DAVIDE NOTO*, BALDASSARE CANINO*, SALVATORE VIENI*****, GIUSEPPA GRACEFFA*****, VINCENZO PATERNÒ**, EUGENIA HOPPS*, EMANUELA FERTITTA**, MARTINO TINAGLIA***, FRANCESCO BROCATO****, ANTONINA GIAMMANCO*, CATERINA URSO*, ANTONELLA CARDELLA*, MAURIZIO R. AVERNA*, ROSARIO SQUATRITO***
*Department of Biomedicine, Internal Medicine and Medical Specialties (DIBIMIS), University of Palermo, Italy -
**Emergency Care Unit, Hospital “San Raffaele Giglio”, Cefalù, Italy -
*** Core Lab Unit, Hospital “San Raffaele Giglio”, Cefalù, Italy -
****Engineering and informatics Unit, Hospital “San Raffaele Giglio”, Cefalù, Italy -
*****Department of Surgical, Oncological and Oral Science, University of Palermo, Italy
Objective: An early diagnosis of pulmonary embolism (PE) improves outcome. Therefore, PE should be diagnosed in Emergency Care Units (ECU) at admission. Clinical algorithms support the clinician in this task, although performance is biased by differences in risk factors prevalent in different populations. The clinical conditions predictive of PE were evaluated in subjects from Southern Italy accessing ECU for dyspnea/chest pain.
Methods: Retrospective clinical data were obtained by electronic retrieving from a hospital database. Data from 8177 patients (age 18-90 years, 54 with PE) were collected from years 2007-2013.
Results: Previous history of PE, thrombosis and/or phlebitis, rheumatic diseases, respiratory failure, low blood pressure, pulse oxymetry rate (SpO2) and high heart rate were associated with PE diagnosis. High white blood count with neutrophilia, C reactive protein, D-dimer, NT-pro-BNP determinations, but not troponin T, were associated with PE. Recalibration of the GENEVA score and its modification, by inclusion of novel risk factors, improved the algorithm performance (GENEVA AROC=0.730, modified GENEVA AROC = 0.792, DeLong’s test p = <0.001).
Conclusions: PE risk factors in a large Sicilian sample are similar to those of other populations. Data from clinical history and clinical features present at admission were used to recalibrate the PE diagnostic algorithm showing that PE predictive power improved by fitting local data into the predictive model.
Pulmonary embolism, diagnostic score, emergency care unit.