Increased incidence of intestinal GvHD with broad-spectrum antibiotic use 

During the process of hematopoietic stem cell transplantation (HSCT), used to treat or potentially cure various hematologic malignancies, patients often become neutropenic due to the conditioning treatment prior to engraftment. In this period, antibiotics are often given to prevent neutropenic infections.

Graft-versus-host disease (GvHD) is a common complication of allogeneic HSCT (allo-HSCT). It has been shown that anaerobic bacteria residing in the intestine and commensal bacteria can mediate the severity of acute GvHD (aGvHD). Preclinical studies also indicate that intestinal microflora can prevent the development of aGvHD, even in the case of major histocompatibility complex (MHC) antigen-mismatch. The use of broad-spectrum antibiotics, therefore, can disrupt this process, either leading to a lack of bacterial diversity, or overgrowth of a dominant organism. This in turn can affect the results of HSCT.

The potential link between the choice of antibiotic, changes to the microbiota, and the influence on aGvHD has not been well evaluated. In this retrospective, single center study, recently published in the Biology of Blood and Marrow Transplantation, Sung-Eun Lee, The Catholic University of Korea, Seoul, KR, and colleagues analyzed the impact of antibiotic usage in patients undergoing allo-HSCT, on aGvHD and microbiota diversity.

Study design and patient characteristics

• Patients (N = 211) with hematological malignancies who received an allo-HSCT at Seoul St Mary’s Hospital, KR, between 2013 and 2014, were included:
  • Median age: 45 years (16–68)
  • Gender (male vs female): 127 vs 84
  • Donor (sibling vs unrelated vs human leukocyte antigen [HLA] mismatched familial): 98 vs 71 vs 42
• Patients were randomized to receive the following, during the neutropenic period:
  • Group 1: no antibiotics, n = 43
  • Group 2: cefepime only (narrow-spectrum), n = 87
  • Group 3: broad-spectrum carbapenem antibiotics (meropenem or prepenem) without prior cefepime therapy, n = 81
  • Median time to antibiotic treatment: eight days post-HSCT (-6 – 22 days)
  • Median duration of antibiotic treatment: eight days (2–37)
• There was an imbalance between groups relating to intensity of conditioning used (myeloablative [MAC] vs reduced-intensity [RIC]), p = 0.046:
  • Total: 112 (53.1%) vs 99 (46.9%)
  • Group 1: 16 (37.2%) vs 27 (62.8%)
  • Group 2: 47 (54%) vs 40 (46%)
  • Group 3: 49 (60.5%) vs 32 (39.5%)
• The following analyses were conducted pre- and post-HSCT:
  • Intestinal microbiota via stool samples
    • Pre-HSCT (one day prior to conditioning therapy): n = 65
    • Post-HSCT (post-engraftment – first of three consecutive days with an absolute neutrophil count >0.5 x 10⁹/L): n = 50
    • Median time to stool collection post-HSCT: 13 days
  • Immunophenotypic analysis via a peripheral blood sample
    • Pre-HSCT (one day prior to conditioning therapy): n = 145
    • Post-HSCT (post-neutrophil engraftment): n = 50
• Treatment regimen:
  • Conditioning (MAC or RIC)
  • Source (bone marrow vs peripheral blood stem cells mobilized by granulocyte colony-stimulating factor [G-CSF]): 33 vs 178
  • GvHD prophylaxis: calcineurin inhibitor (cyclosporine if matched-related and tacrolimus for unrelated and haploidentical) and short-term methotrexate
  • Antibacterial and antifungal prophylaxis was administered
  • G-CSF was given from day +7 after transplantation until neutrophil count was > 3 x 10⁹/L
• Patient characteristics were comparable across the three groups with the exception of conditioning regimen

Results

• All patients engrafted
  • Median time to neutrophil engraftment: 12 days
  • Median time to platelet engraftment: 14 days
• Rates of GvHD and transplant outcomes and are shown in Table 1 and Table 2
• In total, 45% of patients (n= 130) developed aGvHD of grade two or higher, at a median of 33 days post-transplantation
• Fifty-four patients developed intestinal GvHD, of whom, 48.1% had received broad-spectrum antibiotics
• Patients who received a longer duration of antibiotic treatment during the neutropenic period had:
  • Higher treatment-related mortality (TRM), p < 0.001
  • Lower rate of relapse, p = 0.022
  • Showed a trend towards higher rates of intestinal GvHD, p = 0.08 and lower overall survival (OS), p = 0.08
• Multivariate analysis showed significant associations between broad-spectrum antibiotic use and the development of intestinal GvHD (Table 3)

Table 1. Transplantation outcome: GvHD by group (compared to total)

	Total, N (%) N = 211	Group 1, % N = 43	Group 2, % N = 87	Group 3, % N = 81	p value
aGvHD, acute graft-versus-host disease
aGvHD grade ≥2	95 (45)	32.6	43.7	53.1	0.087
aGvHD grade ≥3	45 (21.3)	9.3	23	25.9	0.088
Intestinal GvHD	54 (25.6)	11.6	26.4	32.1	0.044
Steroid-refractory intestinal GvHD	Not reported	20	13	19	0.827

Table 2. Overall outcomes of transplantation by group

Mean ± SD	Total, N (%) N=211	Group 1, % N=43	Group 2, % N=87	Group 3, % N=81	p value
OS, overall survival; SD, standard deviation; TRM, treatment-related mortality
Cumulative incidence of relapse	24.8 ± 4.6	17.7 ± 6.3	30 ± 10.5	24.7 ± 5.9	0.983
Cumulative incidence of TRM	14.4 ± 2.6	9.3 ± 4.5	12.3 ± 3.7	19.3 ± 5	0.527
Event-free survival	60.8 ± 4.8	73± 7.1	57.5 ± 10	56.6 ± 6.7	0.698
OS	69.2 ± 4.1	78.4 ± 7	66.8 ± 6.5	67.5 ± 7.1	0.893

Table 3. Multivariate analysis of factors associated with development of intestinal GvHD (95% CI)

Factor	HR (range)	p value
Age (years), continuous	0.97 (0.95–1)	0.07
GvHD prophylaxis (FK506-based vs cyclosporine-based)	1.83 (0.95–3.55)	0.072
Antibiotic use (cefepime only vs no antibiotics)	2.21 (0.76–6.43)	0.146
Antibiotic use (carbapenem vs no antibiotics)	3.25 (1.13–9.34)	0.029
GvHD, graft-versus-host disease; HR, hazard ratio

Intestinal microbiota

• Analyzing stool samples pre- and post-HSCT showed differences in the following microbiota:
  • Phylum level: Firmicutes and Proteobacteria
  • Genus level:
    • Down regulated post-HSCT: Lachnospira
    • Up-regulated post-HSCT: Parabacteroides, Streptococcus, Lactonifactor and Macellibacteroides
• Analyzing differences based on antibiotic usage found:
  • Patients receiving broad-spectrum antibiotics (group 3) had:
    • Significant reduction in: Bacteroidetes, Cyanobacteria and Verrucomicrobia
    • Expansion in: Melissococcus and Anaerotruncus
• Effect of a loss of diversity on intestinal GvHD:
  • Paired stool samples were available for 21 patients split into:
    • No intestinal GvHD (n= 11)
      • No significant difference in pre-HSCT and post-HSCT microbial diversity
    • Intestinal GvHD (n= 10)
      • There was a difference between pre-HSCT and post-HSCT microbial diversity in patients who developed intestinal GvHD based on the Chao1 richness index
      • Reduction in diversity: reduction in Ruminococcus and increase in Eubacterium and Escherichia

Patients who received broad spectrum antibiotics had a loss of intestinal microbiota diversity. This change in the frequency of intestinal bacteria phyla was associated with increased intestinal GvHD.

Immunophenotypic analysis

• The authors conducted an analysis of the following immune cell subsets by use of antibiotic:
  • CD3⁺, CD4⁺ and CD8⁺ T-cells
  • Regulatory T-cells
  • Mucosal-associated invariant T (MAIT) cells
  • Natural killer (NK), NKT and type 1 NKT cells
  • Myeloid-derived suppressor cells (MDSCs, E- and M-MDSCs)
• Patients receiving broad spectrum antibiotics (group 3):
  • Pre-HSCT: lower frequency of CD4⁺ T-cells
  • Post-HSCT: higher frequency of CD8⁺ and lower ratio of E-MDSCs and M-MDSCs to T-cells

Conclusion

This study has demonstrated that the use of broad-spectrum antibiotics during the neutropenic period post-HSCT is associated with increased rates of intestinal GvHD and a reduced diversity in the microbiome.

In order to reduce the rate of intestinal GvHD, the authors recommend to consider:

• Avoiding the use of broad-spectrum antibiotics where possible
  • This is because the use of narrow spectrum antibiotics may result in a reduced loss of microbiota diversity, which may attenuate the risk of intestinal GvHD.
• Limiting the use of antibiotics in low-risk patients
• Restoring the microbiota after necessary antibiotic treatment
  • For example, by fecal microbiota transplantation or microbiota

Limitations of this study included the timing of stool collection and measurement at a single time point. Changes in microbiota occur rapidly, therefore the full extent of these changes may not have been captured. Further prospective studies to evaluate these approaches are therefore warranted, and should aim to identify if maintaining bacterial diversity would prevent aGvHD developing.