Spinal Infections: Vertebral Osteomyelitis and Spinal Epidural Abscess

Published on 13/03/2015 by admin

Filed under Neurosurgery

Last modified 13/03/2015

Print this page

This article have been viewed 2138 times

Chapter 145 Spinal Infections

Vertebral Osteomyelitis and Spinal Epidural Abscess

Sunil Jeswani, Doniel Drazin, Ali Shirzadi, Paula Eboli, Debraj Mukherjee, Frank L. Acosta

Introduction

The surgeon is frequently called upon to treat vertebral osteomyelitis (VO) and spinal epidural abscess (SEA). To obtain the best patient outcomes, these spinal infections require prompt diagnosis and appropriate interventions.

This chapter reviews the epidemiology, pathogenesis, pathophysiology, and microbiology of these spinal infections. Risk factors are identified and various VO and SEA clinical presentations and laboratory findings are discussed. A review of current, relevant radiographic studies and diagnosis is presented. This chapter also discusses the current thinking regarding the management and treatment of spinal abscess, including antibiotic therapy, the role of immobilization, and the recommended indications for surgery versus nonoperative management. Mortality and outcome statistics indicate that the ability of practitioners to promptly identify and treat VO and SEA reduces patients’ mortality and permanent neurologic deficits.

Epidemiology of Spinal Infection

During recent decades, there has been a steady rise in the incidence of VO and SEA, traditionally felt to be relatively infrequent infections of the vertebral column. Although older series reported rates of VO and SEA of 0.2 to 2 per 10,000 hospitalizations, researchers in more recent series in both Europe and the United States have reported a 10- to 15-fold increase in the incidence of VO and SEA at large tertiary care centers.¹^–⁶

Additionally, though older series reported a biphasic distribution in the age-related incidence of VO and SEA, with incidence peaking in the first and fifth to sixth decades of life, newer series have shown a rising incidence of these infections in adult patients with impaired host defenses, frequent bacteremia, or both.⁷^–¹⁰ In particular, in recent population-based reports, patients with AIDS and intravenous drug users have been shown to have exceedingly high relative rates of VO or SEA.¹¹^–¹³ Patients with diabetes, liver cirrhosis, renal disease on dialysis, and patients on chronic immunosuppression have all shown a proclivity toward the development of spinal infections.¹⁴ The widespread use of MRI technology, coupled with the greater concentration of these patients at tertiary care centers, may account for at least a portion of the greater relative incidence of these spinal infections documented in recent years.

Pathogenesis and Pathophysiology

The primary pathogenesis of VO varies by patient age. In children, the intervertebral disc receives a direct blood supply through collateral vessels associated with the vertebral end plates. Thus, systemic bacteria are able to relatively easily affect and penetrate the vertebral body. By the third decade of life, the collateral vessels present in childhood often involute, such that there is no direct vascular connection between the systemic circulation and the vertebral body. Therefore, in adults, transient bacteremia seeding relatively avascular regions adjacent to subchondral bone, as well as urinary sepsis, become the more prominent means of initiating infection. As a result, spinal infection in adults more often involves the subchondral bone, with secondary involvement of the vertebral end plates and a relative ease of spread between adjacent vertebral bodies.

The pathogenesis of SEA has traditionally been thought to develop through direct extension from a VO primary site, by hematologic seeding, or from direct contamination, such as through surgical intervention. Most cases of SEA develop dorsally, as the spinal epidural space contains relatively loose areolar tissue and a rich venous supply; these factors make the dorsal region more prone to infection, at least to the S2 vertebral level, beyond which dura surrounds the epidural space circumferentially.

Risk Factors for Vertebral Osteomyelitis and Spinal Epidural Abscess

Several risk factors have been associated with osteomyelitis and SEAs. Many of these risk factors are associated with an immunocompromised state and appear to lead to the development of the infection. Diabetes, for example, lowers the immune response and is associated with spinal infections. Studies have shown diabetes to be associated with SEAs in 18% to 54% of cases ³ and 31% of VO cases.¹⁵ IV drug abuse has also been associated with the development of spinal infections. This may be secondary to the sharing of unsterile needles, leading to systemic infections and endocarditis, which then predispose individuals to spinal infection. There also seems to be an association between alcoholism and spinal infections. This association may have multifactorial causes, including diet and hygiene and increased predisposition to spinal trauma.¹⁶ Other associated risk factors that decrease immune system function include malignancy and long-term steroid use.

Since one of the major mechanisms of the development of VO and SEA include hematogenous dissemination of an infection from a remote source, infections and bacteria originating from other parts of the body serve as risk factors for spinal infections. Therefore, spinal infections have been associated with other infections, including bacterial endocarditis, urinary tract infections, skin abscesses, pneumonia, dental infections, and sepsis. In several studies, skin and soft-tissue infections were found to be the most common cause of SEAs via hematogenous spread.^11,¹⁴ In VO, however, the most common source of infection seems to be from the urinary tract.¹⁷ Hematogenous seeding of spinal epidural infections account for approximately 50% of cases.^18,¹⁹

In addition to hematogenous dissemination, SEAs may develop by direct extension from an area of VO. In such cases, the abscess tends to be located in the ventral epidural space. Several studies have shown this association of SEA with VO, especially in ventrally located abscesses. Infections extending to the spinal epidural space can also directly spread from a psoas muscle abscess. This route of spread is responsible for approximately 10% to 30% of cases.¹⁹

Spinal infections may also develop from direct seeding of bacteria via an invasive procedure. Open spinal surgery, including laminectomies, discectomies, and operations involving spinal instrumentation, as well as less invasive procedures, such as lumbar punctures and epidural injections, have been shown to be risk factors for the development of spinal infections. Interestingly, the rate of infection due to less invasive procedures seems to be larger than that due to open spinal surgeries. Researchers in recent studies have found the risk of developing a SEA after epidural catheterization to be from 0.04% to 0.07%.¹⁹ Moreover, leaving catheters and drains in place at the time of surgery seems to increase the risk of acquiring a spinal infection.^18,¹⁹ Trauma to the spine has also been associated with the development of SEA and VO. Trauma may predispose patients to spinal infections because of disruption of anatomic barriers.¹⁸

Microbiology of Vertebral Osteomyelitis and Spinal Epidural Abscess

Bacterial pathogens remain the primary infectious agents responsible for most cases of VO and SEA. Fungal agents occasionally cause disease, particularly in immunosuppressed patients, and some parasitic agents have been associated with VO and SEA in certain endemic geographic regions.

Among bacterial pathogens, Staphylococcus aureus, most often drawn from a primary skin or soft-tissue infection, is by far the most common etiologic agent responsible for VO and SEA. S. aureus accounts for more than one half of all cases of VO and more than three fourths of all SEA cases. Gram-negative bacteria-based infections are increasing as well, with Escherichia coli being responsible for more than one half of all Gram-negative bacteria-based infections, followed by Enterobacter and Proteus infections. Pseudomonal infections are particularly prominent in intravenous drug-using patients. Tuberculosis remains a relatively rare cause of infection in developed countries, but is often the most common source of VO in many endemic regions.

Clinical Presentation

The classic triad of symptoms of SEA is fever, back pain, and neurologic dysfunction. Investigators in some studies, however, have reported that this triad is present in only 10% to 15% of patients at the time of initial presentation.^18,¹⁹ Important features of the clinical presentation of SEA are that the symptoms progress through several stages and that this progression can occur rapidly. A classic paper by Heusner¹ described the stages of clinical SEA presentation:

• Stage I: Back pain and tenderness

• Stage II: Radicular pain

• Stage III: Weakness and sensory loss

• Stage IV: Paralysis

The course of time through which these stages progress is highly variable. A patient may present with symptoms that have been ongoing from days to weeks.¹⁹ An abscess caused by hematogenous spreading tends to have a clinical presentation that progresses more rapidly through these stages as compared to an abscess formed secondarily to local extension. Once neurologic dysfunction is present, however, the average time to complete paralysis is 24 hours and often can occur in even less time. Thus, the onset of neurologic dysfunction should prompt the need for emergent surgical decompression to prevent progression to paralysis.

Localized pain and tenderness over the involved area of the spine is found very commonly as the initial presentation of SEA (71% in one large meta-analysis).¹¹ Oftentimes pain is described as sharp and stabbing¹⁹ This pain reflects the level of the spine involved. The pain may occur in a dermatomal distribution in the lower extremities, for example, or a patient may experience pain in the chest wall or abdomen which may mimic other conditions such as pancreatitis.¹⁹ Neurologic dysfunction secondary to a SEA may include motor weakness, including sphincter dysfunction and sensory loss. Motor deficits were shown to be present in 54% of cases in one study.²⁰ A large meta-analysis showed paralysis affects approximately 34% of patients with SEA.¹¹

Fever is also a common manifestation of SEA, with an incidence between 66% and 75%.¹⁶ Symptoms and signs of meningitis may also be present, including headaches and nuchal rigidity. This may represent a parameningeal reaction. In patients with a chronic infection course, symptoms such as unintended weight loss and malaise can occur. Sometimes the clinical presentation of SEA may be dominated by that of sepsis, making the detection of a SEA more obscure. In children, the symptoms are more likely to be nonspecific, and the child may simply complain of not feeling well.^18,²¹

The clinical manifestations of VO are similar to those of SEA in that localized pain exacerbated by motion and tenderness almost always occurs and is very frequently the initial symptom. Often-times the pain may be associated with paraspinal muscle spasm. In one review, the authors reported a 92% incidence of back pain in VO.¹⁷ Radicular pain in VO has been reported with incidences of up to 25-33%.¹⁵ Neurologic deficits can occur with VO and are due to either associated SEA or spinal instability. Investigators in one study reported neurologic impairment in 28% of patients with VO.²² Neurologic involvement seems to be more common in cases of VO affecting the cervical or thoracic spine as compared to the lumbar spine.²³ Fever can also be present in VO; however, this presentation is not as consistent a finding as back pain and tenderness.

Laboratory Findings

Laboratory findings associated with VO and SEA reflect the underlying inflammatory nature of the disease and are not specific. The erythrocyte sedimentation rate (ESR) is the most consistently elevated laboratory value in spinal infection, with the ESR typically being greater than 40. In one meta-analysis, the average ESR was 77.^2,⁸ ESR has been shown to be greater than 20 in 95% of cases in some studies.^3,¹⁹ The ESR is also helpful in monitoring the disease process during therapy, with declining values of ESR to be expected with appropriate therapy. The C-reactive protein (CRP), an acute phase reactant, has also been found to be elevated in patients with spinal infections.^2,⁸ Obtaining CRP levels may be useful because CRP levels have been shown to rise more quickly during inflammation and to return to normal more quickly than ESR levels.²⁴ In addition, one study showed CRP values obtained over 2 weeks after surgery for SEAs can provide prognostic information regarding outcomes.²⁰

The peripheral white blood cell count may be elevated in patients with spinal infections, but many times may be normal. An elevation of the peripheral white blood cell count may be more indicative of a systemic infectious process. The average peripheral white blood cell count in one large meta-analysis of SEAs was 15,700/μL.¹¹

In patients with SEA, a cerebrospinal fluid (CSF) sample may show an elevated white blood cell count and protein level, which are indicative of a parameningeal process. In addition, the CSF values may be reflective of concomitant encephalitis or meningitis, which may present along with a spinal infection in a significant proportion of cases.² However, a lumbar puncture is often inadvisable in the setting of a SEA because of the risk of seeding bacteria into the thecal sac, especially in abscesses in the lumbar region that are located in the dorsal epidural space.

Blood cultures are often positive in the setting of a spinal infection. The organism grown on blood culture may be the same as the causative organism in a SEA or VO, especially if the spinal infection was caused by a hematogenous seeding. Blood cultures have been reported to be positive in approximately 50% of cases in some studies.¹⁵ Positive blood cultures can be useful for guiding antibiotic therapy in the setting in which direct cultures from the spinal infection cannot be obtained. Because blood cultures may become negative if obtained after the administration of antibiotics, it is important to obtain blood cultures prior to the initiation of antibiotics if possible.

Radiographic Studies

Imaging studies are essential for the diagnosis of both VO and SEA. Plain film examination of the spine should be performed as part of the initial workup. In cases of SEA, x-rays are usually normal unless they are associated with an adjacent vertebral body’s osteomyelitis.²⁵ In cases of VO, plain radiographs are usually helpful in localizing the spinal level and the extent of the lesion and in assessing the presence of deformity. Unfortunately, early in the course of VO, plain radiographs may be entirely normal and provide little or no clue to the diagnosis. Most of the time, radiographic changes lag well behind the clinical presentation and frequently are not helpful in establishing an early diagnosis. Plain x-ray changes usually take 2 to 8 weeks to develop from the onset of infection. The earliest changes are loss of the cortical end plate margins and loss of the disc’s space height.²⁵

Narrowing of the intervertebral disc space can be seen in plain films as early as 2 weeks from the onset of symptoms but usually appear at week 6 or 8.^15,²⁶^–²⁸ As the infection progresses, there is an increase in vertebral body destruction followed, in some cases, by subsequent spinal deformity. Lytic lesions, demineralization, and scalloping of the end plates may occur.²⁵ As the infection resolves and healing takes place, new bone formation and subsequent fusion can occur, but this may not be evident for as long as 1 year.

For the above-described reasons, computed tomography (CT) has become an important diagnostic tool in the diagnosis of spinal infections, especially for VO. Among its benefits, CT provides excellent detail of the bony anatomy, and it clearly demonstrates the lytic lesions as well as gas within the disc space or soft tissues (Fig. 145-1). Although identification of SEA has been reported using CT alone, it is generally thought that plain CT is not particularly sensitive for visualizing SEA.²⁹ Indirect findings such as intraspinal gas have been described in the presence of SEA.²⁵