Obesity increases the odds of multiple pregnancies: A nationwide register‐based cohort study in Finland

To investigate the effects of increased pre‐pregnancy body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters) on the risk of having spontaneous multiple pregnancies using a nationwide register‐based study sample.


| INTRODUC TI ON
The effects of overweight and obesity on the risk of multiple pregnancies have been studied. Multiple pregnancies are relatively rare, making the phenomenon challenging to study without proper data.
Only a few studies have investigated the effects of a higher body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters) on the risk of multiple pregnancies. A 2005 study that analyzed the US population found that women with a higher BMI had an increased risk of having twins. 1 The same study found that the risk of dizygous twins was higher, but the risk of homozygous twins was not increased. 1 Another study using a twin register in the Netherlands also found higher risk for multiple pregnancies in the obese population. 2 According to a large National Birth Cohort in Denmark, in which women were recruited early in pregnancy and responded to two questionnaires before delivery, increasing BMI and height correlated with twinning among women without fertility treatment. 3 A Swedish cohort study reported a higher rate of multiple pregnancies in higher BMI groups when compared with a normal weight population. 4 However, to the best of our knowledge after these studies, no other study has focused on the association between a higher BMI and the risk of multiple pregnancies.
Women with higher BMI are known to be at higher risk for infertility. 5 Overweight women have a higher incidence of menstrual dysfunction and anovulation. 5 According to a review of the literature, obese women are more likely to have ovulatory dysfunction due to dysregulation of the hypothalamic-pituitary-ovarian axis. 6 Therefore, assisted reproductive technologies are more common among women with obesity and also a poorer outcome has been reported among these women. 7,8 Based on our hypothesis, the possible increased risk of multiple pregnancies might be caused by disorders in hormonal activity due to obesity. 9 In obese women, spontaneous multiple pregnancies are suggested to be related to the increased levels of follicle-stimulating hormone. 10 Additionally, the prevalence of obesity and rates of obese parturients giving birth are increasing, meaning that obesityrelated risks must be more thoroughly studied and acknowledged. 11 Hence, the aim of this study is to investigate the effects of increased BMI on the risk of having multiple pregnancies using a nationwide register-based study sample.

| MATERIAL S AND ME THODS
In this nationwide retrospective register-based cohort study, data from the National Medical Birth Register (MBR) was used to evaluate the effects of a higher pre-pregnancy BMI on the risk of multiple pregnancies. Multiple pregnancy was defined as a pregnancy with two or more fetuses. The MBR is maintained by the Finnish Institute for Health and Welfare. The study period was from 1 January 2004 to 31 December 2018.
The MBR contains information on pregnancies, delivery statistics and the perinatal outcomes of all births with a birthweight of 500 g or more or a gestational age of 22 +0 weeks or longer. The MBR has high coverage and quality (the current coverage is nearly 100%). 12,13 We included all pregnancies recorded in the MBR between 2004 and 2018. Stillbirths were also included. A total of 697 710 pregnancies with a known BMI status were included in this study. Cases with missing BMI statuses were excluded. The weights collected were either the pre-pregnancy weight or the weight measured at the first visit to the maternity clinic during weeks 6-8 of pregnancy.
The study groups were formed using the classifications given by the World Health Organization. 14 Based on their pre-pregnancy BMI, women were divided into six groups: underweight (BMI less than 18.5), overweight (BMI 25.0-29.9), obesity class I (BMI 30.0-34.9), obesity class II (BMI 35.0-39.9), obesity class III (BMI more than 40.0), and a reference group comprising normal weight women (BMI 18.5-24.9). A total of 20 784 (4.2%) women in the reference group were underweight. Pregnancies that were started using fertility treatments were excluded from the analysis, as the chances of multiple pregnancies are higher among these pregnancies. 15 The process used to form the study groups is shown as a flowchart in Figure  Continuous variables were interpreted as means with standard deviations or as a median with an interquartile range based on the distribution of the data. The categorical variables are presented as absolute numbers and percentages. A logistic regression model was used to assess the primary outcomes. The exposure variable was the categorized BMI class, and the primary outcome was multiple pregnancies. Women in higher BMI classes were compared with women in the non-overweight BMI class. Adjusted odds ratios (aOR) with 95% confidence intervals (CI) between the groups were compared.
The model was adjusted for maternal age (continuous), maternal height (continuous), and maternal smoking status during pregnancy (yes/no), as these are known to be possible risk factors for multiple pregnancies. 3,16,17 Adjustments were made by choosing the variables for a multivariate model using directed acyclic graphs (DAG) constructed using the free online software DAGitty (dagit ty.net). 18 The DAG approach can be used to help choose which covariates should be included in traditional statistical approaches to minimize the magnitude of the bias in the estimate produced. 19 The variables included in the DAGs were chosen based on known risk factors and by hypothesized causal pathways. DAGitty automatically suggests possible adjustment variable sets that can influence the main outcome (see Figure S1). A P value less than 0.05 was considered statistically significant in all analyses. Statistical analyses were performed using R version 4.0.3 for Windows (R Foundation for Statistical Computing).

| DISCUSS ION
The main finding of this study is that the odds of multiple pregnancies increase with increasing BMI. The odds of multiple pregnancies were a little higher with a higher BMI, except for obesity class III, where the odds were not increased. However, the number of patients in this group was relatively small, and the results may be imprecise.
Based on previous studies, a higher BMI increases the risk of multiple pregnancies, 1,3,4 which is also supported by our results featuring a large nationwide register-based study sample. The increase in risk for multiple pregnancies was relatively restrained in our study when compared with these previous studies. The Swedish study with a large national birth cohort reported adjusted odds of 1.44 (95% CI 1.13-1.83) among women with BMI greater than 30, 3 which is higher than observed in our results. However, because of the crude nature of our data, the exact reason for this slightly increased risk remains unknown, and more research is warranted. The results of this study should provide important information regarding the risks of multiple pregnancies in overweight and obese mothers.
The research behind the association between obesity and multiple pregnancies is important, because obesity and multiple pregnancies are known to be risk factors for adverse pregnancy outcomes, both for the neonate and the mother. 20,21 According to a large cohort study consisting of only twin pregnancies, higher BMI was associated with gestational hypertension, pre-eclampsia, and gestational diabetes. 22 In addition, as multiple pregnancies are associated with increased risk for gestational diabetes 23,24 and obesity is known to be one of the biggest risk factors for the development of gestational diabetes, 25 studies assessing the association among these factors are important in the prevention of adverse maternal and fetal outcomes.

F I G U R E 1
Flowchart of the study population. Women with different classes of obesity were compared with women with normal body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters).
In addition, the rapidly increasing prevalence of obese parturients makes this an important research topic. 11,26 Compared with previous studies, the unique strength of our study is the large nationwide register with a BMI variable registered for nearly all pregnancies that occurred during the study period, making it the most comprehensive data set on the BMI of women. 12,13 Compared with previous studies in which BMI was assessed using questionnaires that are inevitably vulnerable to recall or reporting bias, the register data used in our study are routinely collected using structured forms and national instructions, which ensures good coverage and reduces possible reporting and selection biases.
Furthermore, the coverage of both registers included in this study was high. The main limitation for this study is that we do not have information on the type of multiple pregnancy (e.g. homozygous and dizygous). In addition, the information on the fertility treatments is not routinely collected, meaning that only treatments brought up by the mother or treatment providers are registered in the MBR.
However, as our study sample is large and the need for fertility treatments is still a relatively small percentage of all pregnancies, 27 we believe that this does not have an important effect on the main results. However, it is good to acknowledge that the total number of fertility treatments reported by the Finnish Institute for Wealth and Welfare is higher than observed in the MBR, meaning that all fertility treatments are not available in the register.
In summary, based on our results, the odds of multiple pregnancies increase slightly with obesity, and it should be acknowledged as a minor risk factor for multiple pregnancies.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared. Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters). a Data are presented as mean ± standard deviation or as number (percentage).