Adequate stratification and scoring of risk is essential to optimize clinical practice; the ability to
predict operative mortality and morbidity is important to choose the gold therapy and a proper use of resources. The
Physiological And Operative Severity Score For The Evaluation Of Mortality And Morbility (POSSUM) has been proved
to be the most appropriate scoring system in providing an estimation of postoperative mortality for patients undergoing
abdominal surgery. The aim of our study was to verify the predictive accuracy of POSSUM in patients undergoing elective
Patients and Methods:
Our study included 189 patients, all admitted to Intensive Care Unit (ICU) at S.Andrea Hospital in
Rome. The sample was divided in 2 groups: group A, abdominal surgery; group B, extrabdominal surgery.
All types of surgery were included except for cardiac surgery, pediatric surgery and urgent surgery. For each patient was
determined the POSSUM score. Age, sex and preoperative information, surgical diagnosis, severity of the procedure ,
time of hospitalization, post operative complications were also recorded.
The two groups were divided in classes, based on the rate of morbidity and mortality. For each class was calculated the relation
between predicted and observed deceases, and the relation between predicted and observed complications, scoring 1
if present and 0 if not present, in order to assign a reliability score to each group. Furthermore, basing on the length of
stay, were individuated three subgroups of patients: hospitalization < 3 days, between 4 and 7 days, > 7 days. For each
group was evaluated the correlation between observed and predicted morbidity.
Above 189 patients, 49 developed postoperative complications, with a real morbidity rate of 25,9% (95%CI:
20%-32%) and a predicted rate of 51,3%; the relation between observed and predicted complications (O/P ratio) was 0,50.
For the A group the O/P ratio was 0,45, with a real morbidity rate of 25,53% (95% CI: 17%-34%) over a predicted rate of
56,4%. In the B group the O/P ratio was 0,53, with a real morbidity rate of 26,31% (95%CI: 17,2%-34,8%), and a predicted
rate of 49%. Above 189 patients, 5 deceased, with a real mortality rate of 2,64% (95% CI: 0,4%-4,8%), over a predicted
rate of 16,4%; the relation between observed and predicted deaths (O/P ratio) was 0,16. For the A group, O/P ratio
was 0,11, with a real mortality rate of 2,1% (95% CI: 0-5%) over a predicted rate of 18,08%; for the B group, O/P ratio
was 0,25, with a real rate of 3,16% (95% CI: 0-436,7%) over a predicted rate of 12,63%. Correlation between time of
hospitalization and POSSUM morbidity has an average strength expressed by a ρ value = 0,4.
Observed data confirm the possibility to extend the POSSUM score in the stratification of patients undergoing
extrabdominal surgery. Even though it overpredicted the value of morbidity and mortality, as widely known, POSSUM
score shows an impressive uniformity and concordance between the main groups A and B. Furthermore, there is a
good correlation between time of stay in ICU and the POSSUM morbidity.
Our results suggest that POSSUM can be an adequate perioperative mean, recommended to evaluate the real condition of
a patient after surgery and to determine the requirement of an admission to the postoperative ICU.