By Michael P. Pignone, M.D., M.P.H.a, Christopher J. Phillips, M.D., M.P.H.b, David Atkins, M.D., M.P.H.c, Steven M. Teutsch, M.D., M.P.H.d, Cynthia D. Mulrow, M.D., M.Sc.e and Kathleen N. Lohr, Ph.D.f
Address correspondence to: Michael Pignone, M.D., M.P.H., University of North Carolina at Chapel Hill, School of Medicine, Department of Medicine, 5039 Old Clinic Building, CB #7110, UNC Hospitals, Chapel Hill, NC 27599-7110. E-mail: pignone@med.unc.edu.
This article originally appeared in the American Journal of Preventive Medicine. Select for copyright and source information.
The summaries of the evidence briefly present evidence of effectiveness for preventive health services used in primary care clinical settings, including screening tests, counseling, and chemoprevention. They summarize the more detailed Systematic Evidence Reviews, which are used by the third U.S. Preventive Services Task Force (USPSTF) to make recommendations.
Abstract
Introduction
Epidemiology
Prior Recommendations
Methods
Results
Availability of Effective Screening tests
Reliability of Screening Tests
Lipid Levels and CHD Risk
Acceptability of Screening to Patients or Parents
Feasibility for Providers
Triglyceride Measurement
Adverse Effects of Screening for Lipid Disorders
Summary of Characteristics of Screening Tests
Frequency of Screening
Effectiveness of Drug Therapy
Effectiveness of Diet Therapy
Effectiveness of Exercise
Discussion
Middle-aged Men
Postmenopausal Women
Elderly Men and Women
Young Adults
Special Populations
Future Research
Summary: Whom to Screen and Treat
Acknowledgments
References and Notes
Context: Screening and treatment of lipid disorders in people at high risk for future coronary heart disease (CHD) events has gained wide acceptance, especially for patients with known CHD, but the proper role in people with low to medium risk is controversial.
Objective: To examine the evidence about the benefits and harms of screening and treatment of lipid disorders in adults without known cardiovascular disease for the U.S. Preventive Services Task Force.
Data Sources: We identified English-language articles on drug therapy, diet and exercise therapy, and screening for lipid disorders from comprehensive searches of the MEDLINE database from 1994 through July 1999. We used published systematic reviews, hand searching of relevant articles, the second Guide to Clinical Preventive Services, and extensive peer review to identify important older articles and to ensure completeness.
Data Synthesis: There is strong, direct evidence that drug therapy reduces CHD events, CHD mortality, and possibly total mortality in middle-aged men (35 to 65 years) with abnormal lipids and a potential risk of CHD events greater than 1% to 2% per year. Indirect evidence suggests that drug therapy is also effective in other adults with similar levels of risk. The evidence is insufficient about benefits and harms of treating men younger than 35 years and women younger than 45 years who have abnormal lipids but no other risk factors for heart disease and low risk for CHD events (less than 1% per year). Trials of diet therapy for primary prevention have led to long-term reductions in cholesterol of 3% to 6% but have not demonstrated a reduction in CHD events overall. Exercise programs that maintain or reduce body weight can produce short-term reductions in total cholesterol of 3% to 6%, but longer-term results in unselected populations have found smaller or no effect. To identify accurately people with abnormal lipids, at least two measurements of total cholesterol and high-density lipoprotein cholesterol are required. The role of measuring triglycerides and the optimal screening interval are unclear from the available evidence.
Conclusions: On the basis of the effectiveness of treatment, the availability of accurate and reliable tests, and the likelihood of identifying people with abnormal lipids and increased CHD risk, screening appears to be effective in middle-aged and older adults and in young adults with additional cardiovascular risk factors.
Keywords: cardiovascular disease; cholesterol; coronary disease; hyperlipidemia; preventive health services; evidence-based medicine; MEDLINE; methods; lipids; mass screening; mortality; drug therapy.
Elevated low-density lipoprotein cholesterol (LDL-C) and low levels of high-density lipoprotein cholesterol (HDL-C) are important risk factors for coronary heart disease (CHD) (1,2,3). CHD is the leading cause of morbidity and mortality in the United States, causing nearly 500,000 deaths each year and requiring nearly 12 million hospital days of care per year. It is the leading cause of disabled life-years and is second only to injuries as a cause of life-years lost (4). The lifetime risk of having a CHD event, calculated at age 40 years, is estimated to be 49% for men and 32% for women in the United States (5). CHD accounted for $78 billion in health care costs in 1995 (4).
Lipid disorders are common in the United States and other Western, developed countries. Data from the National Center for Health Statistics collected from 1988 through 1994 show that 17.5% of U.S. men and 20% of U.S. women aged 20 to 74 years had total cholesterol (TC) levels greater than 240 mg/dL (6). After adjusting for the effect of other risk factors, an analysis from a large U.S. cohort study estimated that 27% of CHD events in men and 34% in women were attributable to TC levels greater than 200 mg/dL (7).
Figure 1 (11 KB) shows mean TC levels by age for men and women. In adults, mean TC increases with age for both men and women (8). In men, mean TC increases steadily from early adulthood to middle age and then reaches a plateau, falling only in men older than age 75 years. Mean TC is initially lower in premenopausal women than in men, but it rises at a similar rate. After menopause, however, women experience an additional 10- to 20-mg/dL rise, and their mean TC remains higher than for men throughout the remainder of life. HDL-C levels do not change greatly throughout adulthood and are consistently higher in women than in men (9). Mean TC is similar for those identifying themselves as Caucasian or African American (10). HDL-C is higher for African Americans than for Caucasians (Figure 2, 11 KB).
Large observational cohort studies have found a strong, graded relationship between increasing levels of LDL-C or decreasing levels of HDL-C and increasing risk of CHD events (1,2). The increased risk for CHD events is continuous, linear, and graded: No clear "cut-off" value separates normal from abnormal values. A 50-year-old man with a blood pressure of 120/80 mmHg, a TC of 180 mg/dL, and an HDL-C of 40 mg/dL has a 10-year risk for CHD events of 7%. If the same man had a TC of 240 mg/dL and an HDL-C of 30 mg/dL, his 10-year risk would be 14%, a relative risk of 2.0 and an absolute risk difference of 7% (7).
The total excess risk for CHD from lipid disorders depends on the presence of other risk factors. A 50-year-old man with hypertension (blood pressure of 160/90 mmHg) who smokes and has a TC of 180 mg/dL and an HDL-C of 40 mg/dL has a 10-year risk for CHD events of 17%. If the same man had a TC of 240 mg/dL and an HDL-C of 30 mg/dL, his risk would increase to 29%, an absolute difference of 12%.
Observational studies suggest that lipid disorders confer less relative risk of CHD events in the elderly than in other age groups. The absolute risk of CHD is higher for the elderly, however, and thus the total number of potentially preventable CHD events remains high for the elderly (11).
The second edition of the Guide to Clinical Preventive Services from the U.S. Preventive Services Task Force (USPSTF) gave a "B" recommendation to "periodic" screening for high TC in men aged 35 to 65 years and women aged 45 to 65 years (12). The USPSTF at that time found that the evidence was insufficient to recommend for or against TC screening in asymptomatic adults older than 65 years of age, young adults, adolescents, and children. They also found evidence to be insufficient to recommend for or against screening for other lipid abnormalities such as low HDL-C or elevated triglycerides.
The National Cholesterol Education Program Adult Treatment Panel II (ATP II) recommended screening all adults aged 20 years and older every 5 years with serum TC and with serum HDL-C "if accurate results are available" (3). New recommendations from the ATP III are to be published in 2001. The Canadian Task Force on Preventive Health Care in 1994 recommended "case-finding" in all men aged 30 to 59 years who present to their health care providers and clinical judgment in other cases (13). The American College of Physicians found "periodic" screening for men aged 35 to 65 years and women aged 45 to 65 years to be "appropriate but not mandatory"; screening young men and women was recommended only when the history or physical examination suggested a familial disorder or when the person had at least two other risk factors (14,15). The American Diabetes Association recommended screening all adults with diabetes yearly with TC, LDL-C, HDL-C, and triglycerides (16).
To examine the role of practice-based screening for lipid disorders in adults without known cardiovascular disease, we first developed an analytic framework and key questions (Figure 3, 11 KB). The four key questions were:
We next identified English-language articles on drug therapy, diet and exercise therapy, and screening for lipid disorders from comprehensive searches of the MEDLINE database from 1994 through July 1999. We used published systematic reviews, hand searching of relevant articles, the second Guide to Clinical Preventive Services (12), focused searches of MEDLINE from 1966 through 1993, and extensive peer review to identify important older articles and to ensure completeness.
We included all randomized trials of at least 1 year's duration that examined drug or diet therapy among patients without previously known CHD and that measured clinical end points, including total mortality, CHD mortality, and nonfatal myocardial infarctions (MIs), as well as randomized trials of diet or exercise therapy that measured change only in cholesterol levels. We included articles that examined the epidemiology and natural history of lipid levels and lipid disorders and ones that measured the accuracy, reliability, acceptability, and feasibility of screening. We also included any articles that examined adverse effects and harms of screening or therapy for lipid disorders.
Full details of the methods and results are available in the Systematic Evidence Review available from the Agency for Healthcare Research and Quality (17).
Several different screening strategies have been proposed for identifying lipid disorders, including screening with TC alone, the ratio of TC to HDL-C (TC/HDL-C), and the ratio of LDL-C to HDL-C (LDL-C/HDL-C). These measures can be used alone to determine risk and the need for treatment. Alternatively, they can be combined with information about the presence or absence of other CHD risk factors, as has been done with the ATP II guidelines (3). They can also be incorporated into a quantitative risk-based screening strategy; in this approach, each person's overall risk for CHD is calculated by using a risk assessment table or computer program, and treatment is recommended for risk levels above a defined risk threshold.
TC measurements from venous blood samples generally have good reliability. The analytic variability for TC is less than 3%; the mean total biologic variability for TC is about 6% (18). Two separate measurements are required to determine a patient's TC level within 10% of the true value. TC levels do not vary substantially between fasting and nonfasting periods; hence, TC can be measured clinically at any time.
HDL-C has higher analytic (6%) and biologic (7.5%) variation than total cholesterol. Two or three values are required to estimate confidently the true level within 10% to 15%. HDL-C in the nonfasting state is lower by 5% to 10% than in the fasting state. Nonfasting measurement may, therefore, slightly overestimate CHD risk, but it is considered sufficiently accurate for use in screening (19). Combined measures such as the TC/HDL-C ratio will be less reliable than each individual measure, but it can also be improved by averaging two or more individual values.
Triglycerides change by 20% to 30% between fasting and nonfasting states. Because LDL-C is routinely calculated indirectly by measuring TC, HDL-C, and triglycerides (TG) and then applying the Friedewald equation (TC = HDL-C + LDL-C + [TG/5]), accurate calculation of the LDL cholesterol level requires a fasting sample to ensure accurate measurement of triglycerides (18). The Friedewald equation produces inaccurate results when triglyceride levels exceed 400 mg/dL, so patients with very high triglyceride levels may need special techniques (e.g., ultracentrifugation) to measure LDL-C accurately.
Capillary blood samples that are used to measure total and HDL-C (so-called "point of care" testing) appear to have similar reliability under optimal conditions to venous samples but may be less reliable if proper attention is not paid to calibration and proper testing technique (20).
An important objective in screening for lipid disorders is to identify accurately which patients are (or are not) at high risk of experiencing CHD events. The amount of CHD risk attributable to abnormal lipids depends on the degree of lipid abnormality and the presence of other CHD risk factors. Several means of assessing the extent of lipid abnormality are available, including measurement of individual lipid components (TC, HDL-C, LDL-C) or ratios of such components (e.g., TC/HDL-C).
Strategies that explicitly consider a person's other CHD risk factors in addition to his or her lipid levels are more accurate than those that measure only lipid levels (7). Grover et al. (21) found that a Framingham-based coronary risk model was the best predictor of CHD mortality. The ATP II guidelines, the LDL-C/HDL-C ratio, and the TC/HDL-C ratio performed approximately equally well. TC alone was the least accurate (Table 1, Text Version).
The acceptability of screening for lipid disorders in adults has been quite high. Obtaining a nonfasting sample (for measurement of TC, HDL-C, or both) at a regular health care visit is the easiest method. Obtaining a fasting sample (which may require a separate visit or change in usual eating habits) is somewhat more taxing, but apparently most patients (more than 80%) will return for such testing when requested to do so (23). The acceptability to patients of the ATP II screening guidelines or an explicit risk-based approach is presumably no different than a nonfasting blood draw alone because the extra work is required of the physician, not the patient.
Screening for lipid disorders by measuring cholesterol levels in adult patients is quite feasible for physicians because it involves only ordering a blood test. Providers appear to have achieved high levels of lipid screening based on population-based patient survey data. Data from primary care practices, however, suggest that screening may not be directed preferentially to those patients who are at highest risk and thus most likely to benefit from treatment (24). The feasibility of routinely using the ATP II guidelines or a risk-based screening tool may be lower, as each requires providers to collect and integrate several pieces of health information (25).
The question of whether an elevated triglyceride level is an independent risk factor for CHD remains controversial (26,27). Even if elevated triglycerides are independently associated with an increased risk of CHD, the question of whether treating people with isolated increased triglycerides will reduce future CHD events is still unclear.
Screening for and identifying lipid disorders in adults do not appear to have important psychological sequelae or to produce important changes in indices of mental health. The research to date has not been sufficient, however, to rule out important changes in small subsets of patients or to detect subtle changes in anxiety (28-31). Patients who are identified as having acceptable lipid levels may have a theoretical disincentive to follow or to adopt healthy dietary habits, which could adversely affect their risk for other illnesses not mediated through lipid levels, but this effect has not been well studied.
Nonfasting TC alone is the least expensive and easiest test to perform for both patient and provider, but its accuracy is lowest. The TC/HDL-C ratio alone is also easy for patients to obtain and moderately easy for providers to interpret. It performs as accurately as the ATP II guideline-based strategy. The LDL-C/HDL-C ratio or ATP II-based predictions perform no better than the TC/HDL-C ratio and may be more difficult for patients and providers.
Risk-based algorithms, such as those based on the Framingham cohort study, that directly incorporate age, the presence and magnitude of other risk factors, and measures of TC and HDL-C are the most accurate approach to screening, but they are more difficult for providers to implement without assistance because they require them to integrate several different pieces of information (7). Using a supplemental table such as the Sheffield Tables (32) or a simple computer program (33) may improve the feasibility of a risk-based strategy.
Good data directly comparing the prospective performance, costs, and marginal cost-effectiveness of the different approaches are not currently available. As initial screens, for example, we cannot say definitely whether the extra accuracy gained by universally measuring HDL-C and calculating the TC/HDL-C ratio justifies the cost difference between this measure and the use of TC alone.
No direct data inform the question of appropriate frequency of screening. Chiefly for that reason, previous USPSTF recommendations did not state a preferred interval (12). By contrast, ATP II recommendations suggested a 5-year interval for people with previous normal results and more frequent screening for those who have borderline values (3).
Several factors enter into a decision about screening frequency. These factors include the usual rates of change in cholesterol levels over time, the variability of individual cholesterol measurements, the likelihood of finding a result that would lead to a change in management (particularly values that are close to treatment thresholds), and the feasibility and costs of different frequencies of screening. A universal 5-year interval, for example, is simple to implement, but it may impose more frequent screening than is necessary on patients with few or no other risk factors and low-risk values on previous screening measurements. Using a more variable algorithm in which patients' frequency of screening would be related to their previous results could be more efficient for diagnosis, but this approach may be confusing or difficult to implement.
Effects of drug therapy on CHD events. We identified four trials of drug therapy for lipid disorders in the primary prevention of CHD. These include two older (pre-1995) trials: one using the bile-acid binding resin cholestyramine (Lipid Research Council [LRC] trial) (34) and one (Helsinki Heart Study [HHS]) using the fibric acid derivative gemfibrozil (35). The other two trials were published after 1995 and used hepatic 3-methylglutaryl coenzyme A reductase inhibitors or "statin" drugs: The West of Scotland Coronary Prevention Study (WOSCOPS) used pravastatin (36) and the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS-TexCAPS, hereafter AFCAPS) used lovastatin (37). Table 2 (Text Version) describes the study design and patient characteristics for these four trials; Table 3 (Text Version) provides key results.
The four trials were conducted mainly among middle-aged men of European descent. The LRC, HHS, and WOSCOPS trials enrolled patients with elevated levels of TC and LDL-C, whereas the AFCAPS study included men and women with TC levels close to the U.S. average but low levels of HDL-C. Few diabetic patients were enrolled in any of the four trials. The trials lasted from 5 to 7 years. All examined the effect of drug therapy on the incidence of CHD events, including CHD mortality, using a placebo-controlled, double blind methodology. In each trial, the intervention and control groups both received low-intensity dietary interventions.
The two trials employing statin drugs (WOSCOPS and AFCAPS) had larger initial decreases in TC (20% and 18%) than the LRC or HHS (15% and 9%). The relative risk reductions for CHD events ranged from 19% to 37% and for CHD mortality from 20% to 28%. No trial was designed with sufficient power or duration to address confidently the question of whether drug therapy reduces total mortality.
WOSCOPS, which examined the highest-risk population among the four studies, demonstrated that treating middle-aged men with elevated LDL-C and a baseline risk of CHD events of about 1.5% per year decreased the relative risk of CHD events by 31% and total mortality by 22%. The absolute risk reduction for total mortality, however, was small (0.9%), suggesting that approximately 111 patients at similar risk would need to be treated for 5 years to prevent one death (36).
Meta-analysis. The combined results of the four main trials suggest that drug therapy decreases the risk of total CHD events (defined as the sum of nonfatal MIs and deaths from CHD) by 30% (95% confidence interval [CI]=20% to 38%) (38). Drug therapy also reduces the risk of CHD death by 26% (95% CI=2% to 43%). Drug therapy appears to have little overall effect on total mortality for the 5 to 7 years over which these trials were conducted (odds ratio [OR]=0.91; 95% CI=0.78 to 1.07). However, the overall result may mask a total mortality benefit in higher-risk patients. The WOSCOPS trial found a 22% relative reduction in total mortality at borderline statistical significance (p=0.051). In the other three trials, drug therapy appeared to confer no total mortality benefit. Repeat analyses, using data from the two statin trials alone, produced slightly larger estimates of effect on CHD events and CHD mortality but still no clear effect on total mortality.
Effect of drug therapy on strokes. Drug therapy reduces the incidence of total strokes in people with known CHD by about 30% (39). A meta-analysis of three primary prevention studies found a 20% decrease in total stroke in incidence (OR=0.80; 95% CI=0.54 to 1.16) that did not reach statistical significance (40). Another meta-analysis of statin trials conducted before the AFCAPS trial was published produced a similar result for total strokes in primary prevention trials (OR=0.85; 95% CI=0.57 to 1.28) (41).
Harms of drug therapy for lipid disorders. On the basis of data from multiple clinical trials and 10 years of experience with adverse drug reporting, statins appear to have few important short- or medium-term (initiation to 5 years) adverse effects (17). Myopathy and muscle pain appear to occur infrequently (in about 1 in 500 to 1 in 1000 users). Elevations in liver enzyme levels, which some studies have noted, have not been found in recent large trials and do not seem to produce clinically important consequences.
In observational studies, hemorrhagic stroke appears to occur more frequently in patients with low TC levels, but it has not been sufficiently studied in treatment trials to conclude that it is increased in patients who have had their cholesterol levels lowered with statins or other drug therapy. Data from one recent secondary prevention study suggest that, although the incidence of total stroke is decreased by drug therapy, the rate of hemorrhagic stroke may be increased (approximate relative Risk=1.7; 95% CI=0.8 to 3.2) (41).
The safety of statin drugs in the long run remains unclear because long-term experience is insufficient to rule out rare but serious consequences of prolonged therapy. Other agents used for lipid disorders, including gemfibrozil, niacin, and bile-acid binding resins, have some minor adverse effects (e.g., gastrointestinal upset for gemfibrozil or bile-acid binding resins; flushing for niacin) or rare major effects (e.g., liver failure for extended-release niacin). The safety experience for bile-acid binding resins and niacin, however, is based on a longer period of time than is the case for the statin drugs.
Summary of drug therapy effects. Drug therapy for lipid disorders reduces the relative risk for CHD events by 30%. Statin drugs have produced larger reductions in cholesterol and appear to reduce events more than the older drugs. The absolute risk reduction with drug therapy depends on the underlying risk in the person or population being treated. Drug therapy appears to have little effect on total mortality after 5 to 7 years of treatment in lower-risk patients (risk of CHD events less than 1.5% per year), but mortality may be reduced in higher-risk populations or with longer follow-up. Short- to medium-term adverse effects appear uncommon with statins, but long-term effects are unknown. Women, elderly people (older than 70 years), and people of non-European descent appear to have similar relative risk reductions for total CHD events with drug treatment, although they have been studied less than middle-aged men.
The relationships among diet, cholesterol, and heart disease have been demonstrated in numerous ecologic and observational studies. In the United States, broad changes over the past 30 years in dietary patterns, particularly the consumption of saturated fat, have been accompanied by reductions in the population's average TC levels (10). In addition, individualized dietary interventions (some, but not all, of which lower TC) have been shown to reduce CHD events in patients with known cardiovascular disease or who have been treated in institutional settings (17). In this section, we examine the effectiveness of diet therapy for preventing CHD events and for reducing cholesterol levels among free-living people without previously diagnosed CHD.
Effect of diet therapy on CHD events. No studies in primary care settings examined the effect of dietary advice therapy on actual CHD events among patients with abnormal lipids but no previous history of CHD. Ebrahim and Smith (42) performed a systematic review and meta-analysis of nine multiple risk factor intervention randomized trials of at least 6 months' duration that examined the effect of diet therapy on CHD events and lipid levels. The median duration was 5 years. The interventions did not reduce total mortality (OR=0.97; 95% CI=0.92 to 1.02), CHD mortality (OR=0.96; 95% CI=0.89 to 1.04), or nonfatal MIs (OR=1.0; 95% CI=0.92 to 1.07). The net effect on serum cholesterol was a reduction of 5.4 mg/dL.
Effect of diet therapy in reducing total cholesterol. It is clear that alterations in diet can affect cholesterol levels. A systematic review of studies conducted on metabolic wards found that dietary therapy can produce short-term TC decreases of 10% to 20% when patients are fed a controlled low-fat diet (43). However, the ability of outpatient dietary interventions to produce sustained reductions in TC is less clear (44). Tang et al. (45) performed a meta-analysis of single intervention dietary trials conducted among free-living adults and published before 1996. Trials of patients with known CHD and trials conducted in non-primary-care settings were included; trials of specific dietary supplements (e.g., oat bran, garlic) and multiple risk factor trials were excluded. For trials of at least 6 months' duration, the mean reduction in cholesterol at 12 months was 5.3%. The subset of studies using the American Heart Association Step I diet, advocated as the first intervention for patients with no previous CHD, produced an average reduction of 3.0%. Brunner et al. (46) found a similar result (mean reduction of 3.7%) in their meta-analysis of 17 studies. We identified a subset of six studies that specifically examined the effect of diet therapy provided in primary care settings. They found mean TC decreases of 2% to 3% (47-54).
Effect of learning one's cholesterol level on the effectiveness of diet therapy. A proposed rationale for screening for lipid disorders, particularly in young adults, has been that knowledge of one's cholesterol level may improve adherence to dietary advice and may increase its effect on lipid levels. Four trials published between 1992 and 1998 examined the effect of learning one's cholesterol level on the effectiveness of dietary therapy to lower TC (55-58). In three studies, subjects were volunteers recruited from work sites; in the fourth, subjects were patients in a British primary care clinic. In three trials, subjects learning their cholesterol level had no net improvement in TC with dietary therapy compared with subjects who were not given their results. In the trial by Elton et al. (55), subjects with high cholesterol (mean cholesterol 277 mg/dL) on initial screening had modest TC reductions with feedback compared with controls (3.9% net reduction), but patients with more modest levels did not.
Summary of diet therapy effects. To date, diet therapy has not been demonstrated to reduce CHD events in free-living primary prevention populations (42). Controlled studies have generally achieved only modest long-term reductions in TC (3% to 6% for trials longer than 6 months), despite relatively intensive interventions. The small cholesterol reductions in primary prevention appear to be a result of incomplete adherence (45). Data are insufficient to determine in advance which patients are most likely to achieve and maintain important reductions in cholesterol. Knowledge of one's cholesterol level does not appear to affect the overall effect of dietary therapy, although people with elevated cholesterol may be slightly better able to reduce their total cholesterol.
Observational epidemiologic studies have found that people who are physically active have lower rates of CHD than people who are inactive (58). Whether these findings can be translated into successful and feasible interventions to lower CHD risk is not clear; no trials of exercise done in primary prevention settings have found decreased CHD events among those assigned to exercise.
Many studies have examined the effect of exercise on CHD risk factors, including lipid disorders. A meta-analysis of 95 studies found that subjects assigned to exercise had TC levels after intervention that were 7 mg/dL to 13 mg/dL (3% to 6%) lower than controls (59). The larger reductions occurred among patients who were able to lose weight; the smaller reductions occurred among those with no weight change. Those reporting weight gain had a small (3 mg/dL), statistically nonsignificant increase in TC. HDL cholesterol levels increased by an average of 2 mg/dL and were not affected by the amount of weight loss.
Exercise interventions have not been adequately evaluated as a means of reducing CHD events in primary prevention. They do not appear to have a large effect on lipid levels, although some studies employing rigorous activity prescriptions and producing weight loss have shown changes in lipid profiles that may be clinically meaningful. These programs, however, have been difficult to implement widely.