Background In patients with nontraumatic osteonecrosis of the femoral head (ONFH), implantation of bone marrow aspirate concentrate (BMAC) could delay the progression of osteonecrosis and improve symptoms in pre-fracture ONFH. However, the BMAC content, especially in osteoblastic stem cells, could have an important individual variability. An autologous osteoblastic cell product could improve the effect of such cell-based therapy. Questions/purposes (1) Does autologous osteoblastic cell therapy decrease the likelihood of progression to subchondral fracture with or without early collapse corresponding to Association Research Circulation Osseous (ARCO) classification Stage III or higher, and provide a clinically important pain improvement compared with BMAC treatment alone? (2) Were patients treated with osteoblastic cell therapy less likely to undergo subsequent THA? (3) What proportion of patients in the treatment and control groups experienced adverse events after surgery? Methods Between 2004 and 2011, we treated 279 patients for Stage I to II hip osteonecrosis (ON) with surgery. During that time, our general indications for surgery in this setting included non-fracture ON lesions. To be eligible for this randomized, single-blind trial, patients needed to have an ONFH Stage I or II; we excluded those with traumatic ONFH, hemoglobinopathies and positive serology for hepatitis B, C or HIV. Of those treated surgically for this diagnosis during the study period, 24% (67) agreed to participate in this randomized trial. Hips with pre-fracture ONFH were randomly treated with a core decompression procedure associated with either implantation of a BMAC (BMAC group; n = 26) or osteoblastic cell (osteoblastic cell group; n = 30). The groups were not different in terms of clinical and imaging characteristics. The primary study outcome was treatment response, defined as the absence of progression to subchondral fracture stage (ARCO stage III or higher) plus a clinically important pain improvement defined as 1 cm on a 10-cm VAS. The secondary endpoint of interest was the frequency in each group of subsequent THA and the frequency of adverse events. The follow-up duration was 36 months. We used an as-treated analysis (rather than intention-to-treat) for our efficacy endpoint, and an intention-to-treat analysis for adverse events. Overall, 26 of 26 patients in the BMAC group and 27 of 30 in the osteoblastic cell group completed the trial. Results At 36 months, no clinically important differences were found in any study endpoint. There was no difference in the proportion of patients who had progressed to fracture (ARCO stage III or higher; 46% of the BMAC hips [12 of 26] versus 22% in the hips with osteoblastic cells [six of 27], hazard ratio, 0.47 [95% CI 0.17 to 1.31]; p = 0.15). There was no clinically important difference in VAS pain scores. No differences were found for either the WOMAC or the Lequesne indexes. With the numbers available, there was no difference in the proportion of patients in the groups who underwent THA at 36 months 15% (four of 27) with osteoblastic cells versus 35% (nine of 26) with BMAC; p = 0.09 With the numbers available, we found no differences between the treatment and control groups in terms of the frequencies of major adverse events. Conclusions We found no benefit to osteoblastic cells over BMAC in patients with pre-collapse ONFH; side effects were uncommon and generally mild in both groups. This study could be used as pilot data to help determine sample sizes for larger (presumably multicenter) randomized controlled trials. However, this novel treatment cannot be recommended in routine practice until future, larger studies demonstrate efficacy. Level of Evidence Level II, therapeutic study.