POSSUM: A re‐evaluation in patients undergoing surgery for rectal cancer

BACKGROUND: The problem of directly comparing morbidity and mortality rates between institutions without some sort of adjustment for case mix is well documented. Scoring systems have been developed to allow comparisons to be made. The Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM) is one such system. It was designed to predict operative mortality and morbidity in differing settings and to be independent of case mix. The present study examines the use of POSSUM in colorectal practice in Saudi Arabia. METHODS: Patients referred to King Faisal Specialist Hospital between 1990 and 1998 for primary management of an histologically proven rectal cancer were identified. POSSUM mortality and morbidity scores and Portsmouth-Physiological and Operative Severity Score (P-POSSUM) mortality scores were calculated separately for each patient, and predicted rates were compared with observed rates in the patients studied. RESULTS: There were 70 men (mean age: 55.6 years; range: 25-87) and 75 women (mean age: 52.8 years; range: 26-84). One hundred and six patients underwent 'curative' surgery. Abdominoperineal resection was the most frequently performed procedure. Major anastomotic leakage following anterior resection occurred in two of fifty patients. One patient developed a pulmonary embolism but no patient developed postoperative myocardial infarction. Two patients died. The median and mean physiological and operative severity scores were 13 (range: 12-37) and 17 (range: 8-37) and 14.68 and 18.36, respectively. The overall POSSUM-predicted (using median scores) morbidity and mortality rates were 35.4% and 6.7%. The P-POSSUM-predicted (using mean scores) mortality rate was 3.5%. Observed morbidity and mortality rates were 54.5% and 1.4%. CONCLUSION: POSSUM failed to predict outcomes accurately in patients undergoing surgery for rectal cancer in Saudi Arabia. P-POSSUM also overpredicted mortality but to a lesser extent. Patient's 'wellness' and the previously identified inability of POSSUM to accurately predict death in low-risk populations may explain these findings. Care must be exercised in using the POSSUM formulae for risk adjustment in different settings.