Probability assessment of pulmonary embolism using clinical and laboratory variables in hospitalized patients: A single-center, retrospective observational study

Abstract

Introduction: Pulmonary embolism (PE) probability assessment relies on clinical scoring systems, which have limitations for certain patient populations. We aimed to investigate the use of laboratory values for PE probability assessment. Materials and methods: This retrospective single-center observational study included patients with suspected PE. Nineteen variables were examined. Logistic regression analysis adjusted for confounding factors was performed, and significant variables were used to develop a scoring method. Receiver operating characteristic (ROC) curves were used to detect PE and determine the optimal cutoff value. Well's scores were also estimated. Results: The model achieved an accuracy of 84.6 %. Hypocapnia, fever, alkaline phosphatase (ALP), D-dimer, and lactate levels had predictive values. The slope was negative for hypocapnia, ALP, and lactate, and positive for fever and D-dimer levels. Fever, with an adjusted odds ratio (OR) of 1.995, received a score of 2 for values above the cutoff, whereas the remaining variables were assigned a score of 1. Patients with PE had significantly higher scores (mean ± SD: 2.07 ± 0.91) than those without PE (1.80 ± 1.13; P = 0.001). The area under the ROC curve was 0.585 (95 % confidence interval: 0.563–0.606; P = 0.001). Using a cutoff score of 1.5 based on the maximum Youden's index, the scoring system achieved a sensitivity of 73.1 % and specificity of 43.4 %. The Well's score demonstrated a sensitivity of 51.1 % and specificity of 75.1 %. Conclusion: This study showed statistically significant laboratory values for the probability assessment of PE and the tentative scoring system (PAPEL score). Larger prospective multicenter studies are required to validate this scoring method in a wider population.

Keywords

• Pulmonary embolism
• Probability assessment
• Laboratory value
• Logistic models
• Prediction model
• Scoring system: sensitivity and specificity