Do pleural fluid pH and lactate improve classification beyond Light's criteria?: A single-center retrospective study

Accurate differentiation between transudative and exudative pleural effusions is essential for determining the underlying etiology and guiding management. Although Light's criteria remain the standard diagnostic approach, their performance may be limited in borderline cases. Biochemical markers such as pleural fluid pH and lactate have been proposed as adjunct tools, yet the extent to which they provide additional diagnostic improvement beyond Light's criteria remains unclear. This single-center retrospective study specifically evaluated whether adding pleural fluid pH and lactate to Light's criteria improves the ability to distinguish transudative from exudative pleural effusions, quantifying the additional diagnostic value of these markers. We retrospectively analyzed 281 patients who underwent diagnostic thoracentesis between November 2020 and March 2024. Receiver operating characteristic-derived cutoff values and corresponding area under the curve (AUC)s for pH and lactate were generated and assessed within the same cohort. Multivariable logistic regression models incorporating Light's criteria alone and in combination with pH and/or lactate were constructed to assess incremental discrimination (ΔAUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI). Of the 281 patients, 62 (22.1%) had transudative and 219 (77.9%) had exudative effusions. Pleural fluid pH ≤ 7.44 showed an AUC of 0.767 (95% CI, 0.696-0.839), sensitivity 71.4%, and specificity 70.2%. Lactate ≥ 2.05 mmol/L yielded an AUC of 0.701 (95% CI, 0.627-0.775), sensitivity 61.7%, and specificity 63.2%. Adding pH to Light's criteria increased diagnostic accuracy (Model-2 AUC 0.985; ΔAUC + 0.010; NRI 0.904; IDI 0.035). Lactate also improved discrimination, though to a lesser extent (Model-3 AUC 0.982; ΔAUC + 0.007; NRI 0.607; IDI 0.025). Incorporating both biomarkers provided the highest incremental metrics (AUC 0.986; ΔAUC + 0.011; IDI 0.053), demonstrating a modest additive benefit across models. Pleural fluid pH and lactate, when added to Light's criteria, provide distinct yet modest incremental gains in diagnostic performance. These findings clarify their potential as an adjunct, though further external validation is necessary before routine clinical application.