Heart failure (HF) is one of the leading causes of hospitalization and death in the United States and throughout Europe. Although a higher risk for HF with antecedent myocardial infarction (MI) has been reported in offspring whose parents had MIs before age 55 years, it is unclear whether adherence to healthful behaviors can mitigate that risk. The aim of the present study was therefore to prospectively examine if adherence to healthy weight, regular exercise, moderate alcohol consumption, and abstinence from smoking can attenuate such increased HF risk. Information on parental history of MI and lifestyle factors was collected using questionnaires. Subjects adhering to ≥3 healthy lifestyle factors were classified as having good versus poor lifestyle scores. Incident HF was assessed via yearly follow-up questionnaires and validated in a subsample. During an average follow up of 21.7 ± 6.5 years, 1,323 new HF cases (6.6%), of which 190 (14.4%) were preceded by MI, occurred. Compared to subjects with good lifestyle scores and no parental histories of premature MI, multivariate adjusted hazard ratios for incident HF with antecedent MI were 3.21 (95% confidence interval 1.74 to 5.91) for subjects with good lifestyle score and parental histories of premature MI, 1.52 (95% confidence interval 1.12 to 2.07) for those with poor lifestyle score and no parental histories of premature MI, and 4.60 (95% confidence interval 2.55 to 8.30) for those with poor lifestyle scores and parental histories of premature MI. In conclusion, our data suggest that even in subjects at higher risk for HF because of genetic predisposition, adherence to healthful lifestyle factors may attenuate such an elevated HF risk.
The present study was designed to examine the extent to which adherence to healthy lifestyle factors might attenuate the increased risk for heart failure (HF) with antecedent myocardial infarction (MI) in subjects with parental histories of premature MI.
Methods
The Physicians’ Health Study (PHS) I is a completed randomized, double-blind, placebo-controlled trial designed to study the effects of low-dose aspirin and β-carotene on cardiovascular disease and cancer in United States male physicians. In 1997, a second randomized trial, PHS II, was started and included 7,641 physicians from PHS I along with 7,000 new physicians. Detailed descriptions of PHS I and PHS II have been published elsewhere. Participants from PHS I, including the 7,641 subjects who participated in PHS II, are subsequently referred to as PHS subjects. Of the 22,071 original subjects in the PHS, we excluded those with missing information on the 12-month questionnaire (n = 60), those with missing data on parental history of MI for both parents or missing parental age at MI (n = 1,416), those with prevalent HF (n = 25) at 12 months after randomization, and those with missing information on covariates (n = 510). Thus, a final sample of 20,060 participants was selected for the present analyses. Each participant gave written informed consent, and the institutional review board at Brigham and Women’s Hospital approved the study protocol.
During the 12-month follow-up questionnaire, information on parental history of MI was collected using the following questions: “Has your mother ever had a documented myocardial infarction? If yes, at what age?” and “Has your father ever had a documented myocardial infarction? If yes, at what age?” Possible answers for the initial question were “No,” “Yes,” and “Don’t know.” In cases in which both parents had histories of MI, we used the minimum age at the diagnosis of parental MI for analyses. A parental history of premature MI was defined as having an MI before 55 years of age.
Cardiovascular end points including HF in the PHS have been ascertained using annual follow-up questionnaires. A questionnaire was mailed to each participant every 6 months during the first year and has been mailed annually thereafter to obtain information on compliance with the intervention and the occurrence of new medical diagnoses, including HF. The validation of HF in the PHS has been previously reported.
Self-reported baseline weight and height were used to compute body mass index (weight in kilograms divided by height in meters squared). Each lifestyle factor was dichotomized as follows: noncurrent smoker versus current smoker, normal weight (body mass index <25 kg/m 2 ) versus overweight or obese (body mass index ≥25 kg/m 2 ), weekly alcohol use versus more or less frequent alcohol use, and regular exercise (weekly) versus infrequent or no exercise (less than weekly). These cut points were chosen on the basis of previous associations between individual lifestyle factors and HF risk in this cohort or public health recommendations. Each participant received 1 point for each of the following: body mass index <25 kg/m 2 , noncurrent smoking, weekly alcohol use, and exercise weekly or more often. Thus, study participants were categorized according to the number of desirable lifestyle factors (0, 1, 2, 3, or 4). A score >2 was defined as a good lifestyle score, while a score ≤2 was defined as poor lifestyle score.
Demographic data were collected at baseline (1982 to 1984). Age was used as a continuous variable. Diagnosis of diabetes mellitus was made on self-reports via questionnaires mailed to each participant every 6 months during the first year and annually thereafter. Diagnosis of coronary heart disease (CHD) was initially made on self-reports via follow-up questionnaires. An end point committee reviewed medical records to confirm the diagnosis of CHD. Hypertension was defined as anyone who self-reported the diagnosis, by blood pressure >140/90 mm Hg per Joint National Committee seventh report guidelines, or on the basis of medication lists provided by the study participants. All types of atrial fibrillation (AF) were assessed via follow-up questionnaires. These self-reports of AF have been validated in the same cohort using a more detailed questionnaire on the diagnosis of AF and the review of medical records in a previous study.
Compared to subjects with good lifestyle scores and no parental histories of premature MI as a reference, we classified the remaining participants into those with good lifestyle scores and parental histories of premature MI, poor lifestyle scores and no histories of parental premature MI, and poor lifestyle scores and parental histories of premature MI. We computed person-time of follow-up from 12 months after randomization until the first occurrence of (1) HF, (2) death, or (3) date of receipt of last follow-up questionnaire. We used Cox proportional-hazard models to compute multivariate-adjusted hazard ratios with corresponding 95% confidence intervals (CIs) using subjects with good lifestyle scores and no parental histories of premature MI as the reference group. We assessed confounding by established risk factors for HF. The initial model adjusted only for age. A second model additionally controlled for diabetes, hypertension, AF, and aspirin randomization status.
Kaplan-Meier survival curves were also used to compare HF-free and MI-free survival for each of the 4 groups considered in our analysis. Additionally, we conducted a secondary analysis stratified by each decade of age at entry to assess effect modification by age. Assumptions for proportional-hazard models were tested using product terms of log (time) factor and variables of interest and were met (all p values >0.05). All analyses were completed using SAS version 9.3 (SAS Institute Inc., Cary, North Carolina). The significance level was set at 0.05.
Results
Among 20,060 participants in PHS I, the mean age at 12 months after randomization was 55 ± 9.5 years. Table 1 lists baseline characteristics of the study participants. Although 94% of the subjects did not report parental histories of premature MI, 6% of the participants had ≥1 parent who had an MI before the age of 55 years. In addition, 60% of the subjects had good lifestyle scores. Subjects with parental histories of premature MI were noted to be younger than those with no parental histories of premature MI. As expected, the prevalence rates of diabetes and hypertension was also noted to be higher in subjects with poor lifestyle scores than in those with good lifestyle scores.
| Parameter | No Parental History of Premature MI | Parental History of Premature MI | ||
|---|---|---|---|---|
| Good Lifestyle Score | Poor Lifestyle Score | Good Lifestyle Score | Poor Lifestyle Score | |
| (n = 11,336) | (n = 7,581) | (n = 697) | (n = 446) | |
| Age (years) | 54 ± 9.5 | 56 ± 9.4 | 51 ± 7.8 | 53 ± 8.0 |
| Body mass index (kg/m 2 ) | 24 ± 2.3 | 26 ± 2.9 | 24 ± 2.3 | 26 ± 2.9 |
| Randomized to aspirin | 50% | 50% | 51% | 46% |
| Current exercise | 91% | 45% | 90% | 47% |
| Never smoker | 55% | 42% | 55% | 40% |
| Past smoking | 43% | 35% | 42% | 32% |
| Current smoker | 2.9% | 23% | 2.7% | 28% |
| Moderate drinking (weekly) | 68% | 21% | 69% | 25% |
| Hypertension | 20% | 29% | 19% | 27% |
| Diabetes mellitus | 2.2% | 4.7% | 2.7% | 4.0% |
| AF | 1.8% | 2.0% | 0.9% | 1.6% |
| Coronary heart disease | 8.4% | 13% | 14% | 19% |
During an average follow-up period of 22 years, 1,323 new cases of HF (6.6%) were documented in this cohort. Of these, 190 cases (14%) occurred in subjects with antecedent MIs. Compared to subjects with good lifestyle scores and no parental histories of premature MI, multivariate-adjusted hazard ratios for incident HF with antecedent MI were 3.21 (95% CI 1.74 to 5.91) for subjects with good lifestyle scores and parental histories of premature MI, 1.52 (95% CI 1.12 to 2.07) for those with poor lifestyle scores and no histories of parental premature MI, and 4.60 (95% CI 2.55 to 8.30) for those with poor lifestyle scores and parental histories of premature MI ( Table 2 ).
| Parameter | No Parental History of Premature MI | Parental History of Premature MI | ||
|---|---|---|---|---|
| Good Lifestyle Score | Poor Lifestyle Score | Good Lifestyle Score | Poor Lifestyle Score | |
| Cases/person-years | 79/253,925 | 86/156,559 | 12/16,304 | 13/9,449 |
| Crude incidence rate (per 10,000 person-years) | 3.1 | 5.5 | 7.4 | 13.8 |
| Unadjusted model | 1.0 | 1.83 (1.35–2.49) | 2.31 (1.26–4.25) | 4.52 (2.51–8.12) |
| Age-adjusted model | 1.0 | 1.68 (1.23–2.28) | 3.25 (1.76–5.98) | 5.21 (2.90–9.38) |
| Multivariate model ⁎ | 1.0 | 1.52 (1.12–2.07) | 3.21 (1.74–5.91) | 4.60 (2.55–8.30) |
⁎ Age, diabetes mellitus, hypertension, AF, and aspirin randomization.
When stratified by age, the influence of healthful lifestyle factors appeared to be stronger with younger age at entry ( Table 3 ). Kaplan-Meier survival curves comparing HF-free and MI-free survival among different groups are shown in Figure 1 .
