記住我
ORIGINAL ARTICLE
Year : 2022 | Volume
: 11
| Issue : 4 | Page : 371-377
Standardization of a stool concentration method for Mycobacterium tuberculosis detection in the pediatric population
Priya Rajendran1, Baskaran Murugesan1, Sarath Balaji2, Sivakumar Shanmugam1, Sivaraman Palanisamy1, Thirumalani Ramamoorthy1, Sindhu Hasini1, Bella Devaleenal3, Basilea Watson4
1 Department of Bacteriology, ICMR - National Institute for Research in Tuberculosis, Chennai, Tamil Nadu, India
2 Department of Pediatric Pulmonology, Institute of Child Health, Chennai, Tamil Nadu, India
3 Department of Clinical Research, ICMR - National Institute for Research in Tuberculosis, Chennai, Tamil Nadu, India
4 Department of Electronic Data Processing, ICMR - National Institute for Research in Tuberculosis, Chennai, Tamil Nadu, India
Correspondence Address:
Priya Rajendran
ICMR - National Institute for Research in Tuberculosis, No. 1. Mayor Sathyamoorthy Road, Chetpet, Chennai - 600 031, Tamil Nadu
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijmy.ijmy_126_22
Background: The inability of young children to expectorate sputum and paucibacillary status of Mycobacterium tuberculosis (MTB) increases its diagnostic complexity. In this study, we aimed to standardize a stool concentration method for the detection of MTB and its drug resistance by line probe assay (LPA). Methods: The stool from 10 healthy children spiked with H37Rv in five different dilutions (1:1, 1:10, 1:100, 1:1000, and 1:10,000), and stool from 10 confirmed TB and 54 clinically diagnosed TB children were subjected to an in-house stool concentration protocol. All the processed filtrates were subjected to smear microscopy, solid culture, Xpert ultra testing, and LPA. Results: Of 10 control samples, growth was seen in four samples (neat 1:1). In smear microscopy, bacilli could be seen in eight samples (1:1 and 1:10). Xpert ultra testing could detect MTB in eight samples in all dilutions with different loads. LPA could detect MTB in all samples and dilutions. In microbiologically confirmed children, seven out of 10 stool samples tested were positive. Out of 54 children with clinically diagnosed TB, 4 (7.4%) could be confirmed by microbiological diagnosis. Conclusion: The protocol standardized in this study proves to be better working in the molecular detection of MTB.
Keywords: Concentration, line probe assay, pediatric tuberculosis, stool
According to the reports of the WHO, globally in 2019, children accounted for 12% of total tuberculosis (TB) incidence, higher than their share of estimated cases, indicating poorer access to diagnosis and initiation of treatment.[1] With appropriate diagnosis and treatment at the earliest, children generally have good TB treatment outcomes; however, immunological immaturity, especially in young children, leads to rapid progression of TB disease due to missed or delayed diagnosis.[2],[3] In addition to the importance of diagnosis, especially in children, the increase in drug-resistant TB worldwide[1] calls for improved efforts to track all patients at risk of drug resistance, for effective treatment and the prevention of Mycobacterium tuberculosis (MTB) strain transmission.[4],[5] However, the inability of young children to expectorate sputum and the paucibacillary status of the pathogen increases the diagnostic complexity of MTB in the pediatric population.[6]
The substitute to expectorated sputum includes induction of sputum, collection of gastric aspirates, and bronchoalveolar lavage from young children, and is considered to be resource exhaustive and relatively invasive. Moreover, these specimens usually have a low mycobacterial burden, resulting in reticent detection by existing available phenotypic and genotypic methods.[7] Although multiple specimens might improve detection yield,[8],[9],[10] the process turns out to be tedious and costly. Furthermore, the testing typically has to occur over consecutive days, which limits its feasibility, emphasizing the need for alternative specimens for TB detection. It has been documented that although children and infants may be unable to expectorate sputum samples, most of the sputum is swallowed, and TB DNA may remain intact after intestinal passage.[10] Moreover, children with prolonged cough were more likely to have a positive stool polymerase chain reaction (PCR), probably because the chance of sputum production may be higher with increased bronchus infection, which when swallowed by children improves MTB detection rate.[11] Since stool is easily collected from children, it can be used for the detection of MTB, present in swallowed sputum, using both culture and molecular methods.[11],[12],[13],[14] With appropriate preprocessing methods, Xpert testing of stool may prove to be a promising method for the detection of MTB and its rifampicin resistance in the pediatric population.[15],[16],[17] However, the current diagnostic algorithm requires further drug susceptibility testing (DST) for isoniazid and other antituberculous drugs. In this study, we aimed to standardize a stool concentration method for the detection of MTB and its drug resistance to rifampicin and isoniazid by line probe assay (LPA).
MethodsThe study was conducted in the pediatric population aged 3–12 years at ICH, Egmore, as a cross-sectional pilot study in three stages during the years of 2021–2022. Prior informed consent from the parent or guardian of the children was obtained for conducting the study. Oral assent was obtained from children aged above 7 years. The study was approved by the Institutional Ethics Committee (IEC NO: 2019039).
Study definition
Confirmed tuberculosis
The definition of confirmed TB includes MTB confirmed microbiologically (culture or Xpert MTB/RIF assay) from at least one respiratory specimen.[18]
Unconfirmed tuberculosis
The definition of unconfirmed TB or clinically diagnosed TB includes the population that is not confirmed bacteriologically and has at least two of the following:
Symptoms/signs suggestive of TBChest X-ray consistent with TBClose exposure to TB or immunologic evidence of MTB infectionPositive response to TB treatment (requires documented positive clinical response to TB treatment).Sampling
The study samples were categorized as control population (10) to standardize the stool concentration protocol, the confirmed TB population (10) to determine the feasibility of the protocol, and presumptive TB population (54) to validate the protocol. Based on the results of a previous study,[19] the sensitivity of Xpert for the detection of pulmonary TB using a stool sample was 88.9%. Assuming an absolute precision of 10% and a confidence interval of 95%, the required sample size is 38. Accounting for a contamination rate of 40% based on the reference, the required minimum sample size for the presumptive TB population was 54.
Study population
Stage 1: A total of 10 healthy children were recruited as controls, and the isolation rate of MTB from their stool samples was measured by spiking them with H37RV and further subjecting them to concentration technique. The stool samples collected from these children were of soft-solid consistency (Type 4 as per the Bristol stool scale).[20]
Stage 2: A total of 10 children with microbiologically confirmed pulmonary TB from the National Tuberculosis Elimination Program (NTEP) diagnostics were recruited, and their stool samples were tested to assess the feasibility of the concentration technique for isolation of MTB. Of the 10 stool samples collected, three were of semisolid consistency (Type 5), three were of liquid consistency (Type 7), two were of soft-solid consistency (Type 4), one was semisolid with mucus (Type 6), and one was of hard lump consistency (Type 2).
Stage 3: A total of 54 clinically diagnosed TB children were recruited, and their stool samples were tested for validation of the standardized protocol. Sputum samples were also collected from these children to compare the yield between the samples. Of the 54 stool samples collected, 34 were solid (Type 4), 12 semisolid (Type 5), five liquid (Type 7), and three semisolid with mucus consistency (Type 6).
Study procedures
Stage 1
Sample preparation by spiking with H37RV
The in-house MTB control (H37Rv) suspension was used for spiking the control stool samples in five different dilutions of 1:1 (N), 1:10 (N1), 1:100 (N2), 1:1000 (N3), and 1:10,000 (N4) imitating 3+, 2+, 1+, scanty for the last two dilutions, respectively. A volume of 0.2 g of stool sample was suspended in 1 ml of distilled water in 5-ml bijou bottles containing glass beads. The emulsification was done initially by to-and-fro movement of the disposable loop in the bottle and homogenized by vigorous mixing in a vortex.[21]
Sample decontamination
About 5 ml of the stool filtrate was mixed with three volumes of 1% chlorhexidine digluconate (catalog number: C9394), vortexed for 15 min at room temperature, washed in phosphate buffered saline (PBS), and centrifuged at 3000 g for 20 min at room temperature. The pellet was suspended in 1 ml PBS.[22]
Sample concentration and filtration technique
A novel procedure of concentration by sedimentation of human feces using the nonionic detergent polyoxyethylene lauryl ether (Brij-35) demonstrated its efficacy in isolating the trophozoites of Entamoeba histolytica.[23] We implemented this procedure in isolating the MTB from the pediatric stool samples spiked with H37RV.
Preparation of Brij-35 (catalog number: 8019620250)
The stock solution of Brij-35 was prepared at 30% w/v (30 g of Brij-35 in 100 mL of distilled water). A working solution of Brij-35 was prepared by adding 0.5 mL of Brij-35 stock solution to 1000 mL of isotonic saline solution (ISS).
Procedure
The decontaminated sample was homogenized with ISS at a ratio of 1:10 (stool: ISS). The sample was centrifuged for 2 min at 1500 rpm, the supernatant discarded, and the sediment suspended in 8 mL of working solution and centrifuged again at 1500 rpm for 2 min. The sediment was suspended with 0.5–1.0 mL working solution, depending on the amount of sediment obtained. The suspended sediment was filtered using glass wool syringe filters, and the filtered suspension was divided into three parts.
Smear, culture, and Xpert ultra
The first part of the filtered suspension was subjected to Ziehl–Neelsen staining and cultured on Löwenstein–Jensen (LJ) medium that was used as a reference standard. The second part of the suspension was subjected to Xpert ultra testing as the comparator.
DNA extraction
The third part of the filtered suspension was subjected to DNA extraction by QIAamp DNA microbiome kit (catalog number: 51704), and the extracted DNA was subjected to LPA using GenoType MTBDRplus (catalog number: 30496 A) (Hain lifesciences). The extracted DNA was added to the amplification mixture containing AMP A and AMP B and subjected to PCR. The hybridization was carried out using the TwinCubator as per the manufacturer's instructions.
Stage 2
Stool samples from 10 children with microbiologically confirmed pulmonary TB (by Xpert MTB/RIF and TrueNat with a different load of bacilli) [Table 1] in sputum were used in Stage 2. The samples were decontaminated with chlorhexidine and concentrated as mentioned in the Stage I protocol. The processed specimens were subjected to Ziehl–Neelsen staining, solid culture in LJ medium, Xpert ultra testing, and LPA after DNA extraction.
Table 1: Results obtained by different methods in stool samples from children with microbiologically confirmed tuberculosisStage 3
Stool and sputum samples from 54 children (24 females and 30 males) with clinically diagnosed TB were used in Stage 3. Sputum smear was prepared and stained by auramine phenol method. Decontamination of the sputum specimens was done by the NaLC-NaOH method using 1.5% NaOH as the final concentration. The final pellet, thus obtained, was washed with 0.067M of PBS and inoculated into a liquid medium (MGIT culture). The stool samples were processed as described earlier and subjected to Ziehl–Neelsen staining, solid culture in LJ medium, Xpert ultra testing, and LPA after DNA extraction.
ResultsStage 1
Culture
Of 10 control samples, growth was seen in four samples (1:1) at the 3rd week of inoculation with 3–8 colonies in two samples and 1+ growth in two samples [Figure 1]. There was no growth in other dilutions.
Figure 1: 1+ growth of MTB in LJ inoculated with 1:1 dilution of processed stool. MTB: Mycobacterium tuberculosisSmear
In Ziehl–Neelsen-stained smears of stool, bacilli could be seen in four samples at dilutions 1:1 and 1:10 and in four samples at 1:1 dilution with 1+ grading.
Xpert ultra
MTB could be detected in all dilutions of stool with “MTB detected low” in 1:1 and 1:10, “MTB detected very low” in 1:100 and 1:1000, and “MTB detected trace” in 1:10,000 in eight samples. MTB was not detected in lower dilutions (1:1000 and 1:10,000) in two samples [Table 2].
Table 2: Results obtained by different methods of Mycobacterium tuberculosis detection at various dilutions (N<1:1, N1<1:10, N2<1:100, N3<1:1000, and N4<1:10,000) of stool samples from control childrenLine probe assay
LPA testing of DNA extracted from stool samples of all dilutions could detect MTB and its DST pattern to rifampicin and isoniazid as sensitive [Figure 2] and [Table 2].
Figure 2: Banding patterns of samples at different dilutions in LPA. (STBC 09 and STBC 10 control samples spiked with H37RV in different dilutions – N = 1:1, N1 = 1:10, N2 = 1:100, N3 = 1:1000, N4 = 1:10,000). LPA: Line probe assayStage 2
Of 10 microbiologically confirmed TB stool samples tested, three samples were positive by smear, culture, Xpert ultra (2 low and 1 high), and LPA, 1 sample was positive for Xpert ultra (very low) and LPA, and 3 samples were positive by LPA alone. Totally, seven out of 10 stool samples tested were positive by either one of the methods. These 10 samples were collected from children who were positive for MTB in sputum by either GeneXpert (8) or TrueNat (2) [Table 1].
Stage 3
Of 54 stool samples from clinically diagnosed TB children tested, one sample turned out to be positive by smear (scanty), two samples were positive by Xpert ultra (MTB detected very low) and LPA, and one sample was positive for LPA alone. In total, of 54 samples, four were positive for MTB (7.4%). Of these four positive children, one was positive for TB meningitis by Xpert testing in CSF. The sputum samples tested from these 54 children turned out to be MTB negative by all four methods (smear, Xpert, culture, and LPA).
DiscussionWith the given limitations of detecting pediatric TB in respiratory samples, stool may prove to be an alternative sample. However, a major setback in stool testing for MTB that needs attention is the decontamination and the sample processing methods involved in decontamination. Stool samples will require wary and unique processing methods for molecular analysis since PCR inhibitors and particulate matters can interfere with the assay.[16],[24],[25] Finally, in adding up to the improvement of MTB diagnosis in stool, the DST needs an equal consideration due to the increasing incidence of drug-resistant TB globally. Although there are numerous studies demonstrating its usefulness in MTB detection by Xpert testing in stool,[19],[26],[27] there are only limited studies from Indian children. Xpert gives rapid results, including rifampin resistance, but the current diagnostic algorithms on stool samples still require further DST to isoniazid and other antituberculous drugs. To address this lacuna, in this study, we developed a protocol for stool concentration, and the concentrated stool was used for DNA extraction to perform LPA for MTB detection and its drug resistance to rifampicin and isoniazid.
We used stool from healthy controls and spiked them with MTB (H37RV) and could get results by LPA in all dilutions of spiking. We used 0.2 g of stool sample for spiking experiments to ensure that inhibitors present in the stool sample do not hinder the PCR in LPA. However, a study from South Africa in 2016 documented that an increased amount of stool (0.6–1.2 g) improved the sensitivity of GeneXpert testing for MTB without any PCR inhibition.[16] The limited detection limit of MTB by Xpert ultra even in higher dilutions (1:1 and 1:10) could be attributed to the amount of low volume of sample used. However, further studies with a larger sample size are needed to prove that an increased amount of stool will increase the yield of MTB in molecular tests. On contrary, in LPA, we could detect MTB in all dilutions with the same amount of stool sample (0.2 g).
To validate the results obtained in the control samples, we selected the children who had varied results of Xpert testing in sputum; one each with high and medium load, respectively, two samples were of low and four of very low load. In addition, we used stool samples of children who were MTB positive by their sputum TrueNat testing (n = 2). However, we could not retrieve the CFU/ml value of MTB detected by TrueNat in two sputum samples due to machine shifting from the site tested. The stool sample that had a medium load of bacilli in sputum testing turned out to be MTB positive by all four methods tested and indeed had a high load of bacilli by Xpert ultra. On contrary, a child with a high load of bacilli in sputum as detected by Xpert had a low load of bacilli in stool by Xpert ultra testing. This discrepancy in results may be due to the consistency of the stool sample since the retrieval of MTB could be tedious in hard stools. Therefore, we need further large-scale studies to determine the role of stool consistency in MTB detection.
Of the four samples with MTB detected very low by sputum Xpert testing, none of them could be retrieved by Xpert ultra, but three could be retrieved by LPA in the stool sample. This increased detection ability could be attributed to the further processing of stool filtrates by proteinase K and an efficient DNA extraction protocol. Studies have reported that preprocessing of stool specimens with proteinase K leads to a better yield of DNA during extraction.[28] This is facilitated by the enzyme's ability to degradation of proteins present in the human sample by cleaving peptide bonds at the carboxyl sides of amino acids present in the sample. Surprisingly, the stool samples of the patients with TrueNat positive for MTB by sputum turned out to be negative by all the methods tested.
Of 54 clinically diagnosed with TB, four children were MTB-positive in stool, while none of them were positive by sputum testing. Of the four positives, two were by both LPA and GeneXpert, 1 by LPA, and 1 by smear. All the LPA-positive samples were sensitive to rifampicin and isoniazid. The sample that was positive by LPA and not by Xpert ultra could be due to its testing ability in the paucibacillary sample. Testing of paired samples could have improved MTB detection since it has been reported that Xpert's sensitivity was increased to 70% from 44% when a paired sample was used.[15] We could not retrieve MTB by culture in any of the samples. It is to be noted that the yield of culture is reported to be <40% in the pediatric population, with slightly higher sensitivity and faster results in liquid mycobacterial growth in tubes compared with solid culture media.[29] Studies conducted in 1996 and 2011 also demonstrated a lesser sensitivity of stool culture in pediatric stool samples with a higher contamination rate.[30],[31] In the future, the use of a stringent decontamination protocol and appropriate processing method for subjecting stool to culture might improve its sensitivity.
ConclusionTo our knowledge, this is the first study that used stool samples for LPA testing by standardization of a concentration technique that has improved the MTB detection yield. As part of the NTEP, national reference laboratories act as service providers for DST by LPA for both first-line and second-line drugs. With the difficulty faced in confirming TB disease in the pediatric population using the current diagnostic algorithm, stool may be used as an alternative sample. The proposed processing method in this study would improve the yield of MTB-DNA for LPA testing. However, a study with a larger sample size and simultaneous comparison with a sputum sample is needed before scaling up or confirming the validity of our findings before incorporating them into the program.
Limitations of the study
Statistical analysis for sensitivity and specificity calculation was not done due to the minimal sample size since we did this as a pilot study to standardize the developed protocol.Paired sample testing was not done due to the reluctance of parents in providing stool samples.Ethical statement
Instituitional ethics committee of ICMR-National Institute for research in tuberculosis.
Consent form
Children aged more than 7 years were explained about the study in detail, and the assent form was signed by them with their willingness. Parents of children aged <7 years were explained about the study and requested to sign the consent form once their willingness was confirmed. The benefits and risks associated with the study were clearly explained. There was no risk to the participants during sample collection.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
留言 (0)