CHAPTER SYNOPSIS:

Postoperative spinal infections require early recognition and aggressive management for successful treatment. Traditional treatment methods included aggressive irrigation and debridement with delayed primary or secondary closure and multiple dressing changes. The goal of debridement should be the removal of all necrotic tissue and should proceed from the superficial to deep layer of the wound. In the presence of instrumentation, the fusion mass should be explored. In the event that the fusion is solid, then the instrumentation can be removed; however, if the fusion is not solid, and removal of the instrumentation places the spine at risk for instability, then the instrumentation should be retained.

IMPORTANT POINTS:

• •
  

  Generally, the infection can be eradicated whereas salvaging the implants, although this partly depends on the type of bacteria responsible for the infection.

  
  
• •
  

  The advent of vacuum-assisted wound closure systems provides an alternative to traditional wet-to-dry dressing changes until wound closure.

CLINICAL/SURGICAL PEARLS:

• •
  

  A high index of suspicion should be maintained for infection in the presence of any wound drainage.

  
  
• •
  

  Aggressive irrigation and debridement with delayed primary or secondary closure is the treatment of choice.

  
  
• •
  

  If the wound is closed after debridement, consideration should be given to closure over drains.

  
  
• •
  

  Unless the instrumentation is obviously loose, it should be left in place if its removal could lead to spinal instability.

  
  
• •
  

  Antibiotic treatment should be dependent on the results of culture and sensitivity, and is generally continued for 6 weeks.

CLINICAL/SURGICAL PITFALLS:

• •
  

  Nonoperative management of wound infections is rarely indicated and should be reserved for those patients with severe medical co-morbidities with significantly increased risk from additional surgery.

  
  
• •
  

  With surgical debridement, care must be taken to remove all necrotic tissue from the superficial to the deep layer of the wound.

  
  
• •
  

  The number of necessary debridements may be greater in cases of retained instrumentation.

VIDEO AVAILABLE:

N/A

RISK FACTORS AND PREVENTION

Reported rates of infection range from 1% to upward of 15% from some individual institutions. Taking into account the type of procedure can segregate infection rates further, the Centers for Disease Control and Prevention identify infection rates as low as 1.25% after laminectomy and up to 2.1% after arthrodesis; however, adding instrumentation, and more likely greater operative time and blood loss can lead to greater rates.

A retrospective review of 1629 procedures reported that a postoperative infection developed in 48 patients (4.4%). Most infections occurred during the early postoperative period (less than 3 months). Age older than 60 years, smoking, diabetes, previous surgical infection, increased body mass index, and alcohol abuse were statistically significant preoperative risk factors. A similar type of review of neurosurgical procedures (similar overlap in pathology and cases) further identified postoperative incontinence, a posterior approach, tumor resection surgery and morbid obesity as independent risk factors.

The most recent retrospective case–control review of 2316 solely orthopedic procedures from 1998 to 2002 at a tertiary referral care center demonstrated an incidence rate of 2.2%. A greater number of deep infections (43%, 20 cases) than superficial incisional infections (39%) was reported. Consistent with previous reports, the median amount of time until infection was 11 days. Diabetes, specifically an increased preoperative or postoperative blood glucose level, was the greatest independently associated risk factor. Other independent risk factors included suboptimal timing of preoperative antibiotics, and two or more surgical residents participating in the case.

These factors also seem equitable to the increased risk seen when treating vertebral fractures with stabilization procedures. The average incidence of postoperative SSI in patients with polytrauma treated with posterior surgery for thoracolumbar spine fractures at Shock Trauma Center at the University of Maryland (Baltimore City, MD) in the fiscal year 2006 was 13 in 166 cases (7.8%), which was almost 3 times the rate for elective cases during the same period.

Extrapolating the work from the total joint literature helps elucidate the cause for postoperative wound infections in the spine. Similar to spinal wounds, total hip arthroplasty (THA) wounds are horizontally positioned, open to the surrounding environment and surrounded typically by more than one surgeon working over wound edges for a prolonged period. Knobben et al. published the results correlating the results of intraoperative cultures from bone particles during preparation of the acetabulum and femur in THA with postoperative infection cultures. They demonstrate a positive predictive value between 81% and 90%, and show that despite lamellar airflow rooms, space suits, and other external measures to control contamination, THA postoperative infections stem from intraoperative contamination of the wound.

Once bacteria contaminate wound edges, the adjacent postoperative hematoma serves as an environment conducive for growth and free from perioperative systemic antibiotics unable to reach the devascularized “dead space.” Bacteria can readily adhere to local spinal instrumentation forming biofilms difficult to eradicate by antibiotics with proven efficacy against the same pathogens in vitro.

In our opinion, these data suggest that attempting to control the local milieu before closure of the incision may be an additional technique for reducing or preventing postoperative wound infections of the lumbar spine. Clearly, any means of increasing local antibiotic concentration to bactericidal levels should help sterilize the surgical wound perioperatively. Conversely, toxic antibiotic levels should be avoided as one study demonstrated that gentamicin at levels greater than 100 μg/mL can lead to reduced osteoblast activity (clinical studies with gentamicin-impregnated polymethylmethacrylate beads showed concentrations up to 345 μg/mL).

DIAGNOSIS

Patients typically have signs and symptoms of an infection at a mean of 15 days from the index procedure, and the most common presenting symptom (93%) is wound drainage. Less than one third of the patients had a temperature greater than 37.5° C at the initial examination.

Because of the lack of pathognomonic symptoms or signs, laboratory studies are a useful adjunct to forming the correct diagnosis. A recent prospective observational study of 149 patients in whom 20 patients experienced development of infectious complications showed a postoperative peak in C-reactive protein in 78% of patients compared with only 48% of patients with a similar peak in erythrocyte sedimentation rate.

Plain radiography can assist diagnosis through indirect signs of postoperative infection such as implant loosening or large amounts of soft tissue swelling. Computed tomography and advanced imaging modalities such as magnetic resonance imaging can detect fluid collections; however, they cannot differentiate between sterile or infected collections.

The addition of gadolinium contrast can assist the magnetic resonance imaging diagnosis if rim enhancement of an abscess is present. However, in the frequent setting of spinal instrumentation, these advance imaging modalities are hampered by artifact from the implants. Therefore, diagnostic imaging can only augment clinical diagnosis of infection without definitive identification.

MICROBIOLOGY

Weinstein et al.’s retrospective review of postoperative wound infections in a single surgeon’s 9-year experience demonstrated that 63% (29/46) of patients’ intraoperative cultures grew Staphylococcus aureus alone. The remaining patients’ cultures grew different single organisms or multiple organisms.

TREATMENT

The treatment of a postoperative SSI in the lumbar spine focuses on eradication of the infection, whereas maintaining spinal stability and preserving neurologic function. Rarely is nonoperative treatment of postoperative spinal wound infections indicated, it is reserved only for severe medical contraindications.

Notably, no prospective, randomized, clinical trial has ever been performed that evaluates different treatment regimens, and the available literature on this topic consists of retrospective reviews of single-treatment regimens.

Traditional treatment methods included aggressive irrigation and debridement with delayed primary or secondary closure and multiple dressing changes. The goal of debridement should be the removal of all necrotic and devitalized tissue, and should proceed from the superficial to deep layer of the wound.

Ideally, the infection can be eradicated whereas salvaging the implants, although this partly depends on the type of bacteria responsible for the infection. Slime-forming bacteria such as coagulase-negative staphylococci are able to embed themselves in the protective biofilm and have been implicated as the cause of persistent infections in at least 50% of patients with retained instrumentation with infections. Others have reported that, in patients with retained instrumentation, the average number of irrigation and debridements approaches five per patient. Picada et al. performed a retrospective study of 817 patients who underwent instrumented lumbosacral posterior fusion with 26 patients acquiring infection, and 24 of these 26 patients treated successfully with sequential debridements without removal of implants.

The clinical appearance of the patient and wound together with the causative organism can determine the need for repeat debridement. Weinstein et al. had success treating 46 postoperative spinal infections in 2391 patients with debridement and irrigation with 9 L bacitracin-saline solution. Superficial infections were closed over drains, but all deep infections were packed with gauze and left open, followed by repeat debridement and irrigation after 48 hours. Viable-appearing bone graft material was left in place, and instrumentation was not removed. Deep infections were treated with a 6-week course of intravenous antibiotic agents. All wounds healed without further sequelae.

Bone morphogenetic protein (BMP) is a naturally occurring protein that has multiple roles but has been demonstrated to be important in the development of bone. Clinically, recombinant forms of BMP (rhBMP) have been extensively investigated, and found to be safe and effective for use in spinal fusions. More recently, some interest has been in the use of rhBMP as an adjuvant for treatment of spinal infections.

In 2007, Allen et al. reported on a retrospective case series of 14 patients with medically nonresponsive pyogenic vertebral osteomyelitis that was treated by anterior and posterior debridement and instrumented fusion together with rhBMP-2. The authors conclude that, in combination with antibiotics and circumferential instrumented fusion, rhBMP-2 appeared to be safe and effective for the surgical treatment of medically nonresponsive pyogenic vertebral osteomyelitis (Allen, 2007). Similarly, in 2008, O’Shaughnessy and colleagues retrospectively reviewed their institution’s experience with the treatment of 20 patients with vertebral osteomyelitis who underwent surgical debridement and the use of rhBMP-2 as the primary graft material for spinal fusion. In this series, the authors found that all 20 patients demonstrated clinical and radiographic evidence of spinal fusion at the time of follow-up.

These findings are particularly interesting in light of the fact that the original indications listed spinal infections as a contraindication for use of rhBMP-2. However, as more and more preclinical and clinical studies become available that describe the safety and effectiveness of BMPs in the treatment either alone or as an adjuvant to traditional management of spinal infections, its use will likely become more common. However, until that time, it should be considered experimental and an off-label use of rhBMP-2.