Intended for healthcare professionals

Clinical Review Lesson of the week

Osteomyelitis in the well looking afebrile child

BMJ 2002; 324 doi: https://doi.org/10.1136/bmj.324.7350.1380 (Published 08 June 2002) Cite this as: BMJ 2002;324:1380
  1. L P Ferguson, senior house officer,
  2. T F Beattie, consultant in accident and emergency care (Louise.Cowan{at}luht.scot.nhs.uk)
  1. Department of Accident and Emergency Medicine, Royal Hospital for Sick Children, Edinburgh EH9 1LF
  1. Correspondence to: T F Beattie
  • Accepted 5 November 2001

Osteomyelitis should be considered even if there is a history of trauma

Musculoskeletal pain and limp are common childhood presentations to general practitioners and accident and emergency departments. The differential diagnosis is broad and includes trauma, bone or joint sepsis, primary or metastatic tumours of bone, and intentional injury.1 Diagnosis of bone or joint sepsis should be straightforward in patients with limb pain who are systemically unwell and have fever. Diagnostic difficulty arises when the patient presents early or with a history of trauma. We describe two children presenting with limp after trauma who looked well but were found to have osteomyelitis.

Case reports

Case 1

A 10 year old girl presented to our department with a painful right knee. The pain had started when she struck her knee on the step of a swimming pool ladder while climbing out. She had continued to swim and play that day, but her mother noticed a limp. The following day she became unwell, with malaise, anorexia, and fever. She attended her doctor. Otitis media was diagnosed, and amoxicillin was prescribed. The knee pain was attributed to soft tissue injury. The pain worsened that night and she was brought to our department the next morning. On examination she appeared well and was afebrile. She walked with an antalgic gait. Mild swelling and tenderness was present over the right medial tibial plateau but there was no effusion. She had full range of movement of the knee. The tympanic membranes were normal, and general examination was unremarkable. Plain radiographs of the knee showed no evidence of fracture but some subtle patchy changes in the density of the proximal tibia. She had a normal full blood count (white cell count 6.6×109/l; normal range 5.0-12.0×109/l). However the erythrocyte sedimentation rate was 60 mm in the first hour (normal range 0-20 mm in the first hour) and serum C-reactive protein was 27 mg/dl (normal range 0-1 mg/dl). Osteomyelitis of the proximal tibial metaphysis was confirmed by magnetic resonance imaging (figure). She was admitted and treated with intravenous flucloxacillin and benzylpenicillin. Blood cultures yielded no growth. She developed a low grade fever during treatment. Repeat magnetic resonance imaging seven days after admission showed an intraosseous abscess. The abscess was drained surgically, and culture of the pus grew Staphylococcus aureus sensitive to flucloxacillin. She had no symptoms at six months' follow up.

Figure1

Short tau inversion recovery magnetic resonance image showing abnormal high signal in proximal tibial metaphysis in case 1. (Reproduced with patient's permission)

Case 2

A 7 year old boy was referred by his doctor with a two day history of pain in his left ankle. The pain had started during a game of football in which he had been struck several times. On examination he had mild swelling and tenderness around the medial malleolus of his left ankle. He had full range of movement of the ankle and was weight bearing with an antalgic gait. Plain radiographs showed no bony abnormality. He was discharged with a diagnosis of ankle sprain and given instructions to mobilise. The following day he became unwell. He became anorexic and felt hot. He returned to our department the next day because the ankle swelling had increased. He looked well but had a more noticeable limp. His axillary temperature was 37.6°C. There was increased soft tissue swelling, with tenderness of the medial and anterior distal tibial metaphysis. He had full range of movement of the ankle and no groin lymphadenopathy. His general examination was normal. The erythrocyte sedimentation rate was 84 mm in the first hour and serum C-reactive protein concentration was 2.2 mg/dl. A full blood count was normal (white cell count 10.3 × 109/l). Blood cultures yielded no growth. He was admitted. Osteomyelitis of the distal tibial metaphysis was confirmed with magnetic resonance imaging. His fever spiked, but this settled with intravenous flucloxacillin and benzylpenicillin. He made a full recovery.

Discussion

Osteomyelitis may occur at any age but is most common in childhood. It remains an important disease in paediatric orthopaedic practice. The most common presentation of osteomyelitis in children is acute haematogenous osteomyelitis. In children the metaphyses of rapidly growing long bones are most often involved, and S aureus is the commonest causative bacterium.2 Acute haematogenous osteomyelitis may be complicated by growth arrest septic arthritis, and chronic infection, but morbidity can be noticeably reduced if the condition is treated shortly after its onset.2 Delay in diagnosis and treatment is not uncommon, however, and in one study 24 of 28 children with acute haematogenous osteomyelitis had an initially incorrect diagnosis despite all presenting with musculoskeletal pain.3

The classic textbook description of acute haematogenous osteomyelitis includes severe pain, fever, and toxaemia.4 Yet presenting symptoms are not uncommonly mild, and our children were not unusual in appearing well. Also, children with bone sepsis often do not have an increased temperature or white cell count.57 Patients may be afebrile at presentation owing to the prior administration of an analgesic with an antipyretic effect. The link between musculoskeletal pain and systemic disturbance is often not made until the infection worsens, causing severe localised pain or refusal to weight bear. The non-specific nature of the systemic disturbance in early osteomyelitis may be easily attributed to a viral illness.

Early diagnosis of osteomyelitis is more difficult in patients with limb pain and a history of injury, where pain is more likely to be linked to the trauma. A history of injury is common in osteomyelitis; in one review 35% of patients had had trauma to the affected bone.8 The history of antecedent trauma may be coincidental and unrelated, although it has become increasingly recognised that trauma can be a causal factor for osteomyelitis. 9 10 In patients with suspected osteomyelitis, it is useful to measure inflammatory markers. The erythrocyte sedimentation rate is usually increased and serum C-reactive protein concentration is normal in only 2% of patients at the time of admission.7 In patients with normal inflammatory markers, diagnosis may be made by early follow up and further investigation if symptoms fail to settle.11 Normal plain radiographs do not exclude osteomyelitis. Actual bony changes are not evident for seven to 10 days, but local oedema may cause displacement of the muscle plane and obliteration of the normal intermuscular fat planes evident on earlier films.12

Early diagnosis of bone sepsis is important to prevent complications. Osteomyelitis should be consid-ered in the differential diagnosis of patients presenting with a painful limb or limp even if there is a history of trauma. A careful history avoids ascribing continued pain and swelling to the initial injury, and inflammatory markers can help confirm clinical suspicion. Patients with osteomyelitis may look well and be afebrile. Patients who are discharged with a diagnosis of trauma should be told to expect improvement in their symptoms. They should be reviewed early if they become systemically unwell or their symptoms do not settle.

Acknowledgments

We thank M McPhillips for her interpretation of the radiological investigations, L Cumming for production of the figure, and JE Robb and IH Annan, the consultants who managed cases 1 and 2 in hospital, respectively.

Contributors: LPF was involved in the diagnosis of both cases and was the main author. TFB was the consultant in charge of both cases in the emergency department and was involved in revising the manuscript. TFB will act as guarantor for the paper.

Footnotes

  • Funding None.

  • Competing interests None declared.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
View Abstract