The clustering of dyslipidaemia, hypertension and glucose intolerance, predominantly in overweight individuals, has been ascribed many names, including syndrome X and the metabolic syndrome. In Reaven's original description of syndrome X, a central aetiological role was attributed to insulin resistance, and this assumption has remained as the dominant paradigm for the metabolic syndrome. There are a number of conceptual problems in such a model, particularly those arising from observations that several novel markers, including measures of endothelial dysfunction and of low-grade inflammation, are as closely related to insulin resistance as are the classic components of the syndrome. Because it is difficult to envisage how these traits might develop as a consequence of insulin resistance, such observations indicate the need for a new paradigm to explain the mechanisms of association better. It has been proposed that a state of low-grade inflammation, consequent upon the production of adipocytokines, particularly from truncal fat, explains the observed relationships between insulin resistance and endothelial dysfunction better than does a model revolving around insulin resistance. Furthermore, the inflammatory cytokines generated from adipose tissue may influence vessel endothelial function without elevations in circulating concentrations. This review alludes to several problems inherent in the epidemiological method in understanding disease mechanisms. These include crude biological measures, the use of venous systemic fasting samples, imprecision of assays, naive physiological models, simplistic statistical approaches and, without clinical trials, an inability to test causation. Integrated systems biology needs more complex approaches to investigate disease mechanisms, involving cell, organ, whole organism and population studies.