Secondary Logo

Journal Logo

Implementation Science

Behavioral Interventions can Mitigate Adverse Pregnancy Outcomes Among Women Conceiving on ART and Those Initiated on ART During Pregnancy: Findings From the MOTIVATE Trial in Southwestern Kenya

Onono, Maricianah MBChB, MSc, PhDa; Odwar, Tobias RNa; Wahome, Samuel BSca; Helova, Anna MPHb; Bukusi, Elizabeth Anne MBChB, MPH, PhDa; Hampanda, Karen PhDc; Turan, Janet PhDb; Abuogi, Lisa MDd

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: January 1, 2021 - Volume 86 - Issue 1 - p 46-55
doi: 10.1097/QAI.0000000000002521
  • Free



In 2015, WHO recommended combination antiretroviral therapy (ART) for all pregnant and breastfeeding women living with HIV (WLWH).1,2 The introduction of ART has greatly reduced vertical transmission of HIV from 220,000 new infections in children globally in 2014 to 160,000 in 2018.3–5 Unfortunately, these 160,000 new HIV infections in children is 4 times the target of 40,000 set forth by the Start Free Stay Free AIDS Free initiative underscoring the fact that vertical HIV transmission is still very much an urgent global health issue.6 To strengthen the elimination of vertical transmission of HIV efforts, the World Health Organization (WHO) recommends offering pre-exposure prophylaxis (PrEP) to HIV-negative pregnant women in HIV high-burden settings.7 As such, the number of women in HIV high-burden settings using ART in pregnancy for HIV treatment and prevention is substantially increasing.

Although the importance of ART for prevention of vertical transmission of HIV and maternal health is clear, significant concerns have been raised regarding the possible association between in utero ART exposure and adverse pregnancy outcomes (APOs).8–13 There is increasing global concern regarding APOs among women with HIV, given that over a million WLWH already on ART are becoming pregnant annually, and newly diagnosed women are initiating ART while pregnant.13 An increasing number of studies have shown a higher risk of preterm delivery (PTD),14–16 low birth weight (LBW),17 small for gestational age (SGA),13,14,18 stillbirth,11,13,16 miscarriage,19 and neonatal deaths among women receiving ART,11,12,17,20 Conversely, some studies have not shown an association between in utero ART exposure and APO.11,21–24 With national treatment programs in high-burden countries such as Kenya implementing ART for all women with HIV, and providing PrEP HIV-negative pregnant women,25 data on how ART exposure may affect birth outcomes are urgently required.

As the availability of ART expands in sub-Saharan Africa, and the numbers of ART-exposed pregnancies increase, it is important to understand the impact that ART may have on maternal and infant outcomes as well as explore potential interventions to prevent these adverse outcomes. In this study, we evaluate pregnancy outcomes among women initiating ART before pregnancy vs. those initiating during pregnancy, identify risk factors associated with APO, and examine the effect of 2 behavioral interventions (community mentor mothers [CMMs] and text messaging) on APO among over 1300 pregnant women prospectively followed from pregnancy up to 12 months postpartum.


Study Design

This is a secondary data analysis of the Mother and Infant Visit Adherence and Treatment Engagement (MOTIVATE) study conducted between December 2015 and May 2019 in southwestern Kenya. The MOTIVATE study is a cluster randomized 2X2 factorial trial assessing the impact of community-based mentor mothers (CMMs) or automated text messages, or a combination of both interventions on ART adherence and retention in the context of ART for pregnant women with HIV. CMMs conducted home visits to assist participants with safe disclosure, supported safe infant feeding, promoted safer sex and family planning, encouraged early infant testing and follow up, and promoted ART adherence and return for HIV care visits. Mobile phone text messages involved bidirectional communication between providers and mothers and encouraged uptake of important maternal and child health behaviors and services (good nutrition, antenatal clinic attendance, birth planning, and skilled delivery) and engagement and adherence to HIV care.26


The study was conducted in 24 Ministry of Health public antenatal clinics (ANC) providing PMTCT services in Homa Bay, Kisumu, and Migori counties in Kenya. Local partners funded through the US Centers for Disease Control—the President's Emergency Plan for AIDS Relief (CDC-PEPFAR), supported ART services. This region is considered to be an HIV hyperendemic area with an overall adult HIV prevalence of 13.0% in Migori, 17.5% in Kisumu, and 19.6% in Homa Bay counties compared with 4.9% nationally in 2018. PMTCT service coverage is high with nearly 90% of women with access to PMTCT interventions.5


We enrolled pregnant WLWH who were 18 years or older regardless of gestation age and previous antenatal clinic attendance who were willing to have home visits (or meetings with a CMM in an alternate location), lived in the facility catchment area, and had access to a mobile phone. We included women who were already living with HIV and on ART as well as women who were being newly initiated on ART while pregnant. In line with the prevailing Option B+ (test and treat) approach for individuals testing HIV positive at the time, ART was initiated at first contact at ANC.1,2 Gestational age was determined by the participant's self-reported last menstrual period date documented on the first ANC medical record, as ultrasounds are almost completely unavailable in this population. A total of 1338 WLWH were enrolled in the parent trial; however, for this analysis, we included 1275 women after excluding false pregnancy ([n = 7] women who despite a history of amenorrhea and enrolment into antenatal care had a confirmatory negative human chorionic gonadotropin [hCG] test conducted in ANC), twin pregnancies (n = 31), women who withdrew consent to participate in the study before ascertainment of pregnancy outcomes (n = 7), and women without a date of ART initiation (n = 18). Pregnancy outcome analyses were restricted to singleton births given increased risk of APO in pregnancies with multiple gestations.27

Data Sources and Measurement

Data were abstracted monthly from routinely collected program medical records by study staff and uploaded to a web-based data server. Primarily, pregnancy outcomes were ascertained from the highly exposed infant (HEI) card and the corresponding facility HEI register. Although these are programmatic data, the implementing local partners did rigorous quality assurance and control procedures to clean the data, ensure consistency and completeness of data. Where participants transferred out or lost to follow-up, we obtained data through participant self-report to CMMs (during home visits) or through follow-up phone calls. The study coordinator, statistician, and investigators periodically reviewed the data for accuracy, consistency, and completeness.


The main outcome variable in this analysis was adverse pregnancy outcome defined as having a PTD, LBW, miscarriage, or stillbirth. As per WHO definitions, PTD was defined as delivery <37 weeks of gestational age, and LBW was defined as a birth weight below 2.5 kilograms (kg).28 Still birth was defined as no signs of life at birth > 28 weeks gestation and miscarriage as loss of fetus before 28 weeks gestation.

We explored the association of several possible predictor variables available from medical records and APO including (1) maternal age in years at the first ANC visit, (2) ART timing (preconception vs. postconception) using ART initiation date documented in patient medical records and estimated gestational age in weeks at the time of ART initiation which was derived from the last normal menstrual period documented in the participants' medical records, (3) cumulative time on ART, (4) baseline CD4+ count (cells/mL) (ie, CD4 result taken closest to the study enrollment date), (5) viral load (copies/mL) (measured from tests conducted 9 months before study enrollment up to 9 months poststudy enrollment); (6) baseline maternal anemia defined as a hemoglobin level <11.0 g/dL (no anemia >11.0 g/dL, mild anemia 10.0 g/dL–10.9 g/dL, moderate anemia 7.0 g/dL–9.9 g/dL, and severe anemia < 7.0 g/dL), 29 (7) type of ART (protease inhibitor [PI]-based, non-nucleoside reverse transcriptase inhibitor [NNRTI]-based, and tenofovir [TDF] vs. zidovudine [AZT] containing regimens), and (8) intervention type (CMM, text intervention, combined CMM and text message) versus control.

At the time of the study, the following ART drug combinations were allowed for pregnant and lactating women with HIV: zidovudine (AZT) + lamivudine (3TC) + nevirapine (NVP), AZT + 3TC + efavirenz (EFV), tenofovir (TDF) + 3TC + NVP, TDF + 3TC + EFV, AZT + 3TC + lopinavir (LPV/r), AZT + 3TC + atazanavir (ATV/r), TDF + 3TC + LPV/r, and TDF + 3TC + ATV/r. ART drug combinations with NNRTI backbone were the recommended first line ART, with TDF containing regimes being the most predominant at the time.30

Statistical Methods

Descriptive statistics were used to characterize baseline maternal characteristics in women with APO versus those without. Chi square tests were used to evaluate the association of APO and independent variables such as maternal age, parity, place of residence, ART regimen, ART timing, gestation age at time of ART initiation, baseline CD4 cell count, viral load copies/mL, and hemoglobin level. Covariates with a P value of <0.20 in bivariate analyses were considered for inclusion in multivariable models. We ran log binomial models for estimation of relative risks (RRs) (accounting for the clustered nature of the data) to evaluate 3 exposure comparisons (1) those initiating ART before pregnancy vs. those initiating during pregnancy in the whole sample of women, (2) among women initiating ART during pregnancy, comparisons across gestational ages at ART initiation (first trimester exposure to ART versus second and third trimester exposure), and (3) intervention (intervention vs. control and then different types of intervention). We further ran a sensitivity analysis limited to those initiating ART before pregnancy vs. those initiating ART in the first trimester of pregnancy. All analyses were conducted using Stata version 13 software and 2-sided P values of <0.05 were considered statistically significant.

Compliance With Ethical Standards

The ethics committees of the Kenya Medical Research Institute, the University of Colorado, Denver, and the University of Alabama at Birmingham gave scientific and ethical approval for the study. Written informed consent was obtained from each woman before enrollment.


A total of N = 1275 women of median age 31 (1QR 27–35) years were included in the analyses: 893 (70.0%) initiated ART before pregnancy and 382 (30.0%) during the pregnancy. Approximately 30.0% of these women experienced an APO (n = 388): 306 preterm births, 38 LBW infants, 33 stillbirths, and 11 miscarriages (20 women had a combined preterm and LBW APO). Among those initiating ART during pregnancy: 117 (9.2%) initiated during the first trimester, 220 (17.3%) during the second, and 45 (3.5%) in the third. The combination of tenofovir (TDF), lamivudine (3TC), and efavirenz (EFV) (TDF + 3TC + EFV) was the most commonly used regimen (N = 920 (71.15%)), whereas 4.4% were on a PI-based regimen. Figure 1 below shows the cohort disposition through delivery.

Cohort disposition.

Table 1 compares demographic and clinical characteristics of women at study enrollment by timing of ART initiation. Women initiating ART before pregnancy were older (31 years IQR 27–35) and had higher parity (P < 0.001). Additional differences by timing of ART initiation included current ART regimen, CD4 cell count nearest pregnancy, number of ANC visits, and receipt of a study intervention.

TABLE 1. - Characteristics of Pregnant Women at Study Entry Stratified by the Exposure Status
Totals N = 1293 Initiation Before Pregnancy N = 893 Initiation During Pregnancy N = 382 P*
First Trimester N = 117 Second Trimester N = 220 Third Trimester N = 45
Maternal characteristics
 Age, median (IQR) yr 31 (27–35) 26 (24–31) 26 (24–31) 26 (22–31) <0.001
  ≤24 yrs 254 115 (12.9) 43 (36.8) 71 (32.3) 19 (42.2) <0.001
  25–29 368 238 (26.7) 38 (32.5) 73 (33.1) 10 (22.2)
  ≥30 671 540 (60.4) 36 (30.7) 76 (34.6) 16 (35.6)
Obstetric characteristics, gestation, wk
 Median EGA at first ANC (IQR) 19 (12–24) 16 (11–20) 20 (16–24) 26 (22–28) 0.14
  Parity, mean (SD) 2.7(1.7) 1.9 (1.4) 1.9 (1.5) 2.0 (1.7) 0.05
  0 105 45 (5.0) 16 (13.7) 34 (15.5) 8 (17.8) <0.001
  1 295 165 (18.5) 37 (31.6) 73 (33.2) 12 (26.7)
  ≥2 893 683 (76.5) 64 (54.7) 113 (51.3) 25 (55.5)
 ANC visits, mean (SD) 2.1(1.4) 2.7 (1.9) 2.2 (1.3) 1.8 (0.8) <0.001
  1 580 425 (47.6) 40 (34.2) 88 (40.0) 20 (44.5) 0.001
  2 314 204 (22.8) 30 (25.6) 61 (27.7) 17 (37.8)
  3 175 119 (13.3) 12 (10.3) 32 (14.6) 6 (13.3)
  ≥4 224 145 (16.2) 35 (29.9) 39 (17.7) 2 (4.4)
Hemoglobin level
 No anemia 178 119 (32.3) 26 (45.5) 24 (23.3) 6 (31.6) 0.12
 Mild anemia 110 69 (18.8) 12 (21.1) 18 (17.5) 5 (26.3)
 Moderate anemia 102 63 (17.1) 7 (12.3) 27 (26.2) 2 (10.5)
 Severe anemia 170 117 (31.8) 12 (21.1) 34 (33.0) 6 (31.6)
HIV-related characteristics
  NNRTI 1237 840 (94.1) 115 (98.3) 219 (99.6) 45 (100.0) <0.001
  PI 56 53 (5.9) 2 (1.7) 1 (0.4) 0 (0.0)
 AZT vs. TDF
  AZT 132 123 (14.3) 4 (3.4) 2 (0.9) 2 (4.4) <0.001
  TDF 1129 738 (85.7) 113 (96.6) 218 (99.1) 43 (95.6)
Baseline CD4 cell count, cells/mm3
  Median (IQR) 443 (295–633) 538 (408–723) 533 (319.5–683) 512 (303–651) 0.01
  ≤200 160 112 (13.4) 10 (9.5) 29 (17.2) 5 (14.3) <0.001
  201–350 209 175 (30.0) 11 (10.5) 17 (10.1) 5 (14.3)
  351–500 281 215 (25.8) 25 (23.8) 30 (17.9) 7 (20.0)
  >500 508 332 (39.8) 59 (56.2) 92 (54.8) 18 (51.4)
Baseline viral load, copies/mL
  Median (IQR) 0 (0–0) 0 (0–49) 0 (0–90) 0 (0–331) 0.002
  <1000 908 687 (92.7) 72 (87.8) 120 (87.0) 18 (85.7) 0.06
  ≥1000 87 54 (7.3) 10 (12.2) 18 (13.0) 3 (14.2)
 CMM 323 212 (23.7) 35 (29.9) 60 (27.3) 14 (31.2) 0.002
 Text 324 204 (22.8) 30 (25.6) 65 (29.6) 11 (24.4)
 CMM + text 322 259 (29.0) 22 (18.8) 31 (14.1) 10 (22.2)
 Control 324 218 (24.4) 30 (25.5) 64 (29.0) 10 (22.2)
*P-values refer to the comparisons across all exposure categories: using χ2 and Kruskal–Wallis tests.
Eighteen women did not report on whether having pre-ART or post-ART conception, all variables with the exception of age, parity, ANC visits, NNRTI vs. PI ART regimen and Intervention had missing data, that is, hemoglobin had 57% (MCAR), CD4 had 10% (MNAR), and viral load had 23% (MNAR), the proportion of missing data was uneven across the comparison groups. Women initiated on ART prior to conception had more missing hemoglobin data compared to those initiated on ART post conception. There were more missing data on CD4 and viral load among women initiated on ART postconception comparedwith those initiated on ART before conception.
Efavirenz, (EFV); Interquartile range, (IQR); Lamivudine, (3TC); Protease Inhibitor, (PI); SD; Tenofovir Disoproxil Fumarate, (TDF); Zidovudine, (AZT).

Comparison of Characteristics of APOs Between Women With Preconception and Postconception ART

Overall, there was no difference observed in types of APO by timing of ART initiation with the exception that women initiating ART postconception had higher rates of very LBW (VLBW) babies (Table 2).

TABLE 2. - Comparison of Adverse Pregnancy Outcomes Between Women Initiating ART Before Versus During Pregnancy
APO Outcomes Totals Initiation before Pregnancy n = 893 Initiation During Pregnancy n = 382 P*
Gestational age (wk)
 Term ≥37 wk 799 567 (70.8) 218 (70.3) 0.88
 Any preterm (<37) 329 234 (29.2) 92 (29.7) 0.88
 Late preterm (34–36) 162 119 (14.8) 41 (13.3) 0.57
 Moderately preterm (32–33) 73 57 (7.1) 16 (5.2) 0.28
 Very preterm (28–31) 61 38 (4.7) 22(7.1) 0.13
Birth weight
 Normal (≥2.5) 1026 731 (96.7) 280 (95.6) 0.36
 Any LBW (<2.5 kg) 38 25 (3.3) 13 (4.4) 0.36
 LBW (2500–1500) 68 47 (6.2) 20 (6.8) 0.78
 VLBW (<1500) 2 0 (0.0) 2 (0.7) 0.07
 Mean (SD) 7.3 (62.7) 3.3 (0.6)
 Stillbirth 33 25 (58.1) 8 (47.1) 0.56
 Miscarriage 11 10 (23.3) 1 (5.9) 0.15
*P values refer to unadjusted Fisher's exact comparisons between women who initiated ART before pregnancy vs. during pregnancy (not expanded into the 3 time periods). All the variables with an exception of moderately preterm, any LBW, VLBW, stillbirth, and miscarriage had <5% missing data, with similar proportions of missing data across comparison groups.

Most of the preterm deliveries were late preterm occurring between 34 and 36 weeks gestation, with preconception ART initiation having the highest number of late preterm deliveries 119(74.4%). In the postconception ART initiation group, most of the late preterm deliveries occurred among women who initiated on ART in the second trimester 30(18.8%), whereas third trimester ART initiation had the lowest numbers of preterm deliveries.

Most LBW deliveries 47(70.2%) occurred among women who initiated ART before conception. Among women initiating ART in the postconception period, mostof the LBW deliveries occurred in those who initiated ART in the first and second trimester (8[11.9] and 9[13.4%]). Only 2 infants were VLBW, and both were born to women who initiated ART postconception.

A total of 33 stillbirths occurred among women included in this analysis. Most of the stillbirths 25(75.7%) occurred among women who conceived while already on ART. Among women who initiated ART postconception, those initiating ART in the second trimester recorded majority of the stillbirths occurring 6(18.2%); and zero stillbirths occurred among women initiating ART in the third trimester.

Similarly, almost all miscarriages occurred among women who initiated ART before conception 10(90.9%) with only 1 occurring among women who initiated ART after conception and in the second trimester.

Risk Factors During Pregnancy Associated With an APO

Table 3 shows the unadjusted and adjusted risk factors associated with APOs. In multivariate models, home delivery (aRR 1.52; 95% CI: 1.48 to 1.56) and viral load ≥1000 copies/mL taken closest to pregnancy (aRR1.57; 95% CI: 1.54 to 1.60), increased the risk of having an adverse pregnancy outcome. Having a higher CD4 count between 350 and 500 was protective (aRR 0.68; 95% CI: 0.64 to 0.78).

TABLE 3. - Risk Factors Associated With Adverse Pregnancy Outcomes Among Women on ART in Kenya
No Adverse Pregnancy Outcome N (%) Median Adverse Pregnancy Outcome N (%) Median P Univariate RR (95% Confidence Interval) P Multivariable aRR (95% Confidence Interval) P
Maternal age (yr)
 ≤24 190 (30.0) 64 (16.5) 0.03 Reference Reference
 25–29 266 (29.4) 102 (26.3) 1.11 (0.97 to 1.23) 0.11 0.93 (0.60 to 1.45) 0.75
 ≥30 449 (49.6) 222 (57.2) 1.31 (1.14 to 1.50) <0.001 1.25 (0.95 to 1.64) 0.09
Place of delivery
 Home 45 (5.8) 32 (9.3) 0.03 1.39 (1.24 to 1.56) <0.001 1.52 (1.48 to 1.56) <0.001
 Hospital 736 (94.2) 312 (90.7) Reference Reference
Hemoglobin level, g/dL
 No anemia 134 (33.8) 44 (27.0) 0.02 Reference
 Mild anemia 85 (21.4) 25 (15.3) 0.92 (0.82 to 1.02) 0.12
 Moderate anemia 71 (17.9) 31 (19.0) 1.23 (1.08 to 1.38) 0.001
 Severe anemia 107 (26.9) 63 (38.7) 1.50 (1.02 to 2.18) 0.03
CD4 count
 ≤200 115 (14.2) 45 (12.9) 0.52 0.87 (0.71 to 1.07) 0.21 0.88 (0.83 to 1.10) 0.20
 201–350 146 (18.0) 63 (18.1) 0.93 (0.82 to 1.07) 0.37 0.96 (0.71 to 1.35) 0.80
 351–500 204 (25.2) 77 (22.2) 0.85 (0.72 to 1.01) 0.07 0.68 (0.64 to 0.78) <0.001
 ≥500 345 (42.6) 163 (46.8) Reference
Viral load (copies/mL)
 <1000 650 (92.2) 258 (89.0) 0.10 Reference Reference
 ≥1000 55 (7.8) 32 (11.0) 1.29 (1.23 to 1.35) <0.001 1.57 (1.54 to 1.60) <0.001
Differentiated ART initiation timing
 Preconception 612 (68.8) 281 (73.0) 0.05 Reference Reference
 First trimester 87 (9.8) 30 (7.8) 0.87 (0.59 to 1.29) 0.51 0.86 (0.61 to 1.22) 0.41
 Second trimester 153 (17.2) 67 (17.4) 0.97 (0.74 to 1.27) 0.85 1.50 (1.30 to 1.73) <0.001
 Third trimester 38 (4.2) 7 (1.8) 0.43 (0.19 to 0.95) 0.04 0.73 (0.60 to 0.88) 0.001
Intervention arms
 CMM alone 241 (26.6) 82 (21.1) <0.001 0.79 (0.69 to 0.92) 0.002 0.71 (0.67 to 0.77) <0.001
 Text alone 244 (27.0) 80 (20.6) 0.78 (069 to 0.86) <0.001 0.76 (0.66 to 0.89) 0.01
 CMM plus text 199 (22.0) 123 (31.7) 1.20 (1.11 to 1.30) <0.001 1.28 (1.03 to 1.59) 0.02
 Control 221 (24.4) 103 (26.6) Reference Reference

Initiating ART during the second trimester, postconception was associated with increased risk of APOs (aRR 1.50; 95% CI: 1.30 to 1.73), whereas initiation during the third trimester was protective (aRR 0.73; 95% CI: 0.60 to 0.88) when compared with preconception initiation.

Receiving the CMM intervention alone (aRR 0.71; 95% CI: 0.67 to 0.77) and text message intervention alone (aRR 0.76; 95% CI: 0.66 to 0.89) were protective of experiencing an adverse pregnancy outcome; however, the combined intervention of CMM plus text messages (aRR 1.28; 95% CI: 1.03 to 1.59) was associated with increased risk.

Adverse Birth Outcomes by ART Timing and Intervention Allocation

As shown in Table 4, no difference was observed in the composite pregnancy outcomes by timing of ART initiation after adjusting for age group, parity, and place of delivery. We further explored timing of ART before pregnancy versus during pregnancy on individual APO categories and did not find any significant associations (results not shown). Sensitivity analysis (Table 5) limited to women conceiving while on ART and those initiating ART in the first trimester of pregnancy showed no difference in the composite pregnancy outcomes by timing of ART initiation. However, in further subanalysis looking at individual APO categories, initiation of ART in the first trimester was associated with lower risk of having a LBW baby.

TABLE 4. - Adjusted Associations of Adverse Birth Outcomes With ART Timing and Intervention Allocation
Outcome  Preconception ART vs. Postconception (Ref Category is Before Pregnancy) ART Initiated During Pregnancy at 3 Time Points (Ref Category is First Trimester) Intervention Allocation (Ref Category is Control)
RR (95% CI) aRR (95% CI) P RR (95% CI) aRR (95% CI) P RR (95% CI) aRR (95% CI) P
PTD (<37 wk) 0.85 (0.64 to 1.14)
1.04 (0.87 to 1.23) 0.64 TM2 1.12 (0.59 to 2.09) 1.11 (0.66 to 1.87) 0.68 CMM 0.73 (0.67 to 0.79) 0.70 (0.62 to 0.80) <0.001
Text 0.82 (0.72 to 0.93) 0.79 (0.71 to 0.89) <0.001
TM3 0.38 (0.04 to 3.55) 0.51 (0.14 to 1.84) 0.30 Text + CMM 1.15 (0.92 to 1.43) 1.12 (0.92 to 1.35) 0.24
Any intervention 0.90 (0.81 to 0.99) 0.87 (0.76 to 0.99) 0.03
LBW (<2.5 kg) 1.22 (0.83 to 1.76) 1.04 (0.82 to 1.31) 0.77 TM2 1.24 (0.62 to 2.46) 1.86 (1.37 to 2.52) <0.001 CMM 1.72 (0.64 to 4.58) 1.52 (0.61 to 3.79) 0.36
Text 1.42 (0.33 to 6.06) 1.36 (0.34 to 5.49) 0.66
TM3 2.65 (0.90 to 7.83) 3.69 (1.43 to 9.53) 0.007 Text + CMM 1.29 (0.53 to 3.13) 1.37 (0.55 to 3.40) 0.49
Any intervention 1.48 (0.49 to 4.44) 1.42 (0.51 to 4.02) 0.49
Miscarriage 0.23 (0.01 to 4.75) 0.22 (0.01 to 5.96) 0.37 TM2 CMM 0.66 (0.17 to 2.61) 0.67 (0.16 to 2.72) 0.58
Text 0.36 (0.06 to 1.89) 0.33 (0.06 to 1.86) 0.21
TM3 Text + CMM 1.67 (0.67 to 4.20) 1.68 (0.64 to 4.39) 0.29
Any intervention 0.89 (0.53 to 1.48) 0.89 (0.52 to 1.51) 0.67
Stillbirth 0.74 (0.29 to 1.91)
0.75 (0.32 to 1.79) 0.52 TM2 1.59 (0.33 to 7.69) 1.55 (0.33 to 7.69) 0.57 CMM 1.12 (0.69 to 1.82) 1.15 (0.69 to 1.93) 0.57
Text 0.5 (0.14 to 1.84) 0.50 (0.13 to 1.83) 0.29
TM3 Text + CMM 1.50 (0.59 to 3.86) 1.51 (0.54 to 4.18) 0.42
Any intervention 1.04 (0.59 to 1.84) 1.05 (0.59 to 1.87) 0.85
0.86 (0.72 to 1.03) 1.01 (0.87 to 1.16) 0.87 TM2 1.11 (0.66 to 1.85) 1.02 (0.64 to 1.63) 0.90 CMM 0.79 (0.69 to 0.92) 0.74 (0.71 to 0.76) <0.001
Text 0.77 (0.69 to 0.86) 0.79 (0.70 to 0.89) <0.001
TM3 0.49 (0.15 to 1.61) 0.65 (0.32 to 1.33) 0.24 Text + CMM 1.20 (1.11 to 1.30) 1.13 (0.97 to 1.32) 0.11
Any intervention 0.92 (0.89 to 0.95) 0.88 (0.81 to 0.97) 0.01
Comparison A: preconception ART vs. ART initiated during pregnancy. Comparison B: ART initiated during pregnancy at 3 time points: TM1 (first trimester), TM2 (second trimester), and TM3 (third trimester). Comparison C: intervention allocation: adjusted for age group, parity, and place of delivery.

TABLE 5. - Sensitivity Analysis: Adjusted Associations of Adverse Birth Outcomes With ART Timing Restricted to Preconception and First Trimester
Outcome Preconception ART vs. First Trimester*
RR (95% CI) aRR (95% CI) P
PTD (<37 wk) 0.91 (0.55 to 1.49) 0.9 (0.63 to 1.30) 0.59
LBW (<2.5 kg) 0.91 (0.35 to 2.36) 0.56 (0.44 to 0.71) <0.001
Stillbirth 1.14 (0.34 to 3.77) 1.06 (0.35 to 3.16) 0.9
Any APO 0.87 (0.59 to 1.29) 1.09 (0.82 to 1.44) 0.53
*Ref category is preconception.

Similarly, among women initiating ART in the postconception period, there was no difference in the composite adverse pregnancy outcome during the different trimesters even after adjusting for age, parity, and place of delivery. However, women initiated on ART in the second and third trimester had a higher risk of LBW infants (aRR 1.86: 95% CI: 1.37 to 2.52) and (aRR 3.69: 95% CI: 1.43 to 9.53), respectively (Table 4).

Women who were randomized to receive CMM visits alone and text messages alone had lower risks of experiencing any APO (aRR 0.74: 95% CI: 0.71 to 0.76) and (aRR 0.79: 95% CI: 0.70 to 0.89), respectively. Specifically, women receiving these interventions had decreased risk of experiencing a PTD (aRR 0.70: 95% CI: 0.62 to 0.80) and (aRR 0.79: 95% CI: 0.71 to 0.89), respectively. There was however no effect of the interventions on LBW, miscarriages, and stillbirths.


In this report of 1275 pregnant WLWH on ART in Kenya receiving HIV and antenatal services at public sector facilities, we found nearly 1 in 3 women experienced an APO, the majority being preterm birth. This high rate of APO has been observed in other studies conducted in similar settings, which have shown between 33.7% and 40% APO among pregnant women on ART.16,31 Unsuppressed HIV viral load, and moderate and severe anemia at baseline were associated with increased risk of adverse pregnancy outcome. Although over 70% of women were on ART before pregnancy, we found no appreciable differences in adverse composite pregnancy outcomes when compared with those initiating ART during pregnancy. However, in sensitivity and subgroup analysis, initiation of ART in the first trimester was associated with lower risk of having a LBW baby when compared with conceiving while on ART and initiation of ART in the second and third trimester postconception as compared with earlier in pregnancy bore a high risk of delivering a LBW baby. Receiving home visits by peer lay health workers (CMMs) or receiving weekly text messages promoting ART adherence and maternal and child health services decreased the risk of experiencing a PTD in this cohort. Surprisingly, the combination of the 2 interventions increased the risk of experiencing an APO.

Risk factors associated with APOs in this report have also been identified in other studies. High HIV viral load is directly related to risk of HIV transmission to infants and APOs such as pregnancy losses, prematurity, and perinatal death.32,33 Anemia has previously been identified as a major risk factor for APOs, especially in malaria-endemic regions such as this.34,35 Several studies conducted in sub-Saharan Africa have reported that HIV-infected women are at increased risk of maternal anemia and APOs.36,37 This finding calls for further anemia mitigating strategies among HIV-infected women so as to improve their overall health and improve birth outcomes.

Our finding in the main analysis of no difference in incidence of APOs in women with HIV regardless of timing of ART initiation is reassuring but inconsistent with several previous meta-analysis of studies from African populations that have shown that the pooled incidence of PTD is 20% higher and risk of LBW infant 30% higher among those on ART before conception.38 These contrasting findings may be related to differences in sample sizes or to differences in other background risk factors for APO among these varied cohorts.39 In sensitivity analysis on individual APO categories and limited to those conceiving while on ART and those initiating ART in the first trimester of pregnancy, initiation of ART in the first trimester was associated with lower risk of having a LBW baby when compared with women conceiving while on ART which is in line with studies performed in similar settings mentioned above. However, this finding is important in that we provide critical data on effect of periconception exposure to ART on pregnancy outcomes, which is largely understudied,40 and is particularly relevant to programs that are introducing newer ART regimen such as dolutegravir. Still, it is important to note that the controversial association of ART during pregnancy and APOs can be partially explained by the fact that investigating the effect of ART on adverse outcomes in isolation is often prone to confounding by indication. The duration of ART exposure may be a proxy for duration of HIV infection and not necessarily an indication of medication effect on outcomes. It is likely, for example, that women conceiving while on ART have a history of more severe HIV illness.

In our cohort, nearly one-quarter of women experienced a preterm birth. These findings raise concern because of the fact that preterm birth is the leading cause of mortality for children under 5 years of age in low-income and middle-income countries.41,42 Nonetheless, a large proportion of the preterm births occurred later in gestation >32 weeks, which is somewhat reassuring. We also found that postconception ART in the second and third trimester was associated with increased risk of experiencing LBW babies. Similar to other studies we identified a significant number of fatal pregnancy losses (stillbirths and miscarriages).12,13,43 This is particularly worrying because it is likely that the number of miscarriages in this study is underestimated given that the average gestational age at study enrollment was 24 weeks; which is after most miscarriages occur, by definition.

We found that supportive behavioral interventions provided in the parent MOTIVATE Study, when provided as separate interventions were protective in reducing the odds of APOs among those receiving them by nearly 20%. Although the exact mechanism by which these interventions reduced adverse outcomes is not known, other studies using mHealth and peer led interventions have demonstrated that these interventions can provide timely, relevant, useful, and supportive information to pregnant women and new mothers resulting in improved maternal, fetal, and infant health outcomes.44,45 In the same vein, our interventions encouraged uptake of important maternal and child health behaviors and services (good nutrition, antenatal clinic attendance, birth planning, and skilled delivery) and engagement and adherence to HIV care, potentially increasing engagement in health care services known to improve maternal and infant outcomes. Of note, however, the combination of CMM and text messages in this trial provided conflicting as opposed to synergistic effects. This might be related to “information overload” which often results in recipients ignoring both interventions due to exhaustion or “cherry-picking” parts of the interventions which inhibits the ability to experience the full benefit of the intervention.46 This deleterious effect of multiple interventions presented at the same time points toward the need for interventions to be delivered in an incremental, sequential, and adaptive manner as opposed to an “overload” or a one size fits all approach.

This study, which uses routine programmatic data for surveillance of APO among WLWH using ART, provides a good representation of the current landscape for pregnant WLWH in sub-Saharan Africa, where most are on ART at the time of conception or initiated rapidly during pregnancy. The utilization of routine programmatic data for APO surveillance is in line with the WHO consolidated recommendations for ART drug toxicity monitoring to provide program managers with the information needed to improve services in real time.47,48 The study was, however, not without limitations. For example, although the parent study was a cluster-randomized trial, the results presented are observational in nature and may be limited by residual biases. We do not, for example, have a comparable group of women without HIV to determine if APOs are higher in women with HIV on ART than the general population. Furthermore, we acknowledge that the use of the last normal menstrual period to date pregnancies may overestimate the preterm birth rate, which can lead to biased associations.49 Given the real world setting of the research and nature of the data collected, it was prone to missing data which we mitigated by constant review and validation by the coordinator, investigators, and statistician. In addition, we do not have data available on incidence of malaria, sexually transmitted infections including syphilis, and malnutrition, which may contribute to APOs. We also did not distinguish between spontaneous versus provider-initiated preterm deliveries, which is critical in research that examines obstetrical outcomes. Provider-initiated preterm deliveries, however, are generally very infrequent in public facilities where the study was conducted as only 4 sites had neonatal intensive care or capability to perform cesarean sections. Aggregating spontaneous and indicated preterm deliveries into a single outcome has the potential to bias associations toward the null as conditions that increase the risk of spontaneous PTD (eg, infection, inflammation, multiple fetuses, and previous spontaneous preterm birth) are distinct from those that typically lead to provider-initiated PTD (eg, hypertensive disorders, intrauterine fetal demise, placenta previa, or accreta). Finally, this study used data abstracted from medical records alone, so we were not able to take into consideration important psychosocial predictors of APO, such as socioeconomic status, intimate partner violence, or depression. Most women initiating ART postconception started antenatal care during or after the second trimester, limiting detection of miscarriages in this group. Finally, we could not establish pregnancy outcomes of some of the women who were lost to follow-up.


This study expands the existing limited literature on APOs among WLWH in 2 ways. First, the study contributes to the knowledge base by demonstrating that supportive behavioral interventions may mitigate the risk of experiencing an APO among WLWH on ART. Second, we demonstrate that programmatic data for surveillance of APO among women in HIV high-burden settings using ART in pregnancy for HIV treatment and prevention (eg, PrEP) is a rich resource of data that can be used to systematically monitor outcomes of pregnancy exposure. With the number of women in HIV high-burden settings using ART in pregnancy for HIV treatment and prevention increasing and the general reluctance to include pregnant women in ART drug safety and efficacy studies, we recommend the use of programmatic data for surveillance of APO including for the newer ART regimens. Additional research studies that use mixed methods are needed to investigate the effectiveness and mechanisms of action of these behavioural interventions in reducing APO not only in WLWH on ART but also pregnant women on PrEP.


The authors are grateful to all the women who participated in this study. The authors acknowledge the large team of staff, collaborators, local communities, and community advisory boards who contributed, supported, and provided guidance during the study. The authors acknowledge the support of the Kenya Medical Research Institute (KEMRI), the Director of KEMRI, and the Director of the Centre for Microbiology Research.


1. World Health Organization. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection. Geneva, Switzerland: World Health Organization; 2013.
2. Tsague L, Abrams EJ. Commentary: antiretroviral treatment for pregnant and breastfeeding women-the shifting paradigm. AIDS. 2014;28(suppl 2):S119–S121.
3. UNAIDS. Global HIV & AIDS Statistics—2019 Fact Sheet. Geneva, Switzerland: UNAIDS; 2019. Available at: Accessed September 25, 2019.
4. United Nations Children's Fund. For Every Child, End AIDS—Seventh Stocktaking Report. New York, NY: UNICEF; 2016.
5. National AIDS Control Council. Kenya HIV Estimates—Report 2018. Nairobi, Kenya: Ministry of Health; 2018.
6. UNICEF, UNAIDS, and WHO. Key Considerations for Programming and Prioritization. Going the “Last Mile” to EMTCT: A Road Map for Ending the HIV Epidemic in Children. UNICEF: New York, NY; 2020.
7. World Health Organization. WHO Technical Brief: Preventing HIV during Pregnancy and Breastfeeding in the Context of Pre-exposure Prophylaxis (PrEP). Geneva, Switzerland: World Health Organization; 2017.
8. Robinson JL, Hughes C. Prevention of in utero and intrapartum transmission of human immunodeficiency virus infection. JAMMI. 2018;3:155–160.
9. Maartens G, Celum C, Lewin SR. HIV infection: epidemiology, pathogenesis, treatment, and prevention. Lancet. 2014;384:258–271.
10. Sibiude J, Mandelbrot L, Blanche S, et al. Association between prenatal exposure to antiretroviral therapy and birth defects: an analysis of the French perinatal cohort study (ANRS CO1/CO11). PLoS Med. 2014;11:e1001635.
11. Malaba TR, Phillips T, Le Roux S, et al. Antiretroviral therapy use during pregnancy and adverse birth outcomes in South African women. Int J Epidemiol. 2017;46:1678–1689.
12. Li N, Sando MM, Spiegelman D, et al. Antiretroviral therapy in relation to birth outcomes among HIV-infected women: a cohort study. J Infect Dis. 2016;213:1057–1064.
13. Mofenson LM. Antiretroviral therapy and adverse pregnancy outcome: the elephant in the room?. J Infect Dis. 2016;213:1051–1054.
14. Chen JY, Ribaudo HJ, Souda S, et al. Highly active antiretroviral therapy and adverse birth outcomes among HIV-infected women in Botswana. J Infect Dis. 2012;206:1695–1705.
15. Lopez M, Figueras F, Hernandez S, et al. Association of HIV infection with spontaneous and iatrogenic preterm delivery: effect of HAART. AIDS. 2012;26:37–43.
16. Stringer EM, Kendall MA, Lockman S, et al. Pregnancy outcomes among HIV-infected women who conceived on antiretroviral therapy. PLoS One. 2018;13:e0199555.
17. Ekouevi DK, Coffie PA, Becquet R, et al. Antiretroviral therapy in pregnant women with advanced HIV disease and pregnancy outcomes in Abidjan, Cote d'Ivoire. AIDS. 2008;22:1815–1820.
18. Snijdewind IJM, Smit C, Godfried MH, et al. Preconception use of cART by HIV-positive pregnant women increases the risk of infants being born small for gestational age. PLoS One. 2018;13:e0191389.
19. UN Committee on the Rights of the Child (CRC). General comment No. 20 (2016) on the implementation of the rights of the child during adolescence; New York, NY; United Nations, 2016. Available at:
20. Short CE, Douglas M, Smith JH, et al. Preterm delivery risk in women initiating antiretroviral therapy to prevent HIV mother-to-child transmission. HIV Med. 2014;15:233–238.
21. Zash R, Souda S, Chen JY, et al. Reassuring birth outcomes with tenofovir/emtricitabine/efavirenz used for prevention of mother-to-child transmission of HIV in Botswana. J Acquir Immune Defic Syndr. 2016;71:428–436.
22. Msukwa MT, Keiser O, Jahn A, et al. Timing of combination antiretroviral therapy (cART) initiation is not associated with stillbirth among HIV-infected pregnant women in Malawi. Trop Med Int Health. 2019;24:727–735.
23. Chetty T, Thorne C, Coutsoudis A. Preterm delivery and small-for-gestation outcomes in HIV-infected pregnant women on antiretroviral therapy in rural South Africa: results from a cohort study, 2010-2015. PLoS One. 2018;13:e0192805.
24. Pintye J, Baeten JM, Celum C, et al. Maternal tenofovir Disoproxil fumarate use during pregnancy is not associated with adverse perinatal outcomes among HIV-infected east african women: a prospective study. J Infect Dis. 2017;216:1561–1568.
25. Kinuthia J, Pintye J, Abuna F, et al. Pre-exposure prophylaxis uptake and early continuation among pregnant and post-partum women within maternal and child health clinics in Kenya: results from an implementation programme. Lancet HIV. 2020;7:e38–e48.
26. Odeny TA, Onono M, Owuor K, et al. Maximizing adherence and retention for women living with HIV and their infants in Kenya (MOTIVATE! study): study protocol for a randomized controlled trial. Trials. 2018;19:77.
27. Santana DS, Silveira C, Costa ML, et al. Perinatal outcomes in twin pregnancies complicated by maternal morbidity: evidence from the WHO Multicountry Survey on Maternal and Newborn Health. BMC Pregnancy Childbirth. 2018;18:449.
28. Organisation WH. Stillbirths. nd; 2020, Available at: Accessed May 14, 2020.
29. World Health Organization. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Vitamin and Mineral Nutrition Information System. WHO/NMH/NHD/MNM/111. Geneva, Switzerland; World Health Organization; 2011. Available at Accessed May 14, 2020.
30. Ministry of Health NAaSCP. Guidelines on Use of Antiretroviral Drugs for Treating and Preventing HIV Infection in Kenya. Nairobi: Kenya Ministry of Health, National AIDS and STI Control Programme; 2016.
31. Fowler MG, Qin M, Fiscus SA, et al. Benefits and risks of antiretroviral therapy for perinatal HIV prevention. New Engl J Med. 2016;375:1726–1737.
32. Ezechi OC, Gab-Okafor CV, Oladele DA, et al. Pregnancy, obstetric and neonatal outcomes in HIV positive Nigerian women. Afr J Reprod Health. 2013;17:160–168.
33. Kim HY, Kasonde P, Mwiya M, et al. Pregnancy loss and role of infant HIV status on perinatal mortality among HIV-infected women. BMC Pediatr. 2012;12:138.
34. Rahman MM, Abe SK, Rahman MS, et al. Maternal anemia and risk of adverse birth and health outcomes in low- and middle-income countries: systematic review and meta-analysis. Am J Clin Nutr. 2016;103:495–504.
35. Fowkes FJI, Moore KA, Opi DH, et al. Iron deficiency during pregnancy is associated with a reduced risk of adverse birth outcomes in a malaria-endemic area in a longitudinal cohort study. BMC Med. 2018;16:156.
36. Mehta S, Manji KP, Young AM, et al. Nutritional indicators of adverse pregnancy outcomes and mother-to-child transmission of HIV among HIV-infected women. Am J Clin Nutr. 2008;87:1639–1649.
37. González R, Rupérez M, Sevene E, et al. Effects of HIV infection on maternal and neonatal health in southern Mozambique: a prospective cohort study after a decade of antiretroviral drugs roll out. PLoS One. 2017;12:e0178134.
38. Uthman OA, Nachega JB, Anderson J, et al. Timing of initiation of antiretroviral therapy and adverse pregnancy outcomes: a systematic review and meta-analysis. Lancet HIV. 2017;4:e21–e30.
39. Theron G, Brummel S, Fairlie L, et al. Pregnancy outcomes of women conceiving on antiretroviral therapy (ART) compared to those commenced on ART during pregnancy. Clin Infect Dis. 2020:ciaa805 [epub ahead of print].
40. Lockman S, De Gruttola V. Outcomes following pregnancy conception on antiretroviral therapy: a call for more data. Clin Infect Dis. 2019;68:280–281.
41. Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet. 2012;379:2162–2172.
42. UNICEF. Under-Five Mortality—UNICEF DATA 2019 2019. Available at: Accessed December 7, 2019, 2019.
43. Veroniki AA, Antony J, Straus SE, et al. Comparative safety and effectiveness of perinatal antiretroviral therapies for HIV-infected women and their children: systematic review and network meta-analysis including different study designs. PLoS One. 2018;13:e0198447.
44. Coleman J, Eriksen J, Black V, et al. The mobile alliance for maternal action text message-based mHealth intervention for maternal care in South Africa: qualitative user study. JMIR Hum Factors. 2020;7:e14078.
45. Wynn A, Rotheram-Borus MJ, Leibowitz AA, et al. Mentor mothers program improved child health outcomes at A relatively low cost in South Africa. Health Aff (Millwood). 2017;36:1947–1955.
46. Murayama K, Blake AB, Kerr T, et al. When enough is not enough: information overload and metacognitive decisions to stop studying information. J Exp Psychol Learn Mem Cogn. 2016;42:914–924.
47. World Health Organization. WHO Implementation Tool for Monitoring the Toxicity of New Antiretroviral and Antiviral Medicines in HIV and Viral Hepatitis Programmes. Geneva, Switzerland: World Health Organization; 2018.
48. World Health Organization. Consolidated HIV Strategic Information Guidelines: Driving Impact through Programme Monitoring and Management. Geneva, Switzerland: World Health Organization; 2020.
49. Rappazzo KM, Lobdell DT, Messer LC, et al. Comparison of gestational dating methods and implications for exposure-outcome associations: an example with PM2.5 and preterm birth. Occup Environ Med. 2017;74:138–143.

Africa; antiretroviral therapy; birth outcome; HIV; risk factors; pregnancy

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.