Thirty-eight patients (40 hips) randomly received either an uncemented fully porous-coated composite stem (Epoch; Zimmer, Warsaw, Ind) or an uncemented proximally porous-coated solid stem (Anatomic; Zimmer). Patients were followed up for 7 years using radiostereometry, dual-energy x-ray absorptiometry, conventional radiography, the Harris Hip Score, and a pain questionnaire. Both stem designs achieved excellent outcome for fixation (stem subsidence and stem rotations close to zero) and clinical outcome, without any difference between the 2 groups (P > .12). Median wear rates were low despite use of conventionally gamma-sterilized polyethylene. No stem was radiographically loose on the postoperative radiographs. The low-modulus composite stem had positive effects on early proximal bone remodeling in Gruen regions 1, 2, 6, and 7 (P < .04). However, at 7 years, this bone-sparing effect persisted in only the calcar region (Gruen region 7).