# The incremental value of tuberculosis detecting African giant pouched rats over Smear microscopy and Xpert MTB/RIF for Tanzanian TB detection

## TL;DR

African giant pouched rats trained to detect tuberculosis can significantly increase TB case detection beyond current methods like smear microscopy and Xpert.

## Contribution

This study quantifies the incremental TB detection value of trained rats over Xpert and smear microscopy in Tanzania.

## Key findings

- Trained rats detected 48% more TB cases than smear and Xpert combined.
- Rats identified 44% more cases at Xpert sites and 59% more at smear sites.
- Rats were 47% more likely to detect TB in low bacillary load samples compared to smear.

## Abstract

Prior studies that used Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling (APOPO)-trained African giant pouched rats reported an annual average of 40% increase in sputum-smear-positive tuberculosis (TB) detection. Sputum-smear microscopy (smear) is being replaced by Xpert MTB/RIF (Xpert)-based diagnostic algorithms in many countries. We evaluated the incremental values of rat-based case detection over smear and Xpert.

From January to December 2023, sputum samples were collected from presumptive TB patients at 69 health facilities that use Xpert or smear-based diagnostic algorithms. An average of five rats were used at APOPO to re-evaluate the second sample from Xpert-negative patients and the smear-negative sample. Concentrated smears and Ziehl-Neelsen staining were used to confirm samples that were rat-positive (samples indicated positive by  ≥ 1 rat). We analyzed the increase in case detection (incremental yield) by the rats over smear and Xpert and compared the rise against Mycobacterium bacillary load using the Chi-square test.

A total of 43,153 samples from 34,565 patients were collected and re-evaluated by rats. There were 6,717 bacteriologically confirmed TB cases; 4,541 (68%) of these cases were detected at health facilities (3,239 from Xpert and 1,302 from smear sites); the remaining 2,176 (32%) were found by trained rats, representing 48% (2,176/4,541) overall incremental value. Of which, 65% (1,409/2,176) and 35% (767/2,176) were among Xpert and smear-negative patients, respectively. The Incremental yield by rats at Xpert facilities was 44% (1,409/3,239), while it was 59% (767/1,302) at smear facilities, Odds Ratio, OR = 0.74, 95% confidence interval, CI:0.66–0.82. The rats were 47% more likely to identify TB among Acid Fast Bacilli smear 1 +  or scanty from Xpert-negative samples than smear-negative [91% (1,276/1,409) versus 87% (665/767), OR = 1.47, 95% CI:1.12–1.94]. The difference between children and adults, however, was not substantial, [69% (68/98) versus 67% (1,138/1,701), OR = 1.12, 95% CI:0.72–1.74].

Our finding indicated that trained rats reasonably added benefit over DOT TB case findings. According to our data, switching from smear to Xpert sites lowered the added value by more than a quarter, even though the Xpert sites were likely to benefit more case finding among TB patients with lower bacillary loads.