NARRATIVE REVIEW
Surgical Site Infections in Spinal Surgery Barrett S. Boody, MD, Tyler J. Jenkins, MD, Sohaib Z. Hashmi, MD, Wellington K. Hsu, MD, Alpesh A. Patel, MD, and Jason W. Savage, MD
Abstract: Surgical site infections (SSIs) are a potentially devastating complication of spine surgery. SSIs are defined by the Centers for Disease Control and Prevention as occurring within 30 days of surgery or within 12 months of placement of foreign bodies, such as spinal instrumentation. SSIs are commonly categorized by the depth of surgical tissue involvement (ie, superficial, deep incisional, or organ and surrounding space). Postoperative infections result in increased costs and postoperative morbidity. Because continued research has improved the evaluation and management of spinal infections, spine surgeons must be aware of these modalities. The controversies in evaluation and management of SSIs in spine surgery will be reviewed. Key Words: spine surgery, infections, SSI (J Spinal Disord Tech 2015;28:352–362)
S
urgical site infections (SSIs) can be a devastating and costly complication after surgery. SSIs are defined by the Centers for Disease Control and Prevention as occurring within 30 days of surgery or within 12 months of placement of foreign bodies, such as spinal instrumentation.1 SSIs are commonly categorized by the depth of surgical tissue involvement (ie, superficial, deep incisional, or organ and surrounding space).2 The total costs of management vary widely, from $30,000 for reoperation and hospitalization related to complicated deep space infections.3,4 On average, postoperative SSIs extend the length of hospital stay by 9.7 days and increase costs by $20,842 per admission.5 DeLissovoy et al reported that SSIs are responsible for nearly $1.6 billion in excess medical expenditures.5 Postoperative SSI rates after spine surgery average 2.1% depending on the diagnosis (1.4% for procedures for degenerative etiology, 4.2% for deformity correction procedures, and 9.4% for surgical management of acute traumatic spinal cord injury).6,7 Surgical, patient, and pathology-related risk factors have been extensively researched due to the significant clinical and financial impact
Received for publication June 24, 2015; accepted October 7, 2015. From the Department of Orthopaedic Surgery, Northwestern Memorial Hospital, Chicago, IL. The authors declare no conflict of interest. Reprints: Barrett Boody, MD, Northwestern Memorial Hospital, Department of Orthopaedic Surgery, 676 N. Saint Clair, Suite 1350, Chicago, IL 60611 (e-mail:
[email protected]). Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
352 | www.jspinaldisorders.com
of SSI. The rate of SSI varies substantially based on patient comorbidities (0.7%–4.3% without risk factors vs. 2.0%– 10.0% with obesity, diabetes, and neuromuscular disease).8 Multiple intraoperative factors such as blood loss >1 L, prior SSI, and increased operative time have also been proven to be independent risk factors for increased SSI rates.9 These risk factors should be considered during the formation of the operative plan, which can help communicate preoperative expectations, and address modifiable factors in the perioperative period to decrease the overall risk of developing an SSI.
PATIENT RISK FACTORS Multiple patient-related factors increase the risk of SSI. Diabetes mellitus (DM) is a prevalent comorbidity in patients undergoing spinal surgery (5.6%–9.7%).10–12 Diabetes exerts a significant effect on wound healing through impaired collagen synthesis, decreased cellular proliferation, altered immune function, and diminished perfusion from macrovascular and microvascular disease.13 Diabetics are at increased risk of SSI, cardiopulmonary complications, increased length of stay, and increased mortality with spinal surgery, with poor glycemic control further increasing complication rates.14,15 Although often found coexisting with obesity, Olsen et al16 found that either preoperative blood glucose >125 mg/dL or postoperative blood glucose >200 mg/dL are independent risk factors for SSI [odds ratio 3.5 (P = 0.004) and 3.3 (P < 0.001), respectively]. Hikata and colleagues demonstrated a significantly higher SSI rate in diabetics compared with nondiabetic control groups (16.7% vs. 3.2%, respectively, P = 0.0005) in thoracolumbar fusions. In addition, 19 patients with hemoglobin A1c levels 7% resulted in a 35.3% SSI rate (6 of 17 patients, P = 0.006). However, they reported that preoperative glucose levels did not correlate significantly with SSI.17 Obesity is also considered to be a modifiable risk factor for SSI18; however, recent data suggest a complex relationship between increasing body mass index (BMI) and increasing SSI rates. Increased skin to lamina distance and increased thickness of subcutaneous fat have been found to significantly increase the risk of developing an SSI for cervical and lumbar spinal fusion surgery, yet BMI and DM were not found to be independent predictors of SSI.19,20 These studies expose the significant impact of body habitus on SSI rates, with increasing thickness of subcutaneous fat likely contributing to J Spinal Disord Tech
Volume 28, Number 10, December 2015
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Spinal Disord Tech
Volume 28, Number 10, December 2015
increased surgical wound dead space and the need for more forceful, prolonged intraoperative retraction potentially causing tissue necrosis.19,20 Conversely, other investigators have suggested a positive correlation between BMI and SSI rates. Rihn and colleagues reported that obesity (BMI > 30) led to a higher SSI rate (5% vs. 1%) with twice the reoperation rate at 4-year follow-up (20% vs. 11%, P = 0.01) with surgical management of degenerative spondylolisthesis. Despite increased SSI and reoperation rates with surgical management, the obese cohort reported similar 4-year subjective clinical outcomes compared with the nonobese cohort.21 In a meta-analysis of 34 articles investigating the effect of BMI on spine SSI, Abdallah et al22 published a significant positive correlation with increasing BMI and SSIs and defined an adjusted SSI risk increase of 21% for every 5-unit increase in BMI. Smoking leads to wound healing detriments through decreased tissue oxygenation and diminished inflammatory responses, and therefore has been identified as a significant risk factor for SSI.23 Durand et al24 reviewed a prospective cohort of 3908 patients undergoing orthopedic procedures with implants, and demonstrated an adjusted relative risk of 2.2 for SSIs in smokers, with the majority of the infections occurring during the index procedure hospital admission. Pahys and colleagues reported smoking as independently associated with increased risk for SSI in posterior spinal surgery. They noted that smokers comprised 78% of SSIs versus 25% of noninfected cases, reporting smoking to have an odds ratio of 2.6 for SSI (P = 0.008).25 Smoking cessation has been shown to decrease the risk of developing an SSI, and should be strongly encouraged before undergoing spinal surgery.26 Adequate preoperative nutrition is crucial for wound and soft tissue healing and affects SSI rates in spinal surgery.27 Klein et al28 reviewed surgical management of vertebral osteomyelitis and spinal cord injuries, noting serum albumin levels 30 [OR 6.76 fusion (95% CI, 2.9–15.7)], diabetes [OR 3.2 (95% CI, 1.2–8.4)], blood loss >500 mL [OR 2.8 (95% CI, 1.02–6.53)] as independent risk factors Laminectomy, Cervical, Multivariate analysis identified presence of prolonged discectomy, thoracolumprocedure duration (defined by >75th percentile of spinal fusion bar, sacral cases, [OR 4.7 (95% CI 1.6–14)], obesity (defined by medical diagnosis, average weight 98 kg, [OR 4.0 (95% CI 1.6–10)], lumbar-sacral level [OR 2.9 (95% CI, 1.2–7.1)], razor shaving of hair [OR 3.6 (95% CI, 1.2–11)] as independent risk factors. Study also reports intraoperative administered fraction of inspired oxygen 35 [OR 5.2 (95% CI, 1.9–14.2)], posterior approach [OR 8.2 (95% CI, 2–33.53)], and tumor resection indication for surgery [OR 6.2 [95% CI, 1.7–22.3)] Laminectomy, Cervical, Independent risks factors identified in this study discectomy, thoracic, included diabetes [OR 3.5 (95% CI, 1.2–10)] subspinal lumbar optimal prophylactic antibiotics timing [OR 3.4 (95% arthrodesis CI, 1.2–10)], preoperative serum glucose >125 or postoperative serum glucose >200 [OR 3.3 (95% CI, 1.4–7.5)], BMI > 30–35 [OR 2.2 (95% CI, 1.1–4.7)]. Study also reports participation of >2 residents is associated with SSI [OR 2.2 (95%CI, 1–4.7)] Various Cervical, Estimated blood loss >1 L, previous history of SSI, and thoracic, diabetes were independent risk factors indentified for lumbar, sacral SSI. Further, diabetes, obesity (BMI > 30), previous history of SSI, and surgical duration of 2–5 h, surgical duration >5 h were found to be independent risk factors for of deep SSI. Obesity was the only independent risk factor for superficial SSI. Spinal deformity Cervical, Obesity was identified as the only independent risk (kyphosis/ thoracic, factor for superficial SSI. Obesity and staged anterior scoliosis) lumbar, sacral and posterior procedures were both independent risk factors for deep SSI. History of previous SSI was reported to be an independent risk factor for SSI. Decompression, Unknown Multivariate analysis identified presence of insulin fusion or dependent diabetes (OR 1.5), current smoking (OR 1.19), operative duration of 3 to 6 h (OR 1.33), and
Laminectomy
Primary Spine Diagnosis/ Procedure
J Spinal Disord Tech
Prospective Obesity/BM Diabetes Veeravaqu et al 752/24, 774 (3.04%)— total SSI 287/24, 774 cohort Duration of 200938 (1.16%)—deep SSI 468/ procedure Smoking
Olsen et al16
Obesity/BMI Diabetes
46/2316 (2%)- total SSI
Retrospective Adult case-control
41/1413 (2.8%)— incisional/organ space SSI
Obesity/BMI Choice of surgical procedure
Olsen et al 200336
Retrospective Adult case-control
Maragakis et al 104/3894 (2.6%)—total SSI 53/104 (51%)— 200935 deep SSI
Obesity/BMI Duration of procedure Intraoperative oxygenation Spinal surgical levels
Obesity/BMI Koutsoumbelis 84/3218 (2.6%)—deep SSI Retrospective Adult case-control Diabetes et al 201134 Intraopertive blood loss
Obesity/BMI Diabetes Spinal surgical levels
Independent Risk Factors/Topics
TABLE 1. Evidentiary Table—Modifiable Risk Factors
Boody et al Volume 28, Number 10, December 2015
2015 Wolters Kluwer Health, Inc. All rights reserved.
Volume 28, Number 10, December 2015
Copyright
r
Diabetes Smoking Schimmel et al History of previous 201040 SSI
Obesity/BM Duration Fang et al of procedure 200539 Diabetes Smoking History of previous SSI
36/1615 (2.2%)—deep SSI Retrospective Adult case-control
Spinal fusion
24, 774 (1.89%)— instrumentasuperficial SSI tion 48/1095 (4.4%)—deep SSI Retrospective Adult Laminectomy, case-control pediatric discectomy, spinal fusion
operative duration >6 h (OR 1.4) as statistically significant risk factors of SSI Cervical, Statistically significant preoperative risk factors included thoracic, age >60 y [OR 2.94 (95% CI, 1.3–6.6)], smoking [OR lumbar, sacral 2.47 (95% CI, 1.1–5.6)], diabetes [OR infinite (95% CI, 1.0–inf)] (no control group patients had diabetes), previous history of SSI [OR 6.64 (95% CI, 1.25–35.1)], increased BMI > 25.5, alcohol abuse [OR 8.55 (95% CI, 0.91–81.5)]. Statistically significant intraoperative risk factors included staged anterior and posterior procedures, estimated blood loss >2.5 L, and operative duration. Lumbar Multivariate analysis identified statistically significant risk factors including diabetes [OR 5.92 (95% CI, 1.23–28.5)], smoking [OR 1.19 (95% CI, 1/02–5.32)], previous SSI [OR 3.7 (95% CI, 1.6–8.6)], and >7 levels fused [OR 6.21 (95% CI, 1.37–28.2)]
J Spinal Disord Tech
2015 Wolters Kluwer Health, Inc. All rights reserved.
Surgical Site Infections in Spinal Surgery
wound before closure has gained recent interest.48–50 Many recent retrospective reviews suggest excellent safety profiles of intrasite VP in spinal surgery. Evaluating 87 pediatric fusion and instrumentation cases, Gans et al51 identified no complications from VP without significant changes in creatinine levels and negligible systemic vancomycin levels. Furthermore, Sweet et al52 reported no wound complications or adverse clinical outcomes in 821 posterior lumbar instrumented fusions. Despite several reviews supporting the safety of intrasite VP, rare complications of anaphylaxis with circulatory collapse and sterile seromas have been reported with intrasite VP.53,54 A recent in vitro study demonstrated a dose-dependent decrease in migration potential of osteoblasts with increasing volumes of applied VP, suggesting further clinical studies to determine the minimal local concentration of VP necessary for infection prophylaxis.55 A substantial amount of clinical intrasite VP studies have been recently conducted due to the acceptable reported safety profile of VP and the significant morbidity of SSI. Kang et al56 performed a systematic review of the available literature, citing a total of 19 studies with 1 randomized control trial and 1 prospective cohort trial. Although the meta-analysis reported decreased incidence of SSI with intrasite VP use, the authors caution the interpretation of evidence due to potential bias and methodologic flaws throughout the retrospective studies, commonly lacking appropriate controls for confounding and reporting inconsistent definitions of SSI.56 Another similar recent metaanalysis of studies investigating the efficacy of VP in spine procedures yielded 10 studies including 1 randomized control trial. This paper concluded that, although significant heterogeneity existed among publications, VP reduced SSI rates from 4.1% to 1.3%, with a 70% relative risk reduction within the instrumentation subgroup.57 Conversely, Tubaki et al58 performed a recent prospective randomized controlled trial assessing the addition of intrasite VP to systemic preoperative antibiotic prophylaxis in 907 patients undergoing operative treatment for various spinal pathologies. At 18-month follow-up, authors found no statistical significance in control versus VP groups, reporting 1.68% and 1.61% postoperative SSI rates, respectively, with no adverse events associated with intrasite VP use reported. The study suggested that use of VP might not be effective when incidence of infection is low. Although there is controversy involving the prophylactic, off-label use of intrasite VP powder within spinal wounds, it has been widely adopted in an effort to reduce the infrequent, but devastating and costly complication of deep SSI. Several recent cost-effectiveness studies suggest that further reduction in the already infrequent deep SSI occurrence with intrasite VP can lead to substantial cost savings. Highlighting the relatively low cost of intrasite VP (12–34 US dollars) and high costs of deep SSI management (33,705–40,992 US dollars), the potential cost savings with intrasite VP through avoiding reoperations and additional hospitalizations can be significant (244,402–438,165 US dollars per 100 index complex spinal procedures).4,59,60 (Table 2). www.jspinaldisorders.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
355
J Spinal Disord Tech
Boody et al
Volume 28, Number 10, December 2015
TABLE 2. Evidentiary Table—Prophylactic Intrasite Vancomycin Powder Author/ Year Sweet et al52 O’Neill et al 201156 Molinari et al 201262
No. Subjects Trial Design
Patient Population
911
Retrospective cohort
Adult
56
Retrospective cohort Retrospective case series
Adult
1512
Adult
Strom et al 201363
156
Retrospective cohort
Adult
Gans et al 201351 Strom et al 201364
87
Prospective cohort Retrospective cohort
Pediatric
Caroom et al 201365 Heller et al 201366
40
Retrospective cohort
Adult
342
Retrospective cohort
Adult
433
Prospective cohort
Adult
Tubaki et al 201358
79
Adult
Spine Location
Degenerative, Posterior, deformity, thoracolumbar trauma, tumor Trauma Posterior, cervical, thoracolumbar Degenerative, Posterior, cervical, deformity, thoracolumbar, trauma, anterior, cervical, tumor thoracolumbar Degenerative, Posterior, lumbar trauma, tumor
Intraoperative Location
Dosage
Deep Infection Rate
Bone 2g graft+deep+superficial
2/911 (0.2%)*, HC 21/821 (2.6%) (P < 0.0001)
Deep+superficial
1g
0/56 (0%)
Deep
1g
Bone graft+deep
Instrumented: 8/663 (1.2%), Uninstrumented: 7/849 (0.8%) 1g 0/156 (0%)*, HC 11/97 (11%) (P = 0.0000182), Instrumented: 0/88 (0%)*, Uninstrumented: 0/68 (0%) 500 mg 3/87 (3.4%)
Deep+superficial
1g
Deep
1g
Degenerative, Posterior, cervical, trauma, thoracolumbar tumor Degenerative, Posterior, occiput, trauma cervical, thoracolumbar
Superficial
Degenerative, Posterior, cervical, trauma, thoracolumbar tumor Trauma Posterior, cervical, thoracolumbar
Deep+superficial
Deformity
Posterior, thoracolumbar Degenerative, Posterior, cervical trauma, tumor Degenerative Posterior, cervical
Deep+superficial
2/79 (2.5%)*, HC 10/92 (10.9%) (P = 0.0384) 0/40 (0%), HC 11/72 (15%)
Retrospective Adult+pediatric Degenerative, Posterior, cervical, case series trauma, thoracolumbar deformity Retrospective Pediatric Deformity Posterior, case series thoracolumbar
Deep
0.5–2 g 4/342 (1.1%)*, HC: 13/341 (3.8%) (P = 0.029) 1g Instrumented: 6/ 302 (1.99%), HC: 6/304 (1.98%) (P = NS), Uninstrumented: 1/131 (0.76%), HC: 2/170 (1.18%) (P = NS) 1g 0/34 (0%)*, HC: 5/40 (12.5%) (P < 0.033) 1g 0/56 (0%)*, HC: 7/45 (13%) (P = 0.02) 1–6 g 66/981 (6.71%)
Deep+superficial
1g
0/25 (0%)
150
Retrospective cohort
Adult
Degenerative
Posterior, cervical, thoracolumbar
Deep+superficial
1-2 g
156
Retrospective cohort
Adult
Deformity
Posterior, thoracolumbar
Deep+superficial
2g
96
Retrospective cohort
Adult
Degenerative
Posterior, thoracolumbar
Deep+superficial
1g
0/150 (0%)*, HC: 6/150 (4%) (P = 0.0297) 8/156 (5.3%), HC: 8/150 (5.1%) (P = NS) 0/96 (0%), HC: 7/ 207 (3.4%)
Kim et al 201367
34
Retrospective cohort
Adult
Godil et al 20134
56
Retrospective cohort
Adult
Ghobrial et al 201468 Armaghani et al 201469 Hill et al 201470
981
Martin et al 201471 Emohare et al 201459
Primary Spine Diagnosis
25
Deep+superficial
Deep+superficial
*Statistically significant value. HC indicates historical control; NS, not significant.
356 | www.jspinaldisorders.com
Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Spinal Disord Tech
Volume 28, Number 10, December 2015
To further reduce SSI, spine surgeons commonly implement a variety of controversial measures that currently lack confirmatory objective data of their ability to reduce SSI. Although preoperative shaving and iodophor-impregnated drapes are commonly used, no available data have demonstrated their effectiveness in reducing SSI.61,62 Multiple studies reviewing drain usage have found no benefit to reducing SSI, but may increase postoperative blood transfusions.63–66 Reviewing 57 SSIs identified out of 1587 posterior spinal fusions (3.6%), Rao et al67 reported prolonged drain duration as a significant risk factor for SSI [odds ratio (OR) 1.6 per day; 95% confidence interval 1.3–1.9]. Furthermore, limited objective data are available to support recommendations for postoperative surgical dressing regimens. Glennie and colleagues’ recent systematic review of dressings and drains in posterior spine surgery reported a 16.3% versus 6.7% wound complication rate for nonocclusive and occlusive dressings, respectively (OR 2.09; 95% CI, 1.44–3.02). Owing to significant clinical heterogeneity, they offered a low-grade recommendation for the use of postoperative occlusive dressings.68 The duration of postoperative prophylactic antibiotic use is also controversial. Hellbusch et al69 reviewed patients receiving either no postoperative antibiotics (117 patients) or 7 days of postoperative cephalexin dosing (116 patients), noting a reduced rate of SSI with prolonged antibiotics (4.1% vs. 1.7%) yet not reaching statistical significance due to the low event rate. However, continuing prophylactic postoperative antibiotics beyond 24 hours for patients with drains has not been shown to decrease infection rates. Takemoto et al70 recently reported on multilevel thoracolumbar fusions with postoperative drains, noting similar rates of SSI for both 24 hours and duration of drain retention antibiotic dosing schedules (12.4% of 170 patients vs. 13.2% of 144 patients, respectively, P = 0.48).
CLINICAL EVALUATION/IMAGING OF SSIs A thorough history and physical examination is crucial for guiding accurate diagnoses and appropriate treatment. Most commonly, patients with an SSI will present with complaints of increased surgical site or back pain with examination findings of wound tenderness, edema, erythema, and drainage. However, classic infectious presentation findings, such as fevers and constitutional symptoms may not be present.72 A thorough neurological examination compared with prior preoperative and postoperative reported physical examinations is essential to identify new or worsening neurological deficits. Routine laboratory analyses with complete blood count with differential, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) are helpful in the infectious workup but can be misleading in the early postoperative course. In the routine postoperative patient, the CRP commonly returns to baseline within 2 weeks of surgery, whereas the ESR may stay elevated for Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Surgical Site Infections in Spinal Surgery
upwards of 6 weeks postoperatively.72 Trending the values of inflammatory markers can be of use in evaluating early SSI, analyzing for continued postoperative increases in inflammatory marker levels more consistent with infection as opposed to downtrending with routine postoperative inflammation. After laboratory evaluation of infection, imaging is a crucial component of the SSI workup. Routinely, standard radiographs, including weight-bearing films, are obtained to evaluate the status of retained hardware and evidence of instability. The use of magnetic resonance imaging (MRI) has improved our understanding and diagnosis of SSI, yet early, aseptic postoperative inflammation and edema can be difficult to distinguish from an SSI. Studies have demonstrated MRI sensitivity as high as 96% and specificity of 93% in evaluation of epidural abscess, vertebral osteomyelitis, and discitis.73 Zarrouk et al74 performed a prospective study of 29 patients with documented bacterial spondylodiscitis, noting continued imaging abnormality after clinical and biological response to antibiotic treatment regimen. In a recent retrospective review of 98 patients with adult spinal infection including epidural and paraspinal abscesses, Baxi et al75 found that routine follow-up MRI may not correlate with clinical improvements. Although there is insufficient evidence supporting serial MRI use to monitor treatment efficacy of SSI, advanced imaging may be useful in further understanding recalcitrant infections or with worsening clinical conditions.
ETIOLOGY AND MICROBIOLOGY OF SSIs SSIs are categorized by the US Centers for Disease Control into superficial incisional, deep incisional, and organ/space infections. Oftentimes, the source of the SSI is unclear, relating to a variety of host and surgical factors. Bacteria often contaminate surgical wounds during the operative procedure, with the skin as the most common source of contaminants. Early SSIs (< 3 mo) commonly suggest a wound contaminant from the index procedure, whereas late SSIs (> 3 mo) may result from a delayed presentation of an insidious, low-virulence wound contaminant (such as Priopriobacterium acnes) or a remote source such as hematogenous origins from systemic bacteremia or direct inoculations from local extension or subsequent spinal procedures. The microbiology of SSIs is predominantly grampositive bacteria, with S. aureus and S. epidermidis accounting for 45.2% and 31.4%, respectively.76 Gramnegative bacteria (Enterococcus, E. coli, Propriobacterium, and Peptostreptococcus) were found in 30.5% of SSIs and were more common in sacral surgical sites.76 Methicillin-resistant strains accounted for 34% of S. aureus infections, increasing to 47.4% for revision cases.76 Thakkar et al77 found an increased MRSA rate for nasal carriers of MRSA compared with noncarrier controls, with SSI rates of 8% in the nasal MRSA-positive group versus 0.61% in the negative group. P. acnes is an important pathogen to consider for indolent, late SSI. Hahn www.jspinaldisorders.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
357
J Spinal Disord Tech
Boody et al
et al78 reported P. acnes as responsible for 6 of their 7 late infections in adolescent scoliosis instrumented fusions. It is important to communicate with the microbiology laboratory concerning the length of incubation for culture specimens, as P. acnes often requires prolonged (up to 1–2 wk) incubation periods to produce positive cultures. False negatives can occur if the culture is not followed for sufficient time.
MANAGEMENT OF SSIs Debridement Early surgical debridement is at the core of SSI management, with eradication and drainage of the abscess or fluid collection, copious irrigation of the wound bed, and debridement of nonviable soft tissues. Cultures should be obtained intraoperatively before antibiotic administration and should be held for 10–14 days, as low virulence organisms such as P. acnes may not be identified in a usual 5-day incubation period.79 In addition, whereas 5–6 intraoperative cultures have been previously recommended, 6 or more specimens may be required to accurately identify insidious organisms.78–81 If soft tissues allow, primary closure of the wound may be undertaken after debridement. However, the use of vacuum-assisted wound closure or early consultation with plastic surgery colleagues are viable options for patients with tenuous soft tissues or those requiring serial debridements.82 Accurately predicting the number of required debridements to eradicate the SSI can be challenging. Thalgott and colleagues identified that initial debridement culture results and patient’s comoribidities, including systemic disease, immunocompromise, and malnourishment, are prognostic for number of debridements required. Healthy patients with less virulent bacteria commonly required a single debridement, whereas immunocompromised hosts, multiple and/or more virulent organisms, and polymicrobial infections often require multiple debridements.83 DiPaola et al84 evaluated risk factors predicting multiple debridements, identifying MRSA and distant site infection as the strongest predictors and DM, the presence of instrumentation, use of allograft, and posterior lumbar spine location also displaying significant associations.
Retention of Hardware The primary goals of treating postoperative spine SSIs are to eradicate the infection, maintain stability, and achieve fusion (when warranted). While the decision to retain existing instrumentation in the setting of an acute infection may be necessary for maintaining stability or promoting fusion, this may jeopardize the surgeon’s ability to completely eradicate the SSI. The preponderance of available evidence suggests that the ability to both retain hardware and successfully eradicate the infection depends on the acuity of the presentation. Several studies have demonstrated successful eradication of infection with debridement and hardware retention for early-onset SSI. Ahmed et al85 reviewed
358 | www.jspinaldisorders.com
Volume 28, Number 10, December 2015
surgical debridement and retention of instrumentation in 17 patients with SSI after spinal arthrodesis ranging from 1 to 6 weeks after the index procedure, noting eradication of infection in all patients and successful fusion in 15 of 17 (88.2%) patients. Sierra-Hoffman and colleagues reported successful instrumentation retention with early-onset (< 30 d) SSIs with debridement and long-term antibiotics alone, noting eradication of infection in 17 of 19 (89.5%) patients. However, 6 of the 7 late infections (> 30 d) ultimately required instrumentation removal for eradication of the infection.86 Pull ter Gunne and Cohen noted that their management of SSI involved aggressive debridement (89.3%) with hardware retention if stable and revision of hardware if unstable, followed by an average of 40 days of antibiotics. With this protocol, 76% of their deep infections were eradicated with a single debridement, although no comment is made about the chronicity of the SSI before reoperation.9 Conversely, delayed diagnoses of SSI commonly require implant removal for successful infection eradication. Hedequist et al80 found that all 26 cases with SSIs presenting >3 months postoperatively required implant removal to definitively clear the infection. Ho and colleagues reviewed their experience with pediatric SSI after instrumented fusion for scoliosis, noting that 43 of 53 (81%) patients had retained implants at their first irrigation and debridement. They found a significant increase in secondary debridements required with implant retention (47%) in comparison with implant removal at the first irrigation and debridement (20%). However, implant removal was associated with a 10-degree or greater curve progression in 60% of patients.87 Balancing the need for spinal stability and prevention of deformity progression or pseudarthrosis against a more complete eradication of infection remains a case-by-case decision guided by surgeon experience. Mok et al88 reviewed the functional impact of infection after posterior spinal fusion with 12 early (< 90 d) and 4 late (> 90 d) SSIs undergoing debridement with retention of instrumentation and reported no significant difference in long-term SF-36 outcomes compared with noninfected controls at an average follow-up of 56.7 months. Kuhns and colleagues similarly compared quality-of-life (QOL) scores between infected posterior cervical fusions requiring reoperation with noninfected matched controls. Whereas the total projected costs were increased ($21,778 vs. $9159) and 6-month QOLs were significantly lower for the infected cohort, no significant differences were found in QOL outcomes at 12-month follow-up.89
Antibiotics Antibiotics play a critical role in the treatment of SSI, and often consultation with an Infectious Disease specialist is helpful in selecting the appropriate intravenous antibiotic along with dosing, route, and duration. To increase intraoperative culture yields, antibiotics are commonly held preoperatively if the patient does not appear septic. Whereas most studies reviewing antibiotic Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Spinal Disord Tech
Volume 28, Number 10, December 2015
Surgical Site Infections in Spinal Surgery
FIGURE 1. A 21-year-old woman with a history of T4-L2 posterior spinal fusion presented to our orthopedic clinic 18 months after surgery with worsening back pain that began 6 months postoperatively. Workup revealed an elevated erythrocyte sedimentation rate (ESR) at 67 and C-reactive protein (CRP) at 7.8. Magnetic resonance imaging revealed multiple posterior paraspinal fluid collections adjacent to the pedicle screws and rods. Axial T1 postcontrast sequence (A) demonstrates peripheral enhancement of a thoracic paraspinal fluid collection (arrow). Axial T2 sequences (B) of the lumbar spine reveal a hyperintense, well-demarcated paraspinal fluid collection (arrow) with adjacent edema. Sagittal T1 sequences (C) show a hypointense, well-defined collection (arrows) posterior to the spinal instrumentation. Sagittal T2 sequences (D) reveal multiple lumbar paraspinal hyperintense fluid collections (arrows) with adjacent edema. She underwent irrigation and debridement of the posterior spine with removal of grossly loose instrumentation, with the prior fusion sites appearing solidly fused. Postoperatively, she underwent a prolonged course of IV antibiotics. Posoperative standing films (E) and 1-year follow-up standing films (F) demonstrate residual scoliotic deformity with mild deformity progression despite clinical evidence of fusion.
treatment for SSI describe the duration of treatment, no studies were identified that compare outcomes based on duration of treatment. Depending on the antibiotic of choice, the reported duration of treatment ranges from 4 to 20 weeks, with parenteral administration commonly used for 48 hours to 3 weeks.79,81 However, there is no standardized antibiotic protocol available for the treatment of spine SSIs. A prudent approach would be close collaboration with the Infectious Disease service, assessing the need for instrumentation explantation versus retention, microbiology of the infecting organism, and patient comorbidities. In general, treatment of postoperative spine SSIs is considered successful when the Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
wound is healed, there is improvement in the patient’s clinical symptoms, and the inflammatory markers (WBC, ESR, and CRP) have returned to normal. Long-term suppressive oral antibiotic therapy is sometimes required, especially in the setting of retained instrumentation, and should be based on the recommendation of an infectious disease specialist (Fig. 1).
AUTHORS’ PREFERRED TREATMENT The authors truly believe that the first step in minimizing the risk of developing an SSI in spine surgery is understanding the importance of preoperative optimization www.jspinaldisorders.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
359
J Spinal Disord Tech
Boody et al
and utilizing other “modifiable factors” that help prevent this complication. All patients undergoing surgery at our institutions are seen by a “perioperative” medical team that aids in the assistance of medical optimization. Although we do not have a BMI cutoff for surgery, obese patients are strongly encouraged to undergo a significant weight loss program before elective surgery. Patients with a BMI > 35 should be strongly counseled about their increased risk for infection and elective surgery should ideally be avoided. Smoking cessation is required, and a preoperative nicotine test is used before any cervical or lumbar fusion procedure. DM must be well controlled, and Hgb A1c levels are checked and preferably should be