Reducing pleural fluid mischaracterization – Using demographics to improve Light’s criteria.

<bold>Introduction:</bold> Light’s criteria has a high sensitivity (up to 98%) in detecting exudative pleural effusions but a specificity of 80%. Multiple predictive models in medicine use demographics, but Light’s criteria relies solely on laboratory parameters. We aimed to assess the association of demographics in characterizing pleural effusion. <bold>Methods:</bold> This is a retrospective study of 1,155 consecutive patients with pleural effusion who underwent a thoracentesis. SPSS 25.0 was used to analyze the data to calculate mean, median, univariate, and multivariable logistic regressions. We adjusted for confounders like age, race, sex, comorbidities, and type of effusion (parapneumonic). ROC curves were generated to calculate the diagnostic accuracy. <bold>Results:</bold> Age was the most significant factor associated with transudative effusion [Table 1]. Female gender [Adjusted OR: 2.071 (1.151-3.728), p=0.015] and Arab population [Adjusted OR: 3.015 (1.693-5.369), p<0.001] were strong predictors of transudative effusion even after adjusting for age, comorbidities and parapneumonic cause. CHF [Adjusted OR: 4.540 (1.840-11.200), p=0.001] and CKD [Adjusted OR: 2.994 (1.435-6.248), p=0.003] were associated with transudate. On ROC, pleural fluid LDH was found to have the highest diagnostic accuracy for diagnosing exudative effusion (91.9%), followed by pleural fluid protein (86.4%) and pleural fluid WBC (74.1%), while pleural fluid glucose had 90.5% accuracy for transudative effusion. <bold>Conclusion:</bold> Our study found an association between sociodemographic factors like age, gender, and race with the type of pleural effusion. Further, well-designed studies may improve the diagnostic accuracy of pleural fluid characterization.