Difficulties in diagnosing vertebral osteomyelitis in a child

Abstract

Vertebral osteomyelitis is a rare diagnosis and often delayed diagnosis in children. This is a case of a child presenting with fever, back pain and raised C reactive protein who was found to have a Staphylococcus aureus (S.aureus) bacteraemia. Initial imaging with CT, MRI of the spine and abdominal ultrasound failed to demonstrate a vertebral osteomyelitis or identify another source of the bacteraemia. Due to the high clinical suspicion of a spinal source of the infection, second-line investigations were arranged. A bone scan identified an area of increase metabolic activity in the 12th thoracic vertebrae (T12) and subsequently a diagnosis was confirmed with a focused MRI of T12. This serves as an opportunity to discuss the diagnostic difficulty presented by paediatric vertebral osteomyelitis and more generally the need for clinicians to pursue their clinical suspicion in the face of false negative results to make an accurate and timely diagnosis.

Background

Diagnoses are made through clinical assessment with a history and examination and when required confirmatory investigations. Tests in all modalities can provide false negative results leading doctors to question their conclusions. In such cases, it is important to be guided by one’s clinical acumen. We report a case where the clinical assessment indicated a spinal infection which was not confirmed by an initial spinal MRI. Reassessment through alternative imaging with a bone scan demonstrated a vertebral osteomyelitis confirming the original clinical suspicion.

Case presentation

A previously well 13-year-old Maori girl presented with a 4-day history of fever and 1 week of lower back pain which radiated to her abdomen, increased with movement and was associated with acute constipation, not opening her bowels for 5 days. She could pass urine without difficulty. She had developed multiple skin abscesses on her right forearm and knee 3 weeks prior to presentation. She had returned from New Zealand 3 weeks earlier and had no known sick contacts. She previously had suffered boils but had no other significant medical history and no risk factors for developing spinal or recurrent skin infections. She did not take any regular medications and had no allergies. Her immunisations were up to date as per the Australian standard vaccination schedule. On examination, she had tenderness along the spine particularly in the lumbosacral region and reduced power, due to pain from the muscle spasm rather than neurological deficit, with preserved reflexes. She demonstrated an antalgic gait, although pain prevented a detailed gait assessment. Assessment of anal tone was not performed. Her right forearm and knee had patches of healed impetigo. Her heart sounds were dual without murmurs and a non-displaced apex. She had no lymphadenopathy nor hepatosplenomegaly. Her ear, nose and throat examination was unremarkable. Subsequent neurological examination that day elicited no neurological deficits and no abnormalities were appreciated on funduscopy.

Investigations

Initial investigations revealed a white cell count of 15.2×10⁹/L with neutrophilia of 12.05×10⁹/L, a C reactive protein (CRP) of 114 mg/L and a blood film with moderate toxic changes. The first two blood cultures grew a methicillin-sensitive Staphylococcus aureus (MSSA). After the commencement of appropriate intravenous antibiotics no further blood cultures grew bacteria. Computer tomography of the abdomen and lumbar spine revealed soft tissue induration of the anterior paravertebral region at the thoracic vertebrae 11–12 level. MRI of the spine showed no evidence of a paravertebral collection, discitis or vertebral osteomyelitis (figure 1A, B). An echocardiogram found no evidence for infective endocarditis. Given the high clinical suspicion of a spinal source of the MSSA bacteraemia, a Technetium-99m labelled bone scan was performed revealing increased uptake in 12th thoracic vertebrae consistent with vertebral osteomyelitis (figure 2). This diagnosis was subsequently confirmed with repeat MRI, 8 days after the first, focused on the 12th thoracic vertebrae (figure 3).

Figure 1.

(A) T2-weighted coronal Short-T1 Interval Recovery (STIR) MRI of the thoracolumbar spine showing no vertebral lesion in T12 vertebra/T11–T12 disc space. (B) T2-weighted axial MRI showing no vertebral lesion in T11 vertebra.

Figure 2.

Tc-99M labelled bone scan showing increased uptake at T12 vertebra.

Figure 3.

Repeat focused T2 diffusion-weighted coronal STIR MRI showing vertebral lesion at T12.

Differential diagnosis

This child presented with low back pain, fever and an MSSA bacteraemia. This puts a spinal infection high on the differential, particularly in the context of a sterile urine and normal chest X-ray, meaning a urine tract infection or occult pneumonia are unlikely. The potential spinal infections would include vertebral osteomyelitis, discitis or an infection of the spinal canal, such as a spinal epidural abscess, subdural empyema or spinal cord abscess. Other occult infections should be considered in this context, particularly infective endocarditis and meningitis.

Treatment

The child was commenced on empiric intravenous antibiotics, including flucloxacillin, on admission. She completed 10 days of intravenous therapy for both the skin infection and vertebral osteomyelitis. Lincomycin was added once the MSSA bacteraemia was identified. She showed a good clinical response with resolution of her skin infection and was ambulatory by day 4 of her admission. She was discharged after a 12-day admission to complete further 6 weeks of oral flucloxacillin.

Outcome and follow-up

She was seen by her general practitioner with complete resolution of her symptoms after completion of the antibiotic course.

Discussion

Epidemiology

Spinal infections in children are uncommon with an incidence of 0.3 per 100 000 in under 20 years in the developed world of which vertebral osteomyelitis accounts for only 1%–2%.¹ Unlike discitis, which typically affects children between 1 and 5, the mean age of presentation of vertebral osteomyelitis is over 7 years.¹ The vertebrae can be infected anywhere along the spine, although there is a predilection for the lower aspect with lumbosacral and thoracolumbar regions affected in 64% and 29% of cases, respectively, while the cervical spine is only involved 7% of the time.² The route of infection is most commonly through haematogenous spread from an isolated organism which, because it enters through the posterior spinal artery, may affect adjacent vertebrae supplied by the same artery.³ However, it should be remembered that the infection can occur through both contiguous spread and direct inoculation, making it important to elicit a history of recent infections, bites, trauma, previous surgery, especially spinal or prostheses. Ideally an infective agent would, in most cases, be identified. This is not always possible as initial culture may prove sterile, while antibiotics are commenced prior to obtaining these results. When identified the most common pathogens are S. aureus and streptococcus species, meaning good gram-positive organism cover is essential to empirical antibiotic therapy.⁴

Diagnosis and diagnostic delay

Given the risks of sepsis and spinal cord injury, pyogenic vertebral osteomyelitis is an important yet challenging diagnosis to make. However, a delayed diagnosis is common as the non-specific nature and the indolent course of the condition can lead to late presentation and slow medical investigation.^{5 6} It is, therefore, important to know when to consider vertebral osteomyelitis as a diagnosis, to that end the infections disease society of America have published guidelines for adults. This recommends consideration under the following circumstances:

1. New or worsening back or neck pain and fever.

2. New or worsening back or neck pain and elevated erythrocyte sedimentation rate or CRP.

3. New or worsening back or neck pain and bloodstream infection or infective endocarditis.

4. Fever and new neurological symptoms with or without back pain.

5. New localised neck or back pain, following a recent episode of S. aureus bloodstream infection.⁷

However, children are not adults and their clinical presentation is not necessarily the same. In children, the most common presenting symptom is back pain, although it may present with neck, shoulder, rib and abdominal pain, followed by fever and neurological deficit, including sphincter dysfunction.² It should be noted that the presentation of pain will depend on a child’s verbal skills, a child under 3 years may well present with refusal to walk or a limp. Clinical examination most frequently finds loss of lumbar lordosis with the coin test and ‘quarter’ sign (difficulty when picking up a coin and irritation of the hip when held in extension) being useful to evoke.^{6 8} However, even if the presentation indicates consideration of the diagnosis, the initial investigations can impact significantly on the time to diagnosis. In adult data, performing a spinal X-ray actually delayed diagnosis as did a CRP <63. Conversely a CRP >63 or a bacteraemia, as were present in our case, expedite the diagnosis.⁵ When clinical history, examination and first line investigations indicate a potential spondylodiscitis, a diagnosis is made using MRI, which has a sensitivity and specificity of 96% and 94%, respectively.⁹ MRI demonstrates the vertebral oedema caused by the osteomyelitis as a T1 hypo intense signal with a T2 hyper intense signal within the vertebrae and provides further information on the surrounding tissues.⁹ This is essential as, when there is extension of the infection, especially involvement of the spinal canal, surgical intervention may be indicated.³ It is important to be aware that MRI is less accurate early in the disease process, suggesting a diagnosis in only 36%–55% of cases presenting within 2 weeks of symptom onset.⁶ Despite this low sensitivity, MRI is still the first choice of investigation early in the disease process as it is non-invasive and pertains no ionising radiation exposure to the child.¹⁰ However, in the setting of the clinical suspicion of spinal infection, with a negative or inconclusive MRI a radionuclide scan should be arranged. If this scan is negative a spinal infection can be excluded and if positive it can be confirmed.⁹ In our case, the child presented with fever, back pain and constipation while initial investigations revealed a raised CRP >63 and S. aureus bacteraemia raising a high suspicion of spondylodiscitis. The initial MRI was performed within 2 weeks of the onset of symptoms, potentially explaining the false negative result. Subsequently, a radionuclide scan was arranged to search for evidence of occult infection, including spondylodiscitis, which lead to the final diagnosis.

Treatment

The mainstay of treatment is antimicrobial therapy with intravenous antibiotics, commenced after collecting blood and urine cultures. Conversion from intravenous to oral therapy, in immunocompetent patients with a native vertebral infection and good clinical response, is likely safe after 2 weeks of intravenous antibiotics. The only independent predictor of a successful conversion is a lower CRP at 2 weeks compared with baseline.¹¹ In our case, we continued intravenous treatment for 1 week before conversion to oral antibiotics, after advice from the tertiary paediatric infectious disease team. Surgery is rarely required and only indicated if there is loss of neurological function, spinal instability, collections or failed conservative therapy. Follow-up imaging may be unhelpful showing stable or increased vertebral body enhancement, bone marrow oedema and disc space enhancement despite clinical improvement.¹²

Patient’s perspective.

My daughter fell ill during her first week of school and complained about a lower back pain that I mistook for a period pain. 72 hours later a visit to our local hospital was needed. She was sent home to take nurofen and panadol as required, however her pain continued throughout the evening into morning where I returned to our local hospital and her journey began with many physicians diagnosing her symptoms. Through consistent and thorough care from all medical staff my daughter recovered well. I was so pleased with the care she received and how concerned all medical staff were with my daughter.

Learning points.

• Vertebral osteomyelitis is a rare, often delayed and difficult diagnosis to make in children.

• It should be considered in the context of back pain associated with fever, an elevated C reactive protein or bacteraemia.

• MRI is the gold-standard investigation, but in the first 2 weeks of symptoms is less accurate and may give a false negative result.

• A radionuclide scan is the best second line of investigation in such cases.

• Doctors should reasonably pursue their diagnostic suspicion, in the face of unexpectedly negative results and rely on their clinical acumen rather than purely on investigations.