Discussion
The purpose of this study was to determine the combined effects of aerobic exercise and a prudent diet on lipid and lipoprotein concentrations in adults. The overall results suggest that the combined effects of both are efficacious for reducing concentrations of TC, the ratio of TC to HDL-C, LDL-C and TG, but not increasing HDL-C, in adults. These findings appear to be important from a practical perspective, especially in relation to the prevention of coronary heart disease (CHD), a disease with an annual estimated incidence rate in the United States of more than 1.4 million people [
1]. For example, it has been reported that every 1% reduction in population levels of TC results in an approximate 2% reduction in the rate of CHD [
46]. Based on the findings of the current study, this would result in a 15% reduction in the rate of CHD as a result of aerobic exercise and a prudent diet. The decreases observed in the ratio of TC to HDL-C also appear to be important with changes based on previous research equivalent to reductions of approximately 21% in CHD risk [
47]. In addition, Wilson et al [
48], estimated that every 1% population decrease in LDL-C would decrease the 12-year incidence of CHD by 1%. Congruent with the changes observed for LDL-C in the current meta-analysis, this would be equivalent to a reduction of approximately 7% in the 12-year incidence of CHD. Improvements in LDL-C may be particularly important given that LDL-C is currently the primary target of lipid-lowering therapy in adults [
2]. Furthermore, based on previous work, the decreases in serum TG observed in the current investigation would be equivalent to decreases in the relative risk of CHD of 2% in men and 4% in women [
49]. Finally, the observed results may be an underestimate of the true effects of aerobic exercise combined with a prudent diet on lipid and lipoprotein concentrations in adults given that participants tend to overestimate their adherence to exercise and diet recommendations [
50,
51].
The changes observed in TC, LDL-C and TG in the current investigation appear to be greater than those achieved with aerobic exercise alone. For example, previous meta-analytic work addressing the effects of aerobic exercise on lipid and lipoprotein concentrations in women [
52] and men [
53] yielded significant reductions in TC, LDL-C and TG that were less than half those observed in the current investigation. The ratio of TC to HDL-C was not assessed in either study [
52,
53]. In contrast, prior meta-analytic work that examined the effects of Step 1 and 2 diets from the American Heart Association reported improvements in lipid and lipoprotein concentrations that were similar to or larger than the current investigation with respect to TC, ratio of TC to HDL-C and LDL-C [
54]. However, changes in TG were less in this prior work [
54]. The results of the prior meta-analysis should be viewed with caution as it appears that traditional statistical approaches versus those specific to the conduct of meta-analysis were used [
54]. Consequently, the reported changes in lipid and lipoprotein concentrations may be exaggerated.
While statistically significant and practically important improvements were observed for TC, the ratio of TC to HDL-C, LDL-C and TG, no such differences were observed for HDL-C. The lack of effect on HDL-C may not be surprising given that aerobic exercise alone has been shown to increase concentrations of HDL-C in both women [
52] and men [
53] while low total and saturated fat diets have been shown to significantly decrease HDL-C [
54]. Thus, it appears that the positive effects of aerobic exercise on HDL-C may not override the lowering effect of diets generally low in total and saturated fat but may help to mitigate these changes given that the current meta-analysis found a non-significant decrease in HDL-C of 1.3% while low-fat only diets resulted in a statistically significant decrease of 7%. Given that participants with low concentrations of HDL-C may respond less to exercise than those with high concentrations, it is important to note that results remained consistent when each study was excluded from the model once, including the one study that enrolled participants with initially low concentrations of HDL-C [
15]. With the former in mind, other forms of therapy, for example, fibrates or niacin [
55], may be necessary for raising HDL-C in adults who exercise aerobically and consume a low-fat diet. However, as previously noted, the results of the low-fat diet meta-analysis may have been exaggerated [
54]. Based on this observation, it appears that a need exists for an updated meta-analysis of randomized controlled trials to determine the effects of low-fat only diets on lipid and lipoprotein concentrations, especially HDL-C, in adults.
The changes observed in the current study for TC, LDL-C and HDL-C are generally less than those reported for statin therapy and equal to or greater than changes reported for TG [
56]. For example, previous meta-analytic research on the effects of statins on lipid and lipoprotein concentrations reported improvements ranging from 17% to 31%, 7% to 12%, 26% to 46% and 10% to 18% respectively, for TC, HDL-C, LDL-C and TG [
56]. Generally speaking, our findings support current recommendations regarding the use of aerobic exercise and a prudent diet as a first line strategy for maintaining optimal concentrations of lipids and lipoproteins in adults [
2]. If optimal levels cannot be attained, aerobic exercise and diet should still be recommended with the possible addition of fibrates or niacin [
55] for increasing HDL-C and/or a statin for improving all other lipid and lipoprotein concentrations, especially LDL-C [
56].
Statistically significant reductions were found for body weight and BMI while changes in the direction of benefit were observed for WHR as a result of the aerobic exercise and diet intervention. In addition, a statistically significant increase in VO2max in ml.kg-1.min-1 was observed. These findings are not surprising given that weight loss and increases in cardiorespiratory fitness are common changes that occur as a result of aerobic exercise and a prudent diet. Given that significant reductions were found for total kilocalories, total fat, saturated fat and cholesterol, it appears, overall, that participants were successful in adhering to the assigned diet.
While the results of this study are encouraging, they must be viewed with respect to several issues. First, while simple meta-regression analyses yielded several significant associations with selected concentrations of lipids and lipoproteins, these should be viewed with caution given the small sample size. In addition, since studies are not randomly assigned to predictors, such analyses are considered to be observational in nature [
57]. Consequently, such analyses do not support causal inferences [
57]. Rather, the validity of these findings would need to be tested in large, well-designed randomized controlled trials. This may be especially relevant for determining whether changes in selected lipid and lipoprotein concentrations are the result of the combined effects of aerobic exercise and a prudent diet or the weight loss associated with a combination of aerobic exercise and a prudent diet.
Second, with the exception of the WHR, statistically significant heterogeneity and a large amount of inconsistency was observed for all of our primary and secondary outcomes. While random effects models account for heterogeneity, resulting in narrower confidence intervals when heterogeneity is present, others have suggested that it is not appropriate to reach conclusions based on aggregate findings when there is significant heterogeneity and/or inconsistency [
28]. However, heterogeneity (
Q) and inconsistency (
I
2
) statistics do not guarantee that the dispersion in results are large enough to be of practical or theoretical importance [
58]. In addition, no significant differences were found for any of our outcomes when each study was deleted from the model once as well as when selected results were adjusted for publication bias. Furthermore, based on our cumulative meta-analyses, the direction of results for all of our outcomes has remained stable for more than a decade.
A third issue has to do with the fact that overlapping prediction intervals were observed for all outcomes except changes in body weight and cholesterol intake. From a practical perspective, prediction intervals may be more relevant since they provide an approximation of the expected treatment effect in a new trial [
37,
38]. However, they should not be used to determine whether confidence intervals from a random effects model are correct or incorrect since prediction intervals are based on a random mean effect while confidence intervals are not [
37,
38].
A fourth issue has to do with the lack of studies available after the investigative team's strict inclusion criteria were applied. This was especially surprising given the fact that aerobic exercise and a diet low in total and saturated fat are common first line strategies for treating dyslipidemia [
2,
3]. In addition, it was surprising to see that the most recent study that met our strict inclusion criteria had been published in 2002 [
8,
10,
44]. While this may have been the result of possible search error, there may also be a general belief that the beneficial effects of this intervention are well established and that no further research in this area is necessary. However, this may be shortsighted. For example, all of the included studies appeared to conduct 'as treated' analyses in determining the combined effects of aerobic exercise and a prudent diet on lipid and lipoprotein concentrations in adults. While such an approach can determine the efficacy of findings, it cannot determine the effectiveness of findings [
59]. Therefore, it would appear plausible to suggest that future studies are needed that include both efficacy (as-treated) and effectiveness (intention-to-treat) analyses and that such studies focus on the Therapeutic Lifestyle Changes currently recommended by the National Cholesterol Education Program and American Heart Association [
2,
3]. Such knowledge would address the issue of whether the currently recommended treatment works (efficacy) as well as whether it works in the real world (effectiveness
) [
59]. Knowledge of both is especially important with respect to the allocation of resources for treating dyslipidemia.
A fifth issue has to do with the general lack of reporting for certain information in the included studies. For example, it was difficult to assess the risk of bias with respect to allocation concealment, incomplete data and outcome reporting. It is suggested that future studies report this information, including a study protocol identification number. Future studies should also ensure that all participants were not exercising on a regular basis prior to enrollment and include a definition of such. In addition, future research on this topic should include data on the race/ethnicity of subjects, number of men and women who started and completed the study, medication status of participants, including hormone replacement therapy, cigarette smoking, alcohol intake, percentage of body fat as an outcome variable, average intensity of training for aerobic exercise, compliance to the aerobic exercise protocol, whether the exercise habits of participants changed outside of the intervention, protein, fiber, and trans-fat intake as well as the season(s) in which lipids and lipoproteins were assessed and the number of hours that exercise was avoided prior to assessment. Finally, since non-HDL-C has been shown to be a better predictor of cardiovascular morbidity and mortality than LDL-C [
60,
61], currently, the primary target of lipid lowering therapy in adults [
2], the inclusion of data for non-HDL-C, including dispersion statistics, is also recommended in future studies.
A sixth issue has to do with the use of traditional meta-analytic models that were employed in the current meta-analysis. While more recent meta-analytic methods have been proposed [
62‐
68], it was the investigative team's position that (1) it would be imprudent to overreact to newly proposed models and (2) that many of the newer alternatives proposed as a means of dealing with specific problems in meta-analysis may have problems of their own, possibly with substantially worse impact than the problems they were intended to solve (Dr. Michael Borenstein and Dr. Larry Hedges, personal communication, 22 February and 25 February 2011).
Finally, the focus of the current meta-analysis was on the combined effects of aerobic exercise and a prudent diet on lipid and lipoprotein concentrations in adults. However, there may also be an interest in future meta-analytic work aimed at determining the independent effects of aerobic exercise, dietary fat reduction and weight loss on lipid and lipoprotein concentrations in adults. Ideally, this would best be accomplished by limiting the inclusion of studies to randomized controlled trials in which all of these intervention arms are included within the same study.
In conclusion, aerobic exercise and a prudent diet are efficacious for improving TC, TC:HDL-C, LDL-C and TG, but not HDL-C concentrations, in adults. However, a need exists for additional studies on this topic, including effectiveness studies using intention-to-treat analysis.
Authors' contributions
GAK was responsible for the conception and design, acquisition of data, analysis and interpretation of data, drafting the initial manuscript and revising it critically for important intellectual content. KSK was responsible for the conception and design, acquisition of data, and reviewing all drafts of the manuscript. SR was responsible for the conception and design, interpretation of data and reviewing all drafts of the manuscript. WH was responsible for the conception and design, interpretation of data and reviewing all drafts of the manuscript. All authors read and approved the final manuscript.