Cardiovascular disease in people born to unmarried mothers in two historical periods: The Helsinki Birth Cohort Study 1934–1944

Aims:Socio-economic conditions in early life are important contributors to cardiovascular disease – the leading cause of mortality globally – in later life. We studied coronary heart disease (CHD) and stroke in adulthood among people born out of wedlock in two historical periods: before and during World War II in Finland. Methods: We compared offspring born out of wedlock before (1934–1939) and during (1940–1944) World War II with the offspring of married mothers in the Helsinki Birth Cohort Study. The war affected the position of unmarried mothers in society. We followed the study subjects from 1971 to 2014 and identified deaths and hospital admissions from CHD and stroke. Data were analysed using a Cox regression, adjusting for other childhood and adulthood socio-economic circumstances. Results: The rate of out-of-wedlock births was 240/4052 (5.9%) before World War II and 397/9197 (4.3%) during World War II. Among those born before World War II, out-of-wedlock birth was associated with an increased risk of stroke (hazard ratio (HR)=1.44; 95% confidence interval (CI) 1.00–2.07) and CHD (HR=1.37; 95% CI 1.02–1.86). Among those born out of wedlock during World War II, the risks of stroke (HR=0.89; 95% CI 0.58–1.36) and CHD (HR=0.70; 95% CI 0.48=1.03) were similar to those observed for the offspring of married mothers. The p-values for interaction of unmarried×World War II were (p=0.015) for stroke and (p=0.003) for CHD. Conclusions: In a society in which marriage is normative, being born out of wedlock is an important predictor of lifelong health disadvantage. However, this may change rapidly when societal circumstances change, such as during a war.


Introduction
Globally, cardiovascular disease (CVD) is the leading cause of death and years of life lost; close to 18 million people die of CVD every year [1,2]. Risk factors for CVD include lower socio-economic position (SEP) throughout the life course [3][4][5].
In life-course epidemiology, childhood SEP is most commonly indicated by parental occupation and educational attainment [3]. These indicators, however, may not capture all aspects of inequality. One important group whose deprivation may be overlooked by these indicators is children born out of wedlock, that is, to an unmarried mother [6][7][8][9]. The few studies that have followed up these children show that they seem to carry this disadvantage throughout their lives [7,8,10,11]. At least two previous studies using Nordic data from the early part of the Cardiovascular disease in people born to unmarried mothers in two historical periods: The Helsinki Birth Cohort Study 1934-1944 twentieth century have reported higher CVD risk in adult life for children from a single-parent background [8,11].
However, due to the small number of studies, the role of the specific sociohistorical context in interpreting the association between a single-parent background and CVD in later life remains unclear. It is likely that rapid changes in society, such as war, may affect the context of single-parent background by changing social norms and socio-economic resources.
Particularly in societies where marriage is the parental norm, mothers and their out-of-wedlock children form a non-standard minority that is exposed to social inequality or stigma [7,8]. In such circumstances, out-of-wedlock birth can have a detrimental impact on the child's health throughout the life course through biological, psychosocial and/or behavioural pathways. First, socio-economic adversity can affect adult health through environmental exposures, such as malnutrition or stress during critical periods of growth and development during gestation and childhood. Second, being born or growing up in a non-standard family can cause stress and have negative effects on parent-child interaction, well-being and social relationships. Finally, a stigma associated with out-of-wedlock birth could narrow professional prospects and have effects on health in later life [4,5,[12][13][14].
Aims using the Helsinki Birth Cohort Study (HBCS) data across the life course, we investigated the associations between out-of-wedlock birth and coronary heart disease (CHD) and stroke in adulthood. The study participants were born in Helsinki, Finland, from 1934 to 1944 in a society where marriage was normative. Because the studied period includes World War II, we investigated whether the association between out-of-wedlock birth and CHD and stroke differed between those born before and those born during the war. As a secondary analysis, we performed separate analyses for men and women.

Historical context
In Helsinki, 9% of babies were born out of wedlock from 1934 to 1944 [15]. The nuclear family predominated, and families headed by an unmarried mother were non-standard and formed a disadvantaged, potentially stigmatised group in several ways [7,8]. The infant mortality rate among the infants born out of wedlock was 75/1000 live births compared to 46/1000 live births among babies born to married mothers, and this increased during the war (118 vs. 52 deaths per 1000 live births) [15].
Moreover, children born out of wedlock were overrepresented among child welfare clients and those receiving economic assistance or being placed in out-of-home care [15].
Maternal and child health care developed rapidly during the interwar period [16]. In Helsinki, 97% of all births took place in hospitals in the late 1930s [17]. Attendance at maternal and child welfare clinics was voluntary and free of charge [16,18]. The first maternity grants were provided in 1938 for lowincome mothers.
Almost 70% of HBCS participants were born during World War II, when Finland became involved in the Winter War (November 1939-March 1940 and the so-called Continuation War (june 1941-September 1944), during which half a million men served in the military forces, that is, 65% of all Finnish men aged 20-49 years [19]. Although air raids caused damage and killed people, Finland was never occupied by the enemy army, unlike many other European countries. However, overcrowding, poor hygienic conditions and a shortage of food made everyday life difficult. Accordingly, mothers and small children were prioritised in food rationing [20,21]. The exceptional circumstances that existed during the war may have evened out socio-economic differences that had previously existed between population groups.

Methods
The HBCS includes 13,345 people born in Helsinki in either of the two public maternity hospitals: Helsinki university Women's Hospital or Helsinki City Maternity Hospital. In addition, to be included in the cohort, one had to visit a child welfare clinic in the city and live in Finland in 1971, when a unique identification number was allocated to each resident of the country. The majority of the study participants also attended school in Helsinki. The original cohort has been described elsewhere [22,23]. The study was approved by the Ethics Committee of the National Public Health Institute. Register data were linked with permission from Statistics Finland and the National Institute for Health and Welfare.

Early-life information
Data obtained from hospital birth records included detailed birth measures and maternal characteristics. We defined the mother's marital status as married or unmarried at the time of giving birth based on information from birth records as described [7]. We excluded mothers who were recorded as dead (n=6), widowed (n=22) or divorced (n=17) or who had unclear entries (n=13). We also excluded 32 married mothers with no information on the male caregiver's occupation and six unmarried mothers with no information on the mother's occupation. Our final analyses included 12,612 children born to married mothers and 637 (5.9%) children born out of wedlock. Information on breastfeeding was obtained from child welfare clinic records and dummy coded as follows: not breastfed, breastfed less than three months, breastfed three to six months, breastfed six months or more or no information on breastfeeding available. We also used the mother's highest occupational status which was obtained from birth, childhood welfare clinic and school records [7]. Occupational status was grouped according to the classification system of Statistics Finland [24] and dummy coded as follows: clerical (employer, self-employed, senior and junior clerical), other (housewife, student, unknown profession) and manual worker, with the latter serving as the reference group [7,25]. As a covariate in our analyses, we used information on 1770 study participants who, as children, were evacuated without their parents to temporary foster homes in Sweden or Denmark during World War II [26].

Mortality and morbidity
using personal identification numbers, we followed the subjects through national registers from 1971 to 2014 and identified deaths and hospital admissions from CHD and stroke. We used International Classification of Diseases (ICD) codes 410-414 in ICD-8 and ICD-9 and I21-25 in ICD-10 to define CHD and codes 430-439 in ICD-8 and ICD-9 and I60-69 in ICD-10 to define stroke [27,28]. In Finland, all inpatient hospital admissions are recorded in the Care Register for Health Care. All-cause mortality data were obtained from the National Mortality Register.

Other adulthood information
using the personal identification number, we obtained information from Statistics Finland on the subjects' educational attainment, occupational status, income and marital status at five-year intervals between 1970 and 2000. We used the highest educational attainment, occupational status and personal income level achieved. Participants were split into two groups based on marital status in adulthood: ever married (married, divorced or widowed at any of the time points) and never married (none of these entries at any time points) as described [7,25].

Statistical analyses
We used Cox proportional hazard models as the primary analytical tool with which to examine the associations between maternal marital status at birth and the occurrence of CHD or stroke. Participants were followed from 1971 until the first of these events: death, migration from Finland, hospitalisation for stroke or CHD or reaching 1 january 2014 alive. First, we calculated hazard ratios (HR) separately for two birth periods: those born before World War II (1934)(1935)(1936)(1937)(1938)(1939) and those born during World War II (1940)(1941)(1942)(1943)(1944). Second, we calculated HR for all people born from 1934 to 1944. Men and women were also analysed separately.
The data were analysed with three Cox models. Each Cox model was stratified for year of birth using the Cox regression strata function, and models including both sexes were also stratified for sex. Model 1 was an unadjusted model, with mother's marital status at birth being the only predictor. Model 2 included the following early life factors associated with out-of-wedlock birth: length of gestation, birth weight, birth order, duration of breastfeeding, evacuation abroad during World War II and the mother's age at giving birth and occupational status in childhood. In model 3, we further added the subject's own educational attainment and marital status as adulthood covariates. In addition, in Supplemental Table  SI, we also report the independent effects of covariates that were studied in relation to CHD and stroke outcomes.
We tested potential interactions between sex and mother's marital status, as well as between period of birth and mother's marital status in association with cardiovascular outcomes.
As supplementary analyses, we repeated our previously published analyses related to out-of-wedlock birth and offspring's SEP in adulthood, now by period of birth [7]. In these analyses, we used multinomial logistic regression analysis with two models including similar covariates as in the Cox models. Model 1 included child's year of birth and sex. Model 2 also included length of gestation, birth weight, birth order, duration of breastfeeding, evacuation abroad during World War II and mother's age at birth and occupational status in childhood.

Maternal and childhood characteristics
unmarried mothers were on average younger, shorter and had a lower body mass index in late pregnancy than married mothers (Table I). Regardless of marital status, the proportion of middle-class mothers, rather than mothers working in manual professions, increased during the follow-up, as did the duration of breastfeeding. In general, children born out of wedlock were born at an earlier gestational age, were more often firstborn and were shorter and lighter  than their peers born to married mothers. These differences were independent of birth period.

Cardiovascular health in adulthood based on period of birth
We compared offspring's risk of developing CHD or stroke in later life based on whether they were born before or during World War II. Table II shows HRs and 95% confidence intervals (CIs) for CHD and stroke among children born out of wedlock compared to children born to married mothers according to period of birth and sex. Children born out of wedlock before World War II had a higher risk of developing either CHD or stroke compared to those born to married mothers during the same time period. The interaction between mother's marital status at birth and period of birth was statistically significant for stroke (p=0.015) and for CHD (p=0.003). <3; 3 to 6 and >6 months; no information available), mother's occupational status in childhood (dummy coded to housewives, students and others; employers, self-employed, senior and junior clericals; workers) and information regarding whether subject was evacuated abroad without parents during World War II. Model 3 adjusted as in model 2+highest achieved educational status in adulthood and marital status in adulthood (dummy coded to ever married if subject had an entry indicating subject was married, divorced or widowed at any of the time points or to never married if subject had none of these entries at any time points). CHD: coronary heart disease. Bolded p<0.05.
We performed a secondary analysis separately among men and women. Among men born before World War II, the association was more robust for stroke, while for CHD, the adjustments for childhood SEP attenuated the association to a borderline significant level. Among women, point estimates of HRs were higher for those born before World War II, although they did not reach statistical significance. No associations between out-of-wedlock birth and CHD or stroke in adulthood were seen among those born between 1940 and 1944. We found no interaction between sex and mother's marital status at birth. However, the interaction between mother's marital status at birth and period of birth was statistically significant among men for stroke (p=0.032) and CHD (p=0.005).

Cardiovascular health in adulthood among people born from 1934 to 1944
We found no association between out-of-wedlock birth and CHD or stroke in adulthood among people born from 1934 to 1944 (Table II). univariate effects of covariates with CHD and stroke outcomes are presented in Supplemental Table SI.

SEP in adulthood
As supplementary analyses, we repeated our previously published [7] analyses related to out-of-wedlock birth and SEP in adulthood by period of birth (Tables III and IV). Table IV presents adjusted odds ratios and 95% CIs predicting the highest educational attainment, occupational status and maximum income in thirds in adulthood according to the period of birth and mother's marital status at birth. These novel analyses confirm that in terms of SEP in adulthood, children born out of wedlock fared worse than children born to married mothers in both periods. However, for educational attainment and occupation but not income, the associations were stronger among those born during the pre-war years. lower adult SEP was also independently associated with increased risk of both stroke and CHD in later life (Supplemental Table SII).

Discussion
Our study shows that people born out of wedlock in a peacetime society in which marriage was the parental norm seem to have a lifelong health disadvantage, as illustrated by higher rates of CHD and stroke in adulthood. This association is only partly explained by other childhood socio-economic indicators or by adult SEP or marital status. However, this relative disadvantage was no longer observed during World War II when society faced a severe strain and exceptional circumstances, such as shortage and rationing. These changes seem to have evened out inequalities that existed in peacetime society. We also repeated our previous analyses related to adult SEP among those born out of wedlock based on period of birth. In terms of adulthood education and occupation but not income, the disadvantage of being born to an unmarried mother was again largely limited to those born before World War II. The most disadvantaged group consisted of children born out of wedlock during the pre-war period.
Previous studies suggest that a mother's marital status may involve a different dimension of social stratification regarding her offspring than SEP based on parental education or occupation. According to these studies, out-of-wedlock origin contributes to later life health disparities [7,[8][9][10][11][12]14] and may increase the risk of CVD [8,11]. However, methodological differences prevent precise comparisons between these studies. A study from Sweden consisted of men who were born from 1915 to 1929. According to this study, out-of-wedlock birth was associated with higher ischaemic heart disease mortality in adulthood among men who remained unmarried themselves. However, compared to our cohort, these men were less affected by the war because they were born in a stable society that remained neutral in World War II [8]. Another study by Elo et al. showed that living in a mother-only family in 1950 was associated with higher CVD mortality among men and women born in Finland from 1936 to 1950. However, in that study, period of birth also predicted mortality: CVD mortality was higher for people born before the war (HR=1.47) and during the war (HR=1.19) than after the war, when family type was taken into account [11].
A novel finding of our study is that the increased risk of CHD or stroke was found only among those who were born out of wedlock during the pre-war period. The interaction between out-of-wedlock birth and period of birth was statistically significant both for stroke and CHD, implying that the difference between birth periods is very unlikely to have arisen by chance. It is obvious that the war had a broad impact on society. The disordered conditions and temporary separation precluded, temporarily or permanently, some marriages that would have normally followed a conception. Public attitude towards unmarried mothers may have become more permissive [29]. Narrowing income inequalities, shortages and rationing may have evened out socio-economic differences that existed between population groups before the war. Moreover, evacuation to the countryside to a safer environment and better nutritional care, women's increased contribution to the labour market and maternity grants most likely benefitted mothers from lower social classes. For these reasons, an out-of-wedlock birth during wartime did not represent the same kind of disadvantage as it did in pre-war society.
This study has several strengths. We measured early life characteristics such as mother's marital status, if available, from reliable records. When measuring mother's occupation, we used the highest occupational status during childhood instead of a single measurement from one specific time point. The register data that we used have been validated and found to be accurate for epidemiological studies. However, some limitations to the present study must also be acknowledged. Data on cohabitations without marriage, other family members or children's place of residence during the war were not available. Occupational data were classified based on modern standards and do not differentiate between skilled and unskilled workers. These data limitations could be expected to increase random error and result in a more conservative estimate. Moreover, the proportion of out-of-wedlock children in our study was lower than in Helsinki during the time period in question. This would again be expected to lead to a more conservative estimate, that is, a more favourable outcome for those born out of wedlock. Due to high infant mortality among children born out of wedlock, many died before our follow-up. This could in theory diminish the difference between those born out of wedlock and those born to a married mother during World War II, but we believe it is unlikely to cause a large enough bias to explain our results.
Finally, even though no direct analogy can be drawn between contemporary and early society in Helsinki, a large number of the people at a higher risk of CVD currently, much like our study participants, are those who are born into societies in which marriage remains the parental norm. Furthermore, from a public health perspective, modern societies may contain minority groups of children experiencing social and economic hardship or stigma with lifelong health effects, similar to those of children born out of wedlock 80 years ago.

Conclusions
Our study shows an association between birth out of wedlock and CVD in adulthood among people born before World War II. These results indicate that in a society in which marriage is the parental norm, the mother's marital status may be an important indicator of early life disadvantage, and that this may vary rapidly when societal circumstances change. Together with other measures, such as parental occupational status, it reflects unfavourable early life circumstances from an alternative viewpoint. The disadvantage that children of unmarried mothers experienced 80 years ago is likely to persist throughout their entire lives.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/ or publication of this article.