Introduction Metabolic derangements are common after traumatic brain injury (TBI), but few studies have used multi-tracer positron emission tomography (PET) to interrogate underlying pathophysiology. We examined relationships between glucose metabolism (measured using 18F-fluorodeoxyglucose (FDG) PET) and cerebral blood flow (CBF; measured using H215O PET). Methods Twenty-six TBI patients with median (range) pre-intubation Glasgow Coma Score of 6 (3-12) underwent combined 15O and FDG PET on 34 occasions; 10 and 18 healthy volunteers (controls) underwent 15O and FDG PET respectively. FDG rate constants were determined with an irreversible two-compartment model: transport across BBB (K1,k2), hexokinase activity (k3), and influx rate (Ki). Regions of interest (ROIs) were defined for haemorrhagic lesion (core), hypodense tissue (penumbra), 1 cm border zone of normal appearing tissue (peri-penumbra), and remote normal appearing tissue (normal). Results K1 was significantly reduced within core, but was comparable to control values in peri-penumbra and normal ROIs. While k3 and Ki were lower than control values, regional increases were found in the vicinity of lesions and some structurally normal areas. Glucose delivery (measured by K1) was unrelated to CBF when values were above 30ml/100ml/min, and was significantly below 95% confidence intervals for control values when CBF was below 25ml/100ml/min. There was a clear linear relationship between K1 and CBF in core ROIs (R²=0.75, p<0.01). Ki exhibited similar relationships to CBF, but showed more variability due to changes in k3. Conclusion We found impaired FDG (glucose) transport within the vicinity of lesions that was linearly related to CBF below a threshold value of 30ml/100ml/min. While a similar relationship was found for FDG influx (Ki), variability in k3 suggests different metabolic derangements are responsible. Further analyses will explore how such regional heterogeneity relates to evidence of necrosis, classical ischemia, and other metabolic derangements.