Concentrations of pyridoxal phosphate and pyridoxic acid were measured in fasting plasma samples from British men and women aged 65 years and over, participating in a National Diet and Nutrition Survey during 1994-5, selected to be representative of the population of mainland Britain. In this population, the concentration of pyridoxal phosphate declined, whereas pyridoxic acid rose, with increasing age and frailty; however, both status indicators were strongly and directly (with a positive coefficient) correlated with estimates of vitamin B6 intake. This was little affected by the inclusion of food energy and protein intakes in the model. Forty-eight percent of the participants living in the community and 75% of those living in institutions had plasma pyridoxal phosphate concentrations below a range considered normal from other studies. In a univariate regression model, plasma pyridoxal phosphate concentrations were inversely correlated with plasma homocysteine concentrations, consistent with the hypothesis that vitamin B6 status may influence plasma homocysteine levels, and hence vascular disease risk. However, this relationship was partly attenuated in a multiple regression model including age, sex, domicile and biochemical status indices, including those of folate and vitamin B12. There was evidence that plasma pyridoxal phosphate was sensitive to metabolic conditions associated with inflammation and the acute-phase reaction, and that plasma pyridoxic acid was sensitive to renal function. Thus, neither index is an ideal predictor of vitamin B6 status in older people, unless these confounding factors are allowed for. Since poor vitamin B6 status may have health implications, e.g. for immune function, cognition, and for essential intermediary metabolic pathways in older people, it needs to be investigated as a possible public health problem.