Background

Cardiometabolic multimorbidity, is defined as the simultaneous coexistence of two or more cardiometabolic diseases (CMDs) [1]. Cardiometabolic multimorbidity is most commonly characterized by coronary heart disease (CHD), stroke, hypertension, and diabetes and has become the predominant multimorbidity profile [2, 3]. Individuals with a single CMD have a higher likelihood of developing another, resulting in the rapid rise of the prevalence of cardiometabolic multimorbidity [4]. Cardiometabolic multimorbidity is strongly associated with adverse health outcomes and has progressively become more common worldwide in recent years. There is substantial evidence that cardiometabolic multimorbidity is related to lower quality of life, greater healthcare costs, reduced life expectancy, higher disability, and all-cause mortality [5]. This poses significant medical burden to the healthcare system and therefore, greater attention is needed to broaden the single-disease framework that dominates the priorities in research.

In addition to well-known risk factors such as behavioral, socioeconomic, lifestyle factors [6], reproductive factors unique to women may also be implicated in cardiometabolic risk [7]. CMDs are the leading cause of mortality in women and have been estimated to be responsible for approximately one-third of female deaths worldwide [8, 9]. Pregnancy loss is a common adverse pregnancy outcome that affects 12–24% of clinically recognized pregnancies [10]. Existing evidence has shown that women with a history of pregnancy loss are at a higher long-term onset of cardiovascular disease, diabetes, and consequently, CMDs-related morbidity and mortality [11,12,13,14,15,16,17,18,19,20,21,22,23,24]. The majority of studies have reported that a history of or recurrent spontaneous abortion are associated with an increased risk of individual CMDs, such as CHD [7, 11, 16, 17, 19, 20], hypertension [12, 22], and diabetes [12, 23] rather than stroke [7, 16, 17, 19]. However, evidence relating induced abortion or stillbirth to CMDs is scant and inconsistent for the same CMD [11,12,13,14, 16, 17, 20, 22,23,24] and while pregnancy loss has been well studied in relation to a single CMD, the association between the type of pregnancy loss with the risk of cardiometabolic multimorbidity is less understood. Therefore, understanding how adverse pregnancy outcomes are associated with multiple CMDs comorbidities risk factors may be useful in cardiometabolic multimorbidity prevention in women.

Using data from the China Kadoorie Biobank (CKB), we assessed the relationship between pregnancy loss (including its major subtypes: spontaneous abortion, induced abortion, and stillbirth) with the risk of cardiometabolic multimorbidity, according to the presence of cardiometabolic multimorbidity, as well as the number and type of CMD.

Methods

Study design and participants

The CKB study is a population-based prospective cohort study designed to investigate the complex interplay of genetic, environmental and lifestyle behavior factors with risks of various chronic diseases. Detailed information about the study design and survey methods of the CKB database have been previously described elsewhere [25, 26]. Briefly, the baseline survey was conducted between 2004 and 2008 in five rural and five urban geographically diverse areas across China. This cross-sectional study analyzed baseline data from the CKB study. From an initial cohort of 512,715 participants aged from 30 to 79 years old, male participants (n = 210,205), female participants who reported having never been pregnant (n = 2,881), and female participants with missing data (n = 47) were excluded. The remaining 299,582 women with a history of pregnancy were included in the present cross-sectional analysis.

Ethics approval was obtained from the Oxford University Tropical Research Ethics Committee (approval number: 025–04, 3.2.2005) and the Chinese Center for Disease Control and Prevention Ethical Review Committee (approval number: 005/2004, 9.7.2004). All participants provided written informed consent in accordance with the Declaration of Helsinki for participation [26].

Data collection

A standardized questionnaire using a laptop-based direct data-entry system was performed by trained health workers at recruitment, which covered, but was not limited to, demographic characteristics (e.g. age, and sex), socio-economic status (e.g. education, and household income), lifestyle factors (e.g. smoking, alcohol consumption, and physical activity), personal and family medical history. In the present study, education was categorized based on the highest level of attained education into the following groups: elementary school and below, middle and high school, and university and above. Annual household income was categorized follows: <5,000, 5,000–19,999, and ≥ 20,000 yuan, which were determined to balance the requirement for the minimum sample size in each category and also to enable comparisons among various transitional groups from lowest to highest income. Reproductive information, including the number of pregnancies, livebirths, spontaneous abortions, induced abortions, stillbirths, and menopause status were collected for female participants. Participants were considered to have pregnancy loss if they experienced spontaneous abortion, induced abortion, or stillbirth. After the interview, a range of physical measurements, such as weight, height, and waist-hip circumference were recorded for each participant using standard instruments and protocols. The physical activity level was measured by adding up metabolic equivalent of tasks (METs) for daily work or leisure activities.

For the purposes of this study, cardiometabolic multimorbidity was defined as the presence of ≥ 2 of the following CMDs in the same individual: CHD, stroke, hypertension, and diabetes. Cardiometabolic multimorbidity was additionally defined by the combinations of CMDs. For the CMD combinations, participants with cardiometabolic multimorbidity were categorized into 11 mutually exclusive groups based on the presence of each CMD (Supplementary Table 3). CMDs were identified if participants reported the specific disease as diagnosed by physicians or if they received treatment for CMDs. Each participant was asked: “Has a doctor ever told you that you had the following disease?” followed by a list of chronic diseases, including CHD, stroke, hypertension, and diabetes. Participants who answered ‘yes’ to this question or reported having received any treatment of these diseases were classified as having a self-reported chronic condition.

Statistical analyses

The baseline characteristics of participants are presented as means (standard deviation, SD) for continuous variables and as number (percentages) for categorical variables, respectively. The exposures were previous history of pregnancy losses (yes or no), including spontaneous abortion, induced abortion, and stillbirth. All types of pregnancy losses were further categorized into 0, 1 and 2 or more, based on the number of occurrences. The outcomes were divided into various categories by the presence of cardiometabolic multimorbidity (yes or no), the number of CMDs (0, 1, and ≥ 2 CMDs) and the combinations of CMDs.

Logistic regression was performed to obtain the odds ratio (OR) and 95% confidence intervals (95% CI) for the association between pregnancy loss, including total pregnancy loss, spontaneous abortion, induced abortion, and stillbirth with the risk of cardiometabolic multimorbidity. The associations between pregnancy loss and the number of CMDs (categorized as 0, 1, and ≥ 2) were also assessed, using participants without CMDs as reference. Similarly, the associations between pregnancy loss with 11 different CMD combinations was assessed, using participants without pregnancy loss as reference. The models were adjusted for age, province (10 study areas), region (urban and rural), body mass index (BMI), educational level, household income, MET, smoking, alcohol consumption, number of live births, and menopause status. Analyses for spontaneous abortion, induced abortion, and stillbirth were additionally adjusted for the number of spontaneous abortions, induced abortions, and stillbirths, as appropriate. Subgroup analyses by age (< 50, and ≥ 50 years), region (rural, and urban), BMI (< 25, and ≥ 25 kg/m2), educational level (elementary school and below, middle and high school, and university and above), annual income (< 5,000, 5,000–19,999, and ≥ 20,000 yuan), MET (< 20, 20-29.9, and ≥ 30 h/day), smoking (smoker, and non-smoker), alcohol consumption (alcohol drinker, and non-alcohol drinker), number of live births (< 2, and ≥ 2), and menopause status (pre-/peri-menopause, and post-menopause) were also performed to obtain the ORs and 95% CIs for the association between pregnancy loss and cardiometabolic multimorbidity. The effect of pregnancy loss and participants’ baseline characteristics on the outcome of cardiometabolic multimorbidity was assessed using an interaction term. Analyses were performed using SAS (version 9.4; SAS Institute Inc., NC, USA), with statistical significance set at a two-sided p value < 0.05.

Results

Participants’ characteristics

The baseline characteristics of the participants are presented in Table 1. Among the 299,582 women who reported having ever been pregnant, a total of 61.52% had a history of pregnancy loss, of which 9.06% had a history of spontaneous abortion, 52.48% had a history of induced abortion, and 5.69% had a history of stillbirth. All participants had a mean age of 51.46 years, mean BMI of 23.82 kg/m2, mean MET values of 20.44 h/day, and had 2.24 livebirths on average. The majority of participants resided in rural areas (55.59%), had an education of primary education or below (56.85%), had household incomes between 5000 and 19,999 yuan (49.13%), did not smoke (94.95%) and did not drink alcohol (63.63%). The number of participants with 0, 1, and ≥ 2 CMDs was 251,883 (84.08%), 37,862 (12.64%), and 9,837 (3.28%), respectively. Women with a history of pregnancy loss were more likely to come from urban areas, to have a higher education level or household income compared with those without pregnancy loss.

Table 1 Baseline characteristics of the study participants
Full size table

Pregnancy loss and risk of cardiometabolic multimorbidity

The associations between pregnancy loss with the risk of cardiometabolic multimorbidity are presented in Table 2; Fig. 1. Compared to women without a history of pregnancy loss, women with a history of pregnancy loss were significantly more likely to be diagnosed with cardiometabolic multimorbidity, OR 1.13 (95% CI 1.08–1.19). An increase in the number of pregnancy losses was significantly associated with an increased risk of cardiometabolic multimorbidity; OR 1.10 (95% CI 1.05–1.16) and OR 1.16 (95% CI 1.10–1.22) for one, and two or more pregnancy losses, respectively. When stratified by participants’ characteristics, the association was more pronounced among women aged ≥ 50 years, those with lower levels of education, lower MET values, never smokers, and never alcohol drinkers. Moreover, this relationship was apparent in women regardless of their region of residence, BMI, household income, number of live births, and menopause status (Supplementary Table 1). However, no significant interactions were found in the association between pregnancy loss and the risk of cardiometabolic multimorbidity in subgroup analyses.

Table 2 Effect estimates of the association between pregnancy loss and risk of cardiometabolic multimorbidity
Full size table
Fig. 1
figure 1

Effect estimates of the association between pregnancy loss and risk of cardiometabolic multimorbidity. Adjustments are as in Table 2 model 2. **p < 0.01, *p < 0.05

Full size image

The associations were similar in direction for the various types of pregnancy losses; OR 1.10 (95% CI 1.03–1.18), OR 1.13 (95% CI 1.08–1.19), and OR 1.03 (95% CI 0.95–1.11) for a history of spontaneous abortion, induced abortion, and stillbirth, respectively. However, the association between stillbirth and the risk of cardiometabolic multimorbidity was not significant in the fully adjusted model. An increasing number of induced abortions was also found to be significantly associated with an increasing risk of cardiometabolic multimorbidity; OR 1.09 (95% CI 1.03–1.15), and OR 1.19 (95% CI 1.13–1.26) for one, and two or more induced abortions, respectively.

Pregnancy loss and risk of CMD by the number of CMDs

The associations between pregnancy loss and the number of CMDs are presented in Fig. 2 and Supplementary Table 2. Women with a history of pregnancy loss had increased odds of being diagnosed with 1 CMD (OR 1.11, 95% CI 1.08–1.14). Similarly, women with a history of spontaneous abortion, induced abortion and stillbirth have increased odds of being diagnosed with 1 CMD; OR 1.01, 95% CI 0.98–1.05, OR 1.12, 95% CI 1.10–1.15, and OR 1.02, 95% CI 0.98–1.07, respectively. Interestingly, the odds of being diagnosed with 2 or more CMDs is higher than being diagnosed with 1 CMD in women with a history of pregnancy loss (OR 1.16 (95% CI 1.11–1.21). Similarly, the odds of being diagnosed with 2 or more CMDs was higher than being diagnosed with 1 CMD in women with a history of spontaneous abortion and induced abortion; OR 1.09, 95% CI 1.01–1.16, and OR 1.17, 95% CI 1.12–1.23, respectively. However, this was not the case for women with a history of stillbirth, OR 1.01, 95% CI 0.93–1.12.

Fig. 2
figure 2

Effect estimates of the association between pregnancy loss and risk of CMD by CMD count. Adjustments are as in Table 2 model 2. **p < 0.01, *p < 0.05

Full size image

An apparent dose-response association was observed between the number of pregnancy losses and the increased odds of being diagnosed with 1 CMD (OR 1.10, 95% CI 1.07–1.13, and OR 1.12, 95% CI 1.09–1.16 for one, and two or more pregnancy losses, respectively) (p for trend < 0.01) and 2 or more CMDs (OR 1.13, 95% CI 1.07–1.19, and OR 1.19, 95% CI 1.13–1.26 for one, and two or more pregnancy losses, respectively) (p for trend < 0.01). This dose-response relationship was consistent among women with induced abortions; OR 1.11 (95% CI 1.08–1.14), and OR 1.14 (95% CI 1.11–1.18) for one, and two or more induced abortions associated with 1 CMD (p for trend < 0.01) and OR 1.12 (95% CI 1.06–1.18), and OR 1.24 (95% CI 1.17–1.31) for one, and two or more induced abortions associated with 2 or more CMDs (p for trend < 0.01). However, this dose-response was not apparent between the number of spontaneous abortions or stillbirths with the risk of CMDs (p for trend > 0.05).

Pregnancy loss and risk of cardiometabolic multimorbidity by CMD combinations

The associations between pregnancy loss and the risk of cardiometabolic multimorbidity by the combinations of CMDs are presented in Fig. 3 and Supplementary Table 4. Of the 11 CMD combinations, women with total pregnancy loss, in particular, induced abortion were significantly associated with a higher risk of cardiometabolic multimorbidity across most CMD combinations. Similarly, women with a history of spontaneous abortion were more likely to be diagnosed with cardiometabolic multimorbidity, although the association was largely non-significant, except for the combinations of CHD and hypertension.

Fig. 3
figure 3

Effect estimates of the association between pregnancy loss and risk of cardiometabolic multimorbidity by CMD combinations. Adjustments are as in Table 2 model 2. **p < 0.01, *p < 0.05

Full size image

In addition, women with a history of stillbirth were significantly associated with higher risk of cardiometabolic multimorbidity, including the combinations of (1) CHD and hypertension, and (2) CHD and diabetes. However, the association between women with a history of stillbirth and most CMD combinations were inverse, although the associations were non-significant.

Discussion

To our knowledge, this large study of almost 300,000 women in China is among the first to evaluate the relationships between pregnancy loss and cardiometabolic multimorbidity. Our findings demonstrated that pregnancy loss, in particular, spontaneous or induced abortions, were associated with higher risk of cardiometabolic multimorbidity. However, the association was not significant for pregnancy loss due to stillbirth. Our findings also showed that an increased number of pregnancy losses, in particular induced abortions, was associated with an elevated risk of cardiometabolic multimorbidity.

Much of the evidence on the role of pregnancy loss for health comes from studies on individual CMDs, rather than cardiometabolic multimorbidity. These studies have yielded conflicting results on the association of pregnancy loss with cardiovascular disease and diabetes. A meta-analysis of ten studies reported that a history or recurrent spontaneous abortion was associated with a greater risk of CHD and stroke, although the association was weaker for stroke [7]. Similarly, two studies using data from the CKB database and the Nurses’ Health Study II (NHSII) reported that spontaneous abortion and induced abortion was associated with increased risk of CHD and stroke [11, 20]. Data from the NHSII also demonstrated greater rates of hypertension in women with a history of spontaneous abortion and stillbirth. However, interestingly, the study also found reduced rates of hypertension in women with a history of induced abortion [12]. Another population-based cohort in Denmark of more than one million women found that a history of spontaneous abortion and stillbirth was associated with a higher risk of stroke and hypertension, with stronger associations observed in women with repeated spontaneous abortions [22]. Moreover, a prior study conducted in the United States on 77,701 women reported that a history of 1 and 2 or more spontaneous abortion, and a history of 1 or more stillbirth was associated with a greater risk of CHD but not stroke [17]. Similarly, a Scottish study on 60,105 female participants demonstrated that women who had a history of two or more spontaneous abortion had an increased risk of CHD but not stroke [19], and a study using data from the UK Biobank on 267,440 women reported that a history of spontaneous abortion was associated with an increased risk of CHD, but not stroke. Interestingly, however, the study reported that a history of stillbirth was associated with an increased risk of stroke, but not CHD [16].

Few studies have examined the relationship between pregnancy loss and the onset of diabetes, with inconsistent findings. The EPIC-Heidelberg study on 13,612 women found that a history or recurrent spontaneous abortion was associated with a greater risk of diabetes. In contrast, no significant associations were observed between induced abortion and stillbirth with the risk of diabetes [23]. A study using data from the NHSII on 60,651 women reported that spontaneous abortion and stillbirth was associated with increased risk of type 2 diabetes, whereas induced abortion was associated with lower risk of type 2 diabetes [12]. An Italian study on 15,404 women reported that stillbirth increased the risk of diabetes by approximately two-fold [24], and a recent study utilizing data from the CKB database reported that pregnancy loss, particularly induced abortion, was associated with an increased risk of diabetes, although the associations were not significant for spontaneous abortion and stillbirth [13].

Our analyses extended upon the study of a single CMD in the existing literature to cardiometabolic multimorbidity, demonstrating that a history of pregnancy loss was associated with an increased risk of cardiometabolic multimorbidity. Similar to most studies, we found that pregnancy loss, particularly recurrent pregnancy loss, was not only associated with an increased risk of a single CMD, but also with a higher risk of cardiometabolic multimorbidity. The findings of this study suggest that women with a history of pregnancy loss may benefit from early screening for risk factors of cardiometabolic multimorbidity. Furthermore, our findings demonstrated that the association between pregnancy loss and cardiometabolic multimorbidity was more pronounced in older women (aged ≥ 50 years). However, no significant differences were observed in the effect of pregnancy loss on the risk of cardiometabolic multimorbidity between different age groups. These findings suggest that pregnancy loss may be an important risk factor for cardiometabolic multimorbidity, which is consistent with the expectation that metabolic processes deteriorate with age due to factors such as cumulative oxidative stress, thereby contributing to the occurrence of CMDs [27,28,29]. The association was also more pronounced in women with lower levels of education and MET hours. Previous studies have reported that women with lower educational achievement tend to have a higher risk of pregnancy loss [30]. There is also evidence that physical activity during pregnancy is beneficial for both the woman and the fetus, with women who exercise during pregnancy demonstrating a reduced risk of pregnancy loss [31]. Moreover, we observed a lack of association between pregnancy loss and cardiometabolic multimorbidity in smokers and alcohol drinkers. This may be because smokers and alcohol drinkers are known to take a longer time to conceive, and due to the limited sample size of this subgroup of women, the association between pregnancy loss and the risk of cardiometabolic multimorbidity may be masked.

In addition, we found a high prevalence of pregnancy loss, especially induced abortion, in this population. Given that the baseline data collection of the CKB study was conducted from 2004 to 2008, this period was particularly important as it overlapped with the implementation of the one-child policy in China, where pregnancy loss was a common outcome of pregnancy during the one-child policy. According to data from national statistics and nationally representative sample surveys, during the one-child policy era, the rate of induced abortion peaked at 56.07% in the 1980s, followed by a gradual downward trend in the 2000s, with an average annual rate of 28.95% among married women aged 20–49 years [32]. Our study revealed a relatively higher prevalence of induced abortion compared with the national study. Notably, the participants in our study, aged 30–79 years with a mean age of 51.46 years, were much older than the mean age of the participants in the national survey. The higher rate of induced abortion observed in our study population may be partly due to the longer cumulative exposure to the risk among older individuals.

There are several possible explanations for the mechanisms underlying the link between pregnancy loss and cardiometabolic multimorbidity. We postulate that both pregnancy loss and cardiometabolic multimorbidity may be due to endothelial dysfunction, which results in placental dysfunction, pregnancy loss, and the development of both single CMD and cardiometabolic multimorbidity [33, 34]. Studies have demonstrated that women with recurrent pregnancy loss have a higher rate of endothelial dysfunction compared to those with uncomplicated pregnancies [33]. Autoimmune disorders and consequently systemic inflammation have been implicated in the occurrence of pregnancy loss [35, 36], and there is growing evidence indicating that inflammatory processes play a role in the development of cardiometabolic multimorbidity [22, 37, 38]. However, further studies are necessary to verify our findings and to elucidate the underlying mechanisms by which pregnancy loss may be involved in the pathophysiology of cardiometabolic multimorbidity.

The present study has several strengths. First, the analysis was based on a large representative population sample from ten diverse areas in China, which strengthened the generalizability of our findings. Second, the comprehensive collection of data allowed for the analysis of the correlations between various types of pregnancy losses and combinations of CMDs, as well as the control of a variety of potential confounders, which contributed to the emerging literature on associations that were limited to a single disease.

However, the study has several limitations. First, as a cross-sectional study, endogeneity issues inherent in such studies are inevitable. The temporal and causal relationships between pregnancy loss and cardiometabolic multimorbidity cannot be discerned, which may lead to spurious associations. Further studies using a longitudinal cohort design or identifying appropriate instruments are warranted, which can address the endogeneity in our results and provide unbiased estimates of the observed associations. Second, pregnancy loss and CMDs were self-reported, which may lead to recall bias. Third, although our methodology facilitated the effective examination of specific CMD combinations, we did not use unsupervised methods to create the CMD combination groups. This may have restricted the identification of other potential groupings or patterns. Last, although we adjusted for a wide range of covariates, residual confounding, such as dietary habits, which were not available, may still remain.

Conclusion

In conclusion, findings from the present study demonstrated that pregnancy loss, specifically spontaneous and induced abortion, was associated with a higher risk of cardiometabolic multimorbidity in Chinese women. Furthermore, increases in the number of pregnancy losses were associated with a greater risk of cardiometabolic multimorbidity.