Stool banking for fecal microbiota transplantation: methods and operations at a large stool bank

Objectives: Fecal microbiota transplantation (FMT) is a recommended therapy for recurrent Clostridioides difficile infection and is being investigated as a potential therapy for dozens of microbiome-mediated indications. Stool banks centralize FMT donor screening and FMT material preparation with the goal of improving the safety, quality, convenience, and accessibility of FMT material. Although there are published consensuses on donor screening guidelines, there are few reports about the implementation of those guidelines in functioning stool banks. Methods: To help inform consensus standards with data gathered from real-world settings and, in turn, to improve patient care, here we describe the general methodology used in 2018 by OpenBiome, a large stool bank, and its outputs in that year. Results: In 2018, the stool bank received 7,536 stool donations from 210 donors, a daily average of 20.6 donations, and processed 4,271 of those donations into FMT preparations. The median time a screened and enrolled stool donor actively donated stool was 5.8 months. The median time between the manufacture of an FMT preparation and its shipment to a hospital or physician was 8.9 months. Half of the stool bank's partner hospitals and physicians ordered an average of 0.75 or fewer FMT preparations per month. Conclusions: Further knowledge sharing should help inform refinements of stool banking guidelines and best practices.


INTRODUCTION 23
Fecal microbiota transplantation (FMT), the transfer of minimally manipulated stool and its 24 associated microbiota from a healthy donor into the gastrointestinal tract of the patient, 1,2 is a 25 recommended investigational therapy for recurrent Clostridioides difficile infection. 3-7 C. difficile 26 infection is the most common healthcare-associated infection in the United States, with 462,000 27 cases and 20,500 deaths in the US in 2017. 8,9 FMT's reported safety profile and efficacy in 28 preventing recurrence of C. difficile infection, approximately 80-90%, 10-18 has inspired research 29 using FMT to treat a wide range of microbiome-mediated indications. 2,19 30 31 There are two main models supplying stool for FMT: patient-selected donors and stool 32 banking. 20-25 33 34 Under the patient-selected donor model, the patient or their guardian identifies their own stool 35 donor candidate. The treating physician screens the candidate and processes the donor's stool 36 into an FMT preparation. The donor typically donates material for only that single patient. This 37 approach places substantial logistical burden on the physician 26 and creates delays between the 38 determination that FMT is indicated and the delivery of therapy. For example, if a patient's first 39 candidate donor fails the screen, another must be found, who may also fail the screen, all 40 before the patient can be treated. 27,28 Furthermore, the patient-directed model poses certain 41 risks. First, different practitioners may use variable screening standards, potentially exposing 42 the patient to substandard screening. Second, the donor stool may be processed into an FMT 43 treatment in an uncontrolled, ad hoc workspace, like a physician's office, increasing the risk of 44 contamination. Barriers to prompt access to FMT have also been reported as reasons for 45 patients to seek "do-it-yourself" (DIY) FMT, which comes with significant risk to patient safety. 29 46 47 . 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 September 9, 2020. ; https://doi. org/10.1101org/10. /2020 candidate to review and clarify their answers. Next, the clinical staff member measures the 126 candidate's body temperature, blood pressure, heart rate, respiratory rate, body mass index, 127 and waist circumference. Finally, a supervising physician reviews the questionnaire and vital 128 signs. 129

130
Candidates who pass the in-person clinical assessment undergo a three-part laboratory 131 screening: blood, stool, and nasal swab. Samples are sent to external laboratories for testing 132 (Table 1), including tests suggested by the European FMT Workgroup guidance and 133 international consensus recommendations. 31,41 Any abnormality is reviewed by the bank's 134 supervising clinician. Candidates who pass the laboratory screens are accepted into the stool 135 donation program. Candidates who fail the laboratory screens are either temporarily deferred or 136 permanently excluded from the program. For example, if any screened multi-drug resistant 137 organism is detected, the candidate is permanently excluded as a stool donor, but a patient 138 carrying a transient enteric pathogen may be invited to re-screen after a temporary deferral. 139 Candidates and donors are informed of any significant incidental findings observed during the 140 health evaluation and monitoring process and referred to the appropriate healthcare services. 141 To protect the security and quality of the donor program, candidates are generally not informed 142 about the reason for their deferral or exclusion other than to share significant incidental findings. 143 144

145
Donors provide donations by visiting the stool bank's collection facility, where they are given a 146 stool collection kit. After passing the stool on site, donors close the container's lid and place it in 147 a resealable plastic bag for secondary containment. A staff member labels the donation with a 148 donor identification number and the time of passage. The donor's health status is re-assessed 149 at each donation as described below. Donors are remunerated for the time and travel required 150 to provide each processed donation. 151 . 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 September 9, 2020. The bank produces two liquid preparations and one capsule preparation. The first liquid 170 preparation is 250 mL diluted at a 10:1 ratio (i.e., approximately 22.7 g of stool per preparation) 171 and is intended for delivery by colonoscopy, sigmoidoscopy, or enema delivery. The second 172 liquid preparation is 30 mL diluted at a 5:2 ratio (i.e., approximately 8.6 g of stool) and is 173 intended for delivery by esophagogastroduodenoscopy or nasoenteric tube. Liquid FMT 174 preparations are transferred to sterile polyethylene terephthalate bottles using sterile, 175 disposable serological pipettes. 176 177 . 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 September 9, 2020. ; https://doi. org/10.1101org/10. /2020 The capsule formulation is designed to be swallowed and resist degradation from stomach acid 178 until reaching the small intestine. 43  including fever, cough, congestion, change in bowel habit, nausea, vomiting, seasonal allergies, 197 medical or dental procedures, bodily injury, and use of oral or topical over-the-counter 198 medications. If an illness or travel event is reported, the donor is further evaluated by a member 199 of the clinical staff, who determines if the donor should be temporarily deferred or permanently 200 excluded from the stool donation program. For example, travel to certain countries entails a high 201 risk for acquiring antibiotic-resistant bacteria and may lead to temporary deferral or permanent 202 . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint exclusion. Any donor who is temporarily deferred undergoes a partial or complete re-screening 203 before being re-admitted to the program. Fourth, donors must agree to undergo random health checks. A clinician examines the donor's 220 vital signs and assesses the donor's health status, with a focus on bowel habits, infectious risk 221 factors, and new behaviors that may impact the microbiome. If the clinician detects any clinical 222 concerns, the donor may be temporarily deferred or permanently excluded. Any donor who is 223 temporarily deferred undergoes a partial or complete re-screening before being re-admitted to 224 the program. 225 226 Finally, every donor repeats the complete set of clinical and laboratory assessments, the same 227 set that they passed when first enrolling as a donor, approximately every 60 days. These re-228 . 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 September 9, 2020. assessments "bookend" 60-day collection periods. FMT preparations produced from stool 229 donations during a collection period are released from quarantine only after the donor passes 230 the second full assessment at the end of the collection period. It is not the case that every 231 donation is screened. Instead, the assessments of a donor's health at the beginning and end of 232 a 60-day period and a review of all clinical data collected during the collection window are taken 233 as sufficient evidence that the intervening donations are fit for use as FMT preparations. 24 infections are transient, and donors are temporarily deferred. As stated above, temporarily 253 . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint deferred donors must undergo a partial or complete re-screening before being re-admitted to 254 the program. 255 256

257
The stool bank provides FMT preparations to gastroenterologists and infectious disease 258 physicians treating patients with recurrent C. difficile infection not responsive to standard 259 therapies. 37 The bank also fulfills requests for FMT material for research purposes, but 260 procedures for fulfillment and patient safety monitoring in that case differ. For clarity, only the 261 enforcement discretion procedures are described here. 262

263
The stool bank's website provides information on FMT regulations in the United States, 264 description of FMT treatment modalities, and clinical guidance, as well as registration forms that 265 interested facilities must complete and submit in order to receive material from the bank. 266 Institutional shipping and billing information, as well as contact information for material control, 267 the overseeing physician, and adverse event reporting, are collected during registration. When 268 registered partners submit a written purchase order, FMT preparations are removed from 269 storage at -80 °C and shipped via overnight air on dry ice in Styrofoam containers. Each 270 container is shipped with a temperature indicator verifying that the preparations remained frozen 271 during shipping. 272 273 Registered physicians or healthcare facilities who received stool bank material submit three 274 kinds of information back to the stool bank. First, they must complete a log confirming that each 275 FMT preparation was frozen upon arrival and report whether it remains in inventory, was used in 276 treatment, or destroyed due to expiration or other reasons. Second, they are asked to complete 277 a clinical follow-up form for each patient within 8 weeks of the FMT procedure, indicating the 278 severity and subtype of C. difficile infection that was treated and the patient's outcome following 279 . 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 September 9, 2020.

285
When a clinician reports a serious adverse event to the bank, the bank begins an investigation. 286 Safety staff appraise the event with respect to seriousness, severity, expectedness, and 287 relatedness. Safety staff engage subject matter experts, including the stool bank's clinical 288 advisory board, which is an independent group of gastroenterologists, infectious diseases 289 specialists, and other subject matter experts. 290 291 Barcoding, tracking, and quality assurance systems allow the bank to identify the donation and 292 donor associated with an adverse event. If the adverse event presents a risk to other individuals 293 who would receive material from the same donor, material from that donor is immediately 294 placed in quarantine, and no more is shipped until safety and quality staff have determined that 295 it is safe to do so. If the event involves an infectious pathogen, safety and quality staff can 296 retrieve the safety aliquot taken from that donation and test it for the presence of the pathogen. 297 Depending on the investigation's findings, material from the donor may be destroyed or recalled. 298 The donor may be permanently excluded from the donation program. 299 300 Findings from the investigation are entered into a safety database with MedDRA terminology. 301 (MedDRA, the Medical Dictionary for Regulatory Activities terminology, is the international 302 medical terminology developed under the auspices of the International Council for 303 Harmonisation of Technical Requirements for Pharmaceuticals for Human Use.) Data are 304 analyzed through an analysis of similar events to detect safety signals which may indicate a 305 . 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 September 9, 2020. Sciences or the MedWatch 46 system. Regular reviews of safety data inform additional safety 308 practices and risk management strategies in a continuous improvement process. purposes, to assess the donors' health. 36% of donations (2,749/7,536) were rejected due to 321 the donation's low weight, poor Bristol stool score, visual stool pathology, or because the 322 donation could not be processed within 6 hours of passage. The remaining 57% of donations 323 (4,271/7,536) were processed into liquid or capsule FMT preparations ( Figure 3A). Donors 324 varied in their productivity and the proportion of their samples that were rejected or processed 325 ( Figure 3B). The time between a donation's passage and when it was processed into an FMT 326 formulation varied but was always less than 6 hours ( Figure 3C 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint quarantine releases, very few FMT preparations are shipped less than 4 months after they are 331 produced. The median time between production and fulfillment was 8.9 months ( Figure 3D We shared these methods to help inform consensus standards with data gathered from real-352 world settings, to support the advancement of similar operations, and to invite broader 353 participation in the improvement of these methods. To help explore the thematic areas in which 354 stool banking protocols could be improved and adapted, we lay out and discuss five scenarios 355 about a hypothetical donor's lifecycle. 356 . 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 September 9, 2020. Third, because stool from any donor in the bank's program appears equally efficacious, the 379 stool bank does not direct material from particular donors to particular patients except as part of 380 specific research protocols. However, it may at some point become clear that individual donors 381 ("super donors") 60 or donations with particular characteristics ("superstool") 61 yield more 382 . 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 September 9, 2020. 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint reliably determines whether a donor's material carries the minimum infective dose of a 409 pathogen. In practice, the fact that a donor or their material screens negative for a pathogen on 410 an available assay does not fully negate the risk that their donated material carries that 411 pathogen, and patients should be counselled on the risk of acquiring pathogens during the 412 informed consent process. 413 414 Scenario 2: A donor fails the initial screen. 415 In this scenario, the potential donor is excluded during the pre-screen survey, on-site evaluation 416 by a clinical staff, or laboratory testing. The candidate never provides a stool sample. This 417 scenario highlights two major areas of ongoing development. 418 419 First, just as blood donation regulations aim to reduce risk "to the lowest level reasonably 420 achievable without unduly decreasing the availability of [blood]", 70 so stool donor screening 421 procedures aim to limit the risk of FMT without unduly restricting access to FMT. However, the 422 optimal screening battery remains an area of active development, and screening for every 423 possible pathogen may not be the best approach. For example, given the high prevalence of 424 prior exposure to Epstein-Barr virus (EBV) and cytomegalovirus (CMV) as well as the lack of 425 reported patient events related to transmission of these viruses via FMT, international guidelines 426 do not recommend that donors be excluded based on their exposure to these viruses. 31 Instead, 427 patients at risk of CMV or EBV infection should therefore be appropriately counselled on the 428 risks, and alternatives to FMT should be considered. The label on the bank's shipped material 429 includes a disclaimer to this effect. 430 431 Second, the optimal screening battery may be different for different patient populations or 432 geographies. As a hypothetical example, it may become clear that testing for some pathogen or 433 risk factor is important for FMT safety, but only when used in a certain patient population, such 434 . 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 September 9, 2020. In this scenario, a donor self-reports, tests positive for, or is diagnosed with an acute infection, 459 such as a viral upper respiratory infection or acute gastrointestinal illness. The donor is 460 . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint temporarily deferred and any material from the current stool collection period is destroyed. The 461 deferral is maintained at least until the donor is asymptomatic, usual stool patterns have 462 returned, and the donor passes a partial or complete re-screening, which includes tests collect large amounts of real-world data 76 that may be more representative of relevant patient 486 . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint populations than data collected from clinical trials, 77 but improved reporting of adverse events to 487 stool banks remains an important challenge for the field. Data from each of these 3 sources are 488 valuable and complementary, and they should be used in concert to improve donor screening 489 and manufacturing protocols for all human-derived microbial therapeutics. Optimal stool banking methods likely depend on the target patient population, evolving 509 regulations, and emerging scientific understanding about the microbiome. Here we described 510 and discussed the stool banking protocol used by a large stool bank in 2018 to illustrate current 511 . 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 September 9, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 practice and to highlight areas for potential improvement. We hope this report is a step toward 512 analyses that can help inform evidence-based stool banking protocols. No external funding was received for this work. 537 . 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 September 9, 2020. 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 September 9, 2020. factors that applied to them. 776 . 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 September 9, 2020. During 2018, 120 donors made at least 1 donation that was processed into an FMT preparation. 780 For each of those donors, enrollment duration was calculated as the time between a donor's 781 first donation (which may have taken place before 2018) and their last donation in 2018. The 782 minimum enrollment duration was 21 days, and the maximum was 3.4 years. The median 783 enrollment duration was 5.8 months (interquartile range 3.2 months to 12.1 months). 784 . 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 September 9, 2020. were processed within six hours of passage. The fastest time to processing was 47 minutes. 801 The longest time was 5.99 hours. The median processing time was 3.9 hours (interquartile 802 range 3.2 to 4.7 hours). 803 (D) By 1 July 2020, the bank had shipped 15,323 of the FMT preparations produced in 2018. 804 Preparations were shipped between 2.2 and 28 months after production. The median time to 805 shipment was 8.9 months (interquartile range 6.3 to 13.3 months). Because not all material 806 produced in 2018 had been shipped as of 1 July 2020, the data are skewed towards earlier 807 shipment. 808 809 . 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 September 9, 2020. The number of preparations per month was computed by dividing the total number of shipped 812 units by 12. The median number of ordered preparations across physicians/hospitals was 9 per 813 year, or 0.75 treatments per month (interquartile range 0.25 months to 1.4 months). The 814 maximum number of preparations ordered was 11.75 per month. The majority of 815 physicians/hospitals ordered an average of less than 1 FMT preparation per month. 816 . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint . 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 September 9, 2020. ; https://doi.org/10.1101/2020.09.03.20187583 doi: medRxiv preprint