Incidence of subarachnoid hemorrhage - Role of region, year, and rate of computed tomography: A meta-analysis

Background and Purpose The incidence of subarachnoid hemorrhage (SAH) has been estimated for many years al 10 to 15 per 100 000 person-years, but the most recent studies yield lower figures, of 6 to 8 per 100 000 person-years. To investigate the cause of this apparent decline, we studied the influence of year of study, rate of CT, and region. Methods We performed a meta-analysis of SAH incidence studies from 1960 onward and assessed the influence of the three factors by means of univariate and multivariate Poisson regression analyses. Results Eighteen studies fulfilled predefined inclusion criteria. In three Finnish studies, the pooled incidence was 21.4 per 100 000 person-years (95% confidence interval [CI], 19.5 to 23.4); in 15 non-Finnish studies, it was 7.8 per 100 000 person-years (95% CI, 7.2 to 8.4). With univariate analysis, in non-Finnish studies the incidence decreased 0.96% for each percentage point increase of patients investigated with CT (rate ratio, 0.9904; 95% CI, 0.9878 to 0.9930). With 100% CT scanning, the incidence of SAH outside Finland is estimated at 6 per 100 000 person-years. The rate ratio for year of study was 0.952 (95% CI, 0.935 to 0.969) for each later year in the period 1960 to 1994. In multivariate analysis, only the use of CT was independently related to SAH incidence. For the Finnish studies. the rate ratios for use of CT and year of study were not statistically significant. We also found in six studies that incidence for women was 1.6 (95% CI, 1.1 to 2.3) times higher than that for men (7.1 [95% CI, 5.4 to 8.7] and 4.5 [95% CI, 3.1 to 5.8], respectively). Conclusions The actual incidence of SAH has remained stable over the last three decades; the apparent decline in incidence is entirely explained by the greater proportion of patients investigated with CT. The incidence of SAH in Finland is almost three times as high as in other parts of the world.