%0 Journal Article %A Isao Ito %A Tadashi Ishida %A Hiromasa Tachibana %A Hiromi Tomioka %A Seizo Kadowaki %A Naoya Tanabe %A Michiaki Mishima %T Prediction of prognosis in healthcare-associated pneumonia %D 2012 %J European Respiratory Journal %P P2473 %V 40 %N Suppl 56 %X Rationale: Healthcare-associated pneumonia (HCAP) is a new category of pneumonia. Since patients with HCAP are at risk for infection with drug-resistant pathogens and increased mortality compared to patients with community-acquired pneumonia (CAP), predicting their prognosis is an important issue in HCAP.Purpose: To apply prognositic scoring systems of CAP (PSI, CURB-65 and Japanese A-DROP systems) in prospectively collected patients with CAP and HCAP.Methods: Patients admitted in three educational hospitals in Japan were analyzed. Receiver operator characteristic curve (ROC) analyses were performed for the three scoring systems in CAP and HCAP. Further, better system was sought in HCAP.Results: 927 cases with CAP (mean age 73.2 years) and 469 cases with HCAP (81.8 years) were enrolled. Compared to HCAP, CAP showed larger values of area under the curve (AUC) in all scoring systems (CAP vs HCAP: PSI 0.77 vs 0.64, CURB-65 0.76 vs 0.65, A-DROP 0.80 vs 0.65). To develop a better scoring system for HCAP, candidate factors for predicting prognosis were extracted by univariate analysis followed by stepwise method. By logistic regression analysis, serum value of albumin (Alb) was related to the prognosis. Each of AUC in prognostic ROC analysis increased by around 0.05 when one point for low Alb (≤3.0mg/dL) was added to each scoring system.Conclusion: Existing scoring systems were not appropriate for predicting mortality in HCAP. Although incorporation of presence of low Alb value into the systems improved prediction of prognosis, more systematic reconstruction of scoring system adapted to HCAP would be needed. %U https://erj.ersjournals.com/content/erj/40/Suppl_56/P2473.full.pdf