Association between history of fertility problems and pregnancy and birth complications: A longitudinal population-based cohort study

JAG conceived the idea, and JAG, DS, and AMH contributed to the design of the study. DS analyzed the data. CAT interpreted the results with assistance from JAG, AMH, LP‐I and DS. CAT drafted the manuscript, and JAG, DS, AMH, and LP‐I critically revised it. All authors approved the final manuscript.

ALSWH has ongoing ethical approval from the Human Research Ethics Committees of the Universities of Newcastle and Queensland (approval numbers H‐076‐0795, granted 26 July 1995, and 2 004 000 224, granted April 2004, respectively). All participants provided written consent to participate and have been able to withdraw or suspend their participation at any time with no need to provide a reason.

No specific funding was received for this work. DS is supported by the National Institute for Health and Care Research (NIHR) through an NIHR Advanced Fellowship (NIHR302955) and the NIHR Southampton Biomedical Research Centre (NIHR203319). CT is supported by a Monash IVF PhD stipend.

The baseline characteristics of the 5653 women in the 1973–79 cohort of the ALSWH who reported a live birth are shown in Table  1 . The mean (SD) age and BMI were 30 (3.2) years and 24.7 (5.0) kg/m 2 , respectively. 897 women (15.9%) reported a history of fertility problems for which the majority had sought help (71.6%). Most of the women were born in Australia (93.5%). While 84.2% of the women lived in urban areas, a higher proportion of women living in rural or remote areas who had fertility problems did not seek help. Baseline characteristics of women in the 1973–79 cohort of the ALSWH who reported a live birth between 2003 and 2018, n  = 5653. Abbreviations: ARFS, Australian Recommended Food Score; MET, metabolic equivalent of task. p ‐Values from Chi‐square test or ANOVA. Overall, 63.1% of the women had a BMI 30 kg/m 2 , while 19.1% with a BMI >30 kg/m 2 did seek help. Among those without a history of fertility problems, 12.0% of women had a BMI >30 kg/m 2 . Endometriosis was reported by 8.4% of the women who had a live birth, making up 6% of the women who had no problems with fertility and 37.8% of those who did have problems with fertility. Likewise, 8.3% of the cohort reported a history of PCOS, with this group comprising 5.2% of the group with no problems conceiving and 42.8% of those with fertility problems. A higher percentage of women who sought help with fertility problems had known associated factors, including endometriosis (24%), PCOS (28.7%), and irregular menstrual periods (26.5%). The mean ARFS scores and levels of physical activity were similar across women who did and did not seek help for fertility problems. Preterm birth was the most widely reported pregnancy complication, experienced by 10.8% of women in 7.2% of the pregnancies ( n  = 705 cases: 9821 pregnancies). Low birthweight was reported by 7.7% of the women in 4.8% of pregnancies ( n  = 480 cases; 10 061 pregnancies). 8.7% of women experienced GDM, in 5.4% of pregnancies ( n  = 522 cases; 9656 pregnancies) while 8.4% of women were affected by HDP in 5.2% of pregnancies ( n  = 479; 9152 pregnancies). In model 1, after adjusting for age, country of birth, area of residence, and highest educational qualification, women with a history of fertility problems had a higher risk of each of the pregnancy complications examined, including GDM [RR: 1.26; 95% CI (1.02 to 1.55)], HDP [RR: 1.26; 95% CI (1.00 to 1.60)], low birthweight [RR: 1.44; 95% CI (1.15 to 1.80)], and preterm birth [RR: 1.42; 95% CI (1.17 to 1.73)] (Table  2 ). However, after additional adjustments in model 2 (age at menarche, irregular menstrual periods, BMI, nulliparity, and multiple birth), the risks became non‐significant. The relative risks were attenuated further in model 3, after adjustment for PCOS and endometriosis: GDM [RR: 1.00; 95% CI (0.79, 1.18)], HDP [RR: 1.08; 95% CI (0.83 to 1.41)], low birthweight [RR: 1.10; 95% CI (0.86 to 1.40)], and preterm birth [RR: 1.04; 95% CI (0.84 to 1.28)], and remained non‐significant with adjustments in model 4 (physical activity, smoking status, total energy intake, ARFS score). Associations between history of fertility problems and pregnancy and birth complications. Note : Model 1: adjusted for age, country of birth, area of residence, and highest educational qualification. Model 2: adjusted for variables in Model 1 + age at menarche, irregular menstrual periods, body mass index, nulliparity, and multiple birth. Model 3: adjusted for variables in Model 2 + polycystic ovary syndrome and endometriosis. Model 4: adjusted for variables in Model 3 + physical activity, smoking status, total energy intake, and total Australian Recommended Food Score. Bold indicate p ≤ 0.05. Bold indicate p ≤ 0.01 .

In this study of women living in Australia and who had a live birth, we report that 15.9% had a history of fertility problems. Around 5% of women who had no issues with fertility reported a pregnancy complication and this percentage was slightly higher in those who did report problems conceiving. Unadjusted analyses revealed women with fertility problems had a higher risk of adverse pregnancy outcomes; however, after adjustment for a range of health and lifestyle variables, no statistically significant association was found between problems with fertility and adverse pregnancy outcomes, including GDM, HDP, preterm birth, and low birth weight. The major strength of this study is the large population‐based cohort in a representative sample of women living in Australia, with high retention rates of participants across the surveys. Participants were recruited using stratified random sampling from the national Medicare database, resulting in a cohort broadly representative of Australian women across urban, rural, and remote regions. Although exclusions and non‐response reduced the sample size for the current analysis, previous modeling has shown minimal bias in longitudinal estimates of associations between risk factors and health outcomes. 25 Additionally, the questions used to measure socio‐demographic, health, and pregnancy outcomes were repeated at each survey, providing confirmation of previous responses and allowing us to adjust for a range of time‐varying known confounders. It also includes all women from the cohort who reported a live birth; those who had no problems conceiving, those who took a long time to conceive but did so spontaneously, and those who used some form of treatment to conceive. This is noteworthy for this type of prospective study, which would typically concentrate on only one of these groups, and hence, this study provides a more representative picture of adverse pregnancy outcomes in the population. We also acknowledge that this introduces a limitation. Although we do have data on which women used ART, this information is not related to specific timeframes, so we are unable to relate the use of ART with specific pregnancy outcomes. All data from this survey were self‐reported, including health characteristics and pregnancy outcomes, which may introduce bias. However, a 2015 data linkage study found strong agreement between the mother's self‐reported perinatal data from the ALSWH and linked medical records, with reliability of >92–>95% for all outcomes (HDP, GDM, preterm birth, low birthweight). 26 We also acknowledge that some health data (PCOS, endometriosis, hypertension) may have been undiagnosed and hence underreported, which could influence the results, likely weakening the associations. Further, we acknowledge the recently expanded, more inclusive definition of infertility. 27 However, due to the survey design, we have used the WHO definition for the purpose of this study. Finally, this study cannot evaluate the influence of male infertility on pregnancy outcomes. As the ALSWH focuses on women's health and was not designed specifically as a reproductive health study, the survey questions do not allow us to fully examine the contribution from men. While we excluded women who reported any male factor contributing to their fertility problems (partner cannot have children, has had a vasectomy, has no sperm), the numbers excluded on this basis were lower than would be expected and potentially are not an accurate reflection of male factor infertility in the population. Finally, the prevalence of PCOS and endometriosis in this cohort was modest, limiting power to isolate their specific contribution to pregnancy outcomes. As such, we were unable to reliably model these conditions separately with adequate precision, and their role should be interpreted within the broader set of reproductive and metabolic factors included in multivariable adjustment. In this study, we did not find evidence of an independent association between subfertility and adverse pregnancy outcomes after full adjustment for underlying health and lifestyle characteristics. While initial models adjusting only for age and socio‐demographic factors suggested that women with a history of fertility problems were at 26%–44% higher risk of complications such as preterm birth and low birthweight, outcomes with significant long‐term health implications for both mother and baby, the associations became non‐significant in fully adjusted models. Notably, the risk appeared greater among women who sought fertility assistance, potentially reflecting a higher burden of underlying health conditions contributing to both infertility and adverse outcomes. These findings are consistent with a large Danish population‐based cohort study which reported a similar increased risk of preterm birth (38%) and low birthweight (44%) for women who took >12 months to conceive. 12 Other studies also report associations between subfertility and higher risk of preterm birth and low birthweight in both IVF 28 , 29 , 30 , 31 and spontaneous pregnancies. 12 , 32 However, most prior studies adjusted only for age and parity, whereas our inclusion of a broader range of factors, such as BMI, PCOS, endometriosis, smoking, and maternal nutrition, enhances the validity of our findings and suggests that these underlying factors, rather than subfertility itself, may be driving the observed associations. This underscores the importance of comprehensive preconception and antenatal care that addresses modifiable health and lifestyle factors, particularly for women with a history of fertility problems, and supports the clinical relevance of identifying and managing these risks early to improve pregnancy outcomes. Higher risk for GDM was found only in model 1, with the greatest risk observed among those women who had not sought help with their infertility. However, this association did not persist after adjustment for underlying reproductive and metabolic health factors, suggesting that infertility may not be an independent risk factor for GDM. These findings are in contrast to previous studies which described significant associations between infertility and GDM. For example, women with a history of infertility who participated in the Nurses' Health Study II had a 39% overall greater risk of GDM, 33 with a higher risk found in sub‐groups of women with ovulation disorders such as PCOS (52%) and tubal blockages (83%). 33 This highlights the contribution of specific reproductive pathologies to metabolic risk in diabetes. Additionally, a systematic review found that women who conceived via ART had a 53% higher risk of GDM compared with those who had a spontaneous conception. 11 Importantly, this review assessed women with PCOS separately, further highlighting the influence of underlying pathologies for infertility and GDM. For HDP, the association was weak and limited to model 1. This contrasts with previous studies describing strong associations between time to pregnancy and risk for preeclampsia, one type of HDP. 10 , 30 , 32 These studies adjusted for similar confounders, including age and parity, but not PCOS or endometriosis, which may partly explain the difference in findings. Underlying metabolic disturbances, such as inflammation, hyperinsulinemia, and insulin resistance, have been implicated in several conditions related to infertility, including PCOS, endometriosis, and obesity. 34 They are also associated with altered endocrine profiles and even changes in the gut microbiome. 35 Chronic low‐grade inflammation present in these conditions is also associated with GDM, HDP, and small and large for gestational age, 36 suggesting a possible shared etiology, but the mechanism is not yet fully understood. The interpretation of these findings warrants caution. In large observational studies with multiple outcomes and repeated measures, statistically significant associations may arise due to multiplicity or small effect sizes rather than reflecting clinically meaningful differences. This is particularly relevant for the associations observed in the minimally adjusted models, which were attenuated after comprehensive adjustment for reproductive, metabolic, and lifestyle factors. The absence of statistically significant associations in fully adjusted models should therefore be interpreted as a lack of evidence for an independent effect of fertility problems, rather than definitive evidence of no effect. Overall, the pattern of attenuation across models suggests that associations previously reported in other studies may potentially reflect shared underlying maternal health characteristics rather than infertility itself. Taken together, these findings support the interpretation that adverse pregnancy outcomes are more closely related to the underlying causes of infertility than to infertility itself. In our cohort, women with a history of fertility problems had a higher prevalence of obesity, irregular periods, PCOS, or endometriosis, factors that are independently associated with both subfertility and adverse pregnancy outcomes. This concurs with Luke et al., who found that the underlying pathology contributing to infertility is an important factor in pregnancy and birth outcomes, and supports the recommendation to change the research focus from infertility per se as a risk factor for adverse pregnancy and birth outcomes to further understanding the contribution of the underlying infertility diagnosis. 16 Further research is needed to clarify the relative contribution of individual factor/s (e.g., BMI, endometriosis, PCOS) and to elucidate the mechanisms linking these conditions to adverse pregnancy outcomes. Addressing modifiable risk factors in the preconception period is an essential approach to reducing the risk of pregnancy complications and optimizing maternal and neonatal health.

In this large population‐based cohort of women living in Australia, associations between a history of fertility problems and adverse pregnancy outcomes observed in minimally adjusted models were attenuated after comprehensive adjustment for key maternal health and lifestyle factors. The absence of statistically significant associations in fully adjusted analyses suggests that previously reported risks are more likely to reflect shared underlying maternal health characteristics rather than infertility itself. While these findings may offer reassurance to women (and couples) experiencing fertility difficulties that a history of infertility does not appear to independently confer a higher risk of GDM, HDP, preterm birth, or low birthweight once pregnant, they should be interpreted cautiously in the context of the study's observational design, large sample size, and multiple comparisons. Overall, these results highlight the importance of focusing clinical care and public health strategies on identifying and managing modifiable underlying health and lifestyle factors to optimize maternal and neonatal outcomes.

Infertility is a global problem, affecting 1 in 6 couples trying to conceive. 1 The World Health Organization defines infertility as a disease of the male or female reproductive system defined by the failure to achieve pregnancy after twelve months or more of regular unprotected sexual intercourse. 1 The underlying reasons for female infertility are diverse and include maternal age, 2 polycystic ovary syndrome (PCOS), 3 endometriosis, 4 and obesity, 5 along with lifestyle factors such as diet, 6 , 7 physical activity, 8 and smoking. 9 There is good evidence linking a history of fertility problems with increased risks of pregnancy and birth complications. Women who experience a longer time to conception are more likely to develop pregnancy complications than those who conceive easily. One study reported a 50% higher risk of preeclampsia in women who have difficulty conceiving, 10 while a systematic review including nearly two million women from 37 studies, found a 53% higher risk for gestational diabetes mellitus (GDM) in women who used assisted reproductive technologies (ART). 11 A history of infertility has also been associated with increased risks for preterm birth, 12 low birthweight, 13 and other adverse obstetric and perienatal outcomes, 14 with a meta‐analysis of 14 studies reporting that women with a longer time to pregnancy had a 38% increased risk of preterm birth and a 30% higher risk of low birthweight compared to women without fertility problems. 15 While these associations are evident, shared underlying risk factors may partly explain the observed relationships between infertility and adverse outcomes. Previous studies generally adjusted for age, parity, and sometimes BMI, all of which are independently associated with both infertility and adverse pregnancy outcomes. However, many studies lack comprehensive data on relevant health conditions, such as PCOS and endometriosis, or on lifestyle factors which could confound the observed associations between infertility and adverse outcomes. Other studies focus solely on women who conceived using ART, making it difficult to disentangle the effects of infertility from effects of the treatment. Stern et al., 16 however, considered underlying infertility diagnoses, including ovulation disorders and inflammatory conditions, and reported increased risks of adverse pregnancy outcomes in both ART ( n  = 3689) and non‐ART ( n  = 4098) pregnancies among women with a longer time to pregnancy. They concluded that these risks were related to the infertility diagnoses rather than the treatment. Importantly, that study did not consider lifestyle factors, included only women who had fertility issues, and was limited to a single U.S. center of 305 774 women, indicating the need for further research in other populations. Clarifying the nature of obstetric and neonatal risks associated with infertility is essential for informing clinical care and public health strategies. It has implications not only for maternal and child health, but also for the development of targeted antenatal interventions that address modifiable preconception risk factors. Therefore, to strengthen the evidence base and improve our understanding of the contribution of health and lifestyle factors to adverse pregnancy outcomes among women with fertility problems, this study aimed to examine associations between fertility problems and pregnancy and birth outcomes (GDM, hypertensive disorders of pregnancy (HDP), low birth weight, and preterm birth) in a nationally representative sample of women, who reported a live birth, from the Australian Longitudinal Study on Women's Health (ALSWH).

None declared.

This study is a secondary analysis of data from the 1973 to 1978 cohort of the ALSWH, an ongoing longitudinal population‐based study. The study was initiated in 1996 when 14 247 women aged 18–23 years were recruited. Participants were sampled at random from the national Medicare health insurance database, which includes all Australian citizens and permanent residents. Women living in rural and remote areas were intentionally oversampled to allow sufficient statistical power to analyse data by area of residence. Further details on recruitment, retention, and survey methods have been published previously 17 and can be found online at http://www.alswh.org.au/ . After the first survey in 1996, women completed surveys every 3–4 years. For the current study, data were taken from the 2003 survey when validated dietary intake data were first collected (Survey 3, age 25–30 years) through to the most recent survey completed in 2018 (Survey 8, age 40–45 years). Of the 9854 women who completed at least two surveys between 2003 and 2018, women were excluded if they reported they could not have children and did not report a live birth (had tubal ligation, hysterectomy or cannot have children, or partner had vasectomy or cannot have children) ( n  = 273), did not report a live birth between 2003 and 2018 ( n  = 3442), had implausible energy intake defined as daily energy intake  14 700 kJ/day ( n  = 142), 18 or had missing data on history of fertility problems ( n  = 16) or on confounders included in the final analysis ( n  = 317). Based on the sample of 5664 women eligible for analysis, we further excluded women with missing data on pregnancy and birth outcomes, and we also excluded women with pre‐existing diabetes for the analysis on GDM and pre‐existing hypertension for the analysis on HDP. This left 5355 women for analysis on GDM, 5078 for HDP, 5453 for preterm birth, and 5653 for low birthweight (Figure  S1 ). At every survey, women were asked if they and their partner (current or previous) ever had problems with fertility, defined as having tried unsuccessfully to get pregnant for 12 months or more. Women selected one of four response options: “no, never tried to get pregnant,” “no, had no problem with fertility,” “yes, but have not sought help/treatment,” and “yes, and have sought help/treatment.” Those who answered yes to either of the questions were identified as having fertility problems and were classified into two groups, those who did and did not seek treatment. Those who had never tried to get pregnant were excluded. From Survey 4 onwards, women were asked about previous births (including all births prior to Survey 4). For each birth, women reported the date of birth for each child, if they had experienced a low birthweight baby (<2500 g) or preterm birth (<37 weeks' gestation), and if they were diagnosed or treated for GDM and/or HDP. Diagnostic criteria for GDM included a 1‐h plasma glucose level ≥7.8 mmol/L after a 50 g glucose load (morning, non‐fasting) or 1‐h plasma glucose level ≥8.0 mmol/L after a 75 g glucose load (morning, non‐fasting). Diagnosis was confirmed after a 75 g oral glucose tolerance test (fasting) with a plasma glucose level at 0 h of ≥5.5 mmol/L and/or at 2 h of ≥8.0 mmol/L. 19 Diagnostic criteria were updated in 2013 with a positive test after a 75 g oral glucose tolerance test (fasting) defined as plasma glucose level at 0 h of ≥5.1 mmol/L and/or at 1 h of ≥10.0 mmol/L and/or at 2 h of ≥8.5 mmol/L. 20 HDP include gestational hypertension (new‐onset hypertension after 20 weeks of gestation: ≥140 mmHg systolic or ≥90 mmHg diastolic blood pressure) and preeclampsia (gestational hypertension with involvement of ≥1 other organ systems and/or the fetus). 21 At every survey, women responded to questions about socio‐demographic factors (age, area of residence, and highest educational qualification), health characteristics and conditions (irregular menstrual periods, polycystic ovary syndrome, endometriosis, BMI, nulliparity [i.e., not given birth before] and multiple birth), and health behaviors (physical activity, dietary intake, and smoking). Country of birth was reported at Survey 1, and age at menarche at Survey 2. BMI was calculated as weight (kg) divided by the square of height (m) and categorized as normal weight (BMI < 25.0 kg/m 2 ), overweight (BMI 25.0–29.9 kg/m 2 ), or obesity (BMI ≥ 30.0 kg/m 2 ). Less than 2.5% of women had an underweight BMI < 18.5 kg/m 2 and were therefore combined with normal weight. Physical activity scores were derived from validated questions on frequency and duration of walking (for recreation or transport) and reported information on moderate‐ and vigorous‐intensity physical activity in the last week. Level of physical activity was categorized as sedentary/low (<600 metabolic equivalent of task (MET)‐min/week), moderate (600–<1200 MET‐minutes/week), or high (≥1200 MET‐min/week). 22 Dietary intake was assessed at Survey 3 and 5 using a validated semi‐quantitative food frequency questionnaire (FFQ): the Dietary Questionnaire for Epidemiological Studies (DQES) Version 2. 23 The DQES uses a 10‐point scale for women to report usual consumption of 74 foods and beverages over the past 12 months. Serving sizes were adjusted using portion size photographs. Respondents were asked further questions about the total number of daily serves of fruit, vegetables, bread, dairy products, eggs, fat spreads, and sugars. Diet quality was estimated by the Australian Recommended Food Score (ARFS), which has been described in detail elsewhere. 24 Baseline characteristics were described for the overall study population, and according to history of fertility problems (no, yes—not sought help, yes—sought help). Differences were examined using ANOVA or chi‐square tests. Generalized estimating equations (GEE) with an exchangeable correlation structure were used to examine associations of history of fertility problems (reported at Surveys 3 to 7) with pregnancy and birth complications (based on pregnancies reported since the previous survey at Surveys 4 to 8). GEE analyses account for the correlations between multiple pregnancies, and unequally numbered measures (i.e., missing data), obtained over time for each woman. History of fertility problems was included in the models as a categorical variable (no [reference], yes—not sought help, yes—sought help) or a binary variable (no [reference], yes), and pregnancy and birth complications as a binary variable (no [reference], yes) to estimate relative risks (RR). Confounders were identified using a directed acyclic graph (Supporting Information Figure  S2 ), based on a priori and expert knowledge of causal relationships between fertility problems and pregnancy and birth complications, to minimize the risk of overmatching. Multi‐collinearity was assessed using Variance Inflation Factor: All values were <2.0, indicating moderate but acceptable correlation. In model 1, we adjusted for socio‐demographic characteristics (country of birth [Survey 1], age, area of residence, and highest educational qualification [Surveys 3 to 7]). Model 2 additionally included health characteristics and pregnancy‐related variables (age at menarche [Survey 2], irregular menstrual periods, BMI [Surveys 3 to 7], nulliparity, and multiple birth [at time of index pregnancy]). Model 3 additionally included PCOS and endometriosis (Surveys 3 to 7), and model 4 was additionally adjusted for health behaviors (physical activity, smoking status, total energy intake, and total ARFS score [Surveys 3 to 7]). Dietary intake data collected at Survey 3 was applied to Survey 4, and dietary intake data collected at Survey 5 was applied to Surveys 6 and 7. Associations were considered statistically significant if the 95% confidence interval for the relative risk excluded 1.00. While formal hypothesis testing was not conducted, the interpretation of associations was guided by both statistical and clinical relevance.

Figure S1. Flowchart. Figure S2. Directed acyclic graph. Table S1. Variables used for the main analysis. Table S2. Associations between history of fertility problems and pregnancy and birth complications in singleton pregnancies.