Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 May 1.
Published in final edited form as: Biol Blood Marrow Transplant. 2017 Feb 14;23(5):845–852. doi: 10.1016/j.bbmt.2017.02.006

Microbiota disruption induced by early use of broad spectrum antibiotics is an independent risk factor of outcome after allogeneic stem cell transplantation

Daniela Weber 1, Robert R Jenq 2, Jonathan U Peled 3, Ying Taur 4, Andreas Hiergeist 5, Josef Koestler 5, Katja Dettmer 6, Markus Weber 7, Daniel Wolff 1, Joachim Hahn 1, Eric G Pamer 4, Wolfgang Herr 1, André Gessner 5, Peter J Oefner 6, Marcel RM van den Brink 3, Ernst Holler 1
PMCID: PMC5546237  NIHMSID: NIHMS887671  PMID: 28232086

Abstract

In allogeneic stem cell transplantation (ASCT), systemic broad-spectrum antibiotics are frequently used for treatment of infectious complications, but their effect on microbiota composition is still poorly understood. Here, in a retrospective analysis of 621 patients, who underwent ASCT at the University Medical Center of Regensburg and Memorial Sloan Kettering Cancer Center in New York, we assessed the impact of timing of peri-transplant antibiotic treatment on intestinal microbiota composition as well as transplant-related mortality (TRM) and overall survival. Early exposure to antibiotics was associated with lower urinary 3-indoxyl sulfate levels (p<0.001) and a decrease in fecal abundance of commensal Clostridiales (p=0.03) compared to late antibiotic treatment, which was particularly significant (p=0.005) for Clostridium cluster XIVa in the Regensburg group. Earlier antibiotic treatment prior to ASCT was further associated with a higher TRM (34%, n=79/236) compared to post-ASCT (21% n=62/297, p=0.001) or no antibiotics (7% n=6/88, p<0.001). Timing of antibiotic treatment was the dominant independent risk factor for TRM (HR 2.0, p≤0.001) in multivariate analysis beside increase age (HR 2.15, p=0.004), reduced Karnofsky performance status (HR 1.47, p=0.03) and female donor/ male recipient sex combination (HR 1.56, p=0.02) A competing-risk analysis revealed the independent effect of early initiation of antibiotics on GvHD-related TRM (p=0.004) in contrast to infection-related TRM and relapse (p=ns). The poor outcome associated with early administration of antibiotic therapy that is active against commensal organisms, and specifically the possibly protective Clostridiales calls for the use of Clostridiales-sparing antibiotics and rapid restoration of microbiota diversity after cessation of antibiotic treatment.

Keywords: Allogeneic stem cell transplantation, acute intestinal GvHD, treatment with broad spectrum antibiotics, intestinal microbiome, outcome

Introduction

Allogeneic stem cell transplantation (ASCT) is a curative treatment option for a variety of hematopoietic malignancies and other severe hematologic and genetic diseases.1 Its success, however, is still limited by a significant risk of life-threatening complications, including acute Graft-versus-Host Disease (GvHD) and infection. As neutropenia and mucosal injury following myeloablative conditioning regimens often lead to neutropenic infections in ASCT recipients,2 the majority of patients require treatment with broad-spectrum antibiotics. However, there is increasing evidence that use of broad-spectrum antibiotics may exert a detrimental impact on intestinal microbiota composition and, subsequently, the outcome of ASCT.3 Recent studies have demonstrated that loss of intestinal microbiome diversity and, in particular, commensal Clostridiales, appears to contribute not only to the pathogenesis of acute gastrointestinal (GI) GvHD, but also to increased transplant-related mortality (TRM).4,5 Similarly, urinary concentration of 3-indoxyl sulfate (3-IS), a tryptophan metabolite of commensal colonic bacteria, has been identified as an indirect marker of a balanced microbiota and predicts outcome at the time of ASCT.6

In the present study we investigated the association of the timing of antibiotic treatment with intestinal microbiota composition and clinical outcome in a cohort of 621 patients undergoing ASCT in two centers with different practices of antibiotic prophylaxis.

Patients, material and methods

Patients

A total of 621 adult patients undergoing ASCT in Regensburg, Germany (n=380), and New York (n=241), USA, were included in our retrospective analysis. Inclusion criteria were hemato-oncologic disease requiring ASCT with an age above 18 years and receipt of non-T cell depleted grafts. The Regensburg cohort consisted of 380 consecutive ASCT recipients enrolled between September 2008 and June 2015, whereas the New York cohort comprised 241 patients who underwent ASCT between October 2009 and May 2015 at the Memorial Sloan Kettering Cancer Center in New York. Further, with approval by the Ethics Committee of the University Medical Center of Regensburg and after receipt of written informed consent, urinary and stool specimens were collected in a total of 130 and 26 patients, respectively, at a minimum of six different time-points between admission and day 28 after ASCT: prior to admission, at least once between days −2 to +2, +2 and +10, +11 to +17, +18 to +24 and +25 to +30. In addition, stool samples from five healthy stem cell donors served as controls. Similarly, for the New York group, stool specimens from 146 ASCT recipients were obtained within 4 days of day 12 after ASCT. All specimens were stored at −80°C until analysis.

All patients received prophylactic antibiotics from the start of conditioning until engraftment, but the type of prophylactic antibiotics differed between the two cohorts. In Regensburg, ciprofloxacin 500 mg twice daily and metronidazole 400 mg three times daily were administered orally to n=189 (49.7%) patients until March 2012, before prophylaxis was switched to oral rifaximin 200 mg twice daily in the remaining n=191 (50.3%) patients to control the emergence of vancomycin-resistant enterococci. In the New York group, all 241 patients received prophylactic intravenous ciprofloxacin 400 mg twice daily and those receiving myeloablative preparative regimens were additionally given intravenous vancomycin at a starting dose of 1 g twice daily.

Although there were differences in prophylactic antibiotic regimens between the New York and Regensburg cohorts, treatment of neutropenic fever/infections with additional antibiotics was comparable between the two cohorts. In the Regensburg cohort, 350 of 380 (92%) patients received additional antibiotic treatment beyond prophylactic regimens. Piperacillin/tazobactam at a thrice-daily dose of 4.0/0.5 g was used as empiric first-line therapy in 275 of 350 patients (78%), while meropenem 1.0 g thrice daily and vancomycin 1.0 g twice daily served as second-line therapy in 55 of 350 patients (16%). In cases of penicillin intolerance, twenty patients (6%) received alternative first-line treatment with vancomycin and/or ceftazidime (in 13 patients after ciprofloxacin/metronidazole prophylaxis; in seven patients after rifaximin prophylaxis). In the New York cohort, piperacillin/tazobactam and imipenem/cilastatin were administered four times a day at doses of 4.0/0.5 g and 500 mg, respectively, as first- and second-line therapy. Overall, 183 of 241 patients (76%) in New York required antibiotics to treat neutropenic fever/infections, among whom n=103/183 (56%) received first-line therapy only, whereas n=80/183 (44%) required second-line therapy. Fifty-eight of 241 ASCT recipients (24%) required no therapeutic antibiotics. The clinical criteria for initiation of first- and second-line antibiotics were comparable to the Regensburg group. At both centers, patients received first- and second-line treatment according to international guidelines for treatment of neutropenic fever/infections.7, 8

All patients were classified into three groups according to the timing of additional antibiotic (AB) initiation: (i) early exposure to antibiotics between day −7 and day 0 (early AB group: 38%, n=236), (ii) exposure to antibiotics on day 0 or thereafter (late AB group: 48%, n=297), and (iii) no systemic antibiotic treatment during the course of ASCT beyond prophylactic regimens (no AB group: 14%, n=88, Table 1). Since acute leukemia patients usually have a history of several courses of cytotoxic treatment and repeated infections treated with broad spectrum antibiotics, we additionally defined two subgroups and divided patients into an acute leukemia (n=296) versus non-acute leukemia (n=325) group. Patients’ characteristics in the three antibiotic subgroups are shown in Table 2. .

Table 1.

Beginning of antibiotic treatment at the two centers.

Total cohort (n=621) Regensburg (n=380) New York (n=241)
No AB group 14% (n=88) 8% (n=30) 24% (n=58)
Early AB group 38% (n=236) 50% (n=190) 19% (n=46)
Late AB group 48% (n=297) 42% (n=160) 57% (n=137)
Open in a new tab

Table 2.

Summary of patient characteristics in the three AB subgroups.

Early AB group (n= 236) Late AB group (n= 297) No AB group (n= 88)
Type of underlying disease
 - Acute leukemia 57.6% (n=136) 43.8% (n=130) 34.1% (n=30)
 - No acute leukemia 42.4% (n=100) 56.2% (n=167) 65.9% (n=58)
Age of the patient, yrs
 - < 20 1.3 % (n=3) 0.7 % (n=2) 0 % (n=0)
 - 20–40 16.1 % (n=38) 19.5 % (n=58) 13.6% (n=12)
 - > 40 82.6 % (n=195) 79.8 % (n=237) 86.4% (n=76)
Donor type
 - HLA-identical sibling donor 19.5% (n=46) 25.3% (n=75) 53.4% (n=47)
 - Unrelated donor 80.5% (n=190) 74.7% (n=222) 46.6% (n=41)
Donor-recipient sex combination
 - Donor female, male recipient 16.9% (n=40) 22.2% (n=66) 21.6% (n=19)
 - All other 83.1% (n=196) 77.8% (n=231) 78.4% (n=69)
Disease stage
 - Early 20.8% (n=49) 33.6% (n=100) 43.2% (n=38)
 - Intermediate 32.2% (n=76) 36.4% (n=108) 36.4% (n=32)
 - Late 47.0% (n=111) 30.0% (n=89) 20.4% (n=18)
Time interval from diagnosis to transplant, months
 - < 12 64.4% (n=152) 55.2% (n=164) 52.3% (n=46)
 - ≥12 35.6% (n=84) 44.8% (n=133) 47.7% (n=42)
Karnofsky performance status
 - ≥90% 62.7% (n=148) 66.3% (n=197) 62.5% (n=55)
 - < 90% 37.3% (n=88) 33.7% (n=100) 37.5% (n=33)
Open in a new tab

Analysis of urinary 3-IS

Urinary 3-IS and creatinine levels were determined by reversed-phase liquid chromatography-electrospray ionization-tandem mass spectrometry as previously described.6

Quantification of Clostridium Cluster XIVa 16S rRNA gene copies by RT-PCR

Using Clostridium cluster XIVa group-specific primers and SYBR Green I Master (Roche) qPCR reagents, 16S rRNA gene copy numbers of Clostridium cluster XIVa species were determined in fecal DNA preparations by real-time quantitative PCR on a LightCycler 480 II instrument (Roche). Full-length 16S rDNA amplicons of Clostridium cluster XIVa bacteria cloned into the pGEM T-Easy vector served as quantification standards (Invitrogen).

Clostridial abundance measurement

At Memorial Sloan Kettering Cancer Center, stool specimens collected within 4 days of day 12 after ASCT were analyzed for clostridial abundance, as described previously.9 Briefly, stool specimens were subjected to mechanical disruption (bead-beating) and DNA was extracted with phenol-chloroform.10 DNA samples were analyzed by the Illumina MiSeq platform to sequence the V4-V5 region of the 16S rRNA gene. Sequence data were compiled and processed using mothur version 1.34,11 screened and filtered for quality,12 and classified to the species level13 using a modification of the Greengenes reference database.14

Bioinformatics and data analysis

Normally and non-normally distributed continuous data are presented as mean (±standard deviation) or median (range), respectively. Accordingly, group comparisons were performed by two-sided t, Mann-Whitney U or Kruskal-Wallis tests. Absolute and relative frequencies were given for categorical data and compared between study groups by chi-square or Fisher’s exact tests. All hypotheses were tested in an explorative manner on a two-sided 5% significance level. Factors associated with early use of systemic antibiotics were assessed using logistic regression analysis. Kaplan-Meier analysis was performed to assess survival and non-relapse mortality, and Cox regression was used for multivariate assessment of risk factors. Competing-risk analysis15 for GvHD related TRM, infectious TRM and relapse was performed using software package R 3.2.2 (The R Foundation of Statistical Computing, Vienna, Austria). Otherwise, IBM SPSS Statistics 22 (SPSS Inc, Chicago, IL, USA) was used for analysis.

Results

1. Factors associated with early antibiotic treatment in ASCT patients

In the study cohort, multivariate analysis identified several factors associated with higher likelihood of early antibiotic exposure before day 0: advanced stage of underlying disease (HR 2.1, 95% CI 1.7–2.7, p<0.001), matched unrelated donor (HR 2.2, 95% CI 1.5–3.4, p<0.001), and interval from first diagnosis to ASCT < 12 months (HR 2.6, 95% CI 1.8–3.8, p<0.001). Patient age and donor/recipient sex ratio did not affect timing of antibiotic treatment. The interval from diagnosis to ASCT was shorter in patients suffering from acute leukemia with a median of 5 months (1–206 months) than in non-acute leukemia patients requiring ASCT with a median of 18 months (2–156, p<0.001).

2. Timing of antibiotics affects intestinal microbiota disruption early after ASCT

Several lines of evidence suggest an impact of early antibiotics on loss of commensal bacteria and, therefore, raise the hypothesis that antibiotic-induced intestinal dysbiosis was involved in the poor outcome of patients receiving early antibiotic treatment.

Since urinary levels of 3-IS directly correlate with the abundance of Clostridiales,7 we measured the levels of this microbial biomarker in patients with evaluable samples from the Regensburg cohort. At baseline, no differences in 3-IS levels were observed prior to ASCT between the three AB subgroups (early AB group 30.3, STD 46.1 μmol/mmol creatinine, late AB group 24.9, STD 39.4 μmol/mmol creatinine, no AB group 19.0, STD 8.0 μmol/mmol creatinine, p=ns). In contrast, significantly lower mean urinary concentrations of 3-IS between days 0 and 28 after ASCT were found in early AB patients (n=62, 5.5, STD 13.6 μmol/mmol creatinine) than in late (n=57, 9.9, STD 8.2 μmol/mmol creatinine, p<0.001) and no AB patients (n=11, 11.1, STD 10.5 μmol/mmol creatinine, p=0.02), respectively, in the Regensburg cohort. The trend of 3-IS levels in the different subgroups from pre-transplant to day 28 after ASCT is shown in Figure 1.

Figure 1. Impact of antibiotic treatment on the course of urinary 3-IS levels during the first 28 days after ASCT.

Figure 1

Open in a new tab

Mean urinary 3-IS levels between days 0 and 28 after ASCT indicate signficant long-term suppression of commensal Clostridiales by early antibiotic treatment.

In the New York cohort, clostridial abundance was analyzed by 16S amplicon sequencing and was found to be higher in the late AB group (2.3%) than in the early AB group (0.74%, p=0.03). Patients that had received no therapeutic antibiotics showed the highest clostridial abundance (14%, p=0.003, Figure 2).

Figure 2. Comparison of Clostridial abundance in stool specimens in relation to timing of antibiotics.

Figure 2

Open in a new tab

Clostridial abundance was higher in the late or no AB group compared to the early AB group.

In a small subgroup of Regensburg patients who received rifaximin prophylaxis, we performed direct quantification of Clostridium Cluster XIVa 16S rRNA gene copies by RT-PCR in stool specimens collected from early (n=13) and late (n=13) AB patients. Between days 0 and 7 after ASCT, significantly lower copy numbers were found in early AB patients (2.4×104) compared to patients with late antibiotic treatment (2.6×108, p=0.005), indicating directly that early antibiotic treatment markedly reduced this subset of commensal intestinal bacteria. A direct comparison with healthy donors (n=5) showed significantly higher copy numbers with 3.4×109 copy numbers.

3. Early start of systemic antibiotics is associated with poor outcome after ASCT

Regarding outcome and early systemic antibiotic treatment, we found a significantly (log rank = 0.001) increased TRM in early AB (34%, n=79/236) compared to late AB patients (21% n=62/297). The lowest rate of TRM (7% n=6/88) was found in patients, who had not received additional antibiotic therapy. Their rate of TRM differed significantly from both early (log rank < 0.001) and late AB patients (log rank = 0.005, Figure 3a). Subgroup analysis revealed similar results in the Regensburg and New York cohorts (Table 3). Similarly, OS was lower in the early (51% n=121/236) than in the late (67% n=200/297, log rank < 0.001) and no AB group (73% n=64/88, log rank = 0.001), respectively.

Figure 3.

Figure 3

Figure 3

Figure 3

Open in a new tab

Figure 3a. Kaplan-Meier curves of transplant-related mortality in relation to initiation of antibiotic treatment. Patients receiving antibiotic therapy before day 0 showed highest TRM after ASCT.

Figure 3b. Kaplan-Meier curves of transplant-related mortality in relation to initiation of antibiotic treatment in the non-acute leukemia group.. For T patients without acute leukemia TRM was highest in the early AB group followed by the late and no AB group.

Figure 3c. Kaplan-Meier curves of transplant-related mortality in relation to initiation of antibiotic treatment in the acute leukemia group. TRM was highest in the early AB group compared to the late and no AB group in patients with acute leukemia.

Table 3. Transplant-related mortality in relation to the beginning of antibiotic treatment for neutropenic fever at the two centers.

TRM was significantly increased in early AB patients compared to late AB patients in each group. The lowest rate of TRM was found in patients without additional antibiotic therapy.

Total cohort (n=621) Regensburg (n=380) New York (n=241)
No AB group 7% (n=6/88) 3% (n=1/30) 9% (n=5/58)
Early AB group 34% (n=79/236) 33% (n=62/190) 37% (n=17/46)
Late AB group 21% (n=62/297) 19% (n=31/160) 23% (n=31/137)
Open in a new tab

In the Regensburg cohort, we observed a significantly higher rate of death from severe acute and/or chronic GvHD in the early (23%, n=44/190) compared to the late AB (16%, n=26/160) and the untreated group (3%, n=1/30, p=0.003). The same association was seen in the New York cohort: Death from severe acute and/or chronic GvHD was observed in 26% (n=12/46) in the early, in 12% (n=16/137) in the late and 5% (n=3/58) in the no AB group (p=0.01).

As previously reported in a smaller cohort of patients, we also observed an impact of antibiotic prophylaxis on TRM.16 Patients prophylaxed with ciprofloxacin/metronidazole experienced a significantly higher rate of TRM (n=43/96, 44.8%) than patients prophylaxed with rifaximin (n=19/94, 20.2%, p<0.001). It is possible that more heavily pretreated patients are both more susceptible to inferior long-term outcomes and to early peri-transplant fevers that necessitate early antibiotic treatment. To address this potential confounder we classified patients into acute leukemia versus non acute-leukemia groups and analyzed TRM rates. For non-acute leukemia patients, TRM was highest in the early AB group (37.0%, n=37/100) followed by the late (25.1%, n=42/167) and no AB group (6.9%, n=4/58, p<0.001, Figure 3b). This held also true for acute leukemia patients independently (30.9%, n=42/136, 15.4%, n=20/130, 6.7%, n=2/30, p=0.002, Figure 3c). In a multivariate analysis we identified early start of antibiotic treatment of neutropenic fever (p<0.001), high patients’ age ( p=0.004), low Karnofsky performance status (p=0.03) and female donor/ male recipient sex combination (p=0.02) as independent risk factors associated with TRM after ASCT (Table 4). The independent effect of early antibiotic treatment on GvHD-associated TRM (p=0.004) could be demonstrated in a competing-risk analysis, whereas infection-related TRM and relapse were not associated with start of systemic antibiotic treatment (p=ns).

Table 4. Multivariate risk factor analysis for TRM after ASCT.

Early initiation of antibiotic treatment, patients’ age more than 40 years, Karnofsky performance status < 90% as well as donor female/ male recipient sex combination were significantly associated with increased risk of TRM.

Risk factor P HR 95% CI for HR
Beginning of antibiotic treatment
Early AB group		 ≤0.001 2.0 1.42–2.83
Type of underlying disease
Acute leukemia		 0.85 0.97 0.68–1.38
Patient’s age
age>40 yrs		 0.004 2.15 1.27–3.64
Karnofsky performance status
< 90%		 0.03 1.47 1.05–2.06
Donor type
MUD		 0.06 1.53 0.99–2.36
Time from first diagnosis to ASCT
<12 months		 0.22 1.27 0.87–1.85
Stage of underlying disease
Advanced		 0.09 1.22 0.97–1.54
Donor-recipient sex combination
Donor female-recipient male		 0.02 1.56 1.06–2.29
Open in a new tab

In the table, numbers for high-risk groups are indicated for categorical variables. Significance level <0.05. CI, confidence interval; HR, hazard ratio; MUD, matched unrelated donor;

Discussion

Our study reveals a novel association between the timing of antibiotic treatment, microbiota disruption and outcome after ASCT. Risk factors for early antibiotic exposure included stage of underlying disease, donor type and time interval from diagnosis to ASCT. The association between unrelated-donor ASCT and early antibiotic treatment may be due to an indirect association with ATG serotherapy, which is commonly given in the last days prior to an unrelated donor ASCT and often produces fever. This fever can trigger empiric antibiotic treatment due to the inability to differentiate between a real infection and a side effect of ATG in high-risk neutropenic patients. In the Regensburg cohort, ATG was routinely administered prior to unrelated-donor transplantation, which might explain the higher proportion of patients in the early AB group compared to the New York cohort. Similarly, the correlation between early AB exposure and short interval from diagnosis to ASCT might reflect the higher proportion of acute leukemia patients in the short-interval group who are more prone to neutropenic fever requiring early antibiotic treatment.

In this large study, early antibiotic treatment was associated with significantly altered microbiota composition. This lends further support to an association between early antibiotic-induced shifts in gut microbiota composition and long-term outcome after ASCT. As reported recently, urinary levels of 3-IS correlate with the abundance of Clostridiales in fecal specimens,6 and this effect on abundance of Clostridiales was confirmed in the New York group using a different approach of Clostridial quantification. These associations were confirmed by Clostridium cluster XIVa species PCR in a small subgroup of patients.

Early antibiotic treatment was also correlated with a higher TRM and worse OS compared to patients with late or no additional broad-spectrum antibiotics. TRM rates in relation to initiation of antibiotic treatment were similar in the Regensburg and New York cohort, despite the use of two different antibiotic prophylaxis regimens. Furthermore, subgroup analysis showed that the correlation between early start of antibiotics and high TRM held true in both acute and non-acute leukemia subgroups . Although these data would need confirmation by a prospective study, competing-risk analysis identified the independent interaction of early antibiotic treatment and GvHD-associated TRM, whereas early exposure to antibiotics had no impact on infectious TRM. This analysis suggests that early exposure to antibiotics is not simply a surrogate parameter observed in more vulnerable patients who are at risk of early death from infectious complications. Rather, our data support the hypothesis that early antibiotic treatment induces microbiota disruptions and is itself a risk factor for worse outcome.

The lower TRM observed in patients prophylaxed in patients with rifaximin compared with ciprofloxacin/metronidazole is concordant with published observations of high antimicrobial activity of metronidazole against Clostridium Cluster XIVa species17 and of the major shifts in intestinal microbiota composition induced by ciprofloxacin.18 Rifaximin exerts little effect on overall gut microbiota diversity and has in fact been reported to reduce expression of bacterial virulence factors, inhibit bacterial attachment to the intestinal mucosa, and decrease production of pro-inflammatory cytokines in the intestinal mucosa.19

Our observation extends data reported on suppression of commensal Blautia species by antibiotic treatment in ASCT patients9 and supports clinical and experimental data indicating that antibiotics with activity against anaerobic Clostridiales increase GvHD-related mortality.20 Moreover, recent findings revealed that treatment of neutropenic fever with piperacillin-tazobactam or imipenem-cilastatin was associated with aggravated gut microbial perturbation and a significantly higher GvHD-related mortality compared to the use of aztreonam or cefepime, both antibiotics with reduced activity against anaerobic commensal bacteria in the gut.21 The data of the present study add the important observation that not only does the type of antibiotic drug matter but in particular the timing of antibiotic initiation has a crucial impact on microbiota composition and outcome after ASCT. We hypothesize that an early loss of protective bacteria in the gut due to early administration of systemic antibiotics affecting clostridia may result in a disturbed environment before the time of first donor T cell contact with recipient gastrointestinal tissues. The configuration of the intestinal microbiota in the immediate peri-transplant period may have a more profound impact than a later microbiota disruption.

Several mechanisms have been identified regarding potential protective effects of commensal bacteria and may be relevant for our observation: In patients with early antibiotic treatment, the increased incidence of death from both chronic as well as acute GvHD indicates that long-term development of intestinal tolerance is impaired. Antibiotic treatment has been shown to suppress production of Reg3α in Paneth cells, which prevents colonization by pathogens;22, 23 similarly, metabolites of commensal Clostridiales mediate colonization resistance and maintain a microbiota supporting integrity of the intestinal epithelial barrier. Furthermore, commensal Clostridiales have been identified as important producers of short chain fatty acids (SCFAs) such as butyrate, which maintain intestinal regulatory T cells.24–26 In addition, SCFAs can induce interleukin-22 (IL-22) responses in innate lymphoid cells (ILCs), which exert anti-inflammatory and protective effects on intestinal epithelial cells.27 Loss of these mechanisms may explain some of the long-term effects on GvHD-related TRM in our study. The time-dependent effects further indicate that a balanced GI microbiota may be particularly critical in the early period following ASCT as donor lymphocytes migrate to peripheral target organs. The importance of a balanced intestinal microbiota for immunoregulation and the detrimental effects of antibiotics during critical periods is increasingly recognized for a variety of other human diseases: most convincingly, antibiotic exposure in early neonatal life has been shown to favor development of autoimmune diseases,28, 29 and more recently even a negative effect of antibiotics on the efficacy of cytotoxic drugs such as cyclophosphamide or doxorubicin has been reported, as the microbiota can stimulate effector cells of the immune system that mediate optimal activity of cytotoxic treatment.30–32

If confirmed in multicenter analyses, our observations have several implications. Since commensal Clostridiales are gram-positive anaerobic bacteria, first the use of Clostridiales-sparing β-lactam antibiotics that target exclusively or mostly gram-negative bacteria for antibiotic therapy could be considered. Depending on the clinical response of the patients, shorter durations of antibiotics seems feasible and might also contribute to more restricted use of antibiotics in low-risk patients. Second, prospective trials of antibiotic prophylaxis should be initiated to investigate different strategies or even no gut decontamination aiming at the protection of gut microbiota. Third, strategies to overcome the impact of antibiotics may be developed either based on pro- or on prebiotics applied in parallel to antibiotics. Alternatively, rapid restoration of microbiota diversity early after cessation of antibiotic treatment by either fecal microbiota transplantation or artificially composed microbiota seems a further feasible option. Beyond ASCT, our observations warrant a more critical use of antibiotics in general, not only to prevent the development of antibiotic resistance, but also due to the observed indirect effects that commensal microbiota may have on long-term health in general.

Highlights.

  • Intestinal microbiota play an important role in patients with allogeneic SCT

  • Early use of antibiotics affects intestinal microbiota composition

  • Early use of antibiotics influences the outcome of allogeneic SCT recipients

Acknowledgments

This study was supported in part by grants from the German Research Foundation (DFG, KFO’s 234 and 262), the Regensburg Center for Interventional Immunology, the Marie Curie Initial Training Network Celleurope, the ReForM program of the Regensburg School of Medicine, the National Institutes of Health award numbers R01-HL124112 (R. R. Jenq), R01-HL069929 (M.R.M. van den Brink), R01-AI080455 (M.R.M. van den Brink), R01-AI101406 (M.R.M. van den Brink), P01-CA023766 (R. J. O’Reilly), Project 4 of P01-CA023766 (M.R.M. van den Brink), P30-CA008748 (S. M. Devlin), R01-AI042135 (E. G. Pamer), R01-AI095706 (E. G. Pamer), and K23-AI095398 (Y. Taur). Support was also received from the U.S National Institute of Allergy and Infectious Diseases (NIAID Contract HHSN272200900059C), The Experimental Therapeutics Center of MSKCC funded by Mr. William H. Goodwin and Mrs. Alice Goodwin, The Lymphoma Foundation, Alex's Lemonade Stand, The Geoffrey Beene Cancer Research Center at MSKCC, The Susan and Peter Solomon Divisional Genomics Program, the Lucille Castori Center for Microbes, Inflammation, and Cancer, and the Tow Foundation. This project has received funding from the European Union’s Seventh Programme for research, technological development and demonstration under grant agreement No. 602587. In addition, D. Wolff and E Holler received support from the German José Carreras Foundation. Authors acknowledge the help of Heike Bremm, Constanze Winter and Yvonne Schumann in collecting and cryopreserving patient specimens as well as Nadine Nuernberger in performing 3-IS analyses.

Footnotes

Declaration of interests: All authors declare no competing financial and personal interests.

Authorship contribution: D.W. and E.H were involved in conception and design of the study, D.Wo. and J.H. were responsible for collection of specimens. R.R.J., J.U.P. and Y.T. were responsible for data assessment and data analysis of the New York cohort. P.J.O. and K.D. performed measurements of 3-IS levels. A.H., J.K. and A.G. conducted qRT-PCR of Clostridium cluster XIVa species. M.W. contributed to statistical data analysis. D.W. and E.H. collected and analyzed clinical data and wrote the manuscript. W.H., M.R.M.B. and E.P. were involved in interpretation and discussion of study results according to the current literature. All authors read, revised and approved the final draft.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373(9674):1550–1561. doi: 10.1016/S0140-6736(09)60237-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Gratwohl A, Brand R, Frassoni F, et al. Cause of death after allogeneic haematopoietic stem cell transplantation (HSCT) in early leukaemias: an EBMT analysis of lethal infectious complications and changes over calendar time. Bone marrow transplantation. 2005;36(9):757–769. doi: 10.1038/sj.bmt.1705140. [DOI] [PubMed] [Google Scholar]
  • 3.Holler E, Butzhammer P, Schmid K, et al. Metagenomic analysis of the stool microbiome in patients receiving allogeneic SCT: Loss of diversity is associated with use of systemic antibiotics and more pronounced in gastrointestinal GvHD. Biol Blood Marrow Transplant. 2014 doi: 10.1016/j.bbmt.2014.01.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Jenq RR, Ubeda C, Taur Y, et al. Regulation of intestinal inflammation by microbiota following allogeneic bone marrow transplantation. J Exp Med. 2012;209(5):903–911. doi: 10.1084/jem.20112408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Taur Y, Jenq RR, Perales MA, et al. The effects of intestinal tract bacterial diversity on mortality following allogeneic hematopoietic stem cell transplantation. Blood. 2014;124(7):1174–1182. doi: 10.1182/blood-2014-02-554725. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Weber D, Oefner PJ, Hiergeist A, et al. Low urinary indoxyl sulfate levels early after ASCT reflect a disrupted microbiome and are associated with poor outcome. Blood. 2015 doi: 10.1182/blood-2015-04-638858. [DOI] [PubMed] [Google Scholar]
  • 7.Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the infectious diseases society of america. Clin Infect Dis. 2011;52(4):e56–93. doi: 10.1093/cid/cir073. [DOI] [PubMed] [Google Scholar]
  • 8.Averbuch D, Orasch C, Cordonnier C, et al. European guidelines for empirical antibacterial therapy for febrile neutropenic patients in the era of growing resistance: summary of the 2011 4th European Conference on Infections in Leukemia. Haematologica. 2013;98(12):1826–1835. doi: 10.3324/haematol.2013.091025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Jenq RR, Taur Y, Devlin SM, et al. Intestinal Blautia Is Associated with Reduced Death from Graft-versus-Host Disease. Biol Blood Marrow Transplant. 2015;21(8):1373–1383. doi: 10.1016/j.bbmt.2015.04.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Turnbaugh PJ, Hamady M, Yatsunenko T, et al. A core gut microbiome in obese and lean twins. Nature. 2009;457(7228):480–484. doi: 10.1038/nature07540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Schloss PD, Westcott SL, Ryabin T, et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 2009;75(23):7537–7541. doi: 10.1128/AEM.01541-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Schloss PD, Gevers D, Westcott SL. Reducing the effects of PCR amplification and sequencing artifacts on 16S rRNA-based studies. PLoS One. 2011;6(12):e27310. doi: 10.1371/journal.pone.0027310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73(16):5261–5267. doi: 10.1128/AEM.00062-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.DeSantis TZ, Hugenholtz P, Larsen N, et al. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol. 2006;72(7):5069–5072. doi: 10.1128/AEM.03006-05. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Scrucca L, Santucci A, Aversa F. Competing risk analysis using R: an easy guide for clinicians. Bone marrow transplantation. 2007;40(4):381–387. doi: 10.1038/sj.bmt.1705727. [DOI] [PubMed] [Google Scholar]
  • 16.Weber D, Oefner PJ, Dettmer K, et al. Rifaximin preserves intestinal microbiota balance in patients undergoing allogeneic stem cell transplantation. Bone marrow transplantation. 2016;51(8):1087–1092. doi: 10.1038/bmt.2016.66. [DOI] [PubMed] [Google Scholar]
  • 17.Goldstein EJ, Citron DM, Tyrrell KL. Comparative in vitro activities of SMT19969, a new antimicrobial agent, against 162 strains from 35 less frequently recovered intestinal Clostridium species: implications for Clostridium difficile recurrence. Antimicrob Agents Chemother. 2014;58(2):1187–1191. doi: 10.1128/AAC.02184-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Dethlefsen L, Huse S, Sogin ML, Relman DA. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLoS Biol. 2008;6(11):e280. doi: 10.1371/journal.pbio.0060280. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.DuPont HL. Review article: the antimicrobial effects of rifaximin on the gut microbiota. Aliment Pharmacol Ther. 2016;43(Suppl 1):3–10. doi: 10.1111/apt.13434. [DOI] [PubMed] [Google Scholar]
  • 20.Shono Y, Docampo MD, Peled JU, Perobelli SM, Jenq RR. Intestinal microbiota-related effects on graft-versus-host disease. International journal of hematology. 2015 doi: 10.1007/s12185-015-1781-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Shono Y, Docampo MD, Peled JU, et al. Increased GVHD-related mortality with broad-spectrum antibiotic use after allogeneic hematopoietic stem cell transplantation in human patients and mice. Sci Transl Med. 2016;8(339):339ra371. doi: 10.1126/scitranslmed.aaf2311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Eriguchi Y, Nakamura K, Hashimoto D, et al. Decreased secretion of Paneth cell alpha-defensins in graft-versus-host disease. Transpl Infect Dis. 2015;17(5):702–706. doi: 10.1111/tid.12423. [DOI] [PubMed] [Google Scholar]
  • 23.Brandl K, Plitas G, Mihu CN, et al. Vancomycin-resistant enterococci exploit antibiotic-induced innate immune deficits. Nature. 2008;455(7214):804–807. doi: 10.1038/nature07250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Narushima S, Sugiura Y, Oshima K, et al. Characterization of the 17 strains of regulatory T cell-inducing human-derived Clostridia. Gut Microbes. 2014;5(3):333–339. doi: 10.4161/gmic.28572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Atarashi K, Tanoue T, Oshima K, et al. Treg induction by a rationally selected mixture of Clostridia strains from the human microbiota. Nature. 2013;500(7461):232–236. doi: 10.1038/nature12331. [DOI] [PubMed] [Google Scholar]
  • 26.Furusawa Y, Obata Y, Fukuda S, et al. Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells. Nature. 2013;504(7480):446–450. doi: 10.1038/nature12721. [DOI] [PubMed] [Google Scholar]
  • 27.McDermott AJ, Huffnagle GB. The microbiome and regulation of mucosal immunity. Immunology. 2014;142(1):24–31. doi: 10.1111/imm.12231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Reynolds LA, Finlay BB. A case for antibiotic perturbation of the microbiota leading to allergy development. Expert Rev Clin Immunol. 2013;9(11):1019–1030. doi: 10.1586/1744666X.2013.851603. [DOI] [PubMed] [Google Scholar]
  • 29.Kuo CH, Kuo HF, Huang CH, Yang SN, Lee MS, Hung CH. Early life exposure to antibiotics and the risk of childhood allergic diseases: an update from the perspective of the hygiene hypothesis. J Microbiol Immunol Infect. 2013;46(5):320–329. doi: 10.1016/j.jmii.2013.04.005. [DOI] [PubMed] [Google Scholar]
  • 30.Viaud S, Daillere R, Yamazaki T, et al. Why should we need the gut microbiota to respond to cancer therapies? Oncoimmunology. 2014;3(1):e27574. doi: 10.4161/onci.27574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Viaud S, Saccheri F, Mignot G, et al. The intestinal microbiota modulates the anticancer immune effects of cyclophosphamide. Science. 2013;342(6161):971–976. doi: 10.1126/science.1240537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Mattarollo SR, Loi S, Duret H, Ma Y, Zitvogel L, Smyth MJ. Pivotal role of innate and adaptive immunity in anthracycline chemotherapy of established tumors. Cancer Res. 2011;71(14):4809–4820. doi: 10.1158/0008-5472.CAN-11-0753. [DOI] [PubMed] [Google Scholar]

RESOURCES