Validation of MPI and PIA II in two different groups of patients with secondary peritonitis.

BACKGROUND/AIMS: There are several scoring systems designed to predict mortality in patients with peritonitis, which need validation in different patient populations. Our aim was to evaluate Mannheim Peritonitis Index (MPI) and Peritonitis Index of Altona (PIA II) in patients with postoperative peritonitis and other causes of secondary peritonitis.

METHODOLOGY

The records of patients operated for intraabdominal infection between 1987-1996 in Hacettepe University Department of General Surgery, were reviewed retrospectively. A total of 473 patients were included in the study; 75 of them had postoperative peritonitis (POSTOP group) and the remaining 398 had secondary peritonitis due to other causes (OTHER group). Using multiple logistic regression, MPI and PIA II were combined in an equation and this new variable was called combined peritonitis score (CPS); CPS = -9 + (0.3 x MPI) + (-1.2 x PIA II). All patients were scored according to MPI, PIA II and CPS. Receiver-operator characteristic (ROC) curves and sharpness of scores were compared. Also mean scores in both groups, proportions of correct predictions of outcome according to scores and correlation of scores with mortality were compared.

RESULTS

Overall mortality was 17.8% in OTHER group and 33.3% in POSTOP group (P = 0.0018). Higher MPI scores, lower PIA II scores and higher CPS scores were associated with higher mortality in both groups (P < 0.0001). Mean MPI values were higher, mean PIA II values were lower and mean CPS values were higher in POSTOP group (P < 0.001). The areas under ROC curves of CPS were bigger than MPI and PIA II in both groups. Sharpness of CPS was higher in both groups compared to MPI and PIA II (P < 0.05). Proportion of correct predictions of outcome was highest in CPS among the three scores (P = 0.0074). CPS had the best correlation with observed mortality.

CONCLUSIONS

POSTOP group patients had higher MPI, lower PIA II and higher CPS values ending up with higher mortality. This may be because of the delay in diagnosis and treatment, resulting with higher organ failure rates. Generally the results of evaluations for MPI and PIA II are similar. When these two peritonitis scores are combined and used together in the form of CPS, all the parameters improve.