A New Classification of Complications in Neurosurgery

Objective

To define and grade neurosurgical and spinal postoperative complications based on their need for treatment.

Methods

Complications were defined as any deviation from the normal postoperative course occurring within 30 days of surgery. A four-grade scale was proposed based on the therapy used to treat the complications: grade I, any non–life-threatening complications treated without invasive procedures; grade II, complications requiring invasive management such as surgical, endoscopic, and endovascular procedures; grade III, life-threatening adverse events requiring treatment in an intensive care unit (ICU); and grade IV, deaths as a result of complications. Each grade was classified as a surgical or medical complication. An observational test of this system was conducted between January 2008 and December 2009 in a cohort of 1190 patients at the Hospital Italiano de Buenos Aires.

Results

Of 167 complications, 129 (10.84%) were classified as surgical, and 38 (3.19%) were classified as medical complications. Grade I (mild) complications accounted for 31.73%, grade II (moderate) complications accounted for 25.74%, and grade III (severe) complications accounted for 34.13%. The overall mortality rate was 1.17%; 0.84% of deaths were directly related to surgical procedures.

Conclusions

The authors present a simple, practical, and easy to reproduce way to report negative outcomes based on the therapy administered to treat a complication. The main advantages of this classification are the ability to compare surgical results among different centers and times, the ability to compare medical and surgical complications, and the ability to perform future meta-analyses.

Introduction

The assessment of surgical complications is an important tool in neurosurgical practice (11) because it can improve safety and quality of patient treatment (7, 14, 27, 43). The different views on and definitions of what is considered a complication (11), coupled with the absence of a widely accepted classification of postoperative adverse events, may lead to a subjective interpretation of surgical negative outcomes (17, 41, 42). Comparisons between two distinct time periods in a single hospital or between different institutions are almost impossible because no standard reporting system exists (17, 19, 25).

To illustrate the problem, we focus on three extensive reports of complications derived from epilepsy surgery (9, 45, 48), which offer very different criteria for complications. In two of the reports, complication severity was judged minor (transient) when the complication resolved within 1 year of the surgical procedure and major (permanent) when the complication lasted for more than 1 year postoperatively (9, 48). In the third report, events that resolved within 3 months were regarded as minor complications, whereas events that extended for more than 3 months were considered to be major complications (45). There is disagreement about what is a medical or a surgical complication; two reports considered postoperative pneumonia, pulmonary embolism, and deep vein thrombosis to be surgical complications (9, 45), whereas the other report included these complications in a miscellaneous group (48).

Black (11) presented neurosurgeons with the results of a survey on what was listed as a complication at their institutions. Diverse definitions were provided. Black asked whether neurosurgeons could and should adopt a uniform definition of complication.

In 1992, Clavien et al. (17) proposed a classification for general surgery complications that focused on a therapy-oriented, four-level severity grading system. In 2004, Dindo et al. (19) revised and modified this classification for improved accuracy and acceptability in the surgical community and proposed a therapy-based five-grade classification. They showed the reproducibility of their classification through a worldwide survey sent to 10 surgical centers. Our hospital was included in this study and used this system in general surgery for more than 6 years.

In 2001, Bonsanto et al. (14) standardized general adverse neurosurgical postoperative events to conform to a three-category classification: (i) neurosurgical complications, (ii) neurosurgically complicated courses, and (iii) medical nonsurgical complications. In 2009, Houkin et al. (29) presented a quantitative analysis of complications in neurosurgery and classified adverse events into five types, on the basis of adverse event avoidance predictability and possibility.

In 2010, Lebude (32) conducted a survey with more than 200 spine surgeons to establish what was considered a complication. Based on the survey results, Lebude presented a binary definition of complications, by virtue of which they were classified into minor and major adverse events.

We based our classification on the proposal of Clavien (17) and Dindo (19) and modified it to conform to neurosurgical and spine procedures and outcomes more suitably. This classification focuses on general postoperative morbidity. We used a four-grade severity scale based on the therapy administered to treat a postoperative adverse event and considered how it related to the surgical procedure to come up with a simple, practical, and easy to reproduce way to report negative outcomes.