Peritonitis Outcome Prediction using Mannheim Perit onitis Index at St. Francis Hospital Nsambya, Kampala - Uganda

Background: A wide variety of disease states give rise to intra-abdominal infection¹. While varying according to age, gender and geography, the three most common causes of generalized peritonitis in low-income countries are probably appendicitis, perforated duodenal ulcer and typhoid perforations, in no particular order².The management of peritonitis for a long time has presented a challenge to surgeons despite different advancements in the field of medicine. This led to the development of disease severity grading systems that would aid in classifying patients by individual risk factors and hence appropriately predict possible outcome. Mannheim Peritonitis Index (MPI) which was developed by Wacha and Linder in 1983 is one of the scoring tools being used to predict outcome. MPI was used to assess outcome in patients with peritonitis at Nsambya hospital. The objective was to predict outcome using the MPI in patients with peritonitis at Nsambya Hospital.
Methods: Retrospective and Prospective evaluation of the MP I score was performed for patients with peritonitis who underwent surgical treatment at Nsambya Hospital for 15 months (Jan 2012- march 2013).
Results: Of the 62 patients included, 29 patients were retrospective and 33 patients prospective, 46 (74.2%) were males while 16 (25.8%) were females, giving a male to female ratio of 3:1. The mean age of presentation was 30.79 (Sd15.55) years, the youngest being 11 years and the oldest 77 years. The mean MPI was 21.92±6.02 points with 10 points as the lowest score and 37 points as the highest score. 58.1% of our patients had an MPI score < 21and 29% MPI 21-29 and 12.9% MPI>29. No death was noted at MPI<21, 2 deaths at MPI 21-29 and 1 death at MPI >29. The most significant predictive factors for morbidity/mortality in this study were the presence of organ failure and female gender. The ROC curve for morbidity showed a predictive power of 0.875 with a sensitivity of 84.2% and a specificity of 90.7% at MPI of ≥26 points. The predictive power of the MPI for mortality was 0.579 with a sensitivity of 15.8% and a specificity of 100% at MPI score of ≥26 points. The positive predictive value and negative predictive value for morbidity and mortality at MPI≥26 points were 75.9%, 94.4% and 100%, 95.9% respectively.
Conclusion: The MPI score is a good tool in predicting morbidity than mortality at Nsambya hospital and can be used to evaluate outcome in patients with peritonitis.