Children’s orthopaedics

Closed reduction for DDH: worth a thought?

Developmental dysplasia of the hip (DDH) is one of the leading causes of early osteoarthritis and the subsequent need for hip arthroplasty. A spectrum of disease exists, from mild dysplasia to frank dislocation; the treatment goal is to obtain a concentric and stable reduction to allow normal acetabular growth and remodelling. The driver for DDH also varies considerably with moulding deformities, genetic, and neuromuscular causes. The role of closed reduction (CR) and hip spica casting is being increasingly questioned due to the mixed results following failure of reduction, and the risk of iatrogenic avascular necrosis of the femoral head (AVN). Researchers from the International Hip Dysplasia Institute (IHDI) in Orlando, Florida (USA) have, therefore, established and enrolled patients into a multicentre and multinational cohort of patients to examine the long-term outcomes following this procedure, but here they report the short-term results of the study.¹ Data were prospectively collected for patients with infantile DDH treated between 2010 and 2014. In all, 87 hips in 78 patients were evaluated with a median age at initial reduction of eight months; most also underwent adductor tenotomy at that time. Overall, 79 hips were initially successfully treated with CR and, at the most recent follow-up at a median of 22 months, 72 remained stable giving an overall failure rate of 9%. Likelihood of failure of treatment was not affected by initial reducibility, age at initial CR, or previous brace treatment. More concerningly, 25% developed radiological evidence of AVN. Yet, the risk of AVN was unaffected by the presence of an ossific nucleus, by previous brace treatment, age at CR, or pre-reduction reducibility. Duration of brace treatment was, however, predictive of AVN with a mean period of 15 weeks in those who developed AVN versus 12 weeks in those who did not. The acetabular index and IHDI grade were recorded to measure residual dysplasia. The mean acetabular index at latest follow-up was 25° (sd 6°). Indeed, eight of the 72 hips that were initially successfully reduced subsequently closed and were treated for residual dysplasia with femoral and/or acetabular osteotomies. Moreover, an older age at initial CR was predictive of this. Overall, this is a really interesting study that removes a lot of the methodological issues of previous work.

Intermediate to long-term results of femoral neck lengthening (Morscher osteotomy)

Coxa brevis or short femoral neck, in association with overgrowth of the greater trochanter, are thought to be caused by an ischaemic event in the proximal femoral epiphysis. This may be caused by a number of pathologies such as Perthes disease, developmental dysplasia of the hip, or following infection. The condition is, in itself, not common, but when it does occur the biomechanical consequences are relatively predictable, with the reduced lever arm causing abductor insufficiency and Trendelenburg gait, as well as a mild leg-length discrepancy and impingement of the trochanters on surrounding structures. Given the mechanical nature of the problem and predictable biomechanical consequences, there ought to be a simple solution with an appropriate osteotomy. In the early 1980s, Morscher developed a femoral neck lengthening osteotomy consisting of a double osteotomy of the shaft and greater trochanter, lateralizing the femur along the line of the neck and restoring the abductor tension. This paper from Atlit (Israel) was performed by surgeons in three centres who followed what is probably the largest series of patients treated using this technique.² Harris Hip Scores, leg-length discrepancy, and satisfaction scores at long-term follow-up were assessed, as well as the proportion of hips undergoing subsequent arthroplasty. Overall, 20 hips in 18 patients underwent treatment with the Morscher osteotomy and were followed up to a median of seven years. The underlying pathology was Perthes disease in ten patients, DDH in four, and avascular necrosis in the other four. The median age at surgery was 16 years and all patients had a limp and Trendelenburg gait.

Harris Hip Score improved from 72 preoperatively to 94 postoperatively, leg-length discrepancy was reduced in 17 of the 20 hips and the Trendelenburg test became negative in 14 hips. Three patients had progression of osteoarthritis with one undergoing total hip arthroplasty after four years, and another two patients undergoing arthroplasty after ten years. Patient satisfaction scores were good to excellent in 12 patients, fair in four, and bad in two. There were also implant failures including blade migration in one patient and wire breakage in two patients. There is little evidence regarding long-term prognosis in the literature, and the cohort described are still intermediate term. The surgery is technically challenging and, therefore, should be performed infrequently, but these results give an indication of what is achievable in specialist hands.

Does late hip dysplasia occur after normal ultrasound screening in breech babies?

Contralateral slip SCFE: what’s the risk?

Another paper, another controversial topic, this time concerning contralateral pinning of hips with unilateral slipped capital femoral epiphysis (SCFE). Patients presenting with a single sided slip can be prophylactically pinned on the contralateral side, which obviously carries the risk of complications in those patients who would not go on to develop a slip. If only there was a method of predicting the risk of contralateral slip, then a better judgement could be formed of the risks and benefits of this second, prophylactic intervention.

This is precisely what the authors of this paper from Boston (USA) have attempted to do.⁴ They designed a study that sought to examine if the morphology of the femoral head-neck junction had any influence on the risk of subsequent contralateral slip, and also to determine what the threshold measurements and number needed to treat (NNT) to facilitate informed discussion with parents on the risks and benefits of contralateral intervention were. This is a radiological prognostic study based around measurements of the morphology of the femoral neck and head in a large cohort of patients. Both previously identified risk factors and some less well-known measurements were examined as potential prognosticators for contralateral slip. The authors reviewed all of the radiographs for over 300 patients and established the epiphyseal tilt and epiphyseal extension ratio, as well as the alpha angle and epiphysis angle of the uninvolved hip. The posterior slip angle, as visualized on a frog lateral, is a known risk factor of SCFE; however, the concept of the superior epiphyseal extension being protective is less well known – although it has previously been hypothesized as not being present in slipped hips. Multivariate logistic regression was performed and receiver operating characteristic (ROC) curves were constructed to determine the optimum measurements predictive of contralateral slip. Overall, 318 patients were included over a 17-year period with a mean age of 12 years and a minimum follow-up of 18 months; male to female split was essentially even. The lateral tilt angle and superior epiphyseal extension ratio were identified as independent factors associated with contralateral slip. Interestingly, for each degree, the posterior tilt increased contralateral slip probability by 8% and for each 0.01 increase in the superior epiphyseal extension ratio the risk of contralateral slip decreased by 6%. This work also found that a posterior tilt of > 10° confirmed a 49% risk of contralateral slip compared with a 19% chance of slip in < 10°. The epiphyseal height was found to be an independent risk factor, but a cut-off score was not able to be quantified. Interestingly, this paper found that neither alpha angle nor epiphyseal angle were risk factors for slip, which is a contrarian position when considered alongside other published research. While this is interesting work, the measurements do strike us as quite subjective, and we suspect strong advocates of prophylactic pinning or otherwise will be relatively unmoved. Yet, all efforts to inform the debate should be welcomed.

Late detection of DDH in England? A 26-year national study

Another worthy read on DDH is included in this issue of 360, again from the team in Southampton (UK).⁵ As we know, DDH diagnosed later than the age of a few months old is associated with poorer outcomes and increased complications. As such, since 1986 selective screening has been mostly standard practice across the UK . However, the experience in the UK is that late presentation of DDH continues to occur, as many of these children do not have the risk factors required to trigger a scan. Indeed, other European countries have universal screening programmes. The question asked by this thoughtful paper is are we right to do so? This universal screening approach does lead to a lower rate of late detected DDH (not zero as outlined by the previous paper) – but is costly, resource intensive, and may lead to overtreatment. This study uses ‘big data’ to examine the incidence of DDH diagnosed after the age of one year in England, and to assess for geographic distribution and any age or sex variability. The UK Clinical Practice Research Datalink (CPRD) and Hospital Episode Statistics (HES) databases were used to identify an ‘at risk’ population of over 3.5 million patients and a study population of 754 patients whose DDH was detected between the ages of one and eight years old. The UK Clinical Practice Research Datalink (CPRD) covers around 1:20 patients in the UK, while the HES database covers all treatment episodes. For the purposes of this paper, DDH due to neuromuscular diseases, syndromes, and trauma were excluded from the analysis. The current incidence of late diagnosis is 1.28 per thousand live births, which is higher than the rate recorded in some areas of England almost 40 years ago prior to the introduction of selective screening. This is, quite understandably, a huge concern. Unsurprisingly, the pickup rate was highest in the age group of one to two years as the diagnosis is often made as children start to walk and 86% of cases were detected by the age of three. There was no significant geographic variation between different regions of England, indicating that the problem remains a national one. Naturally, this paper suffers from many of the criticisms levelled at papers utilizing large databases, including accuracy of the primary coding and data integrity. Interestingly, one limitation introduced an underestimation, as cases diagnosed after six months of age could be construed to be late diagnoses but were unable to be studied due to the binary one-year age categorization. Overall, the authors conclude that the selective screening programme appears to have had little impact on the rates of late diagnosed DDH, which remains a significant cause of disability and litigation. The debate surrounding the optimal means of screening for DDH will continue, but this paper means that the debate is now better informed.

Multicentre study of physical abuse and limb fractures in young children

Physical child abuse is a sad fact of life and, while we cannot prevent all cases, it is vital that when such cases do present to hospital the opportunity is not missed for detection and intervention. This retrospective review across a whole geographical region in East Anglia (UK) led by a team of researchers from Peterborough (UK) aimed to examine if high standards of practice were occurring uniformly across the region and, if not, which areas of practice required improvement.⁶ Data was gathered over a four-year study period from seven different hospitals and included all high risk paediatric fractures. This was defined as all fractures in those age one year or less and all humeral and femoral fractures in children under three years. Supracondylar fractures were excluded. Review by a paediatrician was recorded, as were skeletal surveys. Multiagency case conference decisions were used as the definition of whether physical abuse had occurred. Relatively unsurprisingly, the at-risk groups for the diagnosis of non-accidental injury were all children under 12 months (10% patients) as well as femoral fractures under 18 months and humeral fractures under 24 months. Interestingly, there was a significant variation between paediatrician review and skeletal survey being undertaken between the hospitals; in one centre, only 46% of children who were under 12 months at time of fracture were reviewed by the paediatrics team. A decision of non-accidental injury was more common in centres with higher rates of paediatric review and use of skeletal survey; this does seem to suggest that the other cases are going undetected. This is the first multicentre review of non-accidental injury in the UK and it is of huge concern that there is such variation seen in practice. From their findings, the authors suggest that to minimize false negative assessments, all high-risk children should be assessed by a paediatrician prior to discharge from the emergency department.

Their proposed criteria seem very reasonable and includes any child under the age of 12 months with any fracture, under 18 months with a femur fracture, and under 24 months with a non-supracondylar humeral shaft fracture. Statistically, this should increase our detection rate, but we should all continue to be aware of the possibility of non-accidental injury in every case that we see.

The reproducibility of ultrasound-based metrics for assessing DDH

Ultrasound assessment is the mainstay radiological method of assessing developmental dysplasia of the hip (DDH), but its utility depends on the reproducibility of the examination as well as the accuracy, which incorporates both reliability and validity. There are a variety of imaging-based metrics reported for the assessment of DDH, the most widespread in clinical use being the Graf acetabular alpha angle. One challenge of assessing the diagnostic accuracy of this test has been the historical variation in what constitutes normal. Indeed, many studies use the outcomes of later examinations such as plain film radiograph at two years of age as a proxy, which assumes that spontaneous resolution of dysplasia is not possible, which of course it is. This study from Vancouver (Canada) aimed to perform a systematic review of reported DDH diagnosis using ultrasound, and hence to estimate and compare the levels of reproducibility in the available DDH metrics.⁷ Following an extensive literature review, 28 studies were included in the overall analysis and the overall evidence quality of these was graded as moderate. Of all the metrics assessed, Graf’s alpha angle demonstrated the least inter-examination variability, which was 10% lower than the variability of the beta angle measurement, the next best available. Both of these measurements had lower variability when contrasted with other metrics of dysplasia. However, despite this finding, the authors found the actual variability in measurement was sufficiently high to be clinically problematic, a conclusion that was supported by the low reported levels of intraclass correlation and Kappa coefficient values. Furthermore, evidence of a deterioration of reproducibility over time was suggested, which is in itself concerning.

What is reported here is somewhat disheartening; overall, there was high variability and low agreement in dysplasia metrics. The failure to maintain standards over three decades indicates that work is urgently needed in this area to make improvements, as the implication of not accurately measuring what we are looking for has huge implications for a test that is the mainstay of early screening programmes.

Are inflatable play structures really safe for our children?