Exposure to childhood out-of-home care (foster family and residential care) is associated with an increased risk of ill-health and disability in adulthood, but the risk for cardiovascular disease has not previously been studied longitudinally.
This was a national cohort study generated from linkage of a range of population-based registers, resulting in a national cohort of 881 731 of whom 26 310 (3.0%) had a history of out-of-home care. The study population, born 1972 to 1981, was followed from age 18 to age 39 to 48 years for hospitalizations and death.
After adjusting for year of birth and maternal education, individuals with a history of childhood out-of-home-care experienced a doubling of the risk for coronary disease (hazard ratio; 95% confidence interval: 2.05; 1.74–2.41) and stroke (hazard ratio 1.85; 1.59–2.15), compared with the general population, with similar estimates for men and women. Women with a history of out-of-home care had a more than doubled risk for cigarette smoking in early pregnancy, with a relative risk of 2.26; (2.18–2.34) and a moderately increased risk for gestational diabetes relative risk 1.49 (1.19–1.86). There was marked attenuation (40% to 90%) in effect estimates for disease and risk factors after further control for cohort members educational achievement at age 15–16 years.
A history of childhood out-of-home care was associated with a doubled risk of early cardiovascular disease events. Cigarette smoking and educational underachievement were the main identified risk factors.
Exposure to out-of-home care in childhood is associated with increased risk of ill-health and disability in middle age, but the contribution of cardiovascular disease to this risk is unknown.
Out-of-home care in childhood was associated with a doubling of the risk of coronary disease and stroke at age 18 to 48 years. Educational underachievement and cigarette smoking were identified as the main risk factors to be targeted in prevention.
A history of out-of-home care (foster family and residential care) in childhood has been shown to be associated with an increased risk of ill-health and disability in adulthood.1,2 A recent meta-analysis of prospective cohort studies, for instance, found that people with a history of childhood care experienced a doubling in the rate of total mortality compared with same-aged, unexposed peers.3
Swedish child welfare is a combination of child protection, provision of family services, and interventions targeting juvenile delinquents. The latter group fall under the jurisdiction of social services in the Scandinavian countries.4 Historically, Swedish policymakers and child welfare professionals have had strong confidence in assumed long-term compensatory effects of out-of-home care. However, a host of national cohort studies, and in addition a few sibling studies, have by and large falsified this claim showing poor health and educational outcomes in adulthood.5,6 A longitudinal study in Swedish adults born in the 1950s with a history of out-of-home care demonstrated increased all-cause mortality rates between ages 20 and 56 years, with a particularly pronounced risk among those who were placed in care in adolescence.7
Multiple studies have shown associations between exposure to adverse childhood experiences and poor health outcomes and an unhealthy lifestyle in adulthood, particularly psychiatric disorders8,9 and use of substances and tobacco,10,11 and ultimately premature mortality.12 Recent studies have suggested that exposure to chronic stress in childhood may influence health later in the life course through changes in the nervous, endocrine, and immune systems.13 Thus, exposure to adverse childhood experiences may influence later health through physiologic mechanisms as well as persistence of damaging health behaviors.
The physiologic strain of pregnancy may reveal underlying predispositions to disease that would otherwise have remained hidden for a long time. Pregnancy complication such as gestational hypertension, preeclampsia, gestational diabetes, placental abruption, spontaneous preterm birth, stillbirth, and miscarriage have all been associated with the development of chronic disease later in life.14 Some of these, like gestational diabetes and gestational hypertension, have been shown to be associated with an increased risk of cardiovascular disease later in life.15,16
Although there are a few studies linking adverse childhood experiences to cardiovascular disease,10 these studies have been cross-sectional with retrospective report of childhood exposures.17 In the current study we used data from Swedish national registers to investigate cardiovascular disease in middle age with a history of out-of-home care in a longitudinal design. We also investigated educational achievement as a risk factor for the development of cardiovascular disease, and the risk of hypertension, diabetes and cigarette smoking during pregnancy for the women in the cohort who had given birth.
Methods
This study is based on information from the Swedish national registers, containing high quality data.18 These registers are based on the unique personal identity number assigned to all Swedish residents at birth (or time of immigration), which allows for linkage of data from different registers with practically no attrition.18 The study was approved by the ethics committee in the Stockholm region in 2020 (No. 2020–00250).
The study population comprised 457 055 men and 424 676 women born in Sweden between 1972 and 1981, who were alive and resident in the country on their 18th birthday and had no record of emigration according to the Register of the Total Population.19 Birth mothers of these individuals were identified in the Multigeneration Register.20
Exposures
The study population was linked to records of out-of-home care via the Swedish Child Welfare Intervention Register.21 Any care recorded before the age of 18 denoted exposure to out-of-home care. There were 13 389 men and 12 462 women who fulfilled these criteria. Accumulated time in out-of-home care and age at first entry into care were calculated with dates of start and termination of episodes in out-of-home care in the register.
Outcomes
Coronary disease and stroke were identified in inpatient care in the Patient Register22 and in deaths in the Causes of Death Register.23 Cardiovascular disease (CVD) was coded according to the International Classification of Diseases (ICD)24,25 as myocardial infarction, other coronary disease, or stroke based on main or complimentary diagnoses and underlying causes of death. For coding see Supplemental Table 6. The study population was followed in these 2 registers from their 18th birthday until December 31, 2020, when they were 39 to 48 years old.
Indicators of Pregnancy Risk Factors for CVD
Women in the study population were linked to their first birth recorded in the Swedish Medical Birth Register from 1986 to 2017. This register includes 97% of all live and still-born births in Sweden.26 Based on the information in this register, we created an indicator of cigarette smoking based on report to midwives during the first visit to the maternity clinic in early pregnancy. During pregnancy urine glucose and blood pressure are routinely measured by midwives in maternal health care.27 If the woman is found to fulfill the criteria for hypertension or diabetes, these diagnoses are reported to the medical birth register as gestational hypertension, pregestational hypertension, gestational diabetes, and pregestational diabetes (Supplemental Table 7).
Educational Achievement
Data on average grades for both men and women, during the final compulsory school year in Sweden (age 15–16), was extracted from The National School Register.28 Previous Swedish studies have shown this indicator to be an effective predictor of social adjustment socioeconomic and health status in adulthood.1,28 The average grade point was categorized into quintiles, with quintile 1 denoting the lowest scoring group. A missing value in the register was coded in quintile 1 since this indicated not having completed the compulsory education or having completed it within a special educational unit for disabled children.
Socioeconomic Covariates
The highest achieved education for the birth mother was retrieved from the Longitudinal Integration Database for Health Insurance and Labor Market Studies in 1990.29
Statistical Analysis
Cox proportional hazard regression was used to analyze the association between out-of-home care experience and CVD. Cox regression is a regression model used for a time-to-event outcome involving 2 components: whether an event has occurred or not and the time leading up to the event (time at risk). Associations between independent variables and outcome are expressed as hazard ratios (HR). A HR >1 indicates an increased risk.
Age-adjusted HRs with 95% confidence intervals were estimated for any coronary disease and stroke (defined above) as the 2 outcome variables. The proportional hazards assumption was confirmed in Kaplan-Meier tables for both outcomes. Time at risk was calculated from the 18th birthday until the first hospital admission, date of death, or December 31, 2020, whichever came first. All hazard ratios were adjusted for year of birth. Maternal education was considered a confounder, reflecting the socioeconomic status of the birth family, whereas own school performance at age 15 to 16 was considered a proxy for socioeconomic status in adulthood and thus a potential mediator.
Cox regression with a constant time variable provides a robust proxy for likelihood for both high and low frequency variables.30 This analysis was used to compute relative risk ratios (RR) and 95% confidence intervals for the risk of the risk factors during pregnancy in the out-of-home care study group compared with the general population. These analyses were adjusted for maternal age and year of the birth of the first child as continuous variables to control for the secular trends and the maternal age patterns of these pregnancy indicators. As in the analyses described above, the analysis was further adjusted for maternal education as a confounder and own school performance as a potential mediator. Statistical analysis was conducted using SPSS (IBM SPSS Statistics version 27.0, [SPSS, Inc, IBM Corp, Armonk, NY, USA]).
Data Access
Because of Swedish legislation, we cannot share our register-based data with second parties.
Results
There were 881 731 individuals that fulfilled the criteria of the study, of whom 26 310 had a history of out-of-home care before the age of 18. Socio-demographic information about the study groups is presented in Table 1. Individuals with a history of out-of-home care were more likely to have a mother with low educational level and themselves a low grade average at age 15 to 16 compared with the general population.
Socio-Demographic Characteristics of the Study Population
. | Women . | Men . | ||
---|---|---|---|---|
. | Out-of-home Care, % . | General Population, % . | Out-of-home Care, % . | General Population, % . |
. | N = 12 711 . | N = 411 965 . | N = 13 599 . | N = 443 456 . |
Year of birth | ||||
1972–1974 | 33.0 | 32.9 | 34.3 | 33.0 |
1975–1977 | 28.7 | 29.6 | 28.2 | 29.6 |
1978–1981 | 38.2 | 37.5 | 37.6 | 37.5 |
Maternal education (1990) | ||||
≤9 y | 32.5 | 15.7 | 32.9 | 15.6 |
10–12 y | 26.6 | 19.8 | 25.6 | 19.7 |
13–15 y | 32.7 | 39.7 | 33.3 | 39.6 |
16+ y | 8.3 | 24.9 | 8.2 | 25.1 |
Own educational performance (15–16 y) | ||||
First quintile (lowest) | 54.3 | 15.4 | 69.1 | 27.5 |
Second | 17.1 | 14.2 | 14.0 | 18.6 |
Third | 12.5 | 19.3 | 8.7 | 19.8 |
Fourth | 10.8 | 27.3 | 6.0 | 21.4 |
Fifth | 5.2 | 23.8 | 2.2 | 12.8 |
. | Women . | Men . | ||
---|---|---|---|---|
. | Out-of-home Care, % . | General Population, % . | Out-of-home Care, % . | General Population, % . |
. | N = 12 711 . | N = 411 965 . | N = 13 599 . | N = 443 456 . |
Year of birth | ||||
1972–1974 | 33.0 | 32.9 | 34.3 | 33.0 |
1975–1977 | 28.7 | 29.6 | 28.2 | 29.6 |
1978–1981 | 38.2 | 37.5 | 37.6 | 37.5 |
Maternal education (1990) | ||||
≤9 y | 32.5 | 15.7 | 32.9 | 15.6 |
10–12 y | 26.6 | 19.8 | 25.6 | 19.7 |
13–15 y | 32.7 | 39.7 | 33.3 | 39.6 |
16+ y | 8.3 | 24.9 | 8.2 | 25.1 |
Own educational performance (15–16 y) | ||||
First quintile (lowest) | 54.3 | 15.4 | 69.1 | 27.5 |
Second | 17.1 | 14.2 | 14.0 | 18.6 |
Third | 12.5 | 19.3 | 8.7 | 19.8 |
Fourth | 10.8 | 27.3 | 6.0 | 21.4 |
Fifth | 5.2 | 23.8 | 2.2 | 12.8 |
A maximum of 31 years of mortality and hospital surveillance in the registers gave rise to 5519 CVD events. The incidence of having had any CVD event was highest in men with a history of out-of-home care, 15.3 of 1000 compared with 7.5 of 1000 in the general population (Table 2). Women had lower incidences ranging from 9.6 of 1000 in those with a history of out of home care to 4.8 to 1000 in the general population. The patterns were similar for stroke and coronary disease (Table 2).
Incidence of Cardiovascular Diseases by Study Group
. | Out-of-home Care (N = 12 711) . | General Population (N = 411 965) . |
---|---|---|
Women | 1/1000 | 1/1000 |
Any cardiovascular disease | 9.6 | 4.8 |
Cardiovascular death | 1.1 | 0.2 |
Any coronary disease | 3.6 | 1.6 |
Myocardial infarction | 2.0 | 1.1 |
Other coronary disease | 2.3 | 1.0 |
Stroke | 6.1 | 3.3 |
Men | (N = 13 599) | (N = 443 456) |
Any cardiovascular disease | 15.3 | 7.5 |
Cardiovascular death | 1.0 | 0.6 |
Any coronary disease | 8.3 | 3.8 |
Myocardial infarction | 5.5 | 2.7 |
Other coronary disease | 4.8 | 2.2 |
Stroke | 7.5 | 3.9 |
. | Out-of-home Care (N = 12 711) . | General Population (N = 411 965) . |
---|---|---|
Women | 1/1000 | 1/1000 |
Any cardiovascular disease | 9.6 | 4.8 |
Cardiovascular death | 1.1 | 0.2 |
Any coronary disease | 3.6 | 1.6 |
Myocardial infarction | 2.0 | 1.1 |
Other coronary disease | 2.3 | 1.0 |
Stroke | 6.1 | 3.3 |
Men | (N = 13 599) | (N = 443 456) |
Any cardiovascular disease | 15.3 | 7.5 |
Cardiovascular death | 1.0 | 0.6 |
Any coronary disease | 8.3 | 3.8 |
Myocardial infarction | 5.5 | 2.7 |
Other coronary disease | 4.8 | 2.2 |
Stroke | 7.5 | 3.9 |
Cox Regression Models of Cardiovascular Disease
Table 3 presents the Cox regression analysis of stroke and coronary disease. The HRs associated with out-of-home care were 1.95 (1.67–2.26) for stroke and 2.27 (1.93–2.67) for coronary disease compared with the general population in Model 1, adjusted for age and gender only. Adjusting for maternal education as a proxy for childhood family socioeconomic status in Model 2 attenuated these HRs slightly to 1.85 (1.59–2.15) and 2.05 (1.74–2.41) respectively, whereas further adjusting for own educational achievement at age 15 to 16 attenuated the HRs considerably to 1.47 (1.26–1.72) and 1.56 (1.33–1.85).
Hazard Ratios of Cardiovascular Disease
. | Stroke . | Coronary Disease . | ||||
---|---|---|---|---|---|---|
. | Model 1 . | Model 2 . | Model 3 . | Model 1 . | Model 2 . | Model 3 . |
. | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . |
Study groups | ||||||
General population | 1 | 1 | 1 | 1 | 1 | 1 |
Out-of-home care | 1.95 (1.67–2.26) | 1.85 (1.59–2.15) | 1.47 (1.26–1.72) | 2.27 (1.93–2.67) | 2.05 (1.74–2.41) | 1.56 (1.33–1.85) |
Maternal education | ||||||
0–9 y | — | 1.37 (1.23–1.53) | 1.11 (0.99–1.25) | — | 1.37 (1.23–1.53) | 1.49 (1.30–1.70) |
10–12 y | — | 1.36 (1.22–1.51) | 1.11 (0.99–1.24) | — | 1.36 (1.22–1.51) | 1.43 (1.25–1.63) |
13–14 y | — | 1.27 (1.15–1.40) | 1.10 (1.00–1.22) | — | 1.27 (1.15–1.40) | 1.25 (1.11–1.42) |
15+ y | — | 1 | 1 | — | 1 | 1 |
Own educational performance (15–16 y) | ||||||
Quintile 1 (Lowest) | — | — | 2.24 (1.98–2.53) | — | — | 2.89 (2.47–3.39) |
Quintile 2 | — | — | 1.54 (1.35–1.76) | — | — | 1.91 (1.61–1.26) |
Quintile 3 | — | — | 1.29 (1.13–1.47) | — | — | 1.44 (1.21–1.71) |
Quintile 4 | — | — | 1.11 (0.98–1.27) | — | — | 1.17 (0.98–1.39) |
Quintile 5 | — | — | 1 | — | — | 1 |
. | Stroke . | Coronary Disease . | ||||
---|---|---|---|---|---|---|
. | Model 1 . | Model 2 . | Model 3 . | Model 1 . | Model 2 . | Model 3 . |
. | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . | HR (95% CI) . |
Study groups | ||||||
General population | 1 | 1 | 1 | 1 | 1 | 1 |
Out-of-home care | 1.95 (1.67–2.26) | 1.85 (1.59–2.15) | 1.47 (1.26–1.72) | 2.27 (1.93–2.67) | 2.05 (1.74–2.41) | 1.56 (1.33–1.85) |
Maternal education | ||||||
0–9 y | — | 1.37 (1.23–1.53) | 1.11 (0.99–1.25) | — | 1.37 (1.23–1.53) | 1.49 (1.30–1.70) |
10–12 y | — | 1.36 (1.22–1.51) | 1.11 (0.99–1.24) | — | 1.36 (1.22–1.51) | 1.43 (1.25–1.63) |
13–14 y | — | 1.27 (1.15–1.40) | 1.10 (1.00–1.22) | — | 1.27 (1.15–1.40) | 1.25 (1.11–1.42) |
15+ y | — | 1 | 1 | — | 1 | 1 |
Own educational performance (15–16 y) | ||||||
Quintile 1 (Lowest) | — | — | 2.24 (1.98–2.53) | — | — | 2.89 (2.47–3.39) |
Quintile 2 | — | — | 1.54 (1.35–1.76) | — | — | 1.91 (1.61–1.26) |
Quintile 3 | — | — | 1.29 (1.13–1.47) | — | — | 1.44 (1.21–1.71) |
Quintile 4 | — | — | 1.11 (0.98–1.27) | — | — | 1.17 (0.98–1.39) |
Quintile 5 | — | — | 1 | — | — | 1 |
Adjusted for gender and year of birth. —, variable not in Model.
In an analysis of low educational achievement as a risk factor for having any cardiovascular disease within the out-of-home care study group, the HR was 2.27 (1.99–2.58) for average grade in the lowest quintile at age 15 to 16 compared with HR 1.31 (1.06–1.62) for grade average in quintile 2 to 5 relative to the general population after adjustment for maternal education, gender, and age.
Men and women with a history of out-of-home care had similar HRs in interaction analyses in Model 1 with P values of interaction effects of 0.96 for coronary disease and 0.71 for stroke. See Supplemental Tables 7 and 8 for gender stratified models.
Cardiovascular Risk Factors During Pregnancy
Of the women in the cohort, 376 369 were identified in the Medical Birth Register after having given birth, whereas 10 539 had been exposed to out-of-home care (82.9% of all women exposed to out-of-home care) and 365 830 (88.8%) in the general population. Cigarette smoking during early pregnancy was associated with both stroke and coronary diseases with HRs between 1.9 to 3.0. Gestational and pregestational hypertension were associated with stroke with HRs of 1.8 to 3.6 and coronary disease with HRs of 3.2 to 8.6. Gestational and pregestational diabetes were associated with coronary disease only with HRs of 4.5–24.6 (Supplemental Table 9). Cigarette smoking during early pregnancy was most common in the youngest mothers and decreased over time (Supplemental Table 10).
Mothers with a history of out-of-home care were younger when giving birth, with 14.2% giving birth already as teenagers compared with 2.8% in the general population. Cigarette smoking during early pregnancy was far more common among women with a history of out-of-home care with 34.8% being reported to smoke at least 1 cigarette a day, compared with only 7.9% in the general population. In the out-of-home-care study group, the prevalence of hypertension and diabetes before the pregnancy was similar in women exposed to out-of-home care and the general population. Gestational hypertension, including eclampsia and pre-eclampsia was less common in the out-of-home care group (P = .01) compared with the general population, a difference that was no longer significant after adjustment for cigarette smoking (P = .33). The prevalence of gestational diabetes was 0.8% in the out-of-home care study group compared with 0.6% in the general population (Table 4).
Gestational Risk Factors for Cardiovascular Disease by Exposure to Out-of-home Care in Childhood
. | Out-of-home Care (N = 10 539), % . | General Population (N = 365 830), % . | P . |
---|---|---|---|
Age at first birth | |||
12–19 | 14.8 | 2.8 | <.001a |
20–24 | 35.8 | 17.3 | |
25–29 | 25.4 | 34.8 | |
30–34 | 17.0 | 34.3 | |
35+ | 6.6 | 11.0 | |
Smoking in early pregnancy | |||
No | 59.8 | 87.0 | <.001a |
1–9 cigarettes per day | 23.8 | 6.1 | |
10+ cigarettes per day | 11.0 | 1.8 | |
Missing data | 5.4 | 5.2 | |
Metabolic risk factors | |||
Prepregnancy hypertension | 0.2 | 0.3 | .777a |
Pregnancy hypertension | 4.4 | 5.7 | .014,a .334b |
Pregestational diabetes | 0.5 | 0.6 | .508a |
Gestational diabetes | 0.8 | 0.6 | <.001a |
. | Out-of-home Care (N = 10 539), % . | General Population (N = 365 830), % . | P . |
---|---|---|---|
Age at first birth | |||
12–19 | 14.8 | 2.8 | <.001a |
20–24 | 35.8 | 17.3 | |
25–29 | 25.4 | 34.8 | |
30–34 | 17.0 | 34.3 | |
35+ | 6.6 | 11.0 | |
Smoking in early pregnancy | |||
No | 59.8 | 87.0 | <.001a |
1–9 cigarettes per day | 23.8 | 6.1 | |
10+ cigarettes per day | 11.0 | 1.8 | |
Missing data | 5.4 | 5.2 | |
Metabolic risk factors | |||
Prepregnancy hypertension | 0.2 | 0.3 | .777a |
Pregnancy hypertension | 4.4 | 5.7 | .014,a .334b |
Pregestational diabetes | 0.5 | 0.6 | .508a |
Gestational diabetes | 0.8 | 0.6 | <.001a |
Adjusted for maternal age and year of giving birth.
Adjusted for maternal age, year of giving birth and smoking habit.
Table 5 shows the multivariate analyses of the pregnancy risk factors cigarette smoking and gestational diabetes, with the analysis of cigarette smoking in early pregnancy, excluding the 5.2% with missing information. After adjusting for maternal age and year of giving birth, individuals with a history of out-of-home care were more likely to smoke; RR 2.35 (2.27–2.34) compared with the general population. Adjusting for maternal education attenuated this estimate slightly to RR 2.26 (2.18–2.34), whereas adjusting for own educational achievement at age 15 to 16 attenuated this estimate greatly to RR 1.65 (1.59–1.71).
Relative Risk of Cardiovascular Risk Factors
. | Cigarette Smoking (N = 362 605) . | Gestational Diabetes (N = 382 468) . | ||||
---|---|---|---|---|---|---|
. | Model 1 . | Model 2 . | Model 3 . | Model 1 . | Model 2 . | Model 3 . |
. | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . |
General population | 1 | 1 | 1 | |||
Out-of-home care | 2.35 (2.27–2.43) | 2.26 (2.18–2.34) | 1.65 (1.59–1.71) | 1.57 (1.27–1.95) | 1.49 (1.19–1.86) | 1.19 (0.95–1.49) |
Maternal education | — | |||||
0–9 y | — | 1.37 (1.23–1.53) | 1.11 (0.99–1.25) | — | 2.01 (1.75–2.31) | 1.68 (1.46–1.94) |
10–12 y | — | 1.36 (1.22–1.51) | 1.11 (0.99–1.24) | — | 1.61 (1.40–1.84) | 1.35 (1.18–1.55) |
13–14 y | — | 1.27 (1.15–1.40) | 1.10 (1.00–1.22) | — | 1.44 (1.30–1.60) | 1.29 (1.16–1.44) |
15+ y | — | 1 | 1 | — | 1 | 1 |
School grades at age 15–16 y | ||||||
Quintile 1 (lowest) | — | — | 2.24 (1.98–2.53) | — | — | 2.43 (2.12–2.79) |
Quintile 2 | — | — | 1.54 (1.35–1.76) | — | — | 1.82 (1.57–2.10) |
Quintile 3 | — | — | 1.29 (1.13–1.47) | — | — | 1.52 (1.33–1.74) |
Quintile 4 | — | — | 1.11 (0.98–1.27) | — | — | 1.31 (1.15–1.49) |
Quintile 5 | — | — | 1 | — | — | 1 |
. | Cigarette Smoking (N = 362 605) . | Gestational Diabetes (N = 382 468) . | ||||
---|---|---|---|---|---|---|
. | Model 1 . | Model 2 . | Model 3 . | Model 1 . | Model 2 . | Model 3 . |
. | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . | RR (95% CI) . |
General population | 1 | 1 | 1 | |||
Out-of-home care | 2.35 (2.27–2.43) | 2.26 (2.18–2.34) | 1.65 (1.59–1.71) | 1.57 (1.27–1.95) | 1.49 (1.19–1.86) | 1.19 (0.95–1.49) |
Maternal education | — | |||||
0–9 y | — | 1.37 (1.23–1.53) | 1.11 (0.99–1.25) | — | 2.01 (1.75–2.31) | 1.68 (1.46–1.94) |
10–12 y | — | 1.36 (1.22–1.51) | 1.11 (0.99–1.24) | — | 1.61 (1.40–1.84) | 1.35 (1.18–1.55) |
13–14 y | — | 1.27 (1.15–1.40) | 1.10 (1.00–1.22) | — | 1.44 (1.30–1.60) | 1.29 (1.16–1.44) |
15+ y | — | 1 | 1 | — | 1 | 1 |
School grades at age 15–16 y | ||||||
Quintile 1 (lowest) | — | — | 2.24 (1.98–2.53) | — | — | 2.43 (2.12–2.79) |
Quintile 2 | — | — | 1.54 (1.35–1.76) | — | — | 1.82 (1.57–2.10) |
Quintile 3 | — | — | 1.29 (1.13–1.47) | — | — | 1.52 (1.33–1.74) |
Quintile 4 | — | — | 1.11 (0.98–1.27) | — | — | 1.31 (1.15–1.49) |
Quintile 5 | — | — | 1 | — | — | 1 |
Adjusted for maternal age and year of first birth. —, variable not in Model.
After adjusting for maternal age and year of giving birth, individuals with a history of out-of-home care were more likely to have gestational diabetes, RR 1.57 (1.27–1.95) compared with the general population. Adjusting further for maternal education attenuated this estimate slightly to RR 1.49 (1.19–1.86), whereas adjusting further for own educational achievement attenuated this estimate to RR 1.19 (0.95–1.49), see Table 5.
Risk Factors Associated With the Out-of-home Care
The risk of having had any CVD was higher in those who entered out-of-home care late (at 11–18 years), HR 2.31 (1.97–2.70) compared with those who entered care at age 1 to 10 years, HR 1.84 (1.58–2.15), relative to the general population, after adjustment for maternal education and year of birth, whereas there was no association with CVD and time in care (Supplemental Table 11). Of women who had entered out-of-home care after age 11 years, 43.2% reported cigarette smoking compared with 30.8% in those who entered care at a younger, adjusted HR 1.27 (1.18–1.36), whereas there was no association between age at entry into care and gestational diabetes. Care leavers with a cumulated time of more than 2 years in care had a slightly higher risk of cigarette smoking than those with shorter durations of care (Supplemental Table 12).
Discussion
In this study in a Swedish national cohort of 881 731 individuals, we investigated the risk of cardiovascular events up to the age of 48 years associated with exposure to out-of-home care during childhood. We found a doubled risk of both stroke and coronary disease. In women who had given birth, we found a strong association between cigarette smoking in early pregnancy and having been exposed to out-of-home care and a more modest association with gestational diabetes. Poor educational achievement was a strong risk factor for CVD within the out-of-home care study group for both genders and attenuated the associations of cigarette smoking and gestational diabetes with CVD for women exposed to out-of-home care.
To the best of our knowledge, this is the first study to have investigated the occurrence of cardiovascular disease events in adults exposed to out-of-home care after World War II. Our results indicate that CVD contributes to the increased rate of ill-health and disability previously described in adults with a history of out-of-home care.3,31 Furthermore, they suggest that an unhealthy lifestyle, as indicated by cigarette smoking, and lower educational achievement were important risk factors for CVD among care leavers.
The finding of a modestly higher risk of gestational diabetes in the care leaves gives some support to the hypothesis suggested by Brown32 of a metabolic pathway for the multiple risks associated with out-of-home care. Somewhat unexpected, gestational hypertension had a lower prevalence in care leavers. However, this is in line with previous studies that have demonstrated that cigarette smokers have a lower risk of pre-eclampsia and eclampsia during pregnancy.33
The prevalence of cigarette smoking found among women with a history of out-of-home care in this study was 2 to threefold higher than that in the general population, a differential found also among care leavers in studies from the United Kingdom10 and the United States.34 A similar risk increase has been described also for substance abuse in this cohort of adults with a history of out-of-home care in Sweden. As for the cigarette smoking results in the current study, substance abuse in care leavers in that study was predicted by low educational achievement.28
Low educational achievement, as measured by school marks at age 16, was related to a higher risk of CVD in men and women and cigarette smoking in women with a history of out-of-home care. This adds to the array of negative outcomes to be associated with low educational achievement among Swedish care leavers, which also includes public welfare dependency, severe criminality, substance abuse, disability pension, mental ill health, suicidality, and all-cause mortality.35 –37 Swedish studies of children in out-of-home care have shown a systematic educational underachievement of educational achievement in relation to their cognitive potential.35 Thus, there is a window of improvement for educational achievement with better support for this vulnerable child population, which could be exploited for prevention.28
Our study found a slightly higher risk of CVD and cigarette smoking during early pregnancy in care leavers who were 11 years or older when entering care. This is congruent with other Swedish studies that have indicated particularly pronounced negative social and health outcomes among adult care leavers who were placed in adolescence.7,38 It seems credible that this to some extent is related to behavioral problems coupled with a destructive lifestyle (eg, substance abuse) being a more common reason for placement in adolescence than in those placed at a younger age.7
This study used multiple high quality Swedish national registers to create the large cohort necessary to study a low frequency outcome like CVD in people who are under 48 years of age. A particular strength with a register design relative to field-based studies is the minimal and unselected attrition.
An obvious limitation of this study is that the metabolic risk factors for CVD and the indicator of cigarette smoking only included women, and among women only those who had given birth. Further studies of metabolic risk factors and cigarette smoking for CVD in adults with a history of out-of-home should also include men.
Conclusions
This study shows a doubled risk for CVD in middle age associated with out-of-home care with low educational achievement and cigarette smoking as the main identified risk factors. Thus, a public health policy for children in out-of-home care should pay particular attention to education. Children’s academic progress should be monitored and appropriate educational supports instigated while in care.39 Efforts should also be directed toward smoking prevention for children in care during their teens,40 an age when initiation of cigarette smoking peaks.4 Further studies are needed to clarify whether metabolic pathways contribute to the association between childhood adversity and cardiovascular disease in adults.
Dr Brännström initiated and funded the study; Dr Hjern created the dataset from multiple register sources, analyzed the data, and wrote the first draft of the article; and all authors interpreted the results, revised the manuscript, and approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: Swedish Research Council for Health, Working Life and Welfare (grant 2019-00057). The Swedish Research Council for Health, Working Life and Welfare had no role in the design and conduct of the study.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
Comments