Avascular necrosis of the femoral head often progresses to femoral head collapse if not treated. Conservative treatment yields highly variable results and is not standardised, mainly because it is typically evaluated in small patient populations. This led us to conduct a large retrospective comparative study with the goals of 1) analysing survival and functional outcomes, 2) looking for differences in survival between core decompression techniques (standard versus augmented), and 3) studying the risk factors for femoral head collapse and revision by arthroplasty.

Core decompression limits the number of patients who suffer femoral head collapse requiring arthroplasty at 2 years’ follow-up.

This multicentre, comparative, retrospective study analysed 330 patient records (1975–2016) where at least 2 years’ follow-up was available. Sixty-two patients were excluded from the analysis: 5 had a stage III with collapse, 5 were lost to follow-up, 2 died within 24 months of the procedure and 50 had incomplete data. The study included 263 patients with a mean age of 42 years (15.7–70). In the Ficat classification, there were 51 cases of stage I necrosis, 186 cases of stage II and 22 cases of stage II with crescent sign (transition stage). The Kerboull angle on radiographs was between 5° and 20° in 40 patients, between 20° and 40° in 107 patients, between 40° and 60° in 52 patients and more than 60° in 29 patients. A standard core decompression was done in 79 patients and an augmented one in 184 patients. The more severe AVN cases (stage II) were more likely to be treated by augmented CD (160/184 patients, 87%) than by standard CD (48/79 patients, 61%) (p<0.001).

In the 263 patients, the overall survival (no arthroplasty at 2 years) was 73% (196/263). At 2 years, the survival rate (without arthroplasty) was 71% (56/79) in the standard CD group versus 76% (140/184) in the augmented CD group. This difference was significant when adjusted for Ficat stage and Kerboull angle [HR=0.457, 95% CI (0.247–0.844) (p=0.012)]. When the survival data was adjusted to the Ficat stage, augmented CD was better than standard CD with 10-year survival of 58.1% vs. 57.9% (p=0.0082). More than 30% necrosis volume increased the risk of failure [HR=3.291 95%CI (1.494–7.248) (p=0.0031)]. Also, a Kerboull angle above 60° increased the risk of failure [HR=3.148 95%CI (1.346–7.5) (p=0.0083)].

After 2 years, CD for non-collapsed femoral head AVN prevents collapse and revision to arthroplasty in 73% of cases (196/268). Augmented CD improves the 2-year survival and the long-term survival after adjusting for preoperative characteristics (Kerboullangle and Ficat stage). The risk of collapse and need for arthroplasty is greater in patients with 30% necrosis volume on MRI and Kerboull angle above 60°.

III; retrospective case-control study.