Reproducibility and Diagnostic Utility of a Simplified Oil Red O Test in Infant Bronchoalveolar Lavage Samples

Aspiration in infants is a diagnostic challenge. The lipid-laden macrophage index (LLMI) developed in 1987 has been used as a supportive test; however, numerous recent studies have questioned its value and reproducibility. We evaluated a simplified LLMI in bronchoalveolar lavage (BAL) specimens from a pediatric cohort to assess its diagnostic utility.

BALs from infants were prospectively collected over a 6-month period for Oil Red O (ORO) staining to evaluate aspiration. BALs from adults with non-aspiration pathologies were simultaneously collected for comparison. Clinical and demographic data were gathered to assess the diagnostic accuracy of the test. Only samples containing ≥ 100 evaluable macrophages and free of obscuring blood or inflammation were included. Positive staining was assessed at low magnification (10×), with only clearly positive cells (Colombo-Hallberg scores 3 and 4) considered. A dichotomous threshold of < 50% or ≥ 50% positive macrophages was established through multidisciplinary consensus. To ensure consistency, a training session was conducted for the entire cytopathology division on the newly developed interpretation criteria.

88/134 (66%) pediatric BAL samples with suspected aspiration and 63/75 (84%) adult samples with various non-aspiration pathologies were adequate for analysis. Aspiration status in children was determined using multidisciplinary aerodigestive group evaluation (MAGE) and videofluoroscopic swallow study (VFSS). Test performance was assessed at various cutoffs. In the pediatric cohort (mean age 16.5 months, 58% male), aspiration was diagnosed in 47% by MAGE. Strong associations were seen with atopia/asthma (83%), functional dysphagia (64%), and congenital/developmental disorders (43%). A significant difference in ≥ 50% lipid-laden macrophage involvement was observed between pediatric (12%) and adult (51%) samples (p < 0.00001). Using MAGE and VFSS as gold standards, the test showed poor discriminatory power for detecting aspiration in infants (AUC 0.506-0.587). A 10% cutoff yielded the best performance (AUC 0.587, sensitivity 27%, specificity 93%), while a 50% cutoff offered practical advantages in workflow and reproducibility.

The modified LLMI demonstrates limited diagnostic value for aspiration in infants. While a 10% cutoff offers slightly improved performance, the test may be phased out in favor of more reliable diagnostic methods.