Spinal epidural abscess (SEA) is a serious condition with potentially devastating sequelae, including paralysis and death.1 The incidence of SEA has increased substantially in the past 20 years, likely due to a growing elderly population, increasing use of spine instrumentation, and increasing intravenous drug use.2 High clinical suspicion and advances in neuroimaging have improved mortality rates over the past century, from 81% in the 1920s to 7% in 2017.1,3,4 However, diagnostic delays and misdiagnosis, which can lead to irreversible neurologic deficits, are still common, with rates ranging from 11% to 75%.5,6 Treatment of SEA remains controversial, with some studies supporting emergent surgical decompression and antibiotics while others recommend a trial of medical management with intravenous antibiotics, depending on presenting neurologic status.1,7–10
Because SEA is an uncommon condition, there is a paucity of appropriately powered studies examining mortality following operative intervention.2 Uncontrolled sepsis and evolution of meningitis have been reported as common causes of mortality in patients with SEA.2 However, to the authors’ knowledge, no study to date has assessed which SEA patients are at highest risk for mortality following operative intervention. Identifying the incidence, associated complications, and risk factors for mortality following surgery for SEA may assist surgeons with clinical decision making and improve quality of care for this condition.
The purpose of this study was to determine rate, timing, risk factors, and complications associated with 30-day mortality following surgery for SEA. We hypothesized that (A) the incidence of mortality following surgery for SEA is high, (B) most cases of mortality will occur in the immediate postoperative period, (C) delayed surgery is a risk factor for mortality for SEA, (D) preoperative risk factors can be used to predict mortality following surgery for SEA, and (E) septic shock in SEA is associated with increased mortality.
MATERIALS AND METHODS
The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was utilized in this study.11,12 The NSQIP database is a prospectively collected surgical outcomes database. Data from major surgical cases from member hospitals were collected. Ambulatory surgery centers and Veterans Affairs hospitals are not included in the database. Demographic, comorbidity, perioperative, and 30-day postoperative complication data were collected for each case. Dedicated data collectors at each site are employed that undergo formal training and routine audits by the NSQIP oversight committee to maintain data reliability of at least 95%.12
Cases in the NSQIP database from 2011 to 2016 were included for analysis. Cases were selected based on International Classification of Diseases 9th and 10th Revisions (ICD-9 and ICD-10, respectively). Patients with a primary postoperative diagnosis of spinal epidural abscess (ICD-9: 324.1, ICD-10: G06.1) with surgery performed by orthopedic surgeons and neurosurgeons were identified (n = 1097). Cases were manually reviewed and cases with additional procedures on other anatomic sites were excluded (n = 3).
Outcomes of Interest
The primary outcome of interest was mortality within 30 days of index procedure. Cause of mortality was not documented in the NSQIP database and was not evaluated. NSQIP data coordinators are able to communicate directly with the hospital to confirm mortality.
Postoperative complications were tracked within 30 days following index procedure. Postoperative complications were evaluated for association with 30-day mortality, and include unplanned readmission, return to operating room, deep vein thrombosis, pulmonary embolism, sepsis, septic shock, surgical site infection (superficial, deep, and organ/space infection), wound dehiscence, pneumonia, urinary tract infection, myocardial infarct, cardiac arrest, and stroke.
Risk Factors of Interest
Patient demographics, comorbidities, preoperative laboratory values, perioperative details, and 30-day postoperative outcomes were compared between cohorts. Demographic factors assessed included age, sex, and ethnicity. Comorbidities assessed included body mass index (BMI), American Society of Anesthesiologists (ASA) class, diabetes (insulin and noninsulin dependent), hypertension requiring medication, smoking status, cardiac disease (history of congestive heart failure or dyspnea), respiratory disease (history of ventilator-assisted respirations 48 hours prior to surgery or chronic obstructive pulmonary disease), renal disease (history of renal failure or previous dialysis), bleeding disorders (conditions increasing bleeding risk, including vitamin K deficiency, hemophilia, thrombocytopenia, chronic anticoagulation not discontinued prior to surgery), corticosteroid use for chronic conditions, and disseminated cancer. Laboratory values assessed included serum sodium, white blood cell (WBC) count, hematocrit, platelet count, serum albumin, and international normalized ratio. We defined hyponatremia as serum sodium <135 mEq/L, hypernatremia as serum sodium >145 mEq/L, leukocytosis as WBC>12,000/mm3, leukopenia as WBC<4000/mm3, anemia as hematocrit <40.6%, thrombocytopenia as platelets<150,000/μL), and hypoalbuminemia as albumin <3.5 g/dL as consistent with prior literature.13–16 Operative details assessed were surgical subspecialty, days from admission to surgery, operative time, wound class, occurrences of blood transfusion (intraoperative or postoperative within 72 h of index procedure), and length of hospital stay. Prolonged operative time (surgery ≥ 310 min) was also assessed.17 Procedural details, including procedures performed, region of surgery, approach, number of levels, revision cases, dural tear repairs, and use of microscope were assessed. Current Procedural Terminology codes (CPT) codes used for cohort construction are provided in supplementary Table 1, http://links.lww.com/BRS/B390.
Categorical variables are expressed as percentages and continuous variables are expressed as mean ± standard deviation. Univariate analysis comparing study cohorts was performed using two-tailed Student t tests with Levene test for equality of variance or chi-squared/Fisher exact tests as appropriate.
Cox proportional hazard models were created to identify independent predictors for unplanned readmission. Time to event (index surgery to mortality) was used in this survival model. Demographic, comorbidity, laboratory value, and perioperative details that were significant on univariate analysis were included. Patients were censored at 30 days following index procedure. Using a forward, stepwise procedure, variables with a P≤0.10 remained in the final model, with significant variables having P < 0.05. A simplified predictive algorithm based on number of independent risk factors was created and assessed by receiver operator curve (ROC) analysis.
Complications found significantly associated with 30-day mortality on univariate analysis were further evaluated adjusting for associated risk factors identified by Cox regression analysis. Forward-stepwise binary logistic models were used to evaluate the adjusted association of mortality with each complication. Variables with a P≤0.10 remained in the final model, with significant variables having P < 0.05.
A Bonferroni corrected P value of P = 0.003 was applied to assessment of complications associated with 30-day mortality to adjust for multiple comparisons. For all other analyses, a P value <0.05 was considered statistically significant. Statistical analysis was performed using SPSS (SPSS 25.0, IBM, Armonk, NY).
Incidence and Timing of Mortality
There were 1094 patients included for analysis. The 30-day mortality rate following spine surgery for epidural abscess was 3.7% (n = 40, Figure 1). Of the mortalities that occurred within 30 days of surgery, the majority occurred within 2 weeks postoperatively (70% within 15 days, Figure 2).
Risk Factors Associated With Mortality
Univariate analysis was performed evaluating demographic, comorbidity, laboratory values, and operative details. Older age, diabetes, hypertension, respiratory disease, renal disease, and a history of disseminated cancer were associated with 30-day mortality following epidural abscess surgery on univariate analysis (Table 1). Analysis of laboratory values identified preoperative thrombocytopenia as associated with postoperative mortality (P < 0.001).
Univariate analysis of perioperative risk factors identified perioperative blood transfusions as associated with postoperative mortality (P = 0.005, Table 2). There were no differences in mortality based on surgical specialty, days from admission to surgery, operative time, or hospital length of stay. Type of spine surgery performed, region of surgery, surgical approach, number of levels of surgery, revision surgery, dural tear repair, and use of microscope were not associated with 30-day mortality.
Multivariate analysis identified age>60 years (hazard ratio [HR]: 2.147, P = 0.027), diabetes (HR: 2.242, P = 0.015), respiratory comorbidities (HR: 2.416, P = 0.037), renal comorbidities (HR: 2.556, P = 0.022), disseminated cancer (HR: 5.219, P = 0.001), and preoperative thrombocytopenia (HR: 3.276, P = 0.001) as independent risk factors for 30-day mortality following spine surgery for epidural abscess (Table 3).
Predictive Algorithm for Mortality
A simplified predictive algorithm for 30-day mortality following surgery for SEA is presented in Table 4. The algorithm predicts a 0.3% mortality for patients with zero risk factors up to 37.5% for patients with 4 or more risk factors. A ROC area under curve (AUC) was 0.761, signifying a fair predictor (95% CI: 0.683–0.839, P < 0.001) (Figure 3).
Complications Associated With Mortality
There were a total of 15 perioperative complications assessed (Table 5). Complications associated with mortality included septic shock (P < 0.001), postoperative pneumonia (P = 0.001), and cardiac arrest (P < 0.001). These factors remained significant on multivariate analysis upon controlling for risk factors of death found previously (Figure 4).
Rates of mortality for patients presenting with SEA have decreased over the past century.1,3,4 Due to the rarity of the condition, there is a paucity of research on mortality following SEA. Utilizing a large national database (NSQIP), we report a 30-day mortality rate of 3.7% following surgery for SEA, with the majority of mortality events occurring within 2 weeks of operative intervention. Septic shock, cardiac arrest, and pneumonia were complications associated with mortality following surgery for SEA. Risk factors for mortality included older age, history of diabetes, respiratory comorbidities, renal comorbidities, disseminated cancer, and thrombocytopenia. A predictive algorithm utilizing these risk factors demonstrated fair predictive value. To the authors’ knowledge, this is the first investigation examining risk factors for 30-day mortality following spine surgery for SEA in the United States. Our study provides benchmark data that can be used to develop quality-control initiatives to decrease rates of mortality.
We report a 30-day mortality rate of 3.7% following surgery for SEA. This is comparable to rates reported by other recent studies.1,2,18 Of the mortalities that occurred within 30 days of surgery, the majority (70%) occurred in the first 15 days following surgery, implying that the rate of mortality decreases after the 2 week benchmark. Decisions regarding end of life care are particularly difficult to navigate, and are often made with limited data. Understanding timing of mortality following surgery for SEA is important for physicians, patients, and their surrogates to make informed decisions regarding treatment, and these results may help with this delicate process.
Our study identified six independent risk factors for mortality following spine surgery for SEA—older age, diabetes, respiratory comorbidities, renal comorbidities, and metastatic cancer. These have been well explored as risk factors for morbidity and mortality following spine surgery, with proposed mechanisms including an immunocompromised state and increased disease burden.19–25 Our study also found low platelet count (<150,000/μL) to be an independent risk factor for mortality. Bleeding disorders have been associated with increased mortality following spine surgery in multiple studies.26–28 Low platelet count (<150,000/μL) specifically has been reported as a risk factor for transfusions, surgical site infections, and mortality following posterior lumbar fusions.26 We believe that a low platelet count may increase mortality by increasing risk for bleeding complications or that it may be a marker of severe systemic disease. Current guidelines recommend prophylactic platelet transfusion for spine surgery if preoperative platelet count falls below 100,000/μL.29 Further study of the preoperative platelet transfusion thresholds should be performed, as intervention may affect mortality rates following spine surgery. Importantly, none of the risk factors identified in this study were surgeon or procedure-related, further emphasizing the importance of optimizing medical care and multidisciplinary care teams for patients with SEA.
Risk stratification tools have become increasingly used in medicine as prognostic tools to help guide management, assess efficacy of interventions, and assess quality of care between institutions.30–33 In a healthcare system with limited resources, identifying patients at high risk for mortality following SEA may help triage patients requiring intensive care unit (ICU) stays. Interventions addressing these risk factors may decrease risk of mortality, and can be benchmarked against the risk stratification tool. The United States department of Medicare and Medicaid services also uses 30-day mortality as a quality metric to financially incentivize hospitals to reduce mortality rates.34 We present a predictive model using the independent risk factors we identified that predicts a mortality rate of 0.8% for a patient with no risk factors up to a mortality rate of 37.5% for a patient with four or more risk factors (Table 5). This model had fair predictive value on ROC analysis. Further study to identify additional independent risk factors may improve the predictive value of this model. This model should also be prospectively validated.
Reasons for mortality following SEA are not well studied. In a review article, mortality following SEA is attributed to sepsis and evolution of meningitis.2 However, there is scarce evidence-based primary literature on this subject. The NSQIP does not describe direct cause of mortality. However, we found that the most common complications associated with mortality following SEA were septic shock (42.5%), cardiac arrest (35.0%), and pneumonia (17.5%). Although septic shock is a sequelae of uncontrolled disease, less than 50% of mortalities following SEA developed septic shock. Cardiac arrest was frequently associated with mortality following SEA surgery, underscoring the importance of cardiac clearance and optimization to control risk for cardiac complications.35–37 Risk factors for pneumonia following surgery is also a topic of significant study.12,38,39 Although association cannot imply causation, interventions that decrease risk of cardiac and pulmonary complications may also indirectly decrease rates of mortality following spine surgery for SEA. These results also suggest the importance of multidisciplinary care for patients with SEA, which has been shown to improve quality of care.40
There are multiple limitations to this study owing to the nature of the NSQIP database. The number of 30-day mortality cases in this study was relatively low (n = 40) and may have been underpowered to detect certain risk factors. This is owing to the relatively rare incidence of SEA. The NSQIP expanded significantly from 121 sites in 2006 to 693 sites in 2018, and future studies may improve the power of studies on mortality following spine surgery for SEA. The NSQIP is limited to 30-day postoperative complications. Therefore, the actual rate of mortality following surgery for SEA is likely underreported. Further study should extend follow-up and may offer insights into risk factors for early versus late cases of mortality. Although we identified complications associated with mortality, the direct reason for mortality is not recorded in the NSQIP and therefore limits assessment. The database does not assess neurologic deficits or quality of life metrics, which are an important topic of study in patients with SEA.1,2,41 The NSQIP does not document antibiotic regimens, specify preoperative medical management details (including percutaneous drainage), or record cases that successfully completed medical management for SEA.1,7–10 However, urgent surgical decompression is considered by many to be the gold standard for management of SEA.18,42 Finally, it would be of tremendous value to compare mortality rates of surgically versus non-surgically treated SEA, which should be the subject of further study. The findings of this study should be interpreted with these limitations in mind.
SEA is an increasingly common and potentially fatal condition that is often treated with surgical decompression. We found a 30-day mortality rate of 3.7%. Of the mortalities that occurred within 30 days of surgery, the majority occurred within 2 weeks of surgical intervention. We identified 6 independent risk factors for mortality following surgery for SEA, all of which were patient-related. They include older age, diabetes, hypertension, respiratory comorbidities, renal comorbidities, metastatic cancer, and thrombocytopenia. We used these independent risk factors to create a simplified model to predict risk of 30-day mortality. Mortality was independently associated with septic shock, cardiac arrest, and pneumonia. These findings may help medical teams identify potential areas for intervention to decrease mortality and improve physician-patient informed consent for this condition.Key Points30-day mortality following operatively treated spinal epidural abscess was 3.7%, with the majority of deaths occurring within 2 weeks.Independent risk factors for mortality included older age, diabetes, hypertension, respiratory comorbidities, renal comorbidities, metastatic cancer, and thrombocytopenia.Septic shock, cardiac arrest, and pneumonia were associated with mortality.
1. Vakili M, Crum-Cianflone NF. Spinal epidural abscess
: a series of 101 cases. Am J Med
2. Darouiche RO. Spinal epidural abscess
. N Engl J Med
3. Akalan N, Ozgen T. Infection
as a cause of spinal cord compression: a review of 36 spinal epidural abscess
cases. Acta Neurochir
4. Dandy WE. Abscesses and inflammatory tumors in the spinal epidural space (so-called pachymeningitis externa). Arch Surg
5. Davis DP, Wold RM, Patel RJ, et al. The clinical presentation and impact of diagnostic delays on emergency department patients with spinal epidural abscess
. J Emerg Med
6. Tang HJ, Lin HJ, Liu YC, et al. Spinal epidural abscess
—experience with 46 patients and evaluation of prognostic factors. J Infect
7. Shah AA, Ogink PT, Nelson SB, et al. Nonoperative management of spinal epidural abscess
: development of a predictive algorithm for failure. J Bone Joint Surg Am
8. Curry WT Jr, Hoh BL, Amin-Hanjani S, et al. Spinal epidural abscess
: clinical presentation, management, and outcome. Surg Neurol
9. Rigamonti D, Liem L, Sampath P, et al. Spinal epidural abscess
: contemporary trends in etiology, evaluation, and management. Surg Neurol
10. Patel AR, Alton TB, Bransford RJ, et al. Spinal epidural abscesses: risk
factors, medical versus surgical management, a retrospective review of 128 cases. Spine J
12. Fuchshuber PR, Greif W, Tidwell CR, et al. The power of the National Surgical Quality
Improvement Program—achieving a zero pneumonia rate in general surgery patients. Perm J
13. Noureldin M, Habermann EB, Ubl DS, et al. Unplanned readmissions following outpatient hand and elbow surgery. J Bone Joint Surg Am
14. Braun MM, Barstow CH, Pyzocha NJ. Diagnosis and management of sodium disorders: hyponatremia and hypernatremia. Am Fam Physician
15. Costa RT, Nassar AP Jr, Caruso P. Accuracy of SOFA, qSOFA, and SIRS scores for mortality
in cancer patients admitted to an intensive care unit with suspected infection
. J Crit Care
16. Williamson DR, Albert M, Heels-Ansdell D, et al. Thrombocytopenia in critically ill patients receiving thromboprophylaxis: frequency, risk
factors, and outcomes. Chest
17. Schoenfeld AJ, Carey PA, Cleveland AW 3rd, et al. Patient factors, comorbidities, and surgical characteristics that increase mortality
and complication risk
after spinal arthrodesis: a prognostic study based on 5,887 patients. Spine J
18. Darouiche RO, Hamill RJ, Greenberg SB, et al. Bacterial spinal epidural abscess
. Review of 43 cases and literature survey. Medicine (Baltimore)
19. Salmenkivi J, Sund R, Paavola M, et al. Mortality
caused by surgery for degenerative lumbar spine
. Spine (Phila Pa 1976)
20. Bohl DD, Mayo BC, Massel DH, et al. Incidence and risk
factors for pneumonia after posterior lumbar fusion procedures: an ACS-NSQIP study. Spine (Phila Pa 1976)
21. Bohl DD, Ahn J, Rossi VJ, et al. Incidence and risk
factors for pneumonia following anterior cervical decompression and fusion procedures: an ACS-NSQIP study. Spine J
22. Chung AS, Campbell DH, Hustedt JW, et al. Inpatient outcomes in dialysis-dependent patients undergoing elective lumbar surgery for degenerative lumbar disease. Spine (Phila Pa 1976)
23. De la Garza Ramos R, Jain A, Nakhla J, et al. Postoperative morbidity and mortality
after elective anterior cervical fusion in patients with chronic and end-stage renal disease. World Neurosurg
24. Piper K, DeAndrea-Lazarus I, Algattas H, et al. Risk
factors associated with readmission and reoperation in patients undergoing spine surgery
. World Neurosurg
25. De la Garza Ramos R, Goodwin CR, Jain A, et al. Development of a Metastatic Spinal Tumor Frailty Index (MSTFI) using a nationwide database and its association with inpatient morbidity, mortality
, and length of stay after spine surgery
. World Neurosurg
26. Bronheim RS, Oermann EK, Cho SK, et al. Coagulation profile as a risk
factor for 30-day morbidity and mortality
following posterior lumbar fusion. Spine (Phila Pa 1976)
27. Pumberger M, Chiu YL, Ma Y, et al. Perioperative mortality
after lumbar spinal fusion surgery: an analysis of epidemiology and risk
factors. Eur Spine J
28. Worley N, Marascalchi B, Jalai CM, et al. Predictors of inpatient morbidity and mortality
in adult spinal deformity surgery. Eur Spine J
29. Liumbruno GM, Bennardello F, Lattanzio A, et al. Recommendations for the transfusion management of patients in the peri-operative period. I. The pre-operative period. Blood Transfus
30. el-Solh AA, Grant BJ. A comparison of severity of illness scoring systems for critically ill obstetric patients. Chest
31. Chen YC, Tsai MH, Ho YP, et al. Comparison of the severity of illness scoring systems for critically ill cirrhotic patients with renal failure. Clin Nephrol
32. Breslow MJ, Badawi O. Severity scoring in the critically ill: part 1—interpretation and accuracy of outcome prediction
scoring systems. Chest
33. Breslow MJ, Badawi O. Severity scoring in the critically ill: part 2: maximizing value from outcome prediction
scoring systems. Chest
35. Ashley EA, Vagelos RH. Preoperative cardiac evaluation: mechanisms, assessment, and reduction of risk
. Thorac Surg Clin
36. Hollenberg SM. Preoperative cardiac risk
1999; 115 (5 suppl):51S–57S.
37. Leppo JA. Preoperative cardiac risk
assessment for noncardiac surgery. J Nucl Cardiol
38. Chughtai M, Gwam CU, Mohamed N, et al. The epidemiology and risk
factors for postoperative pneumonia. J Clin Med Res
39. Rustagi T, Drazin D, Oner C, et al. Fractures in spinal ankylosing disorders: a narrative review of disease and injury types, treatment techniques, and outcomes. J Orthop Trauma
2017; 31 (suppl 4):S57–S74.
40. Epstein NE. Multidisciplinary in-hospital teams improve patient outcomes: a review. Surg Neurol Int
2014; 5 (suppl 7):S295–S303.
41. Alton TB, Patel AR, Bransford RJ, et al. Is there a difference in neurologic outcome in medical versus early operative management of cervical epidural abscesses? Spine J
42. Hlavin ML, Kaminski HJ, Ross JS, et al. Spinal epidural abscess
: a ten-year perspective. Neurosurgery
cervical spine; infection; lumbar spine; mortality; prediction; quality; risk; risk stratification; spinal epidural abscess; spine surgery
Supplemental Digital Content
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.