Introduction

The incidence of pyogenic spondylodiskitis has increased in recent decades. Whereas tuberculosis was the most important cause of spine infections during the 20th century, pyogenic spondylodiskitis is now the most common cause of primary spine infections. ¹ , ² Aging of the population, increased numbers of immunocompromised patients, and improved diagnostic possibilities are responsible for this increase. ¹ , ² As described in Chapter 5, pyogenic spondylodiskitis is one of the diagnostic challenges in modern medicine. Spondylodiskitis is a unique disease in which a diagnostic delay of up to several months is considered to be a key characteristic of the clinical presentation. ³ , ⁴ Besides the diagnostic challenges, the treatment of spondylodiskitis can be equally complex and often requires a multidisciplinary approach. Although most patients can be treated conservatively, it is important to recognize the cases in which surgical treatment is more appropriate. This chapter provides a systematic and evidence-based algorithm for the diagnosis and treatment of pyogenic spondylodiskitis.

Differential Diagnosis

Pyogenic spondylodiskitis can be difficult to diagnose and may initially be misdiagnosed due to the low frequency of the disease. Because it is often a complication of a distant infection causing bacteremia, the relatively nonspecific array of symptoms of spondylodiskitis may be initially dominated by the primary infection. Consequently, a considerable delay between disease onset and diagnosis may develop. Clinicians should suspect spondylodiskitis in patients with new or worsening back or neck pain accompanied by fever, neurologic symptoms, a recent bacteremia, endocarditis, or other infection symptoms. Pain caused by spondylodiskitis may be misdiagnosed as a degenerative process, back strain, osteoporotic fracture, herniated disk, spinal stenosis, inflammatory conditions, or metastatic disease. Initial misdiagnosis is common in elderly patients, especially in the absence of fever. Clinicians should maintain a high index of suspicion in patients with back and neck pain given the nonspecific array of symptoms.

Hematologic and Biochemical Markers

In general, leukocytosis, neutrophilia, and elevated erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) suggest a pyogenic infection. Although ESR and CRP are sensitive inflammatory markers for infection, and are elevated in the majority of the spondylodiskitis cases, they lack specificity. CRP is a more sensitive parameter than ESR, and it provides a more accurate reflection of the response to the treatment. The leukocyte count is the least useful among the inflammatory markers and is elevated in only one third to one half of the affected patients. Approximately 70% of the spondylodiskitis patients may be anemic, and about half have a raised alkaline phosphatase serum value. Apart from the inflammatory markers, serological tests such as interferon-γ release assay and seroagglutination assay may be performed in suspected tuberculosis and brucellosis infections.

Imaging

The first diagnostic procedure in patients with back or neck pain is often conventional radiography (CR). However, the sensitivity and specificity of radiography for diagnosing spondylodiskitis is low. In the initial phase, there are usually no CR changes. The earliest radiographic signs of spondylodiskitis can develop 2 to 8 weeks after the initial symptoms and may consist of narrowing of the disk space, blurred outlines of the vertebral end plates, loss of end-plate margins, and osteolysis. At later stages, progression of the infection can reveal further destruction of the vertebral body, reactive changes with new bone formation, and sclerosis. Although upright radiographs can be negative in the early phase, the initial CR is useful as a baseline image to detect deformities or ankylosis of the affected levels and to refer to for comparisons during follow-up ( Fig. 6.1 ).

Magnetic resonance imaging (MRI) is the diagnostic procedure of choice and is recommended in all patients with suspected spondylodiskitis. MRI has a high sensitivity (96%) and specificity (92%). ⁵ The MRI protocol can consist of several standard sequences including spin echo (T1- and T2-weighted), gradient echo (T2-weighted), fat signal suppression, short tau inversion recovery (STIR), and contrast-enhanced images. Signs of infection consist of areas of low signal within the vertebral body, loss of end-plate margins, and disruption of cortical continuity on T1-weighted images. The inflammatory tissue generates a strong signal and a hyperintense disk on T2-weighted images in contrast to degenerative changes, where the intervertebral disk is usually hypointense due to dehydration. Fluid-sensitive MRI sequences such as STIR and fat-saturated T2-weighted images are highly sensitive for detecting early inflammatory changes. ⁶ Furthermore, additional pre- and postcontrast fat-suppressed T1-weighted images enable differentiating between vascularized (hyperemic osseous and soft tissue) and nonvascularized, necrotic (abscesses, bone sequesters) inflammatory components. Despite the high sensitivity, MRI findings in early spondylodiskitis may be atypical and overlap with findings in other conditions such as Modic type 1 changes, Schmorl’s node, disk extrusion, malignancy, and acute end-plate fractures. Therefore, it remains important to correlate these early MRI findings with clinical history and laboratory tests ( Fig. 6.2 ).

When it is not possible to request an MRI because the patient has implantable cardiac devices, cochlear implants, or claustrophobia, fluorine-18 fluorodeoxyglucose positron emission tomography (FDG-PET) can be an excellent alternative diagnostic procedure. Because there is hardly any physiological uptake of FDG in the spinal column, the increased glucose metabolism in activated inflammatory cells are imaged as a “hot spot” in the PET scan. Furthermore, FDG-PET can distinguish between initial spondylodiskitis and common degenerative changes in the adjacent bone. Overall, FDG-PET has a very high sensitivity and specificity in diagnosing spondylodiskitis and also facilitates assessment of abscesses and the extent of the infection by combining it with computed tomography (CT). However, tumors also have an increased uptake of FDG, and therefore differentiation between malignant processes and spondylodiskitis may be difficult.

When MRI and FDG-PET are unavailable or impossible, other imaging modalities such as CT or gallium-67/technetium-99 bone scans can be requested. Although CT’s sensitivity and specificity are inferior to those of MRI in detecting spondylodiskitis, CT can provide details on disease progression, bone destruction, sequestration of the end plate, cortical erosions, and sclerosis. CT imaging of nervous structures and soft tissues may be limited. However, contrast enhancement can provide good imaging of paravertebral abscesses, and CT guidance for percutaneous diagnostic needle biopsy and drainage of abscesses can help identify the causative microorganism. Nuclear medicine techniques such as gallium-67 or technetium-99 bone scan or single photon emission computed tomography (SPECT) also have a high sensitivity in diagnosing spondylodiskitis. The inflammatory changes or increased bone remodeling in spondylodiskitis results in spots with increased radiotracer accumulation in the spine. Due to this high sensitivity, a normal bone scan provides reliable evidence for the absence of inflammation. However, degenerative changes or tumors can produce false-positive results. Consequently, the specificity of these techniques is low, and increased uptake can be only suggestive for spondylodiskitis. Combining these techniques with CT imaging results in a higher specificity.

In summary, several imaging techniques can be used to diagnose spondylodiskitis, but each of these techniques has limitations. Whereas MRI remains the preferred diagnostic procedure, alternative methods such as FDG-PET, CT, gallium-67 scans, technetium-99 scans, and CR can be used when MRI is not possible or unavailable ( Fig. 6.3 ).

Cultures

Identification of the responsible microorganism and determination of its susceptibility for antibiotic treatment are essential. Unless clinical circumstances such as severe sepsis dictate otherwise, empiric antibiotic therapy should not be started before every effort is made to obtain adequate cultures. ³ It is recommended to obtain both aerobic and anaerobic blood cultures. ³ In a systematic review, blood cultures resulted in the isolation of the pathogen in 30 to 78% of the patients. ⁷ It has been suggested that blood cultures directly following diskovertebral biopsy may result in a higher yield of positive cultures. ⁸

Another way to identify the microorganism is to perform CT-guided fine-needle biopsies and percutaneous or open surgical biopsies. When blood cultures are negative and clinical circumstances allow postponing antibiotic treatment, local sampling is recommended. ³ Although CT-guided biopsies are the least invasive and the preferred initial method for local sampling, they provide a relatively low quantity of tissue, which can lead to negative results. So far, isolation of the microorganism from biopsies ranges from 30 to 91% in the literature. ⁹ – ¹² Therefore, it is recommended that a second biopsy be performed in patients with suspected spondylodiskitis with negative blood cultures and the original image-guided biopsy. ³ These second biopsies should also be tested for more difficult to grow microorganisms. Extended culture times can help to detect low-virulence organisms. Molecular methods such as 16S rRNA polymerase chain reaction (PCR) and sequencing can also be helpful for diagnosing spondylodiskitis in suspected cases and for determining the pathogen. ¹³ Percutaneous endoscopic diskectomy or open excisional biopsy can be considered because culture positivity is higher with surgical sampling. The diagnostic yield can be further improved by obtaining more than one specimen for culture and histopathological investigation. If antibiotic treatment was already started and clinical improvement stagnates, discontinuation of the antibiotics for some days is recommended before the biopsies are taken. When antibiotic treatment fails, clinicians should also be aware of a possible superinfection by another microorganism. Therefore, it is recommended to obtain new cultures in case of a relapse of pyogenic spondylodiskitis, including cultures for tuberculosis.