# Incremental diagnostic yield of consecutive sputum nucleic acid amplification tests for pulmonary tuberculosis

## TL;DR

Using two sputum NAAT tests improves tuberculosis diagnosis accuracy and speed more than one or three tests in low-burden settings.

## Contribution

Demonstrates the optimal number of sputum NAAT tests for diagnosing PTB in non-high-burden settings.

## Key findings

- A second NAAT increases sensitivity from 53.1% to 63.1% and reduces diagnosis time by 2.6 days.
- A third NAAT provides minimal benefit, increasing sensitivity to 67.2% and reducing time by 0.2 days.
- Two NAATs achieve high sensitivity in smear-positive cases and steady gains in smear-negative cases.

## Abstract

The incremental diagnostic yield of consecutive sputum nucleic acid amplification tests (NAATs) for pulmonary tuberculosis (PTB) is not well established. We aimed to assess the diagnostic performance of one to three consecutive sputum NAATs in a non-high-tuberculosis-burden setting. We retrospectively analyzed data from 4,051 patients with suspected PTB in Japan between 2010 and 2023. Patients with a positive culture or a positive NAAT consistent with clinical findings were classified into the PTB group (n = 290); all others were classified into the non-PTB group (n = 3,761). We evaluated the cumulative sensitivity and specificity of smear microscopy, NAAT, and culture for one, two, and three sputum specimens and compared the time to microbiological diagnosis for culture alone versus culture combined with one, two, or three sputum NAATs. In the PTB group, a second NAAT substantially increased cumulative sensitivity from 53.1% to 63.1% (+10.0%) and shortened the mean time to diagnosis from 10.9 to 8.3 days (−2.6 days). A third NAAT provided minimal additional benefit, increasing sensitivity to 67.2% (+4.1%) and reducing the time to diagnosis to 8.1 days (−0.2 days). In smear-positive PTB, the NAAT sensitivity increased from an already high 84.9% with a single test to 97.8% with a second test. Similarly, in smear-negative patients, sensitivity increased from 23.8% to 31.1%. In a non-high-tuberculosis-burden setting, performing a second sputum NAAT significantly improves diagnostic sensitivity and shortens the time to diagnosis.

The standard diagnostic protocol for pulmonary tuberculosis still relies on collecting three consecutive sputum specimens to maximize the diagnostic yield of smear and culture. While nucleic acid amplification tests (NAATs) have become increasingly important diagnostic tools, the incremental value of this approach for sputum NAAT-based diagnosis has remained uncertain. In a large cohort from a non-high-tuberculosis-burden setting, we show that adding a second sputum NAAT meaningfully improves cumulative sensitivity and shortens the mean time to diagnosis, whereas a third sputum NAAT provides only modest additional benefit. Notably, a two-sputum NAAT strategy optimizes yield irrespective of smear status, achieving saturation in smear-positive cases and securing a modest but steady stepwise gain in sensitivity for smear-negative patients. Therefore, our findings provide crucial evidence to support a two-sputum NAAT strategy to optimize the diagnostic workflow for PTB in non-high-tuberculosis-burden settings.