Late-presenting dural tear: incidence, risk factors, and associated complications.

BACKGROUND CONTEXT: Unrecognized and inadequately repaired intraoperative durotomies may lead to cerebrospinal fluid leak, pseudomeningocele, and other complications. Few studies have investigated durotomy that is unrecognized intraoperatively and requires additional postoperative management (hereafter, late-presenting dural tear [LPDT]), although estimates of LPDT range from 0.6 to 8.3 per 1,000 spinal surgeries. These single-center studies are based on relatively small sample sizes for an event of this rarity, all with <10 patients experiencing LPDT.

PURPOSE: This investigation is the largest yet conducted on LPDT, and sought to identify incidence, risk factors for, and complications associated with LPDT.

STUDY DESIGN/SETTING: This observational cohort study employed the American College of Surgeons National Surgical Quality Improvement Program dataset (years 2012-2015).

PATIENT SAMPLE: Patients who underwent spine surgery were identified based on presence of primary listed Current Procedural Terminology (CPT) codes corresponding to spinal fusion or isolated posterior decompression without fusion.

OUTCOME MEASURES: The primary variable in this study was occurrence of LPDT, identified as reoperation or readmission with durotomy-specific CPT or International Classification of Diseases, Ninth Revision, Clinical Modification codes but without durotomy codes present for the index procedure.

METHODS: Descriptive statistics were generated. Bivariate and multivariate analyses were conducted using chi-square tests and multiple logistic regression, respectively, generating both risk factors for LPDT and independent association of LPDT with postoperative complications. Statistical significance was defined as p<.05.

RESULTS: In total, 86,212 patients were analyzed. The overall rate of reoperation or readmission without reoperation for LPDT was 2.0 per 1,000 patients (n=174). Of LPDT patients, 97.7% required one or more unplanned reoperations (n=170), and 5.7% of patients (n=10) required two reoperations. On multivariate analysis, lumbar procedures (odds ratio [OR] 2.79, p<.0001, vs. cervical), procedures involving both cervical and lumbar levels (OR 3.78, p=.0338, vs. cervical only), procedures with decompression only (OR 1.72, p=.0017, vs. fusion and decompression), and operative duration ≥250 minutes (OR 1.70, p=.0058, vs. <250 minutes) were associated with increased likelihood of LPDT. Late-presenting dural tear was significantly associated with surgical site infection (SSI) (OR 2.54, p<.0001), wound disruption (OR 2.24, p<.0001), sepsis (OR 2.19, p<.0001), thromboembolism (OR 1.71, p<.0001), acute kidney injury (OR 1.59, p=.0281), pneumonia (OR 1.14, p=.0269), and urinary tract infection (UTI) (OR 1.08, p=.0057).

CONCLUSIONS: Late-presenting dural tears occurred in 2.0 per 1,000 patients who underwent spine surgery. Patients who underwent lumbar procedures, decompression procedures, and procedures with operative duration ≥250 minutes were at increased risk for LPDT. Further, LPDT was independently associated with increased likelihood of SSI, sepsis, pneumonia, UTI, wound dehiscence, thromboembolism, and acute kidney injury. As LPDT is associated with markedly increased morbidity and potential liability risk, spine surgeons should be aware of best-practice management for LPDT and consider it a rare, but possible etiology for developing postoperative complications.