Oxygen-15 positron emission tomography (15O PET) can provide important data regarding patients with head injury. We provide reference data on intersubject variability and reproducibility of cerebral blood flow (CBF), cerebral blood volume (CBV), cerebral metabolism (CMRO2) and oxygen extraction fraction (OEF) in patients and healthy controls, and explored alternative ways of assessing reproducibility within the context of a single PET study. In addition, we used independent measurements of CBF and CMRO2 to investigate the effect of mathematical correlation on the relationship between flow and metabolism. In patients, intersubject coefficients of variation (CoV) for CBF, CMRO2 and OEF were larger than in controls (32.9%+/-2.2%, 23.2%+/-2.0% and 22.5%+/-3.4% versus 13.5%+/-1.4%, 12.8%+/-1.1% and 7.3%+/-1.2%), while CoV for CBV were lower (15.2%+/-2.1% versus 22.5%+/-2.8%) (P<0.001). The CoV for the test-retest reproducibility of CBF, CBV, CMRO2 and OEF in patients were 2.1%+/-1.5%, 3.8%+/-3.0%, 3.7%+/-3.0% and 4.6%+/-3.5%, respectively. These were much lower than the intersubject CoV figures, and were similar to alternative measures of reproducibility obtained by fractionating data from a single study. The physiological relationship between flow and metabolism was preserved even when mathematically independent measures were used for analysis. These data provide a context for the design and interpretation of interventional PET studies. While ideally each centre should develop its own bank of such data, the figures provided will allow initial generic approximations of sample size for such studies.