. Under Option B+, all pregnant women living with HIV were initiated on lifelong ART and the first-line ART regimen was tenofovir (TDF)-lamivudine (3TC)-efavirenz (EFV). The trial inclusion criteria were (i) women ≥18 years, (ii) pregnant and in the second trimester (12–27 weeks gestation) by the reported date of the last menstrual period (iii) living with HIV, (iv) were receiving ART, and (v) a normal serum albumin-adjusted calcium (≤2.6 mmol/L). The group randomized to vitamin D3 supplementation received daily 3,000 IU vitamin D3 oral supplements (cholecalciferol) and the placebo group received a matching daily oral placebo supplement. There was no effect of maternal vitamin D3 supplementation on the primary outcomes of maternal HIV progression or death and infant small-for-gestational-age (SGA) birth or stunting, and the secondary outcomes of fetal death, stillbirth, neonatal death, or infant death[14].
Data collection
The detailed trial protocol and results are published elsewhere[14, 15]. Briefly, screening procedures were integrated into antenatal care and included HIV testing. At randomization, all participants received a clinical examination, WHO HIV disease stage assessment, and comorbidities documentation by a physician. Research nurses administered a questionnaire that included information on sociodemographic characteristics and measured maternal height and weight using standardized procedures. Follow-up study visits were conducted every four weeks during pregnancy and pregnant women were reminded to contact the study team at the time of labor. Study nurses/midwives attended labor and delivery to collect data on vital status signs, birthweight, and other clinical outcomes. For mothers who delivered outside of Dar es Salaam, or the delivery was not able to be attended by study staff, information from clinical staff and medical records were obtained. Infant birthweight was measured to the nearest 5 grams using a digital scale (SECA, Hamburg, Germany). During the postnatal period, mother-infant pairs attended study visits at 6 weeks postpartum and monthly thereafter until trial discharge at 1 year postpartum. Infants received an HIV-1 DNA PCR test at 6 weeks and 1 year of age.
Statistical analysis
We first described the proportions and rates of fetal death, stillbirth, neonatal mortality, and infant mortality in the cohort using standard international definitions and metrics. Fetal death was defined by an offspring, irrespective of the duration of pregnancy, that did not breathe or show any other sign of life including a heartbeat, pulsation of the umbilical cord, or movement of voluntary muscles[16]; stillbirth was defined as a fetal death ≥28 weeks gestation[17]. The stillbirth rate was calculated as the number of stillbirths per 1,000 total births with a pregnancy duration ≥28 weeks gestation[17]. The neonatal mortality rate was calculated as the number of deaths that occurred <28 days of life per 1,000 livebirths; the infant mortality rate was the number of deaths that occurred <365 days of life per 1,000 livebirths[18].
We then examined the relationship of sociodemographic, pregnancy, and HIV-related factors with the risk of fetal death, stillbirth, neonatal death, and infant death. The risk factors we analyzed included maternal age (18–24, 25–34, ≥35 years), maternal education (no formal education, primary, secondary/advanced), household wealth (less than or greater than or equal to the median), parity (0, 1–2, 3+), maternal body mass index (BMI) at randomization (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m2), maternal WHO HIV disease stage at baseline (I or II/III/IV), maternal CD4 T-cell at baseline (<350, ≥350 cells/μL), the timing of ART initiation (during this pregnancy or before conception), and multiple gestation (singleton or twin). During the conduct of the trial, the HIV program changed to only provide CD4 T-cell count assessment at ART initiation and discontinued routine testing during ART follow-up visits. As a result, about half of the pregnant women in the study had missing CD4 T-cell counts. Wealth quintiles were created with a principal component analysis of household assets (electricity, generator, running water, sofa/couch, television, radio, refrigerator, fan, bicycle, car, and roof type). Generalized estimating equations (GEEs) with a compound symmetry correlation matrix, a log link, and binomial variance were used to account for multiple gestations (twins) and produce relative risks. Models also included a covariate for the randomized trial regimen (vitamin D or placebo).
We examined the relationship of SVN phenotypes, including LBW, preterm, and SGA with the risk of neonatal and infant death among livebirths. Small-for-gestational age (SGA), an indicator of fetal growth restriction, was defined as a birthweight less than the 10th percentile for gestational age by sex defined by INTERGROWTH-21st standards[19]. We also examined the four mutually exclusive combinations of SVN phenotypes defined by preterm and SGA statuses: (i) term-appropriate-for-gestational age (AGA) [reference group], (ii) term birth-SGA, (iii) preterm-AGA, and (iv) preterm-SGA with neonatal and infant death. Multivariable models for SVN phenotype analyses included adjustment for all previously noted sociodemographic, pregnancy, and HIV-related factors along with the randomized regimen and infant sex. Variables were selected as confounders based on a causal framework due to their theoretical association with prematurity, SGA, and LBW and neonatal and infant mortality and not being a mediator of the potential causal pathway. GEE models were similarly used to account for twin livebirths and to produce relative risk estimates.
We then calculated the partial population-attributable risk percentage (PAR%) for neonatal and infant mortality for the combination of term-SGA, preterm-AGA, and preterm-SGA[20]. The calculation of the PAR% included the prevalence of these SVN phenotypes and the multivariable relative risk estimates for each group. The interpretation of the PAR% is the proportion of neonatal or infant deaths that would not have occurred if we eliminated preterm birth and SGA in the cohort, but all other risks remained the same, and the assumption the relationships were causal. The missing indicator method was used to retain data for covariates that were missing >1%. Statistical analyses were performed using SAS version 9.4 (SAS Institute, North Carolina, USA).
Results
The prospective cohort of PWLHIV were enrolled in the parent vitamin D trial between June 2015 and April 2018 and mother-infant pairs were follow-up until October 2019. Fig. 1 presents the flow chart for PWLHIV and their offspring in the analysis. A total of 3,153 pregnant women were screened of which 2,300 were enrolled in the trial. One woman was administratively withdrawn after testing negative for HIV at the time of labor, resulting in 2,299 PWLHIV included in this analysis. Characteristics at the time of trial enrollment for PWLHIV are presented in Table 1. In terms of follow-up, 3 (0.1%) pregnant women died, 7 withdrew consent (0.3%), and 29 (1.3%) pregnant were lost to follow-up before a birth outcome was ascertained.
Figure 1.
Flow chart of mother and infant follow-up
Table 1.
Baseline characteristics of pregnant women living with HIV at study enrollment at 12–27 weeks gestation (n=2,299)
| n (%) | |
|---|---|
|
| |
| Maternal age | |
| 18 – 24 years | 369 (16.1) |
| 25 – 34 years | 1338 (58.2) |
| 35+ years | 592 (25.8) |
| Maternal education | |
| No formal education | 250 (10.9) |
| Primary | 1343 (58.4) |
| Secondary/Advanced | 704 (30.6) |
| Missing | 2 (0.1) |
| Weeks gestation at randomization | |
| 12–16.9 weeks | 486 (21.1) |
| 17–22.9 weeks | 1086 (47.2) |
| 23–27 weeks | 727 (31.6) |
| Body mass index at randomization | |
| <18.5 kg/m2 | 64 (2.8) |
| 18.5–24.9 kg/m2 | 1041 (45.3) |
| 25.0–29.9 kg/m2 | 751 (32.7) |
| ≥ 30.0 kg/m2 | 425 (18.5) |
| Missing | 18 (0.8) |
| WHO HIV disease stage | |
| I | 1973 (85.8) |
| II | 142 (6.2) |
| III | 166 (7.2) |
| IV | 18 (0.8) |
| CD4 T-cell count, cells per μL | |
| <350 | 474 (20.6) |
| 350–500 | 310 (13.5) |
| ≥500 | 297 (12.9) |
| Missing | 1218 (53.0) |
| Timing of ART initiation | |
| During this pregnancy | 1364 (59.3) |
| Before conception | 935 (40.7) |
| ART regimen | |
| Tenofovir / Lamivudine / Efavirenz | 2272 (98.8) |
| Other | 27 (1.2) |
| Randomized regimen | |
| Vitamin D | 1148 (49.9) |
| Placebo | 1151 (50.1) |
There were 136 fetal deaths (41 miscarriages and 95 stillbirths), including twins, recorded. The cumulative risk of fetal death was 5.9% among all fetuses with a known vital status and the stillbirth rate was 42.0 per 1,000 total births that occurred after 28 weeks gestation (95% CI: 33.7 to 50.3 per 1,000 total births). Sociodemographic, pregnancy, and HIV-related risk factors for fetal death and stillbirth in univariable and multivariable models are presented in Supplemental Table 1 and Table 2, respectively. In multivariable models, nulliparous pregnant women had 1.55 times (95% CI: 1.02–2.35) the risk of fetal death as compared to pregnant women with a parity of 1 or 2 and mothers below the median for household wealth had 1.50 times (95% CI: 1.07–2.09) the risk of fetal death as compared to those at or above median household wealth. There was no relationship of other factors examined with the risk of fetal death. No statistically significant risk factors for stillbirth were identified (Table 2).
Table 2.
Sociodemographic and pregnancy-related risk factors for fetal death, stillbirth, neonatal death, and infant death among mothers living with HIV in Tanzania in multivariable models
| Fetal death (any time during pregnancy) | Stillbirth (>28 weeks gestation) | Neonatal death (<28 days of age) | Infant death (<365 days of age) | |||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| Fetal death / total births (%) | Multivariable* RR (95% CI) | Stillbirths / total births >28 wks (%) | Multivariable* RR (95% CI) | Neonatal deaths / livebirths (%) | Multivariable* RR (95% CI) | Neonatal deaths n events / N livebirths (%) | Multivariable* RR (95% CI) | |
|
| ||||||||
| Maternal age | ||||||||
| 18 – 24 years | 21 / 365 (5.8) | 0.79 (0.48–1.31) | 15 / 359 (4.2) | 0.75 (0.42–1.34) | 7 / 344 (2.0) | 0.63 (0.26–1.53) | 15 / 344 (4.4) | 0.77 (0.43–1.40) |
| 25 – 34 years | 79 / 1342 (5.9) | Ref. | 59 / 1322 (4.5) | Ref. | 38 / 1263 (3.0) | Ref. | 70 / 1263 (5.5) | Ref. |
| 35+ years | 36 / 596 (6.0) | 1.12 (0.74–1.71) | 21 / 581 (3.6) | 0.82 (0.48–1.40) | 12 / 560 (2.1) | 0.64 (0.33–1.26) | 29 / 560 (5.2) | 0.91 (0.58–1.44) |
| Maternal education | ||||||||
| No formal education | 12 / 254 (4.7) | 0.81 (0.44–1.49) | 7 / 249 (2.8) | 0.67 (0.30–1.49) | 10 / 242 (4.1) | 1.88 (0.90–3.90) | 18 / 242 (7.4) | 1.52 (0.91–2.54) |
| Primary | 78 / 1344 (5.8) | Ref. | 52 / 1318 (4.0) | Ref. | 28 / 1266 (2.2) | Ref. | 67 / 1266 (5.3) | Ref. |
| Secondary/Advanced | 46 / 703 (6.5) | 1.10 (0.76–1.59) | 36 / 693 (5.2) | 1.28 (0.83–1.97) | 19 / 657 (2.9) | 1.20 (0.62–2.33) | 29 / 657 (4.4) | 0.75 (0.48–1.18) |
| Household wealth | ||||||||
| < Median | 77 / 1148 (6.7) | 1.50 (1.07–2.09) | 51 / 1122 (4.6) | 1.38 (0.93–2.06) | 26 / 1071 (2.4) | 0.85 (0.46–1.54) | 54 / 1071 (5.0) | 0.86 (0.58–1.28) |
| ≥ Median | 59 / 1153 (5.1) | Ref. | 44 / 1138 (3.9) | Ref. | 31 / 1094 (2.8) | Ref. | 60 / 1094 (5.5) | Ref. |
| Parity | ||||||||
| Nulliparous | 43 / 555 (7.8) | 1.55 (1.02–2.35) | 31 / 543 (5.7) | 1.56 (0.96–2.53) | 16 / 512 (3.1) | 1.50 (0.76–2.94) | 30 / 512 (5.9) | 1.36 (0.85–2.15) |
| 1–2 | 71 / 1357 (5.4) | Ref. | 49 / 1333 (3.7) | Ref. | 32 / 1284 (2.5) | Ref. | 66 / 1284 (5.1) | Ref. |
| 3+ | 20 / 391 (5.1) | 0.91 (0.54–1.53) | 15 / 386 (3.9) | 1.21 (0.65–2.23) | 9 / 371 (2.4) | 0.96 (0.43–2.12) | 18 / 371 (4.9) | 0.85 (0.49–1.49) |
| Maternal BMI at baseline | ||||||||
| <18.5 kg/m2 | 3 / 59 (4.8) | 0.75 (0.24–2.39) | 2 / 61 (3.3) | 0.69 (0.16–2.89) | 1 / 59 (1.7) | 0.70 (0.09–5.35) | 4 / 59 (6.8) | 1.38 (0.50–3.77) |
| 18.5–24.9 kg/m2 | 59 / 1038 (5.7) | Ref. | 41 / 1020 (4.0) | Ref. | 23 / 979 (2.4) | Ref. | 49 / 979 (5.0) | Ref. |
| 25.0–29.9 kg/m2 | 42 / 755 (5.6) | 1.02 (0.70–1.49) | 31 / 744 (4.2) | 1.08 (0.69–1.71) | 21 / 713 (3.0) | 1.25 (0.70–2.25) | 35 / 713 (4.9) | 0.99 (0.64–1.51) |
| ≥ 30.0 kg/m2 | 31 / 430 (7.2) | 1.35 (0.88–2.09) | 20 / 419 (4.8) | 1.29 (0.76–2.20) | 12 / 399 (3.0) | 1.28 (0.62–2.62) | 26 / 399 (6.5) | 1.32 (0.81–2.14) |
| WHO HIV disease stage | ||||||||
| I | 121 / 1976 (6.1) | Ref. | 84 / 1939 (4.3) | Ref. | 47 / 1855 (2.5) | Ref. | 88 / 1855 (5.3) | Ref. |
| II/III/IV | 15 / 327 (4.6) | 0.78 (0.46–1.35) | 11 / 323 (3.4) | 0.86 (0.45–1.62) | 10 / 312 (3.2) | 1.41 (0.72–2.75) | 15 / 312 (4.8) | 0.89 (0.52–1.54) |
| CD4 T-cell count, cells per μL | ||||||||
| <350 | 31 / 475 (6.5) | 1.17 (0.72–1.89) | 24 / 468 (5.1) | 1.44 (0.80–2.58) | 12 / 444 (2.7) | 1.75 (0.75–4.08) | 27 / 444 (6.1) | 1.87 (1.05–3.32) |
| ≥350 | 33 / 605 (5.5) | Ref. | 21 / 593 (3.5) | Ref. | 9 / 572 (1.6) | Ref. | 19 / 572 (3.3) | Ref. |
| Timing of ART initiation | ||||||||
| During this pregnancy | 84 / 1275 (6.2) | 1.03 (0.72–1.48) | 59 / 1334 (4.4) | 1.01 (0.65–1.57) | 31 / 1275 (2.4) | 0.88 (0.52–1.50) | 64 / 1275 (5.0) | 0.88 (0.60–1.28) |
| Before conception | 52 / 944 (5.5) | Ref. | 36 / 928 (3.9) | Ref. | 26 / 892 (2.9) | Ref. | 50 / 892 (5.6) | Ref. |
| Multiple gestation | ||||||||
| Singleton | 134 / 2217 (6.0) | 0.39 (0.10–1.57) | 93 / 2176 (4.3) | Ref. | 51 / 2083 (2.5) | Ref. | 106 / 2083 (5.1) | Ref. |
| Twin | 2 / 86 (2.3) | Ref. | 2 / 86 (2.3) | 0.56 (0.14–2.25) | 6 / 84 (7.1) | 3.45 (1.42–8.40) | 8 / 84 (9.5) | 1.95 (0.93–4.10) |
Multivariable models included all variables in the table plus randomized regimen (vitamin D3 or placebo). Models also accounted for the correlation for twin pregnancies.
There were 2,167 livebirths in the study, 57 neonatal deaths, and 114 infant deaths (inclusive of neonatal deaths) were recorded. The neonatal mortality rate was 26.3 per 1,000 livebirths (95% CI: 19.6 to 33.0 per 1,000) and the infant mortality rate was 52.6 per 1,000 livebirths (95% CI: 44.0 to 62.8 per 1,000). 45 infants were diagnosed with HIV infection during follow-up (2.1% transmission rate) and among these infants there was only one death was recorded. Twin births had 3.45 times (95% CI: 1.42–8.40) the risk of neonatal death as compared to singletons (Table 2). A maternal CD4 T-cell count <350 in pregnancy, as compared to ≥350 cells/μL, was the only characteristic examined that was associated with an increased risk of infant death (RR: 1.88; 95% CI: 1.06–3.34).
Among livebirths, 523 (24.1%) were born preterm. A total of 2,161 livebirths had birthweight recorded, 421 (19.5%) were born SGA <10th percentile, and 217 (10.0%) were LBW. The association of preterm birth, SGA, and LBW with neonatal and infant mortality is presented in Table 3. Preterm birth (RR: 2.46; 95% CI: 1.44–4.18) and LBW (RR: 4.26; 95% CI: 2.32–7.80) were associated with increased risk of neonatal death in multivariable analyses (p-values <0.05); both were also associated with increased risk of infant death, but the magnitude of association was attenuated (preterm RR: 1.82; 95% CI: 1.26, 2.64; LBW RR: 2.43; 1.54; 95% CI: 3.84). In sensitivity analyses restricted to singleton livebirths, preterm birth, and LBW were associated with a similar magnitude of increased risk for neonatal and infant death (Supplemental Table 2).
Table 3.
Association of preterm, small-for-gestational age, and low birthweight births with neonatal and infant death among infants born to women living with HIV
| Neonatal deaths (< 28 days) | Infant deaths (< 365 days) | |||
|---|---|---|---|---|
|
| ||||
| Deaths / N livebirths (%) | Multivariable* RR (95% CI) | Deaths / N livebirths (%) | Multivariable* RR (95% CI) | |
|
| ||||
| Gestational age | ||||
| Preterm (<37 weeks) | 26 / 523 (5.0) | 2.46 (1.44–4.18) | 44 / 523 (8.4) | 1.82 (1.26–2.64) |
| Term (≥37 weeks) | 31 / 1644 (1.9) | Ref. | 70 / 1644 (4.3) | Ref. |
| Birthweight-for-gestational age | ||||
| SGA (<10th percentile) | 16 / 421 (3.8) | 1.57 (0.86–2.87) | 29 / 421 (6.9) | 1.39 (0.90–2.14) |
| AGA (≥10th percentile) | 38 / 1740 (2.4) | Ref. | 81 / 1740 (4.7) | Ref. |
| Birthweight | ||||
| LBW (<2500g) | 17 / 217 (7.8) | 4.26 (2.32–7.80) | 23 / 217 (10.6) | 2.43 (1.54–3.84) |
| Non-LBW (≥2500g) | 37 / 1944 (1.9) | Ref. | 87 / 1944 (4.5) | Ref. |
|
| ||||
| Preterm and SGA status | ||||
| Term-AGA (≥37 weeks and ≥10th percentile) | 18 / 1253 (1.4) | Ref. | 46 / 1253 (3.6) | Ref. |
| Term-SGA (≥37 weeks and <10th percentile) | 12 / 388 (3.1) | 2.07 (1.00–4.28) | 23 / 388 (5.9) | 1.55 (0.94–2.55) |
| Preterm-AGA (<37 weeks and ≥10th percentile) | 20 / 487 (4.1) | 2.87 (1.54–5.35) | 35 / 487 (7.2) | 1.91 (1.24–2.92) |
| Preterm-SGA (<37 weeks and <10th percentile) | 4 / 33 (12.1) | 7.15 (2.11–24.30) | 6 / 33 (18.2) | 5.14 (2.22–11.90) |
AGA: appropriate-for-gestational age; LBW: low birthweight; SGA: small-for-gestational age
Multivariable models included adjustment for maternal age (18–24, 25–34, ≥ 35 years), maternal education (no formal education, primary, secondary/advanced), household wealth (less than or greater than or equal to the median), parity (0, 1–2, 3+), maternal body mass index (BMI) at randomization (<18.5, 18.5–24.9, 25.0–29.9, ≥ 30.0 kg/m2), maternal WHO HIV disease stage at baseline (I or II / III / IV), maternal CD4 T-cell at baseline (< 350, ≥ 350 cells/μL), the timing of ART initiation (during this pregnancy or before conception), multiple gestation (singleton or twin), and randomized regimen (vitamin D or placebo). Models also accounted for correlation for twin infants.
As for SVN phenotypes defined by preterm and SGA status, 58.0% of livebirths were term-AGA, 18.0% term-SGA, 22.5% preterm-AGA, and 1.5% were preterm-SGA. Compared to the reference group of term-AGA livebirths, the relative risks for neonatal mortality for term-SGA, preterm-AGA, and preterm-SGA were 2.07 (95% CI: 1.00–4.28), 2.87 (95% CI 1.54–5.35) and 8.36 (95% CI 2.95–23.75), respectively. There was a similar pattern of increasing risk for the combinations of SGA and preterm births with infant mortality, but the magnitude of the risk tended to be attenuated as compared to neonatal death (Table 3). Among singleton livebirths, preterm-AGA and preterm-SGA were associated with increased risk of neonatal and infant death as compared to term-AGA births, but the measures of association for term-SGA did not reach statistical significance (Supplemental Table 2).
In terms of population-attributable risk percentages, an estimated 42.7% (95% CI: 4.0–70.3%) of neonatal deaths and 29.4% (95% CI: 0.2–53.9%) of infant deaths in the study cohort were attributable to the combinations of preterm and SGA births. In sensitivity analyses restricted to singleton livebirths, the population-attributable risk percentages for combinations of preterm and SGA births were similar for neonatal (44.3%; 95% CI: 10.7–68.8%) and infant death (30.1%; 95% CI: 4.1–52.2%).
Discussion
In this prospective cohort study, we found a high risk of fetal death, stillbirth, neonatal death, and infant death among offspring of PWLHIV in urban Tanzania. Nulliparity and low household wealth were associated with an increased risk of fetal death, twin birth was associated with an increased risk of neonatal death, and a maternal CD4 T-cell count <350 cells per μL in pregnancy was associated with an increased risk of infant death. Preterm and SGA births were associated with an increased risk of neonatal and infant death, and we estimated in the cohort that nearly one-half of neonatal deaths and one-third of infant deaths were attributable to these small and vulnerable newborn phenotypes.
The rate of stillbirth in our cohort was approximately double that of the general population of Tanzania at 22.0 per 1,000 total births[17]. Untreated HIV infection is known to increase the risk of stillbirth[1]; however, ART appears to reduce, but not eliminate, the excess risk of stillbirth[11]. A cohort study in Lesotho found PHWLIV who mostly received TDF-3TC-EFV, the same ART regimen as our study, had a 7.1% risk of fetal death which was over twice the risk as compared to HIV-uninfected pregnant women[3]. This doubling of risk is consistent with the difference in the stillbirth rate found in our study as compared to the estimate for the general population of Tanzania. However, due to the lack of an HIV-uninfected comparison group in our study, we were not able to directly assess fetal death or stillbirth rates in the general population. Nevertheless, it is important to acknowledge that our fetal and infant mortality estimates were directly measured which may explain some of the differences with the modeled national mortality estimates that attempted to take into account underreporting with statistical approaches[17].
We also found a higher rate of neonatal and infant mortality among PWLHIV as compared to that estimated for the general population of Tanzania. The estimated neonatal mortality rate for Tanzania in 2021 was 20.0 per 1,000 livebirths, while the infant mortality rate was 34.1 per 1,000 livebirths[21]; we found approximately 30% and 50% higher mortality rates in our cohort, respectively. Children living with HIV and HIV-exposed uninfected children may be more likely to experience multiple risk factors for death, including preterm birth and childhood infections[22, 23]. Aligned with our findings, twins are known to experience greater neonatal mortality risk as compared to singletons in sub-Saharan Africa and this is likely attributable to the greater risk of pregnancy and intrapartum complications[24]. Studies conducted before the availability of ART in sub-Saharan Africa also documented that low maternal CD4 T-cell counts were associated with an increased risk of child mortality[25, 26]. In addition, maternal ART regimens may modify the risk of infant death[11]. A prior study in Botswana found that exposure to ZDV-3TC-NVP and ZDV-3TC-lopinavir-ritonavir (LPV-R) were associated with increased risk of neonatal death as compared to TDF-FTC-EFV[27]. In the IMPAACT 2010/VESTED trial, EFV–FTC–TDF was associated with a greater risk of neonatal mortality than with newer dolutegravir (DTG)-based regimens[28]; however, observational studies have also found no differences in adverse birth outcomes between DTG and EFV-based regimens[29, 30]. Therefore, further research on the role of ART regimens in birth outcomes and mortality is needed, including studies to elucidate mechanisms of action. However, it is important to recognize that PWLHIV experience many HIV and non-HIV-related risk factors for poor offspring outcomes, such as depression, tobacco smoking, suboptimal weight gain in pregnancy, hypertensive disorders, and other maternal infections, which require research and the development of effective interventions[31–34].
We found about 4 in 10 livebirths in the study cohort were preterm, SGA, or LBW and we found each of these phenotypes was differentially associated with the risk of neonatal and infant mortality. This is consistent with a large and growing body of literature on SVNs among HIV-uninfected pregnancies[31, 32]. Notably, in our study, preterm live births that were also born SGA had the highest risk of neonatal mortality with nearly seven times the risk as the reference term-AGA infant infants, but the prevalence of this phenotype was less than 2% of livebirths. Preterm-AGA infants had the highest prevalence among the SVN phenotypes (22.5%) and were associated with nearly three times the risk of infant death, while term-SGA infants had a slightly lower prevalence (18.0%) but twice the risk of death as compared to term-AGA infants. A recent pooled cohort study of over 200,000 livebirths born predominantly to HIV-uninfected pregnant women in low- and middle-income countries found a similar trend, but the median relative risks for term-SGA, preterm-AGA, and preterm-SGA were slightly elevated at 2.7, 6.0, and 10.4, respectively[35].
Importantly, due to the high prevalence and strong magnitude of risk for death for preterm and SGA, we estimated that a large percentage of neonatal (42.6%) and infant mortality (29.4%) were attributable to the combination of these factors in the study cohort. Our population-attributable risk percentages were slightly lower than estimates from the recent Lancet series on SVNs, which estimated that ~60% of neonatal mortality in the general population of sub-Saharan Africa was attributable to preterm birth and SGA[31]. There are multiple antenatal care interventions for pregnancies recommended that have been shown to reduce the risk of prematurity or fetal growth restriction including but not limited to multiple micronutrient supplementation, calcium supplementation, low-dose aspirin, malaria prevention, smoking cessation, and treatment of sexually transmitted infections[36]. In the context of HIV, early initiation and optimizing ART regimens may also potentially reduce the risk of preterm or fetal growth restriction[11].
The major strengths of our study included the relatively large sample size and low loss-to-follow-up; however, our study also had several limitations. First, although we put substantial effort into differentiating fetal deaths and neonatal deaths, there remains some risk of misclassification for deaths occurring at delivery. Second, HIV testing was first performed for children at 6 weeks of age and therefore we were unable to assess the relationship between early HIV infection on the risk of neonatal mortality. Universal HIV testing at birth, which has shown to be programmatically feasible in similar contexts[37–39], should be considered for implementation in Tanzania to allow for early initiation of ART for infants with in-utero or intrapartum HIV infection. We also did not have complete data on maternal viral load in pregnancy since pregnant women who initiated ART in pregnancy received their first viral load test after 6 months of treatment, largely after delivery, in the program. Further, about half of pregnant women were missing data on CD4 T-cell counts at enrollment in the study which may result in some degree of residual confounding in analyses of the relationship between birth outcomes and mortality. In addition, while the sample size of our pregnancy cohort was large, fetal and infant deaths were rare (<5%), and therefore we had limited statistical power to detect risk factors that had moderate to small effect sizes and the confidence intervals for the relationship of birth outcomes with neonatal and infant mortality were relatively wide. Gestational age dating was estimated maternal-reported last menstrual period and therefore it is known that some degree of measurement error occurred. Last, we conducted this study in urban Tanzania among PWLHIV who were enrolled in a clinical trial, and therefore our findings may not be directly generalizable to other settings or among women not as strongly linked to antenatal care. In addition, in our study, PWLHIV also received EFV-based ART regimens, and therefore the findings may not be directly generalizable in the context of newer DTG-based regimens.
In our cohort study, we found that PWLHIV in Tanzania has a notably high risk for fetal death, stillbirth, neonatal mortality, and infant mortality that exceeds that expected for the general population of pregnant women in Tanzania. We identified several pregnancy-related and HIV-related factors associated with these adverse outcomes, including nulliparity, low socioeconomic status, twin births, and low maternal CD4 T-cell counts in pregnancy. We also found that preterm birth, SGA, and low birthweight were also strongly associated with the risk of neonatal and infant mortality, and due to their high prevalence in the cohort, we estimated nearly half of the neonatal deaths and about one-third of infant deaths were attributable to these adverse birth outcomes. As a result, the development of new interventions along with effective implementation of recommended antenatal care and newborn survival interventions may significantly improve the survival of infants born to PWLHIV in the context of sub-Saharan Africa.
Supplementary Material
Acknowledgments
Author contributions: KPM and CRS developed the research question and designed the study. KPM, WWF, and CRS obtained funding for the parent trial. KPM, AM, CPD, FMAA, NU, WWF, and CRS oversaw the conduct of the parent trial data collection. CRS conducted the statistical analysis. KPM and CRS produced the initial manuscript, all authors critically reviewed the manuscript for important intellectual content and all authors approved the submission of the final version of the manuscript.
We would like to thank the field teams (including physicians, nurses, supervisors, and laboratory staff), administrative staff, and study participants, all of whom made this study possible.
Source of Funding:
The study was funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) under grant R01HD083113. CPD was partially funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant P30DK040561. CRS and WWF were partially supported by the Fogarty International Center of the National Institutes of Health grant D43TW009775.
Footnotes
Conflicts of Interest: None to declare.
References
- 1.Wedi CO, Kirtley S, Hopewell S, Corrigan R, Kennedy SH, Hemelaar J. Perinatal outcomes associated with maternal HIV infection: a systematic review and meta-analysis. Lancet HIV 2016; 3(1):e33–48. [DOI] [PubMed] [Google Scholar]
- 2.Teasdale CA, Marais BJ, Abrams EJ. HIV: prevention of mother-to-child transmission. BMJ Clin Evid 2011; 2011. [PMC free article] [PubMed] [Google Scholar]
- 3.Tukei VJ, Hoffman HJ, Greenberg L, Thabelo R, Nchephe M, Mots’oane T, et al. Adverse Pregnancy Outcomes Among HIV-positive Women in the Era of Universal Antiretroviral Therapy Remain Elevated Compared With HIV-negative Women. Pediatr Infect Dis J 2021; 40(9):821–826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Shinar S, Agrawal S, Ryu M, Walmsley S, Serghides L, Yudin MH, Murphy KE. Perinatal outcomes in women living with HIV-1 and receiving antiretroviral therapy-a systematic review and meta-analysis. Acta Obstet Gynecol Scand 2022; 101(2):168–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wang L Determinants of child mortality in LDCs: empirical findings from demographic and health surveys. Health Policy 2003; 65(3):277–299. [DOI] [PubMed] [Google Scholar]
- 6.Simmons RA, Anthopolos R, O’Meara WP. Effect of health systems context on infant and child mortality in sub-Saharan Africa from 1995 to 2015, a longitudinal cohort analysis. Sci Rep 2021; 11(1):16263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Muhihi A, Sudfeld CR, Smith ER, Noor RA, Mshamu S, Briegleb C, et al. Risk factors for small-for-gestational-age and preterm births among 19,269 Tanzanian newborns. BMC Pregnancy Childbirth 2016; 16:110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Evans C, Jones CE, Prendergast AJ. HIV-exposed, uninfected infants: new global challenges in the era of paediatric HIV elimination. The Lancet infectious diseases 2016; 16(6):e92–e107. [DOI] [PubMed] [Google Scholar]
- 9.Jiang W, Ronen K, Osborn L, Drake AL, Unger JA, Matemo D, et al. Predictors of adverse pregnancy outcomes among Kenyan women with HIV on antiretroviral treatment in pregnancy. Aids 2022; 36(7):1007–1019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chen JY, Ribaudo HJ, Souda S, Parekh N, Ogwu A, Lockman S, et al. Highly active antiretroviral therapy and adverse birth outcomes among HIV-infected women in Botswana. J Infect Dis 2012; 206(11):1695–1705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Eke AC, Mirochnick M, Lockman S. Antiretroviral Therapy and Adverse Pregnancy Outcomes in People Living with HIV. N Engl J Med 2023; 388(4):344–356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Filteau S The HIV-exposed, uninfected African child. Trop Med Int Health 2009; 14(3):276–287. [DOI] [PubMed] [Google Scholar]
- 13.Ashorn P, Ashorn U, Muthiani Y, Aboubaker S, Askari S, Bahl R, et al. Small vulnerable newborns-big potential for impact. Lancet, 2023; 401(10389):1692–1706. [DOI] [PubMed] [Google Scholar]
- 14.Sudfeld CR, Manji KP, Muhihi A, Duggan CP, Aboud S, Alwy Al-Beity FM, et al. Vitamin D3 supplementation during pregnancy and lactation for women living with HIV in Tanzania: A randomized controlled trial. PLoS Med, 2022; 19(4):e1003973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sudfeld CR, Manji KP, Duggan CP, Aboud S, Muhihi A, Sando DM, et al. Effect of maternal vitamin D(3) supplementation on maternal health, birth outcomes, and infant growth among HIV-infected Tanzanian pregnant women: study protocol for a randomized controlled trial. Trials 2017; 18(1):411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.World Health Organization International Statistical Classification of Diseases and related health problems: Alphabetical index. World Health Organization; 2004. [Google Scholar]
- 17.Lawn JE, Blencowe H, Waiswa P, Amouzou A, Mathers C, Hogan D, et al. Stillbirths: rates, risk factors, and acceleration towards 2030. Lancet, 2016; 387(10018):587–603. [DOI] [PubMed] [Google Scholar]
- 18.Sharrow D, Hug L, You D, Alkema L, Black R, Cousens S, et al. Global, regional, and national trends in under-5 mortality between 1990 and 2019 with scenario-based projections until 2030: a systematic analysis by the UN Inter-agency Group for Child Mortality Estimation. Lancet Glob Health, 2022; 10(2):e195–e206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Papageorghiou AT, Kennedy SH, Salomon LJ, Altman DG, Ohuma EO, Stones W, et al. The INTERGROWTH-21st fetal growth standards: toward the global integration of pregnancy and pediatric care. Am J Obstet Gynecol 2018; 218(2):S630–S640. [DOI] [PubMed] [Google Scholar]
- 20.Spiegelman D, Hertzmark E, Wand HC. Point and interval estimates of partial population attributable risks in cohort studies: examples and software. Cancer Causes Control, 2007; 18(5):571–579. [DOI] [PubMed] [Google Scholar]
- 21.United Nations Inter-agency Group for Child Mortality Estimation (UN IGME) Levels & Trends in Child Mortality: Report 2022, Estimates developmed by the United Nations Inter-agency Group for Child Mortality Estimation. New York: United Nations Children’s Fund; 2023. [Google Scholar]
- 22.Slogrove AL, Esser MM, Cotton MF, Speert DP, Kollmann TR, Singer J, Bettinger JA. A Prospective Cohort Study of Common Childhood Infections in South African HIV-exposed Uninfected and HIV-unexposed Infants. Pediatr Infect Dis J 2017; 36(2):e38–e44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Xiao PL, Zhou YB, Chen Y, Yang MX, Song XX, Shi Y, Jiang QW. Association between maternal HIV infection and low birth weight and prematurity: a meta-analysis of cohort studies. BMC Pregnancy Childbirth 2015; 15:246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Monden CWS, Smits J. Mortality among twins and singletons in sub-Saharan Africa between 1995 and 2014: a pooled analysis of data from 90 Demographic and Health Surveys in 30 countries. Lancet Glob Health, 2017; 5(7):e673–e679. [DOI] [PubMed] [Google Scholar]
- 25.Brahmbhatt H, Kigozi G, Wabwire-Mangen F, Serwadda D, Lutalo T, Nalugoda F, et al. Mortality in HIV-infected and uninfected children of HIV-infected and uninfected mothers in rural Uganda. J Acquir Immune Defic Syndr 2006; 41(4):504–508. [DOI] [PubMed] [Google Scholar]
- 26.Kuhn L, Kasonde P, Sinkala M, Kankasa C, Semrau K, Scott N, et al. Does severity of HIV disease in HIV-infected mothers affect mortality and morbidity among their uninfected infants? Clin Infect Dis 2005; 41(11):1654–1661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zash R, Jacobson DL, Diseko M, Mayondi G, Mmalane M, Essex M, et al. Comparative Safety of Antiretroviral Treatment Regimens in Pregnancy. JAMA Pediatr, 2017; 171(10):e172222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lockman S, Brummel SS, Ziemba L, Stranix-Chibanda L, McCarthy K, Coletti A, et al. Efficacy and safety of dolutegravir with emtricitabine and tenofovir alafenamide fumarate or tenofovir disoproxil fumarate, and efavirenz, emtricitabine, and tenofovir disoproxil fumarate HIV antiretroviral therapy regimens started in pregnancy (IMPAACT 2010/VESTED): a multicentre, open-label, randomised, controlled, phase 3 trial. Lancet (London, England) 2021; 397(10281):1276–1292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zash R, Jacobson DL, Diseko M, Mayondi G, Mmalane M, Essex M, et al. Comparative safety of dolutegravir-based or efavirenz-based antiretroviral treatment started during pregnancy in Botswana: an observational study. The Lancet Global health 2018; 6(7):e804–e810. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Patel K, Huo Y, Jao J, Powis KM, Williams PL, Kacanek D, et al. Dolutegravir in Pregnancy as Compared with Current HIV Regimens in the United States. N Engl J Med 2022; 387(9):799–809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lawn JE, Ohuma EO, Bradley E, Idueta LS, Hazel E, Okwaraji YB, et al. Small babies, big risks: global estimates of prevalence and mortality for vulnerable newborns to accelerate change and improve counting. Lancet, 2023; 401(10389):1707–1719. [DOI] [PubMed] [Google Scholar]
- 32.Katz J, Lee AC, Kozuki N, Lawn JE, Cousens S, Blencowe H, et al. Mortality risk in preterm and small-for-gestational-age infants in low-income and middle-income countries: a pooled country analysis. Lancet, 2013; 382(9890):417–425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Dekker GA, Lee SY, North RA, McCowan LM, Simpson NA, Roberts CT. Risk factors for preterm birth in an international prospective cohort of nulliparous women. PLoS One 2012; 7(7):e39154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lambert JS, Watts DH, Mofenson L, Stiehm ER, Harris DR, Bethel J, et al. Risk factors for preterm birth, low birth weight, and intrauterine growth retardation in infants born to HIV-infected pregnant women receiving zidovudine. Pediatric AIDS Clinical Trials Group 185 Team. AIDS 2000; 14(10):1389–1399. [DOI] [PubMed] [Google Scholar]
- 35.Hazel EA, Erchick DJ, Katz J, Lee ACC, Diaz M, Wu LSF, et al. Neonatal mortality risk of vulnerable newborns by fine stratum of gestational age and birthweight for 230 679 live births in nine low- and middle-income countries, 2000–2017. BJOG, 2024. [DOI] [PubMed] [Google Scholar]
- 36.Hofmeyr GJ, Black RE, Rogozińska E, Heuer A, Walker N, Ashorn P, et al. Evidence-based antenatal interventions to reduce the incidence of small vulnerable newborns and their associated poor outcomes. Lancet, 2023; 401(10389):1733–1744. [DOI] [PubMed] [Google Scholar]
- 37.Teasdale CA, Tsiouris F, Mafukidze A, Shongwe S, Choy M, Nhlengetfwa H, et al. Birth Testing for Infant HIV Diagnosis in Eswatini: Implementation Experience and Uptake Among Women Living With HIV in Manzini Region. Pediatr Infect Dis J 2020; 39(9):e235–e241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Wexler C, Nazir N, Maloba M, Brown M, Goggin K, Gautney B, et al. Programmatic evaluation of feasibility and efficiency of at birth and 6-week, point of care HIV testing in Kenyan infant. PLoS One 2020; 15(10):e0240621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Boeke CE, Joseph J, Wang M, Abate ZM, Atem C, Coulibaly KD, et al. Point-of-care testing can achieve same-day diagnosis for infants and rapid ART initiation: results from government programmes across six African countries. J Int AIDS Soc 2021; 24(3):e25677. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.