It is unclear which treatment should be given after stopping teriparatide therapy for severe osteoporosis. In a prospective, randomized, controlled, 2-yr study, we compared BMD effects and clinical safety of three follow-up treatments (anabolic with teriparatide, antiresorptive with raloxifene, or no active treatment) after 1 yr of teriparatide. Postmenopausal women with osteoporosis and a recent fragility fracture received open-label teriparatide (20 microg/d) for 12 mo before they were randomized (3:1:1) to continue teriparatide (n = 305), switch to raloxifene 60 mg/d (n = 100), or receive no active treatment for the second year (n = 102). All patients received calcium and vitamin D supplementation. Changes in areal BMD from baseline to 24 mo were analyzed using mixed-model repeated measures. Daily teriparatide treatment for 2 yr significantly increased spine BMD by 10.7%. Patients receiving raloxifene in year 2 had no further change in spine BMD from year 1 (change from baseline, 7.9%), whereas patients receiving no active treatment had a BMD decrease of 2.5% in year 2 (change from baseline, +3.8%). At the total hip, BMD increases from baseline at 2 yr were 2.5% with teriparatide, 2.3% with raloxifene, and 0.5% with no active treatment; the respective changes at the femoral neck were 3.5%, 3.1%, and 1.3%. The study had insufficient power to assess antifracture efficacy. In conclusion, BMD increases progressively over 2 yr of teriparatide therapy in women with severe osteoporosis. After discontinuation of teriparatide, raloxifene maintains spine BMD and increases hip BMD.