Can a Patient’s Breath Detect Post-Transplant Gastrointestinal Graft-Versus-Host Disease?

Analyzing the volatile organic compounds (VOCs) expelled in the breath can help clinicians detect and monitor gastrointestinal graft-versus-host disease (GI GVHD) in patients who have undergone an allogeneic hematopoietic cell transplantation (alloHCT). The results of this pilot study were published in Blood Advances.

“The ultimate hope is that, in the future, breath analysis will be a novel and easy test that can identify and monitor changes in GVHD,” said lead study author Betty Hamilton, MD, of the Cleveland Clinic. “Breath analysis can be done in the clinic and does not require any invasive procedures or hospitalizations and may potentially substitute endoscopies with biopsies as a tool to monitor early changes in GVHD.”

VOCs generated during pathologic processes have been used to diagnose a variety of other diseases, Dr. Hamilton explained. Given the role of the microbiome in GVHD, the investigators hypothesized that micro-organisms producing volatile metabolites may alter VOCs expelled in breath in patients with GI GVHD. These alterations could be detected on mass spectrometry to determine the presence and changes in GI GVHD.

This prospective, single-institution, pilot study enrolled 29 patients (age ≥12 years) who had no active respiratory infection and had received alloHCT for any diagnosis, with any conditioning, and with any donor source. The study included:

  • 19 patients who had grade 2-4 GI GVHD
  • 10 patients with no GVHD at day 100 post-alloHCT

As an additional control group, researchers also collected breath samples from 10 healthy controls who were present in a GVHD environment (e.g., health care staff and caregivers) to evaluate potential background environmental factors.

The samples were then analyzed for the presence of several prespecified VOCs, including acetaldehyde, benzene, ethanol, and ammonia.

There were no significant differences in baseline characteristics between patients with and without GI GVHD, the researchers noted, with groups well balanced in regards to sex (p=0.41), race (p=0.53), age at transplant (p=0.89), or comorbidity index (p=0.54).

All patients received a standard protocol of prophylactic antibiotics (trimethoprim-sulfamethoxazole at the start of conditioning through engraftment, followed by amoxicillin or azithromycin during immunosuppression).

There were five VOCs that differentiated between GI GVHD and no GVHD, including 2-propanol, acetaldehyde, dimethyl sulfide, isoprene, and 1-decene. A model that incorporated these five VOCs correctly classified 89% of patients with GI GVHD and 90% of patients without GI GVHD. However, there appeared to be no correlation between VOCs and GVHD severity.

“We hope that, in the future, breath analysis will be a novel and easy test that can identify and monitor changes in GVHD.”

—Betty Hamilton, MD

Samples also were analyzed to determine whether breath compounds were associated with malignant disease response after alloHCT. Of 19 patients, 10 achieved complete response (CR) and three had a partial response, while six patients had no response. Breath analysis identified two VOCs (pentane and ammonia) that distinguished between patients with a CR and less than a CR (p=0.011). A model incorporating these two factors correctly classified 90% of patients with CR and 78% of patients with less than a CR. Higher levels of acetone also appeared to be associated with a lower likelihood of response (odds ratio = 0.22 per 1-unit increase; 95% CI 0.06-0.82; p=0.024).

“We certainly acknowledge that this was a pilot study primarily demonstrating feasibility of breath analysis in a post-transplant population,” Dr. Hamilton said. “We were able to establish a signal of a ‘breath signature’ unique to GVHD, which may help clinicians “identify patients at higher risk for developing high-risk or severe GVHD and intervene earlier in their disease course.”

Limitations of the study include its small sample size, the recruitment of patients from a single institution, the pilot study design, and the determination of VOCs at only one point in time.

“We are now performing a follow-up study evaluating serial breath samples (pre- and post-transplant), as well as stool and saliva samples, to further evaluate the microbiome in these body fluids as it correlates to the breath signature,” Dr. Hamilton concluded.

The authors disclosed no conflicts of interest.


Hamilton B, Rybicki L, Grove D, et al. Breath analysis in gastrointestinal graft-versus-host disease after allogeneic hematopoietic cell transplantation. Blood Adv. 2019.

Gastrointestinal (GI) graft-versus-host disease (GVHD) is a dangerous complication of allogeneic transplantation with a high mortality rate. Identifying GVHD early and noninvasively and distinguishing it from other causes of GI symptoms common in allogeneic transplantation would be a major advance in practice. It is tantalizing to think about a breath test – like a roadside breathalyzer for blood alcohol content or a urea breath test for H. pylori – that could provide diagnostic information that otherwise requires endoscopy and biopsies. Earlier diagnosis would be beneficial if it avoids extensive tissue damage before symptoms develop. The authors conclude that breath analysis in patients undergoing allogeneic transplantation is feasible and there may be differences between those with GI GVHD and those who do not have it. However, as acknowledged by the authors, this is the first preliminary look into whether breath tests are useful diagnostic tools. As noted in the article, there are many reasons why patients with and without GI GVHD could exude different volatile organic compounds having nothing to do with GVHD but rather due to differences in medications, hospitalization status, time since the transplant, food intake, or any number of epiphenomena that could not be controlled in this study. The authors are to be congratulated for conducting this study, for their conservative interpretation of their results, and for planning a larger, more definitive study.

Stephanie Lee, MD, MPH
Fred Hutchinson Cancer Research Center
Seattle, WA