Epidemiology & Social
Pregnancy and birth rates among HIV-infected women in the United States: the confounding effects of illicit drug use
Forsyth, Brian W. C.a; Davis, Julie A.c; Freudigman, Kimberly A.a; Katz, Karol H.b; Zelterman, Danielb
From the aDepartment of Pediatrics, and bCenter for Interdisciplinary Research on AIDS, Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven, CT, USA; and cDepartment of Pediatrics, Oregon Health Sciences University, Portland, OR, USA.
Correspondence to: Brian W.C. Forsyth, MB ChB, Department of Pediatrics, Yale University School of Medicine, 333 Cedar Street, PO Box 208064, New Haven, CT 06520-8064, USA. Tel: +1 203 688 2475; fax: +1 203 785 3932; e-mail: firstname.lastname@example.org
Received: 26 October 2000;
revised: 20 September 2001; accepted: 26 September 2001.
Objective: To determine the effect of HIV infection on pregnancy and birth rates and assess the potentially confounding effect of illicit drug use.
Design: A retrospective record review of matched cohorts examining pregnancy outcomes for HIV-positive women and two HIV-negative comparison groups (one matched by drug use).
Methods: Ninety HIV-positive women who gave birth in a US city between 1989 and 1993 were matched to HIV-negative women by race, age, parity and date of index birth (group 1, N = 180) and also by the type of illicit drug used (group 2, N = 90). Data were abstracted on tubal ligations and pregnancies occurring before April 1996.
Results: A total of 63% of HIV-positive women used cocaine during the index pregnancy and 26% also used opiates. HIV-positive women had fewer tubal ligations than group 1 (38.9% versus 51.1%, P = 0.058), but there was no difference when matching included drug use (38.9% in group 2). HIV infection was associated with a decrease in the number of pregnancies; this decrease was most marked when matching included drug use (18.0 versus 32.1 pregnancies per 100 woman-years, P < 0.01). There were no significant differences in spontaneous or therapeutic terminations. Poisson regression analysis demonstrated that HIV infection and older age were associated with fewer pregnancies, and cocaine use with an increased pregnancy rate.
Conclusion: This study confirms that HIV infection is associated with a decrease in the number of pregnancies, but also illustrates the confounding effects of illicit drug use among women in the United States.
Previous studies examining the effect of HIV infection on pregnancy and birth rates among HIV-infected women have provided conflicting results. In general, studies performed in the earlier part of the epidemic suggested that HIV infection did not alter the likelihood of women becoming pregnant [1–3], nor did it affect the rates of either spontaneous abortions or therapeutic terminations of pregnancy [4–6].However, more recent studies have suggested significant differences in both the rates and outcomes of pregnancy for HIV-infected women [7–16]. Of note is the fact that most of the studies reported in the earlier years of the epidemic were performed in north America, whereas more recent studies have been performed in Europe [9,10], Australia , and developing countries in Africa [12–16], suggesting the possibility that regional differences or differences in populations affected by HIV could explain the differences in results. However, recent studies performed in north America have also contradicted the results of earlier studies, and have documented decreased pregnancy rates among HIV-infected women [7,8].
For an HIV-infected woman there are many factors that could contribute both to the likelihood of becoming pregnant and whether or not the pregnancy results in a live birth. First, the duration of time in which she can become pregnant may be shortened because of death or a decision to have a tubal ligation. Other factors to do with the illness may decrease her fertility or her ability to become pregnant. Finally, in some studies [2,9–11,17,18], HIV-infected women have more frequently chosen to terminate a pregnancy. However, to date, little attention has been paid to the potentially confounding effects of drug use, although in the United States there is a substantial link between the AIDS epidemic and the use of illicit drugs, including cocaine [19,20]. The effects of drug use might have a substantial impact on both the likelihood of becoming pregnant and maintaining the pregnancy. For example, women who use cocaine have high rates of unprotected sexual activity [20–23], and tend to have more children than women who are not using drugs . Therefore, if the number of HIV-infected women who use drugs, particularly cocaine, is disproportionately high, then the behaviors associated with drug use might obscure a potential decrease in fertility associated with HIV infection. This might be further complicated by the fact that cocaine users have also been reported to have both spontaneous abortions and therapeutic terminations of pregnancy more frequently . In addition, although substantial attention has been paid to the use of condoms among HIV-infected women, less attention has been paid to decisions about tubal ligation, another area in which drug use might have an important confounding effect. The purpose of this study was thus to compare the rates of pregnancy and of live births among HIV-infected women with those of suitably matched HIV-negative women, and to assess the potentially confounding effects of illicit drug use, with particular attention to the frequency and timing of tubal ligations. We hypothesized that HIV infection may result in a decrease in the number of pregnancies and births, but this finding might be obscured by the fact that drug use among HIV-infected women has an opposite effect. Therefore, when comparing HIV-infected and non-infected women, there may be no differences in pregnancy outcomes if drug use is ignored, but when the groups are matched by drug use, the effect of HIV infection becomes evident.
A matched cohort design was used to compare the outcomes of pregnancy for HIV-positive and HIV-negative women. To assess the potentially confounding effects of drug use, two groups of women were included for comparison: subjects for the first comparison group were selected without attention to drug use, whereas in the second comparison group, drug use and the type of drug being used was included as a matching variable. In addition, to minimize the possibility that other differences between HIV-infected women and non-infected women might affect outcomes and obscure the effects of HIV disease, HIV-infected and comparison subjects were matched by race, age and the number of previous children. The matching of women at a specific time, the birth of a child (referred to as the zero-time birth), was used to control for the potential for future reproduction between groups. The study was conducted by reviewing the medical records of HIV-positive and comparison women who had given birth to children between 1 April 1989 and 31 March 31 1993, and comparing pregnancy outcomes over the subsequent period to 31 March 1996.
Selection and matching of subjects
Potential subjects were women who had given birth in the specified time period at either of the only two hospitals in New Haven, Connecticut, USA, and who were residents of either New Haven or one of three adjacent towns. Index subjects were all women who were known to be HIV positive at the time of delivery. Comparison subjects were women who were either HIV negative or had not been tested for HIV infection at the time of the index delivery and were not found to be positive for HIV during the follow-up period. Subjects in the first comparison group were matched to index women using four matching variables: race, age at the time of the index delivery (± 3 years), parity, and the date of the index birth (± 6 months). For this first comparison group, two HIV-negative women were identified for each index subject (1 : 2 match). For the second comparison group, drug use during pregnancy was included as an additional matching variable: if the index subject had documentation of drug use during the index pregnancy, then a comparison subject was identified (1 : 1 match) who also had documentation of using the same class of illicit drug (cocaine or opiates) during the index pregnancy.
Subjects for the first comparison group were identified from a database of all deliveries at Yale–New Haven Hospital during the specified time period. For the second comparison group, in which drug use during pregnancy was included as a matching variable, subjects were selected from among women for whom identifying data had already been obtained in previous studies of drug use in pregnancy [26–28]. Lists of potential comparison subjects were ordered by birth date and were reviewed by alternately going forwards or backwards from the birth date of the index subject to identify comparison subjects who satisfied the matching criteria. Whenever a perfect match for all variables could not be found, the comparison subject who matched most closely was identified by dropping the matching variables in the following order: proximity of birth dates, age of the mother, parity, and race.
Baseline data included sociodemographic variables, the number of previous pregnancies and births, and data related to the zero-time pregnancy, including the use of drugs during pregnancy. Prenatal care was categorized as adequate if the subject registered before 28 weeks’ gestation and had at least four obstetric visits before delivery. For the HIV-infected subjects, additional information was collected regarding the timing of the first HIV test, documentation of CD4 lymphocyte count, and previous AIDS-defining conditions. AIDS was defined using the 1994 revised definition, which includes a CD4 cell count of less than 200/μl.
Follow-up data were obtained by reviewing all clinic visits and hospital admissions occurring between the zero-time birth and 31 March 1996. Records were reviewed for notations of pregnancies, births, therapeutic abortions, spontaneous abortions, stillbirths and tubal ligations. Any changes in pregnancy notations were examined for consistency. For the HIV-infected women, the progression of disease was documented by abstracting data on all HIV-related illnesses and CD4 cell levels.
The chi-square statistic was used to assess differences in proportions and McNemar's test was applied for comparisons of matched cohorts . The Student's t-test was used to assess differences in continuous data. Conditional logistic regression analysis using a backward elimination technique was used to determine whether potential confounding variables, such as deaths among the HIV-infected women or differences in the timing of tubal ligations, might explain differences between groups in the overall likelihood that women would become pregnant . This analysis takes into account the matching of case and comparison subjects, and determines which variables contribute to discordance in outcomes. For example, in the first comparison in which there were two matched comparison subjects for every index subject, a discordant outcome included the index subject not having a pregnancy when one or both comparison subjects did become pregnant. Poisson regression analysis was then performed to examine the effect of independent variables on the rate of pregnancies during the time women were eligible to become pregnant (i.e., before tubal ligation) .Because this analyses does not take into account the matching process, all three groups were combined in the analysis.
One hundred and six HIV-positive women were identified. Of these, 16 were excluded either because their medical records could not be located (12) or because the medical records contained no follow-up data after the zero-time delivery (four). The study group included 90 HIV-positive subjects (`index group'), 180 comparison subjects for whom drug use was not included as a matching variable (`first comparison group'), and 90 comparison subjects for whom drug use was included as a matching variable (`second comparison group'). Three potential comparison subjects were excluded and replacement subjects found because there were no follow-up data after the zero-time birth.
The characteristics of study subjects used for matching are shown in Table 1. The majority of the women were African-American, their mean age was 28 years, and approximately two-thirds of the women had had at least two children before the zero-time birth. Among the index group, 63.3% used drugs during the pregnancy: 34 women used only cocaine and 23 others used both opiates and cocaine. Only one woman used only opiates. HIV-positive women were significantly more likely to have used drugs during the pregnancy than women in the first comparison group in which drug use was not included as a matching variable (P < 0.0001).
Pregnancy history at baseline
Women in all groups had an average of more than four pregnancies before the baseline pregnancy, and similar proportions of women in each group had previously had spontaneous abortions and therapeutic terminations (see Table 2). There were very few stillbirths to women in all three groups: four HIV-positive subjects (4.4%); three subjects from the first comparison group (1.7%); and one subject from the second comparison group (1.1%) had a history of a stillbirth before the index birth.
For the zero-time pregnancy, significantly fewer HIV-positive women received adequate prenatal care (P < 0.0001), and a greater proportion had premature births (P < 0.01) compared with those in the first comparison group, but these proportions were very similar when subjects were matched by drug use in the second comparison group. A greater proportion of HIV-positive subjects had cesarean sections, although differences between groups were not statistically significant.
Description of HIV-positive subjects and clinical progression
Seventy-two (80%) of the HIV-positive women first learned they were HIV positive during the zero-time pregnancy, whereas the remaining 18 had known they were infected before the pregnancy. Only three of these 18 women had previously had children who were HIV infected. Of the HIV-positive women, 18 (20.0%) had CDC-defined AIDS at the time of the index delivery: seven had had an AIDS-defining illness and an additional 11 had CD4 cell levels of less than 200/μl. During the follow-up period, 22 (24.4%) additional HIV-infected subjects developed AIDS: 21 developed AIDS-defining illnesses and an additional one had CD4 cell levels that dropped below 200/μl. A total of 40 HIV-positive women (44.4%) had AIDS during the study. Nineteen (21.1%) women died during the follow-up period, 10 of whom had been diagnosed with AIDS before the zero-time birth.
Tubal ligations and length of follow-up interval in which women could become pregnant
Fewer HIV-positive women had tubal ligations (38.9%) than did subjects in the first comparison group (51.1%, P = 0.058); however, the proportion having tubal ligations was identical in the second comparison group (38.9%) in which drug use was included as a matching variable. Women who used cocaine during the index pregnancy were less likely to have a tubal ligation (38.5%) than women who did not use cocaine (49.3%, P < 0.05). Despite the differences between groups in the rates of tubal ligations, deaths among the HIV-infected women resulted in the period of time in which women were eligible to become pregnant being similar between the index group (mean 33.3 months) and the first comparison group (36.5). However, in the second comparison, because the proportion of women having tubal ligations was identical in the two groups, deaths among the HIV-positive women resulted in this group having a significantly shorter follow-up interval in which they could become pregnant (mean 33.3 months versus 41.9 months in the second comparison group, P < 0.05).
Pregnancy outcomes and pregnancy rates during follow-up
Table 3 shows the outcomes for the 290 pregnancies documented during the follow-up period. Of these, there were no details in the medical record for 27 (9.3%), but instead information was extrapolated from changes in obstetric history notations: 16 were therapeutic terminations, 10 were spontaneous abortions, and one was a birth that was presumed to have occurred elsewhere.
HIV-positive women had fewer pregnancies than did women in the comparison groups, and these differences were most marked in the second comparison, in which drug use was included as a matching variable. For example, only 34.4% of HIV-infected women had a pregnancy during the period of follow-up, whereas 46.1% (P = 0.068) and 64.4% (P < 0.0001) of women in the first and second comparison groups, respectively, had at least one pregnancy. Pregnancy rates (before tubal ligation) were significantly lower among HIV-infected women (18.0 per 100 woman-years) than either the first comparison group (26.3 per 100 woman-years, P = 0.02) or the second comparison group (32.1 per 100 woman-years, P < 0.01). Similar differences occurred in the number of births in each group, and as before, the difference was most marked in the second comparison. There were no significant differences between groups in the proportion of pregnancies that ended in either spontaneous or therapeutic abortions (see Table 3). There were two stillbirths in the first comparison group and none in the other two groups. (In addition, there were four instances, three in comparison group 1, and one in comparison group 2, in which there was documentation in the medical record regarding a pregnancy but no notation of the outcomes of the pregnancies).
Analyses to determine whether differences between groups might explain why fewer HIV-infected women became pregnant
Conditional regression analyses were conducted to determine whether differences between groups might explain why fewer HIV-infected women became pregnant during the follow-up period (Table 4). These analyses take into account the fact that comparison groups were matched, and only those variables for which there were differences between groups were entered into the models. Most importantly, we wished to examine the extent to which deaths and differences in the timing of tubal ligations might have resulted in fewer HIV-infected women becoming pregnant, thus the period of time lost from follow-up because of death or tubal ligation (the interval between the date of the event and the study termination on 31 March 1996) was included in the analyses. The results confirmed that HIV infection was significantly associated with a decreased likelihood of a woman becoming pregnant, and that this association was not explained by differences in the duration of time in which women were eligible to become pregnant. In the first comparison, which included cocaine and opiate use among the independent variables, two variables, cocaine use and inadequate prenatal care in the zero-time pregnancy, were associated with an increased likelihood of becoming pregnant. In the second comparison, in which drug use had been included in the matching process, only HIV infection remained in the model.
Variables contributing to decreased pregnancy rates
First, to determine whether the progression of disease might contribute to a decrease in the number of pregnancies among HIV-infected women, we compared pregnancy rates for women before and after the diagnosis of AIDS. Of the 45 pregnancies among the HIV-infected women, 32 occurred before the diagnosis of AIDS, and 13 occurred in women who had already been diagnosed with AIDS. However, taking periods of time into account, the rates of pregnancy were similar before and after the AIDS diagnosis: 17.9 and 18.6 pregnancies per 100 woman-years, respectively.
Poisson regression analyses were then conducted to examine which variables contributed to the number of pregnancies a woman might have during the time she was eligible to be pregnant (i.e. before tubal ligation). Unlike conditional regression analysis, Poisson regression does not take into account the matching process, so for this analysis all groups were included in one model, and the independent variables examined included the matching variables (race, age, parity and drug use) and the one variable that was significantly different between cases and comparison subjects: inadequate prenatal care in the zero-time pregnancy. As shown in Table 5, HIV infection and increased maternal age were associated with decreased pregnancy rates and cocaine use in the zero-time pregnancy was associated with an increased pregnancy rate. Opiate use had an opposite effect, and was associated with a decreased pregnancy rate, although not at a level of statistical significance.
The results of this study show that in the second decade of the AIDS epidemic, HIV-infected women in the United States were getting pregnant less often and having fewer children when compared with HIV-negative women, and that this decrease is more striking when the women's history of drug use is taken into consideration. The decrease in the number of births to HIV-infected women is a result of fewer pregnancies among these women, and not to an increase in the number of spontaneous abortions or therapeutic terminations of pregnancies. These results confirm the findings of a number of studies performed in different countries [7–16], and contradict the results of those studies that found no decrease in pregnancy rates among HIV-infected women [1–3].
The effect of drug use
The populations studied in the earlier studies were diverse, as were the research methods used. In our study, because we hypothesized that drug use might have important implications for HIV-infected women in the United States, we used a design that would both control for the effect of drugs as well as illustrate the extent to which illicit drug use might explain earlier results. As hypothesized, cocaine use significantly increased the likelihood that a woman would become pregnant, and differences in pregnancy rates between HIV-infected women and non-infected women were most marked when drug use was controlled for in the matching process. When drug use was ignored, as in the first comparison, differences between groups tend to be minimized. Earlier studies that have attempted to measure the effect of drug use have had conflicting results: Thackway et al.  reported an increased pregnancy rate among drug-using HIV-positive women in Australia, but other studies have not shown significant differences between drug-using and non-drug using women [2,7,8,32]. Our study, however, reveals the importance of the type of drug being used, and also provides some understanding of the complexity of the association. Cocaine users were both less likely to have tubal ligations and also tended to have higher pregnancy rates. The latter fact is almost certainly contributed to by the increased sexual activity among women using cocaine and the exchange of sex for drugs. Even in studies of women being treated for opiate addiction [1,4], cocaine use could be an important confounding factor because cocaine use is probably more prevalent among those who are HIV infected .
The results of this study also revealed the effects of drug use on the frequency of obtaining tubal ligations. When drug use was ignored in the matching process, it appeared that tubal ligations occurred less frequently among HIV-infected women, but when drug use was included as a matching variable, it revealed that the apparent difference between groups was actually caused by drug use. The rate of tubal ligations was high in this study, and had a substantial effect on decreasing the duration of time in which women could become pregnant. Although differences in the use of contraceptives have been reported by other authors [3,12,34,35], the effect of drug use on the frequency of tubal ligations has not previously been emphasized.
The fact that tubal ligations did not appear to have an independent effect on decreasing the likelihood of a pregnancy appears to be counterintuitive, but is probably explained by the fact that tubal ligations in the follow-up period often occurred after a woman had already given birth to another child. In this study, we did not measure how many women decided to have a tubal ligation and then failed to return for the appointed operation. However, the very high rate of inadequate prenatal care among HIV-infected women and the association of this with an increased risk of pregnancy suggests that their poor use of healthcare interfered with their being able to follow through on decisions to limit the number of pregnancies.
Although the results of this study confirmed that HIV disease is associated with decreased fertility, it is surprising that those with a diagnosis of AIDS did not have lower pregnancy rates than those with less advanced disease. This may be explained, however, by the fact that the decrease in fertility occurs long before individuals have progressed to AIDS. In this study it was not possible to know how long women had been infected; however, Lee et al.  have documented decreased pregnancy rates for as long as 7.5 years before the diagnosis of AIDS.
Application of these results to other populations of HIV-infected women
In highlighting the effect that the use of cocaine has on pregnancy outcomes, this study helps to illustrate how differences in population groups might affect the conclusions of studies performed in different countries. The very high prevalence of cocaine use in this study is unlikely to be found elsewhere. In contrast, studies carried out in Europe include populations that are ethnically diverse, and women of African or Caribbean origins who acquired the infection heterosexually have continued to have higher pregnancy rates than women of European background, for whom intravenous drug use is more prevalent [9,36,37]. Studies in Europe have more consistently suggested that HIV-infected women have a high rate of therapeutic terminations of pregnancy [6,9,10,17,18], which was also true in a study performed in Australia  and a recent study reported from the United States . However, these findings contrast with the present study, in which rates of pregnancy terminations were not significantly different between groups. To date, studies performed in sub-Saharan Africa have shown decreased pregnancy rates among HIV-infected women, but in these countries factors such as nutritional status and the lack of antiretroviral therapy could have effects that are not seen in developed countries [12–16].
A potential limitation of this study is that it was conducted in a single city, and therefore the results may not be representative of HIV-infected women in the United States. For example, the rates of tubal ligation and cocaine use could be substantially different elsewhere. However, because the study demonstrated an independent effect of HIV disease on pregnancy rates, while also illustrating the confounding effects of these other variables, the results can probably be applied to other settings. Limiting subject enrolment only to those who had delivered a baby might have introduced a selection bias; women who knew of their HIV infection might already have chosen not to have children, and would thus not have been eligible for enrolment. However, the inclusion of such women in the study population would only have served to emphasize further the decrease in the pregnancy rate among HIV-infected women. This study was conducted by a retrospective review of medical records, thus we were unable to obtain reliable data on the use of contraceptives, and are therefore unable to comment on the extent to which contraceptive use might explain some of the findings of the study. Similarly, although there were no apparent differences between groups in the rate of spontaneous abortions, very early miscarriages were possibly not documented, and therefore would not have been included in the results.
This study concurs with the results of other studies in documenting the effects of HIV disease on decreasing pregnancy rates, but also highlights the fact that differences between countries in risk groups for HIV infection, and between population groups within a country, might have important contributing effects to pregnancy outcomes. Although the methods used in the study might explain why the results contradict earlier studies performed in the United States, it is also possible that with the passage of time and changes in the epidemic, HIV-infected women were making different choices than earlier in the epidemic and were less frequently choosing to become pregnant. Now, with the expanded use of antiretroviral medications for decreasing mother-to-child HIV transmission, there may be a swing in the opposite direction, and HIV-infected women may now more frequently be choosing to have children .
The potential effect of drug use on decreasing a woman's likelihood of having a tubal ligation is concerning and something that can potentially be addressed. Certainly, access to effective substance abuse treatment could have an important effect. In addition, although recognizing the importance of promoting condom use for decreasing HIV transmission, efforts should be increased to help drug-using women who do not wish to have more children have access to tubal ligations and follow through with the procedure.
1. Selwyn PA, Schoenbaum EE, Davenny K. et al
. Prospective study of human immunodeficiency virus infection and pregnancy outcomes in intravenous drug users. JAMA 1989, 261: 1289–1294.
2. Sunderland A, Minkoff HL, Handte J, Moroso G, Landesman S. The impact of human immunodeficiency virus serostatus on reproductive decisions of women. Obstet Gynecol 1992, 79: 1027–1031.
3. Lindsay MK, Grant J, Peterson HB, Willis S, Nelson P, Klein L. The impact of knowledge of human immunodeficiency virus serostatus on contraceptive choice and repeat pregnancy. Obstet Gynecol 1995, 85: 675–679.
4. Selwyn PA, Carter RJ, Schoenbaum EE, Robertson VJ, Klein RS, Rogers MF. Knowledge HIV antibody status and decisions to continue or terminate pregnancy among intravenous drug users. JAMA 1989, 261: 3567–3571.
5. Barbacci M, Chaisson R, Anderson J, Horn J. Knowledge of HIV serostatus and pregnancy decisions.
In:5th International Conference on AIDS
. Montreal, June 1989 [Abstract MBP10].
6. Johnstone FD, Brettle RP, MacCallum LR, Mok J, Peutherer JF, Burns S. Women's knowledge of their HIV antibody state: its effect on their decision whether to continue the pregnancy. BMJ 1990, 300: 23–24.
7. Chu SY, Hanson DL, Jones Jl and the Adult/Adolescent HIV Spectrum of Disease Project Group. Pregnancy rates among women infected with human immunodeficiency virus. Obstet Gynecol 1996, 87: 195–198.
8. Lee LM, Wortley PM, Fleming PL, Eldred LJ, Gray RH. Duration of human immunodeficiency virus infection and likelihood of giving birth in a medicaid population in Maryland. Am J Epidemiol 2000, 151: 1020–1028.
9. Stephenson JM, Griffioen A, and the Study Group for the Medical Research Council Collaborative Study of Women with HIV. The effect of HIV diagnosis on reproductive experience. AIDS 1996, 10: 1683–1687.
10. Van Bethem BHB, DeVincenzi I, Delmas M-C, Larsen C, Van den Hoek A, Prins M, and The European Study on The Natural History of HIV Infection in Women. Pregnancies before and after HIV diagnosis in a European cohort of HIV-infected women. AIDS 2000, 14: 2171–2178.
11. Thackway SV, Furner V, Mijch A. et al
. Fertility and reproductive choice in women with HIV-1 infection. AIDS 1997, 11: 663–667.
12. Allen S, Serufilira A, Gruber V. et al
. Pregnancy and contraception use among urban Rwandan women after HIV testing and counseling. Am J Public Health 1993, 83: 705–710.
13. Desgrées du Loû A, Msellati P, Yao A. et al
. Impaired fertility in HIV-1-infected pregnant women: a clinic-based survey in Abidjan, Côte d'Ivoire, 1997. AIDS 1999, 13: 517–521.
14. Gray RH, Wawer MJ, Serwadda D. et al
. Population-based study of fertility in women with HIV-1 infection in Uganda. Lancet 1998, 351: 98–103.
15. Ross A, Morgan D, Lubega R, Carpenter LM, Mayanja B, Whitworth JAG. Reduced fertility associated with HIV: the contribution of pre-existing sub-fertility. AIDS 1999, 13: 2133–2141.
16. Glynn JR, Buvé A, Caraël M. et al
. Decreased fertility among HIV-1-infected women attending antenatal clinics in three African cities. J Acquir Immune Defic Syndr 2000, 25: 345–352.
17. Nicoll A, Stephenson J, Griffioen A, Cliffe S, Rogers P, Boisson E. The relationship of HIV prevalence in pregnant women to that in women of reproductive age: a validated method of adjustment. AIDS 1998, 12: 1861–1867.
18. Shaw L, Goldberg DJ, Scrimgeaer JB, Maginnis M, et al. Prevalence of HIV among high and low risk pregnant women in Edinburg 1993–97.
In:12th World AIDS Conference
. Geneva, 28 June–3 July 1998 [Abstract 233469].
19. Gawin FH, Ellinwood Jr EH. Cocaine and other stimulants. N Engl J Med 1988, 318: 1173–1182.
20. Edlin BR, Irwin KL, Faruque S. et al
. Intersecting epidemics – crack cocaine use and HIV infection among inner-city young adults. N Engl J Med 1994, 331: 1422–1427.
21. Fullilove RE, Thompson Fullilove M, Bowser BP, Gross SA. Risk of sexually transmitted disease among black adolescent crack users in Oakland and San Francisco, California. JAMA 1990, 263: 851–855.
22. Chiasson MA, Stoneburner RL, Hildebrandt DS, Ewing WE, Telzak EE, Jaffe HW. Heterosexual transmission of HIV-1 associated with the use of smokable freebase cocaine (crack). AIDS 1991, 5: 1121–1126.
23. Cohen, E, Navaline H, Metzger D. High-risk behaviors for HIV: a comparison between crack-abusing and opioid-abusing African-American women. J Psychoactive Drugs 1994, 26: 233–241.
24. Ostrea Jr EM, Brady M, Gause S, Raymundo AL, Stevens M. Drug screening of newborns of meconium analysis: a large-scale, prospective, epidemiologic study. Pediatrics 1992, 89: 107–113.
25. Frank DA, Zuckerman BS, Amaro H. et al
. Cocaine use during pregnancy: prevalence and correlates. Pediatrics 1988, 82: 888–895.
26. Leventhal JM, Forsyth BWC, Qi K, Johnson L, Schroeder D, Votto N. Maltreatment of children born to women who used cocaine during pregnancy: a population-based study. Pediatrics 1997, 100: 2–7.
27. Forsyth BWC, Leventhal JM, Qi K, Johnson L, Schroeder D, Votto N. Health care and hospitalizations of young children born to cocaine-using women. Arch Pediatr Adolesc Med 1998, 152: 177–184.
28. Stanford PA, Forsyth BW, Qi K. Factors affecting length of hospital stay of opiate-exposed neonates.
In:37th Meeting of the Ambulatory Pediatric Association
. Washington, DC, May 1997 [Abstract APA 53].
29. Breslow NE, Day NE. Statistical methods in cancer research
. Lyon: International Agency for Research on Cancer; 1980. p. 165165.
30. Breslow NE, Day NE. Statistical methods in cancer research
. Lyon: International Agency for Research on Cancer; 1980. pp. 202–205.
31. Zelterman D. Models for discrete data
, chapter 4. Oxford: Clarendon Press; 1999.
32. Kline A, Strickler J, Kempf J. Factors associated with pregnancy and pregnancy resolution in HIV seropositive women. Soc Sci Med 1995, 40: 1539–1547.
33. Bux DA, Lamb RJ, Iguchi MY. Cocaine use and risk behavior in methadone maintenance patients. Drug Alcohol Depend 1994, 37: 29–35.
34. Diaz T, Schable B, Chu SY, and the Supplement to HIV and AIDS Surveillance Project Group. Relationship between use of condoms and other forms of contraception among human immunodeficiency virus-infected women. Obstet Gynecol 1995, 86: 277–282.
35. Williams HA, Watkins CE, Risby JA. Reproductive decision-making and determinants of contraceptive use in HIV-infected women. Obstet Gynecol 1995, 39: 333–343.
36. Delmas MC, Anzens B, Pena JM, et al. Incidence and outcome of pregnancies in HIV-infected women.
In:12th World AIDS Conference
. Geneva, June/July 1998 [Abstract 24194].
37. Meyer L, Fourquet F, Le Chenadel J, Nayaus MJ. Incidence of pregnancies in HIV-infected women between 1988 and 1996.
In:12th World AIDS Conference
. Geneva, June/July 1998 [Abstract 24200].
38. Hankins C, Tran T, Lapointe N, Hum L, Samson J, and the Canadian Women's HIV Study Group. Is antiretroviral MCT prophylaxis provoking increased pregnancy incidence in women living with HIV.
In:12th World AIDS Conference
. Geneva, June/July 1998 [Abstract 24199].
This article has been cited 9 time(s).
British Journal of CancerRisk of breast, ovary, and uterine corpus cancers among 85 268 women with AIDSBritish Journal of Cancer
AIDS Patient Care and StdsDesire to have children: Gender and reproductive rights of men and women living with HIV: A challenge to health care in BrazilAIDS Patient Care and Stds
Analyzing multiply matched cohort studies with two different comparison groups: Application to pregnancy rates among HIV plus women
Hospitalizations of pregnant HIV-infected women in the USA prior to and during the era of HAART, 1994-2003
Ciencia & Saude Coletiva
Induced abortion in women of reproductive age living with and without HIV/Aids in Brazil
Ciencia & Saude Coletiva, 14(4):
International Journal of Std & AIDS
Reduced fertility among HIV-infected women associated with viral load in Rakai district, Uganda
International Journal of Std & AIDS, 17():
Journal of Womens HealthHospitalizations of pregnant HIV-infected women in the United States in the era of highly active antiretroviral therapy (HAART)Journal of Womens Health
Obstetrics & GynecologyTrends in Pregnancy Rates Among Women With Human Immunodeficiency VirusObstetrics & Gynecology
Births; cocaine; HIV-infected women; pregnancies; tubal ligations
© 2002 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.