February 2006

An 11-year-old girl with known sickle cell disease is admitted to the hospital for evaluation and treatment of fever and pain. She has had multiple admissions in the past for febrile illnesses and/or vaso-occlusive crises and many more, milder pain crises that were managed at home. Her current problem began with the abrupt onset of thoracolumbar pain on the day of admission. Initially, she had no history of fever, but within a couple of days, she began to spike fevers of over 102º F. Blood cultures were obtained and antibiotic therapy (Ceftriaxone) was started upon admission. She had repeat blood cultures obtained with this fever spike, and vancomycin was added to her empiric therapy.

Examination on admission revealed some point tenderness over her lower thoracic/upper lumbar area, but within a couple of days, some notable, diffuse and painful swelling was seen over the same area as shown in figures 1–3. A complete blood count (CBC), C-reactive protein (C-RP) and an erythrocyte sedimentation rate (ESR) were obtained, revealing a white blood cell count of 13.2 thousand, with 67% granulocytes and no bands, C-RP = 183 and ESR = 73. Plain radiographs of her spine were normal. Selected MRI pictures are shown in figures 4–6.

What’s Your Diagnosis?

1. Staphylococcus aureus
2. salmonella
3. shigella
4. Streptococcus pneumonia

Answer

The answer is B, salmonella. This is a case of osteomyelitis of the spinous process of the first and second lumbar vertebrae. Compare figure 7, an uninvolved vertebra, with figure 4 to see the abnormal findings of the spinous process on MRI. In the 1970s, when I was a resident, we were taught that although salmonella was more common in children with sickle cell disease (SCD), the most common cause of osteomyelitis was still Staphylococcus aureus, as it is in children without SCD. However, since then, at least two studies have shown conclusively that salmonella is indeed not only the most common cause of osteomyelitis in patients with SCD, but it is many times more common in these patients than its closest competitor, S. aureus (Givner LB, et al. J Pediatr. 1981;99:411-413/Burnett MW, et al. Pediatrics. 1998;101:296-297).

Patients with SCD are at increased risk of infection because of impaired splenic function as well as the damage caused by the microinfarcts that occur due to vaso-occlusive ischemia, periodically occurring in the bone as well the gut mucosa. Diagnosis is usually problematic: The signs and symptoms of a vaso-occlusive crisis can be identical to early osteomyelitis, including fever. The lab is also of little help for the same reasons. Bone scan and MRI may also be abnormal.

One difference in the patient presented is the surrounding cellulitis seen on exam and the MRI. As it became clearer that she likely had vertebral osteomyelitis, an attempt to make a diagnosis was done by interventional radiology with a needle aspirate and biopsy (figure 8). However, as expected, the stains and cultures of material removed were negative, probably due to the antibiotic treatment she had been on.

In an attempt to simplify her therapy, vancomycin was discontinued and clindamycin was started to continue to cover for the possibility of resistant Staphylococcus. She was left on ceftriaxone for probable salmonella, and was discharged on home IV therapy for four to six weeks. Orthopedic management may range from immobilizing with a body cast to simply decreasing activities for a couple of months, which was the case here.

When I think of vertebral osteomyelitis, I think of an infection of the vertebral body. I am assuming that if it involves only the spinous process, it still falls into this diagnostic category. Also of interest, salmonella is the second most common cause of vertebral osteomyelitis in patients without SCD, up to 14% of all cases, according to one review (Correa AG, et al. Pediatr Infect Dis J. 1993;12:228-233). Fortunately, it is fairly rare, even in patients with SCD.

So, for those of you preparing for the Pediatric Boards, remember that salmonella is the most common cause of osteomyelitis in patients with SCD, not S. aureus. And empiric therapy for osteomyelitis in SCD patients should include coverage for salmonella as well as S. aureus.

In our geographic area, where methicillin-resistant S. aureus is almost always sensitive to clindamycin, we use a combination of clindamycin and ceftriaxone, pending culture results. Of course, the patient is followed by a team that consists of a hematologist as the “team leader,” along with infectious diseases and orthopedic consultants. Generally, the clinical status is closely monitored along with the inflammatory markers — ESR and CR-P — and follow-up MRI imaging, continuing antibiotic therapy until it is clear that the infection is resolved. Most experts recommend a minimum of four weeks of treatment, with six weeks more likely being used.

I am going to spare you my usual comments. Your brain probably needs a rest from my usual ranting and observations. Please let me know your thoughts on these comment sections: if I’m over the top at times, or if you agree with the subject at hand. Also, if you have an interesting case you would like to be considered for this column, e-mail me at jhbrien@aol.com. A publishable picture of the patient, radiograph or Gram stain is required. If you have a case, but you’re not sure, just drop me a line.