Empiric antibiotic use in allogeneic hematopoietic cell transplantation: should we avoid anaerobe coverage?

Yusuke Shono and Marcel R. M. van den Brink


Bacterial infection remains one of the major causes of morbidity and mortality early after allogeneic hematopoietic cell transplantation (allo-HCT) because of a patient’s immunocompromised state requiring use of antibiotic treatments as prophylactic regimens or to treat manifest infections and neutropenic fever (NF).1,2 Based on preclinical studies performed in early 1970s indicating the benefit of gut bacterial decontamination in preventing the development of acute graft-versus-host disease (GVHD),3,4 gut decontamination (GD) by oral antibiotics such as vancomycin, polymyxin B, metronidazole, or ciprofloxacin as a prophylactic strategy is a common, but not standard practice in allo-HCT. Another indication for using empiric antibiotics is to treat NF that occurs in more than 90% of patients.5 Antibiotics, however, are able to damage the commensal gut microflora to various degrees, depending on the antibiotic class.6,7 This becomes of increasing interest as recent studies with advanced DNA sequencing techniques of gut-resident microorganisms begin to provide detailed relationships between the gut microbiota and clinical outcomes of allo-HCT, including overall survival (OS),8 transplant-related mortality (TRM),9 acute GVHD,8-12 relapse,13 and development of bacteremia.14,15 Given the emerging evidence that the preserved diversity of gut microbiota is correlated with better clinical outcomes after allo-HCT,8,9,12 concerns have been raised regarding the practice of GD for prophylaxis as well as the use of broad-spectrum antibiotics for NF that target anaerobic bacteria of the gut (anaerobes).7

Van der Waaij et al described the role of the anaerobes in colonization resistance in the 1970s.16 The article refers to resistance of the host to colonization with a microbial pathogen triggered by the indigenous commensal microflora of the gut. Since then, the anaerobic flora with the major Clostridia and Bacteroidetes genera have been proven repeatedly to prevent long-term colonization by pathogenic bacteria such as Clostridium difficile, Salmonella enterica, or vancomycin-resistant enterococci.17 In addition to pathogen defense, anaerobic Clostridia, a polyphyletic class of the phylum Firmicutes, including Clostridium as well as other similar genera, also play important anti-inflammatory homeostatic roles, including FoxP3 regulation through the production of the short-chain fatty acid.18-20 Reduced abundance of Clostridiales was observed during GVHD,10 and other clinical data indicate that increased abundance of the Blautia genus, which belongs to the class Clostridia, is significantly associated with less GVHD-related mortality and improved OS.12 Another report showed the loss of diversity in association with reduced production of antimicrobial peptides during GVHD.21 Moreover, a small, prospective case cohort study in 15 pediatric allo-HCT patients found an association between bacteria of Clostridia cluster IV and XIVa with anti-inflammatory properties and suppression of GVHD.22 Importantly, treatment with antibiotics results in loss of bacterial diversity; this loss can affect GVHD-related outcomes and OS after allo-HCT.8,14 Loss of bacterial diversity on day 12 or within 7 days of neutrophil engraftment were associated with increased mortality from GVHD.8,12 This emerging knowledge regarding the role of the microbiota (especially the beneficial roles of anaerobes) in the health of allo-HCT patients raises questions as to the practice of bacterial decontamination as well as ideal choices of antibiotics for NF,23 causing a reappraisal of the manner in which antibiotics are used in allo-HCT.

Jones et al showed in 19713 that the absence of gut microbiota resulted in reduced acute GVHD and prolonged survival by using germ-free mice as recipients in allo-HCT models. Similarly, van Bekkum, in 1974,4 reported that allo-HCT host mice kept in continuous germ-free conditions throughout the course of allo-HCT or decontaminated with high-dose antibiotics (neomycin, streptomycin, and/or bacitracin) starting 1 week before transplant significantly reduced mortality from GVHD. A few years later, a clinical report by Storb et al showed that GD and laminar-airflow isolation decreased the cumulative incidence of grade 2 to 3 GVHD in 130 aplastic anemia patients undergoing allo-HCT from HLA-identical siblings.24 These studies formed the fundamental rationale for the practice of GD for acute GVHD prophylaxis in allo-HCT that is still common practice in several HCT centers.

Efforts to evaluate the effect of GD on acute GVHD prevention in allo-HCT patients are summarized in a review by Whangbo et al.23 Limitations of the clinical studies to date evaluating the benefit of GD for acute GVHD prophylaxis are (1) mostly retrospective and single-center studies, (2) lack control groups that do not receive any GD, (3) lack of comparable GD regimens with varying antibiotic regimens between centers, (4) poor patient compliance, and (5) lack of detailed gut microbiome analyses. These concerns leave interpretations of published clinical studies equivocal. To our knowledge, there are 2 prospective randomized trials25,26 among 7 reported clinical trials24-29 of prophylactic GD and their impact on GVHD since 1978; Buckner et al26 showed no effect of GD on the incidence of GVHD, whereas Beelen et al25 showed significant lower incidence of grade 2 to 4 acute GVHD in the GD group. In the latter study, 134 allo-HCT recipients were randomly assigned to receive either ciprofloxacin or ciprofloxacin plus metronidazole as a regimen used for GD for 5 weeks after allo-HCT. Among HLA-identical sibling transplantation, the combination of ciprofloxacin plus metronidazole was associated with significantly less grade 2 to 4 acute GVHD.25 However, the incidence of acute GVHD did not differ significantly for recipients of partially matched related or matched unrelated transplantation. We do not know the exact impact of the metronidazole on the microbiota in this study; a prospective study with advanced 16S sequencing technology would reveal its benefit in GD improving GVHD outcomes (metronidazole might eliminate mucolytic bacteria that contribute to GVHD11). These studies should be performed together with detailed assessment of bacterial susceptibility to antibiotics because the 16S ribosomal RNA sequence cannot provide this information. Guidelines by the joint committee of National Marrow Donor Program/American Society for Blood and Marrow Transplantation/Infectious Diseases Society of America/European Blood and Marrow Transplant Group do not recommend routine GD for allo-HCT candidates with metronidazole because of insufficient evidence.30 Regarding the mixed reports in the effects of GD on GVHD, we have to consider the emergence of antibiotic-resistant enterococci that made successful GD very difficult. Nowadays, fluoroquinolone is widely used for prophylactic purposes and increasing rates of antibiotic-resistant bacteria in allo-HCT patients14,15 that are also associated with increased incidences of acute GVHD.31

To explore the effect of antibiotic prophylaxis not limited to GVHD-related outcomes, a meta-analysis including 17 prospective randomized trials (performed from 1986 to 2012) with 1453 autologous and allo-HCT patients was performed.5 Systemic antibiotic prophylaxis reduced the incidence of febrile episodes as well as infections and bacteremia; however, an effect on all-cause mortality or infection-related mortality was not observed.

There is accumulating evidence that preserving intestinal anaerobic bacteria after allo-HCT may reduce incidence of acute GVHD and TRM and prolong OS, although studies in randomized, prospective method are lacking. A recent retrospective study from Canadian centers32 showed that GD before allo-HCT resulted in higher incidence of grade 2-4 acute GVHD in 500 patients. We reported the effects of different spectrum antibiotics on the intestinal flora and transplant-related outcomes and found that the use of antibiotics with broad-spectrum activity, such as piperacillin-tazobactam, compared with more narrow antibacterial activity, such as cefepime, for treating NF increased GVHD-related mortality in allo-HCT patients and in a murine model of GVHD.11 Administration of antibiotics with broad-spectrum activity resulted in a loss of commensal flora, including Bacteroidetes, Lactobacillus, and Clostridia. As a mechanistic explanation, we found that treatment with broad-spectrum antibiotics in the murine allo-HCT model increased abundance of the mucolytic bacteria Akkermansia muciniphila, contributing to the degradation of the mucus layer in the colon and thereby impairing mucosal homeostasis. Another report confirmed the deleterious effect of antibiotics in a xenogeneic GVHD mouse model resulting in more aggressive GVHD.33 Regarding the timing of systemic antibiotic administration and the impact on GVHD, a retrospective study of 621 patients showed that early exposure to antibiotics (between days −7 and 0) resulted in significant reduction of Clostridiales with higher GVHD-related mortality.34 These results suggest that, in the setting of allo-HCT, strategies to maintain beneficial anaerobes can contribute to better prognosis of patients. Beneficial effect on allo-HCT outcomes by fecal microbiota transplant also advocates the advantage of preserving anaerobes.35-37

Other important reasons for maintaining anaerobes in allo-HCT are discussed under the aspect of immune reconstitution and antitumor activity. Given that the gut microbiota are considered major regulators of immune homeostasis,23,38 microbiota injury may result in impaired immune reconstitution after allo-HCT, making recipients more susceptible to infections. Several studies showed that intestinal microbiota can influence the immune response to systemic cancer chemotherapy, radiotherapy, and immunotherapy,39 and disruption of the gut microbiota is associated with resistance to cancer therapy.40,41 Our retrospective observational analysis of 541 patients undergoing allo-HCT at a single center indicates high abundance of a cluster of bacteria dominated by the anaerobe Eubacterium limosum is correlated with less relapse after allo-HCT.13 Taken together, these findings also support the deliberate choices of antibiotics that protect potentially beneficial anaerobes in allo-HCT patients.

Current clinical guidelines30,42 recommend use of a fluoroquinolone for neutropenic patients with cancer to prevent febrile events and blood infections. Microbiota analysis can assist with optimizing the choice (with a narrow-spectrum antibacterial activity) and duration of antibiotics in allo-HCT patients (shorter as possible). For example, a retrospective analysis in allo-HCT patients compared rifaximin (a poorly absorbed member of the rifamycin family widely used in inflammatory bowel disease43,44 with less activity against Enterobacteriaceae) with ciprofloxacin plus metronidazole found significantly reduced gut GVHD, reduced TRM and improved overall survival.45 Other therapeutic strategies to rescue anaerobes once the microbiota injury occurs are considered in the concept of probiotics (eg, fecal microbiota transplant), prebiotics, and postbiotics (eg, short-chain fatty acid).46 The data are beginning the emerge regarding the composition of the intestinal microbiota ecology and allo-HCT outcomes. To change current standard care routines for allo-HCT, randomized, prospective clinical studies with multiple centers are indispensable. A randomized, phase 2 clinical trial evaluating the impact of GD with vancomycin plus polymyxin B on microbial diversity after allo-HCT is now under way at the Dana-Farber Cancer Institute (NCT02641236), and a phase 2 randomized clinical trial is being conducted at Memorial Sloan Kettering Cancer Center to evaluate the contribution of broad-spectrum antibiotics and microbiota injury to the development of GVHD by comparing piperacillin-tazobactam vs cefepime as a treatment of NF (NCT03078010). The results of these prospective studies might inform us regarding the risks and benefits of specific antibiotic prophylactic regimen in allo-HCT patients.


The authors thank Christoph Stein-Thoeringer and Jonathan U. Peled for their scientific input and editing of the manuscript.

This research was supported by grants from the National Institutes of Health, National Heart, Lung, and Blood Institute (R01-HL069929 and R01-AI100288) (M.R.M.v.d.B.); National Institute of Allergy and Infectious Diseases (R01-AI101406) (M.R.M.v.d.B.); National Cancer Institute (P01-CA023766) (M.R.M.v.d.B.); and a Memorial Sloan Kettering Cancer Center Support Grant/Core Grant (P30 CA008748). Support was also received from The American Society for Blood and Marrow Transplantation (Y.S.), The Lymphoma Foundation, The Susan and Peter Solomon Divisional Genomics Program, Cycle for Survival, and the Parker Institute for Cancer Immunotherapy at Memorial Sloan Kettering Cancer Center.


Contribution: Y.S. and M.R.M.v.d.B. wrote this manuscript.

Conflict-of-interest disclosure: Memorial Sloan Kettering Cancer Center has filed patent applications related to this work (PCT/US2015/062734 entitled “Intestinal Microbiota and GVHD”; inventors include M.R.M.v.d.B. and Y.S.). M.R.M.v.d.B. received research funding from Seres Therapeutics and M.R.M.v.d.B. and Y.S. received licensing fees from Seres.

Correspondence: Marcel R. M. van den Brink, Department of Immunology, Sloan Kettering Institute, Memorial Sloan Kettering Cancer Center, 1275 York Ave, Box 111, New York, NY 10065; e-mail: vandenbm{at}

  • Submitted June 13, 2017.
  • Accepted October 24, 2017.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.