Background: A severity grading system is essential to reporting surgical complications. In 1992, we presented such a system (T92). Its use and that of systems derived from it have increased exponentially. Our purpose was to determine how well T92 and its modifications have functioned as a severity grading system and to develop an improved system for reporting complications. Methods: 129 articles were studied in detail. Twenty variables were searched for in each article with particular emphasis on type of study, substitution of qualitative terms for grades, grade compression, and cut-points if grade compression was used. We also determined relative distribution of complications and manner of presentation of complications. Results: T92 and derivative classifications have received wide use in surgical studies ranging from small studies with few complications to large studies of complex operations that describe many complications. There is a strong tendency to contract classifications and to substitute terms with self evident meaning for the numerical grades. Complications are presented in a large variety of tabular forms some of which are much easier to follow than others. Conclusions: Current methods for reporting the severity of complications incompletely fulfill the needs of authors of surgical studies. A new system-the Accordion Severity Grading System-is presented. The Accordion system can be used more readily for small as well as large studies. It introduces standard definition of simple quantitative terms and presents a standard tabular reporting system. This system should bring the field closer to a common severity grading method for surgical complications.