key: cord-0274287-3411ijbq
authors: de haas, P.; Yenew, B.; Diriba, G.; Amare, M.; Slyzkyi, A.; Demissie, Y.; Sherefdin, B.; Bedru, A.; Mengesha, E.; Dememew, Z. G.; Kebede, A.; Getahun, M.; Tiemersma, E.; Jerene, D.
title: The Simple One-step stool processing method for detection of Pulmonary tuberculosis: a study protocol to assess the robustness, stool storage conditions and sampling strategy for global implementation and scale-up
date: 2022-02-05
journal: nan
DOI: 10.1101/2022.02.04.22270430
sha: cfcece1aa72345c132ce0340c5491c491181e79e
doc_id: 274287
cord_uid: 3411ijbq

Background Xpert MTB/RIF Ultra (Xpert-Ultra) provides timely results with good sensitivity and acceptable specificity with stool samples in children for bacteriological confirmation of tuberculosis (TB). This study aims to optimize the Simple One-Step (SOS) stool processing method for testing stool samples using the Xpert-Ultra in children and adults in selected health facilities in Addis Ababa, Ethiopia. The study is designed to assess the robustness of the SOS stool method, to help fine-tune the practical aspects of performing the test and to provide insights in stool storage conditions and sampling strategies before the method can be implemented and scaled in routine settings in Ethiopia as well as globally. Methods and design The project "painless optimized diagnosis of TB in Ethiopian children" (PODTEC) will be a cross sectional study where three key experiments will be carried out focusing on 1) sampling strategy to investigate if the Xpert-Ultra M.tuberculosis (MTB) -positivity rate depends on stool consistency, and if sensitivity can be increased by taking more than one stool sample from the same participant, or doing multiple tests from the same stool sample, 2) storage conditions to determine how long and at what temperature stool can be stored without losing sensitivity, and 3) optimization of sensitivity and robustness of the SOS stool processing method by varying sample processing steps, stool volume, and sampling from the stool-sample reagent mixture. Stool samples will be collected from participants (children and adults) who are either sputum or naso-gastric aspiration (NGA) and/or stool Xpert-Ultra MTB positive depending on the experiment. Stool samples from these participants, recruited from 22 sites for an ongoing related study, will be utilized for the PODTEC experiments. The sample size is estimated will be 50 participants. We will use EpiData for data entry and Stata for data analysis purposes. The main analyses will include computing the loss or gain in the Xpert-Ultra MTB positivity rate, and rates of unsuccessful test results. The differences in the positivity rate regarding testing more than one sample per child, different storage, and processing conditions, will be compared to the baseline (on-site) Xpert-Ultra result. Ethics and dissemination The protocol was reviewed and approved by the Ethical Review Board of the Ethiopian Public Health Institute (EPHI-IRB) (Protocol no EPHI-IRB-234-2020). The study results will be shared with the national TB program and stakeholders to the benefit of further roll out of the test in a routine Ethiopian setting. The results will also be disseminated in peer-reviewed scientific journals.

specificity with stool samples in children for bacteriological confirmation of tuberculosis (TB). 23 This study aims to optimize the Simple One-Step (SOS) stool processing method for testing stool 24 samples using the Xpert-Ultra in children and adults in selected health facilities in Addis Ababa, 25 Ethiopia. The study is designed to assess the robustness of the SOS stool method, to help fine- 26 tune the practical aspects of performing the test and to provide insights in stool storage 27 conditions and sampling strategies before the method can be implemented and scaled in routine 28 settings in Ethiopia as well as globally. 29 Methods and design 30 The project "painless optimized diagnosis of TB in Ethiopian children" (PODTEC) will be a 31 cross sectional study where three key experiments will be carried out focusing on 1) sampling 32 strategy to investigate if the Xpert-Ultra M.tuberculosis (MTB) -positivity rate depends on stool 33 consistency, and if sensitivity can be increased by taking more than one stool sample from the 34 same participant, or doing multiple tests from the same stool sample, 2) storage conditions to 35 determine how long and at what temperature stool can be stored without losing sensitivity, and 3) 36 optimization of sensitivity and robustness of the SOS stool processing method by varying sample 37 processing steps, stool volume, and sampling from the stool-sample reagent mixture. 38 Stool samples will be collected from participants (children and adults) who are either sputum or 39 naso-gastric aspiration (NGA) and/or stool Xpert-Ultra MTB positive depending on the 40 experiment. Stool samples from these participants, recruited from 22 sites for an ongoing related 41 study, will be utilized for the PODTEC experiments. The sample size is estimated will be 50 42 participants. 43 We will use EpiData for data entry and Stata for data analysis purposes. The main analyses will 44 include computing the loss or gain in the Xpert-Ultra MTB positivity rate, and rates of 45 unsuccessful test results. The differences in the positivity rate regarding testing more than one 46 sample per child, different storage, and processing conditions, will be compared to the baseline 47 (on-site) Xpert-Ultra result.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted February 5, 2022. Ethics and dissemination 50 The protocol was reviewed and approved by the Ethical Review Board of the Ethiopian Public 51 Health Institute (EPHI-IRB) (Protocol no EPHI-IRB-234-2020). The study results will be shared 52 with the national TB program and stakeholders to the benefit of further roll out of the test in a 53 routine Ethiopian setting. The results will also be disseminated in peer-reviewed scientific 54 journals. 55 56 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted February 5, 2022.

Approximately 1.09 million children globally fall ill with tuberculosis (TB) each year,, of whom 59 only 399.000 are notified (1). Every day, nearly 700 children die from TB, 80% of them before 60 reaching their fifth birthday. Treatment exists that could prevent nearly all these deaths, but less 61 than 5% of children get treatment as childhood TB is difficult to diagnose (2). 62 The recent WHO guidelines recommend stool as non-invasive primary diagnostic samples for 63 testing with Xpert and Xpert Ultra for a diagnosis of TB among children (3, 4)based on evidence 64 from multiple studies summarized in three systematic literature reviews ((5), (6), (7). These 65 reviews showed the pooled sensitivities and specificities of stool Xpert against a microbiological 66 (sputum-based) reference standard of between 50 and 67% and 98-99%, respectively. The 67 reviews also showed that there was high heterogeneity in sensitivity which might be partly due to 68 using varying protocols for stool processing, with differences in reagents and methods of 69 homogenization and filtering (5). Moreover, most of the stool processing methods for Xpert 70 testing were rather quite complex and cannot be performed at the lower levels of the healthcare 71 system. This shows that there is a lack of standardized stool preparation and testing protocols and 72 warrants the optimization and standardization of the stool processing methods that can be used at 73 the decentralized level. The KNCV Tuberculosis Foundation (KNCV) and Ethiopian Public 74 Health Institute (EPHI) developed the Simple One-Step (SOS) stool processing method, which 75 can be applied in any GeneXpert laboratory (8). 76 This SOS stool processing method uses similar steps as sputum Xpert testing and does not 77 require additional materials or equipment other than an applicator to pick the correct stool 78 amount for testing. Since the method is as simple as sputum Xpert processing, it can be 79 performed at all sites where a GeneXpert instrument is functional after providing minimal 80 training to the staff involved in Xpert testing(8). 81 The SOS method has shown to be successful, with a low rate of unsuccessful results reported, in 82 a demonstration study conducted in Ethiopia, in which multiple laboratories were involved (8). 83 Furthermore, a head-to-head comparison study, in which the SOS method is compared to other 84 stool processing methods showed similar sensitivity and specificity. The SOS stool method was 85 most suitable for low-resource settings, because of its low error rate, processing time, and 86 minimal requirements regarding biosafety precautions and laboratory equipment (9) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted February 5, 2022. To gain more knowledge and in-depth experience on how the SOS stool processing method with 88 Xpert-Ultra would behave if included in the routine diagnostic pathway for (childhood) TB and 89 rolled out under the national TB program, we aim to further test and optimize the SOS 90 processing method for the detection of TB in stool by Xpert-Ultra and its ability to tolerate 91 perturbations (robustness). The study will also help to fine-tune standard operating procedures 92 (SOPs) for the SOS method. "S1 Figure" . Therefore, MTB positive children who will be identified in the ASTTIE study will 106 also be used in the current study (PODTEC study). We will also recruit adults with MTB 107 detected in sputum from the same facilities. The study was originally planned to be carried out 108 till the end of 2020. However, due to the COVID-19 pandemic, the study period has been 109 extended. health facilities will constitute the study population. Eligible persons, or their caregivers will be 115 requested to sign (parental) consent or assent, depending on the age of the participant. The 116 exclusion criteria include being critically sick i.e., those who are in coma, terminally ill due to 117 chronic debilitating co-morbidities, or other conditions determined to be "critical" by the treating . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 5, 2022. 122 For children, the facility coordinator from the ASTTIE project will daily retrieve stool Xpert- 123 Ultra results from the study sites and checks for eligibility. Parents of eligible children and 124 eligible adults will be asked for informed consent for participation in this study, see "S1 125 appendix". For children, this is an additional consent to the consent already provided for the 126 ASTTIE study. After enrolment in the study, for children, the remainder of the initial stool 127 sample (stool 1) will be collected and transported to EPHI. Participants will be provided with 128 two (children) or three (adults) large stool containers to allow collection of at least 30 grams of 129 stool. They will be instructed on how to collect and store the samples till delivery at the site. The stool samples will be kept in a cold chain at the site until they are transported to EPHI on the 137 same day. The site will inform the study coordinator which will assign a dedicated transporter for increases when more than one sample from the same participant is tested. It also indicates how 150 homogeneous the Mycobacteria are distributed within the stool samples and across different 151 stools. Furthermore, it will provide insight in repeat testing if required due to unsuccessful test 152 result, whether to advise to perform the repeat test from the same stool or from a fresh stool 153 sample. This will be done by testing three stools from the same participant collected during 154 consecutive days (Experiment 1a) as well as three aliquots (North, South and East/West) from 155 the same stool sample (Experiment 1b). For children, two aliquots will be collected from the first 156 stool sample "S1 Figure 1 ". Aliquots will be tested using Xpert-Ultra, totaling a maximum of 157 nine Xpert-Ultra tests per participant. If the Xpert-Ultra result is unsuccessful (i.e., the result is "invalid" or "error"), the test will NOT 185 be repeated as this is part of the study outcome. 186 The SOS stool method's comprehensive instructions can be found on KNCV website (10). In "S3 187 appendix" a schematic overview of the SOS stool method is provided. Depending on the stool 188 consistency the protocol for solid stool or liquid stool is followed. Each stool sample will be submitted to EPHI together with a stool submission form "S2 221 appendix". Details when conducting the experiment are collected on the experiment form "S4 222 appendix". The forms are stored at EPHI. All data will be entered into pre-structured EpiData 223 files (EpiData version 3.1; www.epidata.dk). A random 10% of the data will be re-entered in a 224 separate file to check the quality of data entry. If more than 3% of errors are found in key 225 variables (experiment conditions and Xpert result), full double data entry will be conducted. or increasing amounts of stool added, will be analyzed using Wilcoxon-like test for trend of 242 across ordered groups using nptrend (11). Logistic regression will be applied to assess factors . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Participants' information will be kept confidential, and the digital files used for analysis will 255 only have the PODTEC laboratory code and ASTTIE unique person identification code (UPIC) 256 and will not contain any names or other personal identifying information of the participant. 257 Participants' information will be kept confidential, and the digital files used for analysis will 258 only have the PODTEC laboratory code and the ASTTIE UPIC and will not contain any names 259 or other personal identifying information of the participant. This is the first study protocol in which the sampling strategy and robustness of a stool 263 processing method will be investigated. Based on the experiment's findings, certain steps in the 264 current SOP of the SOS stool method might be adjusted. The experiments will be performed 265 using samples from presumptive TB cases, so for the patients for whom the test will be used in 266 practice in a country with a relatively high TB burden. Uniquely, stool samples will not be 267 spiked with mycobacteria as in other studies (9), (12). We expect that the MTB distribution is 268 different in the stool samples from TB patients than in the spiked stools. Moreover, we will 269 include children who have mostly have paucibacillary TB and who will benefit most from stool 270 Xpert testing, as they cannot easily provide sputum. The study population will be drawn from 271 locations where the test is expected to be conducted in the future, providing more realistic 272 insights in the possibility of implementing the method in routine and collection of multiple . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 5, 2022. with the sample reagent before Xpert testing, variation in the stool portion size picked for 277 processing, and variation in sample processing. The outcome will provide insights of the 278 robustness of the method and will show how far certain steps can be stretched. It will provide 279 practical outcomes that will enable the laboratory personnel and healthcare professionals 280 involved in the stool testing to implement the most optimal protocol. 281 The main results will be presented both at local and international scientific meetings. The results 282 will also be disseminated in the form of peer reviewed publications and as policy briefs. Key 283 audiences for the dissemination will include global scientific advisory group members, local 284 technical advisory committee (TAC) members and NTP. Study host communities will also be 285 informed about the key results of the study through appropriate popular media.

Acknowledgements 288 We would like to acknowledge the Ethiopian Public Health Institute and KNCV Tuberculosis 289 Foundation. We would also like to thank all healthcare facilities that will participate in the study 290 by recruiting the study participants. We would further like to acknowledge Mamush Sahile from 291 KNCV Ethiopia for his assistance in the study. 

World Health Organisation; 2021. 311 2. WHO. Roadmap towards ending TB in children and adolescents

Rapid communication on updated guidance on the management of tuberculosis in children 315 and adolescents. World Health Organisation

Diagnostic Accuracy 317 of Stool Xpert MTB/RIF for Detection of Pulmonary Tuberculosis in Children: a Systematic Review

Diagnostic accuracy of 320 molecular detection of Mycobacterium tuberculosis in pediatric stool samples: A systematic review and 321 meta-analysis

Variable Diagnostic 323 Performance of Stool Xpert in Pediatric Tuberculosis: A Systematic Review and Meta-analysis. Open 324 Forum Infect Dis

The Simple One-Step 326 (SOS) Stool Processing Method for Use with the Xpert MTB/RIF Assay for a Child-Friendly Diagnosis 327 of Tuberculosis Closer to the Point of Care

Laboratory 329 comparison of stool processing methods for Xpert(®) Ultra

implementation package of the SOS stool methode 2021

A Wilcoxon-type test for trend

development of a simple stool sample processing method diagnosis of pediatric tuberculosis using Xpert 335Ultra. Tuberculosis (Edinb). 2020;125:102002. 336 337 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)The copyright holder for this preprint this version posted February 5, 2022.