Advertisement

Miscarriage among women in the United States Women’s Interagency HIV Study, 1994–2017

      Background

      Relatively little is known about the frequency and factors associated with miscarriage among women living with HIV.

      Objective

      The objective of the study was to evaluate factors associated with miscarriage among women enrolled in the Women’s Interagency HIV Study.

      Study Design

      We conducted an analysis of longitudinal data collected from Oct. 1, 1994, to Sept. 30, 2017. Women who attended at least 2 Women’s Interagency HIV Study visits and reported pregnancy during follow-up were included. Miscarriage was defined as spontaneous loss of pregnancy before 20 weeks of gestation based on self-report assessed at biannual visits. We modeled the association between demographic, behavioral, and clinical covariates and miscarriage (vs live birth) for women overall and stratified by HIV status using mixed-model logistic regression.

      Results

      Similar proportions of women living with and without HIV experienced miscarriage (37% and 39%, respectively, P = .638). In adjusted analyses, smoking tobacco (adjusted odds ratio, 2.0), alcohol use (adjusted odds ratio, 4.0), and marijuana use (adjusted odds ratio, 2.0) were associated with miscarriage. Among women living with HIV, low HIV viral load (<4 log10 copies/mL) (adjusted odds ratio, 0.5) and protease inhibitor (adjusted odds ratio, 0.4) vs the nonuse of combination antiretroviral therapy use were protective against miscarriage.

      Conclusion

      We did not find an increased odds of miscarriage among women living with HIV compared with uninfected women; however, poorly controlled HIV infection was associated with increased miscarriage risk. Higher miscarriage risk among women exposed to tobacco, alcohol, and marijuana highlight potentially modifiable behaviors. Given previous concern about antiretroviral therapy and adverse pregnancy outcomes, the novel protective association between protease inhibitors compared with non–combination antiretroviral therapy and miscarriage in this study is reassuring.

      Key words

      In the current combined antiretroviral treatment (cART) era, pregnancy rates have increased among US women with HIV and are comparable with women without HIV.
      • Haddad L.B.
      • Wall K.M.
      • Mehta C.C.
      • et al.
      Trends of and factors associated with live-birth and abortion rates among HIV-positive and HIV-negative women.
      Miscarriage occurs in an estimated 10–20% of US women.
      • Wilcox A.J.
      • Weinberg C.R.
      • O'Connor J.F.
      • Baird D.D.
      • Schlatterer J.P.
      • Canfield R.E.
      • Armstrong E.G.
      • Nisula B.C.
      Incidence of early loss of pregnancy.
      • Zinaman M.J.
      • Clegg E.D.
      • Brown C.C.
      • O'Connor J.
      • Selevan S.G.
      Estimates of human fertility and pregnancy loss.
      • Wyatt P.R.
      • Owolabi T.
      • Meier C.
      • Huang T.
      Age-specific risk of fetal loss observed in a second trimester serum screening population.
      However, relatively little is known about the frequency and factors associated with miscarriage among women living with HIV (WLHIV).

      Why was this study conducted?

      To provide a more comprehensive and contemporary evaluation of miscarriage, we assessed the frequency and factors associated with miscarriage in both women living with and without HIV enrolled in the Women’s Interagency HIV Study.

      Key findings

      HIV status was not associated with miscarriage risk. Poorly controlled HIV infection and tobacco, alcohol, and marijuana use were associated with an increased miscarriage risk, while protease inhibitors were protective compared with a noncombination antiretroviral treatment.

      What does this add to what is known?

      Our study adds several years of analytic follow-up to previous analyses, confirms previous findings related to HIV status and miscarriage, and identifies new findings related to alcohol use, marijuana use frequency, and the protective effect of protease inhibitors.
      Among the few cART-era analyses, data are conflicting as to whether women with HIV experience an increased risk miscarriage; 3 studies from Africa reported an increased risk,
      • Schwartz S.R.
      • Rees H.
      • Mehta S.
      • Venter W.D.F.
      • Taha T.E.
      • Black V.
      High incidence of unplanned pregnancy after antiretroviral therapy initiation: findings from a prospective cohort study in South Africa.
      • Chen J.Y.
      • Ribaudo H.J.
      • Souda S.
      • et al.
      Highly active antiretroviral therapy and adverse birth outcomes among HIV-infected women in Botswana.
      • Ezechi O.C.
      • Gab-Okafor C.V.
      • Oladele D.A.
      • et al.
      Pregnancy, obstetric and neonatal outcomes in HIV positive Nigerian women.
      while one from the United States did not.
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      Most studies have focused on outcomes of low birthweight, preterm birth, neonatal intensive care unit admission, and neonatal morbidity,
      • van Gelder M.M.
      • Reefhuis J.
      • Caton A.R.
      • Werler M.M.
      • Druschel C.M.
      • Roeleveld N.
      Characteristics of pregnant illicit drug users and associations between cannabis use and perinatal outcome in a population-based study.
      • Fergusson D.M.
      • Horwood L.J.
      • Northstone K.
      Maternal use of cannabis and pregnancy outcome.
      • Wu C.-S.
      • Jew C.P.
      • Lu H.-C.
      Lasting impacts of prenatal cannabis exposure and the role of endogenous cannabinoids in the developing brain.
      • Metz T.D.
      • Allshouse A.A.
      • Hogue C.J.
      • et al.
      Maternal marijuana use, adverse pregnancy outcomes, and neonatal morbidity.
      • Gunn J.K.
      • Rosales C.B.
      • Center K.E.
      • et al.
      Prenatal exposure to cannabis and maternal and child health outcomes: a systematic review and meta-analysis.
      • Hayatbakhsh M.R.
      • Flenady V.J.
      • Gibbons K.S.
      • et al.
      Birth outcomes associated with cannabis use before and during pregnancy.
      • Conner S.N.
      • Bedell V.
      • Lipsey K.
      • Macones G.A.
      • Cahill A.G.
      • Tuuli M.G.
      Maternal marijuana use and adverse neonatal outcomes: a systematic review and meta-analysis.
      • Machado E.S.
      • Hofer C.B.
      • Costa T.T.
      • et al.
      Pregnancy outcome in women infected with HIV-1 receiving combination antiretroviral therapy before versus after conception.
      • Townsend C.L.
      • Cortina-Borja M.
      • Peckham C.S.
      • Tookey P.A.
      Antiretroviral therapy and premature delivery in diagnosed HIV-infected women in the United Kingdom and Ireland.
      • Powis K.M.
      • Kitch D.
      • Ogwu A.
      • et al.
      Increased risk of preterm delivery among HIV-infected women randomized to protease versus nucleoside reverse transcriptase inhibitor-based HAART during pregnancy.
      • Polgár B.T.
      • Nagy E.
      • Mikó E.V.
      • Varga P.T.
      • Szekeres-Barthó J.L.
      Urinary progesterone-induced blocking factor concentration is related to pregnancy outcome.
      • Salas S.P.
      • Marshall G.
      • Gutierrez B.L.
      • Rosso P.
      Time course of maternal plasma volume and hormonal changes in women with preeclampsia or fetal growth restriction.
      • Coleman-Cowger V.H.
      • Oga E.A.
      • Peters E.N.
      • Mark K.
      Prevalence and associated birth outcomes of co-use of Cannabis and tobacco cigarettes during pregnancy.
      • Chabarria K.C.
      • Racusin D.A.
      • Antony K.M.
      • et al.
      Marijuana use and its effects in pregnancy.
      with fewer and more contradictory studies, mostly from sub-Saharan Africa, focused on miscarriage.
      • Schwartz S.R.
      • Rees H.
      • Mehta S.
      • Venter W.D.F.
      • Taha T.E.
      • Black V.
      High incidence of unplanned pregnancy after antiretroviral therapy initiation: findings from a prospective cohort study in South Africa.
      • Chen J.Y.
      • Ribaudo H.J.
      • Souda S.
      • et al.
      Highly active antiretroviral therapy and adverse birth outcomes among HIV-infected women in Botswana.
      • Ezechi O.C.
      • Gab-Okafor C.V.
      • Oladele D.A.
      • et al.
      Pregnancy, obstetric and neonatal outcomes in HIV positive Nigerian women.
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      • Kim H.Y.
      • Kasonde P.
      • Mwiya M.
      • et al.
      Pregnancy loss and role of infant HIV status on perinatal mortality among HIV-infected women.
      • Turner A.N.
      • Tabbah S.
      • Mwapasa V.
      • et al.
      Severity of maternal HIV-1 disease is associated with adverse birth outcomes in Malawian women: a cohort study.
      • Cates J.E.
      • Westreich D.
      • Edmonds A.
      • et al.
      The effects of viral load burden on pregnancy loss among HIV-infected women in the United States.
      • Westreich D.
      • Cates J.
      • Cohen M.
      • et al.
      Smoking, HIV, and risk of pregnancy loss.
      Previous analyses from the Women’s Interagency HIV Study (WIHS), the largest ongoing multicenter prospective cohort study of HIV among women in the United States, have explored factors associated with miscarriage. In the 1994–2002 WIHS cohort, Massad et al
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      found that miscarriage rates were similar for WLHIV and women without HIV and that in WLHIV, miscarriage was associated with prior miscarriage, marijuana use, and non–antiretroviral treatment (ART) use. That analysis did not consider the role of potential correlates including alcohol, frequency of marijuana use, sexually transmitted infections (STIs), specific cART regimens, or women without HIV.
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      In the 1994–2012 WIHS cohort, Cates et al
      • Cates J.E.
      • Westreich D.
      • Edmonds A.
      • et al.
      The effects of viral load burden on pregnancy loss among HIV-infected women in the United States.
      found that HIV viral load during pregnancy predicted pregnancy loss, while Westreich et al
      • Westreich D.
      • Cates J.
      • Cohen M.
      • et al.
      Smoking, HIV, and risk of pregnancy loss.
      found cigarette smoking was associated with pregnancy loss in the 1994–2014 WIHS cohort.
      Neither of these studies reported on factors associated with miscarriage outside viral load and smoking, respectively. To provide a more comprehensive and contemporary evaluation, we assessed the frequency and factors associated with miscarriage in both women living with and without HIV enrolled in the WIHS.

      Materials and Methods

      Cohort

      We analyzed data collected between Oct. 1, 1994, and Sept. 30, 2017, from women enrolled in the WIHS.
      • Barkan S.E.
      • Melnick S.L.
      • Preston-Martin S.
      • et al.
      The Women's Interagency HIV Study. WIHS Collaborative Study Group.
      This cohort is comprised of women living with and without HIV recruited from 10 US sites. In 1994, enrollment began at 6 sites (Bronx, NY; Brooklyn, NY; Chicago, IL; Washington DC; Los Angeles, CA; and San Francisco, CA), and those sites reenrolled during 2001–2002 and 2011–2012.
      The cohort was expanded in 2013–2015 to include 4 additional sites to represent the southern United States (Atlanta, GA; Chapel Hill, NC; Miami, FL; and Birmingham, AL/Jackson, MS). Enrollees were frequency matched on demographics and risk factors for acquiring HIV including drug use and number of sexual partners.
      • Barkan S.E.
      • Melnick S.L.
      • Preston-Martin S.
      • et al.
      The Women's Interagency HIV Study. WIHS Collaborative Study Group.
      • Bacon M.C.
      • von Wyl V.
      • Alden C.
      • et al.
      The Women's Interagency HIV Study: an observational cohort brings clinical sciences to the bench.
      Medical and laboratory examinations (including CD4 T-cell count and HIV viral load) as well as detailed interviews (collecting data including cART use, obstetric and gynecological history, and medication use) are conducted longitudinally at semiannual visits by trained interviewers.

      Ethics

      Each WIHS site and the data center provided institutional review board approval prior to study enrollment. Enrollees provided written informed consent. All data used in this analysis were deidentified.

      Outcome definition

      Women who attended at least 2 WIHS visits and self-reported a pregnancy during longitudinal follow-up were included in this analysis. Because prenatal care, delivery, and other obstetrical services were received outside the study, pregnancy outcomes and dates were self-reported. Interviewers were trained to assess pregnancy and pregnancy outcome information from women by asking at every visit: “since your last study visit, how many times have you been pregnant?” and, for those responding with a non-zero number, “what was the outcome of the pregnancy?”
      Pregnancy outcomes were explained to help reduce misclassification error. Miscarriage was defined as spontaneous loss of a pregnancy before 20 weeks of gestation. Other pregnancy outcomes included stillbirth, defined as a child born dead after 20 weeks of gestation; live birth, defined as a child born alive; ectopic pregnancy, defined as any pregnancy outside the uterus; and abortion, defined as any pregnancy terminated through an elective procedure.
      In total, 755 women reported 1487 pregnancies with known outcomes. We excluded pregnancies that ended in stillbirths (n = 11, 0.7% of total pregnancies ending in stillbirth; n = 5 women with only stillbirths) or ectopic pregnancies (n = 47, 3% of total pregnancies ending in ectopic pregnancy; n = 16 women with only ectopic pregnancies).
      We also excluded pregnancies without information in the visit prior to an outcome (n = 64, 4% of total pregnancies; n = 29 women) and those pregnancies that ended in abortion (n = 454, 31% of total pregnancies ending in abortion; n = 121 women with only abortion outcomes). We excluded women in the 2011–2012 wave (n = 11) and southern sites (n = 25) because of small numbers (comprising 44 pregnancies). In primary analyses, we compared women reporting miscarriage with the referent group of women reporting live births. In sensitivity analyses, the referent group also included abortions.

      Covariates of interest

      Baseline covariates of interest measured at enrollment included the following: race/ethnicity, income, marital status, education, employment status, insurance status, WIHS site and enrollment wave, number of lifetime male sexual partners (categorized based on the distribution of the continuous variable and meaningful groupings), parity, prior miscarriage, and year of cART initiation for women with HIV.
      Because of the small number of self-reported prior STI events (<6 cases were recorded for each STI and none were individually associated with the outcome of interest), a composite variable for any prior STIs included gonorrhea, trichomonas, genital herpes, genital warts, syphilis, and chlamydia.
      Time-varying covariates occur in the visit prior to the outcome of interest (and thus occur either during pregnancy or just prior to conception). Time-varying covariates of interest included the following: age group; alcoholic drinks per week (dichotomized as abstainers vs drinkers); marijuana use (dichotomized as yes/no) and frequency coded as none, less than weekly, 1–6 times a week, and at least daily); current cigarette smoking; any health-related issues (including stroke, cancer, myocardial infarction, high blood pressure pulmonary embolism, diabetes); year of pregnancy; any new STI diagnosis (including gonorrhea, trichomonas, genital herpes, genital warts, syphilis, and chlamydia); yeast infection; bacterial vaginosis infection; and having a loop electrosurgical excision procedure.
      Time-varying HIV-specific covariates of interest included the following: HIV viral load (dichotomized as ≥4 log10 copies/mL of plasma vs <4 log10 copies/mL of plasma for comparability with other WIHS analyses); CD4+ T-cell count (dichotomized as <350 cells/μL vs ≥350 cells/μL); ART regimen (categorized as follows: (1) cART-containing protease inhibitor (PI)–based regimen with nucleoside reverse transcriptase inhibitors (NRTI), (PI-based cART); (2) cART-containing integrase inhibitors and/or non–NNRTI-based regimens (non-PI cART) with NRTIs; and (3) no cART including NRTI therapy alone or no ART); and ever having an AIDS-defining illness.

      Analysis methods

      We describe the frequencies of baseline categorical covariates using counts and percentages stratified by the primary outcome of interest (miscarriage vs live birth), with differences by miscarriage status quantified by χ2 or Fisher exact tests, as appropriate. Frequencies for time-varying variables were calculated over all longitudinal visits. Descriptive analyses were calculated overall and stratified by HIV status.
      To account for multiple pregnancy outcomes per woman, we modeled associations between covariates and miscarriage (vs live birth) for women overall and stratified by HIV status using mixed logistic regression models with a random intercept and an unstructured covariance matrix. Model-based point estimates (adjusted odds ratios) and 95% confidence intervals are reported.
      Covariates significant in bivariate analyses at P < .05 within a given strata (overall, living with or without HIV) were included in the corresponding multivariate models after assessing for multicollinearity. We decided a priori to include HIV status as a covariate in the model for women overall. Possible 2-way interactions between marijuana use and cigarette smoking, marijuana use and drinking, and drinking and cigarette smoking were explored. In sensitivity analyses, models included both live births and abortion in the referent group. Analyses were performed with SAS version 9.4 (SAS Institute, Cary, NC).

      Results

      Miscarriage outcomes

      Among the 548 women included in the primary analysis, 207 (38%) experienced a miscarriage during follow-up and 341 (68%) experienced only live birth. Among 226 women without HIV, 88 (39%) ever experienced a miscarriage and 138 (61%) experienced only live birth. Among 322 WLHIV, 119 (37%) ever experienced a miscarriage and 203 (63%) experienced only live birth. There was no significant difference in the proportion of women ever having a miscarriage between WLHIV vs women without HIV (37% vs 39%, respectively, P = .638).

      Baseline descriptive findings (Table 1)

      The majority of women were African American (58%), earned ≤$12,000 USD per year (55%), had less than a high school education (60%), were unemployed (70%), and had between 1 and 2 children (45%) upon entry into the WIHS. Among all women, those experiencing a miscarriage were more likely (P < .05) to be insured, be recruited from the New York WIHS sites, and have a higher number of lifetime male sex partners. Among WLHIV, those experiencing a miscarriage were more likely (P < .05) to be insured, have a higher number of male sex partners, and have had a prior miscarriage. No differences were observed between baseline covariates and miscarriage in women without HIV.
      Table 1Baseline characteristics stratified by HIV status and miscarriage, WIHS 1994–2017
      VariablesAll women (N = 548)Women living with HIV (N = 322)Women without HIV (N = 226)
      Mis-carriage (n = 207, 38%)Live birth (n = 341, 66%)P valueMis-carriage (n = 119, 37%)Live birth (n = 203, 63%)P valueMis-carriage (n = 88, 39%)Live birth (n = 138, 61%)P value
      n%n%n%n%n%n%
      Race/ethnicity
       White/other22114814.193982914.14013151914.494
       African American130631885579661175851587151
       Hispanic5527105313126572824274835
      Income per year
       ≤$12,0001085518456.951625410956.72646557556.539
       $12,000–24,000472475232724512620242418
       >$24,000432270212522361818213426
      Marital status
       Legally/common-law married28146720.08217144221.11711132518.525
       Unmarried but living with partner361770212118472315172317
       Single/widowed142692046080681145662709065
      Education1256120059.668726111557.49953618562.866
       High school or higher8139140414739884334395238
      Employed
       No1477123770.708907615074.73057658763.792
       Yes6029104302924532631355137
      Insurance
       No482310832.03518155125.04130345742.241
       Yes157772316899851527558667958
      Site group
       New York1055113038.03862527637.07143495439.281
       District of Columbia271353161513381912141511
       California5326111332723613026305036
       Chicago2211471415132814781914
      WIHS enrollment wave
       Original cohort (1994–1995)1065114944.08769589848.70137425137.444
       2001–2002 recruitment101491925650421055251588763
      Number of lifetime male partners
       0–4432110331.01324217337.00119223022.978
       5–249446155465144864343496950
       ≥25673378234135392026303928
      Parity
       062308926.31327234824.53035404130.268
       1–28441161475143974833386446
       ≥3612991274134582920233324
      Prior miscarriage
       No1306323770.1067160149730.01159678864.614
       Yes7737104304840542729335036
      Prior STI
      Includes gonorrhea, trichomonas, genital herpes, genital warts, syphilis, and chlamydia.
       No824013740.853393378390.27943495943.368
       Yes125602026080671236145517957
      Year of cART initiation
       1995–19996154143550.57661549851.092
       2000–20033229722832296132
       2004–20081513281115132111
       2009–4418744116
      P values are 2 tailed. cART, combination highly active antiretroviral therapy; STI, sexually transmitted infection; WIHS, Women's Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.
      a Includes gonorrhea, trichomonas, genital herpes, genital warts, syphilis, and chlamydia.

      Time-varying descriptive findings (Table 2)

      In the visit prior to pregnancy outcomes of interest, the majority of women were older than 30 years (57%), did not use alcohol (71%) or marijuana (83%), did not smoke (65%), and had no STI diagnosis (88%). Among all women, those experiencing a miscarriage were more likely (P < .05) to be alcohol drinkers, marijuana users, and cigarette smokers. WLHIV experiencing a miscarriage were also more likely (P < .05) to have experienced a pregnancy before 1998, have a higher viral load, and not be taking cART.
      Table 2Time-varying characteristics stratified by HIV status and miscarriage, WIHS 1994–2017
      Variables (previous visit)All women (N = 848)Women living with HIV (N = 468)Women without HIV (N = 380)
      Mis-carriage (n = 250, 29%)Live birth (n = 598, 71%)P valueMis-carriage (n = 140, 30%)Live birth (n = 328, 70%)P valueMis-carriage (n = 110, 29%)Live birth (n = 270, 71%)P value
      n%n%n%n%n%n%
      Age group
       <2563259716.01129213812.06234315922.138
       25–305020155262921902721196524
       30–3565261893241291053224228431
       >357229157264129952931286223
      Alcohol use
       Abstain1184945081< .0001765626384< .0001424018777< .0001
       Drinkers12351108196044511663605723
      Alcohol use frequency
       Abstain1184945081< .0001765626384< .0001424018777< .0001
       0–71014289164634451455524418
       >722919314106288135
      Marijuana use
       No1717049288< .00011057728891< .0001666220483< .0001
       Yes72306812322327940384117
      Marijuana use frequency
       None1717149288< .00011057728891.001666420484.001
       Less than weekly1562446411399135
       1–6 times per week22916312972101094
       At least daily32132751410931817187
      Cigarette smoking
       No1204940172< .0001654723173< .0001555217069.003
       Yes12351159287253842751487531
      Health related
      Includes stroke, cancer, myocardial infarction, high blood pressure, pulmonary embolism, and diabetes
       No1937747780.4141087725678.829857722182.334
       Yes5723121203223722225234918
      Pregnancy year
       1994–199750208614.06636264915.02314133714.655
       1998–2001461885142921571817162810
       2002–200574302153639281243835329134
       ≥200679322083536269529433911342
      Any STI diagnosis
      Includes gonorrhea, trichomonas, herpes-2, genital warts, syphilis, and chlamydia.
       No2118649388.5641148326584.688979222893.608
       Yes331467122417501698177
      Yeast infection
       No2008243478.1201118123775.177898419780.509
       Yes4318126232619782517164820
      BV infection
       No1868035685.134977319281.066899016490.873
       Yes462063153627441910101910
      LEEP procedure
       No24899596100.38913899327100.208110100269100n/a
       Yes212021100010
      HIV viral load
       ≥4 log10 copies/ mL45334314< .0001
       <4 log10 copies/ mL906726686
      CD4 count <500 cells/μL
       No574312741.759
       Yes775718359
      cART regimen
       II, NRTI, NNRTI23177122< .0001
       PI282012840
       No cART876312138
      AIDS-defining illness
       No987024775.235
       Yes42308125
      n is given over all study intervals. P values are 2 tailed. BV, bacterial vaginosis; cART, combination highly active antiretroviral therapy; II, integrase inhibitor; LEEP, loop electrosurgical excision procedure; NNRTI, non–nucleoside reverse transcriptase inhibitors; NRTI, nucleoside reverse transcriptase inhibitors; PI, protease inhibitor; STI, sexually transmitted infection; WIHS, Women's Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.
      a Includes stroke, cancer, myocardial infarction, high blood pressure, pulmonary embolism, and diabetes
      b Includes gonorrhea, trichomonas, herpes-2, genital warts, syphilis, and chlamydia.

      Adjusted model: all women (Table 3)

      Factors associated (P < .05) with miscarriage (vs live birth) included the following: being enrolled in the New York WIHS sites vs California sites, drinking alcohol, using marijuana in the previous visit, and cigarette smoking. Older age was protective for miscarriage. HIV status was not associated with miscarriage (P = .280). None of the 2-way interactions between marijuana use and cigarette smoking, marijuana use and drinking, and drinking and cigarette smoking were significant. In sensitivity analyses when elective abortions were included in the referent group, we found similarly adjusted findings (data not tabled).
      Table 3Baseline and time-varying factors associated with miscarriage in all women, WIHS 1994-2017
      FactorsLevelaOR95% CIP value
      HIV statusHIV positive (ref: HIV-)1.230.84–1.79.280
      Insurance (baseline)Yes (ref: No)1.110.72–1.71.628
      Site groupNew York (ref: California)1.841.18–2.86.007
      District of Columbia (ref: California)1.090.58–2.04.789
      Chicago (ref: California)0.800.41–1.56.502
      Number male partners5–24 (ref: 0–4)0.990.63–1.55.962
      ≥25 (ref: 0–4)1.210.72–2.03.471
      Age group (previous visit)25–30 (ref: <25)0.480.28–0.82.007
      30–35 (ref: <25)0.560.34–0.92.022
      >35 (ref: >25)0.630.38–1.05.078
      Alcohol use (previous visit)Drinkers (ref: abstainers)3.972.71–5.81< .0001
      Marijuana use (previous visit)Yes (ref: No)1.681.06–2.67.029
      Cigarette smoking (previous visit)Yes (ref: No)1.981.35–2.91.001
      Mixed models with random intercept and unstructured covariance matrix. P values are 2 tailed.
      aOR, adjusted odds ratio; CI, confidence interval; ref, reference; WIHS, Women’s Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.

      Adjusted model: WLHIV (Table 4)

      Factors associated (P < .05) with miscarriage (vs live birth) included the following: having insurance, drinking alcohol, and smoking cigarettes. Lower log HIV viral load (<4 log10 copies/mL) and using a PI regimen (vs no cART) were protective for miscarriage. Women with HIV viral load <4 log10 copies/mL experienced similar miscarriage proportions compared with women without HIV (data not tabled). None of the 2-way interactions between marijuana use and cigarette smoking, marijuana use and drinking, and drinking and cigarette smoking were significant. In sensitivity analyses when abortion was included in the referent group, we found similarly adjusted findings.
      Table 4Baseline and time-varying factors associated with miscarriage in women living with HIV, WIHS 1994–2017
      VariableLevelaOR95% CIP value
      Insurance (baseline)Yes (ref: no)2.301.12– 4.72.024
      Number of male partners5–24 (ref: 0-4)1.050.55–1.99.885
      ≥25 (ref: 0-4)1.020.49–2.14.955
      Prior miscarriageYes (ref: no)1.250.73–2.14.412
      Alcohol use (previous visit)Drinkers (ref: abstainers)2.961.71–5.12< .001
      Marijuana use (previous visit)Yes (ref: no)1.580.75–3.32.226
      Cigarette smoking (previous visit)Yes (ref: no)2.201.26–3.85.006
      Pregnancy year1998–2001 (ref: 1994–1997)1.020.45–2.34.961
      2002–2005 (ref: 1994–1997)0.800.37–1.75.576
      ≥2006 (ref: 1994–1997)1.300.56–3.05.540
      HIV viral load <4 log10 copies/mL (previous visit)Yes (ref: no)0.450.24–0.84.013
      ART regimencART-containing II, NRTI, NNRTI (ref: no cART)0.690.34–1.40.302
      cART containing PI (ref: no cART)0.400.21–0.79.008
      Mixed models with random intercept and unstructured covariance matrix. P values are 2 tailed. aOR, adjusted odds ratio; cART, combination highly active antiretroviral therapy; CI, confidence interval; NNRTI, non–nucleoside reverse transcriptase inhibitors; NRTI, nucleoside reverse transcriptase inhibitors; PI, protease inhibitor; ref, reference; WIHS, Women's Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.

      Adjusted model: women without HIV (Table 5)

      Factors associated (P < .05) with miscarriage (vs live birth) included the following: drinking alcohol and smoking cigarettes. None of the 2-way interactions between marijuana use and cigarette smoking, marijuana use and drinking, and drinking and cigarette smoking were significant. In sensitivity analyses when abortion was included in the referent group, we found similarly adjusted findings, although marijuana use became statistically significant (adjusted odds ratio [aOR], 1.71; 95% confidence interval, 1.02–2.86, P = .041).
      Table 5Baseline and time-varying factors associated with miscarriage in women without HIV, WIHS 1994–2017
      VariableLevelaOR95% CIP value
      Alcohol use (previous visit)Drinkers (ref: abstainers)4.162.38–7.28< .0001
      Marijuana use (previous visit)Yes (ref: no)1.820.97–3.43.062
      Point estimate is significant in sensitivity analyses including abortion in reference group (aOR, 1.71; 95% CI, 1.02–2.86, P = .041).
      Cigarette smoking (previous visit)Yes (ref: no)1.771.01–3.10.047
      Mixed models with random intercept and unstructured covariance matrix. P values are 2 tailed. aOR, adjusted odds ratio; CI, confidence interval; ref, reference; WIHS, Women's Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.
      a Point estimate is significant in sensitivity analyses including abortion in reference group (aOR, 1.71; 95% CI, 1.02–2.86, P = .041).

      Marijuana frequencies (Figure)

      In addition to dichotomized marijuana use (yes/no), frequency of marijuana use was also significant in bivariate analyses. Among all women, only the highest category of marijuana use (at least daily) reached statistical significance (P = 032) compared with either no marijuana use or none/less than weekly marijuana use. Similar findings were observed for women living with and without HIV, although the adjusted point estimates for frequency of marijuana use did not reach statistical significance after stratifying by HIV status.
      Figure thumbnail gr1
      Figure 1Adjusted associations between marijuana use frequency and miscarriage, WIHS, 1994–2017
      Adjusted association between marijuana use, frequency categories, and miscarriage, WIHS, 1994–2017. Odds ratios are adjusted for the same covariates as the primary analyses. Asterisk indicates P = .0320 comparing items at least daily vs none. Double asteriska indicate P = .0324 comparting items at least daily vs none to less than weekly.
      CI, confidence interval; WIHS, Women's Interagency HIV Study.
      Wall et al. Miscarriage in women with and without HIV. Am J Obstet Gynecol 2019.

      Comment

      Principal Findings

      In this longitudinal study, a similar proportion of women living with and without HIV reported miscarriage (37% and 39%, respectively; P = .638). These proportions are similar to previous (1994–2002) WIHS analyses, which observed 41% of women living with and 34% of women living without HIV reporting miscarriage (among women reporting either a miscarriage or live birth).
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      These proportions are considerably higher than the general population in which miscarriage occurs in an estimated 10–20% of US women,
      • Wilcox A.J.
      • Weinberg C.R.
      • O'Connor J.F.
      • Baird D.D.
      • Schlatterer J.P.
      • Canfield R.E.
      • Armstrong E.G.
      • Nisula B.C.
      Incidence of early loss of pregnancy.
      • Zinaman M.J.
      • Clegg E.D.
      • Brown C.C.
      • O'Connor J.
      • Selevan S.G.
      Estimates of human fertility and pregnancy loss.
      • Wyatt P.R.
      • Owolabi T.
      • Meier C.
      • Huang T.
      Age-specific risk of fetal loss observed in a second trimester serum screening population.
      which was not surprising, given that women in the WIHS are on average of lower socioeconomic status, report more substance use, and have more genital infections relative to the general population, all of which have been associated with miscarriage in other studies.
      • Coleman-Cowger V.H.
      • Oga E.A.
      • Peters E.N.
      • Mark K.
      Prevalence and associated birth outcomes of co-use of Cannabis and tobacco cigarettes during pregnancy.
      • Armstrong B.G.
      • McDonald A.D.
      • Sloan M.
      Cigarette, alcohol, and coffee consumption and spontaneous abortion.
      • Chiodo L.M.
      • Bailey B.A.
      • Sokol R.J.
      • Janisse J.
      • Delaney-Black V.
      • Hannigan J.H.
      Recognized spontaneous abortion in mid-pregnancy and patterns of pregnancy alcohol use.
      • Rasch V.
      Cigarette, alcohol, and caffeine consumption: risk factors for spontaneous abortion.
      • Baud D.
      • Goy G.
      • Jaton K.
      • Osterheld M.-C.
      • et al.
      Role of Chlamydia trachomatis in miscarriage.
      • Oakeshott P.
      • Hay P.
      • Hay S.
      • Steinke F.
      • Rink E.
      • Kerry S.
      Association between bacterial vaginosis or chlamydial infection and miscarriage before 16 weeks' gestation: prospective community based cohort study.
      • Giakoumelou S.
      • Wheelhouse N.
      • Cuschieri K.
      • Entrican G.
      • Howie S.E.M.
      • Horne A.W.
      The role of infection in miscarriage.
      Our study adds several years of follow-up to confirm previous WIHS findings including no association between miscarriage and HIV status,
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      a positive association between miscarriage and higher HIV viral load,
      • Cates J.E.
      • Westreich D.
      • Edmonds A.
      • et al.
      The effects of viral load burden on pregnancy loss among HIV-infected women in the United States.
      and an effect for cigarette use that is stronger among women living with vs without HIV.
      • Westreich D.
      • Cates J.
      • Cohen M.
      • et al.
      Smoking, HIV, and risk of pregnancy loss.
      We also identified new factors associated with miscarriage: alcohol use, marijuana use frequency, and a protective effect for PI vs non-cART use. Interestingly, a history of miscarriage was not associated with miscarriage.

      Results

      Alcohol use was associated with miscarriage in this analysis. Heavy alcohol use in pregnancy has been related to negative birth outcomes, including miscarriage, and the amount and pattern of drinking likely influences these outcomes.
      • Armstrong B.G.
      • McDonald A.D.
      • Sloan M.
      Cigarette, alcohol, and coffee consumption and spontaneous abortion.
      • Chiodo L.M.
      • Bailey B.A.
      • Sokol R.J.
      • Janisse J.
      • Delaney-Black V.
      • Hannigan J.H.
      Recognized spontaneous abortion in mid-pregnancy and patterns of pregnancy alcohol use.
      • Rasch V.
      Cigarette, alcohol, and caffeine consumption: risk factors for spontaneous abortion.
      Although we did not have sufficient sample size to evaluate moderate and heavy drinking, a systematic review reported that low to moderate drinking is not associated with negative birth outcomes including miscarriage.
      • Henderson J.
      • Gray R.
      • Brocklehurst P.
      Systematic review of effects of low-moderate prenatal alcohol exposure on pregnancy outcome.
      However, given that low to moderate alcohol consumption has not been shown to be definitively safe during pregnancy, US medical societies and the Centers for Disease Control and Prevention recommend alcohol avoidance when trying to conceive or during pregnancy/breastfeeding.
      Centers for Disease Control and Prevention
      Alcohol and pregnancy.
      In a previous (1994–2002) WIHS analysis by Massad et al,
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      marijuana use in 2 visits prior to pregnancy in women with HIV was associated with miscarriage (aOR, 6.6 compared with women with no or inconsistent use), but findings for women without HIV were not explored, nor was the effect of frequency of marijuana use for women living with and without HIV.
      In the present analysis, we found that any marijuana use was associated with miscarriage overall, with a trend for increasing marijuana use. Although publications have demonstrated a potential biological mechanism for the effect of marijuana on miscarriage (primarily related to the harmful effect of delta-9-tetrahydrocannabinol, the main psychoactive in marijuana that crosses the placental barrier),
      • Huizink A.
      Prenatal cannabis exposure and infant outcomes: overview of studies.
      • Trabucco E.
      • Acone G.
      • Marenna A.
      • Pierantoni R.
      • et al.
      Endocannabinoid system in first trimester placenta: low FAAH and high CB1 expression characterize spontaneous miscarriage.
      • Friedrich J.
      • Khatib D.
      • Parsa K.
      • Santopietro A.
      • Gallicano G.I.
      The grass isn’t always greener: the effects of cannabis on embryological development.
      several studies failed to show an association between marijuana use and miscarriage.
      Earlier (1987–2002) longitudinal studies of prenatal marijuana exposure in women without HIV found no association between marijuana use during pregnancy and miscarriage.
      • Fergusson D.M.
      • Horwood L.J.
      • Northstone K.
      Maternal use of cannabis and pregnancy outcome.
      • Fried P.A.
      • Makin J.E.
      Neonatal behavioural correlates of prenatal exposure to marihuana, cigarettes and alcohol in a low risk population.
      • Richardson G.A.
      • Ryan C.
      • Willford J.
      • Day N.L.
      • Goldschmidt L.
      Prenatal alcohol and marijuana exposure: Effects on neuropsychological outcomes at 10 years.
      • Kline J.
      • Hutzler M.
      • Levin B.
      • Stein Z.
      • Susser M.
      • Warburton D.
      Marijuana and spontaneous abortion of known karyotype.
      • Wilcox A.J.
      • Weinberg C.R.
      • Baird D.D.
      Risk factors for early pregnancy loss.
      However, contemporary marijuana products have higher levels of delta-9-tetrahydrocannabinol than in the 1980s–1990s.
      • Metz T.D.
      • Stickrath E.H.
      Marijuana use in pregnancy and lactation: a review of the evidence.
      As access to marijuana increases through legalization and stigmatization diminishes, women may be more likely to accurately disclose their marijuana use, and thus, future studies may be more accurate. Given conflicting literature on marijuana use and adverse birth outcomes including miscarriage, low birthweight, preterm birth, and neonatal morbidity,
      • van Gelder M.M.
      • Reefhuis J.
      • Caton A.R.
      • Werler M.M.
      • Druschel C.M.
      • Roeleveld N.
      Characteristics of pregnant illicit drug users and associations between cannabis use and perinatal outcome in a population-based study.
      • Fergusson D.M.
      • Horwood L.J.
      • Northstone K.
      Maternal use of cannabis and pregnancy outcome.
      • Wu C.-S.
      • Jew C.P.
      • Lu H.-C.
      Lasting impacts of prenatal cannabis exposure and the role of endogenous cannabinoids in the developing brain.
      • Metz T.D.
      • Allshouse A.A.
      • Hogue C.J.
      • et al.
      Maternal marijuana use, adverse pregnancy outcomes, and neonatal morbidity.
      • Gunn J.K.
      • Rosales C.B.
      • Center K.E.
      • et al.
      Prenatal exposure to cannabis and maternal and child health outcomes: a systematic review and meta-analysis.
      • Hayatbakhsh M.R.
      • Flenady V.J.
      • Gibbons K.S.
      • et al.
      Birth outcomes associated with cannabis use before and during pregnancy.
      • Conner S.N.
      • Bedell V.
      • Lipsey K.
      • Macones G.A.
      • Cahill A.G.
      • Tuuli M.G.
      Maternal marijuana use and adverse neonatal outcomes: a systematic review and meta-analysis.
      • Coleman-Cowger V.H.
      • Oga E.A.
      • Peters E.N.
      • Mark K.
      Prevalence and associated birth outcomes of co-use of Cannabis and tobacco cigarettes during pregnancy.
      • Fried P.A.
      • Makin J.E.
      Neonatal behavioural correlates of prenatal exposure to marihuana, cigarettes and alcohol in a low risk population.
      • Richardson G.A.
      • Ryan C.
      • Willford J.
      • Day N.L.
      • Goldschmidt L.
      Prenatal alcohol and marijuana exposure: Effects on neuropsychological outcomes at 10 years.
      • Kline J.
      • Hutzler M.
      • Levin B.
      • Stein Z.
      • Susser M.
      • Warburton D.
      Marijuana and spontaneous abortion of known karyotype.
      • Wilcox A.J.
      • Weinberg C.R.
      • Baird D.D.
      Risk factors for early pregnancy loss.
      • Jansson L.M.
      • Jordan C.J.
      • Velez M.L.
      Perinatal marijuana use and the developing child.
      professional societies recommend women discontinue marijuana during pregnancy, while breastfeeding, or when attempting to conceive.
      American College of Obstetrics and Gynecology
      Marijuana use during pregnancy and lactation. ACOG Committee opinion no. 722.
      In a previous WIHS analysis, ART at the visit prior to pregnancy appeared protective against miscarriage (aOR, 0.37).
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      Here we explored specific cART regimens and found that PI-containing regimens vs no cART were significantly protective for miscarriage, while other cART regimens may also be protective (although their association did not reach statistical significance). While data are conflicting, some studies show cART to be associated with adverse pregnancy outcomes including preterm delivery and low birthweight.
      • Machado E.S.
      • Hofer C.B.
      • Costa T.T.
      • et al.
      Pregnancy outcome in women infected with HIV-1 receiving combination antiretroviral therapy before versus after conception.
      • Townsend C.L.
      • Cortina-Borja M.
      • Peckham C.S.
      • Tookey P.A.
      Antiretroviral therapy and premature delivery in diagnosed HIV-infected women in the United Kingdom and Ireland.
      • Powis K.M.
      • Kitch D.
      • Ogwu A.
      • et al.
      Increased risk of preterm delivery among HIV-infected women randomized to protease versus nucleoside reverse transcriptase inhibitor-based HAART during pregnancy.
      A review of studies published between 1980 and 2017 compared women receiving tenofovir vs nontenofovir cART regimens during pregnancy and found no increased miscarriage risk,
      • Nachega J.B.
      • Uthman O.A.
      • Mofenson L.M.
      • et al.
      Safety of tenofovir disoproxil fumarate-based antiretroviral therapy regimens in pregnancy for HIV-infected women and their infants: a systematic review and meta-analysis.
      and a study of zidovudine vs nonzidovudine cART regimens found a protective effect for miscarriage.
      • Vannappagari V.
      • Koram N.
      • Albano J.
      • Tilson H.
      • Gee C.
      Association between in utero zidovudine exposure and nondefect adverse birth outcomes: analysis of prospectively collected data from the Antiretroviral Pregnancy Registry.
      Meanwhile, PIs may lower progesterone levels,
      • Papp E.
      • Mohammadi H.
      • Loutfy M.R.
      • et al.
      HIV protease inhibitor use during pregnancy is associated with decreased progesterone levels, suggesting a potential mechanism contributing to fetal growth restriction.
      possibly leading to adverse pregnancy outcomes including miscarriage,
      • Raghupathy R.
      • Al-Mutawa E.
      • Al-Azemi M.
      • Makhseed M.
      • Azizieh F.
      • Szekeres-Bartho J.
      Progesterone-induced blocking factor (PIBF) modulates cytokine production by lymphocytes from women with recurrent miscarriage or preterm delivery.
      preterm birth,
      • Polgár B.T.
      • Nagy E.
      • Mikó E.V.
      • Varga P.T.
      • Szekeres-Barthó J.L.
      Urinary progesterone-induced blocking factor concentration is related to pregnancy outcome.
      and lower birthweight.
      • Salas S.P.
      • Marshall G.
      • Gutierrez B.L.
      • Rosso P.
      Time course of maternal plasma volume and hormonal changes in women with preeclampsia or fetal growth restriction.
      History of miscarriage was not associated with miscarriage in our adjusted models. This finding contradicts Massad et al
      • Massad L.S.
      • Springer G.
      • Jacobson L.
      • et al.
      Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
      in which miscarriage was significantly associated with prior miscarriage in women with HIV (aOR, 1.94). This difference may be related to our inclusion of an additional 15 years of study data and control for influential covariates such as alcohol and cART. Importantly, we are not differentiating between 1–2 vs recurrent prior miscarriage (3 or more), the latter being much less common and likely reflective of underlying factors that may increase miscarriage risk.

      Research implications

      The protective association with PIs persists when controlling for viral load, indicating that possible mechanism of protection is not conferred by viral suppression. This mechanism warrants further exploration.

      Clinical implications

      Alcohol consumption and marijuana use are potentially modifiable behaviors that providers can counsel on to have an impact on the miscarriage risk. The finding that PIs appeared significantly protective for miscarriage importantly adds to PI safety data.

      Strengths and limitations

      Several limitations warrant discussion. First, miscarriage in the WIHS is self-reported. The risk of miscarriage is highest in the first several weeks of pregnancy
      • Wilcox A.J.
      • Baird D.D.
      • Weinberg C.R.
      Time of implantation of the conceptus and loss of pregnancy.
      when women and their providers may not diagnose early pregnancy.
      • Pandya P.P.
      • Snijders R.J.
      • Psara N.
      • Hilbert L.
      • Nicolaides K.H.
      The prevalence of non-viable pregnancy at 10–13 weeks of gestation.
      • Wilcox A.J.
      • Horney L.F.
      Accuracy of spontaneous abortion recall.
      This could contribute to ascertainment bias. Additionally, use of marijuana, alcohol, tobacco use, and STI could be differentially reported by the outcome of interest, with women experiencing miscarriage perhaps less likely to report exposure. However, self-reported data in other WIHS studies has been strongly correlated with clinical outcomes.
      We did not have sufficient numbers to explore the role of individual STIs or levels of drinking. We also did not have measures of time-varying intimate partner violence, which may be associated with miscarriage. Future analyses could consider the role of various measures of smoking frequency and different time-varying combinations of ART regimens. Although rare, it is possible that a proportion of miscarriages could include early ectopic pregnancies. Although we excluded the 2011–2012 wave and southern because of small numbers, when combined with the analysis cohort, findings did not meaningfully change. As the southern cohort accrues more women, future analyses will be able to draw conclusions from this group. Finally, our results are most generalizable to US urban women living with and without HIV who are primarily older and African American.

      Conclusions

      The WIHS is one of the largest and longest-running cohorts of women living with and without HIV in the world and is well validated with standardized data collection procedures and highly trained staff. Our study adds several years of analytic follow-up to confirm previous findings and identify new findings related to alcohol use, marijuana use frequency, and the protective effect of PIs. These findings highlight additional potentially modifiable behaviors addressable via interventions that could reduce miscarriage risk.

      Acknowledgment

      Data in this manuscript were collected by the Women’s Interagency HIV Study (WIHS). WIHS (principal investigators) include the following: University of Alabama at Birmingham-Mississippi WIHS (Mirjam-Colette Kempf and Deborah Konkle-Parker), grant U01-AI-103401; Atlanta WIHS (Ighovwerha Ofotokun, Anandi, Sheth, and Gina Wingood), grant U01-AI-103408; Bronx WIHS (Kathryn Anastos and Anjali Sharma), grant U01-AI-035004; Brooklyn WIHS (Deborah Gustafson and Tracey Wilson), grant U01-AI-031834; Chicago WIHS (Mardge Cohen and Audrey French), grant U01-AI-034993; metropolitan Washington WIHS (Seble Kassaye and Daniel Merenstein), grant U01-AI-034994; Miami WIHS (Maria Alcaide, Margaret Fischl, and Deborah Jones), grant U01-AI-103397; University of North Carolina WIHS (Adaora Adimora), grant U01-AI-103390; Connie Wofsy Women’s HIV Study, northern California (Bradley Aouizerat and Phyllis Tien), grant U01-AI-034989; WIHS Data Management and Analysis Center (Stephen Gange and Elizabeth Golub), grant U01-AI-042590; and southern California WIHS (Joel Milam), grant U01-HD-032632 (WIHS I–WIHS IV).

      References

        • Haddad L.B.
        • Wall K.M.
        • Mehta C.C.
        • et al.
        Trends of and factors associated with live-birth and abortion rates among HIV-positive and HIV-negative women.
        Am J Obstet Gynecol. 2017; 216: 71.e1-71.e16
        • Wilcox A.J.
        • Weinberg C.R.
        • O'Connor J.F.
        • Baird D.D.
        • Schlatterer J.P.
        • Canfield R.E.
        • Armstrong E.G.
        • Nisula B.C.
        Incidence of early loss of pregnancy.
        N Engl J Med. 1988; 319: 189-194
        • Zinaman M.J.
        • Clegg E.D.
        • Brown C.C.
        • O'Connor J.
        • Selevan S.G.
        Estimates of human fertility and pregnancy loss.
        Fertil Steril. 1996; 65: 503-509
        • Wyatt P.R.
        • Owolabi T.
        • Meier C.
        • Huang T.
        Age-specific risk of fetal loss observed in a second trimester serum screening population.
        Am J Obstet Gynecol. 2005; 192: 240-246
        • Schwartz S.R.
        • Rees H.
        • Mehta S.
        • Venter W.D.F.
        • Taha T.E.
        • Black V.
        High incidence of unplanned pregnancy after antiretroviral therapy initiation: findings from a prospective cohort study in South Africa.
        PLoS One. 2012; 7: e36039
        • Chen J.Y.
        • Ribaudo H.J.
        • 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
        • Ezechi O.C.
        • Gab-Okafor C.V.
        • Oladele D.A.
        • et al.
        Pregnancy, obstetric and neonatal outcomes in HIV positive Nigerian women.
        Afr J Reprod Health. 2013; 17: 160-168
        • Massad L.S.
        • Springer G.
        • Jacobson L.
        • et al.
        Pregnancy rates and predictors of conception, miscarriage and abortion in US women with HIV.
        AIDS (London, England). 2004; 18: 281-286
        • van Gelder M.M.
        • Reefhuis J.
        • Caton A.R.
        • Werler M.M.
        • Druschel C.M.
        • Roeleveld N.
        Characteristics of pregnant illicit drug users and associations between cannabis use and perinatal outcome in a population-based study.
        Drug Alcohol Depend. 2010; 109: 243-247
        • Fergusson D.M.
        • Horwood L.J.
        • Northstone K.
        Maternal use of cannabis and pregnancy outcome.
        BJOG. 2002; 109: 21-27
        • Wu C.-S.
        • Jew C.P.
        • Lu H.-C.
        Lasting impacts of prenatal cannabis exposure and the role of endogenous cannabinoids in the developing brain.
        Future Neur. 2011; 6: 459-480
        • Metz T.D.
        • Allshouse A.A.
        • Hogue C.J.
        • et al.
        Maternal marijuana use, adverse pregnancy outcomes, and neonatal morbidity.
        Am J Obstet Gynecol. 2017; 217: 478.e1-478.e8
        • Gunn J.K.
        • Rosales C.B.
        • Center K.E.
        • et al.
        Prenatal exposure to cannabis and maternal and child health outcomes: a systematic review and meta-analysis.
        BMJ Open. 2016; 6: e009986
        • Hayatbakhsh M.R.
        • Flenady V.J.
        • Gibbons K.S.
        • et al.
        Birth outcomes associated with cannabis use before and during pregnancy.
        Pediatr Res. 2011; 71: 215
        • Conner S.N.
        • Bedell V.
        • Lipsey K.
        • Macones G.A.
        • Cahill A.G.
        • Tuuli M.G.
        Maternal marijuana use and adverse neonatal outcomes: a systematic review and meta-analysis.
        Obstet Gynecol. 2016; 128: 713-723
        • Machado E.S.
        • Hofer C.B.
        • Costa T.T.
        • et al.
        Pregnancy outcome in women infected with HIV-1 receiving combination antiretroviral therapy before versus after conception.
        Sex Transm Infect. 2009; 85: 82-87
        • Townsend C.L.
        • Cortina-Borja M.
        • Peckham C.S.
        • Tookey P.A.
        Antiretroviral therapy and premature delivery in diagnosed HIV-infected women in the United Kingdom and Ireland.
        AIDS (London, England). 2007; 21: 1019-1026
        • Powis K.M.
        • Kitch D.
        • Ogwu A.
        • et al.
        Increased risk of preterm delivery among HIV-infected women randomized to protease versus nucleoside reverse transcriptase inhibitor-based HAART during pregnancy.
        J Infect Dis. 2011; 204: 506-514
        • Polgár B.T.
        • Nagy E.
        • Mikó E.V.
        • Varga P.T.
        • Szekeres-Barthó J.L.
        Urinary progesterone-induced blocking factor concentration is related to pregnancy outcome.
        Biol Reprod. 2004; 71: 1699-1705
        • Salas S.P.
        • Marshall G.
        • Gutierrez B.L.
        • Rosso P.
        Time course of maternal plasma volume and hormonal changes in women with preeclampsia or fetal growth restriction.
        Hypertension (Dallas, Tex: 1979). 2006; 47: 203-208
        • Coleman-Cowger V.H.
        • Oga E.A.
        • Peters E.N.
        • Mark K.
        Prevalence and associated birth outcomes of co-use of Cannabis and tobacco cigarettes during pregnancy.
        Neurotoxicol Teratol. 2018; 68: 84-90
        • Chabarria K.C.
        • Racusin D.A.
        • Antony K.M.
        • et al.
        Marijuana use and its effects in pregnancy.
        Am J Obstet Gynecol. 2016; 215: 506.e1-506.e7
        • Kim H.Y.
        • 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
        • Turner A.N.
        • Tabbah S.
        • Mwapasa V.
        • et al.
        Severity of maternal HIV-1 disease is associated with adverse birth outcomes in Malawian women: a cohort study.
        J Acquir Immune Defic Syndr (1999). 2013; 64: 392-399
        • Cates J.E.
        • Westreich D.
        • Edmonds A.
        • et al.
        The effects of viral load burden on pregnancy loss among HIV-infected women in the United States.
        Infect Dis Obstet Gynecol. 2015; 2015362357
        • Westreich D.
        • Cates J.
        • Cohen M.
        • et al.
        Smoking, HIV, and risk of pregnancy loss.
        AIDS (London, England). 2017; 31: 553-560
        • Barkan S.E.
        • Melnick S.L.
        • Preston-Martin S.
        • et al.
        The Women's Interagency HIV Study. WIHS Collaborative Study Group.
        Epidemiology. 1998; 9: 117-125
        • Bacon M.C.
        • von Wyl V.
        • Alden C.
        • et al.
        The Women's Interagency HIV Study: an observational cohort brings clinical sciences to the bench.
        Clin Diagn Lab Immunol. 2005; 12: 1013-1019
        • Armstrong B.G.
        • McDonald A.D.
        • Sloan M.
        Cigarette, alcohol, and coffee consumption and spontaneous abortion.
        Am J Public Health. 1992; 82: 85-87
        • Chiodo L.M.
        • Bailey B.A.
        • Sokol R.J.
        • Janisse J.
        • Delaney-Black V.
        • Hannigan J.H.
        Recognized spontaneous abortion in mid-pregnancy and patterns of pregnancy alcohol use.
        Alcohol (Fayetteville, NY). 2012; 46: 261-267
        • Rasch V.
        Cigarette, alcohol, and caffeine consumption: risk factors for spontaneous abortion.
        Acta Obstet Gynecol Scand. 2003; 82: 182-188
        • Baud D.
        • Goy G.
        • Jaton K.
        • Osterheld M.-C.
        • et al.
        Role of Chlamydia trachomatis in miscarriage.
        Emerg Infect Dis. 2011; 17: 1630-1635
        • Oakeshott P.
        • Hay P.
        • Hay S.
        • Steinke F.
        • Rink E.
        • Kerry S.
        Association between bacterial vaginosis or chlamydial infection and miscarriage before 16 weeks' gestation: prospective community based cohort study.
        BMJ (Clin Res Ed). 2002; 325: 1334
        • Giakoumelou S.
        • Wheelhouse N.
        • Cuschieri K.
        • Entrican G.
        • Howie S.E.M.
        • Horne A.W.
        The role of infection in miscarriage.
        Hum Reprod Update. 2016; 22: 116-133
        • Henderson J.
        • Gray R.
        • Brocklehurst P.
        Systematic review of effects of low-moderate prenatal alcohol exposure on pregnancy outcome.
        BJOG. 2007; 114: 243-252
        • Centers for Disease Control and Prevention
        Alcohol and pregnancy.
        Centers for Disease Control and Prevention, Atlanta, GA2016 (Available at:) (Accessed December 2018)
        • Huizink A.
        Prenatal cannabis exposure and infant outcomes: overview of studies.
        Prog Neuro-Psychopharmacol Biol Psychiatry. 2014; 52: 45-52
        • Trabucco E.
        • Acone G.
        • Marenna A.
        • Pierantoni R.
        • et al.
        Endocannabinoid system in first trimester placenta: low FAAH and high CB1 expression characterize spontaneous miscarriage.
        Placenta. 2009; 30: 516-522
        • Friedrich J.
        • Khatib D.
        • Parsa K.
        • Santopietro A.
        • Gallicano G.I.
        The grass isn’t always greener: the effects of cannabis on embryological development.
        BMC Pharmacol Toxicol. 2016; 17: 45
        • Fried P.A.
        • Makin J.E.
        Neonatal behavioural correlates of prenatal exposure to marihuana, cigarettes and alcohol in a low risk population.
        Neurotoxicol Teratol. 1987; 9: 1-7
        • Richardson G.A.
        • Ryan C.
        • Willford J.
        • Day N.L.
        • Goldschmidt L.
        Prenatal alcohol and marijuana exposure: Effects on neuropsychological outcomes at 10 years.
        Neurotoxicol Teratol. 2002; 24: 309-320
        • Kline J.
        • Hutzler M.
        • Levin B.
        • Stein Z.
        • Susser M.
        • Warburton D.
        Marijuana and spontaneous abortion of known karyotype.
        Paediatr Perinat Epidemiol. 1991; 5: 320-332
        • Wilcox A.J.
        • Weinberg C.R.
        • Baird D.D.
        Risk factors for early pregnancy loss.
        Epidemiology. 1990; 1: 382-385
        • Metz T.D.
        • Stickrath E.H.
        Marijuana use in pregnancy and lactation: a review of the evidence.
        Am J Obstet Gynecol. 2015; 213: 761-778
        • Jansson L.M.
        • Jordan C.J.
        • Velez M.L.
        Perinatal marijuana use and the developing child.
        JAMA. 2018; 320: 545-546
        • American College of Obstetrics and Gynecology
        Marijuana use during pregnancy and lactation. ACOG Committee opinion no. 722.
        Obstet Gynecol. 2017; 130: e205-e209
        • Nachega J.B.
        • Uthman O.A.
        • Mofenson L.M.
        • et al.
        Safety of tenofovir disoproxil fumarate-based antiretroviral therapy regimens in pregnancy for HIV-infected women and their infants: a systematic review and meta-analysis.
        J Acquir Immune Defic Syndr (1999). 2017; 76: 1-12
        • Vannappagari V.
        • Koram N.
        • Albano J.
        • Tilson H.
        • Gee C.
        Association between in utero zidovudine exposure and nondefect adverse birth outcomes: analysis of prospectively collected data from the Antiretroviral Pregnancy Registry.
        BJOG. 2016; 123: 910-916
        • Papp E.
        • Mohammadi H.
        • Loutfy M.R.
        • et al.
        HIV protease inhibitor use during pregnancy is associated with decreased progesterone levels, suggesting a potential mechanism contributing to fetal growth restriction.
        J Infect Dis. 2015; 211: 10-18
        • Raghupathy R.
        • Al-Mutawa E.
        • Al-Azemi M.
        • Makhseed M.
        • Azizieh F.
        • Szekeres-Bartho J.
        Progesterone-induced blocking factor (PIBF) modulates cytokine production by lymphocytes from women with recurrent miscarriage or preterm delivery.
        J Reprod Immunol. 2009; 80: 91-99
        • Wilcox A.J.
        • Baird D.D.
        • Weinberg C.R.
        Time of implantation of the conceptus and loss of pregnancy.
        N Engl J Med. 1999; 340: 1796-1799
        • Pandya P.P.
        • Snijders R.J.
        • Psara N.
        • Hilbert L.
        • Nicolaides K.H.
        The prevalence of non-viable pregnancy at 10–13 weeks of gestation.
        Ultrasound Obstet Gynecol. 1996; 7: 170-173
        • Wilcox A.J.
        • Horney L.F.
        Accuracy of spontaneous abortion recall.
        Am J Epidemiol. 1984; 120: 727-733