Positron emission tomography (PET) is increasingly being used to study dopamine receptor occupancy and the clinical effects of antipsychotic medication. Dopamine D2 receptor occupancy has been shown to predict several clinical effects of antipsychotic medication including therapeutic response, motor and endocrine side-effects. Plasma levels may be used as a surrogate marker for central occupancy if the relationship between these two measures may be accurately described. This study was designed to test the capacity of a previously derived relationship equation (%D2 occupancy=plasma level/ED50+plasma level, where ED50= 0.40 ng/ml) to predict striatal D2 occupancy from plasma level. Twenty-one patients receiving treatment with low dose haloperidol underwent a 11C-raclopride PET scan to measure D2 occupancy. The D2 occupancy levels were accurately predicted by use of the previously generated equation with only a small degree of error (3.89% CI 0.45-7.33). Predicted and measured D2 occupancy values correlated closely (Pearson's r=0.864, P=0.003). The study indicates that reliable prediction of D2 occupancy from plasma levels is possible. This provides a potentially useful surrogate measure of D2 occupancy for research and possibly clinical practice, as the routine use of PET to measure occupancy levels is not feasible.