To determine the effectiveness of exercise training (aerobic and resistance) in modifying blood lipids and to determine the most effective training programme with regard to duration, intensity and frequency for optimising the blood lipid profile.

Computerised searches of the following databases were conducted from January 1975 to December 1997: MEDLINE, EMBASE and the Science Citation Index. The search included the following expanded terms: 'exercise', 'blood-lipid', 'cholesterol' and 'randomized'. Additional articles were identified through searching the bibliography of relevant trials and previous review articles. Only English-language articles were included.

Study designs of evaluations included in the review

Only randomised and controlled trials (RCTs) were included. Crossover studies were included if the order of treatments was randomised and there were no significant order effects. The mean number of participants per study was 59 +/- 52.4 (range 16-300).

Specific interventions included in the review

Aerobic or resistance training programmes of at least four weeks duration. Programmes comprised various forms of exercise with varying intensities, duration and number of sessions per week. Most studies (27/31) looked exclusively at aerobic exercise programmes, the majority of which included walking, running and /or jogging. The mean length of the programmes was 25.7 weeks (range 9-52wks) and the frequency of training sessions was on average 3.9 +/- 1.1 sessions per week (range 2.5-7.0). Studies that did not report the intensity of programmes were excluded.

Participants included in the review

Hyperlipidemic and normolipidemic adults who were both sedentary and healthy. Individuals rehabilitating from a cardiac event were excluded. The majority of the studies (16/31) reported in the review looked only at men, and only included normolipidemic individuals (17/31 studies). The age of participants ranged from 19 to 83yrs.

Outcomes assessed in the review

Blood lipid profiles reported as either a primary or secondary outcome. Profiles included one or more of the following measurements: total cholesterol (TC), high-density lipoprotein (HDL-C), low-density lipoprotein (LSL-C) and triglycerides (TG). Changes in body mass and body mass index were also reported where available.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The scheme described in the Cochrane Collaboration Handbook (1995) was used to assess study validity. This scheme involves assessing the quality of allocation concealment and rates studies on a three-point scale: 'A' - maximal efforts were made to control selection bias (e.g. central randomisation by an independent third party), 'B' - some effort to control selection bias (e.g. sealed envelopes), and 'C' - little or no effort to control selection bias at entry. The authors' noted however, that in many cases this scheme may reflect the quality of reporting rather than the study itself as many of the trials lacked methodological detail and did not state the method of randomisation. Other aspects of study validity were also examined including sample size, percentage follow-up of randomised participants, adherence rate for session attendance, reporting of information relating to dietary advice and intake. The authors do not state how the papers were assessed for validity, or how many of the authors performed the validity assessment.

The information extracted included: study design, numbers of participants, age and gender of participants, details about exercise programme, control group, outcome measures, body composition measures, dietary instructions to participants and adherence rates. The identity and number of individuals extracting the data were The authors do not state how the data were extracted for the review, or how many of the authors performed the data extraction. The effects of exercise on the individual blood profile components was calculated as the difference (mmol/L) between the mean change in blood lipid (baseline-final value) in the training versus control groups. The variances of the difference between means (assuming that the baseline and final blood lipid measurements were unpaired) were calculated.

How were the studies combined?

Pooled effect sizes were calculated as weighted mean differences (WMD) with 95% confidence intervals (95% CIs) using a random-effects model.

How were differences between studies investigated?

Heterogeneity tests were performed using the Mantel-Haenszel method reported in Cochran 1954 (see Other Publications of Related Interest).

There were 31 studies (30 randomised and/or controlled trials, and 1 crossover study), including a total of 1833 participants.

Quality of studies:

In terms of controlling for selection bias after entry 26 of the 31 studies were graded as level 'C' and 5 as 'B'. No studies were graded as level 'A'. The method of randomisation was only stated in four of the 31 studies. Eleven studies included all of the randomised participants in the final analysis and 12 stated an adherence rate for session attendance. Four studies provided information on the dietary instructions or measurement of dietary intake. Overall the quality of the studies was poor, the major defects were small sample sizes, failure to state method of randomisation and lack of information on how selection bias was controlled after entry.

Effect of aerobic training programmes (28 studies; 1693 participants, 863/1693 hyperlipidemic).

Statistically significant (p < 0.05) changes were identified in TC (WMD = 0.10, 95% CI: 0.02, 0.18), HDL-C (WMD = -0.05, 95% CI: -0.08, -0.02), LDL-C (WMD = 0.10, 95% CI: 0.02, 0.19) and TG (WMD = 0.08, 95% CI: 0.02, 0.14). Overall, the comparison of exercise intensities showed that programmes at intensities greater than 70% VO2 max produced larger changes in TC (WMD = 0.14, 95% CI: 0.07, 0.21) and LDL-C (WMD = 0.16, 95% CI: 0.06, 0.25) while programmes at lower intensities modified TG (WMD = 0.10, 95% CI: 0.01, 0.19) and HDL-C (WMD = -0.06, 95% CI: -0.10, -0.02). Correlations between baseline lipid concentrations and the post-training changes were low; TC 9 = 0.10), TG 9 = 0.23), HDL-C 9 = 0.19) and LDL-C 9 = 0.17). Three exercise sessions per week resulted in greater modifications to all blood lipid levels than more frequent exercise (TC - WMD = 0.18, 95% CI: 0.08, 0.28; HDL-C - WMD = -0.06, 95% CI: -0.09, -0.02; LDL-C - WMD = 0.16, 95% CI: 0.03, 0.29; TG - WMD = 0.13, 95% CI: 0.06, 0.20). The effect of exercise on blood lipids differed, with similar patterns of change occurring between HDL-C and TG, and between TC and LDL-C. The blood lipid status of the individuals (i.e. hyper- vs. normo-lipidemic) was only significant for effects of aerobic training on HDL-C (WMD = -0.07, 95% CI: -0.10, -0.03) and TG (WMD = 0.15, 95% CI: 0.07, 0.23). However, significant heterogeneity was found with all comparisons.

Effect of resistance training programmes (4 studies; 174 participants, 42/174 hyperlipidemic):

One study looked at a combined aerobic and resistance training programme. Overall, a statistically significant (p < 0.05) decrease was identified in LDL-C levels, (WMD = 0.40, 95% CI: 0.07, 0.73) with no differences for levels of TC, TG or HDL-C.

Caution is required when drawing firm conclusions from this study given the significant heterogeneity with comparisons. However, the results appear to indicate that aerobic exercise training produced small but favourable modifications to blood lipids in previously sedentary adults.

This is a clearly presented review based on well-defined inclusion criteria. The literature search used a number of databases; although excluding non-English language publications may have missed relevant data and publication bias may be an issue as no specific attempts were made to locate unpublished material. Details about the identity and number of individuals involved in the selection of studies, extraction of data and assessment of quality were not reported.

Before combining data the validity of individual studies and heterogeneity between studies was assessed. However, despite significant heterogeneity as acknowledged by the authors in the text, pooled effects were presented in the results. In addition the authors often did not distinguish or comment on differences between the two different populations of patients (normo- and hyper-lipidemic). In view of these factors and as the authors' acknowledge in their conclusions, the findings of the review should be interpreted with caution. The recommendations of the review however would appear to be valid.

Practice: The authors state that 'there is insufficient evidence to recommend exercising more than three times per week if the specific intention is to improve the blood lipid profile'.

Research: The authors state that 'the benefits of exercise while small at an individual level, maybe of significant public health benefit if they can be applied broadly throughout the community', however, 'such benefits need confirmation in large prospective controlled trials'.

1. Cochran WG. The combination of estimates from different experiments. Biometrics 1954;10:101-29.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.