- Click here for explanation of classification scheme.

To identify factors associated with remaining free of chronic disease in later life.

• Men and women
• Aged 65 years and older
• Community dwelling
• Did not require a proxy respondent at baseline
• Participants in the Cardiovascular Health Study.

Those who required a proxy respondent at baseline.

Recruitment

Participants were recruited as part of the Cardiovascular Health Study (CHS), using Medicare eligibility lists of the Health Care Finance Administration from four communities.

Design

Prospective cohort study 

Blinding used

Implied with measurements

Intervention

Not applicable

Statistical Analysis

• Age and cohort-adjusted bivariate association were calculated to identify behavioral and other factors that predict remaining healthy in each cohort
• To account for different follow-up times in two different cohorts, survival analysis models were generated using time to first event as the dependent variable
• Baseline hazard function as modeled with a Weibull distribution, and parameters were determined using SAS statistical software
• Stepwise models were run in two stages, first to determine the most significant behavioral factors and then to identify the most significant predictors among traditional risk factors and measures of subclinical disease.

 

Timing of Measurements 

• Data was collected at baseline and every six months over a maximum of seven years. There were two cohorts; one recruited in 1989-1990 and another recruited in 1992-1993. The follow-up on the second cohort averaged 3.5 years of follow-up.
• The baseline exam consisted of a home interview and a clinic examination.

Dependent Variables

The outcomes of interest included cancer, COPD, angina, coronary angioplasty, coronary artery bypass surgery, claudification, congestive heart failure, acute myocardial infarction, stroke, transient ischemic attack or any fatal event.

• Cancer as measured by self-report
• CVD as verified from hospital and physician records: Subclinical measures of CVD considered included carotid artery wall thickness and ankle-arm BP index
• COPD as measured by self-report or a physician's diagnosis
• HTN as defined by blood pressure readings (≥160/95) using a mercury sphygmomanometer or use of hypertensive medication
• Impaired glucose tolerance as defined according to American Diabetes Association guidelines (fasting glucose 110-125mg/dL)
• Diabetes as defined according to American Diabetes Association guidelines (fasting glucose ≥126mg/dL) or participant was taking insulin or oral hypoglycemic agents.

Independent Variables

• Participants were asked to fast for 12 hours and then serum was collected for a complete blood count (CBC), serum chemistry, serum insulin, plasma lipids, and coagulation factors
• Blood pressure was measured
• Anthropometric measurements included height, weight, waist circumference, and hip circumference 
• Functional status was assessed using an interviewer-administered questionnaire. Self-reported difficulty in activities of daily living (ADL) and instrumental activities of daily living (IADL) were evaluated.
• Dietary intake was assessed using a modified Block food frequency questionnaire administered via picture card of foods sorted by frequency of use
• Participants were asked about type and frequency of exercise
• Participants were asked about their past medical history and reported cardiovascular events were confirmed by a medical record review or by contacting the participant's physician
• Behavioral risk factors were evaluated by self-report during the participant interviews and included physical activity, alcohol consumption, smoking status, social support, self-reported health, influenza vaccination status, regular aspirin use, demographic factors, and hormone replacement therapy (HRT) in women.

Control Variables

Age

• Initial N: 5,888 participants took place in the baseline exam. Only 3,342 were included in the data analysis; these were the participants who were disease-fee at baseline and included 2,043 women and 1,299 men. The second cohort included 687 men and women (gender breakdown was not provided).
• Attrition (final N): All participants were accounted for: no attrition was reported
• Age: 65 years of age and older. The majority of the participants (2,341) were between 65 and 74.
• Ethnicity: Unspecified, except that the second cohort was described as African American
• Other relevant demographics: Information on income level and education level was obtained but baseline information was not reported
• Anthropometrics: Baseline information on height, weight, BMI, and waist circumference was obtained but not reported 
• Location: Participants were from four areas of the country:
  • Forsyth County, North Carolina
  • Sacramento County, California
  • Washington County, Maryland
  • Allegheny County, Pennsylvania.

 

Key Findings

• The risk of incident events increased with age and the event rates were higher in men then women. On the average, a majority of participants remained free of CVD, COPD, and cancer and remained alive.
• A number of behavioral factors were associated with continued health in older adults, including high-intensity exercise, never having smoked (or quit smoking), higher educational status (in men) and higher incomes
• Self-reported health was a strong predictor of future health
• Use of HRT in women was not associated with remaining healthy
• Regular use of aspirin was associated with a decreased risk of remaining healthy
• Dietary factors were associated with remaining healthy, with a lower percentage of calories from fat and a higher percentage of calories from carbohydrate (women only) in those who remained healthy
• Wine-drinking was associated with remaining healthy
• CVD risk factors and subclinical markers of disease were related to continued health
• Lower systolic BP and absence of a history of HTN were associated with remaining healthy in both men and women
• Lower C-reactive protein and higher baseline HDL-cholesterol were significantly associated with remaining healthy for both men and women
• The absence of diabetes was significantly related to remaining disease-free.

Proportion of Healthy Years During Follow-Up Associated with Behavioral Risk Factors

Variables	Proportion of Health Years  Associated with Variable	95% Confidence Interval
Age (five years)	0.79*	(0.75, 0.82)
Male gender	0.74*	(0.66, 0.83)
Waist (10cm)	0.94**	(0.90, 0.98)
Exercise intensity  No exercise  Low  Medium  High	1.00* 1.30* 1.37* 1.53*	º
Cigarette smoking  Never  Former  Current	1.00* 0.75* 0.56*	º
Wine drinking	1.11	(0.99, 1.24)
Regular aspirin use	0.82*	(0.73, 0.91)
Income (>$25,000)	1.19**	(1.06, 1.33)
Recent change in finances	0.78**	(0.66, 0.93)

ºCI's were not clear on original article due to formatting issues. Unable to report

*P<0.001

**P<0.01

Behavioral Factors Associated with Remaining Health During Follow-Up

Variables	Proportion of Health Years  Associated with Variable	95% Confidence Interval
Age (five years)	0.82*	(0.78, 0.86)
Male gender	0.82*	(0.72, 0.92)
Waist (10cm)	1.01	(0.96, 1.06)
Exercise intensity  No exercise  Low  Medium  High	1.00** 1.25** 1.34** 1.42**	(1.03, 1.52) (1.09, 1.64) (1.09, 1.85)
Cigarette smoking  Never  Former  Current	1.00* 0.78* 0.66*	(0.69, 0.88) (0.56, 0.78)
Wine drinking	1.04	(0.93, 1.17)
Regular aspirin use	0.83**	(0.75, 0.93)
Income (>$25,000)	1.08	(0.96, 1.22)
Recent change in finances	0.75**	(0.63, 0.90)

*P=0.01

**P=0.001

 Risk Factors, and Subclinical Disease Measures  Associated with Remaining Healthy

Behavioral Factors	Proportion of Healthy Years Associated with Variable	95% Confidence Interval
Diabetes status Impaired fasting glucose Diabetes	0.94* 0.71*	(0.80, 1.10) (0.62, 0.82)
Ankle/arm BP ratio <0.9	0.73*	(0.62, 0.86)
Diastolic BP (1 standard deviation)	0.92**	(0.87, 0.97)
HDL (1 standard deviation)	1.07**	(1.01, 1.14)
Standardized carotid wall thickness (1 standard deviation)	0.85*	(0.81, 0.90)
Ln (CRP) mg/L	0.93**	(0.88, 0.98)

*P=0.01

**P=0.001

The authors conclude that the data suggest that specific behavioral factors and medical interventions are associated with continued health in later life. Remaining physically active, refraining from smoking, consuming moderate amounts of wine, and maintaining weight are important prevention strategies. Controlling cardiovascular risk factors (diabetes, HDL cholesterol, and blood pressure) are associated with maintenance of good health in older adults.

Government:	National Heart, Lung, and Blood Institute
University/Hospital:	Wake Forest University, University of Washington, University of Vermont, Johns Hopkins University, University of Pittsburgh, University of California

• It is unclear to this reviewer if any anthropometric or biochemical analysis were conducted at any follow-up contacts that were made every six months. It appears that conclusions about good health over time were obtained from self-reports and hospital records but not follow-up data obtained as part of the study. Changes in weight, BR, dietary habits, and exercise habits could affect the study outcomes and do not appear to be accounted for.
• This reviewer is concerned that BMI and waist circumferences were not reported at baseline. If the study population was mostly normal or underweight, the results may have been skewed. It is not clear to this review if obesity was one of the risk factors that was controlled for, although it does state "after controlling for traditional risk factors".
• Demographic information, including SES and education level, were not reported at baseline. This information could affect participant's health status. It is not clear if data were controlled for SES or education level.
• For the second cohort, the average length of follow up was 3.5 years. That is not adequate time to assess changes in health status, especially in "young old".
• This reviewer was unclear as to how the two cohorts were integrated for the reporting of results.