Conditional sensitivity and specificity of tuberculous pleurisy criteria by the inflammation status: a simulation study

H. Park^a (Dr), JH. Lee^a (Dr), CM. Choi^a (Prof), YJ. Jung^a (Prof), KW. Jo*^a (Prof)

^a Asan medical center, Seoul, KOREA, REPUBLIC OF

* healthcliff6800@hanmail.net

Background: The diagnostic performance of ADA and lymphocyte/neutrophil (L/N) ratio for tuberculous effusion could vary according to the conditional subgroup whose suspected diagnosis differs from the general population. However, no study was conducted in those subgroups to evaluate the diagnostic role of tuberculous effusion.

Methods: The thoracentesis conducted between 2009 and 2019 at Asan Medical Center was extracted by a data warehouse system. Two independent clinicians manually reviewed the etiology of pleural effusion. The sensitivity and specificity of ADA and L/N ratio criteria were evaluated by random sampling method to five quantile subgroups according to age and inflammatory lab (CRP, WBC, LD).

Results: Using 40IU/L of ADA and 0.75 of L/N ratio, the overall sensitivity and specificity were 77.7% and 93.1%. In the highest inflammatory subgroup (CRP, WBC), sensitivity lowered to around 60%, although the specificity was about 95%. (Figure 1) The highest LD group showed more than 70% of sensitivity according to ADA (40IU/L) and 0.75 L/N ratio criteria. In the age subgroup, the older age group did not show a lower trend of sensitivity. The youngest age group did not increase the sensitivity by over 80% by liberal criteria (ADA 30IU/L and L/N 0.75) due to the high proportion of low L/N ratios in false-negative cases.

Conclusion: Inflammatory status defined by WBC and CRP affects the sensitivity of ADA and L/N ratio criteria for tuberculous effusion. Clinicians should consider the false-negative cases of tuberculous effusion, especially in high-inflammatory cases, and readjust the cut-off level of ADA and age according to age and inflammatory levels.

Disclosure of funding source(s): none