Mild primary hyperparathyroidism (PHPT) is known to affect the skeleton, even though patients usually are asymptomatic. Treatment strategies have been widely discussed. However, long-term randomized studies comparing parathyroidectomy to observation are lacking. The objective was to study the effect of parathyroidectomy (PTX) compared with observation (OBS) on bone mineral density (BMD) in g/cm and T-scores and on biochemical markers of bone turnover (P1NP and CTX-1) in a prospective randomized controlled study of patients with mild PHPT after 5 years of follow-up. Of 191 patients with mild PHPT randomized to either PTX or OBS, 145 patients remained for analysis after 5 years (110 with validated DXA scans). A significant decrease in P1NP (p < 0.001) and CTX-1 (p < 0.001) was found in the PTX group only. A significant positive treatment effect of surgery compared with observation on BMD (g/cm ) was found for the lumbar spine (LS) (p = 0.011), the femoral neck (FN) (p < 0.001), the ultradistal radius (UDR) (p = 0.042), and for the total body (TB) (p < 0.001) but not for the radius 33% (Rad33), where BMD decreased significantly also in the PTX group (p = 0.012). However, compared with baseline values, there was no significant BMD increase in the PTX group, except for the lumbar spine. In the OBS group, there was a significant decrease in BMD (g/cm ) for all compartments (FN, p < 0.001; Rad33, p = 0.001; UDR, p = 0.006; TB, p < 0.001) with the exception of the LS, where BMD was stable. In conclusion, parathyroidectomy improves BMD and observation leads to a small but statistically significant decrease in BMD after 5 years. Thus, bone health appears to be a clinical concern with long-term observation in patients with mild PHPT. © 2017 American Society for Bone and Mineral Research.