A 13-year departmental sample of 34 patients with definite (biopsy-verified) giant cell arteritis (GCA) was reviewed. The mortality of this material was compared to sex-, age- and time-specific death rates in the Danish population. The standardized mortality ratio (SMR) was 1.8 (95% confidence limits, 1.1-2.8). During the same period 146 patients with probable (not biopsied, but clinically diagnosed in the department) GCA and 85 cases of possible (diagnosed and treated before admission) GCA had been admitted to the department. Those two groups did not differ from the biopsy-verified group with respect to SMR, sex distribution or age. In the group of patients with department-diagnosed GCA (definite + probable = 180 patients), the 95% confidence interval for the SMR of the women included 1.0. In all other subgroups there was a significant excess mortality. Excess mortality has been found in two of seven previous studies on survival in GCA. The prevailing opinion that steroid-treated GCA does not affect the life expectancy of patients is probably not correct.