Culturing stool specimens has no added value in diagnosing pulmonary tuberculosis Authors: Dorte Bek Folkvardsen, Victor Naestholt Dahl, Christian Wejse, Erik Svensson, Troels Lillebaek Published: 2024 ABSTRACT Highlights•Evaluates the benefits of stool testing for diagnosing pulmonary tuberculosis (TB).•Analyzes two years of nationwide data from a low-incidence, resource-rich country.•Findings reveal no additional diagnostic value of stool testing.•Suggests stool testing only, could lead to missed TB diagnoses in similar settings.•Advocates for tailored diagnostic strategies based on healthcare context.Keywords: Tuberculosis, Stool, Culture, Diagnostics, TB low-burden Introduction Patients with pulmonary tuberculosis (PTB) may have difficulties producing sputum for diagnostics. However, Mycobacterium tuberculosis (Mtb) can survive in the gastrointestinal tract and potentially be found in the stool due to swallowed sputum. Accordingly, stool has been suggested as an easy-to-obtain, non-invasive alternative to induced sputum, tracheal suction, and bronchoalveolar lavage[1]. A review from 2022, including 13 studies, mainly from tuberculosis (TB) high-burden countries, reported a high PCR sensitivity for identifying Mtb in stool at 89.7 % (95 % confidence interval 81.4–95.9) using sputum culture as a reference[2]. A more recent review focusing on stool for diagnosing TB in adults[3], concluded that more studies are needed. Culturing of sputum remains the gold standard for PTB testing in most resource-rich settings due to the higher sensitivity and the possibility of getting an extended drug susceptibility profile and data for epidemiological investigations[4]. Stool specimens may be difficult and expensive to culture, as they require a supplement of antibiotic mixtures to the media. The bacteriologic yield is reportedly inferior to other specimen types[5]; however, little data is available on this topic. To investigate stool as a suitable alternative to respiratory specimens to diagnose PTB, primarily by culture, in Denmark, we evaluated all diagnostic specimens routinely sent for mycobacterial investigations at the International Reference Laboratory of Mycobacteriology, Copenhagen, Denmark (IRLM) for two years, from July 1st 2021 to June 30th 2023. Methods A retrospective study was performed on data extracted from the Laboratory Information Management System at IRLM, Statens Serum Institut, Copenhagen, Denmark, including all specimens sent for mycobacterial investigations in the whole country. Denmark is a resource-rich TB low-incidence country. In 2021, the incidence of TB was 3.5/100,000 population, and the MDR-TB burden has been low at 3–5 cases per year[6]. In Denmark, all mycobacteria diagnostics (TB and non-tuberculous mycobacteria) have been centralized to one laboratory in the country, the IRLM, Copenhagen. All primary specimens were decontaminated with the N-Acetyl-l-Cysteine–Sodium Hydroxide method using the BD BBL MycoPrep kit (BD, Franklin Lakes, NJ, USA), a droplet was put on a slide for fluorescence microscopy, and the remains were cultured on liquid (MGIT, BD, Franklin Lakes, NJ, USA) and solid (Lowenstein-Jensen, LJ; SSI Diagnostica, Hillerød, Denmark) media. Stool specimens were cultured on three additional LJ media prepared with antibiotic mixtures. The antibiotic mixtures were allowed to soak into the ready-made LJ medium for 14 days. The mixtures consisted of a) Polymyxin B sulfate; 3000 E/ml, Trimethoprim: 600 µg/ml, Lincomycin: 400 µg/ml and Mycostatin: 1600 E/ml; b) 2: 0,1% Nalidixic acid, Penicillin; 2000 E/ml; c) 3: Same as a), but on an LJ media with pH 5.8. PCR was performed with FlouroType MTBDR (Hain Lifescience GmbH, Nehren, Germany) or Xpert MTB/RIF Ultra (Cepheid, USA), depending on how quickly the clinician requested an answer. Results During the study period, IRLM received 29,774 specimens from 18,337 patients, including 562 (1.9 %) stool specimens from 265 patients (Table 1). Among patients with stool sent, 13 had PTB based on respiratory specimens, but only five were stool Mtb culture-positive (in ten specimens). No patients had Mtb culture-positive stool specimens without accompanying Mtb culture-positive respiratory specimens. Six patients had extra-pulmonary TB diagnosed from specimens other than stool and none of them had positive stool specimens. For 71 patients (26.8 %), only stool was tested for Mtb, all were culture and PCR-negative. Based on all specimens sent for mycobacterial culturing (both stool and other specimens), the majority of patients with stool cultured for Mtb (n = 265) had negative or contaminated results (n = 246/265, 93 %; excluding the 13 patients with culture positive PTB and 6 with extra-pulmonary TB (EPTB) Mtb culture-positive patients, of which 14 were found positive in other specimens than stool only). Table 1Overview of Mycobacterium tuberculosis (Mtb) culture results from 562 stool specimens and from all other types of specimens, among 265 patients investigated for Mtb in stool. NTM: Non-tuberculous mycobacteria. PTB: Pulmonary tuberculosis. EPTB: Extra-pulmonary tuberculosis.ResultNo. of patientsNo. of stool specimensNumber of PCR positive/number of PCR testedNumber of smear positivePTBEPTB(Based on other specimen types)(Based on other specimen types)Mtb positive5102/3750Mtb negative2324602/253275Contaminated26880/39011NTM positive240/2200 Overview of Mycobacterium tuberculosis (Mtb) culture results from 562 stool specimens and from all other types of specimens, among 265 patients investigated for Mtb in stool. NTM: Non-tuberculous mycobacteria. PTB: Pulmonary tuberculosis. EPTB: Extra-pulmonary tuberculosis. PCR was performed on some of the specimens according to IRLḾs routine diagnostic flow, which requires all specimens to be cultured on solid and liquid media. PCR for Mtb is be performed on at least one specimen per patient per year and upon request. Therefore, 297 (53 %) of the 562 stool specimens were also analyzed with PCR following the routine flow. Of the 29,212 non-stool specimens in the study period, 22,979 had a PCR performed. Four of 297 PCR-tested stool specimens were Mtb complex PCR positive; two from Mtb culture-positive stool specimens, and two from culture-negative stool specimens. All four Mtb complex PCR-positive stool specimens were from patients with PTB, diagnosed on other specimen types. The two PCR-positive, culture-negative stool specimens were from patients previously diagnosed with PTB, presumably undergoing anti-TB treatment or already cured. All specimens were investigated with smear microscopy. Only 11 (2 %) of the 562 stool specimens were positive. Seven were culture-positive, 2 were culture-negative, and 2 had positive cultures for non-tuberculous mycobacteria (NTM). Thirty (11.3 %) of the 265 patients with stool specimens sent were less than 18 years old. Two had culture-positive stool specimens, but both also had Mtb culture-positive respiratory specimens; a third with a negative stool specimen had NTM isolated from a lung specimen. Eight of the children had no other specimens than stool sent for investigations; five were culture-negative for mycobacteria, and three cultivated specimens were contaminated by other microorganisms. None of the stool specimens analyzed from children (n = 18/30) were PCR-positive for Mtb. The diagnostic performance of stool culture was calculated on a subset of the data for stool specimens with at least one pulmonary specimen sent (Table 2). Table 2Diagnostic performance of stool culture on a subset of the data, including only stool-specimen with at least one pulmonary specimen sent.Value95 % CISensitivity35.71 %18.64 % to 55.93 %Specificity100.00 %98.73 % to 100.00 %Positive Predictive Value100.00 %69.15 % to 100.00 %Negative Predictive Value94.14 %92.41 % to 95.49 % Diagnostic performance of stool culture on a subset of the data, including only stool-specimen with at least one pulmonary specimen sent. Discussion Examining two years of nationwide stool specimens collected for Mtb testing as part of routine diagnostic procedures, we found that stool culture did not identify additional patients diagnosed with PTB. This suggests that stool is not a suitable alternative to sputum or other respiratory specimens for diagnosing PTB in our TB low incidence setting. All five patients with Mtb culture-positive stool also had positive respiratory specimens. Although the positive predictive value was 100 % the very low sensitivity of 35.7 % and the lack of additional cases detected through stool specimen testing, limits its utility considerably. In contrast, seven patients with PTB (and five with extra-pulmonary TB), based on non-stool specimens, had negative stool tests. Consequently, relying solely on stool specimens could result in missed TB diagnoses, potentially affecting patients such as the 71 examined for TB in our cohort based on stool specimens alone. We highly encourage collecting respiratory specimens if PTB is suspected to ensure the highest diagnostic sensitivity and enable antimicrobial susceptibility testing and genotyping. While stool specimens may offer added diagnostic value in specific subpopulations, such as children unable to produce sputum[7], our study did not demonstrate improved diagnostic yield from stool specimens, with most children being Mtb culture-negative. Only two children (out of 30) had Mtb-positive stool specimens, both with Mtb culture-positive pulmonary specimens. Stool analysis may also have diagnostic value in patients with HIV-TB coinfection and low CD4 counts[8]. Also, we acknowledge the possible diagnostic value of investigating stool specimens if Mtb complex PCR is the only available diagnostic method at hand or in remote areas where more invasive sample procedures can be challenging[8]. This study was based on nationwide routine microbiological data, with PCR data available for only half of the specimens. Therefore, the diagnostic value of PCR was not investigated directly. However, among the PCR-positive specimens, two were also culture-positive while the other two were culture-negative. Both the PCR-positive, culture-negative specimens were from patients with recent Mtb-positive specimens, likely patients undergoing anti-TB treatment. Notably, both microscopy and culturing for mycobacteria on solid and liquid media were performed on all specimens. Only 2 % of the specimens were microscopy-positive, underscoring the limitations of smear microscopy in detecting mycobacteria in stool specimens. Culturing, generally regarded as the most sensitive method for detecting mycobacteria[9]. served as the reference standard in this study. Additionally, culturing is essential in our setting for performing drug susceptibility testing and genotyping, which are crucial for contact investigations and outbreak analysis. Despite some limitations, such as limited HIV/TB and drug-resistant cases in our resource-rich, TB low-incidence setting and reliance on register-based laboratory data, we believe our findings are relevant to similar settings worldwide. Although accessing individual hospital cases and additional patient information could have added value, particularly to elucidate indications for stool testing for the 71 patients with no other type of specimens sent, this was beyond the scope of our study. Conclusion In conclusion, based on two years of nationwide data from Denmark, we discourage the general use of stool for PTB screening in resource-rich, low TB incidence settings. The available data in our study is primarily suitable for comparing culturing across different specimen types. However, we advise against relying solely on stool PCR testing in similar settings, as it misses opportunities for antimicrobial susceptibility testing and contact tracing. Instead, we advocate for prospective trials in specific subpopulations to identify the role of stool as an ancillary diagnostic tool for identifying PTB. Finally, we want to emphasize the importance of distinguishing between different settings when choosing the optimal diagnostic methods, both when it comes to diagnostic tests and use of invasive techniques for obtaining specimens. Author contributions. DBF: Conceptualization, Investigation, Data curation, Writing original draft. VND: Conceptualization, Writing −review and editing. CW: Writing −review and editing. ES: Visualization, Writing −review and editing. TL: Visualization, Supervision, Writing −review and editing. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Ethics statement: The study was purely register-based and did not involve any study participants. For those reasons and according to Danish legislation, scientific ethical approval was not required. CRediT authorship contribution statement Dorte Bek Folkvardsen: Writing – original draft, Investigation, Data curation, Conceptualization. Victor Naestholt Dahl: Writing – review & editing, Conceptualization. Christian Wejse: Writing – review & editing. Erik Svensson: Writing – review & editing, Visualization. Troels Lillebaek: Writing – review & editing, Visualization, Supervision. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References