Incidence of Childbirth, Pregnancy, Spontaneous Abortion, and Induced Abortion Among Women With Human Immunodeficiency Virus in a Nationwide Matched Cohort Study

Abstract

Background

Reproductive health in women with human immunodeficiency virus (HIV) (WWH) has improved in recent decades. We aimed to investigate incidences of childbirth, pregnancy, spontaneous abortion, and induced abortion among WWH in a nationwide, population-based, matched cohort study.

Methods

We included all WWH aged 20–40 years treated at an HIV healthcare center in Denmark from 1995 to 2021 and a matched comparison cohort of women from the general population (WGP). We calculated incidence rates per 1000 person-years and used Poisson regression to calculate adjusted incidence rate ratios (aIRRs) of childbirth, pregnancy, spontaneous abortion, and induced abortion stratified according to calendar periods (1995–2001, 2002–2008, and 2009–2021).

Results

We included 1288 WWH and 12 880 WGP; 46% of WWH were of African origin, compared with 1% of WGP. Compared with WGP, WWH had a decreased incidence of childbirth (aIRR, 0.6 [95% confidence interval, .6–.7]), no difference in the incidence of pregnancy (0.9 [.8–1.0]) or spontaneous abortion (0.9 [.8–1.0]), but an increased incidence of induced abortion (1.9 [1.6–2.1]) from 1995 to 2021. The aIRRs for childbirth, pregnancy, and spontaneous abortion increased from 1995–2000 to 2009–2021, while the aIRR for induced abortion remained increased across all time periods for WWH.

Conclusions

From 1995 to 2008, the incidences of childbirth, pregnancy, and spontaneous abortion were decreased among WWH compared with WGP. From 2009 to 2021, the incidence of childbirth, pregnancy, and spontaneous abortion no longer differed among WWH compared with WGP. The incidence of induced abortions remains increased compared with WGP.

General health among women with human immunodeficiency virus (HIV; WWH) has improved significantly since the introduction of combination antiretroviral therapy (cART). In addition, mother-to-child-transmission (MTCT) of HIV has been reduced to <0.5% among WWH with suppressed viral load throughout pregnancy [1, 2]. These factors have led to changes in clinical guidelines for pregnancies and childbirths among WWH. In most high-income countries, termination of pregnancies in WWH was no longer recommended after 2000, and between 2002 and 2008 vaginal delivery replaced caesarean delivery as the recommended mode of delivery in WWH with undetectable viral loads and no obstetric contraindications [2–6]. In addition, since 2008, condomless sex in serodiscordant couples has been considered safe owing to the low risk of HIV transmission from people with HIV with suppressed viral load [7].

As a result of these advancements, previous studies have reported increases in both pregnancies and childbirths among WWH [2, 8–10]. In agreement, the majority of WWH reportedly intend to have children and view motherhood as an important aspect of their lives [11–13]. Conversely, studies have shown that 50% of pregnancies among WWH were unintended, and the use of contraceptives is low [14–17].

Previous studies have investigated the incidences of childbirth, pregnancy, spontaneous abortion, and induced abortion among WWH [2, 8–10, 18–20]. However, no previous study has compared these incidences among WWH with a matched cohort of women from the general population (WGP). Some previous studies report increased birth rates among WWH [2, 8–10], while others found decreased birth rates [20]. Furthermore, previous studies differed on whether the rates of spontaneous abortions and induced abortions changed [8, 9, 18, 19, 21] after cART introduction.

There are approximately 1700 WWH in Denmark, of whom 80% are of childbearing age. The majority of WWH in Denmark are immigrants, primarily from sub-Saharan Africa, and the primary route of HIV transmission is by sexual contact [22]. We aimed to investigate the incidences of childbirth, pregnancy, spontaneous abortion, and induced abortion among WWH compared with an age-matched comparison cohort in different time periods (1995–2001, 2002–2008, and 2009–2021).

METHODS

We performed a nationwide, matched population-based cohort study to investigate the incidence rates of childbirth, pregnancy, spontaneous abortion, and induced abortion among WWH, compared with a comparison cohort of WGP.

Setting

The population of Denmark ranged from 5.2 to 5.8 million in the study period (1995–2021), and the estimated prevalence of HIV infections among adults ranged from 0.07% to 0.1% [21, 23]. Universal healthcare is tax funded and free for all Danish residents [23].

HIV treatment in Denmark is provided at 10 specialized centers. People with HIV have outpatient contacts every 12–24 weeks. The care of WWH who become pregnant is centralized to 4 specialized centers, with assessment by a multidisciplinary team comprising of infectious disease specialist and an obstetrician, according to guidelines [5, 24]. Furthermore, care of infants born to WWH is continued through ≥18 months of age to clarify infant HIV status [25].

Danish guidelines on birth and pregnancy in WWH have changed 3 times during the study period. Abortion was recommended before 2000, owing to the risk of MTCT [2], and caesarean delivery was recommended until 2007. Since 2007, WWH compliant with cART and with HIV-RNA levels <400 copies/mL may deliver vaginally, and natural conception has been regarded as safe in WWH virally suppressed by cART since 2020 [2, 5, 26].

An individual unique 10-digit personal identification number is assigned to all Danish residents at birth or on immigration. We used this number to link data from the Danish HIV Cohort Study and the different Danish National Registries at an individual level [27].

Data Sources

We extracted data on WWH, including dates of HIV diagnosis and routes of transmission, from the Danish HIV Cohort Study, an ongoing nationwide, population-based cohort study of all people with HIV >16 years old, who received HIV treatment at a Danish clinic specialized in HIV after January 1995 [28]. We extracted data on date of birth, sex, date of immigration or emigration, country of origin, dates of liveborn childbirth, and date of death from the Danish Civil Registration System, established in 1967 [27].

We extracted data on diagnoses and surgical procedures from the Danish National Patient Registration System (DNPR). The DNPR contains data on all patients discharged from nonpsychiatric hospitals since 1977 and all patients with outpatient or emergency hospital contact since 1994 [29]. Data on dates and types of diagnosis of pregnancy, spontaneous abortion (loss of pregnancy at or before 22 completed weeks of pregnancy), and induced abortion were coded by the attending physician according to the International Classification of Diseases (ICD; ICD-8 codes until the end of 1993 and ICD-10 codes thereafter) [30]. Data on date and type of induced surgical abortion were coded according to the Nordic Medico-Statistical Committee Classification of Surgical Procedures (NOMESCO) by the operating surgeon [31].

Study Population

We formed a patient cohort of all WWH from the Danish HIV Cohort Study, who were residents of Denmark and were treated at a HIV specialized clinic between 1 January 1995, and 1 September 2021. The date of study inclusion was defined as date of a first-time registered diagnosis of HIV (ICD-10 code, B20-B25) in the DNPR, the date of the first viral load measurement, and date of the patient’s 20th birthday or 1 January 1995, whichever came later. We excluded all WWH who had a hysterectomy or tubal ligation or who were >40 years old before study inclusion. We included women aged 20–40 years, as few childbirths were observed among WWH outside this age range in preliminary analyses.

We randomly extracted 10 WGP without HIV infection or hysterectomy or tubal ligation from the Danish Civil Registration System for each member of the WWH cohort, matched on date of birth to form a comparison cohort. Members of the comparison cohort were assigned the same day of study inclusion as the corresponding member of the WWH cohort.

Outcome Measures

We investigated all childbirths, pregnancies, spontaneous abortions, and induced abortions registered in the DNPR. Induced abortion was defined as a diagnosis of induced abortion or surgical procedure of induced abortion (Supplementary Tables 1 and 2).

Statistical Analyses

We used a cohort design and data from Danish registries to calculate the time until an event of interest, date of emigration, 40th birthday, loss to follow-up, or date of death. We calculated incidence rates of outcomes per 1000 person-years (absolute risk) among WWH and WGP. We used Poisson regression to calculate the incidence rate ratios and 95% confidence intervals (CIs) (relative risk) of outcomes among WWH compared with WGP. Our analyses were stratified according to calendar periods (1995–2001, 2002–2008, and 2009–2021), which corresponded to changes in Danish reproduction guidelines for WWH [2, 5, 26]. All analyses were adjusted for age and prior births before study inclusion to calculate adjusted incidence rate ratios (aIRRs).

We performed 2 sensitivity analyses in which we excluded (1) WWH who reported intravenous drug use as route of HIV transmission and the corresponding WGP and (2) WWH who were born in Africa and the corresponding WGP. We used SPSS software, version 28, and R software, version 4.1.2., for all analyses.

Ethical Approval

All patient information was treated confidentially and with anonymous, coded identification. The Danish HIV Cohort Study is approved by the Danish Data Protection Agency.

RESULTS

We included 1288 WWH and 12 880 WGP. The median age at study inclusion was 30 years (Table 1). Among WWH, 46% of WWH were born in Africa, while the majority of WGP were born in Denmark (88%) (Table 1). An increase in WWH of African origin was seen throughout the study period (Table 2). The primary route of HIV transmission among WWH was heterosexual (75%), followed by intravenous drug use (12%) and other or unknown (13%) (Table 1). Changes in population characteristic over time periods are presented in Table 2.

Childbirth

During the entire study period (1995–2021) we observed a decreased aIRR for childbirth among WWH compared with WGP (0.6 [95% CI: .6–.7]) (Table 3). The incidence of childbirth was decreased among WWH compared with WGP in 1995–2001 (aIRR, 0.3 [95% CI: .2–.4]) and 2002–2008 (0.5 [.4–.6]) but not in 2009–2021 (0.9 [.8–1.0]) (Table 3).

Pregnancy

During the entire study period (1995–2021) we observed an aIRR of 0.9 (95% CI: .8–1.0) for pregnancy among WWH compared with WGP (Table 3). The incidence of pregnancies was decreased among WWH compared with WGP in 1995–2001 (aIRR, 0.7 [95% CI: .5–.9]) but not thereafter (Table 3).

Spontaneous Abortion

During the entire study period (1995–2021) we observed an aIRR of 0.9 (95% CI: .8–1.0) for spontaneous abortions among WWH compared with WGP (Table 3). The incidence of spontaneous abortions was decreased among WWH compared with WGP in 1995–2001 (aIRR, 0.5 [95% CI: .4–.6]) and 2002–2008 (0.8 [.7–.9]) but was slightly increased in 2009–2021 (1.2 [1.1–1.3]) (Table 3).

Induced Abortion

During the entire study period—1995–2021—we observed an aIRR of 1.9 (95% CI: 1.6–2.1) for induced abortions among WWH compared with WGP. The incidences of induced abortions were higher among WWH than among WGP in all 3 time periods (Table 3). We also investigated the incidence rate of >1 induced abortion among WWH compared with WGP. This analysis showed that the risk of multiple induced abortions is 3 times higher among WWH compared with WGP across all time periods (Supplementary Table 3).

Sensitivity Analyses

In the sensitivity analysis excluding WWH whose HIV was transmitted by intravenous drug use and the corresponding WGP, the aIRR was slightly increased for childbirths, pregnancies, and spontaneous abortions and significantly increased for induced abortions among WWH compared with WGP during the entire study period (Supplementary Table 4). In the sensitivity analysis excluding WWH of African origin and the corresponding WGP, we found a decreased aIRR for childbirths but no difference in the aIRR for pregnancies among non-African WWH compared with WGP during the entire study period. The aIRR for spontaneous abortions was slightly decreased among non-African WWH compared with WGP during all 3 time periods, and the aIRR for induced abortions was increased during the entire study period (Table 4). Finally, the results of the unadjusted and adjusted Poisson regression analyses were essentially the same (Table 4 and Supplementary Table 4).

DISCUSSION

In this Danish, nationwide, population-based, matched cohort study, we demonstrated decreased incidences of childbirth, pregnancy, and spontaneous abortion among WWH compared with WGP between 1995 and 2008. Between 2008 and 2021 WWH had reached incidences of childbirths and pregnancies similar to those in WGP, while the incidence of spontaneous abortions was slightly increased among WWH. The incidence of induced abortions remained increased relative to WGP from 1995 to 2021.

We found a decreased incidence of childbirth from 1995 to 2008 but not thereafter, similar to findings from the United States [9, 10]. Before the introduction of cART, studies from the United Kingdom and Ireland reported an MTCT rate of 20% among WWH, decreasing to <1% between 2000 and 2006 among WWH on cART with viral suppression [32, 33]. General improvements in health, as well as better understanding of the reduced risk of MTCT, could be major factors contributing to the increase in childbirth rates among WWH. In addition, specialized support for WWH with fertility wishes has been established in Denmark [3]. The finding of a lower incidence of childbirth among non-African WWH compared with WGP could indicate a need for further clinical reproductive guidance for WWH. Because we did not investigate the intent to have children, this possibility remains speculative.

We observed an increase in pregnancy rates among WWH from the first to the last study period, consistent with the childbirth rates. Previous studies have also reported increases in pregnancy rates among WWH [9, 10]. A Danish study found a 5-fold increase in the total number of pregnancies among WWH who gave birth from 1995 to 2007 [2] but did not examine the incidence rates of pregnancies among WWH. In contrast, a Swiss cohort study did not report an increase in pregnancy rates among Swiss WWH from 2005 to 2019, although it should be noted that the cohort included a lower proportion of WWH of childbearing age [8].

Our study revealed that the incidence of spontaneous abortion among WWH increased slightly from 1995–2001 to 2009–2021. This in contrast to a study from the United States, which reported that abortion rates among WWH did not change from 1994 to 2012 [9]. However, that study did not perform separate analyses for spontaneous abortion and induced abortion, did not include a matched comparison cohort, and relied on self-reported outcomes, which may induce a misclassification bias [9]. In line with our findings, the Swiss cohort study reported an increase in the frequency of spontaneous abortion among WWH, from 12% of pregnancies before 2008 to 17% thereafter [8]. Factors associated with higher rates of spontaneous abortion were detectable viral load, age, and alcohol consumption. The higher maternal age among Swiss WWH was suggested as the most likely reason for the increased risk of spontaneous abortions [8].

We found that the rate of induced abortion was increased among WWH compared with WGP. The incidence of induced abortions decreased slightly from 1995 to 2021, although the CIs were broad. In agreement, previous studies reported a decrease in induced abortions among WWH since the introduction of cART [9, 18, 21, 34]. However, the incidences of induced abortion remained increased among WWH comapared with WGP during all time periods, in accordance with findings from the Swiss Cohort study [8]. In contrast to our findings, the Swiss Cohort Study did not find a decrease in the risk of induced abortion among Swiss WWH after 2005 [8]. Potential explanations for the increased risk of induced abortion among WWH include a higher rate of unintended pregnancies. A systemic review and meta-analysis found a pooled magnitude of 55.9% of unintended pregnancies among WWH [17], and a study from the United States found a similar trend [15]. This may be due to incorrect use of contraceptives among WWH, possibly resulting from drug-drug interactions between certain cART regimens and hormonal contraceptives, reducing protection against pregnancy [35, 36].

In addition, a Swiss questionnaire found that 36% of sexually active WWH did not use contraceptives at all and more than half used oral hormonal contraceptives, which could be inhibited by their cART [16]. A study from the United States reported that although most WWH used contraception, condoms were the preferred choice, and the use of hormonal contraceptives was lower among WWH than among WGP [35]. Improved guidance about contraceptive use for WWH and increased awareness of the interaction between cART and hormonal contraceptives may help reduce the risk of unintended pregnancies and induced abortions among WWH.

In the sensitivity analysis, which excluded African WWH and corresponding WGP, the incidence of childbirth among non-African WWH compared with WGP was decreased in all time periods. Similarly, a Canadian study found that African WWH had a higher birth rate compared with WGP [37]. Conversely, when we excluded WWH of African origin, the increased incidence of spontaneous abortion from 2009 to 2021 disappeared. This indicates that the increased incidence was limited to WWH of African origin.

In a population-based Swedish register study, the risk of stillbirth was increased in women of African and Middle East origin [38]. This may also account for spontaneous abortion. As we found an increase of WWH of African origin throughout the study time (Table 2), the increased incidence of spontaneous abortion from 2009 to 2021 could possibly be explained by an increased risk of spontaneous abortion among African women. Furthermore, immigrants may face several difficulties accessing healthcare systems, including low socioeconomic status, language barriers, and other stressors related to immigration [39]. Notably, women from Africa are known to have lower use of any contraceptive treatment than women from other continents [40].

Finally, differences in cultural and religious practices may influence choices with regard to childbirth, contraception, and induced abortions. The fact that the incidence of induced abortion among WWH is increased compared with WGP highlights the continued need for the clinicians who treat WWH to focus on guidance with regard to contraceptives.

The major strengths of the current study are the study size and the use of an age-matched comparison cohort formed of WGP. Furthermore, the use of nationwide data from the Danish national registries ensured high-quality data on an individual level with a minor risk of incorrect registrations.

The current study does have some limitations. Ethnicity is a potential confounder; almost half of WWH were of African origin, compared with 1% of WGP. This prevented us from performing analyses including only women of African origin. Spontaneous abortions may be underreported in the DNPR, which relies on hospital contacts. Not all women who have spontaneous abortions will be examined in a hospital and correctly categorized in the DNPR. Spontaneous abortions among WWH may be more likely to be recorded owing to regular follow-up in HIV clinics.

As another limitation, we did not determine the ratio of spontaneous abortions to pregnancies among WWH, which would have been a relevant outcome. Pregnancy is a prerequisite for an abortion, and because outcomes were based on assigned hospital diagnoses, both pregnancies and spontaneous abortions are likely underreported, particularly during the first 3 months of pregnancy. Therefore, we did not calculate the ratio of abortions to pregnancies owing to the risks of under- or overreporting.

Finally, other studies found age as a potential factor increasing rates of spontaneous abortions. Because our study included an age-matched comparison cohort, it is unlikely that any difference in spontaneous abortion rates could be explained by age-related factors.

In conclusion, from 1995 to 2001, the incidences of childbirth, pregnancy, and spontaneous abortions were decreased among WWH compared with WGP in Denmark. However, these incidences increased among WWH during the following study periods of 2002–2008 and 2009–2021. By 2009–2021, WWH had reached similar incidences of childbirth and pregnancy, while the incidence of spontaneous abortion was slightly increased compared with that in WGP. Throughout the entire study period, WWH had an almost 2-fold increased incidence of induced abortions and a 3-fold increased incidence of multiple induced abortions compared with WGP. This emphasizes the continued need for clinicians who treat WWH to focus on contraceptive guidance.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Author contributions. All authors were responsible for the concept. Statistical analyses: F. W. P., M. M. T, C. V. V., and L. H. O. Data collection: J. G., G. K., I. S. J., C. S. L., L. W., M. D. P., S. L., L. N. N., and N. O. Drafting of manuscript: F. W. P. Editing of manuscript: M. M. T., C. V. V., J. G., G. K., I. S. J., C. S. L., L. W., M. D. P., S. L., L. N. N., N. W., N. O., L. H. O., and A. M. L.

Acknowledgments. We are grateful to the staff at our clinical departments for their continuous support and enthusiasm.

Financial support. This work was supported by Gilead 2022 (Nordic Fellowship [unrestricted research grant]) and the Preben og Anne Simonsens Fond (unrestricted research grant).

Data availability. The ethical approval of this study by the Danish Data Protection Agency states that the data used in this article cannot be shared publicly. On reasonable request to the corresponding author, the data can be shared and accessed at our institution.

References

Author notes