IntroductionFailure rates for screw fixation in femoral neck fracture in young patients are often high, with risk of aseptic femoral head osteonecrosis and non-consolidation. The present study sought to identify factors for success or failure of internal fixation according to: 1) initial treatment; 2) initial reduction quality; and 3) population characteristics.HypothesisThe study hypothesis was that population, fracture type, initial treatment and reduction quality can predict survival.Material and methodsA retrospective study included all cases of femoral neck fracture in under 65-years-old treated by screwing in our center: i.e., 112 patients. Patient characteristics, time to surgery were collated; surviving patients were followed up at a minimum 24 months. Reduction quality was assessed on X-rays in 3 dimensions and cervico-diaphyseal angle.ResultsMean follow-up was 5.3 ± 3.0 years [range, 2.0–13.6 years]. At 2 years, 23 of the 112 patients (20.5%) had developed complications: 10 osteonecroses (8.9%) and 13 non-unions (11.6%). Known hip osteonecrosis risk factors showed no significant association with survival. Failure rates were significantly higher in unstable (Garden ≥ 3) than stable (Garden ≤ 2) fracture: HR = 2.77 [95%CI: 1.09–7.02]; p = 0.025. There was no significant association with time to treatment (≤ 6 hours): HR = 1.08 [95%CI: 0.46–2.54]; p = 0.86. On 2-year radiographs, mean shortening on the z-axis was 12.3 ± 4.8 mm [−0.7 to 26.2], 8.5 ± 5.0 mm [−6.8 to 23.9] on the x-axis, and 6.4 ± 6.1 mm [−6.3 to 25.3] on the y-axis. There was a significant negative correlation between z shortening and HOOS pain component (r = −0.38; p = 0.005), a non-significant negative correlation with quality of life (r = −0.20; p = 0.16), and a significant negative correlation with sports activity (r = −0.28; p = 0.039).ConclusionThe present series showed lower rates of complications and of arthroplasty than in the literature. Internal fixation seemed to be indicated even at an interval of 6 hours or more.Level of evidenceIV, retrospective study.