Between September 2004 and December 2005 we carried out a prospective study of all cases of sepsis of the hip in childhood at a South African regional hospital with a large local population, and which also took referrals from nine rural hospitals. The clinical, radiological, ultrasound and bacteriological features were assessed. All the hips were drained by arthrotomy and the diagnosis was confirmed microbiologically and histologically. Hips with tuberculosis were excluded. The children were reviewed in a dedicated clinic at a mean follow-up of 8.1 months (3 to 18).

There were 40 hips with sepsis in 38 patients. Two patients were lost to follow-up. Nine (24%) had multi-focal sepsis. Overall, 13 hips (34%) had a full and uncomplicated clinical and radiological recovery and 25 (66%) had complications. All patients treated by arthrotomy and appropriate antibiotics within five days of the onset of symptoms had an uncomplicated recovery. Initial misdiagnosis was associated with a delay to arthrotomy. However, ‘deprivation’, consultation with a traditional healer, maternal educational attainment and distance to a primary health-care facility were not associated with delay to arthrotomy. The early correct diagnosis of this condition, common in the developing world, remains a significant factor in improving the clinical outcome.

We compared the accuracy of the growth remaining method of assessing leg-length discrepancy (LLD) with the straight-line graph method, the multiplier method and their variants. We retrospectively reviewed the records of 44 patients treated by percutaneous epiphysiodesis for LLD. All were followed up until maturity. We used the modified Green–Anderson growth-remaining method (Method 1) to plan the timing of epiphysiodesis. Then we presumed that the other four methods described below were used pre-operatively for calculating the timing of epiphysiodesis. We then assumed that these four methods were used pre-operatively. Method 2 was the original Green–Anderson growth-remaining method; Method 3, Paley’s multiplier method using bone age; Method 4, Paley’s multiplier method using chronological age; and Method 5, Moseley’s straight-line graph method. We compared ‘Expected LLD at maturity with surgery’ with ‘Final LLD at maturity with surgery’ for each method. Statistical analysis revealed that ‘Expected LLD at maturity with surgery’ was significantly different from ‘Final LLD at maturity with surgery’. Method 2 was the most accurate. There was a significant correlation between ‘Expected LLD at maturity with surgery’ and ‘Final LLD at maturity with surgery’, the greatest correlation being with Method 2. Generally all the methods generated an overcorrected value. No method generates the precise ‘Expected LLD at maturity with surgery’. It is essential that an analysis of the pattern of growth is taken into account when predicting final LLD. As many additional data as possible are required.

Cite this article: Bone Joint J 2013;95-B:993–1000.

We retrospectively reviewed the records of 16 children treated for spondylodiscitis at our hospital between 2000 and 2007. The mean follow-up was 24 months (12 to 38). There was a mean delay in diagnosis in hospital of 25 days in the ten children aged less than 24 months. At presentation only five of the 16 children presented with localising signs and symptoms. Common presenting symptoms were a refusal to walk or sit in nine children, unexplained fever in six, irritability in five, and limping in four. Plain radiography showed changes in only seven children. The ESR was the most useful investigation when following the clinical course of the disease. Positive blood cultures were obtained in seven children with Staphylococcus aureus being isolated in five. Antibiotics were used in 14 children and spinal bracing in six. Children with spondylodiscitis often present with a confusing clinical picture leading to late diagnosis.

The early use of MRI in the investigation of children with an atypical picture may avoid unnecessary delay in starting treatment and possibly prevent long-term problems. All except one of our children had made a complete clinical recovery at final follow-up. However, all six children in the > 24-month age group showed radiological evidence of degenerative changes which might cause problems in the future.