SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References

An elevated level of non-high-density lipoprotein cholesterol (non-HDL-C) is a major risk factor for cardiovascular disease. The purpose of this study was to use the meta-analytic approach to examine the effects of walking on non-HDL-C in adults. Twenty-two randomized controlled trials representing 30 outcomes from 948 subjects (573 exercise, 375 control) met our inclusion criteria. Across all designs and categories, random effects modeling resulted in a significantly greater decrease in the walking group when compared with the control group of approximately 4% for non-HDL-C (± standard error of the mean, −5.6±1.8 mg/dL, 95% confidence interval, −8.8 to −2.4 mg/dL). Meta-regression showed a statistically significant association between changes in non-HDL-C and the year of publication, with greater reductions associated with more recent publication year (R2=0.23, p=0.005). The results of this meta-analytic review suggest that walking reduces non-HDL-C in adult humans.

It has been suggested that non-high-density lipoprotein cholesterol (non-HDL-C) may be a better predictor of cardiovascular disease (CVD) morbidity and mortality than low-density lipoprotein cholesterol because it contains all known lipid particles considered to be atherogenic (low-density lipoprotein cholesterol lipoprotein (a), intermediate-density lipoprotein, very-low-density lipoprotein).1,2 Recommendations from the recent National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Report3,4 suggest lifestyle therapies, including increased physical activity, for improving lipid and lipoprotein levels in adults. One type of physical activity that may be particularly appropriate is walking, a nonpharmacologic intervention that can be done without any additional cost, is readily accessible and has a low rate of injury.3,4 In addition, walking is currently the most common type of physical activity that adults in the United States participate in with approximately 34% of both men and women 18 years of age and older reporting regular walking, defined as walking at least 5 times/wk for at least 30 min/session.5 Unfortunately, we are not aware of any randomized controlled trials or metaanalytic reviews of literature that have examined the effects of walking on non-HDL-C levels in adults. However, a number of randomized controlled trials that have examined the effects of walking on lipids and lipoproteins in adults have included assessments of total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-C), values from which non-HDL-C can be calculated.6–34

Thus, given the importance of non-HDL-C as a CVD risk factor, the prevalence of walking, and the absence of randomized controlled trials and metaanalytic reviews dealing with the effects of walking on non-HDL-C, the purpose of this study was to use the meta-analytic approach to examine the effects of walking on non-HDL-C in adult humans aged 18 years and older.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References

Data Sources

Studies for this meta-analysis were retrieved using: 1) computerized databases (MEDLINE, EMBASE, SportDiscus, Current Contents, Dissertation Abstracts International), 2) cross-referencing from retrieved studies and review articles, 3) hand searching selected journals, and 4) examination of our reference database by an independent expert on exercise, lipids, and lipoproteins (Dr. William Haskell, personal written communication, May 7, 2003).

Study Selectiona

The inclusion criteria for this meta-analysis were as follows:

  • 1
    randomized controlled trials;
  • 2
    walking exercise of at least 8 weeks duration as an intervention;
  • 3
    adult humans aged 18 years and older;
  • 4
    studies published in journal, dissertation, or master's thesis format;
  • 5
    studies published in the English language;
  • 6
    studies published from January 1, 1955 to January 1, 2003; and
  • 7
    initial and final assessment of both TC and HDLC, so that non-HDL-C could be calculated.

Data Abstraction

A coding sheet was developed that could hold more than 200 items per study. The major categories of items coded included study, subject, and training program characteristics as well as methods used to assess TC and HDL-C. All studies were coded by the first two authors, independent of each other. The authors then reviewed every item for accuracy and consistency (Cohen's kappa=0.95, 95% confidence interval, [CI]=0.945 to 0.950).35 Any discrepancies that occurred were resolved by consensus. If no consensus could be reached, the third author served as an arbitrator.

Statistical Analysis

Primary Outcome . The primary outcome in this study was changes in non-HDL-C, calculated by taking the difference between TC and HDL-C. Variances for these calculated values were determined using previously developed procedures by Follmann et al.36 Net changes in non-HDL-C were calculated as the difference (exercise-control) of the changes (initial-final) in these mean values. Variances for each group (exercise and control) were calculated from variances at baseline and final measurement using standard methods.36 Pooled outcomes for non-HDL-C were calculated by assigning weights equal to the inverse of the total variance for net changes in non-HDL-C. A random effects model was used for all analyses.37–39 Study quality was examined using an index that has previously been reported to be both valid (face validity) and reliable (r=0.77).40–41 This assessment is a three-item questionnaire designed to assess bias, specifically, randomization, blinding, and withdrawals/dropouts. The minimum number of points possible ranges from a low of 0 to a high of 5, with higher numbers representing greater study quality. To examine the influence of each study on our overall results, analyses were also performed with each study deleted from the model once. Publication bias was examined using the procedure of Duval and Tweedie.42,43

Subgroup analyses were performed for categorical variables using analysis of variance-like procedures designed specifically for meta-analysis.44 Categorical variables that were analyzed included source of study, country in which the study was conducted, menopausal status, presence of diabetes and/or overweight/obesity, gender, and initial non-HDL-C levels (>220 mg/dL vs. =220 mg/dL partitioned according to available data). In addition, we also examined whether the subjects were classified by the authors as apparently healthy vs. having one or more of the following conditions: hyperlipidemia, diabetes, overweight/obese, and/or CVD. We were unable to conduct separate analyses for hyperlipidemia and CVD because of a lack of data. Because of a lack of data, we were unable to examine the effects of drug use for elevated lipids, cigarette smoking, alcohol consumption, diet, and position in which lipid assessment took place on changes in non-HDL-C.

To examine for potential relationships between continuous variables and non-HDL-C, simple, weighted, generalized least-squares meta-regression was performed.44 Variables that were regressed included the year of publication; study quality; percentage of dropout; initial level of non-HDLC; age in years; height; initial, as well as changes in, body weight; body mass index (BMI); percentage of body fat; maximum oxygen consumption (VO2max) in mL/kg-1min-1; number of hours that subjects fasted before lipid assessment; number of hours that exercise was avoided before lipid assessment; length, frequency, intensity, duration, and total minutes of training (length x frequency x duration); and compliance to the exercise protocol. Multiple regression analysis was not possible because of missing data from different studies and different variables.

Secondary Outcomes . Secondary outcomes (TC, HDL-C, BMI, body weight, percentage of body fat, VO2max) were calculated using the same general procedures as those used for examining non-HDL-C.

Data Reporting and Significance Levels . Descriptive data are reported as mean (x̄) ± standard deviation (SD) while outcomes data are reported as x̄± standard error of the mean (SEM). Ninety-five percent CIs were used to determine the statistical significance of our overall results for both primary and secondary outcomes. If the CI did not include zero (0.00), the result was considered to be statistically significant. Because of the multiple comparisons performed for analysis of variance and regression analyses, the a level was set at p<0.01.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References

Study Characteristics

A total of 29 studies met our criteria for inclusion,6–34 however, seven studies were excluded because of missing data necessary for the calculation of non- HDL-C,11,12,14,21,29,30,34 which resulted in a loss of 24% for studies that met our inclusion criteria. Thus a total of 22 studies that included 52 groups (30 exercise, 22 control), 948 subjects (573 exercise, 375 control), and 30 non-HDL-C outcomes were included in the analysis.6–10,13,15–20,22–28,31–33 The number of subjects in each study ranged from 5–100 in the exercise groups (± SD=19±18) and 5–104 in the control groups (± SD=17±20). Studies were conducted in the United States, United Kingdom, Canada, and Japan. The percentage of subjects in which final lipid data were not available ranged from 0%–67% in the exercise groups (x̄± SD=26.2±18.0%) and 0%–40% in the control groups (x̄± SD=13.2±14.6%). All of the studies appeared to use an analysis-by-protocol approach in the analysis of their data. Study quality ranged from 0–5 (x̄± SD=2±1).

Subject Characteristics

A description of the physical characteristics of the subjects is shown in Table I. The subjects included male as well as pre- and postmenopausal females. The majority of subjects included in the studies were white, however, blacks, Hispanics, Japanese, and Asians were also represented.

Table I.  Initial Characteristics of Subjects
VariablenWalkersx̄± SDRangenControlsx̄± SDRange
Age (yr)3048.6±13. 630.0–76.02249.9±13.629.4–74.0
Height (cm)17165.2±4.4159.0–177.011164.7±5.9158.0–177.0
Weight (kg)2871.0±7.560.3–88.42072.3±7.863.1–90.5
BMI (kg/m2)2326.6±3.422.6–36.21627.3±3.722.9–34.0
Body fat (%)1533.1±5.626.2–43.11135.3±6.027.5–43.4
VO2max(mL/kg–1/min–1)2326.5±5.017.3–35.91526.4±6.617.8–35.2
TC (mg/dL)30207.3±22.0178.6–256.722209.9±23.0171.7–264
HDL-C (mg/dL)3055.6±6.144.0–67.32253.8±7.439.4–70.4
Non-HDL-C (mg/dL)30151.8±21.0116.3–204.522156.1±23.5116.4–213.4
N=number or groups reporting data; BMI=body mass index; VO2max=maximum oxygen consumption; TC=total cholesterol; HDL- C=high-density lipoprotein cholesterol; non-HDL-C=non-high-density lipoprotein cholesterol, calculated as total cholesterol- high-density lipoprotein cholesterol6

Also included in the studies are cigarette smokers, hyperlipidemics, non-insulin-dependent diabetics, subjects who consumed alcohol, and overweight/obese subjects. With the exception of one study, which reported that most of the subjects had altered their diet in a manner to include more carbohydrates and less fat,31 no changes in diet that might affect non-HDL-C were reported.

Lipid Assessment Characteristics

The number of hours that subjects fasted before the morning assessment of lipids and lipoproteins ranged from 9–13 (x̄± SD=11.8±1.0 h) while the number of hours that exercise was avoided before lipid assessment ranged from 12–72 (x̄± SD=30.1±19.7 h). Six studies reported that multiple measurements were used in the calculation of lipid and lipoprotein values at the initial and final testing periods.9,22–25,28 Two studies reported the assessment of lipids and lipoproteins in the sitting position,23,25 while one reported the assessment in the supine position.24

Training Program Characteristics

Length of training ranged from 10–104 weeks (x̄± SD=22.5±17.8) and frequency from 3–15 sessions/wk (x̄± SD=4.9±2.6). One study included two groups that performed multiple walking sessions during each day, accounting for an average of 11 and 15 walking sessions/wk.33 Intensity, defined as a percentage of maximum oxygen consumption, ranged from 50%–86% (x̄± SD=64.9±9.3%), while duration of training ranged from 10–75 min/session (x̄± SD=38.4±16.4). Compliance to the walking programs ranged from 52%–100% (x̄± SD=85.0±16.1). Compliance data were available for only 43.3% of the walking groups.

Primary and Secondary Outcomes

Changes in non-HDL-C are shown in Table II. As can be seen, random effects modeling resulted in statistically significant exercise-control decreases that were equivalent to a relative decrease of approximately 4% for non-HDL-C. With each study deleted from the model once, results remained statistically significant, ranging from ± SEM, -4.8±1.7 mg/dL (95% CI, -8.1 to -1.5 mg/dL) to ± SEM, -6.1±1.6 (95% CI, -9.2 to -2.9 mg/dL). Examination for potential publication bias resulted in no adjustments for the numbers and outcomes from missing studies. Thus no publication bias was observed. No statistically significant differences were observed for any of the a priori subgroup analyses performed (p>0.05). Simple meta-regression resulted in a statistically significant association between changes in non-HDL-C and the year of publication, with greater reductions associated with a more recent publication year (R2=0.23, p=0.005). Because of this statistically significant association, we also conducted, post hoc, a simple meta-regression test to examine whether an association existed between the year of publication and initial levels of non-HDL-C, however, no statistically significant association was observed (R2=0.02, p=0.50). A statistically significant association was also observed between changes in non-HDL-C and changes in BMI (R2=0.21, p=0.004), with greater reductions in BMI associated with greater reductions in non-HDL-C, however, no relationship was observed between changes in non-HDL-C and initial levels of BMI (R2=0.02, p=0.53). Because of the relationship between changes in non-HDLC and changes in BMI, we also compared, post hoc, changes in BMI when partitioned according to whether or not subjects had decreased their BMI during the study, however, no statistically significant difference between the two groups was found (Qb=1.52, p=0.22). No other statistically significant associations were observed for any of the other variables examined, including baseline levels of non-HDL-C and length of training (p>0.05).

Table II.  Primary and Secondary Outcomes
Variablenx̄± SEM95% CIChange (%)
 Primary Outcome    
Non-HDL-C (mg/dL)30–5.6±1.6–8.8 to -2.4*–4
 Secondary Outcomes    
TC (mg/dL)30–3.7±1.3–6.2 to −1.1*–2
HDL-C (mg/dL)301.4±0.8–0.1 to 2.83
Weight (kg)23–0.9±0.3–1.6 to −0.3*–2
BMI (kg/m2)9–0.6±0.3–1.1 to −0.4*–3
Body fat (%)16–1.4±0.4–2.2 to −0.7*–5
VO2max(mL/kg–1/min–1)193.9±0.52.8 to 4.9*16
N=number or groups reporting data; BMI=body mass index; VO2max=maximum oxygen consumption; TC=total cholesterol; HDL-C=high-density lipoprotein cholesterol; non-HDL-C=non-high-density lipoprotein cholesterol, calculated as total cholesterol-high-density lipoprotein cholesterol; CI=confidence interval; *significantly different from zero (0)

Secondary outcomes are shown in Table II. As can be seen, there was a statistically significant exercise- control decrease of 2% in TC and a trend for an approximately 3% increase in HDL-C. Statistically significant exercise-control increases of approximately 16% were observed for Vo2max mL/kg-1/min-1. In addition, small but statistically significant exercise- control decreases of approximately 2%, 3%, and 5%, were found, respectively, for body weight, BMI, and percentage of fat.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References

The results of this study suggest that walking decreases non-HDL-C levels in adult humans and that the changes observed appear to be independent of changes in body composition. Given that walking is currently the most preferred form of exercise in the United States and has a low rate of injury when compared with other forms of physical activity,45 the use of walking for reducing non-HDL-C levels appears to be efficacious, especially since numerous other health-related benefits can be derived from such participation.46,47 Since the majority of studies included in our meta-analysis adhered to the general guidelines for participating in physical activity (moderate intensity activity for =30 min/session at least 5 d/wk),48 adherence to these guidelines should generally bring about the decreases in non-HDLC observed in our meta-analysis. In addition, our results also indicated that adherence to such a program should also increase VO2max while reducing body weight, BMI, and percentage of body fat. Counseling by health care professionals may be especially appropriate since it has been suggested that such counseling can cause patients to become more physically active.49 The former notwithstanding, the clinical importance of our approximately 6 mg/dL (4%) reduction in non-HDL-C and its effect on reducing CVD risk could be questioned, however, since reductions in low-density lipoprotein-C of as low as approximately 8 mg/dL that are maintained for =6 years have been shown to reduce the risk of ischemic heart disease events by approximately 21%,50 it would appear plausible to suggest that the reductions in non-HDL-C observed in our study are important, especially if viewed from a population- wide perspective. While beyond the scope of this investigation, it may also be that walking results in the development of larger cholesterol-carrying liproproteins,51 thereby reducing the risk of CVD mortality.52 It would seem appropriate to suggest that future research examine the clinical importance of reductions similar to those observed in our study on non-HDL-C and subsequent CVD risk. Indeed, it may be that the addition of diet modification in the form of reduced saturated fat and cholesterol combined with physical activity may have a greater impact on reducing non-HDL-C levels in adults than either intervention alone.3,4 For some people, the use of pharmacologic intervention, such as statins, along with increased physical activity and diet modification may be necessary.

All of the studies included in our meta-analysis appeared to use the analysis-by-protocol vs. intention-to-treat approach in the analysis of their data. Thus, while walking appears to be efficacious for reducing non-HDL-C levels in adults, the effectiveness of walking on non-HDL-C has not been determined. Therefore, it would appear plausible to suggest that future research dealing with the effects of walking on non-HDL-C levels include the intention- to-treat approach in the analysis of their data. In addition, because of a lack of data provided by the original studies, we were unable to conduct a thorough analysis in relation to diet, drug use (for example, statins), cigarette smoking, and alcohol consumption. This information is important to know because any one of these elements could have had an affect on non-HDL-C. In addition, since the subjects who began an exercise program may have taken on additional positive health habits during the study (for example, a less atherogenic diet), such changes could have confounded our results. Consequently, it is suggested that future studies assess and report complete information for these variables. Finally, it would be interesting for future meta-analytic research to compare walking with other training modalities (for example, swimming, bicycling, running) on changes in non-HDL-C in adults.

The statistically significant association we found between the year of publication and decreases in non-HDL-C suggests that the more recent the year of publication, the greater the decreases in non- HDL-C. Our initial thought was that it may be the result of higher initial non-HDL-C levels with more recent publications, however, since no statistically significant association existed between the year of publication and initial levels of non-HDL-C, other factors need to be considered. One possible factor may have to do with study design differences, including the possibility of increased accuracy in determining lipid and lipoprotein levels. Other possible factors may have to do with the increased use of modern day statins and/or decreased cigarette smoking among subjects included in more recent studies. Alternatively, our results could have been nothing more than a spurious finding. In conclusion, the results of this study suggest that walking reduces non-HDL-C in adult humans.

Acknowledgments:

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References

This study was supported by a grant from the National Institutes of Health-National Heart, Lung and Blood Institute, Award #R01-HL069802 (G.A. Kelley, Principal Investigator).

The authors would like to thank William Haskell, PhD, Stanford University, for reviewing our reference list and providing suggestions for the coding of studies.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments:
  7. References
  • 1
    Cui Y, Blumenthal RS, Flaws JA, et al. Non-high-density lipoprotein cholesterol level as a predictor of cardiovascular disease mortalit. Arch Intern Me. 2001;161:14131419.
  • 2
    Frost PH, Havel RJ. Rationale for use of non-high-density lipoprotein cholesterol rather than low-density lipoprotein cholesterol as a tool for lipoprotein cholesterol screening and assessment of risk and therap. Am J Cardio. 1998;81:26B31B.
  • 3
    Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults: Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III. JAMA. 2001;285:24862497.
  • 4
    National Cholesterol Education Program. National Heart Lung and Blood Institute, National Institutes of Health: Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Repor. Circulatio. 2002;106:31433421.
  • 5
    Eyler AA, Brownson RC, Bacak SJ, et al. The epidemiology of walking for physical activity in the United State. Med Sci Sports Exer. 2003;35:15291536.
  • 6
    Aldred HE, Hardman AE, Taylor S. Influence of 12 weeks of training by brisk walking on postprandial lipemia and insulinemia in sedentary middle-aged wome. Metabolis. 1995;44:390397.
  • 7
    Cauley JA, Kriska AM, LaPorte RE, et al. A two year randomized exercise trial in older women: effects on HDL-cholestero. Atherosclerosi. 1987;66:247258.
  • 8
    Davison RCR, Grant S. The physiological effects of a 14 week walking programme on sedentary middle-aged wome. J Sports Sc. 1995;13:2425.
  • 9
    Duncan JJ, Gordon NF, Scott CB. Women walking for health and fitnes. JAMA. 1991;266:32953299.
  • 10
    Fahlman MM, Boardley D, Lambert CP, et al. Effects of endurance training and resistance training on plasma lipoprotein profiles in elderly wome. J Gerontol A Biol Sci Med Sc. 2002;57:B54B60.
  • 11
    Fukahori M, Aono H, Saito I, et al. Program of exercise training as Total Health Promotion Plan and its evaluatio. J Occup Healt. 1999;41:7682.
  • 12
    Gelin J, Jivegard L, Karlsson J, et al. Treatment efficacy of intermittent claudication by surgical intervention, supervised physical exercise training compared with no treatment in unselected randomised patients I: one year results of functional and physiological improvement. Eur J Vasc Endovasc Sur. 2001;22:107113.
  • 13
    Higashi Y, Sasaki S, Sasaki N, et al. Daily aerobic exercise improves reactive hyperemia in patients with essential hypertensio. Hypertensio. 1999;33:591597.
  • 14
    Hinkleman LL, Nieman DC. The effects of a walking program on body composition and serum lipids and lipoproteins in overweight wome. J Sports Med Phys Fitnes. 1993;33:4958.
  • 15
    Hopewell R. The Effect of Fiber and Exercise on Weight Loss and Blood Lipids in Moderately Overweight Wome. [dissertation. Morgantown, WV: West Virginia University, 1989.
  • 16
    Kaplan RM, Wilson DK, Hartwell SL, et al. Prospective evaluation of HDL cholesterol changes after diet and physical conditioning programs for patients with type II diabetes mellitu. Diabetes Car. 1985;8:343348.
  • 17
    Keller C, Trevino RP. Effects of two frequencies of walking on cardiovascular risk factor reduction in Mexican American wome. Res Nurs Healt. 2001;24:390401.
  • 18
    Miller CW. The Effects of Various Caloric Expenditures Per Week at Moderate Levels of Exercise Training Intensity on Coronary Risk Factors in Wome. [dissertatio.]. Pittsburgh, PA: University of Pittsburgh, 1995.
  • 19
    Nieman DC, Warren BJ, O'Donnell KA, et al. Physical activity and serum lipids and lipoproteins in elderly wome. J Am Geriatr So. 1993;41:13391344.
  • 20
    Nieman DC, Brock DW, Butterworth D, et al. Reducing diet and/or exercise training decreases the lipid and lipoprotein risk factors of moderately obese wome. J Am Coll Nut. 2002;21:344350.
  • 21
    Palank EA, Hargreaves EH. The benefits of walking the golf course: effects of lipoprotein levels and risk ratio. Physician Sportsme. 1990;18:7780.
  • 22
    Ready EA, Drinkwater DT, Ducas J, et al. Walking program reduces elevated cholesterol in women postmenopaus. Can J Cardio. 1995;11:905912.
  • 23
    Ready EA, Naimark B, Ducas J, et al. Influence of walking volume on health benefits in women post-menopaus. Med Sci Sports Exer. 1996;28:10971105.
  • 24
    Santiago MC, Alexander JF, Stull GA, et al. Physiological responses of sedentary women to a 20 week conditioning program of walking or joggin. Scand J Sports Sc. 1987;2:3339.
  • 25
    Santiago MC, Leon AS, Serfass RC. Failure of 40 weeks of brisk walking to alter blood lipids in normolipemic wome. Can J Appl Physio. 1995;20:417428.
  • 26
    Seals DR, Tanaka H, Clevenger CM, et al. Blood pressure reductions with exercise and sodium restriction in postmenopausal women with elevated systolic pressure: role of arterial stiffnes. J Am Coll Cardio. 2001;38:506513.
  • 27
    Sinatra ST, Allen GJ, Camaione DN, et al. Effects of continuous passive motion, walking, and a placebo intervention on physical and psychological well-bein. J Cardiopulmonary Rehabi. 1990;10:279286.
  • 28
    Stensel DJ, Hardman AE, Brooke-Wavell KSF, et al. Brisk walking and serum lipoprotein variables in formerly sedentary men aged 42–59 year. Clin Sc. 1993;85:701708.
  • 29
    Suter E, Marti B, Gutzwiller F. Jogging or walking-comparison of health effect. Ann Epidemio. 1994;4:375381.
  • 30
    Vasankari T, Fogelholm M, Kukkonen-Harjula K, et al. Reduced oxidized low-density lipoprotein after weight reduction in obese premenopausal wome. Int J Obe. 2001;25:205211.
  • 31
    Verity LS, Ismail AH. Effects of exercise on cardiovascular disease risk in women with NIDD. Diabetes Res Clin Prac. 1989;6:2735.
  • 32
    Woolf-May K, Kearney EM, Jones DW, et al. The effect of two different 18-week walking programmes on aerobic fitness, selected blood lipids and factor XII. J Sports Sc. 1998;16:701710.
  • 33
    Woolf-May K, Kearney EM, Owen A, et al. The efficacy of accumulated short bouts versus single bouts of brisk walking in improving aerobic fitness and blood lipid profile. Health Educ Re. 1999;14:803815.
  • 34
    Yeater RA, Ullrich IH, Maxwell LP, et al. Coronary risk factors in type II diabetes: response to low-intensity aerobic exercis. W V Med . 1990;86:287290.
  • 35
    Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credi. Psychol Bul. 1968;70:213220.
  • 36
    Follmann D, Elliot P, Suh I, et al. Variance imputation for overviews of clinical trials with continuous respons. J Clin Epidemio. 1992;45:769773.
  • 37
    Dersimonian R, Laird N. Meta-analysis in clinical trial. Control Clin Trial. 1986;7:177188.
  • 38
    Hunter JE, Schmidt FL. Fixed effects vs. random effects meta-analysis models: implications for cumulative research knowledg. Int J Sel Asses. 2000;8:275292.
  • 39
    NRC Committee on Applied and Theoretical Statistic. Combining Information: Statistical Issues and Opportunities for Researc. Washington, DC: Academy Press; 1992.
  • 40
    Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessar. Control Clin Trial. 1996;17:112.
  • 41
    Moher D, Jadad AR, Nichol G, et al. Assessing the quality of randomized controlled trials: an annotated bibliography of scales and checklist. Control Clin Trial. 1995;16:6273.
  • 42
    Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysi. Biometric. 2000;56:455463.
  • 43
    Duval S, Tweedie R. A nonparametric “trim and fill” method of accounting for publication bias in meta-analysi. J Am Stat Asso. 2000;95:8998.
  • 44
    Lipsey MW, Wilson DB. Practical Meta-Analysi. Thousand Oaks, CA: Sage; 2001.
  • 45
    Hootman JM, Macera CA, Ainsworth BE, et al. Association among physical activity level, cardiorespiratory fitness, and risk of musculoskeletal injur. Am J Epidemio. 2001;154:251258.
  • 46
    Booth FW, Gordon SE, Carlson CJ, et al. Waging war on modern chronic diseases: primary prevention through exercise biolog. J Appl Physio. 2000;88:774787.
  • 47
    Booth FW, Chakravarthy MV, Gordon SE, et al. Waging war on physical inactivity: using modern molecular ammunition against an ancient enem. J Appl Physio. 2002;93:330.
  • 48
    US Department of Health and Human Service. Physical Activity and Health: A Report of the Surgeon Genera. Atlanta, Georgia: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996.
  • 49
    Chakravarthy MV, Joyner MJ, Booth FW. An obligation for primary care physicians to prescribe physical activity to sedentary patients to reduce the risk of chronic health condition. Mayo Clin Pro. 2002;77:165173.
  • 50
    Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and meta-analysi. BM. 2003;326:14231429.
  • 51
    Kraus WE, Houmard JA, Duscha BD, et al. Effects of the amount and intensity of exercise on plasma lipoprotein. N Engl J Me. 2002;347:14831492.
  • 52
    Gardner CD, Fortmann SP, Krauss RM. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and wome. JAMA. 1996;276:875881.