From April 2019 to December 2020, a total of 159 presumptive TB patients were offered mask sampling, of whom nine declined participation and were therefore not included in the study. In total, 150 patients participated in the study, of which 148 had either “MTB detected” or “MTB not detected” for both the mask and the sputum sample. Two patients with “Invalid” Xpert-mask results were excluded from the final analysis; one had “MTB detected” and the other “MTB not detected” by Xpert testing from sputum. No “Invalid” results were obtained for sputum-based Xpert testing.

Of 148 patients with presumed PTB, 135 (91.2%) were new presumptive TB patients and 13 (8.8%) had received previous anti-TB treatment. Additionally, 10 (7.4%) of these new presumptive TB patients were known contacts of individuals who had confirmed TB. HIV was positive in 16 (11.1%) of 143 tested patients. Sputum smear microscopy identified acid-fast bacilli in 47 (31.8%) patients (Table 1).

Xpert on sputum was positive in 54 (36.5%) patients, of whom Xpert on mask was positive in 26 (17.6%, Table 1).

In four patients, mask results were positive while sputum results were negative. This corresponded to 95% observed agreement and a k value of 0.47, suggesting a moderate level of agreement.

The overall mask sensitivity compared to sputum-Xpert testing was 48.1%, with 95.7% specificity. Xpert sputum result was considered the gold standard for sensitivity calculation, (Table 2).

Considering only the 118 patients for whom samples were processed locally at the LNRM, 48 samples were sputum-Xpert positive of which 29 were positive on mask, showing a sensitivity of 52.1% with 94.2% specificity. Of the 30 patients for whom masks samples were processed at ITM, six were sputum-Xpert in Guinea, of which only 1 had a positive FMS-Xpert result, yielding 16.7% sensitivity only (Table 2).

Sput = sputum; + = positive; - = negative; N = number of samples; Sen = sensitivity; Spec = specificity; ITM = Institute of Tropical Medicine, Antwerp, Belgium; LNRM = Laboratoire National de Référence de Mycobactériologie, Conakry, Guinée.

Stratification of FMS-Xpert results by type of Xpert cartridge, showed a relative higher sensitivity for Xpert Ultra compared to Xpert MTB/RIF. Of the 118 paired samples processed at the LNRM, 70 were performed with Xpert Ultra and 48 with Xpert MTB/RIF. Of the 27 sputum-positive on Xpert Ultra, 17 were mask-positive, while four FMS were positive with a negative sputum-Xpert result. Hence, sensitivity for Xpert Ultra mask testing was 62.9% with 90.6% specificity (Table 3a).

For the 48 patients tested by Xpert MTB/RIF, 21 samples were sputum-Xpert positive of which eight were mask-Xpert positive, showing a lower mask-Xpert sensitivity of 38.0% compared with 100% specificity (Table 3b).

Regarding the detected bacillary load, except for one patient (N°147), all patients with positive Xpert results for both sputum and FMS were found to be positive on smear microscopy. Likewise, except for one patient (N° 73), the bacillary load detected with Xpert in sputum-positive samples was consistently higher than the bacillary load in the respective mask-positive samples, irrespective of the Xpert cartridge type used (Table 4, please consider as supplementary table).

As per protocol and ethical approval, only patients who tested sputum positive were initiated on TB treatment. After one year follow-up of the four patients who tested positive only on FMS but negative on sputum, the one with HIV coinfection was clinically diagnosed with lymph node TB two months after FMS positivity, two improved clinically after a two-week amoxicillin-clavulanic acid treatment and were declared to not have TB disease, while the fourth was lost to follow-up.

Regarding the patient’s history, only 13 (8.8%) patients experienced a previous TB episode, of which 3 were diagnosed sputum-positive by Xpert and initiated on TB treatment. Only one of them was mask-positive on Xpert. No mask-positive/sputum-negative previously treated TB patients were identified.

Among the 16 HIV-coinfected patients, ten yielded a sputum-positive Xpert result and seven a mask-positive Xpert result. Only one had a mask-positive/sputum-negative result.