Jia L, Liu X, Bai YY, et al. Short-term effect of cocoa product consumption on lipid profile: a meta-analysis of randomized controlled trials. Am J Clin Nutr. July 2010;92(1) 218-225.

Elevated blood cholesterol levels are a risk factor for coronary artery disease (CAD), one of the leading causes of morbidity, mortality, and disability. Diet is a major factor affecting blood cholesterol levels. Studies have shown that cocoa (Theobroma cacao) and its products, rich in polyphenols, can affect lipid profiles in persons with cardiovascular-related diseases. These authors conducted a review of the scientific literature and a meta-analysis of all published randomized controlled trials that investigated the effects of cocoa on the lipid profile: total blood cholesterol (TC), low-density lipoprotein (LDL) cholesterol, and high-density lipoprotein (HDL) cholesterol.

The authors searched PubMed (from 1950 to May 2009), Embase (from 1966 to May 2009), the Cochrane Library database, and reviews of relevant articles for clinical trials that met these criteria: the use of cocoa products matched by a suitable control arm that allowed any observed effects to be reasonably ascribed to cocoa; the exclusion of children or critically ill subjects with any degree of cardiovascular disease; the use of assignable designs (crossover, parallel, etc.) specifying the treatment type, dose, and duration; the evaluation of elevated blood lipids by estimating the concentrations TC, LDL cholesterol, and HDL cholesterol; and the use of food intake control methods.

The studies were scored on the Jadad scale between 0 and a high score of 5, according to quality of randomization, blinding, reporting of withdrawals, generation of random numbers, and concealment of allocation. Extracted data included study characteristics, population information, and baseline and final concentrations or net changes of TC, LDL cholesterol, and HDL cholesterol. The authors defined the estimated principal effects as the net changes in each study variable, which were calculated as the mean difference (active treatment minus control) in the changes (follow-up minus baseline).

The authors initially identified 1,211 potentially eligible studies. After excluding most of those because they were not clinical trials or because the interventions were not relevant to the purpose of this meta-analysis, they identified eight eligible studies with 215 subjects.

The eight trials varied in size from 15 to 44 subjects. The main sources of cocoa were dark chocolate and cocoa powder. Of the eight studies, four investigated the effect of cocoa on healthy subjects and the others investigated the effect in subjects with cardiovascular risks such as prehypertension and hypertension. The cocoa products varied in their polyphenol content: from 30 to 963 mg/d. Duration of treatment varied from two to 18 weeks. Three trials were double-blinded, and five were crossover trials. Only the three double-blind trials were classified as high quality studies (Jadad scores of 5); the others were of low quality (Jadad scores of 2).

Short-term data from the eight trials showed that LDL cholesterol was significantly lower in the subjects who consumed the cocoa products than in the placebo-treated subjects. LDL cholesterol decreased by 5.87 mg/dL (95% confidence interval [CI]: -11.13, -0.61; P=0.03) in the cocoa group, a 4.82% difference in reduction compared to placebo. Regarding TC levels, a marginal decrease was seen in those using cocoa products (-5.82 mg/dL; 95% CI: -12.39, 0.76; P=0.08), a 3.07% reduction difference.

Subgroup analyses showed that cocoa consumption could significantly reduce TC (P=0.007) and LDL (P=0.04) cholesterol in subjects with cardiovascular risk compared with controls (mean differences of 4.23% and 6.25%, respectively, from placebo); however, cocoa did not affect blood cholesterol levels in the healthy subjects. A subgroup analysis also revealed a marginally significant reduction of LDL cholesterol and a tendency to decrease TC in the subgroup with shorter-term treatment of 6 weeks or less (P=0.07) but not in those with longer-term treatment.

The authors report no evidence of a dose-effect relation or of any change in HDL cholesterol. However, lower daily doses of cocoa polyphenols (<260 mg) significantly decreased TC and LDL cholesterol (P=0.0008 and P=0.03, respectively) compared to controls, whereas middle (260-665 mg) and high levels (>665 mg) of consumption did not. They note that a sensitivity analysis that excluded the lower-quality studies indicated that cocoa consumption did not significantly affect TC or LDL cholesterol levels (P=0.07 and P=0.23, respectively).

The authors explain that polyphenols have been shown to inhibit cholesterol absorption and promote the expression of LDL cholesterol receptors, but that a high dose of polyphenols has been shown to exert cytotoxic effects on liver cells. Therefore, they say, higher polyphenol supplementation may counteract its beneficial biological effects on lipid metabolism.

The authors acknowledge several weaknesses of this meta-analysis and state that the long-term effectiveness and appropriate dose range of cocoa consumption are not clear; however, they suggest that moderate cocoa consumption might be a worthwhile dietary approach for preventing hypercholesterolemia, particularly in specific patient subgroups. "Future research effects should concentrate on higher-quality and more rigorous randomized trials with longer follow-ups to resolve the uncertainty regarding the clinical effectiveness."