|Home | About | Journals | Submit | Contact Us | Français|
Osteomyelitis of the odontoid process is a rare disease and Staphylococcus aureus is the usual causative organism. We present a case caused by a synergistic infection with Staphylococcus aureus and Proteus mirabilis, which has not been reported previously. A high index of suspicion is required for diagnosis, and early institution of broad antimicrobial therapy, including agents active against gram–negative organisms, is mandatory.
Osteomyelitis of the odontoid process is a rare disease that is often difficult to diagnose. Delay in diagnosis can be catastrophic with development of an epidural abscess and consequent death or major neurological disability. The initial presenting symptoms can be nonspecific and definitive neurological signs develop relatively late in the disease. Therefore a high index of clinical suspicion must be exercised. Appropriate neuroimaging combined with the isolation of blood–borne organisms will help clinch the diagnosis.1, 2, 3 Early aggressive treatment is associated with a good outcome.2
A 65–year–old man on regular hemodialysis for long–standing chronic renal failure sought treatment after 2 days of progressively worsening neck pain. The patient described a febrile illness in the days preceding his admission. Examination of his peripheral nervous system was normal other than that he exhibited Lhermitte's sign.4 Based on this history, spinal sepsis was suspected. Blood was drawn for culture and sensitivity and urgent magnetic resonance imaging (MRI) was arranged. Gram–positive cocci were detected in the blood, and empirical therapy with vancomycin, flucloxacillin, and rifampicin was started. MRI of the craniovertebral junction showed an epidural mass around the odontoid process causing cervicomedullary compression (Fig. 1A) and an abnormal hypointensity in the area on the right side of C1 (Fig. 1B). Subsequently, the bacteremia was confirmed to be due to Staphylococcus aureus. Day 3 of admission, the patient developed signs of respiratory distress and carbon dioxide retention. His lower cranial nerve function remained normal, but he exhibited upper motor neuron signs in both lower extremities. This clinical picture suggested worsening cervical epidural compression, and an emergency decompression of the foramen magnum and a C1 laminectomy were performed. At surgery the bone of the first cervical vertebra (Fig. 1B) was found to be soft, and a small amount of epidural pus was obtained. Postoperatively, he was immobilized in a halo head–fixation device. Proteus mirabilis was cultured from both specimens. The patient was then administered gentamicin and ciprofloxacin. Long–term venous access was established, and his response to treatment was monitored by regular measurement of blood inflammatory indices.
Immediately after surgery his neurological signs resolved and his ventilation returned to normal. Over the next few weeks the patient's neck pain resolved, and he had no further febrile episodes. Once inflammatory indices had normalized and he had completed 2 months of antimicrobial therapy, the patient underwent formal occipitocervical fusion with a Ransford loop5, 6 and iliac crest bone graft. Early removal of the halo fixation device was associated with recurrent neck pain and radiological evidence of atlanto–occipital subluxation. Therefore, the halo was replaced until a complete fusion developed 4 months after surgery. In addition to showing bony destruction of the odontoid and right side of C1 (Fig. 2A), computerized tomography (CT) of the craniovertebral junction performed during this period showed that the occipital bone had become involved with the infection (Fig. 2B).
Osteomyelitis of the odontoid process is very rare and therefore requires a high degree of clinical suspicion to diagnose.1, 7 CT, MRI, and white blood cell nuclear scintigraphy may improve diagnostic accuracy.3 Of these diagnostic modalities, MRI is the investigation of first choice.1, 2, 3 Appropriate neuroimaging helps ensure an early diagnosis. Rigid occipitocervical immobilization, high–dose broad–spectrum antibiotics, and surgical stabilization in selected cases form the mainstay of management.2, 3, 8, 9
Our patient represented a number of challenges. His initial antimicrobial therapy was directed toward Staphylococcus aureus isolated from blood cultures. Antibiotics combined with occipitocervical immobilization have been reported to provide satisfactory outcomes without recourse to surgical decompression and fixation.1, 10 However, our patient's clinical condition deteriorated, and synergistic infection with Proteus mirabilis was detected only after surgical specimens were obtained from the craniocervical junction. How the disease might have evolved if antibiotics active against Proteus mirabilis had been included in the initial empirical regime is impossible to say.
Proteus mirabilis is a gram–negative rod. It is motile and produces urease-factors that make it virulent. It is a common uropathogen in individuals with structural abnormalities of the urinary tract.11 It mainly causes cystitis and pyelonephritis, but it can also cause pneumonia, wound infections, and septicemia. It is diagnosed by growth on selected media and is treated with aminoglycosides (gentamicin, amikacin).
Spinal sepsis frequently coexists with other major illnesses or immunocompromised conditions.12 Consequently, such patients are prone to polymicrobial infection. Our case illustrates the importance of obtaining tissue or pus to guide appropriate antimicrobial therapy.
Our patient suffered rapid onset of acute respiratory failure and required emergency decompression late at night. The expeditious surgery likely contributed to his good outcome, a point that has been made previously in relation to epidural sepsis.13 Transoral decompressive surgery is well described for the treatment of this disease and is both safe and effective.9, 14 However, it does require particular surgical and anesthetic expertise as well as specialized instrumentation.15, 16 Transoral surgery is rarely performed at our institution; therefore, it seemed prudent to use the more familiar posterior approach. Although the bulk of the inflammatory tissue lay anterior to the cervicomedullary junction, it was still possible to obtain diagnostic tissue samples and to effect decompression. We suggest that the posterior approach is a safe alternative in an emergency for surgeons less familiar with the transoral approach.
In addition to the destructive effects of osteomyelitis, our patient was subject to a degree of renal bone disease.17, 18 Its presence was reflected by the prolonged period of external fixation required to achieve atlanto–occipital fusion.
In a patient with neck pain and fever with or without neurological symptoms and signs, the possibility of spinal sepsis should be considered, particularly in immunocompromised patients. Early MRI shows the characteristic appearance of this potentially life–threatening condition.1, 2, 3 Empirical antibiotic therapy should cover against gram–positive and gram–negative organisms and probably anaerobes, pending the availability of definite bacteriological sensitivities. Posterior surgical decompression of the cervicomedullary junction is a safe and effective alternative to the transoral route. Early appropriate treatment is associated with a good prognosis.2
Haridas and associates describe their management of a patient with osteomyelitis of the odontoid process with an epidural abscess causing myelopathic symptoms and respiratory distress from neural compression. Initially, the patient was treated with halo immobilization followed by delayed emergent posterior decompression with craniocervical fusion after neurological decline. Of interest is that preoperative blood cultures revealed Staphylococcus aureus whereas intraoperative cultures from the epidural abscess revealed Proteus mirabilis. The patient recovered neurologically and was left in halo immobilization. Some evidence of osteomyelitis of the occipital bone was seen on follow-up computed tomography. The discrepancy between the bacteria cultured from the blood and from the abscess stresses the importance of obtaining cultures from the involved bony territory.
The authors point out that a transoral approach for decompression of the spinal cord in this scenario also would have been possible. However, the posterior approach was favored because it was familiar. Given a rapidly declining patient suffering from an epidural abscess of the odontoid process, this response was appropriate.We would favor the transoral approach because it allows both decompression and extensive debridement of involved bone and resection of the infected odontoid process. As demonstrated in Figure 2B, the bony involvement had extended to the occipital bone. This finding likely reflected insufficient local debridement of osteomyelitic bone, which can occur if treatment with antibiotics alone is insufficient. Whether treatment was planned for the osteomyelitic changes on follow-up is unclear.
Another issue is the timing of the initial decompressive surgery. At presentation the patient had no focal neurological symptoms but displayed a positive Lhermitte's sign. This finding in conjunction with magnetic resonance imaging evidence of epidural compression at the dens suggestive of abscess would have been a sufficient indication to proceed with decompression, debridement, and local cultures combined with antibiotic treatment and immobilization rather than pursuing conservative treatment initially.