A retrospective study based on Danish National Registries 1995-2011 covering diagnoses, medications, as well as clinical outcomes. The main analysis was time dependent poisson regression models with results presented as rate ratios (RR), and incidence per 1000 patient years. Steroid use was defined as minimum one injection during April-July for at least three consecutive years. Treatment with specific immunotherapy against grass, birch or both was used as non-steroid control group. Relative risk of adverse outcomes such as osteoporosis, infections, diabetes and/or tendon rupture was investigated.