In a controlled, prospective, double-blind study, the incidence of accurate injection of the abductor pollicis longus (APL) and extensor pollicis brevis (EPB) tendon compartments was defined and correlated with clinical relief of de Quervain's tendinitis. X-ray dye was included with steroid and lidocaine injections for 19 patients; dye location was immediately checked by 1 radiologist blinded to the clinical results. Dye was confirmed to be within the first dorsal compartment in 16 of 19 cases. There was relief of symptoms in 11 of 19 patients. Four of 5 patients with dye in both the APL and EPB tendon compartments, experienced relief of symptoms, while all 3 with dye in neither compartment experienced no relief. This suggests that accurate injection of steroids is required for relief of de Quervain's tendinitis. The EPB compartment was often missed (13/19 cases), possibly because it was separate or of small size and deep location. This may be a factor in failed injections, just as surgery can fail if a separate EPB compartment is not released.