The non-traditional and familial risk factors for preeclampsia in the FINNPEC cohort

Background Considering the burden of preeclampsia (PE), it is important to understand better the underlying risk factors involved in its etiology. In this nationwide study, we studied the association of background factors with PE with an emphasis on socioeconomic factors, reproductive factors and health history enclosing the parents of pregnant women. Methods In the Finnish Genetics of Pre-eclampsia Consortium (FINNPEC) cohort participants filled in a questionnaire on background information including data on socioeconomic factors, health history and reproductive factors. The questionnaire data was available from 708 women with PE and 724 control women. Two different control groups, healthy controls with uncomplicated pregnancies (n=498) and all controls (n=724, including controls with uncomplicated pregnancies and pregnancy complications other than PE), were established. Background information among PE women and the two different control groups were compared. Results PE women had similar socioeconomic status and more often non-communicable diseases including type I diabetes, chronic hypertension and hyperlipidemia than the two control groups (p<0.05 for all). Depression was more common among PE women (11.4%) than among all controls (7.6%) (p=0.019). Subfertility (estimated by time to pregnancy) was more common among PE women (p=0.013 for healthy controls, p=0.019 for all controls). PE women had earlier menarche (p=0.001 for healthy controls, p=0.022 for all controls). Hypertension was more common in both parents of PE women (p<0.001), stroke in fathers (PE women 6.2 %, healthy controls 3.2 % (p=0.020) and all controls 3.5 % (p=0.022)) and diabetes in mothers (PE women 7.5 %, healthy controls 3.1 % (p=0.001) and all controls 4.3 % (p=0.012)). Mental disorders including depression were more common in mothers of PE women compared to controls (PE women 7.2 %, healthy controls 3.7 % (p=0.013) and all controls 3.9 % (p=0.007)).

Conclusions In this Finnish nationwide FINNPEC cohort, PE women had similar socioeconomic status, more non-communicable diseases and depression, earlier menarche, more subfertility and more parental non-communicable diseases compared to controls. As a novel finding we found more mental disorders including depression in mothers of PE women. Background Preeclampsia (PE) is a pregnancy complication characterized by coexistence of hypertension and one or more of the following new-onset conditions: proteinuria, other maternal organ dysfunction or uteroplacental dysfunction (1). It affects 3-5 % of pregnancies and is a major cause of maternal, fetal and neonatal mortality and morbidity (2). The etiology of PE remains unclear but reduced placental perfusion and inflammation associating with oxidative stress and endothelial dysfunction are considered as central features in its pathogenesis. Maternal genetic, behavioral and metabolic factors are thought to contribute to the PE phenotype (3). Several maternal risk factors for PE have been identified such as previous PE, chronic hypertension, pregestational diabetes, body mass index (BMI) > 30 kg/m², use of assisted reproductive technology (ART), chronic kidney disease, antiphospholipid syndrome and other autoimmune disorders (2,4).
However, known clinical risk factors are predictive only in 30 % of PE women (2).
The information on other characteristics or exposures as risk factors for PE is inconsistent and/or modest. Low socioeconomic status is associated with PE in most (5)(6)(7)(8) but not all studies (9,10). Previously only few studies have assessed the association with childhood socioeconomic status and PE (10,11). An increasing amount of studies have investigated the association between PE and use of medication. In a systematic review antidepressant use during the second trimester was associated with an increased risk of PE (12). Whether antidepressants or underlying depression affect the risk of PE independently remains uncertain, because depression itself is related to PE (13,14). Family history of hypertension and cardiovascular disease (CVD) have been shown to increase risk for PE (15) but to our knowledge, family history of depression has not been studied. Few studies have addressed the association of age at menarche with the risk of PE and the results have been conflicting. Early age at menarche has been associated with PE in a few (16,17) but not all studies (18).
Identifying women at risk for PE is important in early pregnancy because strategies to prevent PE are available. Aspirin initiated at ≤ 14 weeks of gestation reduces the risk of preterm PE in women with high risk for PE (19). Since traditional risk factors predict only 30 % of women who develop PE later in pregnancy, we focused on less explored risk factors: socioeconomic factors, health history including the parents of the study participants and reproductive factors. (2).

Study design
Data was derived from the prospective arm of the Finnish Genetics of Pre-eclampsia Consortium (FINNPEC) cohort. FINNPEC, a cross-sectional case-control multicentre study, was established to set up a nationwide clinical and DNA database on women with and without PE including their partners and infants in order to identify genetic risk factors for PE. Details of the study design, methods and procedures have been described in the FINNPEC cohort profile (20).

Study subjects
In the current study, we included 923 women with PE and 1009 controls recruited during 2008-2011. The inclusion criteria were age above 18 years, a singleton pregnancy and ability to provide an informed consent based on information in Finnish or Swedish. Using the American College of Obstetricians and Gynecologists (ACOG) 2002 criteria, PE was defined as hypertension and proteinuria (21). Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg after 20 weeks of gestation. Proteinuria was defined as the urinary excretion of ≥0.3 g protein in a 24-hour specimen, or 0.3g/l, or two ≥1+ readings on dipstick in a random urine sample with no evidence of a urinary tract infection. Each diagnosis was confirmed independently from medical records by a research nurse and a research physician. The FINNPEC study protocol was approved by the coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa.
The participants were asked to fill a detailed questionnaire on background information including data on socioeconomic factors, health history and reproductive factors.
Information on pre-pregnancy weight and height, obstetric history, pregnancy outcome, delivery and newborn was obtained from the hospital records and maternity cards. Data on smoking were collected from the maternity cards and complemented from the background information questionnaires. The participants filled in the questionnaire during pregnancy, or shortly after delivery.
A detailed questionnaire on background information was available from 708 women with PE (76.7 %) and 724 control women (71.8 %). Two different control groups were established for the current study. A subgroup of healthy controls consisted of 498 healthy women with uncomplicated pregnancies. All controls (n = 724) included also women with other pregnancy complications than PE.

Statistical analysis
Statistical analyses were performed using SPSS Statistics 23.0 (IBM Corp., Armonk, NY, USA). Background information among PE women and the two different control groups were compared separately (PE vs. healthy controls and PE vs. all controls). The normality of distributions was verified graphically and with the Kolmogorov-Smirnov-test. Statistical analyses of continuous variables were performed using the two-sample t-test for normal and Mann-Whitney U-test for skewed distributions. For categorical variables comparisons between the groups were performed with the Chi-square test. P value of <0.05 was considered as statistically significant.

Results
Basic characteristics of the study population Basic characteristics of the study participants are presented in Table 1. Women with PE were more often nulliparous and had a higher BMI than women in the two control groups.
They delivered earlier and more often with a caesarean section. PE women who did not respond to the questionnaire were more often multiparous, had a higher BMI and delivered earlier and more often with a caesarean section compared to responding PE women. There were no differences in the background factors between the respondents and nonrespondents in the control groups (data not shown).

Socioeconomic factors
Socioeconomic factors of the study participants are presented in Table 2. Women with PE were more often employed and they worked more hours/week than women in the two control groups. However, when the women without employment (including housewives, students and unemployed) were excluded, there was no difference in the weekly working hours between the groups. PE women reported more unemployment during the 12 months preceding delivery. Women in the control groups were more often housewives and married. There were no differences in education level, income and area of residence between the PE women and controls.

Health history
Information on the participants' health and pregnancies is shown in Table 3. The subjective health of the women suffering from PE was poorer. PE women had more often pre-existing medical conditions (Table 4) including type I diabetes, chronic hypertension, hyperlipidemia and depression. The use of any regular or seasonal medication did not differ between women with PE and all controls (Table 5). When compared to healthy controls, women with PE used more often selective serotonin reuptake inhibitors (SSRIs) and metformin. Women with PE were more often on insulin and antihypertensive medication, including medications used before pregnancy, than women in the two control groups. Reported oral health was found similar between women with PE and all controls.
However, women with PE were more often treated for tooth decay than the healthy controls (data not shown).

Reproductive factors
Menarche occurred earlier in the women with PE compared to the control groups (Table 3).
It took more time to conceive for women with PE, more than twelve months for 16.7 % of women with PE, for 12.4 % of healthy controls and for 12.8 % of all controls. The use of ART did not differ between the groups.

Study participants' parents
Education and medical history of the index patient's family is presented in Table 6. The fathers of the women in the control groups had higher education than those of the women with PE, whereas there were no differences regarding the education of the participants' mothers. Both parents of PE women suffered more often from hypertension than the parents of the women in the two control groups. Also stroke was more prevalent in the fathers of the women with PE. Diabetes and mental disorders, including depression, were more common in mothers of PE women. Furthermore, PE women were more often born 8 from a pregnancy complicated by PE than the women in the control groups.

Discussion
In this Finnish nationwide FINNPEC cohort, women with PE had similar socioeconomic status, more non-communicable diseases and depression, earlier menarche and more subfertility compared to controls. Hypertension was more common in both parents, stroke in fathers and diabetes and mental disorders including depression in mothers of PE women compared to controls.
The socioeconomic status of the study groups was similar estimated both by education and income which is in line with some previous studies (9,10). In contrast, several previous studies have reported an association between low socioeconomic status and PE (5-8) when maternal education (5,6) or income (7,8) have been used as indicators of socioeconomic status. It remains unclear whether the association between PE and low socioeconomic status found in some studies is attributed to inadequate prenatal care or to low socioeconomic status itself. Inadequate prenatal care is associated with PE, and women with low socioeconomic status are less likely to receive adequate prenatal care (7). In Finland services of the maternity clinics are free of charge and of high-quality and used by more than 90 % of pregnant women. Thus, inadequate prenatal care is not likely to be an intermediating factor in our population. Furthermore, in Finland there are fairly small differences between social classes and education is costless also at university level.
Only few studies have explored the association between socioeconomic status in childhood and adulthood and hypertensive pregnancy disorders. Studies carried out in Sweden and in the UK found that neither childhood nor adulthood socioeconomic status were related to PE (10,11). In our study there was no difference in the socioeconomic status regarding mothers of the study participants, but the fathers of the women in the control groups had a higher education level. The results are not completely comparable with previous studies which used both education, family social class (11) and childhood social class based on occupation of study participants' fathers (10) as indicators of socioeconomic status.
We found that women with PE were more often employed than the women in the control groups. Yet there was more unemployment among PE women during the last 12 months. This is probably due to the fact that women in the control groups were more often multiparous and housewives. Among employed women, the weekly working hours did not differ between the groups, in line with previous studies (22,23). A Dutch cohort study found no difference in the risk of PE between employed and unemployed women. They also compared the risk of PE among women experiencing different types of unemployment and reported that housewives and job-seeking women had a similar risk of PE as employed women. However, when compared to housewives, job-seeking women had a higher risk for PE. (23) The discrepancy with our results may be partly due to a smaller number of PE women in the Dutch cohort. In addition, the population of the Dutch study was somewhat different concerning employment. The percentage of all employed women was 72 % in the Dutch study compared to 66.8 % for PE women, 54.3 % for healthy controls and 56.3 % for all controls in the present study.
Our data suggest that the overall health was worse in the women developing PE compared to the control women. The known clinical risk factors for PE such as pregestational diabetes, chronic hypertension and hyperlipidemia were more prevalent also in this study population (2,4,24).
Our finding on increased incidence of depression among women with PE is in line with the results of two meta-analyses (13,14) which showed that both a history of depression and depressive symptoms during pregnancy are associated with an increased risk of PE.
Depression and PE may share common pathophysiologic pathways since inflammation and oxidative stress are associated both with depression and PE (3,25). The use of SSRIs was higher among women with PE than controls with uncomplicated pregnancy, but there were no differences when the women with PE were compared to all controls. A systematic review of seven studies showed that the use of antidepressants during the second trimester of pregnancy was associated with a modestly increased risk of PE or gestational hypertension (12). The risk of PE appeared to be low in patients using SSRIs or serotonin- The parents of PE patients had more morbidity compared to those of the controls' including hypertension which is in line with some previous studies (15,28,29). Family history of CVDs including stroke (15,29) and type II diabetes (30) have been shown to increase risk of PE. Our findings on fathers of the PE women suffering from stroke and mothers with diabetes are in accordance with previous data. We also found an increased incidence of mental disorders including depression among PE womens' mothers. To our knowledge there is no previous published research on this topic.
Earlier menarche in women with PE compared to the women in control groups is in line with some previous studies (16,17). Higher BMI may partly mediate this association (17).
Early age at menarche has been associated to gestational diabetes (31) and CVD (32).
Gestational diabetes (33) and future cardiovascular morbidity (34) have been shown to be associated with PE as well. In our study women with PE had more often risk factors for metabolic syndrome, like higher BMI, which also associates with earlier menarche. This may partly explain our finding on earlier menarche with PE women. In a recent study Petry et al. found a negative association between age at menarche and subsequent blood pressure in pregnancy, but not independently with the risk of PE (35). Their results were attenuated byincreased BMI and insulin resistance. The researchers suggested that the associations between blood pressure and age at menarche are mediated by obesity and/or insulin resistance. Alternatively, there may be a common pathophysiological mechanism like systemic inflammation that links age at menarche to each of these metabolic risk factors (35). Indeed, a negative correlation between age at menarche and circulating Creactive protein concentrations has been reported suggesting a link between early age at menarche and inflammation (36). Insulin resistance and increased blood pressure may result from low grade systemic inflammation occurring as a result of increased macrophage infiltration of adipose tissue (35). CVD is also associated with systemic inflammation with adipose tissue being a significant contributor to the inflammatory state (37). The associations between PE, early age at menarche, insulin resistance and CVD may be explained by obesity or systemic inflammation or both. Future studies are needed to better understand these associations.
Although infertility treatments were as frequent in all groups, conceiving took more time for women with PE. In line with our findings, previous studies have shown that subfertile women are at increased risk for PE (38)(39)(40)(41)(42). It is well recognised that pregnancies following ART are at higher risk for PE when compared with those after natural conception (43)(44)(45). Similar underlying causes may be behind PE and infertility. Hayashi et al. found that adverse obstetric outcomes in singleton pregnancies after ART may be related to maternal factors associated with infertility rather than the type of ART used (46). Inflammation and/or obesity have been suggested as possible mechanisms linking subfertility with PE (47). These conditions may also associate PE with early age at menarche, CVD and insulin resistance.

Strengths and limitations
The main strength of this study is a nationwide and population based cohort with detailed information from medical records. Further, background information was collected on a wide basis comprising health, lifestyle and health history of the parents of the pregnant women. The response rates to the study's questionnaires were good (PE women 76.7 % and control women 71.8 %) considering that a response rate of 60 % has face validity as a measure of survey quality (48). On the other hand PE nonrespondents differed somewhat from PE respondents and this may have biased some results of the study.Information gained through questionnaires is subjective, which can be considered a limitation of this study. The completeness and accuracy of information varies from case to case.

Conclusions
In this Finnish nationwide FINNPEC cohort, PE women had similar socioeconomic status, more non-communicable diseases and depression, earlier menarche, more subfertility and more parental non-communicable diseases compared to controls. As a novel finding we found more mental disorders including depression in mothers of PE women. More attention should be paid to depression/depressive symptoms and family health history of depression in addition to somatic diseases in antenatal care.