The current interest for the use of n-3 (polyunsaturated) fatty acids in vascular disease can be originally tracked to observations in Greenland Inuits (Eskimos), revealing a lower prevalence of coronary heart disease (CHD) in these populations compared with Scandinavian control subjects (1–4). In a series of pioneering studies, Dyerberg and Bang (5,6) originally showed that Inuits had an attenuated platelet reactivity and a prolonged bleeding time compared with Scandinavian control subjects. This was attributed to the Eskimo diet, with an extremely high content of fish or of fish-derived products (such as seal), abundant in n-3 fatty acids, mainly eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) (7). In other populations with a high consumption of fish, such as the Japanese (8,9) and the Alaskans (10), a similar inverse correlation between fish consumption and mortality from CHD has been subsequently found.

However, in Western populations with a generally low intake of n-3 fatty acids, both protective effects (11–16) and no effects (17–20) of n-3 fatty acids on CHD have been reported. There are good explanations for the lack of uniformity in the epidemiological data, including the difficulty in maintaining constant feeding habits in a population during long observational studies and the influence of other dietary principles, including the simultaneous ingestion of saturated or other unsaturated fatty acids.

Overall, the bulk of epidemiological data suggests the existence of favorable associations between fish consumption and mortality from CHD (21–24), as reflected in a recent health statement from the American Heart Association (25). In addition, a host of in vitro and in vivo studies have provided possible biological explanations for the epidemiological observations. Such studies have demonstrated that 1) diets rich in n-3 fatty acids partially replace n-6 with n-3 fatty acids in membrane phospholipids in all cells at a rate consistent with the turnover of the tissue under study; 2) this produces a modulation in the metabolism of bioactive eicosanoids prostaglandins, thromboxanes, and leukotrienes, arising from long-chain fatty acids; 3) n-3 fatty acids reduce platelet and leukocyte reactivity and blood pressure; 4) n-3 fatty acids reduce atherogenesis and thrombosis in most (but not all) animal studies; 5) n-3 fatty acids reduce plasma triglycerides; and 6) n-3 fatty acids likely possess antiarrhythmic properties.

Patients with diabetes have a three- to fourfold increase in the risk of CHD compared with the general population (21,26–29). Although partially attributable to an increased association with other traditional risk factors (mostly obesity, dyslipidemia, and hypertension), epidemiological studies have demonstrated that diabetes also involves an additional excess of risk (21,27,30). Many properties ascribed to n-3 fatty acids appear to be particularly applicable to diabetic patients, where the need for “nontraditional” preventive or therapeutic measures is even higher than for cardiovascular disease (CVD) in general.

Against this background, we have here reviewed the medical literature on the effects of n-3 fatty acids in relation to diabetes, with particular emphasis on clinical studies. The database for this systematic review includes all published articles on n-3 fatty acids and diabetes retrieved in a PubMed, Embase, and Excerpta Medica current contents search up to November 2006. A few review articles have been selected to summarize less recent studies and the bulk of available literature on some of the covered topics, based on the personal experience of the authors. The search was not restricted by the language of the publication or the publication type. Search terms included n-3 fatty acids, ω-3 fatty acids, or fish oil, and diabetes mellitus, insulin resistance, glycemic control, or hypertriglyceridemia. Additional published or unpublished literature was sought through manual searches of reference lists of included studies, key review articles, and the personal experience of the authors. Only randomized clinical studies were included. Criteria for the assessment of trial quality included the quality of control, if any, the method of randomization, the blinding of investigators, the blinding of enrolled subjects, and any systematic difference in care between the intervention groups. Intervention groups excluded children, acutely ill, or pregnant subjects. For inclusion, the intervention (through dietary changes or n-3 fatty acid supplementation) had to continue for at least 1