Tests for latent tuberculosis in people with ESRD: a systematic review

Abstract

Background: The relative diagnostic accuracy of interferon γ release assays (IGRAs; based on ELISA [enzyme-linked immunosorbent assay] or ELISPOT [enzyme-linked immunosorbent spot], ie, the QuantiFERON and T-SPOT.TB tests, respectively) and the tuberculin skin test (TST) for latent tuberculosis (TB) infection in people with end-stage kidney disease is uncertain and national guidelines for their use are inconsistent.

Study design: Systematic review.

Selection criteria for studies: Evaluated performance of tests for latent TB with clinical risk-factor assessment.

Setting & population: People with end-stage kidney disease (chronic kidney disease stage 5 [eGFR <15] or kidney transplant recipients). No limits on setting.

Index tests: ELISA- or ELISPOT-based IGRAs, TST, assays to detect antimycobacterial antibodies, and flow cytometry-based tests.

Outcomes: Odds of test positivity with clinical risk factor for latent TB, expressed as ORs and relative ORs (RORs).

Results: 47 studies (6,828 participants) were included, but only 30 studies (4,546 participants) contained sufficient data to contribute to meta-analysis. Studies were predominately in the dialysis population (23/30; 3,700 participants) in countries with low to moderate TB prevalence (0.0-50.0 cases/10(5) persons). BCG vaccination rate was variable (2.7%-100.0%). 9 studies compared IGRAs with the TST directly, 17 studies evaluated the TST only, and the other 4 studies evaluated other tests. Compared to a positive TST result, a positive ELISA-based IGRA result was associated more strongly with radiologic evidence of past TB (ROR, 4.29; 95% CI, 1.83-10.3; P = 0.001) and contact with active TB (ROR, 3.36; 95% CI, 1.61-7.01; P = 0.001). Compared to a negative TST result, a negative ELISA-based IGRA result was associated more strongly with BCG vaccination (ROR, 0.30; 95% CI, 0.14-0.63; P = 0.002). There were insufficient data to compare performance of the ELISPOT-based IGRA with the TST or ELISA-based IGRA.

Limitations: 17 of 47 included studies (36.2%) did not contain sufficient data to contribute to meta-analysis.

Conclusions: Compared to the TST, the ELISA-based IGRA was associated more strongly with risk factors for latent TB in end-stage kidney disease.