Acute gastrointestinal graft-versus-host disease in allogeneic hematopoietic stem cell transplanted children and adolescents : clinical aspects of histopathological evaluation and risk factors

Abstract

Background: Acute graft-versus-host disease (aGVHD), following allogeneic hematopoietic stem cell transplantation (HSCT), is a potentially life-threating condition. Gastrointestinal involvement of aGVHD (GI-aGVHD) affects approximately every fourth transplanted child. The diagnosis of GI-aGVHD is primarily symptom-based. However, symptoms associated with GI-aGVHD are nonspecific; thus, histopathological confirmation of the diagnosis, is recommended. The overall objectives of this thesis were: i) to evaluate the influence of two different conditioning regimens on the incidence of GI-aGVHD, and ii) to evaluate clinical aspects of the currently recommended diagnostic approach to GI-aGVHD, i.e., endoscopy-guided histopathological assessment, applied to pediatric HSCT patients.

Patients and methods: Four retrospective cohort studies were included in this thesis. Paper I enrolled all children with HSCT performed during 2000–2010 at Karolinska University Hospital Huddinge who also had underlying diagnoses of juvenile myelomonocytic leukemia (JMML) or myelodysplastic syndrome (MDS). The children were conditioned with busulfan (Bu) and cyclophosphamide (Cy), with or without addition of melphalan (Mel). Paper IIIV included all children who underwent HSCT at any of the four HSCT centers in Sweden between 2000 and 2012 and with endoscopy-guided histopathological assessment performed to confirm symptom-based GI-aGVHD within one-year post-HSCT. In paper III-IV a retrospective, blinded, histopathological assessment (RIHA) was carried out based on the National Institutes of Health (NIH) 2014 criteria for histopathology-based GI-GVHD. Paper IV only included those with at least a biopsy sampling from the rectosigmoid area and the area proximal to the left colonic flexure.

Results: Paper I. Twenty-five children were enrolled. Forty-seven percent (8/17) of the children that received addition of Mel to the BuCy conditioning, versus none (0/8) in the BuCy group, developed GI-aGVHD (stages 2-4) (p<0.05). Paper II. Based on 68 children with 91 endoscopic occasions, treatment changes in response to histopathology reports occurred in 48% (44/91). Paper III. Seventy children with 92 endoscopic occasions were assessed. Histopathology-based GI-GVHD diagnosis was established in 67 of 92 (73%) endoscopic occasions in the RIHA and in 50 of 92 (54%) in the clinical standard histopathological assessment (p=0.014). The risk of a subsequent re-endoscopy within one-year post-HSCT was higher in endoscopic occasions with GI-GVHD solely detected in RIHA versus non-GI-GVHD in both assessments (p=0.005). Paper IV. Forty-four children with 51 endoscopic occasions were analyzed. Biopsies from the rectosigmoid area had 85% sensitivity for RIHA-based GI-GVHD diagnosis. The corresponding figure for combined biopsy sampling from both rectosigmoid area and upper gastrointestinal tract was 97% and was similarly high compared with biopsies collected from complete lower endoscopy.

Conclusions: I) Addition of Mel to the BuCy conditioning increased the incidence of symptom-based GI-aGVHD in children with JMML and MDS. II) Endoscopy-guided histopathological assessment was found to influence the treatment decisions and should therefore be considered in children with GI-aGVHD. III) In children with symptom-based GI-aGVHD, without confirmation of the diagnosis by clinical standard histopathological assessment, a second histopathological assessment based on the NIH 2014 criteria should be considered before performing a re-endoscopy. IV) Sigmoidoscopy combined with upper endoscopy, colonoscopy/ileocolonoscopy, or full upper and lower endoscopy should be considered as preferred choices for the endoscopic procedure in children with clinically suspected GI-aGVHD.