Postoperative complications have been identified as poor prognostic factors for hip survivorship, the most serious being osteonecrosis of the proximal femoral epiphysis. This has been reported to have occurred in 60 % of patients following closed reduction . However, there is considerable variation in the reported incidence, ranging from 0% to 73% [4-8] following closed or open reduction, which is likely to reflect the inconsistency in diagnostic ability. The external validity of the osteonecrosis rates are further undermined by the heterogeneity in the interventions preceding diagnosis, different case definitions, inconsistencies in the outcome measures and variable participant attrition rates. Many studies lack adequate length of follow-up to appreciate radiological signs that may appear after many years .
While the precise mechanism of osteonecrosis is not known, its iatrogenic nature is not disputed as it is not seen in the absence of treatment . Patho-morphological changes of the proximal femur ultimately result in a variable growth disturbance of the proximal femoral physis which can manifest as a lack of femoral head sphericity, shortening and deformity of the femoral neck, abnormality of the metaphyseal region, and overgrowth of the greater trochanter with resultant abductor musculature weakness .
Risk factors for osteonecrosis highlighted in a number of long-term studies include the patient’s age at surgery , perioperative injury to the proximal femoral blood supply , and an eccentric position of the femoral head in plaster .
Osteonecrosis can be devastating as it leads to deformities to the femoral head, with changes in the proximal femoral physis key to predicting residual deformity . Previous work demonstrated osteonecrosis compromises subsequent acetabular remodelling thereby causing secondary abnormalities to the acetabulum . Other deformities resulting from osteonecrosis that are associated with osteoarthritis include shape mismatch in the convexity of the femoral head and concavity of the acetabulum, persistent lateral and proximal subluxation, and medial femoral head irregularity .
Treatments for osteonecrosis concentrate on improving the biomechanics of the hip, as interventions directly addressing early proximal femoral growth disturbances by repairing physeal injuries or altering the blood supply to the physis are not yet available.
Understanding this complication is so important as children and adolescents with this irreversible condition experience early decline in function, pain and premature osteoarthritis; and many patients with ‘severe’ osteonecrosis require hip arthroplasty from late adolescence . Even hips with minor anatomical abnormalities and residual deformity are at risk of premature osteoarthritis because of DDH .
Our previous research showed that the effects of osteonecrosis were benign in childhood: at a mean age of 14 years affected patients demonstrated nearly normal physical functioning, a normal health-related quality of life, and minor limitations of hip function. The present study examined whether this remained the case in adolescence and in young adulthood. To do this, we performed a cross-sectional evaluation of patients at a mean age of 20 years, as well as a longitudinal assessment of 54 patients with a mean age of 22 years who we had studied in 2011.
Changes associated with osteonecrosis (physeal arrest) involve the proximal femur and this could impair hip function. We previously showed that at a mean age of 14 years hip function was minimally affected by osteonecrosis, ranging from 94/100 points on the CHOHES for grade I osteonecrosis, to 78/100 points in grade IV . This study confirmed that, at a mean age of 21 years, it ranged from 86/100 points to 77/100 points. Aguilar et al. , using the CHOHES, found a mean score of 88/100 points in children without any hip problems. This would indicate that the hip function of our young adult patients was reduced in those with osteonecrosis grades III and IV by a degree that was clinically important. However, these differences were no longer seen in the adjusted analysis – scores were above 88/100 points and almost identical across all four grades of osteonecrosis. This suggests that the effects of the osteonecrosis alone did not explain the reduced hip function in grades III and IV. When assessing the CHOHES subscales, ‘physical examination’ scores declined with increasing Bucholz-Ogden grades (Table 3) indicating reduced range of motion in radiographically more severely affected hips. Overall, the isolated effects of the radiographic changes associated with physeal arrest were small in terms of hip-specific function.
In terms of physical function and general health status, the adolescent and young adult patients did well, but those with grade III osteonecrosis showed lower scores than all other grades of osteonecrosis (Table 2). But again, these differences were no longer seen in the adjusted analyses suggesting that regardless of the grade of osteonecrosis, these patients had near normal scores for physical function and for general health status. This is similar to what was found in children at a mean age of 14 years , and with an analysis of normal individuals where ASK scores as low as 80 have been observed in nondisabled children .
From the adjusted analysis it appears that the drivers for low patient-reported outcome scores were not the grades of osteonecrosis, but other factors. Residual acetabular dysplasia is generally associated with less favourable functional outcome scores . Persistent subluxation leads to a reduced contact surface between femoral head and acetabulum, thus increasing the risk for osteoarthritis . Three of the lowest CHOHES scores (<45) in this study were encountered in patients with grade IV osteonecrosis with marked acetabular dysplasia and subluxation. A further three patients with a CHOHES score <45 showed grade III osteonecrosis; all had bilateral DDH and osteoarthritic changes of grades III and IV according to Kellgren-Lawrence . Our comparison group of 32 patients who underwent childhood treatment of DDH further demonstrated that factors other than osteonecrosis drove low outcome scores: 43% of their hips scored less than 85/100 on the CHOHES despite the absence osteonecrosis. Of these, three hips showed a centre-edge angle <20 degrees; three hips had Kellgren-Lawrence grade 0; and six hips had Kellgren-Lawrence grade 1.
In this study we were able to report, for the first time, how patients with osteonecrosis and DDH changed over eight years by re-examining 54 of 72 patients (75%) who took part in a previous study . These changes were minimal, suggesting that patients maintained high levels of hip function, physical function and health status if their hip survived. However, 5 of 72 patients (6%) had needed a hip replacement within this time period, indicating they had severe hip-related disabilities. It is unknown how the remaining 13 patients not included in this follow up study faired over those eight years – this limits our conclusions about how patients change over time. Yet, analysis of the baseline variables does not reveal systematic differences between cohort patients who were or were not recruited for follow-up.
We note other potential limitations of this study. The participants of this study may have been too young to discern the ultimate effects of osteonecrosis on patient-reported outcomes. However, we selected this age group deliberately to gather insight into patients transitioning from paediatric to adult health care services, when activities and demands change after leaving school  often heralding the onset of functional impairments . We utilised two different instruments for measuring ‘physical function’ based on patient age (80% used the HOOS-PS and 20% used the ASK). However, we think this was acceptable because factor analysis showed  that items of the ASK loaded on two distinct factors, ‘activities of daily living’ and ‘play/sport’, which are also the underlying constructs of the HOOS-PS. Both instruments have a very similar method of scoring, but as we could not assume a normal distribution for these response variables (the scores tended to be skewed towards better function), we used quantile regression to establish equivalence of scores.
Gibson and Benson  recognised that following treatment of developmental dysplasia of the hip, poor radiographic appearances of the hip may cause few symptoms until late adolescence or early adult life. By re-assessing patients from our original cohort and by assessing newly recruited patients of a similar age, we sought to further evaluate disease severity associated with osteonecrosis in patients transitioning from paediatric to adult care.