Extensively drug-resistant organisms (XDRO) are a global threat to health. Colonization with XDRO before hematopoietic cell transplantation (HCT) frequently results in delayed delivery of antimicrobials to which the organisms are susceptible and significantly increases non-relapse mortality. Their inherent resistance to available antimicrobial agents coupled with a preponderance to evolve further resistance makes biological approaches attractive. Suppression of pathogenic organisms by fecal microbiome transplantation has previously been demonstrated, and here we explain in detail the use of this approach to successfully suppress XDRO before HCT that permitted an uneventful transplant course in an otherwise high-risk situation.
Non-relapse mortality (NRM) of allogeneic haematopoietic cell transplantation (HCT) has progressively fallen over the last four decades. Better supportive care, particularly in managing infection, has significantly contributed to the improved safety over this period. However, antimicrobial resistance poses a significant global threat to health,1 and the emergence of extensively drug-resistant organisms (XDRO) within HCT units now poses a direct threat to transplant recipients.2 Gut colonisation with XDRO has been associated with increased NRM,3 and the management of XDRO infections during neutropenic periods is complex and associated with high mortality.2 Innovative approaches in preventing and managing them are therefore necessary to avoid reversing much of the progress made in limiting NRM over the last four decades.
A 63-year-old man presented to our institution with a new diagnosis of Philadelphia-positive acute lymphoblastic leukaemia and received treatment following the UKALLXII trial schedule.4 He achieved complete remission after induction chemotherapy along with imatinib. Following intensification of chemotherapy and continuous imatinib, allogeneic HCT was recommended to consolidate his therapy. His treatment course was complicated by two separate episodes of extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli bloodstream infections, two episodes of Clostridium difficile infection (CDI) and central line-related methicillin-sensitive Staphylococcus aureus bacteremia. Each infection was successfully treated with antimicrobials, but he was subsequently found to be colonised with a highly-resistant GES-5 carbapenemase-producing Enterobacteriaceae (CPE), Klebsiella oxytoca, on routine rectal screening (Table 1).
While gut colonisation with XDRO does not pose any significant risk per se, these organisms can cause opportunistic infections during periods of prolonged neutropenia. Rates of spontaneous clearance of these organisms from colonised individuals are low, even in immunocompetent hosts, ranging from 7-30%.5, 6 Treatment options for elimination of XDRO from their site of origin within the intestine are limited; non-absorbable antimicrobial agents often lead to only transient suppression,5 and may precipitate the development of further resistance. Given the success of donor faecal microbiota transplant (FMT) in the management of recurrent/refractory CDI,7 and the apparently acceptable safety profile when used for CDI in the HCT setting,8 there is considerable interest in the potential role of FMT in gut decontamination prior to HCT. Recipients of FMT for CDI have been shown to have fewer antibiotic-resistant organisms within their gut microbiota following transplantation.9 There are emerging clinical reports of successful use of FMT in gut decontamination of a variety of XDRO (including ESBL and CPE),10 even in the setting of haematological disorders.11 Therefore after discussion, this patient was offered FMT prior to allogeneic HCT in an attempt to eradicate the XDRO and C. difficile from its intestinal niche, with the aim of minimising his HCT NRM.
Following informed consent, the patient received the gut preparation with 4 days of oral vancomycin and neomycin, both 500 mg four times daily. Antibiotics were stopped 24 h prior to FMT delivery, and preparation was completed with iso-osmotic bowel purgatives (Kleen Prep). The unrelated donor stool was pre-screened, and tested negative for C. difficile PCR and toxin, as well as for XDRO; other routine donor screening for transmissible infection was also negative.12 Preparation for the transplant was done immediately after stool donation under strict anaerobic conditions using an adapted version of a previously described protocol,13 and stored at −80 °C until required. The FMT product comprised a thawed slurry of around 100 mL homogenised stool, preserved in a mixture of glycerol and phosphate-buffered saline (15:85, v/v) and was delivered via nasogastric tube. Fasting was instituted 6 h before receiving FMT, and treatment with a proton-pump inhibitor (omeprazole) and pro-kinetic (metoclopramide) was administered 1 h prior to FMT delivery. The patient was allowed to eat and drink normally 1 h post administration. Following the procedure, he experienced mild nausea, loose stool and abdominal discomfort, which were resolved after 24 h without any specific intervention. Repeat rectal screening 7 days after FMT showed continued carriage of the ESBL E. coli but no evidence of GES-5 K. oxytoca CPE or C. difficile. By day 16, after FMT, neither the CPE nor ESBL was detected on rectal screening swabs (Table 1).
Two weeks after FMT, the patient underwent a fludarabine-(30 mg/m2; day−7 to day−3) and melphalan-(140 mg/m2; day−2) conditioned reduced-intensity sibling allogeneic HCT, with standard cyclosporine and methotrexate GvHD prophylaxis. The transplant course was complicated by an episode of neutropenic fever on day +5, with isolation of a fully sensitive Enterococcus faecalis from blood cultures (Table 1). Empirical treatment with piperacillin-tazobactam (4.5 g three times daily), amikacin (15 mg/kg once daily), teicoplanin (12 mg/kg twice daily for three doses, followed by 12 mg/kg once daily) as per local policy with the addition of colistin (3 million units twice daily) resulted in prompt resolution of fever within 24 h, and following isolation of the sensitive organism, antimicrobials were de-escalated to piperacillin-tazobactram and teicoplanin. A second episode of neutropenic fever developed on day +10 and responded to a change in antimicrobials from piperacillin-tazobactam to meropenem (1 g three times daily), and cultures remained sterile. Neutrophil engraftment was achieved on day +25 and the patient was discharged from the hospital on day +29. At day +100 he was well, with no evidence of leukaemia, GvHD or XDRO by rectal screening. At 12-months post-transplant the patient remains well and in remission.
Carbapenemase-producing micro-organisms are now endemic in a number of countries1, 14 and the preponderance of these organisms to extend their resistance spectrum is now contributing to the emergence of strains resistant to our last-resort antimicrobials.15 A paucity in novel antimicrobials means that current approaches are restricted to minimising the risk of XDRO colonisation by antimicrobial stewardship and infection control, as well as managing clinical infection with complex, and often more toxic, antimicrobial schedules. Novel strategies are therefore required, and biological approaches would seem most favourable given the weaknesses of our current pharmacological armoury. Resident gut commensals are adapted to the intestinal microenvironment and have developed complex ecological networks, subsequently becoming interdependent on them. Pathogens are equally reliant on their microenvironment, and competition for critical nutrients, alteration of pH or oxygen tension, and production of toxic metabolites are all mechanisms by which healthy commensals are capable of suppressing pathogens.16 While FMT has been reported in decontamination of XDRO in immunocompromised17 patients and those with blood disorders in earlier studies,11 here we detail our use of this biological approach in the suppression of XDROs in order to minimise NRM prior to allogeneic HCT. Our experience supports the use of FMT in this setting as safe and tolerable, and warrants further study of the efficacy of FMT in a randomised fashion. The suppression of XDRO by FMT before HCT is particularly pertinent because rather than simply identifying an additional risk factor for NRM, the presence of XDROs should been considered as a potentially modifiable risk factor, and this distinction is exceptionally important in risk stratification.
AJI and JFA are supported by the National Institute for Health Research Imperial Biomedical Research Centre. BHM is supported by Imperial College Healthcare Charity (grant number 161722).
AJI, BHM, FD, JRM, EB, JFA and JP conceived and implemented the treatment strategy and prepared the manuscript. BHM performed the procedure with the assistance of FF and GA, and the advice of JRM. All authors reviewed and revised the manuscript before approving the final draft.