International Journal of Cardiology 190 (2015) 302–307
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Family history of premature myocardial infarction, life course socioeconomic position and coronary heart disease mortality — A Cohort of Norway (CONOR) study Bendik S. Fiskå a,⁎, Inger Ariansen b, Sidsel Graff-Iversen b, Grethe S. Tell c,d, Grace M. Egeland c,d, Øyvind Næss a,b a
University of Oslo, Faculty of Medicine, PB 1078 Blindern, 0316 Oslo, Norway Division of Epidemiology, National Institute of Public Health, PB 4404, Nydalen, 0403 Oslo, Norway c Department of Global Public Health and Primary Care, University of Bergen, PB 7804, 5018 Bergen, Norway d Department of Health Registries, Norwegian Institute of Public Health, Kalfarveien 31, 5018 Bergen, Norway b
a r t i c l e
i n f o
Article history: Received 27 March 2015 Accepted 18 April 2015 Available online 22 April 2015 Keywords: Coronary heart disease Myocardial infarction Family history Life course socioeconomic position Coronary heart disease mortality
a b s t r a c t Background/objectives: To investigate self-reported family history (FH) of premature myocardial infarction (MI) in first-degree relatives as a risk factor for coronary heart disease (CHD) mortality, and assess whether any observed effect could be explained by current or life course socioeconomic position. Methods: 130,066 participants from Cohort of Norway were examined during 1994–2003. A subgroup (n = 84,631) had additional life course socioeconomic data. Using Cox proportional hazard analyses, we calculated hazard ratios (HR) for CHD mortality, assessed by linkages to the Norwegian Cause of Death Registry through 2009. For subgroup analyses, we created an index of life course socioeconomic position, and assessed its role as a potential confounder in the association of FH with CHD. Results: For men, MI in parents and siblings were both a significant risk factor for CHD mortality after adjusting for established risk factors and current socioeconomic conditions; the highest risk was with MI in siblings (HR: 1.44 [1.19–1.75]). For women, FH constituted significant risk after similar adjustment only for those with MI in parents plus siblings (HR: 1.78 [1.16–2.73]). Adjusting for current and life course socioeconomic conditions only marginally lowered the estimates, and those with FH did not have worse life course socioeconomic position than those without. Conclusions: FH of premature MI is an independent risk factor for CHD mortality that differs in magnitude of effect by the sex of the index person and type of familial relationship. Life course socioeconomic position has little impact on the association between FH and CHD, suggesting the effect is not confounded by this. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Background/objectives A family history (FH) of coronary heart disease (CHD) in first–degree relatives is an established risk factor for CHD [1–12]; however, the true magnitude and pattern of this risk association remains unknown, and concerns have been raised regarding whether the impact of a positive FH is independent from other risk factors [7]. Risk of CHD is associated with socioeconomic position [13–17], and it is possible that a positive FH could be of socioeconomic, rather than biological, origin. Only a few studies on the effect of FH have controlled for socioeconomic position, using single indicators, such as education or occupation [1,12], or a score emphasizing psychological stress [5]. Public health researchers have increasingly emphasized the need to take the full life course into account to understand the origin and population distribution of diseases such as CHD [18,19]. Merely adjusting for single ⁎ Corresponding author. E-mail address: b.s.fi[email protected] (B.S. Fiskå).
http://dx.doi.org/10.1016/j.ijcard.2015.04.160 0167-5273/© 2015 Elsevier Ireland Ltd. All rights reserved.
measures at single points in time, may lead to estimates prone to residual confounding [20]. The extent of this potential residual confounding has not previously been investigated for FH. The aim of this study was to investigate self-reported FH of premature MI in parents and/or siblings as a risk factor for CHD mortality in the index patient, and evaluate whether an effect of a positive FH could be explained by established risk factors, as well as indicators of current and life course socioeconomic position. 2. Methods 2.1. Study population Cohort of Norway (CONOR) is a compilation of ten regional, both urban and rural, community-based health surveys conducted in the period 1994–2003 [21]. Subjects were invited to a baseline health examination which included anthropometric measures, blood sampling and questionnaires. The overall response rate was 58.3%. In eight of the ten
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
surveys (n = 164,432), history of premature myocardial infarction (MI) in family members was surveyed. The study population consisted of participants for whom we had complete covariate information (n = 130,066). Data from the Norwegian censuses of 1960, 1970 and 1980 were available for 84,635 participants, who were included in subgroup analyses of life course socioeconomic conditions. The mean age at time of first census was 15.1 years (SD = 12.5), and our data allowed us to assess socioeconomic position over a period of 34–43 years, depending on time of recruitment to CONOR. A flow-chart presenting the inclusion process is provided in Supplementary Fig. 1. 2.2. Exposure Exposure was self-reported FH of premature MI in first-degree relatives. Using a standardized questionnaire, the respondents were asked to indicate the relatives who have or have had MI before age 60, with the options being mother, father, brother, sister and child. Information on MI in children was not included in our analyses. We created three mutually exclusive categories: FH of premature MI in one or both parents only, one or more siblings only, and one or both parents plus one or more siblings. 2.3. Risk factors In addition to age and sex, the following biological markers recorded at the baseline health examination were considered as relevant risk- or protective factors: systolic and diastolic blood pressure (mm Hg), BMI (kg/m2) and non-fasting serum triglycerides (mmol/l), total cholesterol (mmol/l) and HDL cholesterol (mmol/l). Detailed information about sampling methods is provided elsewhere [21]. Additionally, selfreported information from the CONOR questionnaire on current smoking (yes/no), diabetes (yes/no) and past year level (hours per week) of hard physical activity (causing sweating or breathlessness) was included. To assess current socioeconomic conditions, we used the following indicators: self-reported educational attainment (years) from the CONOR questionnaire, marital status as registered in the Norwegian National Registry at the time of the baseline survey and data on household income from 1990, obtained from the Norwegian Tax Administration. For the life course analysis, we constructed a social index (SI), using data from the Norwegian censuses of 1960, 1970 and 1980 on type of dwelling (apartment block, row- or detached house), rooms per household capita, status of ownership, owning a telephone, having access to a water closet (WC) and bath/shower inside the dwelling. This information was recorded in all three censuses. We also included the baseline data on educational level, marital status and household income. The individual was given 0 or 1 point for each of these criteria. Points were given for the presumed lesser privileged option, such as not having an indoor WC or bath. This gave an index of life course socioeconomic position ranging from 0 to 21, with those with an index score of 0–3 being the low-exposure reference group, and the next three steps being 4–7, 8–11, and ≥ 12. Details on the scoring system and distribution of participants across the SI are available in Supplementary Tables 1 and 2. 2.4. Endpoint The endpoint used was death from CHD, as registered in the Norwegian Cause of Death Registry, through 2009. Causes of death were categorized according to the International Classification of Disease (ICD-9 and ICD-10). CHD mortality was defined by ICD 9: 410–414 and ICD10: I20–I25. 2.5. Statistical analysis We used multivariable Cox proportional hazard analyses to estimate hazard ratios (HR) for CHD mortality by FH of premature MI, using
303
STATA SE12 (StataCorp LP, College Station, TX). Follow-up time was from entry into CONOR until death or end of follow-up through 2009. When calculating HR, the reference group consisted of those with no known FH, which was compared to the mutually exclusive categories of those with FH of premature MI in parent(s), sibling(s) and parent(s) plus sibling(s). The proportional hazard assumption was tested and confirmed on the basis of Schoenfeld residuals. A significance level of 5% was chosen. The analytic strategy was to examine the effect of FH on CHD mortality in three sex-stratified models: model 1 adjusted for age only; model 2 for age, blood pressure, BMI, triglycerides, cholesterol (total and HDL), diabetes, current smoking and physical activity; model 3 adjusted for model 2 covariates plus indicators of current socioeconomic conditions: education, marital status and household income in 1990. In secondary analyses limited to the SI-subgroup, we investigated the importance of life course socioeconomic position. We compared HR for CHD mortality in the four strata of the SI-subgroup, to assess if the index reflected the socioeconomic gradient of CHD, and adjusted for any FH to observe whether those with FH had a worse socioeconomic position than those without. This was further investigated by using logistic regression to compare odds ratios (OR) for having any FH of premature MI in the different levels of the SI. Lastly, we calculated HR for CHD mortality in the SI-subgroup, adjusting for established risk factors and the SI. 3. Results Mean age at baseline for the participants in the main study population was 47.5 years (SD = 14.5). Over a mean follow-up of 11.0 years (SD = 2.9), corresponding to 1,424,426.7 person years, 1289 (2.1%) men and 688 (1.0%) women in the main study population died from CHD. For the SI-subgroup, the mean age was 52.6 years (SD = 12.4) and mean follow-up was 10.9 years (SD = 2.9), giving 920,613.8 person years. In the SI-subgroup 1088 (2.6%) men and 556 (1.3%) women died from CHD. In the main study population, 10.6% of the men had FH restricted to one or both parents, 2.7% in siblings only, and 1.0% in both parent(s) and sibling(s). Among the women 11.3% had premature MI in one or both parents, 2.9% in one or more siblings and 1.0% in parents and siblings. For the SI-subgroup the corresponding percentages were: 10.6%, 2.7% and 1.2% for the men; and 11.5%, 3.7% and 1.3% for the women. Descriptive characteristics of the population at baseline indicated that those with a positive FH on average had a higher risk factor burden than those with no FH, with higher rates of smoking, less favorable lipid profiles and lower education levels (Table 1). For men there was an increased risk of CHD mortality in all three categories of FH, most expressed in those with premature MI in siblings, where HR adjusted for established risk factors and current socioeconomic position (model 3) was 1.44 (1.19–1.75) for siblings only, and 1.44 (0.98–2.11) for the group with CHD in parents plus siblings (Table 2). For women, there was no effect of premature MI restricted to parents, HR 0.99 (0.70–1.38), whereas CHD in siblings was significant in the age-adjusted estimates (model 1), HR 1.39 (1.10–1.76), and attenuated to non-significant levels by adding established risk factors (model 2). Only the most exposed group, with premature MI in parents as well as siblings, had significantly higher CHD mortality in model 3, HR 1.78 (1.16–2.73). For both sexes, adjusting for established risk factors in model 2 only modestly reduced the risk estimates, and adding indicators of current socioeconomic position in model 3 was of negligible importance. This was also the case when adjusting solely for current socioeconomic conditions (data not shown). The effect of a positive FH in the various first-degree relatives also displayed some difference in risk association for men and women (Table 3). For the men, there was in all three models an increased risk of CHD mortality associated with premature MI in mother, father and
304
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
Table 1 Baseline characteristics of the CONOR participants investigated for family history (FH) of premature myocardial infarction. Men
Table 3 Hazard ratios (HR) and 95% confidence intervals (CI) for CHD mortality by relationship of first-degree relatives with a history of premature myocardial infarction.
Women
Family history
No FH
Any FH
No FH
Any FH
Total Age group 18–39 40–49 50–59 60–69 70+
65,730 % (n) 26.5 (17,387) 35.5 (23,317) 12.6 (8285) 11.3 (7450) 14.1 (9291)
10,674 % (n) 19.4 (2072) 37.6 (4011) 16.8 (1790) 15.4 (1645) 10.8 (1156)
74,700 % (n) 27.3 (20,396) 35.7 (26,636) 11.1 (8317) 10.5 (7846) 15.4 (11,505)
13,328 % (n) 20.1 (2672) 34.6 (4611) 16.6 (2211) 15.1 (2011) 13.7 (1823)
Physical activity⁎ None b1 h 1–2 h N3 h
33.0 (18,786) 25.6 (14,549) 24.2 (13,771) 17.2 (9789)
35.1 (3273) 25.7 (2392) 23.9 (2229) 15.4 (1433)
43.8 (26,778) 24.3 (14,843) 22.9 (14,034) 9.0 (5517)
48.1 (5225) 22.8 (2473) 21.5 (2336) 7.7 (836)
Years of education b10 10–12 13–15 16+
26.9 (17,152) 40.7 (25,943) 16.2 (10,302) 16.2 (10,316)
31.3 (3277) 39.1 (4089) 15.5 (1624) 14.1 (1473)
32.1 (23,021) 36.1 (25,904) 16.3 (11,712) 15.6 (11,196)
38.4 (4976) 36.6 (4741) 13.7 (1780) 11.2 (1454)
Household income 1–25 percentiles 26–50 percentiles 51–75 percentiles 76–100 percentiles
22.1 (14,140) 24.8 (15,887) 26.5 (17,019) 26.6 (17,070)
19.3 (2032) 24.7 (2601) 28.0 (2945) 28.0 (2943)
29.4 (21,312) 25.9 (18,758) 23.2 (16,832) 21.5 (15,602)
28.8 (3781) 26.1 (3418) 23.0 (3016) 22.1 (2893)
Married No Yes
38.3 (25,116) 61.7 (40,432)
34.4 (3669) 65.6 (6986)
41.8 (31,172) 58.2 (43,430)
40.1 (5338) 59.9 (7969)
Current smoker No Yes
69.0 (44,825) 31.0 (20,209)
66.0 (6978) 34.0 (3597)
70.0 (51,595) 30.0 (22,119)
64.1 (8468) 35.9 (4735)
Diabetes No Yes BMI Total cholesterol HDL-cholesterol Triglycerides Systolic BP Diastolic BP
97.2 (63,264) 2.8 (1809) 26.3 ± 0.01 5.8 ± 0.00 1.3 ± 0.00 1.9 ± 0.00 136.6 ± 0.07 79.5 ± 0.05
96.4 (10,187) 3.6 (380) 26.8 ± 0.03 6.0 ± 0.01 1.2 ± 0.00 2.1 ± 0.01 137.7 ± 0.2 80.9 ± 0.1
97.5 (71,790) 2.5 (1862) 25.7 ± 0.04 5.8 ± 0.00 1.5 ± 0.00 1.5 ± 0.00 131.0 ± 0.08 75.0 ± 0.04
97.1 (12,772) 2.9 (388) 26.2 ± 0.04 6.0 ± 0.01 1.5 ± 0.00 1.6 ± 0.01 133.8 ± 0.2 77.1 ± 0.1
Variables are presented as percentage (number) and mean ± SD. Missing data b 4%, unless otherwise noted. ⁎ Missing data for men 13.3% (10,182) and women 18.2% (15,986).
brother, but not sister. The highest risk in model 3 was seen with premature MI in a brother: HR 1.47 (1.23–1.77). For the women, we did not find any significant effect of premature MI in any of the parents, but in
Model 1
Model 2
Model 3
HR (95% CI)
HR (95% CI)
HR (95% CI)
Men, n = 62,628 Mother, n = 1571 Father, n = 5965 Sister, n = 403 Brother, n = 2041
1.58 (1.18–2.10) 1.37 (1.10–1.69) 1.35 (0.91–2.00) 1.60 (1.34–1.92)
1.41 (1.06–1.89) 1.30 (1.05–1.61) 1.27 (0.85–1.89) 1.49 (1.24–1.79)
1.40 (1.05–1.86) 1.30 (1.05–1.61) 1.24 (0.83–1.84) 1.47 (1.23–1.77)
Women, n = 67,438 Mother, n = 2046 Father, n = 6697 Sister, n = 608 Brother, n = 2274
1.15 (0.74–1.79) 1.28 (0.94–1.74) 1.64 (1.16–2.32) 1.45 (1.15–1.82)
1.13 (0.73–1.77) 1.24 (0.91–1.69) 1.30 (0.92–1.84) 1.32 (1.05–1.66)
1.14 (0.73–1.78) 1.24 (0.91–1.69) 1.27 (0.90–1.80) 1.29 (1.03–1.63)
Model 1: adjusted for age. Model 2: adjusted for age, blood pressure, total cholesterol, HDL cholesterol triglycerides, BMI, physical activity, smoking and diabetes. Model 3: adjusted for age, blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking, diabetes, education level, marital status and household income in 1990. Note: the four categories in x-axis in this table are not mutually exclusive.
both brother and sister in age-adjusted estimates (model 1), and for CHD in a brother when adjusting for the complete set of risk factors (model 3): HR 1.29 (1.03–1.63). In the SI-subgroup analyses, the SI was associated with CHD mortality in a dose–response manner, with successively higher mortality for each additional level of lifetime socioeconomic disadvantage (Fig. 1). Adjusting for having a FH had a negligible effect on the estimates. After adjustment for established risk factors, there was no significant association between the SI and OR for having a positive FH (Fig. 2). Adjusting for the SI, representing life course socioeconomic position, did not affect HR for CHD mortality in the SI-subgroup (Supplementary Table 3).
4. Discussion 4.1. Main finding In this cohort study with 11-year follow-up, we found that a FH of premature MI was an independent risk factor for CHD mortality in the index person. Premature MI in siblings appeared to be a more pronounced risk factor than premature MI in parents, and those with a FH of premature MI in both parents and siblings were at the greatest risk. FH seemed to be a somewhat more pronounced risk factor for men than women, except for those with premature MI in both parents and siblings.
Table 2 Hazard ratios (HR) and 95% confidence intervals (CI) for CHD mortality by family history of premature myocardial infarction (MI). Family history
Model 1
Model 2
Model 3
HR (95% CI)
HR (95% CI)
HR (95% CI)
Men, n = 62,628 None, n = 53,721 One or both parents, n = 6606 One or more siblings, n = 1707 One or both parents plus one or more siblings, n = 594 Per category
1.0 1.41 (1.15–1.73) 1.54 (1.27–1.87) 1.65 (1.12–2.41) 1.24 (1.15–1.33)
1.0 1.34 (1.10–1.64) 1.45 (1.19–1.75) 1.46 (1.00–2.14) 1.19 (1.11–1.28)
1.0 1.34 (1.09–1.64) 1.44 (1.19–1.75) 1.44 (0.98–2.11) 1.19 (1.10–1.28)
Women, n = 67,438 None, n = 57,133 One or both parents, n = 7618 One or more siblings, n = 1986 One or both parents plus one or more siblings, n = 701 Per category
1.0 0.99 (0.70–1.38) 1.39 (1.10–1.76) 1.98 (1.29–3.03) 1.19 (1.09–1.31)
1.0 0.97 (0.69–1.36) 1.22 (0.96–1.54) 1.83 (1.19–2.80) 1.14 (1.03–1.25)
1.0 0.98 (0.70–1.37) 1.19 (0.94–1.51) 1.78 (1.16–2.73) 1.12 (1.02–1.24)
Model 1: adjusted for age. Model 2: adjusted for age, blood pressure, total cholesterol, HDL cholesterol triglycerides, BMI, physical activity, smoking and diabetes, age. Model 3: adjusted for blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking, diabetes, education level, marital status and household income in 1990. Per category = sum score per any combination of family members having premature MI.
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
305
CHD in both generations compared to CHD in only parents or siblings. The reason for this might be that middle-aged women have a lower background incidence of MI than men in the same age group, and thus require more genetic risk factors than men to develop CHD, consistent with the Carter-effect [3,25]. For both men and women, the risk estimates, especially in model 1, were higher with a FH of premature MI in same sex siblings. Such sex-of-proband/sex-of-relative interactions have been seen in other studies [3]. Whether this is due to a more similar genetic makeup, or living habits and conditions, is not possible to infer from these figures. As middle-aged women, given comparable risk factors, have lower risk of MI than men, a history of MI in the mother or sister may imply a higher CHD risk for the index person, than MI in a male relative [26]. Fig. 1. Hazard ratio (HR) for CHD mortality as a function of the social index (SI) in the SIsubgroup. N = 84,631. FH = Family history of premature myocardial infarction. *Blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking and diabetes. SI 0 = 0–3 points on SI, SI 1 = 4–7, SI 2 = 8–11, SI 3 = 12 and above.
Adjustment for established risk factors attenuated the HR for CHD mortality modestly. Adjusting for socioeconomic status was of negligible importance, and life course analyses had no advantage in explanatory power compared to indicators of current socioeconomic conditions. 4.2. Effect of premature MI in parents compared to siblings The excess risk of CHD mortality associated with a positive FH seems to vary by which family member that was affected [5,8,9,11,12,22]. In this study, premature MI in siblings appeared to be a greater risk factor than parental MI, consistent with previous studies [8,22,23]. Siblings will usually have more environmental exposures in common than parents and offspring, which might lead to more similar patterns of disease. We do not have data on the exact age of relatives' first MI, but given the mean age in this cohort it could be that siblings on average were younger when developing CHD than parents. To illustrate, whether a relative had a premature MI at age 59 or age 40, might be an expression of different degrees of genetic or environmental risk factor load. A younger age at which parent's MI occurred has been shown to increase offspring risk of MI [5]. 4.3. Sex differences in effect of FH Previous studies have found that a FH of CHD carries a higher CHD risk in men than in women [24]. In our study women, unlike men, had no risk increase when only one or both parents had premature MI. However, estimated HR when parents as well as siblings had CHD was higher for women compared to men. For men there was no increased risk with
4.4. Socioeconomic position Most of the studies evaluating the effects of FH have not controlled for socioeconomic position, and those that have typically used single measures such as education [1], occupation [12] or a combined score with emphasis on psychological stress [5]. Educational level is the most widely used indicator of socioeconomic position in cardiovascular epidemiology, and has been shown to be correlated with cardiovascular health [13,27]. However, there are indications that compared to analysis of single measures, using multiple indicators of socioeconomic position may better reveal the dose–response relationship between number of exposures to social disadvantage and the outcome of ill health [17], and that a life course approach is necessary to fully elucidate the effect of the socioeconomic environment on the development of diseases such as CHD [18,19]. Life course socioeconomic position, including research done with data from the 1960 Norwegian Census, has previously been shown to be associated with increased risk of CHD mortality [15, 28–30]. Single measures of social exposures could fail to capture the complex ways that social and behavioral factors confound associations between risk factors and disease, such under-adjustment may leave the analyses open to the effect of residual confounding [20], which for instance was the case in the observed inverse correlation between plasma levels of vitamin C and cardiovascular risk [20,31]. This could equally be an alternative explanation for the relationship between FH and CHD risk, but has until now not been explored. However, in our data we found little evidence for any consistent effect of socioeconomic conditions independent of other risk factors. The inclusion of life course analyses did not increase the explanatory power compared to only adjusting for education, marital status and household income. Even if low socioeconomic position was associated with excess risk of CHD mortality, we did not find that those with a positive FH had worse socioeconomic conditions than those without. The effect of FH has previously been shown to be consistent across regions and socioeconomic strata [5]. In a recent adoption study, the biological father's, but not the adoptive father's, socioeconomic status was associated with higher mortality in adult life, possibly lending support to the idea that genetics or prenatal environment shape future pattern of illness more than the rearing environment [32]. Twin studies have found a moderate genetic contribution to the variation in CHD mortality [11,33]. The genetics CHD is a complex interplay between a few known and many as of yet unknown genes, and epigenetic influences [34–36], and currently unmeasured genetic factors are a potential explanation for the association between FH and CHD [5]. 4.5. Strengths and limitations
Fig. 2. Odds ratio (OR) for any positive family history of premature MI in first-degree relatives as a function of the social index (SI) in the SI-subgroup. N = 84,631. *Blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking and diabetes. SI 0 = 0–3 points on SI, SI 1 = 4–7, SI 2 = 8–11, SI 3 = 12 and above.
The strength of the CONOR study is the large number of individuals who underwent standardized assessments of anthropometry, blood tests, lifestyle characteristics and medical history, which provided the opportunity to control for a large number of potential confounding factors. The linkage between CONOR and several Norwegian censuses was
306
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
a unique source for analysis of the relationship between socioeconomic position and FH. Limitations are that CHD risk factor information among the relatives was not available, and that the CONOR participants were only assessed at a single point in time. There is some uncertainty associated with FH being self-reported. However, self-reported FH of premature MI has been shown to have both a positive and negative predictive value of above 90%, and self-reported FH of MI at any age to have a sensitivity of around 70%, and a specificity of more than 90% [37–39]. The relatively poor sensitivity could mean that a large number of premature MI in relatives went unreported, possibly biasing results towards an underestimation of the effect of a FH. The advantage of self-reported information is the fact that it corresponds to how FH is assessed clinically, making the risk estimates of interest to clinicians. Some precaution is necessary in the interpretation of the results. The number of index persons with the different types of FH probably affected results: For men the point estimates of those with CHD in a sibling, and those with CHD in parents plus siblings, were similar, but with wider confidence interval in the latter group, reflecting the smaller sample size. Similarly, the effect of premature MI in sister and mother on the index person's risk of CHD mortality had a wide confidence interval, due to a low number of participants. 5. Conclusion FH of premature MI is an independent risk factor for CHD mortality that differs in its magnitude of effect by sex of index person and type of familial relationship. Established risk factors only modestly affected the results, and both current and life course socioeconomic conditions were of negligible importance, suggesting the association is not confounded by this. Funding The project was funded by our own institutions. Conflict of interests All authors have completed the ICMJE uniform disclosure form and declare: no conflict of interests. Acknowledgments We acknowledge to contributions of the CONOR Steering Committee and the participating universities and the Norwegian institute of Public Health for letting us analyze the data from the CONOR cohort. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2015.04.160. References [1] M. Andresdottir, G. Sigurdsson, H. Sigvaldason, V. Gudnason, Fifteen percent of myocardial infarctions and coronary revascularizations explained by family history unrelated to conventional risk factors. The Reykjavik Cohort Study, Eur. Heart J. 23 (2002) 1655–1663. [2] J.M. Bachmann, B.L. Willis, C.R. Ayers, A. Khera, J.D. Berry, Association between family history and coronary heart disease death across long-term follow-up in men: the Cooper Center Longitudinal Study, Circulation 125 (2012) 3092–3098. [3] A. Banerjee, L.E. Silver, C. Heneghan, S.J. Welch, L.M. Bull, Z. Mehta, et al., Sex-specific familial clustering of myocardial infarction in patients with acute coronary syndromes, Circulation 2 (2009) 98–105. [4] E. Barrett-Connor, K. Khaw, Family history of heart attack as an independent predictor of death due to cardiovascular disease, Circulation 69 (1984) 1065–1069. [5] C.K. Chow, S. Islam, L. Bautista, Z. Rumboldt, A. Yusufali, C. Xie, et al., Parental history and myocardial infarction risk across the world: The INTERHEART Study, J. Am. Coll. Cardiol. 57 (2011) 619–627.
[6] D.M. Lloyd-Jones, B.H. Nam, R.B. D'Agostino Sr., D. Levy, J.M. Murabito, T.J. Wang, et al., Parental cardiovascular disease as a risk factor for cardiovascular disease in middle-aged adults: a prospective study of parents and offspring, JAMA 291 (2004) 2204–2211. [7] J.M. Murabito, M.J. Pencina, B.-H. Nam, R.B. D'Agostino, T.J. Wang, D. Lloyd-Jones, et al., Sibling cardiovascular disease as a risk factor for cardiovascular disease in middle-aged adults, JAMA 294 (2005) 3117–3123. [8] M. Nielsen, C. Andersson, T.A. Gerds, P.K. Andersen, T.B. Jensen, L. Køber, et al., Familial clustering of myocardial infarction in first-degree relatives: a nationwide study, Eur. Heart J. 34 (2013) 1198–1203. [9] H.D. Sesso, I.-M. Lee, J.M. Gaziano, K.M. Rexrode, R.J. Glynn, J.E. Buring, Maternal and paternal history of myocardial infarction and risk of cardiovascular disease in men and women, Circulation 104 (2001) 393–398. [10] S. Sivapalaratnam, S.M. Boekholdt, M.D. Trip, M.S. Sandhu, R. Luben, J.J. Kastelein, et al., Family history of premature coronary heart disease and risk prediction in the EPIC-Norfolk prospective population study, Heart 96 (2010) 1985–1989. [11] S. Zdravkovic, A. Wienke, N. Pedersen, M. Marenberg, A. Yashin, U. De Faire, Heritability of death from coronary heart disease: a 36‐year follow‐up of 20966 Swedish twins, J. Intern. Med. 252 (2002) 247–254. [12] B. Zöller, X. Li, J. Sundquist, K. Sundquist, Multiplex sibling history of coronary heart disease is a strong risk factor for coronary heart disease, Eur. Heart J. 33 (2012) 2849–2855. [13] G.A. Kaplan, J.E. Keil, Socioeconomic factors and cardiovascular disease: a review of the literature, Circulation 88 (1993) 1973–1998. [14] C. Méjean, M. Droomers, Y.T. van der Schouw, I. Sluijs, S. Czernichow, D.E. Grobbee, et al., The contribution of diet and lifestyle to socioeconomic inequalities in cardiovascular morbidity and mortality, Int. J. Cardiol. 168 (2013) 5190–5195. [15] Ø. Næss, B. Claussen, G.D. Smith, Housing conditions in childhood and cause-specific adult mortality: The effect of sanitary conditions and economic deprivation on 55,761 men in Oslo, Scand. J. Public Health 35 (2007) 570–576. [16] Ø. Næss, B.H. Strand, G.D. Smith, Childhood and adulthood socioeconomic position across 20 causes of death: a prospective cohort study of 800000 Norwegian men and women, J. Epidemiol. Community Health 61 (2007) 1004–1009. [17] G.D. Smith, M.J. Shipley, G. Rose, Magnitude and causes of socioeconomic differentials in mortality: further evidence from the Whitehall Study, J. Epidemiol. Community Health 44 (1990) 265–270. [18] Y. Ben-Shlomo, D. Kuh, A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives, Int. J. Epidemiol. 31 (2002) 285–293. [19] J. Lynch, G.D. Smith, A life course approach to chronic disease epidemiology, Annu. Rev. Public Health 26 (2005) 1–35. [20] D. Lawlor, S.G. Davey, D. Kundu, K. Bruckdorfer, S. Ebrahim, Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet 363 (2004) 1724. [21] O. Naess, A.J. Sogaard, E. Arnesen, A.C. Beckstrom, E. Bjertness, A. Engeland, et al., Cohort profile: cohort of Norway (CONOR), Int. J. Epidemiol. 37 (2008) 481–485. [22] M.F. Ranthe, L. Carstensen, N. Øyen, J. Tfelt-Hansen, M. Christiansen, W.J. McKenna, et al., family history of premature death and risk of early onset cardiovascular disease, J. Am. Coll. Cardiol. 60 (2012) 814–821. [23] M.C. Roncaglioni, L. Santoro, B. D'Avanzo, E. Negri, A. Nobili, A. Ledda, et al., Role of family history in patients with myocardial infarction. An Italian case-control study. GISSI-EFRIM Investigators, Circulation 85 (1992) 2065–2072. [24] Ž. Reiner, A.L. Catapano, G. De Backer, I. Graham, M.-R. Taskinen, O. Wiklund, et al., ESC/EAS Guidelines for the management of dyslipidaemias The Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS), Eur. Heart J. 32 (2011) 1769–1818. [25] C. Carter, The inheritance of congenital pyloric stenosis, Br. Med. Bull. 17 (1961) 251–253. [26] K. Kappert, M. Böhm, R. Schmieder, H. Schumacher, K. Teo, S. Yusuf, et al., Impact of gender on cardiovascular outcome in patients at cardiovascular high risk: analysis of the Telmisartan Randomized AssessmeNT Study in ACE-INtolerant Subjects with Cardiovascular Disease (TRANSCEND) and the Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET), Circulation (2012) 111.086660 (CIRCULATIONAHA). [27] M.A. Winkleby, D.E. Jatulis, E. Frank, S.P. Fortmann, Socioeconomic status and health: how education, income, and occupation contribute to risk factors for cardiovascular disease, Am. J. Public Health 82 (1992) 816–820. [28] G.D. Smith, C. Hart, Life-course socioeconomic and behavioral influences on cardiovascular disease mortality: the collaborative study, Am. J. Public Health 92 (2002) 1295–1298. [29] B. Galobardes, J.W. Lynch, G.D. Smith, Is the association between childhood socioeconomic circumstances and cause-specific mortality established? Update of a systematic review, J. Epidemiol. Community Health 62 (2008) 387–390. [30] B. Galobardes, G.D. Smith, J.W. Lynch, Systematic review of the influence of childhood socioeconomic circumstances on risk for cardiovascular disease in adulthood, Ann. Epidemiol. 16 (2006) 91–104. [31] D. Lawlor, S. Ebrahim, D. Kundu, K. Bruckdorfer, P. Whincup, G.D. Smith, Vitamin C is not associated with coronary heart disease risk once life course socioeconomic position is taken into account: prospective findings from the British Women's Heart and Health Study, Heart 91 (2005) 1086–1087. [32] M. Osler, L. Petersen, E. Prescott, T.W. Teasdale, T.I. Sørensen, Genetic and environmental influences on the relation between parental social class and mortality, Int. J. Epidemiol. 35 (2006) 1272–1277. [33] M.E. Marenberg, N. Risch, L.F. Berkman, B. Floderus, U. de Faire, Genetic susceptibility to death from coronary heart disease in a study of twins, N. Engl. J. Med. 330 (1994) 1041–1046.
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307 [34] S. Kathiresan, D. Srivastava, Genetics of human cardiovascular disease, Cell 148 (2012) 1242–1257. [35] J.M. Lorenzen, F. Martino, T. Thum, Epigenetic modifications in cardiovascular disease, Basic Res. Cardiol. 107 (2012) 245. [36] B. Mayer, J. Erdmann, H. Schunkert, Genetics and heritability of coronary artery disease and myocardial infarction, Clin. Res. Cardiol. 96 (2007) 1–7. [37] F. Kee, L. Tiret, J.Y. Robo, V. Nicaud, E. McCrum, A. Evans, et al., Reliability of reported family history of myocardial infarction, BMJ (Clin. Res. Ed.) 307 (1993) 1528–1530.
307
[38] J.M. Murabito, B.H. Nam, R.B. D'Agostino Sr., D.M. Lloyd-Jones, C.J. O'Donnell, P.W. Wilson, Accuracy of offspring reports of parental cardiovascular disease history: the Framingham Offspring Study, Ann. Intern. Med. 140 (2004) 434–440. [39] J.S. Silberberg, J. Wlodarczyk, J. Fryer, C.D. Ray, M.J. Hensley, Correction for biases in a population-based study of family history and coronary heart disease. The Newcastle Family History Study I, Am. J. Epidemiol. 147 (1998) 1123–1132.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Family history of premature myocardial infarction, life course socioeconomic position and coronary heart disease mortality — A Cohort of Norway (CONOR) study Bendik S. Fiskå a,⁎, Inger Ariansen b, Sidsel Graff-Iversen b, Grethe S. Tell c,d, Grace M. Egeland c,d, Øyvind Næss a,b a
University of Oslo, Faculty of Medicine, PB 1078 Blindern, 0316 Oslo, Norway Division of Epidemiology, National Institute of Public Health, PB 4404, Nydalen, 0403 Oslo, Norway c Department of Global Public Health and Primary Care, University of Bergen, PB 7804, 5018 Bergen, Norway d Department of Health Registries, Norwegian Institute of Public Health, Kalfarveien 31, 5018 Bergen, Norway b
a r t i c l e
i n f o
Article history: Received 27 March 2015 Accepted 18 April 2015 Available online 22 April 2015 Keywords: Coronary heart disease Myocardial infarction Family history Life course socioeconomic position Coronary heart disease mortality
a b s t r a c t Background/objectives: To investigate self-reported family history (FH) of premature myocardial infarction (MI) in first-degree relatives as a risk factor for coronary heart disease (CHD) mortality, and assess whether any observed effect could be explained by current or life course socioeconomic position. Methods: 130,066 participants from Cohort of Norway were examined during 1994–2003. A subgroup (n = 84,631) had additional life course socioeconomic data. Using Cox proportional hazard analyses, we calculated hazard ratios (HR) for CHD mortality, assessed by linkages to the Norwegian Cause of Death Registry through 2009. For subgroup analyses, we created an index of life course socioeconomic position, and assessed its role as a potential confounder in the association of FH with CHD. Results: For men, MI in parents and siblings were both a significant risk factor for CHD mortality after adjusting for established risk factors and current socioeconomic conditions; the highest risk was with MI in siblings (HR: 1.44 [1.19–1.75]). For women, FH constituted significant risk after similar adjustment only for those with MI in parents plus siblings (HR: 1.78 [1.16–2.73]). Adjusting for current and life course socioeconomic conditions only marginally lowered the estimates, and those with FH did not have worse life course socioeconomic position than those without. Conclusions: FH of premature MI is an independent risk factor for CHD mortality that differs in magnitude of effect by the sex of the index person and type of familial relationship. Life course socioeconomic position has little impact on the association between FH and CHD, suggesting the effect is not confounded by this. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Background/objectives A family history (FH) of coronary heart disease (CHD) in first–degree relatives is an established risk factor for CHD [1–12]; however, the true magnitude and pattern of this risk association remains unknown, and concerns have been raised regarding whether the impact of a positive FH is independent from other risk factors [7]. Risk of CHD is associated with socioeconomic position [13–17], and it is possible that a positive FH could be of socioeconomic, rather than biological, origin. Only a few studies on the effect of FH have controlled for socioeconomic position, using single indicators, such as education or occupation [1,12], or a score emphasizing psychological stress [5]. Public health researchers have increasingly emphasized the need to take the full life course into account to understand the origin and population distribution of diseases such as CHD [18,19]. Merely adjusting for single ⁎ Corresponding author. E-mail address: b.s.fi[email protected] (B.S. Fiskå).
http://dx.doi.org/10.1016/j.ijcard.2015.04.160 0167-5273/© 2015 Elsevier Ireland Ltd. All rights reserved.
measures at single points in time, may lead to estimates prone to residual confounding [20]. The extent of this potential residual confounding has not previously been investigated for FH. The aim of this study was to investigate self-reported FH of premature MI in parents and/or siblings as a risk factor for CHD mortality in the index patient, and evaluate whether an effect of a positive FH could be explained by established risk factors, as well as indicators of current and life course socioeconomic position. 2. Methods 2.1. Study population Cohort of Norway (CONOR) is a compilation of ten regional, both urban and rural, community-based health surveys conducted in the period 1994–2003 [21]. Subjects were invited to a baseline health examination which included anthropometric measures, blood sampling and questionnaires. The overall response rate was 58.3%. In eight of the ten
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
surveys (n = 164,432), history of premature myocardial infarction (MI) in family members was surveyed. The study population consisted of participants for whom we had complete covariate information (n = 130,066). Data from the Norwegian censuses of 1960, 1970 and 1980 were available for 84,635 participants, who were included in subgroup analyses of life course socioeconomic conditions. The mean age at time of first census was 15.1 years (SD = 12.5), and our data allowed us to assess socioeconomic position over a period of 34–43 years, depending on time of recruitment to CONOR. A flow-chart presenting the inclusion process is provided in Supplementary Fig. 1. 2.2. Exposure Exposure was self-reported FH of premature MI in first-degree relatives. Using a standardized questionnaire, the respondents were asked to indicate the relatives who have or have had MI before age 60, with the options being mother, father, brother, sister and child. Information on MI in children was not included in our analyses. We created three mutually exclusive categories: FH of premature MI in one or both parents only, one or more siblings only, and one or both parents plus one or more siblings. 2.3. Risk factors In addition to age and sex, the following biological markers recorded at the baseline health examination were considered as relevant risk- or protective factors: systolic and diastolic blood pressure (mm Hg), BMI (kg/m2) and non-fasting serum triglycerides (mmol/l), total cholesterol (mmol/l) and HDL cholesterol (mmol/l). Detailed information about sampling methods is provided elsewhere [21]. Additionally, selfreported information from the CONOR questionnaire on current smoking (yes/no), diabetes (yes/no) and past year level (hours per week) of hard physical activity (causing sweating or breathlessness) was included. To assess current socioeconomic conditions, we used the following indicators: self-reported educational attainment (years) from the CONOR questionnaire, marital status as registered in the Norwegian National Registry at the time of the baseline survey and data on household income from 1990, obtained from the Norwegian Tax Administration. For the life course analysis, we constructed a social index (SI), using data from the Norwegian censuses of 1960, 1970 and 1980 on type of dwelling (apartment block, row- or detached house), rooms per household capita, status of ownership, owning a telephone, having access to a water closet (WC) and bath/shower inside the dwelling. This information was recorded in all three censuses. We also included the baseline data on educational level, marital status and household income. The individual was given 0 or 1 point for each of these criteria. Points were given for the presumed lesser privileged option, such as not having an indoor WC or bath. This gave an index of life course socioeconomic position ranging from 0 to 21, with those with an index score of 0–3 being the low-exposure reference group, and the next three steps being 4–7, 8–11, and ≥ 12. Details on the scoring system and distribution of participants across the SI are available in Supplementary Tables 1 and 2. 2.4. Endpoint The endpoint used was death from CHD, as registered in the Norwegian Cause of Death Registry, through 2009. Causes of death were categorized according to the International Classification of Disease (ICD-9 and ICD-10). CHD mortality was defined by ICD 9: 410–414 and ICD10: I20–I25. 2.5. Statistical analysis We used multivariable Cox proportional hazard analyses to estimate hazard ratios (HR) for CHD mortality by FH of premature MI, using
303
STATA SE12 (StataCorp LP, College Station, TX). Follow-up time was from entry into CONOR until death or end of follow-up through 2009. When calculating HR, the reference group consisted of those with no known FH, which was compared to the mutually exclusive categories of those with FH of premature MI in parent(s), sibling(s) and parent(s) plus sibling(s). The proportional hazard assumption was tested and confirmed on the basis of Schoenfeld residuals. A significance level of 5% was chosen. The analytic strategy was to examine the effect of FH on CHD mortality in three sex-stratified models: model 1 adjusted for age only; model 2 for age, blood pressure, BMI, triglycerides, cholesterol (total and HDL), diabetes, current smoking and physical activity; model 3 adjusted for model 2 covariates plus indicators of current socioeconomic conditions: education, marital status and household income in 1990. In secondary analyses limited to the SI-subgroup, we investigated the importance of life course socioeconomic position. We compared HR for CHD mortality in the four strata of the SI-subgroup, to assess if the index reflected the socioeconomic gradient of CHD, and adjusted for any FH to observe whether those with FH had a worse socioeconomic position than those without. This was further investigated by using logistic regression to compare odds ratios (OR) for having any FH of premature MI in the different levels of the SI. Lastly, we calculated HR for CHD mortality in the SI-subgroup, adjusting for established risk factors and the SI. 3. Results Mean age at baseline for the participants in the main study population was 47.5 years (SD = 14.5). Over a mean follow-up of 11.0 years (SD = 2.9), corresponding to 1,424,426.7 person years, 1289 (2.1%) men and 688 (1.0%) women in the main study population died from CHD. For the SI-subgroup, the mean age was 52.6 years (SD = 12.4) and mean follow-up was 10.9 years (SD = 2.9), giving 920,613.8 person years. In the SI-subgroup 1088 (2.6%) men and 556 (1.3%) women died from CHD. In the main study population, 10.6% of the men had FH restricted to one or both parents, 2.7% in siblings only, and 1.0% in both parent(s) and sibling(s). Among the women 11.3% had premature MI in one or both parents, 2.9% in one or more siblings and 1.0% in parents and siblings. For the SI-subgroup the corresponding percentages were: 10.6%, 2.7% and 1.2% for the men; and 11.5%, 3.7% and 1.3% for the women. Descriptive characteristics of the population at baseline indicated that those with a positive FH on average had a higher risk factor burden than those with no FH, with higher rates of smoking, less favorable lipid profiles and lower education levels (Table 1). For men there was an increased risk of CHD mortality in all three categories of FH, most expressed in those with premature MI in siblings, where HR adjusted for established risk factors and current socioeconomic position (model 3) was 1.44 (1.19–1.75) for siblings only, and 1.44 (0.98–2.11) for the group with CHD in parents plus siblings (Table 2). For women, there was no effect of premature MI restricted to parents, HR 0.99 (0.70–1.38), whereas CHD in siblings was significant in the age-adjusted estimates (model 1), HR 1.39 (1.10–1.76), and attenuated to non-significant levels by adding established risk factors (model 2). Only the most exposed group, with premature MI in parents as well as siblings, had significantly higher CHD mortality in model 3, HR 1.78 (1.16–2.73). For both sexes, adjusting for established risk factors in model 2 only modestly reduced the risk estimates, and adding indicators of current socioeconomic position in model 3 was of negligible importance. This was also the case when adjusting solely for current socioeconomic conditions (data not shown). The effect of a positive FH in the various first-degree relatives also displayed some difference in risk association for men and women (Table 3). For the men, there was in all three models an increased risk of CHD mortality associated with premature MI in mother, father and
304
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
Table 1 Baseline characteristics of the CONOR participants investigated for family history (FH) of premature myocardial infarction. Men
Table 3 Hazard ratios (HR) and 95% confidence intervals (CI) for CHD mortality by relationship of first-degree relatives with a history of premature myocardial infarction.
Women
Family history
No FH
Any FH
No FH
Any FH
Total Age group 18–39 40–49 50–59 60–69 70+
65,730 % (n) 26.5 (17,387) 35.5 (23,317) 12.6 (8285) 11.3 (7450) 14.1 (9291)
10,674 % (n) 19.4 (2072) 37.6 (4011) 16.8 (1790) 15.4 (1645) 10.8 (1156)
74,700 % (n) 27.3 (20,396) 35.7 (26,636) 11.1 (8317) 10.5 (7846) 15.4 (11,505)
13,328 % (n) 20.1 (2672) 34.6 (4611) 16.6 (2211) 15.1 (2011) 13.7 (1823)
Physical activity⁎ None b1 h 1–2 h N3 h
33.0 (18,786) 25.6 (14,549) 24.2 (13,771) 17.2 (9789)
35.1 (3273) 25.7 (2392) 23.9 (2229) 15.4 (1433)
43.8 (26,778) 24.3 (14,843) 22.9 (14,034) 9.0 (5517)
48.1 (5225) 22.8 (2473) 21.5 (2336) 7.7 (836)
Years of education b10 10–12 13–15 16+
26.9 (17,152) 40.7 (25,943) 16.2 (10,302) 16.2 (10,316)
31.3 (3277) 39.1 (4089) 15.5 (1624) 14.1 (1473)
32.1 (23,021) 36.1 (25,904) 16.3 (11,712) 15.6 (11,196)
38.4 (4976) 36.6 (4741) 13.7 (1780) 11.2 (1454)
Household income 1–25 percentiles 26–50 percentiles 51–75 percentiles 76–100 percentiles
22.1 (14,140) 24.8 (15,887) 26.5 (17,019) 26.6 (17,070)
19.3 (2032) 24.7 (2601) 28.0 (2945) 28.0 (2943)
29.4 (21,312) 25.9 (18,758) 23.2 (16,832) 21.5 (15,602)
28.8 (3781) 26.1 (3418) 23.0 (3016) 22.1 (2893)
Married No Yes
38.3 (25,116) 61.7 (40,432)
34.4 (3669) 65.6 (6986)
41.8 (31,172) 58.2 (43,430)
40.1 (5338) 59.9 (7969)
Current smoker No Yes
69.0 (44,825) 31.0 (20,209)
66.0 (6978) 34.0 (3597)
70.0 (51,595) 30.0 (22,119)
64.1 (8468) 35.9 (4735)
Diabetes No Yes BMI Total cholesterol HDL-cholesterol Triglycerides Systolic BP Diastolic BP
97.2 (63,264) 2.8 (1809) 26.3 ± 0.01 5.8 ± 0.00 1.3 ± 0.00 1.9 ± 0.00 136.6 ± 0.07 79.5 ± 0.05
96.4 (10,187) 3.6 (380) 26.8 ± 0.03 6.0 ± 0.01 1.2 ± 0.00 2.1 ± 0.01 137.7 ± 0.2 80.9 ± 0.1
97.5 (71,790) 2.5 (1862) 25.7 ± 0.04 5.8 ± 0.00 1.5 ± 0.00 1.5 ± 0.00 131.0 ± 0.08 75.0 ± 0.04
97.1 (12,772) 2.9 (388) 26.2 ± 0.04 6.0 ± 0.01 1.5 ± 0.00 1.6 ± 0.01 133.8 ± 0.2 77.1 ± 0.1
Variables are presented as percentage (number) and mean ± SD. Missing data b 4%, unless otherwise noted. ⁎ Missing data for men 13.3% (10,182) and women 18.2% (15,986).
brother, but not sister. The highest risk in model 3 was seen with premature MI in a brother: HR 1.47 (1.23–1.77). For the women, we did not find any significant effect of premature MI in any of the parents, but in
Model 1
Model 2
Model 3
HR (95% CI)
HR (95% CI)
HR (95% CI)
Men, n = 62,628 Mother, n = 1571 Father, n = 5965 Sister, n = 403 Brother, n = 2041
1.58 (1.18–2.10) 1.37 (1.10–1.69) 1.35 (0.91–2.00) 1.60 (1.34–1.92)
1.41 (1.06–1.89) 1.30 (1.05–1.61) 1.27 (0.85–1.89) 1.49 (1.24–1.79)
1.40 (1.05–1.86) 1.30 (1.05–1.61) 1.24 (0.83–1.84) 1.47 (1.23–1.77)
Women, n = 67,438 Mother, n = 2046 Father, n = 6697 Sister, n = 608 Brother, n = 2274
1.15 (0.74–1.79) 1.28 (0.94–1.74) 1.64 (1.16–2.32) 1.45 (1.15–1.82)
1.13 (0.73–1.77) 1.24 (0.91–1.69) 1.30 (0.92–1.84) 1.32 (1.05–1.66)
1.14 (0.73–1.78) 1.24 (0.91–1.69) 1.27 (0.90–1.80) 1.29 (1.03–1.63)
Model 1: adjusted for age. Model 2: adjusted for age, blood pressure, total cholesterol, HDL cholesterol triglycerides, BMI, physical activity, smoking and diabetes. Model 3: adjusted for age, blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking, diabetes, education level, marital status and household income in 1990. Note: the four categories in x-axis in this table are not mutually exclusive.
both brother and sister in age-adjusted estimates (model 1), and for CHD in a brother when adjusting for the complete set of risk factors (model 3): HR 1.29 (1.03–1.63). In the SI-subgroup analyses, the SI was associated with CHD mortality in a dose–response manner, with successively higher mortality for each additional level of lifetime socioeconomic disadvantage (Fig. 1). Adjusting for having a FH had a negligible effect on the estimates. After adjustment for established risk factors, there was no significant association between the SI and OR for having a positive FH (Fig. 2). Adjusting for the SI, representing life course socioeconomic position, did not affect HR for CHD mortality in the SI-subgroup (Supplementary Table 3).
4. Discussion 4.1. Main finding In this cohort study with 11-year follow-up, we found that a FH of premature MI was an independent risk factor for CHD mortality in the index person. Premature MI in siblings appeared to be a more pronounced risk factor than premature MI in parents, and those with a FH of premature MI in both parents and siblings were at the greatest risk. FH seemed to be a somewhat more pronounced risk factor for men than women, except for those with premature MI in both parents and siblings.
Table 2 Hazard ratios (HR) and 95% confidence intervals (CI) for CHD mortality by family history of premature myocardial infarction (MI). Family history
Model 1
Model 2
Model 3
HR (95% CI)
HR (95% CI)
HR (95% CI)
Men, n = 62,628 None, n = 53,721 One or both parents, n = 6606 One or more siblings, n = 1707 One or both parents plus one or more siblings, n = 594 Per category
1.0 1.41 (1.15–1.73) 1.54 (1.27–1.87) 1.65 (1.12–2.41) 1.24 (1.15–1.33)
1.0 1.34 (1.10–1.64) 1.45 (1.19–1.75) 1.46 (1.00–2.14) 1.19 (1.11–1.28)
1.0 1.34 (1.09–1.64) 1.44 (1.19–1.75) 1.44 (0.98–2.11) 1.19 (1.10–1.28)
Women, n = 67,438 None, n = 57,133 One or both parents, n = 7618 One or more siblings, n = 1986 One or both parents plus one or more siblings, n = 701 Per category
1.0 0.99 (0.70–1.38) 1.39 (1.10–1.76) 1.98 (1.29–3.03) 1.19 (1.09–1.31)
1.0 0.97 (0.69–1.36) 1.22 (0.96–1.54) 1.83 (1.19–2.80) 1.14 (1.03–1.25)
1.0 0.98 (0.70–1.37) 1.19 (0.94–1.51) 1.78 (1.16–2.73) 1.12 (1.02–1.24)
Model 1: adjusted for age. Model 2: adjusted for age, blood pressure, total cholesterol, HDL cholesterol triglycerides, BMI, physical activity, smoking and diabetes, age. Model 3: adjusted for blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking, diabetes, education level, marital status and household income in 1990. Per category = sum score per any combination of family members having premature MI.
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
305
CHD in both generations compared to CHD in only parents or siblings. The reason for this might be that middle-aged women have a lower background incidence of MI than men in the same age group, and thus require more genetic risk factors than men to develop CHD, consistent with the Carter-effect [3,25]. For both men and women, the risk estimates, especially in model 1, were higher with a FH of premature MI in same sex siblings. Such sex-of-proband/sex-of-relative interactions have been seen in other studies [3]. Whether this is due to a more similar genetic makeup, or living habits and conditions, is not possible to infer from these figures. As middle-aged women, given comparable risk factors, have lower risk of MI than men, a history of MI in the mother or sister may imply a higher CHD risk for the index person, than MI in a male relative [26]. Fig. 1. Hazard ratio (HR) for CHD mortality as a function of the social index (SI) in the SIsubgroup. N = 84,631. FH = Family history of premature myocardial infarction. *Blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking and diabetes. SI 0 = 0–3 points on SI, SI 1 = 4–7, SI 2 = 8–11, SI 3 = 12 and above.
Adjustment for established risk factors attenuated the HR for CHD mortality modestly. Adjusting for socioeconomic status was of negligible importance, and life course analyses had no advantage in explanatory power compared to indicators of current socioeconomic conditions. 4.2. Effect of premature MI in parents compared to siblings The excess risk of CHD mortality associated with a positive FH seems to vary by which family member that was affected [5,8,9,11,12,22]. In this study, premature MI in siblings appeared to be a greater risk factor than parental MI, consistent with previous studies [8,22,23]. Siblings will usually have more environmental exposures in common than parents and offspring, which might lead to more similar patterns of disease. We do not have data on the exact age of relatives' first MI, but given the mean age in this cohort it could be that siblings on average were younger when developing CHD than parents. To illustrate, whether a relative had a premature MI at age 59 or age 40, might be an expression of different degrees of genetic or environmental risk factor load. A younger age at which parent's MI occurred has been shown to increase offspring risk of MI [5]. 4.3. Sex differences in effect of FH Previous studies have found that a FH of CHD carries a higher CHD risk in men than in women [24]. In our study women, unlike men, had no risk increase when only one or both parents had premature MI. However, estimated HR when parents as well as siblings had CHD was higher for women compared to men. For men there was no increased risk with
4.4. Socioeconomic position Most of the studies evaluating the effects of FH have not controlled for socioeconomic position, and those that have typically used single measures such as education [1], occupation [12] or a combined score with emphasis on psychological stress [5]. Educational level is the most widely used indicator of socioeconomic position in cardiovascular epidemiology, and has been shown to be correlated with cardiovascular health [13,27]. However, there are indications that compared to analysis of single measures, using multiple indicators of socioeconomic position may better reveal the dose–response relationship between number of exposures to social disadvantage and the outcome of ill health [17], and that a life course approach is necessary to fully elucidate the effect of the socioeconomic environment on the development of diseases such as CHD [18,19]. Life course socioeconomic position, including research done with data from the 1960 Norwegian Census, has previously been shown to be associated with increased risk of CHD mortality [15, 28–30]. Single measures of social exposures could fail to capture the complex ways that social and behavioral factors confound associations between risk factors and disease, such under-adjustment may leave the analyses open to the effect of residual confounding [20], which for instance was the case in the observed inverse correlation between plasma levels of vitamin C and cardiovascular risk [20,31]. This could equally be an alternative explanation for the relationship between FH and CHD risk, but has until now not been explored. However, in our data we found little evidence for any consistent effect of socioeconomic conditions independent of other risk factors. The inclusion of life course analyses did not increase the explanatory power compared to only adjusting for education, marital status and household income. Even if low socioeconomic position was associated with excess risk of CHD mortality, we did not find that those with a positive FH had worse socioeconomic conditions than those without. The effect of FH has previously been shown to be consistent across regions and socioeconomic strata [5]. In a recent adoption study, the biological father's, but not the adoptive father's, socioeconomic status was associated with higher mortality in adult life, possibly lending support to the idea that genetics or prenatal environment shape future pattern of illness more than the rearing environment [32]. Twin studies have found a moderate genetic contribution to the variation in CHD mortality [11,33]. The genetics CHD is a complex interplay between a few known and many as of yet unknown genes, and epigenetic influences [34–36], and currently unmeasured genetic factors are a potential explanation for the association between FH and CHD [5]. 4.5. Strengths and limitations
Fig. 2. Odds ratio (OR) for any positive family history of premature MI in first-degree relatives as a function of the social index (SI) in the SI-subgroup. N = 84,631. *Blood pressure, total cholesterol, HDL cholesterol, triglycerides, BMI, physical activity, smoking and diabetes. SI 0 = 0–3 points on SI, SI 1 = 4–7, SI 2 = 8–11, SI 3 = 12 and above.
The strength of the CONOR study is the large number of individuals who underwent standardized assessments of anthropometry, blood tests, lifestyle characteristics and medical history, which provided the opportunity to control for a large number of potential confounding factors. The linkage between CONOR and several Norwegian censuses was
306
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307
a unique source for analysis of the relationship between socioeconomic position and FH. Limitations are that CHD risk factor information among the relatives was not available, and that the CONOR participants were only assessed at a single point in time. There is some uncertainty associated with FH being self-reported. However, self-reported FH of premature MI has been shown to have both a positive and negative predictive value of above 90%, and self-reported FH of MI at any age to have a sensitivity of around 70%, and a specificity of more than 90% [37–39]. The relatively poor sensitivity could mean that a large number of premature MI in relatives went unreported, possibly biasing results towards an underestimation of the effect of a FH. The advantage of self-reported information is the fact that it corresponds to how FH is assessed clinically, making the risk estimates of interest to clinicians. Some precaution is necessary in the interpretation of the results. The number of index persons with the different types of FH probably affected results: For men the point estimates of those with CHD in a sibling, and those with CHD in parents plus siblings, were similar, but with wider confidence interval in the latter group, reflecting the smaller sample size. Similarly, the effect of premature MI in sister and mother on the index person's risk of CHD mortality had a wide confidence interval, due to a low number of participants. 5. Conclusion FH of premature MI is an independent risk factor for CHD mortality that differs in its magnitude of effect by sex of index person and type of familial relationship. Established risk factors only modestly affected the results, and both current and life course socioeconomic conditions were of negligible importance, suggesting the association is not confounded by this. Funding The project was funded by our own institutions. Conflict of interests All authors have completed the ICMJE uniform disclosure form and declare: no conflict of interests. Acknowledgments We acknowledge to contributions of the CONOR Steering Committee and the participating universities and the Norwegian institute of Public Health for letting us analyze the data from the CONOR cohort. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2015.04.160. References [1] M. Andresdottir, G. Sigurdsson, H. Sigvaldason, V. Gudnason, Fifteen percent of myocardial infarctions and coronary revascularizations explained by family history unrelated to conventional risk factors. The Reykjavik Cohort Study, Eur. Heart J. 23 (2002) 1655–1663. [2] J.M. Bachmann, B.L. Willis, C.R. Ayers, A. Khera, J.D. Berry, Association between family history and coronary heart disease death across long-term follow-up in men: the Cooper Center Longitudinal Study, Circulation 125 (2012) 3092–3098. [3] A. Banerjee, L.E. Silver, C. Heneghan, S.J. Welch, L.M. Bull, Z. Mehta, et al., Sex-specific familial clustering of myocardial infarction in patients with acute coronary syndromes, Circulation 2 (2009) 98–105. [4] E. Barrett-Connor, K. Khaw, Family history of heart attack as an independent predictor of death due to cardiovascular disease, Circulation 69 (1984) 1065–1069. [5] C.K. Chow, S. Islam, L. Bautista, Z. Rumboldt, A. Yusufali, C. Xie, et al., Parental history and myocardial infarction risk across the world: The INTERHEART Study, J. Am. Coll. Cardiol. 57 (2011) 619–627.
[6] D.M. Lloyd-Jones, B.H. Nam, R.B. D'Agostino Sr., D. Levy, J.M. Murabito, T.J. Wang, et al., Parental cardiovascular disease as a risk factor for cardiovascular disease in middle-aged adults: a prospective study of parents and offspring, JAMA 291 (2004) 2204–2211. [7] J.M. Murabito, M.J. Pencina, B.-H. Nam, R.B. D'Agostino, T.J. Wang, D. Lloyd-Jones, et al., Sibling cardiovascular disease as a risk factor for cardiovascular disease in middle-aged adults, JAMA 294 (2005) 3117–3123. [8] M. Nielsen, C. Andersson, T.A. Gerds, P.K. Andersen, T.B. Jensen, L. Køber, et al., Familial clustering of myocardial infarction in first-degree relatives: a nationwide study, Eur. Heart J. 34 (2013) 1198–1203. [9] H.D. Sesso, I.-M. Lee, J.M. Gaziano, K.M. Rexrode, R.J. Glynn, J.E. Buring, Maternal and paternal history of myocardial infarction and risk of cardiovascular disease in men and women, Circulation 104 (2001) 393–398. [10] S. Sivapalaratnam, S.M. Boekholdt, M.D. Trip, M.S. Sandhu, R. Luben, J.J. Kastelein, et al., Family history of premature coronary heart disease and risk prediction in the EPIC-Norfolk prospective population study, Heart 96 (2010) 1985–1989. [11] S. Zdravkovic, A. Wienke, N. Pedersen, M. Marenberg, A. Yashin, U. De Faire, Heritability of death from coronary heart disease: a 36‐year follow‐up of 20966 Swedish twins, J. Intern. Med. 252 (2002) 247–254. [12] B. Zöller, X. Li, J. Sundquist, K. Sundquist, Multiplex sibling history of coronary heart disease is a strong risk factor for coronary heart disease, Eur. Heart J. 33 (2012) 2849–2855. [13] G.A. Kaplan, J.E. Keil, Socioeconomic factors and cardiovascular disease: a review of the literature, Circulation 88 (1993) 1973–1998. [14] C. Méjean, M. Droomers, Y.T. van der Schouw, I. Sluijs, S. Czernichow, D.E. Grobbee, et al., The contribution of diet and lifestyle to socioeconomic inequalities in cardiovascular morbidity and mortality, Int. J. Cardiol. 168 (2013) 5190–5195. [15] Ø. Næss, B. Claussen, G.D. Smith, Housing conditions in childhood and cause-specific adult mortality: The effect of sanitary conditions and economic deprivation on 55,761 men in Oslo, Scand. J. Public Health 35 (2007) 570–576. [16] Ø. Næss, B.H. Strand, G.D. Smith, Childhood and adulthood socioeconomic position across 20 causes of death: a prospective cohort study of 800000 Norwegian men and women, J. Epidemiol. Community Health 61 (2007) 1004–1009. [17] G.D. Smith, M.J. Shipley, G. Rose, Magnitude and causes of socioeconomic differentials in mortality: further evidence from the Whitehall Study, J. Epidemiol. Community Health 44 (1990) 265–270. [18] Y. Ben-Shlomo, D. Kuh, A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives, Int. J. Epidemiol. 31 (2002) 285–293. [19] J. Lynch, G.D. Smith, A life course approach to chronic disease epidemiology, Annu. Rev. Public Health 26 (2005) 1–35. [20] D. Lawlor, S.G. Davey, D. Kundu, K. Bruckdorfer, S. Ebrahim, Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet 363 (2004) 1724. [21] O. Naess, A.J. Sogaard, E. Arnesen, A.C. Beckstrom, E. Bjertness, A. Engeland, et al., Cohort profile: cohort of Norway (CONOR), Int. J. Epidemiol. 37 (2008) 481–485. [22] M.F. Ranthe, L. Carstensen, N. Øyen, J. Tfelt-Hansen, M. Christiansen, W.J. McKenna, et al., family history of premature death and risk of early onset cardiovascular disease, J. Am. Coll. Cardiol. 60 (2012) 814–821. [23] M.C. Roncaglioni, L. Santoro, B. D'Avanzo, E. Negri, A. Nobili, A. Ledda, et al., Role of family history in patients with myocardial infarction. An Italian case-control study. GISSI-EFRIM Investigators, Circulation 85 (1992) 2065–2072. [24] Ž. Reiner, A.L. Catapano, G. De Backer, I. Graham, M.-R. Taskinen, O. Wiklund, et al., ESC/EAS Guidelines for the management of dyslipidaemias The Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS), Eur. Heart J. 32 (2011) 1769–1818. [25] C. Carter, The inheritance of congenital pyloric stenosis, Br. Med. Bull. 17 (1961) 251–253. [26] K. Kappert, M. Böhm, R. Schmieder, H. Schumacher, K. Teo, S. Yusuf, et al., Impact of gender on cardiovascular outcome in patients at cardiovascular high risk: analysis of the Telmisartan Randomized AssessmeNT Study in ACE-INtolerant Subjects with Cardiovascular Disease (TRANSCEND) and the Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET), Circulation (2012) 111.086660 (CIRCULATIONAHA). [27] M.A. Winkleby, D.E. Jatulis, E. Frank, S.P. Fortmann, Socioeconomic status and health: how education, income, and occupation contribute to risk factors for cardiovascular disease, Am. J. Public Health 82 (1992) 816–820. [28] G.D. Smith, C. Hart, Life-course socioeconomic and behavioral influences on cardiovascular disease mortality: the collaborative study, Am. J. Public Health 92 (2002) 1295–1298. [29] B. Galobardes, J.W. Lynch, G.D. Smith, Is the association between childhood socioeconomic circumstances and cause-specific mortality established? Update of a systematic review, J. Epidemiol. Community Health 62 (2008) 387–390. [30] B. Galobardes, G.D. Smith, J.W. Lynch, Systematic review of the influence of childhood socioeconomic circumstances on risk for cardiovascular disease in adulthood, Ann. Epidemiol. 16 (2006) 91–104. [31] D. Lawlor, S. Ebrahim, D. Kundu, K. Bruckdorfer, P. Whincup, G.D. Smith, Vitamin C is not associated with coronary heart disease risk once life course socioeconomic position is taken into account: prospective findings from the British Women's Heart and Health Study, Heart 91 (2005) 1086–1087. [32] M. Osler, L. Petersen, E. Prescott, T.W. Teasdale, T.I. Sørensen, Genetic and environmental influences on the relation between parental social class and mortality, Int. J. Epidemiol. 35 (2006) 1272–1277. [33] M.E. Marenberg, N. Risch, L.F. Berkman, B. Floderus, U. de Faire, Genetic susceptibility to death from coronary heart disease in a study of twins, N. Engl. J. Med. 330 (1994) 1041–1046.
B.S. Fiskå et al. / International Journal of Cardiology 190 (2015) 302–307 [34] S. Kathiresan, D. Srivastava, Genetics of human cardiovascular disease, Cell 148 (2012) 1242–1257. [35] J.M. Lorenzen, F. Martino, T. Thum, Epigenetic modifications in cardiovascular disease, Basic Res. Cardiol. 107 (2012) 245. [36] B. Mayer, J. Erdmann, H. Schunkert, Genetics and heritability of coronary artery disease and myocardial infarction, Clin. Res. Cardiol. 96 (2007) 1–7. [37] F. Kee, L. Tiret, J.Y. Robo, V. Nicaud, E. McCrum, A. Evans, et al., Reliability of reported family history of myocardial infarction, BMJ (Clin. Res. Ed.) 307 (1993) 1528–1530.
307
[38] J.M. Murabito, B.H. Nam, R.B. D'Agostino Sr., D.M. Lloyd-Jones, C.J. O'Donnell, P.W. Wilson, Accuracy of offspring reports of parental cardiovascular disease history: the Framingham Offspring Study, Ann. Intern. Med. 140 (2004) 434–440. [39] J.S. Silberberg, J. Wlodarczyk, J. Fryer, C.D. Ray, M.J. Hensley, Correction for biases in a population-based study of family history and coronary heart disease. The Newcastle Family History Study I, Am. J. Epidemiol. 147 (1998) 1123–1132.
