High Levels of Fecal Calprotectin and alpha1-Antitrypsin At Diagnosis of Acute Gastro-Intestinal Graft-Versus-Host Disease Are Correlated with Severity and Response to Corticosteroids

Diagnosis of gastrointestinal acute graft-versus-host disease (GI-GVHD) after allogeneic hematopoietic stem cell transplantation (HSCT) is based on clinical symptoms (nausea, vomiting and diarrhea quantification) and histological findings. No biomarkers are currently routinely used to predict the response to treatment whereas 30 to 50% of patients will not respond to the first line therapy. Histology can be an indicator for severity but as it's an invasive procedure, it's not available in all patients. Fecal calprotectin, alpha-1-antitrypsin (α₁-AT) and elastase are noninvasive biomarkers used in the screening of intestinal inflammation, protein-losing enteropathy and exocrine pancreatic dysfunction, respectively. These markers have not been yet validated in the context of HSCT and we aim to evaluate them as diagnostic and prognostic biomarkers of GI-GVHD.

From 09/2008 to 12/2010, 56 patients who developed GI symptoms for the first time after HSCT were prospectively included after inform consent in our center. Twenty patients who developed an acute GVHD without GI symptoms were also included as controls. At time of first symptoms (after a median of 2 days), fecal samples were collected to measure calprotectin, α₁-AT and elastase concentrations. Data regarding GI symptoms, GVHD stage, immunosuppressive treatment, albuminemia and performance status were also collected and tested in statistical model with fecal markers. Prognostic value of markers was evaluated by their association with complete response (CR) and steroid-refractory (SR) GVHD using a multiple logistic regression. Model calibration was assessed by the calibration slope and the bootstrap bias-corrected calibration slope and model discrimination by the c-index with or without bootstrap correction for over-optimism. Cumulative incidence (CI) functions were compared using Gray's test and adjustment was performed using the Fine and Gray method.

Among the 76 patients, median age was 44 years (8–66), a majority of them received a reduced intensity conditioning regimen (62%) and peripheral blood stem cells as source of stem cells (75%) from an unrelated donor (62%). Definitive diagnosis in patients with GI symptoms was GI-GVHD in 51 (histology in 43) and infection in 5 patients.

Calprotectin concentration was higher in patients with GI symptoms than in controls. α₁-AT concentration was higher in patients with GI-GVHD than in those with GI infections or controls. The concentration of elastase in patients with GI-GVHD was decreased but associated with high levels of calprotectin and α₁-AT (76%) attesting from mucosal damage possibly not related to a pancreatic dysfunction. In patients with GI-GVHD, the concentrations of calprotectin and α₁-AT increase with GI-GVHD stage but patients with stage 1 GI-GVHD have similar marker levels than patients without GI-GVHD (values in normal ranges in 87 and 83% for stage 1 vs. 92 and 88% for stage 0).

Regarding CR of GI-GVHD after treatment, initial concentrations of calprotectin ≥100μg/g and α₁-AT > 1.5 mg/g of dry stools decrease the probability of CR after treatment (56% vs. 86%, p=0.005; 67% vs. 90%, p=0.00007, respectively). Similarly, high level of calprotectin and α₁-AT predict a 93% (vs. 33%, p=0.00001) and a 72% (vs. 21%, p=0.00009) probability of SR-GVHD.

Calprotectin and α₁-AT concentration were independently and exclusively associated with a lower probability of CR (HR: 0.47, [p=0.028] and HR: 0.47, [p=0.037], respectively). The model selection procedure also identified a mode with both variables together with stage > 2 GI-GVHD as best predicting SR-GVHD, although marginal testing did not yield a significant odds ratio (OR) for α₁-AT (calprotectin: OR 17.1 [p=0.016]; GI stage III-IV: OR 7.11 [p=0.043]; α₁-AT: OR 3.67 [p=0.11]). Interestingly, excluding GVHD stage from the model, Calprotectin and α₁-AT concentration could also independently predict SR-GVHD (OR :15.8, p=0.015 & OR :4.93, p=0.038).

Our results showed that fecal calprotectin and α₁-AT levels are biomarkers of GI-GVHD, particularly in GI stage > I. Furthermore, an increase level of those markers predicts the response to treatment. Their use as predictors from severe GI-GVHD are very promising including in patients in whom GI-GVHD stage or histology are not available.

Peffault de Latour:Alexion: Consultancy, Research Funding.