Latent tuberculosis diagnostic tests to predict longitudinal tuberculosis during dialysis: a meta-analysis.

Abstract

Tuberculosis (TB) rates in dialysis patients are more than 10 times greater than in the general population. Recent recommendations advise the use of interferon-gamma release assays (IGRAs) over the tuberculin skin test (TST) to aid in the diagnosis of latent tuberculous infection (LTBI); however, their longitudinal predictive ability for TB development has not been assessed. To determine whether the TST or IGRA are able to predict longitudinal TB development in dialysis patients. We performed a systematic review to determine the longitudinal risk of TB in dialysis patients. Random-effects meta-analysis was used to determine the incidence rate ratio (IRR) of longitudinal TB development and the predictive value of such tests. Eight studies were included. An IRR of 2.59 (95%CI 1.20-5.57) for longitudinal TB was seen in patients with a TST ⩾ 10 mm compared to patients with a TST < 10 mm. The positive predictive value (PPV) of a TST ⩾ 10 mm was 11.93% and the negative predictive value was 94.03%. We were unable to analyse the studies that used IGRAs, as only one study had TB events. A TST with a 10 mm cut-off point appears to offer the capability to distinguish long-term risk of TB, with a modest PPV. The predictive value of IGRAs could not be quantified.