Abbreviations: DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; HMG-CoA, 3-hydroxy 3-methylglutaryl CoA.A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many other substances.
Fish Oil Supplementation in Type 2 DiabetesA quantitative systematic review
R E V I E W A R T I C L EOBJECTIVE -To determine the effects of fish oil supplementation on lipid levels and glycemic control in patients with type 2 diabetes.
RESEARCH DESIGN AND METHODS-A comprehensive search of Medline, Embase, Lilacs, the Cochrane Clinical Trials Registry, bibliographies of relevant papers, and expert input updated through September 1998 was undertaken. All randomized placebo-controlled trials were included in which fish oil supplementation was the only intervention in subjects with type 2 diabetes. Three investigators performed data extraction and quality scoring independently with discrepancies resolved by consensus. Eighteen trials including 823 subjects followed for a mean of 12 weeks were included. Doses of fish oil used ranged from 3 to 18 g/day.The outcomes studied were glycemic control and lipid levels.RESULTS -Meta-analysis of pooled data demonstrated a statistically significant effect of fish oil on lowering triglycerides (Ϫ0.56 mmol/l [95% CI Ϫ0.71 to Ϫ0.41]) and raising LDL cholesterol (0.21 mmol/l [0.02 to 0.41]). No statistically significant effect was observed for fasting glucose, HbA 1c , total cholesterol, or HDL cholesterol. The triglyceride-lowering effect and the elevation in LDL cholesterol were most marked in those trials that recruited hypertriglyceridemic subjects and used higher doses of fish oil. Heterogeneity was observed and explained by the recruitment of subjects with baseline hypertriglyceridemia in some studies.
CONCLUSIONS
R e v i e w s / C o m m e n t a r i e s / P o s i t i o n S t a t e m e n t s
1408DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000
Fish oil in type 2 diabetesOutcome measures and data extraction Study selection, data extraction, and assigning of a quality score were undertaken independently by 3 investigators with discrepancies resolved by consensus. When data were not available in a published report, efforts were made to contact the primary investigators. Data related to the effects of fish oil and placebo on glycemic and lipid outcomes were extracted from each trial. One study reported total glycosylated hemoglobin; a formula was used to convert data to HbA 1c (15). Study quality was assessed using the score developed by Jadad et al. (16), which has a possible range of 0-5, with a cutoff of 2 used to designate studies of high versus low quality.
Data analysis and statistical methodsExtracted data were analyzed using the Review Manager 3.1 statistical software developed by the Cochrane Collaboration.A random-effects model was used to pool data. Effect sizes are presented as weighted mean differences with 95% CIs. Heterogeneity was assessed using the 2 method. Publication bias was evaluated using a funnel plot method (1...