Are Varicose Veins a Marker for Susceptibility to Coronary Heart Disease in Men? Results from the Normative Aging Study

Are Varicose Veins a Marker for Susceptibility to Coronary Heart Disease in Men? Results from the Normative Aging Study

Are Varicose Veins a Marker for Susceptibility to Coronary Heart Disease in Men? Results from the Normative Aging Study Thayer E. Scott, MPH,1 Manuel ...

89KB Sizes 0 Downloads 49 Views

Are Varicose Veins a Marker for Susceptibility to Coronary Heart Disease in Men? Results from the Normative Aging Study Thayer E. Scott, MPH,1 Manuel V. Mendez, MD,1 Wayne W. LaMorte, MD, PhD, MPH,1,2 L. Adrienne Cupples, PhD,3 Pantel S. Vokonas, MD,2,4,5 Raul I. Garcia, DMD,6 and James O. Menzoian, MD,1 Boston, Massachusetts

Clinical observations suggest that varicose veins (VV) are less frequent in patients undergoing infrainguinal bypass surgery for femoral artery occlusive disease. While some previous studies support this relationship, others report that VV are more prevalent in coronary heart disease patients (CHD). This study used the Normative Aging Study (NAS) population to examine the association between VV and symptomatic CHD. The incidence of CHD over 35 years of followup was determined in the 2280 initially healthy male volunteers enrolled in the NAS. The incidence of CHD in the VV population and the subjects without VV were compared using KaplanMeier survival curves and the log-rank test. A time-dependent proportional hazards regression method was used to further explore the relationship between VV disease and subsequent development of CHD after adjusting for other cardiovascular risk factors. A total of 569 subjects (24.9%) were diagnosed with VV prior to the development of symptomatic CHD, and 1708 (75.1%) were not. Over 35 years of follow-up, 98 subjects with VV developed symptomatic CHD (17.2%), while 363 of those without VV subsequently developed symptomatic CHD (21.2%). Kaplan-Meier survival curves suggested a reduced risk of symptomatic CHD for subjects with VV (p = 0.0001). Further exploration of this relationship in a proportional hazards multivariate model showed VV to be associated with a 36% decreased risk of symptomatic CHD after adjusting for other recognized cardiovascular risk factors. In the NAS population, men with VV were less likely to develop symptomatic CHD over the 35+ years of follow-up than were subjects without VV.

INTRODUCTION 1

Department of Surgery, Section of Vascular Surgery and Surgical Research Section, Boston Medical Center, Boston, MA, USA. 2 Department of Epidemiology, School of Public Health, Boston University, Boston, MA, USA. 3 Department of Biostatistics, School of Public Health, Boston University, Boston, MA, USA. 4 Normative Aging Study, VA Boston Healthcare System, Boston, MA, USA. 5 Section of Preventive Medicine and Epidemiology, Department of Medicine, Boston University School of Medicine, Boston, MA, USA. 6 Department of Health Policy and Health Services Research, Goldman School of Dental Medicine, Boston University, Boston, MA, USA. Correspondence to: James O. Menzoian, MD, Section of Vascular Surgery, Boston Medical Center, 88 East Newton Street, Boston, MA 02118, USA, E-mail: [email protected]

Ann Vasc Surg 2004; 18: 459-464 DOI: 10.1007/s10016-004-0056-z Ó Annals of Vascular Surgery Inc. Published online: 27 May 2004

An observation during surgery to harvest venous conduit grafts for arterial reconstructive surgery suggested that relatively few of these patients also had varicose veins (VV) and raised the possibility that presence or absence of VV might be a marker for susceptibility to coronary heart disease (CHD). Several previous studies have explored whether an association exists between heart disease and VV, but results have been inconsistent. A cross-sectional study in Trabzon, Turkey, found a lower prevalence of angina in elderly individuals with VV.1 The Kaiser-Permanente Epidemiologic Study of Myocardial Infarction also reported that CHD was less frequent in subjects with either VV or hemorrhoids.2 In contrast, the Framingham Heart Study and the Paris Prospective Study both detected an increased risk for CHD in VV subjects.3,4 459

460 Scott et al.

To further explore this relationship, we used the Normative Aging Study population of 2280 initially healthy male volunteers to determine whether subjects with VV developed heart disease at a lower rate than those without VV over the 35-year course of the study.

SUBJECTS AND METHODS Study subjects were male participants in the U.S. Department of Veteran Affairs (VA) Normative Aging Study (NAS). The NAS is a closed-panel longitudinal study, instituted in 1961, and designed to investigate age-related morbidity in 2280 initially healthy male volunteers from the greater Boston area. At the inception of the study, these self-selected volunteers ranged in age from 21 to 80 years and were free of known chronic medical conditions such as CHD and diabetes. While 97% of subjects were veterans, with few exceptions they were not VA patients, and they have continued to receive all of their medical and dental care in the private sector. Subjects were examined every 3 years by trained VA staff physicians following a systematic protocol, and all information was recorded on standardized forms specifically designed for the NAS. Diseases and conditions were entered into the database according to the 8th Revision of the International Classification of Diseases (ICD-8). The protocol was approved by the Department of Veterans Affairs Committee on Human Studies, and procedures followed were in accordance with institutional guidelines. All subjects conferred their informed consent prior to their entry into the study. All clinical assessments included a medical history, a physical examination, and a variety of biochemical laboratory tests. Total serum cholesterol (mg/dL) was quantified colorimetrically on blood drawn from subjects following an overnight fast. Blood pressure was measured in each arm on seated subjects using standard mercury sphygmomanometry. For analytic purposes, the mean readings from both arms were used for systolic and diastolic blood pressure (mmHg). Obesity was gauged using body mass index (kg/m2). Cigarette smoking status was categorized as ‘‘current smoker,’’ ‘‘former smoker,’’ or ‘‘never smoked.’’ Current smokers were defined as men who smoked one or more cigarettes daily and were still smoking at least 1 month prior to survey, while former smokers had smoked at least one cigarette per day for at least 1 year, but had refrained from smoking for a minimum of one month from examination. Never smokers included subjects whose

Annals of Vascular Surgery

lifetime exposure to smoking was less than 20 packs of cigarettes. Total lifetime cigarette exposure was calculated at the time of subject enrollment as the average number of cigarettes smoked per day (converted into packs), multiplied by the number of years the subject smoked. Daily alcohol consumption was derived from replies to the Cornell Medical Index Health Questionnaire, where subjects responded to whether they usually drank two or more alcoholic drinks per day (yes/no). Maximum level of education completed was dichotomized into less than high school education, or high school education and beyond. Subjects also provided information as to whether any heart trouble ran in their family (yes/ no). Symptomatic CHD was defined as a diagnosis of nonfatal myocardial infarction, angina pectoris, or death from cardiac diseases (ICD8: 410-414). The criteria for myocardial infarction and angina pectoris were those used by the Framingham Heart Study.5 All CHD diagnoses by the staff physicians were confirmed and validated by an NAS cardiologist (P.V.). The data were supplemented by information on cardiovascular illnesses abstracted from hospital records. Subjects were classified as diabetic if their fasting blood glucose was >140 mg/ dL, their 2-hr postprandial blood glucose was >200 mg/dL, or if they were documented as having a clinical diagnosis of diabetes in their medical record. The presence or absence of lower extremity VV was routinely determined by physical inspection at each triennial examination. The standardized research study record included a specific question regarding VV. The examining physician consistently documented presence or absence of VV on the study record checklist. The working definition included dilated clusters of varicosities that were distinctly visible on the lower extremity (usually located on the posterior aspect of the lower leg and popliteal areas), but omitted telangectasia (spider veins) and tiny superficial veins (ICD-8 codes: 454454.1, 451). When stasis dermatitis or an antecedent diagnosis of phlebitis or thrombophlebitis was made, these were noted separately. A diagnosis of VV prior to or at the inception of the study was not considered to be a medical condition, and these individuals were not excluded from study participation. Statistical Analysis Subjects who were diagnosed with VV were compared at study entry to those without VV using

Vol. 18, No. 4, 2004

Varicose veins in men with CHD 461

Table I. Baseline cardiovascular risk factor status in subjects with and without varicose veins Variable

With VV (n = 569)

Without VV (n = 1708)

p

Age (years) Body mass index (kg/m2) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Cholesterol (mg/dL) Diabetes Family history of heart disease Packs of cigarettes smoked over lifetime/1000 Current smoker Two or more drinks of alcohol/day High school education or more

43.8 ± 26 ± 124 ± 77.7 ± 205.7 ± 0.2% 39.2% 5.3 ± 35.5% 11.4% 55.9%

41.8 ± 25.6 ± 123.5 ± 76.8 ± 203.3 ± 0.2% 33% 5.5 ± 41.1% 12.6% 53.4%

0.0001 0.002 0.4 0.02 0.3 1.0 0.007 0.4 0.02 0.4 0.3

chi-squared tests and Student’s t-tests to examine univariate relationships for common cardiovascular risk factors. The incidence of symptomatic CHD in the VV population and the subjects without VV were compared using Kaplan-Meier survival curves and the log-rank test. To investigate how multivariate associations varied over time, a time-dependent proportional hazards model was used to assess risk factors for development of VV and to explore the relationship between VV disease and subsequent development of symptomatic CHD after adjusting for other cardiovascular risk factors. This analysis pooled repeated observations from the subjects’ triennial exams; each exam and its 3-year follow-up were analyzed as one observation in the analysis to determine a current risk profile for the development of disease.6,7 Because subjects’ medical conditions or clinical measurements may change over the course of the study, the following variables were treated as time-varying covariates: diabetes diagnosis, VV diagnosis, systolic blood pressure, diastolic blood pressure, cholesterol level, smoking status, alcohol consumption, and body mass index. Results for groups or are reported as mean ± SEM. All analyses were performed on SAS software licensed to Boston University.

RESULTS The NAS prospectively observed 2280 initially healthy male participants who provided 49,383 person-years of follow-up (range 3 to 35 years; mean of 21.7 years). During this period, 569 subjects (24.9%) were diagnosed with VV prior to the development of symptomatic CHD, and 1708 (75.1%) were not. For subjects diagnosed with VV subsequent to CHD development, analyses considered them to not have VV. Table I shows the distribution of cardiovascular risk factors at the time of subject entry into the

0.4 0.1 0.5 0.3 1.8

0.2

0.2 0.07 0.3 0.2 1.1

0.1

study, comparing subjects diagnosed with VV to those without VV. Subjects with VV were somewhat older, slightly heavier, had somewhat higher diastolic blood pressure, were more likely to report a family history of heart disease, and were more likely to smoke than subjects without VV. The two groups were comparable in frequency of diabetes, cholesterol levels, systolic blood pressure, amount of alcohol consumed, lifetime exposure to smoking, and education level. Results from a multivariate model addressing the effects of various cardiovascular risk factors on VV incidence, as measured at the study visit immediately prior to the diagnosis of VV, are reported in Table II. In this analysis, older subjects, those with lower cholesterol and systolic blood pressure levels, those with a family history of heart disease and smokers were more likely to develop VV. Over the 35 years of the study, 98 subjects with VV developed symptomatic CHD (17.2%) while 363 of the controls were diagnosed with symptomatic heart disease (21.2%). Figure 1 shows the Kaplan-Meier survival curves for survival free of symptomatic CHD according to presence or absence of VV. The incidence of symptomatic CHD was lower for those with VV during the 35-year followup (p = 0.0001, by log-rank test). Table III summarizes the multivariate analysis for the 1570 subjects with complete data. The risk profile for subjects who developed symptomatic CHD included older age, heavier weight, lower education level, hypercholesterolemia, and the report of a family history of heart disease. Those who develop heart disease were more frequently smokers at their most recent exam and reported elevated lifetime smoking exposure at baseline. Even after adjusting for all of the common cardiovascular risk factors listed in Table II, VV subjects had a significantly reduced risk of symptomatic CHD (RR = 0.64, 95% confidence limits 0.5–0.8; p = 0.0004).

462 Scott et al.

Annals of Vascular Surgery

Table II. Risk of varicose veins as measured by the time-dependent proportional hazards model Variable

Risk ratio

95% confidence interval

p

Age at baseline (10 years) Body mass index (5 kg/m2) Systolic blood pressure (10 mmHg) Diastolic blood pressure (10 mmHg) Cholesterol (10 mg/dL) Diabetes Family history of heart disease Baseline packs of cigarettes smoked over lifetime/1000 Current smoker Two or more drinks of alcohol/day High school education or more

1.5 1.07 0.9 1.1 0.97 0.7 1.3 0.99 1.3 1.03 0.93

1.4–1.7 0.9–1.2 0.8–0.93 0.98–1.3 0.95–0.99 0.4–1.06 1.05–1.5 0.98–1.01 1.01–1.6 0.8–1.3 0.8–1.1

0.0001 0.3 0.0001 0.09 0.002 0.09 0.01 0.7 0.04 0.8 0.4

Fig. 1. Kaplan-Meier distributions comparing the probability of survival free of symptomatic coronary heart disease in subjects diagnosed with varicose veins and those not diagnosed with varicose veins (log-rank test, p = 0.0001).

DISCUSSION In this study of 2280 men, Kaplan-Meier survival curves suggested a reduced risk of symptomatic CHD among men with preexisting VV. Further exploration of this relationship in a time-dependent proportional hazards multivariate model showed VV to be associated with a 36% decrease in risk of symptomatic CHD, even after adjusting for many other common cardiovascular risk factors. Several previous studies have explored the relationship between VV and CHD with inconsistent results. In an elderly Turkish population, both

angina and congestive heart failure were less frequent in individuals with VV, however, the latter relationship was not statistically significant.1 Responses to the Kaiser-Permanente Epidemiologic Study of Myocardial Infarction survey also found a negative association for both VV and hernias with myocardial infarction.2 In contrast, two prospective studies detected a positive effect for VV and CHD. The Framingham Heart Study reported a greater incidence of atherosclerotic disease in men with VV in univariate comparisons, however, this relationship disappeared in multivariate models once body mass index and systolic blood pressure were controlled for.3 The Paris Prospective Study also reported an elevated risk for heart disease and intermittent claudication in VV subjects followed for 7 years, but could not detect a statistically significant relationship with angina. However, when the Paris Prospective Study restricted their analysis to only the higher ranked police officers, they found VV to be protective for both coronary heart disease and angina.4 Reanalysis of our study population restricting the follow-up period to 7 years showed VV to be minimally protective and nonsignificant. Only with longer subject observation does the protective effect of VV on CHD development emerge. While the mechanism that explains why VV would prevent CHD development is unknown, one possible mediator between these diseases is estrogen. Estrogen has been suggested to be protective in females because of the decreased risk for CHD in premenopausal women. Elevated estrogen levels have also been associated with VV development in pregnant women.8 While estrogen levels might seem to be irrelevant in men, reports show that the cytochrome P450 enzyme, aromatase, converts testosterone to estradiol in men, resulting in serum levels similar to those found in postmenopausal

Vol. 18, No. 4, 2004

Varicose veins in men with CHD 463

Table III. Risk of symptomatic coronary heart disease as measured by the time-dependent proportional hazards model Variable

Risk ratio

95% confidence interval

p

Age at baseline (10 years) Body mass index (5 kg/m2) Systolic blood pressure (10 mmHg) Diastolic blood pressure (10 mmHg) Cholesterol (10 mg/dL) Diabetes Family history of heart disease Baseline packs of cigarettes smoked over lifetime/1000 Current smoker Two or more drinks of alcohol/day High school education or more Varicose veins

1.6 1.2 0.94 1.01 1.04 0.93 2.1 1.02 1.5 1.008 0.8 0.64

1.4–1.8 1.007–1.4 0.9–1.008 0.9–1.2 1.02–1.06 0.6–1.5 1.75–2.6 1.002–1.04 1.2–2 0.8–1.3 0.7–0.98 0.5–0.8

0.0001 0.04 0.08 0.9 0.0002 0.8 0.0001 0.03 0.002 0.9 0.03 0.0004

women.9 In addition, a recent double-blind clinical trial relates estradiol levels in men to arterial contractility. In this study, 20 healthy men were randomized to an aromatase inhibitor or placebo control. Results showed that the experimental group had lower serum estradiol levels along with an impaired flow-mediated dilation of the brachial artery than did controls.10 While these studies suggest a possible mechanism for how VV might be related to CHD, because we did not measure estrogen levels in our study population, this study provides no direct evidence to support or refute the relationships between estrogen levels, VV, and heart disease in men. A strength of this study is that it was prospectively designed and unbiased by selection for VV, because no preconceived hypothesis was present at the inception of data collection. However, the use of the Normative Aging Study has several limitations. All subjects were males, precluding generalization of the results to females. In addition, study subjects were relatively healthy at entry into the study, and as volunteers in a long-term health study, were likely to be fairly health-conscious. The NAS exam sequence frequency also resulted in early disease detection and intervention, which may explain our inability to detect some previously reported relationships, such as the elevated risk for heart disease in diabetics. In addition, antihypertensive medications might be eliciting preventive affects on CHD development over and above their action on hypertension. Furthermore, since we did not adjust for differences in diet or physical activity, it is possible that there is residual confounding. Finally, as with all large datasets of this type, misclassification and incomplete data can bias the re-

sults, but these problems would likely have occurred randomly and therefore would have biased our results toward the null.

CONCLUSION In the NAS population, men with VV were less likely to develop symptomatic CHD over the 35+ years of follow-up than were subjects without VV.

The VA Normative Aging Study and VA Dental Longitudinal Study are components of the Massachusetts Veterans Epidemiology Research and Information Center, supported by the VA Cooperative Studies Program. Dr. Garcia is a recipient of a VA Career Development Award in Health Services Research from the VA HSR&D Service, and NIH grant K24 DE-00419 from the National Institute of Dental and Craniofacial Research. We are grateful for the expert programming skills of Michael Cohen of the NAS.

REFERENCES 1. Komsuoglu B, Goldeli O, Kulan K, Cetinarslan B, Komsuoglu SS. Prevalence of risk factors of varicose veins in an elderly population. Gerontology 1994;40:25-31. 2. Klatsky AL, Friedman GD, Siegelaub AB. Medical history questions predictive of myocardial infarction: results from the Kaiser-Permanente epidemiologic study of myocardial infarction. J. Chron. Dis. 1976;29:683-696. 3. Brand FN, Dannenberg AL, Abbott RD, Kannel WB. The epidemiology of varicose veins: the Framingham Study. Am. J. Prev. Med. 1988;4:96-101. 4. Ducimetiere P, Richard JL, Pequignot G, Warnet JM. Varicose veins: a risk factor for atherosclerotic disease in middleaged men? Int. J. Epidemiol. 1981;10:329-335. 5. Shurtleff D (1974) Some Characteristics Related to the Incidence of Cardiovascular Disease and Death: Framingham

464 Scott et al.

Study, 18-year Follow-up. Bethesda: US Department of Health, Education, and Welfare, USDHEW publication NIH, 74–599. 6. Cupples LA, D’Agostino RB, Anderson K, Kannel WB. Comparison of baseline and repeated measure covariate techniques in the Framingham Heart Study. Stat. Med. 1988;7:205-218. 7. D’Agostino RB, Mei-Ling L, Belanger AJ, Cupples LA, Anderson K, Kannel WB. Relation of pooled logistic regression

Annals of Vascular Surgery

to time-dependent Cox regression analysis: the Framingham heart study. Stat. Med. 1990;9:1501-1515. 8. Callam MJ. Epidemiology of varicose veins. Br. J. Surg. 1994;81:167-173. 9. Anderson FH, Francis RM, Selby PL, Cooper C. Sex hormones and osteoporosis in men. Calcif. Tissue Int. 1998; 62:185-188. 10. Lew R, Komesaroff P, Williams M, Dawood T, Sudhir K. Endogenous estrogens influence endothelial function in young men. Circ. Res. 2003;93:1127-1133.