Introduction: Pediatric community acquired complicated pneumonia (PCACP) has on average a prolonged clinical course but this may be highly variable. This variability contributes to the lack of consistent evidence based data regarding invasive and fibrinolytic therapy. Unlike in adults, in children there are no good clinical or laboratory parameters which predict a complicated disease course and accordingly tailor treatment.
Objective: To develop and validated a reliable clinical tool for the prediction of prolonged complicated hospitalizations in PCACP.
Methods: We conducted a retrospective cohort study. The derivation cohort consisted of patients with PCACP hospitalized in the three major hospitals in Jerusalem between the years 2001 and 2010. Clinical and laboratory parameters predicting prolonged hospitalizations were evaluated. The validation cohort consisted of patients with PCACP hospitalized between 2011 and 2016 in two major hospitals in Jerusalem.
Results: The derivation cohort included 144 children. After a multivariate analysis, lower levels of glucose (p=0.005) and higher levels of LDH (p<0.001) in pleural fluid were significantly associated with prolonged hospitalization. A stepwise logistic regression model applied on the validation cohort (n=26) yielded a positive predictive value of 85%, negative predictive value of 57%, sensitivity of 78% and a specificity of 67%.
Conclusion: The model established in this study accurately identifies patients with a prolonged disease course. In patients with PCACP, applying an initial predictive approach may allow better therapeutic decisions regarding invasive and fibrinolytic therapy and may allow selection of the more complicated patients for clinical trials.
