Usefulness of gastric aspirates for diagnosing nontuberculous mycobacteriosis

Distinguishing between nontuberculous mycobacterial pulmonary disease (NTM-PD) and pulmonary tuberculosis (TB) is difficult. We aimed to evaluate the usefulness of gastric aspirate examination for NTM-PD diagnosis and for differentiating NTM-PD from other diseases, including pulmonary TB. We retrospectively collected data for 491 patients with negative sputum smears or a lack of sputum production at Fukujuji Hospital. We compared 31 patients with NTM-PD to 218 patients with other diseases (excluding 203 with pulmonary TB). Additionally, we compared 81 patients with NTM cultured from at least one sputum or bronchoscopy sample to the other 410 patients. Gastric aspirate examination for NTM-PD diagnosis showed 74.2% sensitivity and 99.0% specificity for culture positivity. There was no significant difference between the nodular bronchiectatic disease and cavitary disease types for culture positivity (p = 0.515). The significance of NTM isolation from gastric aspirate showed 64.2% sensitivity and 99.8% specificity for culture positivity. Gastric aspirate examination revealed NTM in one TB patient, allowing TB to be ruled out in 98.1% of patients with NTM cultured from gastric aspirates. Gastric aspirate examination is helpful for early-stage NTM diagnosis and ruling out pulmonary TB. This could lead to more accurate and timely treatment.

www.nature.com/scientificreports/ one patients with a positive smear on the 1st sputum examination, eight patients who underwent gastric aspirate examination after receiving anti-TB drugs, 22 patients with a clinical diagnosis of NTM regardless of sputum and gastric aspirate culture negativity, and 35 patients who were undiagnosed and lost to follow-up without a definitive diagnosis were excluded. In total, the data of 491 patients were reviewed. The baseline characteristics of the patients are shown in Table 1 The other disease  group included 203 patients with pulmonary TB, 58 with pneumonia, 36 with malignant diseases, 30 with radiological findings compatible with cured TB, 27 with inflammatory lesions, 15 with diffuse interstitial lung disease,  12 with bronchiectasis, 6 with mycosis, and 23 with other lung diseases.  www.nature.com/scientificreports/ Subject for comparison. To investigate whether gastric aspirate helps to diagnose NTM-PD, 39 patients who did not meet the NTM-PD diagnostic criteria 7 were also excluded, as were 11 of 81 patients included in the NTM culture positivity group; therefore, 31 patients with NTM-PD and 421 patients in the other diseases group were compared (Comparison 1). Next, the NTM-PD patients were classified into nodular bronchiectatic (NB) type and cavitary disease (cavitary NB + FC) based on chest computed tomography (CT) 1 . We targeted NB type and cavitary disease because sputum examination in patients with the cavitary disease type is expected to identify NTM specimens more frequently than that in patients with the NB type (Comparison 2) 8 . Finally, to investigate whether gastric aspirate can distinguish patients with gastric aspirate-isolated NTM species from patients with other diseases, including TB, eighty-one patients in the NTM culture positivity group were compared to 410 patients in the other disease group (Comparison 3). For all patients, the gastric aspirate in the acid-fast bacillus (AFB) smears was subjected to fluorescence staining. Samples from 487 patients were tested using mycobacterial growth indicator tubes (MGITs), and samples from 4 patients were tested using solid medium (egg-based Ogawa) for AFB culture of the gastric aspirate. Samples were incubated at 35 °C in MGITs for 6 weeks and in solid medium for 8 weeks.
Procedure for collecting gastric aspirate. The collection of gastric aspirate was conducted in a negative pressure room. A sterile, flexible catheter was passed through the patient's nose into the stomach under aseptic precautions, and gastric aspirate was obtained by applying negative pressure with a syringe and transferred to a sterile container. We did not apply irrigation water during the procedure. Prior to the procedure, patients were required to fast for at least 8-10 h. Statistical methods. All data were analysed and processed using EZR, version 1.53 9 . The Mann-Whitney U test and Fisher's exact test were used for group comparisons. The sensitivity, specificity, and odds ratio values were calculated. The McNemar test was used to compare the utility of gastric aspirates and the 1st sputum culture. The level of statistical significance was set at p = 0.05 (2-tailed).

Results
The The diagnostic accuracy of gastric aspirate for NTM-PD was as follows: 38.7% sensitivity and 88.8% specificity for smear positivity and 74.2% sensitivity and 99.0% specificity for culture positivity. The odds ratios of gastric aspirate smear and culture for the diagnosis of NTM-PD were 5.00 (95% Cl 2.07-11.7) and 276.5 (95% Cl 74.6-1356.5), respectively ( Table 2). The sensitivity of the 1st sputum culture for patients with NTM-PD was 64.5%, which was lower than that of gastric aspirates, but there was no significant difference (p = 0.606). Among 31 patients with NTM-PD, 1 patient was diagnosed by bronchoscopy, and 5 patients were diagnosed by 4 or more sputum examinations. In Comparison 3, the significance of gastric aspirate for NTM cultured from at least one sputum or bronchoscopy sample was as follows: 32.1% sensitivity and 88.5% specificity for smear positivity and 64.2% sensitivity Table 2. Diagnostic accuracy of gastric aspirate examination for NTM-PD. NTM Nontuberculous mycobacteria.

Discussion
This study first demonstrates the usefulness of gastric aspirates for NTM-PD diagnosis at an early stage among cases with first smear negativity or those without sputum production. By comparing CT findings, gastric aspiration was useful regardless of the presence or absence of cavity lesions. In addition, it also indicated that NTM isolation from gastric aspirate can distinguish patients from patients with other diseases, including pulmonary TB. Gastric aspirate can be useful for diagnosing NTM-PD because of its high sensitivity, specificity, and odds ratio. Generally, gastric aspirates are not used for diagnosing NTM-PD because the diagnostic criteria do not include gastric aspirate evaluations 7 . However, in our study, gastric aspirate in NTM-PD was identified as concordant with sputum examination except for one patient who showed different NTM species between the gastric aspirate and sputum, which is consistent with a previous report 10 . Furthermore, NTM-causing organisms were cultured from gastric aspirates of 78.3% (18 of 23 patients) for whom three repeated sputum samples were negative, and 33.3% (6 patients) of those cases fulfilled the diagnostic criteria. Accordingly, the present study indicated that gastric aspirate could be one of the diagnostic tools for NTM-PD for whom diagnosis is difficult based on sputum assessment.
In contrast, when patients swallow contaminated water, their gastric aspirate results can even be smearpositive for AFB regardless of their sputum results 10 ; hence, NTM can be isolated due to environmental contamination 7 . Therefore, gastric aspirate for diagnosing NTM should be used in combination with sputum examination to confirm the concordance. Additionally, patients with the NB type generally have a lower bacterial load than those with the cavitary disease type 1 . However, our data showed no significant difference in the NTM positivity rate between the NB type and cavitary type; therefore, it seems that gastric aspirate might be helpful for diagnosing mild NB-type.
The present study also demonstrated the significance of occasional isolation of NTM from gastric aspirates. The study investigated patients who were suspected of having mycobacterial infection but whose first sputum smear was negative or who did not produce sputum, making it difficult to diagnose whether they had pulmonary TB. Our results showed that false-positive gastric aspirates of patients with pulmonary TB were very low (4.9%), similar to sputum examination. Therefore, we believe that gastric aspirate could be used instead of sputum in patients who are suspected of having mycobacterial infection but who cannot provide a sputum sample for examination. If gastric aspirate identifies NTM-causing species, pulmonary TB can be ruled out with high probability, and those patients can be monitored without more invasive examinations such as bronchoscopy or empiric anti-tuberculous treatment. Furthermore, in the present study, six of the 29 patients (21%) who underwent three repeated sputum examinations tested negative, but those for whom gastric aspirate showed a positive result were ultimately diagnosed with NTM-PD by bronchoscopy or by 4 or more sputum examinations.
This investigation had several limitations. The study was conducted retrospectively at a single centre. Gastric aspirates were collected from patients suspected of having a mycobacterial infection. However, the selection of  www.nature.com/scientificreports/ patients who underwent gastric aspirates was made by the clinician, even though it was based on patients with difficulty in obtaining sputum, resulting in variations in patient backgrounds and timing of sample collection. Second, gastric aspirate collection is usually performed in a negative pressure private room in our hospital; however, the risk of contamination with NTM might be increased if gastric aspirate collection is performed where water pollution is present. Therefore, data may vary depending on local water quality. Last, gastric aspiration was performed once in all patients; thus, repeated positive cultures from gastric aspiration were not confirmed. Validation among clinical settings will be needed. Finally, nucleic acid amplification tests of the M. avium complex were not performed.

Conclusions
Gastric aspirate examination is useful for diagnosing NTM-PD at an early stage regardless of the radiological presentation, and it can also help to rule out pulmonary TB in patients at an early disease stage if gastric aspirate shows NTM positivity.

Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.