AIMS: To determine lipid profiles and associations with other metabolic disease in a representative British elderly population. METHODS: Part of a prevalence survey of dementia in all 75+ year olds conducted from the large general practice serving the town and surrounding area of Melton Mowbray, Leicestershire (the M-old study). Patients (n = 224) aged from 75 to 98 years, and representative of the overall population, also provided pre-prandial blood samples on which various age and nutrition related analytes were determined. These included documented medical history, thyroid stimulating hormone (TSH), glucose, immunoglobulins, and lipid profile in plasma. RESULTS: Cholesterol and lipid variables showed wide scatter, with some negative trends but no significant associations with age for total cholesterol, high density lipoprotein (HDL) cholesterol, the ratio of total to HDL cholesterol or triglycerides. Women had significantly higher concentrations of total and HDL cholesterol at all ages. Serum TSH was above 6.0 mU/1 in 10/205 patients, random glucose was above 11.2 mmol/l in nine of 207 patients, borderline dysglobulinaemia was present in four of 210 patients, all without correlation with cholesterol concentrations. CONCLUSION: This British data is consistent with an inverse correlation between survival and cholesterol, but wide scatter restricts reliance on single result lipid data in individual patient management. Random lipid screening is also unhelpful, inefficient and without added value in revealing other age related and unrecognised occult metabolic disease.