Adverse Perinatal Outcomes Associated With HAART and Monotherapy

Clara Portwood; Claudia Murray; Harriet Sexton; Mary Kumarendran; Zoe Brandon; Bradley Johnson; Shona Kirtley; Joris Hemelaar

Disclosures

AIDS. 2022;36(10):1409-1427.

Abstract and Introduction

Abstract

Objectives: Assess adverse perinatal outcomes in pregnant women living with HIV (WLHIV) receiving HAART or zidovudine (ZDV) monotherapy, compared with antiretroviral therapy (ART)-naive WLHIV and HIV-negative women.

Design: Systematic review and meta-analysis.

Methods: We conducted a systematic literature review by searching PubMed, CINAHL, Global Health, and EMBASE for studies published between 1 January 1980 and 20 April 2020. We included studies reporting on the association of pregnant WLHIV receiving HAART or ZDV monotherapy with 11 perinatal outcomes: preterm birth (PTB), very PTB, spontaneous PTB (sPTB), low birth weight (LBW), very LBW, term LBW, preterm LBW, small for gestational age (SGA), very SGA (VSGA), stillbirth, and neonatal death. Random-effects meta-analyses were conducted.

Results: Sixty-one cohort studies assessing 409 781 pregnant women were included. WLHIV receiving ZDV monotherapy were associated with a decreased risk of PTB [relative risk 0.70, 95% confidence interval (CI) 0.62–0.79] and LBW (0.77, 0.67–0.88), and comparable risk of SGA, compared with ART-naive WLHIV. WLHIV receiving ZDV monotherapy had a comparable risk of PTB and LBW, and an increased risk of SGA (1.16, 1.04–1.30) compared with HIV-negative women. In contrast, WLHIV receiving HAART were associated with a comparable risk of PTB and LBW, and increased risk of SGA (1.38, 1.09–1.75), compared with ART-naive WLHIV. WLHIV receiving HAART were associated with an increased risk of PTB (1.55, 1.38–1.74), sPTB (2.09, 1.48–2.96), LBW (1.79, 1.51–2.13), term LBW (1.88, 1.23–2.85), SGA (1.80,1.34–2.40), and VSGA (1.22, 1.10–1.34) compared with HIV-negative women.

Conclusion: Pregnant WLHIV receiving HAART have an increased risk of a wide range of perinatal outcomes compared with HIV-negative women.

Introduction

In 2020, 37.7 million people worldwide were living with HIV, including 19.3 million women of childbearing age.^[1] Each year, an estimated 1.3 million women living with HIV (WLHIV) are pregnant, of whom the vast majority reside in sub-Saharan Africa.^[1] Antiretroviral therapy (ART)-naive maternal HIV infection is associated with an increased risk of preterm birth (PTB), low birthweight (LBW), small for gestational age (SGA), and stillbirth compared with HIV-negative women.^[2] Adverse perinatal outcomes are major contributors to neonatal and child mortality and morbidity, with the highest rates found in sub-Saharan Africa.^[3–5] The United Nations' Sustainable Development Goal 3 (SDG3) target 3.2 aims to reduce neonatal and under-5 mortality to 12 and 25 per 1000 live births, respectively, by 2030.^[6]

In 1994, a randomized controlled trial (RCT) demonstrated that antenatal zidovudine (ZDV) monotherapy reduced the risk of mother-to-child HIV transmission.^[7] In the past, WHO guidelines recommended ZDV monotherapy in pregnant WLHIV to prevent vertical HIV transmission (option A), or HAART for pregnant WLHIV requiring treatment for their own health as well as prevention of vertical HIV transmission (option B).^[8] From 2013, WHO recommended that all pregnant WLHIV should receive HAART during pregnancy.^[9] In 2015, this recommendation was updated so that all people living with HIV should initiate HAART as soon as possible after diagnosis, including pregnant WLHIV.^[10] Consequently, the proportion of pregnant WLHIV receiving ZDV monotherapy decreased from 31 to 0% in the period 2011–2020, whereas the proportion of pregnant WLHIV receiving HAART increased from 27 to 83% during the same period.^[1] These trends were accompanied by a 41% reduction in vertical HIV transmission globally during 2010–2018.^[11] However, the impact of these changes in antenatal ART regimens on other important perinatal outcomes is unknown.

Several studies suggest adverse perinatal outcomes are associated with ART exposure during pregnancy but evidence is conflicting regarding different regimens.^[12–15] A recent network meta-analysis of seven randomized controlled trials (RCTs), which compared ART regimens initiated during pregnancy, showed that a number of HAART regimens were associated with an increased risk of LBW, very LBW and PTB, compared with ZDV monotherapy.^[16] Some cohort studies report that HAART exposure is associated with increased risk of PTB and LBW in pregnant WLHIV, whereas ZDV monotherapy is not.^[17,18] However, others report no significant association.^[19]

As the number of pregnancies exposed to HAART increases and ZDV monotherapy has been phased out, it is important to understand the effects of HAART and ZDV monotherapy on perinatal outcomes in WLHIV, and whether either therapy restores the risk of perinatal outcomes to levels seen in HIV-negative women. We conducted a systematic review and meta-analysis of cohort studies examining the risk of 11 specific perinatal outcomes in WLHIV receiving HAART or ZDV monotherapy, compared with ART-naive WLHIV and HIV-negative women.

1 2 3 4

Next Section

Abstract and Introduction
Methods
Results
Discussion
References

Table 1. Characteristics of studies included in the systematic review and meta-analysis.
Study: first author (year) [ref]	Country	Country income status	Cohort study design	Recruitment period	Number of women analysed	Population characteristics^a	Method to correct for confounders	Method to estimate gestational age	Quality assessment
Ai-Jie (2013) [26]	China	Middle	Retrospective	1/2006 to 3/2008	155	Twins excluded, rural and urban setting	None	No description	Poor
Albert (2020) [27]	Canada	High	Retrospective	1/1/1997 to 31/1/2018	477	Twins excluded, women recruited from a provincial surveillance database, 46.1% smoking, 23.3% alcohol use, 26% IDU	Risk factor analysis	Ultrasound in first and/or second trimester	Average
Azria (2009) [28]	France	High	Retrospective	1/2003 to 6/2007	300	Twins excluded, women recruited from a level III maternity unit, urban setting, hospital deliveries, 4.3% smoking during pregnancy, 1.7% history of IDU	Risk factor analysis, matching	First day of LNMP, corrected if needed by routine first trimester ultrasound	Average
Bailey (2013) [29]	Ukraine	Middle	Retrospective	2008 to 2010	3535	First born twin included, hospital deliveries, 14.7% history of IDU	None	LNMP and ultrasound (unspecified)	Poor
Balogun (2018) [30]	Canada	High	Prospective	9/2010 to 12/2015	104	Twins excluded, women recruited from four sites in Toronto, 0% smoking	Risk factor analysis, matching	LNMP confirmed by ultrasound (unspecified)	Average
Bengtson (2020) [31]	South Africa	Middle	Prospective	3/2013 to 8/2015	1116	Twins excluded, women recruited from antenatal care clinics in Gugulethu Cape Town, urban setting, 17.2% alcohol use	None	Ultrasound (unspecified), LNMP, or symphysis-fundal height	Poor
Boer (2006) [32]	Netherlands	High	Retrospective	12/1997 to 7/2003	294	First born twin included, women recruited from an academic medical centre, 12.9% smoking, 1.7% history of IDU	Regression analysis, matching	LMNP confirmed by first trimester ultrasound	Poor
Boyajian (2012) [33]	Canada	High	Retrospective	1/1/2003 to 10/1/2010	364	Second twin excluded, women recruited from tertiary pregnancy referral centre, hospital deliveries, 6.3% smokers, 1.4% IDU	Regression analysis, risk factor analysis, matching	No description	Average
Carceller (2009) [34]	Canada	High	Retrospective	1997 to 2005	412	Recruited from a tertiary hospital in Montreal, urban setting, hospital deliveries	None	No description	Poor
Chagomerana (2017) [35]	Malawi	Low	Retrospective	1/4/2012 to 15/11/2015	3074	Twins excluded, urban setting, hospital deliveries	Regression analysis	LNMP	Average
Chen (2012) [36]	Botswana	Middle	Retrospective	1/5/2009 to 30/4/2011	33148	First born twin included, hospital deliveries, 5.3% alcohol use, 1.7% smoking	Regression analysis, risk factor analysis	LNMP, symphysis-fundal height, or ultrasound (unspecified)	Average
Chibwesha (2016) [37]	Zambia	Low	Retrospective	1/2/2006 to 31/12/2012	200557	First born twin included, women recruited from MNCH health system, urban setting	None	LNMP and symphysis-fundal height	Poor
Cooper (2002) [38]	USA	High	Prospective	1/1990 to 6/2000	1542	Twins excluded, 31% IDU	Risk factor analysis	LNMP, ultrasound (unspecified), symphysis-fundal height, or neonatal assessment (unspecified)	Poor
Cotter (2006) [39]	USA	High	Prospective	1/1990 to 12/2002	1337	Twins excluded, 5.4% alcohol use, 11.2% smoking, 17.8% IDU, women recruited from medical centre, hospital deliveries	Regression analysis, risk factor analysis	LNMP and/or ultrasound (unspecified)	Poor
Dadabhai (2019) [40]	Malawi	Low	Prospective	1/2016 to 9/2017	1299	Twins excluded, 96% of deliveries occurred in healthcare facilities, urban setting	Regression analysis	Ballard score and LNMP	Average
De Souza (2000) [41]	USA	High	Retrospective	1/1/1990 to 31/12/1994	403	First born twin included, women recruited from a tertiary hospital, 18.9% IDU	Risk factor analysis	No description	Average
Garcia-Otero (2019) [42]	Spain	High	Prospective	12/2014 to 3/2017	94	Women recruited from hospital and hospital clinic, urban setting, 20.2% smoking, 3.2% IDU	Risk factor analysis	No description	Average
Goetghebuer (2019) [43]	Belgium	High	Prospective	12/2010 to 11/2013	255	Women recruited from hospital antenatal clinic, urban setting, 9.2% smoking, 10.1% alcohol use	Risk factor analysis	Ballard score	Average
Haeri (2009) [44]	USA	High	Retrospective	1/2000 to 1/2007	453	Women recruited from two tertiary care centres, 13.3% smoking	Regression analysis, risk factor analysis, matching	LNMP and ultrasound (unspecified)	Average
Hernandez (2017) [45]	Spain	High	Prospective	6/2006 to 12/2007	56	Twins excluded, women recruited from maternofetal medicine department of hospital, urban setting, 25% smoking, 0% alcohol use, 0% IDU	Risk factor analysis, matching	No description	Average
Hu (2019) [46]	China	Middle	Prospective	10/2009 to 5/2018	802	Twins included, urban setting	Regression analysis, risk factor analysis	First or second trimester ultrasound, in the absence of ultrasound LNMP used	Average
Joseph (2011) [47]	Nigeria	Middle	Retrospective	1/2008 to 6/2009	249	Twins excluded, women recruited from a tertiary referral centre, hospital deliveries	Risk factor analysis	No description	Average
Jumare (2019) [48]	Nigeria	Middle	Prospective	2013 to 2017	424	Twins included, women recruited from a specialist hospital, urban setting	Risk factor analysis	LNMP	Average
Kakkar (2015) [49]	Canada	High	Prospective	1988 to 2011	589	Twins excluded, women recruited from a tertiary referral centre and the largest maternal-health centre in the province	Regression analysis, risk factor analysis	LNMP and ultrasound (unspecified)	Average
Kowalska (2003) [50]	Poland	Middle	Prospective	1/1995 to 2/2003	102	Twins included, women recruited from an outpatient HIV clinic, 47.1% IDU	Risk factor analysis	LNMP	Poor
Li (2016) [13]	Tanzania	Low	Prospective	11/2004 to 9/2011	3314	Women recruited from hospitals, health centres and dispensaries, urban setting	Risk factor analysis	LNMP and symphysis-fundal height	Poor
Li (2020) [51]	China	Middle	Prospective	10/2014 to 9/2017	1449	Twins excluded, women recruited from midwifery hospitals	Regression analysis, risk factor analysis	LNMP or ultrasound (unspecified)	Average
Liff (2020) [52]	Botswana	Middle	Prospective	4/2016 to 4/2017	179	Twins excluded, women recruited from 8 nationwide delivery sites	Risk factor analysis	Second trimester ultrasound	Poor
Lopez (2012) [53]	Spain	High	Retrospective	1/1986 to 6/2010	1557	Twins excluded, women recruited from a tertiary hospital, urban setting, hospital deliveries, 55.2% smoking	Regression analysis, risk factor analysis, matching	Second trimester ultrasound	Poor
Malaba (2017) [54]	South Africa	Middle	Prospective	4/2013 to 8/2015	1793	Twins excluded, recruited from large community primary care facility, urban setting	Regression analysis, risk factor analysis	LNMP and symphysis-fundal height	Average
Malaba (2018) [55]	South Africa	Middle	Prospective	4/2014 to 10/2016	1787	Twins excluded, women recruited from a large primary care antenatal clinic, urban setting	Regression analysis	LNMP and symphysis-fundal height	Average
Mandelbrot (1998) [56]	France	High	Retrospective	1/9/1985 to 31/12/1996	2834	Twins excluded, 31% IDU, recruited from obstetrical services, hospital deliveries	None	LNMP, confirmed by first trimester ultrasound	Poor
Marazzi (2011) [57]	Malawi and Mozambique	Low	Retrospective	7/2005 to 6/2009	3273	Twins included, women recruited from DREAM centres	Regression analysis	LNMP and clinical exam (unspecified)	Average
Marti (2007) [58]	Spain	High	Prospective	1/1/1997 to 31/12/2003	167	Twins excluded, women recruited from hospital, hospital deliveries, urban setting, 1% IDU	None	No description	Poor
Matheson (1995) [59]	USA	High	Prospective	3/1986 to 12/1993	321	Twins excluded, 41.7% IDU	Risk factor analysis	Ballard score	Average
Mehta (2019) [60]	South Africa	Middle	Retrospective	7/10/2013 to 6/10/2014	10293	Twins included, women recruited from hospital, urban setting, hospital deliveries, 0.09% smoking, 0.2% alcohol use, 0.04% IDU	Risk factor analysis	LNMP, ultrasound (unspecified)	Average
Moodley (2016) [61]	South Africa	Middle	Retrospective	7/2011 to 12/2011, 1/2014 to 6/2014	9847	Twins excluded, data abstracted from maternity registers of a regional hospital	Regression analysis, risk factor analysis	LNMP and/or ultrasound (unspecified)	Average
Moseholm (2019) [62]	Denmark	High	Retrospective	1/1/2000 to 31/12/2016	2980	Twins excluded, women recruited from specialized clinical centres for treatment and care of pregnant women living with HIV, 7.6% smoking during pregnancy	Risk factor analysis, matching	No description	Average
Olagbuji (2010) [63]	Nigeria	Middle	Prospective	1/2007 to 12/2008	406	Twins excluded, women recruited from a tertiary referral centre, all delivered in a healthcare facility	Risk factor analysis	No description	Poor
Phiri (2015) [64]	USA	High	Retrospective	1/1/1994 to 31/12/2009	790	6.7% alcohol use, 25% smoking, 11% IDU	Regression analysis	LNMP, ultrasound (unspecified), and clinical assessment	Poor
Ramokolo (2017) [65]	South Africa	Middle	Retrospective	10/2012 to 5/2013	8778	Women recruited from primary health facilities	Risk factor analysis	LNMP	Average
Rempis (2017) [66]	Uganda	Low	Retrospective	2/2013 to 12/2013	412	Twins excluded, all deliveries in a private referral hospital	Risk factor analysis	No description	Poor
Rudin (2011) [67]	Switzerland	High	Prospective	1984 to 2007	1040	Twins excluded, 22% smoking, 26% IDU	None	No description	Poor
Santosa (2019) [68]	South Africa	Middle	Prospective	28/5/2013 to 20/7/2016	633	Twins excluded, women recruited from hospital, 98.7% hospital deliveries, urban setting, 6.4% smoking, 8.2% alcohol	Regression analysis, risk factor analysis	Ultrasound <14 weeks	Good
Saums (2019) [69]	USA	High	Retrospective	2011 to 2018	3729	Women recruited from hospital, urban setting, hospital deliveries, 11.5% smoking, 2.9% alcohol use, 13.4% IDU	Risk factor analysis	No description	Average
Schulte (2007) [70]	USA	High	Retrospective	1989 to 2004	11231	27.6% history of IDU	Regression analysis	LNMP, ultrasound (unspecified), neonatal assessment (unspecified)	Poor
Sebitloane (2017) [71]	South Africa	Middle	Retrospective	1/4/2011 to 30/4/2014	1461	Twins excluded, women recruited at a regional hospital, urban setting, hospital deliveries	None	No description	Poor
Short (2014) [72]	United Kingdom	High	Retrospective	1996 to 2010	331	Twins included, women recruited from a HIV antenatal clinic, urban setting, deliveries in a tertiary hospital, 13% smoking	None	No description	Poor
Silverman (2010) [73]	Zambia	Low	Retrospective	Unspecified	1238	Twins included	Risk factor analysis	No description	Poor
Simonds (1998) [74]	USA	High	Retrospective	1985 to 12/1995	1366	Twins excluded, 18.4% IDU	None	Ballard score	Poor
Snijdewind (2018) [75]	Netherlands	High	Retrospective	1/1997 to 2/2015	10795	Twins excluded, women recruited from 26 nationwide sites, 10.8% smoking, 11.7% alcohol use, 0.6% IDU	Risk factor analysis	Early ultrasound or LNMP	Average
Tiam (2019) [76]	Lesotho	Middle	Prospective	6/2014 to 2/2016	1594	Women recruited from 14 mixed setting study centres across 3 districts, 91.6% delivered in a health facility	None	LNMP	Poor
Townsend ECS (2010) [18]	Belgium, Denmark, Germany, Italy, Netherlands, Poland, Spain, Sweden, United Kingdom	High	Prospective	1990 to 2006	4253	Twins excluded, 35.4% IDU	Regression analysis	LNMP and/or ultrasound (unspecified)	Poor
Townsend NSHPC (2010) [18]	United Kingdom, Ireland	High	Prospective	1990 to 2006	6426	Women recruited from 205 hospitals across UK and Ireland, 4.4% IDU	Regression analysis	No description	Poor
Tuomala (2002) [19]	USA	High	Retrospective	1/1/1990 to 1998	3266	Twins excluded, women recruited from PACTS and WITS studies, and three single site studies, 39.9% tobacco use during pregnancy, 26.9% alcohol use during pregnancy, 28.7% IDU use during pregnancy	Regression analysis, risk factor analysis	LNMP and/or ultrasound (unspecified), or neonatal assessment (unspecified)	Average
Van der Merwe (2011) [77]	South Africa	Middle	Retrospective	10/2004 to 3/2007	1630	Twins excluded, women recruited from HIV referral centres including a tertiary hospital, 3.7% smoking, 3.9% alcohol use	Regression analysis, risk factor analysis	LNMP, ultrasound (unspecified), symphysis-fundal height, neonatal assessment (unspecified)	Poor
Von Linstow (2010) [78]	Denmark	High	Retrospective	1/6/1994 to 30/6/2008	255	Twins included, women recruited from six centres nationwide, all hospital deliveries, 15.4% smoking, 2.2% IDU	None	Late ultrasound at 18–20 weeks	Poor
Watts (2013) [79]	USA and Puerto Rico	High	Retrospective	2007 to 31/10/2010	1869	Twins excluded, 17% smoking, 8% alcohol use, 8% IDU	Regression analysis	Clinical method (unspecified) and ultrasound (unspecified)	Average
Yu (2012) [80]	China	Middle	Retrospective	6/2006 to 7/2010	194	Twins excluded, 8.8% IDU	Risk factor analysis	No description	Poor
Zash (2018) [81]	Botswana	Middle	Retrospective	15/8/2014 to 15/8/2016	57005	Twins excluded, women recruited from eight government hospitals, hospital deliveries, 8.3% alcohol or smoking in pregnancy	Regression analysis	LNMP and/or ultrasound (unspecified), or symphysis-fundal height	Average
Ziske (2013) [82]	Tanzania	Low	Prospective	9/2008 to 9/2009	144	Twins excluded, women recruited from antenatal care (HIV+ receiving ART) or maternity ward (HIV+ no ART), rural setting, hospital deliveries	Risk factor analysis	No description	Poor

^aDetails on the inclusion of twins, recruitment centre, urban/rural setting, deliveries at home/hospital, smoking, alcohol use, and IDU were sought and reported here if provided by each study. ART, antiretroviral therapy; IDU, illicit drug use; LNMP, last normal menstrual period.

Table 2. Antiretroviral therapies, HIV/antiretroviral therapy comparisons, and perinatal outcomes reported by studies included in the systematic review and meta-analysis.
Study: first author (year) [ref]	ART regimens	WLHIV with monotherapy vs. ART-naive WLHIV	WLHIV with monotherapy vs. HIV-negative women	WLHIV with HAART vs. ART-naive WLHIV	WLHIV with HAART vs. HIV-negative women	Perinatal outcomes
Ai-Jie (2013) [26]	77.4% ZDV monotherapy, 22.6% NNRTI-based HAART (ZDV-3TC-NVP)	Yes	No	Yes	No	LBW
Albert (2020) [27]	4.5% mono/dual/triple NRTI therapy, 17.7% NNRTI-based HAART, 73.7% PI-based HAART, 4.1% INSTI-based HAART	No	No	Yes	No	sPTB
Azria (2009) [28]	PI-based HAART (LPV/r)	No	No	No	Yes	PTB, VPTB, SGA, VSGA, NND
Bailey (2013) [29]	91.3% ZDV monotherapy, 1.2% dual therapy, 7.5% HAART (91% PI-based HAART)	Yes	No	Yes	No	PTB
Balogun (2018) [30]	PI-based HAART (50.7% LPV/r, 31.8% ATV/r, 4.8% DRV/r)	No	No	No	Yes	sPTB, SGA
Bengtson (2020) [31]	NNRTI-based HAART (TDF-FTC/3TC-EFV)	No	No	No	Yes	PTB, SGA, VSGA
Boer (2006) [32]	PI-/NNRTI-based HAART (proportion unspecified)	No	No	No	Yes	PTB, LBW, VLBW
Boyajian (2012) [33]	75% PI-based HAART, 25% non-PI based HAART	No	No	No	Yes	PTB, LBW, SGA
Carceller (2009) [34]	85.4% PI-based HAART, 14.6% non-PI based HAART	No	No	No	Yes	PTB, Term LBW
Chagomerana (2017) [35]	NNRTI-based HAART (TDF-3TC-EFV)	No	No	Yes	No	PTB, VPTB
Chen (2012) [36]	58.4% ZDV monotherapy, 2.9% PI-based HAART (LPV/r-ZDV-3TC), 33.5% NNRTI-based HAART (NVP-ZDV-3TC), 5.2% unspecified HAART	No	Yes	No	Yes	PTB, SGA
Chibwesha (2016) [37]	66.6% ZDV monotherapy, 33.4% HAART [unspecified drug class(es)]	Yes	Yes	Yes	Yes	LBW
Cooper (2002) [38]	62% ZDV monotherapy, 16.2% dual therapy (96.8% two NRTIs, 2.2% NRTI-NNRTI, 0.5% two NNRTIs), 21.8% HAART (NNRTI-based, PI-based, or NNRTI-PI-based)	Yes	No	Yes	No	PTB, LBW
Cotter (2006) [39]	49.3% ZDV monotherapy, 37.3% non-PI-based HAART, 13.4% PI-based HAART	Yes	No	No	No	LBW, VLBW
Dadabhai (2019) [40]	NNRTI-based HAART (TDF-3TC-EFV)	No	No	No	Yes	PTB, LBW, Term LBW, Preterm LBW, SGA, VSGA
De Souza (2000) [41]	ZDV monotherapy	Yes	No	No	No	PTB
Garcia-Otero (2019) [42]	HAART (29.8% NNRTI-containing, 66% PI-containing, 14.9% INSTI-containing)	No	No	No	Yes	PTB, SGA, NND
Goetghebuer (2019) [43]	77.3% PI-based HAART, 12.9% NNRTI-based HAART, 5.3% NRTI-based HAART, 4.5% other regimen	No	No	No	Yes	PTB, LBW
Haeri (2009) [44]	HAART (94% NRTI-containing, 20% NNRTI-containing, 74% PI-containing)	No	No	No	Yes	PTB, sPTB, Term LBW, SGA
Hernandez (2017) [45]	4.2% ZDV monotherapy, 33.3% NNRTI-based HAART, 58.3% PI-based HAART, 4.2% NRTI-based HAART	No	No	No	Yes	SGA
Hu (2019) [46]	20.1% ZDV monotherapy/ZDV-3TC dual therapy, 79.9% HAART (NNRTI-based/PI- based)	No	No	Yes	No	PTB, SGA
Joseph (2011) [47]	NNRTI-based HAART (NVP)	No	No	Yes	No	LBW
Jumare (2019) [48]	HAART [drug class(es) unspecified]	No	No	No	Yes	LBW
Kakkar (2015) [49]	16.8% ZDV monotherapy, 14.5% NRTI-/NNRTI- containing dual therapy/HAART, 68.7% PI-based HAART	Yes	No	Yes	No	PTB
Kowalska (2003) [50]	43.2% ZDV monotherapy, 22.2% PI-based HAART, 34.6% non-PI-based HAART	Yes	No	Yes	No	PTB, LBW
Li (2016) [13]	61.8% ZDV monotherapy, 35.5% NNRTI-based HAART, 0.6% PI-based HAART, 2.1% unspecified HAART	Yes	No	Yes	No	PTB, LBW, SGA, VSGA
Li (2020) [51]	24.2% mono/dual therapy, 75.8% HAART (drug class(es) unspecified)	No	No	Yes	Yes	PTB, LBW, SGA
Liff (2020) [52]	78% NNRTI-based HAART, 12% INSTI-based HAART, 10% other HAART	No	No	No	Yes	PTB
Lopez (2012) [53]	HAART (98.7% NRTI-containing, 51.3% NNRTI-containing, 59.7% PI-containing)	No	No	No	Yes	PTB, sPTB
Malaba (2017) [54]	71.6% NNRTI-based HAART, 2.3% PI-based HAART, 26.1% other HAART	No	No	No	Yes	PTB, VPTB, LBW, VLBW, SGA
Malaba (2018) [55]	92.5% NNRTI-based HAART, 2.8% PI-based HAART, 4.7% other HAART	No	No	No	Yes	PTB, SGA
Mandelbrot (1998) [56]	ZDV monotherapy	Yes	No	No	No	PTB
Marazzi (2011) [57]	NRTI-/NNRTI-based HAART (proportion unspecified)	No	No	Yes	No	PTB
Marti (2007) [58]	15.1% ZDV monotherapy, 13.8% NRTI dual therapy, 7.9% NNRTI-based HAART, 61.8% PI-based HAART, 1.4% NRTI-based HAART	Yes	No	Yes	No	PTB, LBW
Matheson (1995) [59]	ZDV monotherapy	Yes	No	No	No	PTB
Mehta (2019) [60]	98% NNRTI-based HAART, 0.9% PI-based HAART, 1.1% unspecified HAART	No	No	No	Yes	PTB, LBW, SGA, NND
Moodley (2016) [61]	27.5% ZDV monotherapy, 72.5% NNRTI-based HAART	Yes	Yes	Yes	Yes	PTB, LBW, SGA
Moseholm (2019) [62]	13.6% NNRTI-based HAART, 78.4% PI-based HAART, 5.7% NRTI-based HAART, 2.3% unspecified HAART	No	No	No	Yes	PTB
Olagbuji (2010) [63]	NNRTI-based HAART (ZDV/3TC/NVP)	No	No	No	Yes	LBW
Phiri (2015) [64]	20% ZDV monotherapy, 15.3% NRTI-NNRTI dual therapy, 21.3% NRTI dual therapy/HAART, 43.4% PI-based therapy (unspecified regimen complexity)	Yes	No	No	No	PTB, SGA
Ramokolo (2017) [65]	38.5% ZDV monotherapy, 61.5% NNRTI-based HAART (TDF-3TC/FTC-NVP)	Yes	Yes	Yes	Yes	PTB, LBW, SGA
Rempis (2017) [66]	NNRTI-based HAART (TDF-3TC-EFV)	No	No	No	Yes	SGA
Rudin (2011) [67]	26.4% ZDV mono/dual therapy, 61.8% PI-based HAART, 11.8% non-PI-based HAART	No	No	Yes	No	PTB, VPTB
Santosa (2019) [68]	1.6% ZDV monotherapy, 96% HAART, 2.4% unspecified regimen	No	No	No	Yes	PTB, VPTB, LBW, VLBW, SGA, VSGA, Stillbirth, NND
Saums (2019) [69]	10.9% NNRTI-based HAART, 54.7% PI-based HAART, 34.3% INSTI-based HAART	No	No	No	Yes	PTB
Schulte (2007) [70]	42.1% ZDV monotherapy, 16.7% dual therapy, 12.6% PI-based HAART, 28.6% non-PI-based HAART	Yes	No	Yes	No	PTB, LBW
Sebitloane (2017) [71]	36.6% ZDV monotherapy, 63.4% NNRTI-based HAART	No	Yes	No	Yes	PTB
Short (2014) [72]	20.1% ZDV monotherapy, 2.2% NRTI dual therapy, 42.4% NNRTI-based HAART, 29.8% PI-based HAART, 1.5% NRTI-based HAART, 4% unspecified HAART	Yes	No	Yes	No	PTB
Silverman (2010) [73]	PI-based HAART (ZDV-3TC-LPV/r)	No	No	Yes	No	LBW
Simonds (1998) [74]	ZDV monotherapy	Yes	No	No	No	PTB, LBW
Snijdewind (2018) [75]	31.5% NNRTI-based HAART, 66.7% PI-based HAART, 1.8% other HAART	No	No	No	Yes	PTB, VPTB, LBW, VLBW, SGA
Tiam (2019) [76]	96.5% NNRTI-based HAART, 2.3% other HAART, 2.2% no ART	No	No	No	Yes	PTB, LBW, VLBW
Townsend ECS (2010) [18]	27.8% ZDV monotherapy, 11.8% NRTI dual therapy, 36.2% PI-based HAART, 24.2% non-PI-based HAART	Yes	No	Yes	No	PTB
Townsend NSHPC (2010) [18]	16.3% ZDV monotherapy, 3.2% dual therapy, 42% PI-based HAART, 38.5% non-PI-based HAART	Yes	No	Yes	No	PTB
Tuomala (2002) [19]	74.8% ZDV monotherapy, 6.5% PI-based dual/HAART, 18.7% non-PI-based dual/HAART	Yes	No	No	No	PTB, VPTB, LBW, VLBW
Van der Merwe (2011) [77]	42.8% NNRTI-based HAART, 44.5% PI-based HAART, 12.7% unspecified HAART	No	No	Yes	No	PTB, LBW, VLBW
Von Linstow (2010) [78]	12.1% ZDV monotherapy/dual therapy, 87.9% NNRTI-/PI-based HAART	No	No	Yes	No	LBW
Watts (2013) [79]	7.6% mono/dual therapy, 8.8% NNRTI-based HAART, 72.9% PI-based HAART, 10.7% NRTI-based HAART	No	No	Yes	No	PTB, sPTB
Yu (2012) [80]	NNRTI-based HAART	No	No	Yes	No	PTB, LBW
Zash (2018) [81]	72.7% NNRTI-based HAART (TDF-FTC-EFV), 27.3% INSTI-based HAART (TDF-FTC-DTG)	No	No	No	Yes	PTB, VPTB, SGA, VSGA, NND
Ziske (2013) [82]	ZDV monotherapy	Yes	No	No	No	PTB

3TC, lamivudine; ART, antiretroviral therapy; ATV/r, ritonavir-boosted atazanavir; cART, combination antiretroviral therapy (≥2 antiretroviral drugs); DRV/r, ritonavir-boosted darunavir; EFV, efavirenz; FTC, emtricitabine; HAART, highly active antiretroviral therapy (at least three antiretroviral drugs); INSTI, integrase strand transfer inhibitor; LBW, low birthweight; LPV/r, ritonavir-boosted lopinavir; NND, neonatal death; NNRTI, nonnucleoside transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; NVP, nevirapine; PI, protease inhibitor; PTB, preterm birth; SGA, small for gestational age; sPTB, spontaneous preterm birth; TDF, tenofovir disoproxil fumarate; VLBW, very low birthweight; VPTB, very preterm birth; VSGA, very small for gestational age; WLHIV, women living with HIV; ZDV, zidovudine.

Table 3. Perinatal outcomes associated with women living with HIV receiving HAART or monotherapy compared to antiretroviral therapy-naive women living with HIV and HIV-negative women.
(a) All studies
		Perinatal outcomes
		PTB	VPTB	sPTB	LBW	VLBW	Term LBW	Preterm LBW	SGA	VSGA	Stillbirth	NND
		RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)	RR (95%CI)
WLHIV with monotherapy vs ART-naïve WLHIV		0.70 (0.62, 0.79)	0.96 (0.63, 1.47)		0.77 (0.67, 0.88)	0.63 (0.24, 1.64)			1.08 (0.90, 1.29)	1.43 (0.97, 2.10)
WLHIV with monotherapy vs HIV-negative women		1.04 (0.84, 1.29)			1.08 (0.88, 1.33)				1.16 (1.04, 1.30)
WLHIV with HAART vs ART-naïve WLHIV		0.85 (0.66, 1.10)	0.89 (0.23, 3.51)	0.46 (0.32, 0.67)	0.92 (0.79, 1.08)	0.36 (0.16, 0.78)			1.38 (1.09, 1.75)	2.35 (1.60, 3.46)
WLHIV with HAART vs HIV-negative women		1.55 (1.38, 1.74)	1.72 (0.75, 3.96)	2.10 (1.48, 2.96)	1.79 (1.51, 2.13)	1.70 (0.63, 4.57)	1.88 (1.23, 2.85)	1.17 (0.64, 2.14)	1.80 (1.34, 2.40)	1.22 (1.10, 1.34)	0.88 (0.34, 2.32)	1.27 (0.75, 2.15)
(b) Subgroup analysis by country income status
WLHIV with monotherapy vs ART-naïve WLHIV	High income countries	0.77 (0.70, 0.84)	0.96 (0.63, 1.47)		0.85 (0.78, 0.93)	0.63 (0.24, 1.64)			0.91 (0.55, 1.51)
WLHIV with monotherapy vs ART-naïve WLHIV	Low- and middle-income countries	0.65 (0.45, 0.92)			0.73 (0.59, 0.91)				1.08 (0.86, 1.36)	1.43 (0.97, 2.10)
WLHIV with monotherapy vs HIV-negative women	High income countries
WLHIV with monotherapy vs HIV-negative women	Low- and middle-income countries	1.04 (0.84, 1.29)			1.08 (0.88, 1.33)				1.16 (1.04, 1.30)
WLHIV with HAART vs ART-naïve WLHIV	High income countries	0.99 (0.79, 1.24)	1.86 (0.78, 4.46)	0.46 (0.32, 0.67)	0.79 (0.72, 0.87)	0.36 (0.16, 0.78)
WLHIV with HAART vs ART-naïve WLHIV	Low- and middle-income countries	0.73 (0.48, 1.09)	0.47 (0.29, 0.77)		1.01 (0.78, 1.30)				1.38 (1.09, 1.75)	2.35 (1.60, 3.46)
WLHIV with HAART vs HIV-negative women	High income countries	1.92 (1.51, 2.43)	2.12 (0.38, 12.04)	2.09 (1.48, 2.96)	2.49 (1.56, 3.98)	1.84 (0.14, 23.58)	1.91 (1.10, 3.29)		3.78 (1.29, 11.08)	1.20 (0.29, 4.92)		0.40 (0.02, 8.21)
WLHIV with HAART vs HIV-negative women	Low- and middle-income countries	1.33 (1.21, 1.45)	1.11 (0.82, 1.49)		1.56 (1.35, 1.80)	1.38 (0.77, 2.45)	1.83 (0.95, 3.53)	1.17 (0.64, 2.14)	1.41 (1.16, 1.72)	1.22 (1.10, 1.34)	0.88 (0.34, 2.32)	1.34 (0.76, 2.35)

Relative risk (RR) and 95% confidence interval (95% CI) of random-effects meta-analyses. For forest plots see Appendix, pp. 30–45, https://links.lww.com/QAD/C513. RR greater than 1 indicates that women living with HIV receiving ART are associated with an increased risk of the corresponding perinatal outcome. ART, antiretroviral therapy; LBW, low birthweight; NND, neonatal death; PTB, preterm birth; SGA, small for gestational age; sPTB, spontaneous preterm birth; VLBW, very low birthweight; VPTB, very preterm birth; VSGA, very small for gestational age; WLHIV, women living with HIV.