Introduction

Hematopoietic stem cell transplantation (HSCT) is increasingly performed for acute and chronic leukemia and other hematologic malignancies, and marrow failure states.¹ Gastrointestinal (GI) symptoms including diarrhea, nausea, vomiting, abdominal cramping, anorexia and gastrointestinal bleeding are common within the first 100 days following HSCT,^{2, 3} and may be severe enough to warrant gastroenterological evaluation in up to 40% of patients.⁴ While these symptoms may be caused by acute graft-versus-host disease (GVHD), they may also be attributable to chemoradiation toxicity, medication side effects, or a variety of bacterial, fungal, viral and parasitic infections.^{2, 5} Endoscopic findings in GVHD, linked to stage of disease, range from normal mucosa to erythema, edema, erosions, ulcerations and mucosal sloughing.^{5, 6, 7} In a retrospective study of patients undergoing upper gastrointestinal (UGI) endoscopy and biopsy between days 20 and 100 following HSCT, no significant differences in symptoms or endoscopic appearance were observed between patients with or without UGI GVHD.⁸ Given the nonspecific symptoms and endoscopic findings of acute GVHD and the significant side effects of immunosuppressive medications used to treat GVHD, a histologic diagnosis should be sought before treatment is initiated.^{2, 8} However, the reported mucosal site with the highest diagnostic yield (upper and/or lower) varies between studies, and thus the best diagnostic approach to GVHD remains undefined.

To better evaluate the utility of proximal colonic and ileal biopsies, we performed colonoscopy with left sided, right sided and ileal biopsies in patients with diarrhea within 20–100 days of transplant. We sought to compare the yield of upper and lower gastrointestinal biopsies to determine the optimum initial endoscopic evaluation of patients suspected of having GVHD with diarrhea developing within 20–100 days of HSCT. We also graded endoscopically visualized lesions as proposed by Kreisel et al.¹ to examine the concordance of endoscopic and histological grading of acute GVHD.

Patients and methods

From May 1999 until May 2003, all patients at the University of Alabama at Birmingham following bone marrow or stem cell transplantation were identified prospectively from among those patients referred for gastrointestinal evaluation for diarrhea. Patients with a prior history of GI GVHD were excluded as were patients developing diarrhea prior to 20 days following HSCT since cytoreductive therapy may mimic the histology of acute GVHD.⁹ All patients were interviewed by one of the investigators (CMW) prior to the endoscopic evaluation and the following data were recorded: age, gender, underlying disease and transplantation type, duration of diarrhea, stools per day, stool volume, weight loss, anorexia, nausea, vomiting, odynophagia, dysphagia, gastrointestinal bleeding, fevers, skin rash, and prior or current history of GVHD of the skin or liver. Laboratory studies including liver chemistry tests and results of prior skin or liver biopsy were also recorded. A complete blood count and liver profile were obtained on all patients on the day of endoscopic examination. Stool studies, which included fecal leukocytes, bacterial culture, C. difficile toxin and ova and parasite examination, were performed on all patients prior to endoscopic examination and were negative in all patients. During the study period, cytomegalovirus (CMV) antigenemia was monitored up to twice weekly during the first 3 months after transplant for CMV seropositive patients while the patient was receiving steroids, and once monthly for seronegative patients. The study was approved by our institutional review board, and written informed consent was obtained from each patient for both the study and the endoscopic procedures.

Histocompatibility and stem cell source

The patients enrolled were recipients of a matched (n=19) (MRD) or partially matched sibling related transplant (n=3) or a matched unrelated donor transplant (n=2) (MUD). Of the three patients who were only partially matched, two were a four of six antigen match and one patient was a 9 of 10 antigen match. Peripheral blood hematopoietic stem cells were collected from donors in all but three cases. The remaining three donors included the two donors identified through the National Marrow Donor Registry, and had their hematopoietic stem cells harvested from bone marrow. The remaining bone marrow donor was a child donating for a sibling.

Preparative regimen and GVHD prophylaxis

Myeloablative conditioning regimens varied considerably, as listed in Table 1. The BuCy+/-VP regimens were in use prior to 2000 at which time we changed to a busulfan/fludarabine preparative regimen. Two patients with relapsed disease underwent a second myeloablative hematopoietic transplant and received either thiotepa and fludarabine, or melphalan and TIB. Cyclosporine and standard dose methotrexate^{10, 11} was initiated in all adult patients for GVHD prophylaxis. Pediatric patients received cyclosporine and 'short course' methotrexate.^{12, 13, 14}

Table 1 - Preparative regimens used for transplantation.

Full table

Endoscopic protocol

All endoscopies were performed or supervised by one of the authors (CW) using Olympus (Lake Success, NY, USA) endoscopes. Ileal intubation was attempted at the time of colonoscopy in all patients. A minimum of four biopsies was taken from the left colon, right colon and ileum. Upper endoscopy (EGD) was performed following lower tract examination. At the time of upper endoscopy, four duodenal biopsies were taken followed by a minimum of four biopsies from the antrum, and two from the body and fundus. Where present, biopsies were taken from regions of endoscopic mucosal abnormalities. Abnormal endoscopic findings were graded as proposed by Kreisel et al.¹

Histological evaluation

Biopsies from all sites were graded based on a scale adapted from McDonald and Sale¹⁵ as follows:

Grade I – Single cell necrosis (apoptosis) noted on medium power.
Grade II – Evidence of epithelial damage by crypt/glandular abscesses, epithelial flattening or glandular/crypt dilation.
Grade III – Dropout of one or more crypts/glands.
Grade IV – Total epithelial denudation.

Evidence of viral infection was assessed by immunohistochemistry for CMV (on stomach, small bowel and colon) and by immunohistochemistry for adenovirus (on small bowel and colon).

The results are reported as a percentage of positive biopsies based upon biopsy site, with a calculation of 95% confidence limits around this point estimate. A positive biopsy for GVHD at any site was used as the gold standard to diagnose GVHD.

Results

During the study period, 135 patients underwent allotransplantation at our center, including 93 matched related donors, 31 matched unrelated donors and 11 haploidentical donors. A total of 40 patients without a prior history of GI GVHD undergoing endoscopic evaluation for GI symptoms suggestive of GVHD were identified. Sixteen patients were excluded as they had undergone BMT more than 100 days previously. Demographic information, symptoms and laboratory results of the remaining 24 patients are presented in Table 2. Endoscopic and histologic findings are summarized in Table 3. Twenty one of the 24 patients (88%) had GVHD diagnosed by histology from at least one biopsy site. The three patients with negative biopsies did not subsequently have GVHD diagnosed and two patients had subsequent spontaneous resolution of diarrhea. The percentage of positive biopsies based upon anatomic location is provided in Table 4. The distal colon was the site with the highest diagnostic yield for GVHD.

Table 2 - Baseline characteristics of the study group.

Full table

Table 3 - Endoscopic features, histologic grading, and stage of gut disease.

Full table

Table 4 - Percent positive biopsies by site.

Full table

Percentage of positive biopsies for patients with nausea or vomiting

The percentage of positive gastric, duodenal, ileal, right colonic and left colonic biopsies in patients with nausea or vomiting ranged from 64 to 82% (Table 5). One of three patients with negative gastric biopsies had duodenal biopsies diagnostic of GVHD. One of three patients with negative duodenal biopsies had gastric biopsies diagnostic of GVHD. Of the two patients with negative gastric and duodenal biopsies, one patient had positive left colonic biopsies and one patient only had positive biopsies from the ileum.

Table 5 - Percent positive by biopsy site in patients with concomitant nausea or vomiting.

Full table

Percentage of positive biopsies of single endoscopy and combined strategies

Positive biopsies from EGD, flexible sigmoidoscopy, colonoscopy, colonoscopy with ileal intubation and EGD combined with flexible sigmoidoscopy are shown in Table 6. Both colonoscopy with ileoscopy and EGD with flexible sigmoidoscopy had positivity rates of greater than 90%. These strategies were also compared in patients with nausea or vomiting (Table 7). Both colonoscopy with ileoscopy and EGD with flexible sigmoidoscopy had biopsy positive rates of 93%.

Table 6 - Percentage of positive biopsies based upon various diagnostic strategies.

Full table

Table 7 - Percentage of positive biopsies based upon various diagnostic strategies in patients with concomitant nausea or vomiting.

Full table

Comparison of endoscopic and histological grading and clinical staging

The endoscopic and histological grade of each biopsy site was tabulated (Table 3). Overall endoscopic and histological grading was in agreement in 38.9%. The difference in endoscopic and histological grading was 1 in 33%. Discordance of 2, 3 and 4 was present in 19.4%, 7.8% and 1%, respectively. The difference between the endoscopic and histological grading was 0–1 in 71.9% of biopsy sites. Seventeen of 38 (44.7%) biopsy sites that were endoscopically normal were consistent with acute GVHD by histology. Clinical staging of GVHD was largely discordant with endoscopic and histologic findings (Table 3). There were only four cases (16.7%), where the clinical, endoscopic and histopathologic findings were concordant.

In three cases, patients lacked clinical staging evidence of GVHD as published by Glucksberg et al.¹⁶ yet had histopathologic evidence of GVHD. In two of these cases, the patients had transient diarrhea but also had persistent nausea and vomiting. Using the modified criteria for clinical staging of GVHD,¹⁷ these two patients would have clinical stage 1 GVHD. The third patient had upper and lower GI symptoms but was also found to have C. difficile toxin and CMV colitis. He had rapid resolution of all of his symptoms, was never treated with systemic steroids and was not classified as having GVHD by the attending physician.

Discussion

There is a lack of information regarding the utility of both colon and ileal biopsies in patients with diarrhea following BMT at risk for acute GVHD. Thus, we evaluated the yield of right and left sided colonic biopsies and attempted ileal biopsies in all patients undergoing endoscopic evaluation for diarrhea following BMT. Our study revealed that both colonoscopy with ileal intubation and EGD with flexible sigmoidoscopy were equally effective in diagnosing acute GVHD and were more sensitive than either EGD or colonoscopy alone. This held true in patients with nausea or vomiting in addition to diarrhea. However, performing sigmoidoscopy and biopsy had a very high yield and may represent an ideal initial strategy. The overlapping confidence intervals for the various diagnostic strategies suggest that further studies may help in both confirming and extending our findings.

Previous studies have addressed the yield of gastric, duodenal and rectosigmoid biopsies in the diagnosis of acute GVHD. In a prospective study of patients with the onset of diarrhea in addition to UGI symptoms from day 20 to 100 post-transplant, 26 patients were identified with biopsy-proven intestinal GVHD by upper endoscopy and sigmoidoscopy. Biopsies were taken from the stomach, duodenum and rectosigmoid. Gastric biopsy was positive in 85% while duodenal and rectosigmoid biopsy were each positive in 58%.⁵ In a retrospective study of patients with GI symptoms from day 20 to 100 post-transplant, 77 patients underwent upper endoscopy and sigmoidoscopy either on the same day or within 7 days. Seventy-five percent of patients studied had anorexia, nausea or vomiting; 27% had upper abdominal pain; and 21% had diarrhea. Of the 42 patients with biopsy-proven GVHD, 81% had positive UGI biopsies and 67% had positive rectal biopsies.⁸ Although it has been suggested that gastric biopsy is more sensitive than rectosigmoid biopsy in the diagnosis of GVHD,^{5, 8} we are unaware of any prior nonautopsy studies examining the diagnostic utility of proximal colonic and ileal biopsies in acute GVHD.

Previous studies have suggested that UGI biopsies are superior to rectal or rectosigmoid biopsies in the diagnosis of GVHD;^{5, 8} however, we found EGD and flexible sigmoidoscopy had similar sensitivities. The increased yield of left colon biopsies in our study may have resulted from obtaining biopsies from regions of the sigmoid colon in comparison to the rectum alone. Importantly, in one of the previous studies only 21% of patients undergoing lower tract evaluation had diarrhea.⁸

A previous study found a difference of 0–1 between macroscopic and histological grading of acute GVHD in 97.6% of endoscopies.¹ However, we noted a difference of 2–4 grades between endoscopic and histological grading in 28.2% biopsy sites. A histological diagnosis of acute GVHD was made in 44.7% of endoscopically normal biopsy sites emphasizing the need to biopsy normal as well as abnormal-looking mucosa at the time of endoscopy. Histopathologic grading must not be used for overall grading of GVHD in HSCT patients. Outcome predictions regarding responses and transplant-related mortality are determined utilizing clinical staging and grading criteria for GVHD.

Several limitations of our study deserve comment. We had a surprisingly high rate of GVHD and believe this may be related to the heightened clinical suspicion of the treating physicians. Given the side effects of immunosuppressive therapy, we rely on the histological diagnosis of GVHD to institute long-term immunosuppressive therapy for GVHD and therefore feel that histology does alter our management. Also, as our study was designed to examine patients with diarrhea, we cannot comment on the prevalence of GVHD in patients with isolated UGI symptoms. Positive biopsies of the stomach or duodenum demonstrating GVHD are required to make the diagnosis of GI GVHD in patients without diarrhea (revised Glucksberg grading). However, our study shows for the first time that colonoscopy with ileal intubation and biopsy is an acceptable alternative to currently used strategies for the evaluation of post-BMT diarrhea. Lastly, we also examined a large number of MRD as compared to MUD and the results could perhaps be different between the two groups.

Based on our findings, we recommend sigmoidoscopy and biopsy for patients at risk for GVHD who complain of diarrhea given its diagnostic yield, safety, ease of performance and cost. If endoscopic examination of the distal colon is abnormal and the endoscopic findings are highly suggestive of GVHD, no further evaluation may be needed pending biopsy results. Sigmoidoscopy could be supplemented with upper endoscopy especially in the patient with upper GI symptoms warranting investigation, and this strategy, along with colonoscopy with ileal biopsy, were equivalent for the diagnosis and had the highest overall diagnostic yields. Biopsies should be performed of normal appearing mucosa as a positive histological diagnosis can still be established.

References

1. Kreisel W, Herbst EW, Schwind B, Ochs A, Olschewski M, Kochling G et al. Diagnosis and grading of acute graft-versus-host disease following allogeneic bone marrow transplantation by sigmoidoscopy. Eur J Gastroenterol Hepatol 1994; 6: 723–729.
2. Bombi JA, Nadal A, Carreras E, Ramirez J, Munoz J, Rozman C et al. Assessment of histopathologic changes in the colonic biopsy in acute graft-versus-host disease. Am J Clin Pathol 1995; 103: 690–695. | PubMed | ChemPort |
3. Ferrara JL, Deeg HJ. Graft-versus-host disease. N Engl J Med 1991; 324: 667–674. | PubMed | ISI | ChemPort |
4. Fallows G, Rubinger M, Bernstein CN. Does gastroenterology consultation change management of patients receiving hematopoietic stem cell transplantation? Bone Marrow Transplant 2001; 28: 289–294. | Article | PubMed | ChemPort |
5. Cox GJ, Matsui SM, Lo RS, Hinds M, Bowden RA, Hackman RC et al. Etiology and outcome of diarrhea after marrow transplantation: a prospective study. Gastroenterology 1994; 107: 1398–1407. | PubMed | ISI | ChemPort |
6. Goker J, Haznedaroglu IC, Chao NJ. Acute graft-versus-host disease: pathobiology and management. Exp Hematol 2001; 29: 259–277. | Article | PubMed | ISI | ChemPort |
7. Ponec RJ, Hackman RC, McDonald GB. Endoscopic and histologic diagnosis of intestinal graft-versus-host disease after marrow transplantation. Gastrointest Endosc 1999; 49: 612–621. | Article | PubMed | ChemPort |
8. Roy J, Snover D, Weisdorf S, Mulvahill A, Filipovich A, Weisdorf D. Simultaneous upper and lower endoscopic biopsy in the diagnosis of intestinal graft-versus-host disease. Transplantation 1991; 51: 642–646. | PubMed | ChemPort |
9. Epstein RJ, McDonald GB, Sale GE, Shulman HM, Thomas ED. The diagnostic accuracy of the rectal biopsy in acute graft-versus-host disease: a prospective study of thirteen patients. Gastroenterology 1980; 78: 764–771. | PubMed | ChemPort |
10. Santos GW, Owens AH. Production of graft-versus-host disease in the rat and its treatment with cytotoxic agents. Nature 1966; 210: 139–140. | Article | PubMed | ChemPort |
11. Storb R, Epstein RB, Graham TC, Thomas ED. Methotrexate regimens for control of graft-versus host disease in dogs with allogeneic marrow grafts. Transplantation 1970; 9: 204–246.
12. Storb R, Deeg HJ, Pepe M, Doney K, Appelbaum F, Beatty P et al. Graft-versus-host disease prevention by methotrexate combined with cyclosporine compared to methotrexate alone in patients given marrow grafts for severe aplastic anaemia: long term follow-up of a controlled trial. Br J Haematol 1989; 72: 567–572. | PubMed | ChemPort |
13. Storb R, Deeg HJ, Whitehead J, Appelbaum F, Beatty P, Bensinger W et al. Methotrexate and cyclosporine compared with cyclosporine alone for prophylaxis of acute graft-versus-host disease after marrow transplantation for leukemia. N Eng J Med 1986; 314: 729–735. | ChemPort |
14. Storb R, Deeg HJ, Pepe M, Appelbaum F, Anasetti C, Beatty P et al. Methotrexate and cyclosporine versus cyclosporine alone for prophylaxis of graft-versus-host disease in patients given HLA-identical marrow grafts for leukemia: long-term follow-up of a controlled trial. Blood 1989; 73: 1729–1734. | PubMed | ChemPort |
15. McDonald GB, Sale GE. The human gastrointestinal tract after allogeneic marrow transplantation. In: Sale GI, Shulman HM (eds). The Pathology of Bone Marrow Transplantation. Masson: New York, 1984, p. 83.
16. Glucksberg H, Storb R, Fefer A, Buckner CD, Neiman PE, Clift RA et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A matched sibling donors. Transplantation 1974; 18: 295–304. | PubMed | ISI | ChemPort |
17. Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 1994 consensus conference on acute GVHD grading. Bone Marrow Transplant 1995; 15: 825–828. | PubMed | ISI | ChemPort |