The utility of a newly-developed Mycobacterium tuberculosis-specific enzyme-linked immunosorbent spot (ELISPOT) assay for diagnosis of tuberculous peritonitis (TBP) has not been fully assessed.

All patients with suspected TBP in a tertiary care hospital in an intermediate TB burden country were prospectively enrolled over a 30-month period. ELISPOT assays were performed on peripheral mononuclear cells (PBMC) and mononuclear cells from peritoneal fluid (PF-MC).

Sixty-four patients with suspected TBP were enrolled. Of these, 30 (47%) were classified as having TBP (27 confirmed and 3 probable cases), and 25 (39%) were classified as not having active tuberculosis. The remaining 9 (14%) with possible TBP were excluded from the final analysis. Five (8%) of the total 64 patients gave indeterminate PBMC ELISPOT results and 7 (18%) of 39 patients who underwent PF-MC ELISPOT assay revealed indeterminate PF-MC ELISPOT results. The sensitivity and specificity, respectively, of the tested methods for diagnosing TBP were as follows: PBMC ELISPOT (≥6 spots), 86% and 67%; PF-MC ELISPOT (≥14spots), 92% and 86%; PF-MC ELISPOT/PBMC ELISPOT ratio (≥ 2), 75% and 93%; and PF ADA levels (≥38IU/L), 95% and 100%. The areas under the receiver operating characteristics curves were as follows: PF-MC ELISPOT, 0.96; PF ADA, 0.96; PBMC ELISPOT, 0.88; and PF-MC ELISPOT/PBMC ELISPOT ratio, 0.87, respectively.

Although the ELISPOT assay does not outperform PF ADA, the ELISPOT assay using PBMC and PF-MC is a useful adjunct for diagnosing TBP, especially for a rule-in test when PF/MC/PBMC ELISPOT ratio (≥2) is used. However, the relatively high proportion of indeterminate results limits test utility, so further studies are needed to develop an optimized assay prototype.