**Background:** Acute graft-versus-host disease (a-GvHD) is a major complication of allogeneic stem cell transplantation (allo-SCT). Gastrointestinal (GI) involvement presents with symptoms that overlap with infections, chemotherapy effects, and drug toxicities, making diagnosis challenging. While histological evaluation of mucosal biopsies can confirm a-GvHD, the role and optimal approach to GI endoscopy in children remains debated. This study aimed to describe clinical, endoscopic, and histological findings in children undergoing GI endoscopy for suspected GI a-GvHD, evaluate rates of steroid-resistant GvHD and transplant-related mortality (TRM), and assess procedure safety.
**Methods:** This observational retrospective study was conducted at IRCSS Istituto G. Gaslini in Genova, Italy, including all children who underwent GI endoscopy for suspected a-GvHD after allo-SCT between January 2000 and December 2017. Among 348 allo-SCTs performed, GI involvement was clinically diagnosed in 50 children (24%), and 26 of those (52%) underwent endoscopy. Indications for endoscopy were: (a) clinical signs of GI a-GvHD without other organ involvement, or (b) persistent GI symptoms after partial resolution of multiorgan a-GvHD. Procedures included esophagogastroduodenoscopy (EGD), pan-colonoscopy (PC), flexible sigmoidoscopy (FS), or rectal suction biopsy, chosen based on presenting symptoms and risk assessment. Preconditions included platelet count >50,000/mmc and normal coagulation. Histological grading used a 4-point validated scale (adapted from Cruz-Correa et al.): grade I (increased crypt apoptosis), grade II (apoptosis with crypt abscess), grade III (individual crypt necrosis), grade IV (total mucosal denudation). Viral PCR was performed on blood and mucosal biopsies.
**Key Results:** Among the 26 patients (median age 9.5 years, 65% male), 10 (38.5%) underwent endoscopy for isolated GI symptoms and 16 (61.5%) for persistent symptoms after multiorgan therapy. Procedures included: lower GI tract in 19 (73%; PC in 11, FS in 3, rectal suction biopsy in 5), upper GI tract (EGD) in 4 (15.4%), and combined in 3 (11.6%). Diarrhea was the most common symptom (84.6%). Median time from SCT to biopsy was 55.5 days (range 21–148). Clinical GI a-GvHD grades at endoscopy were: grade 1 (4%), grade 2 (38%), grade 3 (12%), grade 4 (46%). Histology confirmed a-GvHD in 19 patients (73.1%); 7 (26.9%) had non-specific findings. Among confirmed cases, histological grades were: grade I (11 patients), grade II–III (4 patients), grade IV (4 patients). Viral PCR on mucosal samples was positive in 6 patients (23.1%; CMV, adenovirus, EBV, HHV6, or combinations), all of whom also had histological confirmation of a-GvHD. Sensitivity of histology was 57.1% for upper endoscopy and 81.8% for lower endoscopy. High clinical grade (3–4) was significantly associated with histological confirmation (P = 0.026). Steroid-resistant GvHD occurred in 78.9% of histology-positive vs. 28.6% of histology-negative patients (P = 0.028). At median follow-up of 94.4 months, 10 patients died (38.4%). Three-year OS was 61.5% (95% CI, 40.3–77.1); 3-year TRM was 22.1% (95% CI, 9.8–45.3). In histology-negative patients, 3-year OS was 71.4% (95% CI, 25.8–92.0) and TRM was 0%, compared to 57.9% (95% CI, 33.2–76.3) and 28.9% (95% CI, 13.2–56.2) in histology-positive patients. One child developed a duodenal intraparietal hematoma after combined endoscopy, which resolved spontaneously within 2 months; no other acute adverse events occurred.
**Clinical Implications:** This study demonstrates that GI endoscopy with biopsy is feasible and safe in children after allo-SCT, with a 73.1% histological confirmation rate for clinically suspected GI a-GvHD. Lower GI endoscopy, particularly flexible sigmoidoscopy, showed higher diagnostic sensitivity (81.8%) than upper endoscopy (57.1%) and was not associated with complications. Histological confirmation of GI a-GvHD was strongly associated with steroid-resistant disease and higher TRM (28.9% vs. 0% at 3 years), suggesting that histology provides important prognostic information. The authors recommend flexible sigmoidoscopy with rectal biopsy as the preferred initial approach due to its favorable risk-benefit profile, especially in children. Limitations include the retrospective design, selection bias (only 52% of clinically diagnosed patients underwent endoscopy), and small sample size. Prospective studies are needed to define the precise role of histology in guiding treatment decisions.