Post-Transplant Complications Gastroparesis Following Bone Marrow Transplantation

Bone Marrow Transplantation (2001) 28, 59–62  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt Post-transplant complications Gastroparesis following bone marrow transplantation

DA Eagle1, V Gian2, GY Lauwers1, JC Manivel3, JS Moreb1, S Mastin1 and JR Wingard1

1University of Florida College of Medicine, Gainesville, FL; 2Sarah Cannon Cancer Center, Nashville, TN; and 3University of Minnesota College of Medicine, Minneapolis, MN, USA

Summary: Keywords: gastroparesis; hematopoietic cell transplant; gastric emptying; nausea and vomiting Patients often develop nausea, vomiting and bloating after bone marrow transplantation (BMT). These symptoms may interfere with nutrition and the ability to take oral medications. Gastroparesis is a recognized cause of Following bone marrow transplantation (BMT), some these symptoms in non-transplant patients but less is patients develop persistent nausea, vomiting, bloating, early known about patients who undergo BMT. Between satiety and decreased tolerance to oral nutrition and medi- January 1996 and March 1997, a total of 151 patients cations. These symptoms frequently develop weeks after underwent BMT. Eighteen patients (12%) developed transplantation and inhibit weight gain, optimal nutrition persistent symptoms suggestive of gastroparesis and the transition from intravenous to oral medications. (persistent nausea, vomiting or bloating). Scintigraphic Gastroparesis is a disorder of motor function of the stom- gastric emptying studies were performed to assess for ach in which gastric emptying is delayed. Patients with gas- gastroparesis. Prokinetic agents were administered at troparesis often develop symptoms of nausea, vomiting, the time of study. The records on these patients were bloating and distension. Gastroparesis has many known compared with those of all other patients undergoing causes which include atrophic gastritis, diabetes, hypothy- BMT during the same time period without these symp- roidism, uremia, scleroderma, anticholinergics and stress.1 toms. Nine patients who demonstrated delayed gastric Additionally, cytomegalovirus (CMV), herpes simplex emptying were further evaluated with esophagastroduo- virus (HSV) and other viral infections have been implicated denoscopy and biopsy. Biopsy samples were reviewed as causes of delayed gastric emptying.1,2 We performed for evidence of graft-versus-host disease (GVHD). Four- gastric emptying studies on patients presenting with persist- teen of 18 patients demonstrated delayed gastric empty- ent nausea, vomiting and bloating following allogeneic or ing and most responded to prokinetic agents given at matched unrelated donor transplantation to assess for the time of study. Age, conditioning regimen, cytomega- underlying gastroparesis and to ascertain which factors lovirus antigenemia and acute GVHD did not appear were associated with the development of gastroparesis. to be associated with the development of gastroparesis. Allogeneic BMT recipients were at higher risk than autologous BMT patients (26% vs 0%, P Ͻ 0.0001). Of Methods allogeneic BMT recipients, there was a nonsignificant trend of patients receiving tacrolimus to be less likely Patients to experience gastroparesis than those receiving cyclosporine (27% vs 48%, P = 0.08). For the nine patients Between January 1996 and March 1997, all patients who undergoing upper endoscopy, GVHD on gastric biopsy presented after resolution of the acute toxicities of the con- was an uncommon finding and was mild when present. ditioning regimen with persistent symptoms of nausea, Gastroparesis appears to be a common cause of nausea, vomiting, bloating or early satiety at the University of Flor- vomiting and bloating following allogeneic BMT. This ida BMT clinic or inpatient unit were evaluated with Tech- may occur less often with tacrolimus than cyclosporine netium 99m-labeled egg scintigraphy. The study was per- because of the former agent’s prokinetic properties. formed by the method described by Johnson et al3 with Patients usually respond to prokinetic drugs at the time modifications by Fahey et al.4 of scintigraphy. GVHD and CMV infection do not A delay in gastric emptying was defined as a gastric emp- appear to be major contributing factors. Bone Marrow tying half-time of greater than 90 min. Intravenous erythro- Transplantation (2001) 28, 59–62. mycin and/or metochlopramide doses were frequently administered during these studies to assess their effect on gastric motility.5 A new gastric emptying half-time was Correspondence: Dr JR Wingard, Division of Hematology, University of Florida College of Medicine, 1600 SW Archer Road, PO Box 100277, determined after administration of these agents. If the slope Gainesville, FL 32610–0277, USA of the gastric emptying half-time did not change, then this Received 6 July 2000; accepted 24 January 2001 was reported as no response. Nine of the patients who were Gastroparesis following BMT DA Eagle et al 60 found to have delayed gastric emptying were further evalu- on scintography compared with those who were found to ated with esophagogastroduodenoscopy and biopsy. The have normal gastric emptying (54 vs 112 days; P = 0.011) decision to perform endoscopy was based on the rec- (Mann–Whitney U test used for calculation of differences ommendations of the consulting gastroenterologist and was with small numbers). often to look for possible infectious etiologies. Degree of the delay in gastric emptying, as well as magnitude of response to prokinetic agents for patients demon- Pathology strating delayed gastric emptying are given in Table 2 and illustrated in Figure 1. All 11 patients who received Biopsy samples were reviewed by pathologists at the Uni- erythromycin responded (100%), while eight of 12 patients versity of Florida and the University of Minnesota for the who received metoclopramide responded (67%). For most presence and grading of possible GVHD. All gastric patients, erythromycin accelerated gastric emptying much biopsies were fixed in buffered formalin and processed rou- more than metoclopramide. tinely. At least two hematoxylin and eosin-stained slides Nine patients with delayed gastric emptying by scinti- were reviewed for each case. The diagnosis of GVHD was graphic assessment underwent upper endoscopy. These made following the criteria put forth by Snover et al6 results are presented in Table 3. Most patients had endos- including apoptosis with karyorrhectic debris in glandular copic evidence of gastritis (n = 6) or other types of epithelium, gland destruction, dilated gastric glands con- inflammation (n = 3) on visual inspection. Two patients had taining eosinophilic granular debris and a sparse lympho- mild acute GVHD and two others had histologic findings cytic infiltrate. The degree of GVHD was quantified using suggestive of but not definitive of GVHD. None showed a three-tier grading system of mild, moderate and severe. histologic evidence of CMV or HSV gastritis. No consistent biopsy findings were apparent for patients demonstrating Data collection gastroparesis. Use of total body irradiation in the conditioning regimen, The records on all patients who underwent scintigraphic age, acute clinical GVHD at any site and CMV antigenemia assessment were subsequently reviewed. Information were not associated with the development of gastroparesis extracted from the records included patient age, time to (Table 4). Of patients developing gastroparesis, only two development of symptoms, development of acute clinical of 14 (14%) patients had undergone transplantation from GVHD at any site, conditioning regimen, immunosuppres- an unrelated donor (Table 5). Of patients not developing sive regimen, CMV antigenemia within 8 weeks of scinti- gastroparesis, 18 of 40 (45%) had undergone transplan- graphic assessment and biopsy results from endoscopy, if tation from an unrelated donor. Patients developing gastro- performed. Additionally, the records were reviewed for all paresis were less likely to have undergone transplantation other patients undergoing allogeneic and matched unrelated from an unrelated donor: 14% vs 45% (P = 0.038). Of the donor transplant during the same time period. Data 20 patients who underwent transplantation from an unre- extracted from the records for these patients included the lated donor, 16 (75%) of these patients received tacrolimus. type of conditioning regimen, immunosuppressive regimen, Of the 34 patients who underwent matched sibling trans- presence of acute clinical GVHD at any site and cytomega- plant, six (17%) received tacrolimus. Of the patients with lovirus antigenemia. gastroparesis, three of 14 (21%) received tacrolimus. Of the patients who did not develop gastroparesis, 19 of 40 (47%) received tacrolimus. There was a trend toward less gastro- Results paresis for patients receiving tacrolimus as part of the immunosuppressive regimen: 21% vs 47% (P = 0.08). A total of 151 patients aged 17 years or older underwent BMT at the University of Florida during the study period. Ninety-seven received autologous or stem cell transplan- Discussion tation, 34 received allogeneic sibling matched transplantation and 20 received matched unrelated donor (MUD) Following matched sibling or matched unrelated donor transplantation. Eighteen patients (12%) developed signs transplantation, patients occasionally develop persistent and symptoms suggestive of gastroparesis following trans- nausea, vomiting, bloating and lack of appetite long after plantation. Clinical characteristics and results of scinti- the acute toxicities of the conditioning regimen abate. This graphic assessment of these patients are presented in study suggests that these symptoms are often due to delayed Table 1. gastric emptying and that gastroparesis should be con- Fourteen of the 18 patients demonstrated delayed gastric sidered in the differential diagnosis. Twenty-six percent of emptying on scintigraphy and were confirmed to have gas- allogeneic patients who underwent bone marrow transplan- troparesis. None of the 97 autologous BMT recipients tation during the study period demonstrated delayed gastric developed gastroparesis compared to 14 of the 54 allo- emptying by scintigraphy, indicating that this entity is geneic BMT recipients (26%, P Ͻ 0.00001). Median onset common. was 54 days following transplantation, but presentation was Delayed gastric emptying has many potential etiologies. as early as 34 days or as late as 113 days following trans- Those etiologies which have been previously described and plant. The median time to development of symptoms in may be causative in the setting of BMT include stress, viral symptomatic patients evaluated for gastroparesis was infection including both CMV and HSV, and medication shorter for those confirmed to have delayed emptying time side-effects such as anticholinergic effects or anti-motility

Bone Marrow Transplantation Gastroparesis following BMT DA Eagle et al 61 Table 1 Clinical characteristics and gastric emptying study results in patients with persistent symptoms suggestive of gastroparesis

Patient Sex Age Conditioning Onset (days Immuno- Acute CMV Type of Technetium egg Response to Number regimen post- suppressive GVHD antigenemia transplant result prokinetics transplant) regimena

1 M 54 Cy/TBI 35 Tacrolimus No Neg Sibling Delayed Y 2 F 43 VP-16/TBI 113 Prednisone/CsA Yes Neg Sibling Delayed Y 3 F 31 Melph/TBI 76 CsA No Neg Sibling Delayed Y 4 F 33 VP-16/TBI 48 CsA Yes Neg Sibling Delayed Y 5 M 46 VP-16/TBI 59 Prednisone/CsA Yes Neg Sibling Delayed Y 6 F 33 Cy/TBI 108 CsA Yes Pos Sibling Delayed ND 7 M 53 Cy/Ara-C/TBI 60 CsA Yes Neg Sibling Delayed Y 8 F 57 Cy/Busulfan 49 CsA No Neg Sibling Delayed Y 9 M 31 Cy/Ara-C/TBI 38 CsA No Neg Sibling Delayed Y 10 M 40 VP-16/TBI 48 CsA Yes Pos Sibling Delayed Y 11 M 39 VP-16/TBI 48 Tacrolimus Yes Pos MUD Delayed Y 12 M 47 Cy/TBI 61 Tacrolimus Yes Neg Sibling Delayed Y 13 M 36 Vp-16/TBI 34 CsA Yes Pos Sibling Delayed Y 14 M 48 Cy/Ara-C/TBI 78 CsA Yes Neg Sibling Delayed Y 15 F 31 VP-16/TBI 128 CsA Yes Neg Sibling Normal 16 F 39 Melph/TBI 87 CsA No Neg Sibling Normal 17 M 20 Cy/Ara-C/TBI 133 CsA No Pos Sibling Normal 18 M 58 Cy/TBI 97 Prednisone/Tacrolimus Yes Pos MUD Normal

Cy = cyclophosphamide; TBI = total body irradiation; Melph = Melphelan; Ara-C = cytarabine; CsA = cyclosporine; MUD = matched-unrelated; ND = not done; Y = yes. aShort-course methotrexate was also given to all patients.

Table 2 Gastric emptying times and response to prokinetics 40596.00 Metoclopramide given Patient Initial emptying Emptying half-time Emptying half-time 30447.00 number half-time (min) following following Erythromycin given metoclopramide erythromycin (min) (min) 20298.00

1 134 43 5 2 132 45 22 Counts per second 10149.00 3 150 Not given 10 4 135 No response 10 5 300 50 15 0.00 6 212 Not given Not given 0 21 42 63 84 105 7 240 No response 12 Time (in min) 8 134 No response 18 9 180 70 32 Figure 1 The degree of delay in the gastric emptying, as well as the 10 240 68 62 magnitude of response to prokinetic agents, for patients demonstrating 11 150 10 Not given delayed gastric emptying. In this scan, there was no emptying spon- 12 No emptying 3 Not given taneously or after a metoclopramide challenge, as evidenced by no drop 13 134 50 15 in counts, but a prompt, precipitous emptying after erythromycin adminis- 14 165 No response 43 tration.

faster in patients receiving tacrolimus compared with effects of opiates.1 Other plausible etiologies related to patients receiving cyclosporine.11 In that study, gastric transplantation include autonomic neurologic damage sec- emptying times were normal in all patients studied and ondary to the preparative regimen or GVHD. Gastroparesis cyclosporine-treated patients only had relatively slower has been described in recipients of solid organ transplants gastric emptying. This result may have been secondary to and in patients undergoing autologous transplantation.7–10 prokinetic properties of tacrolimus which is a macrolide.11 No clear etiology has been found in patients receiving solid Only three of 14 patients in our study with delayed gastric organ transplants while chemotherapy effects have been emptying received tacrolimus, with a trend toward less gas- postulated as a potential cause in stem cell transplant troparesis for patients receiving tacrolimus when compared patients.7–10 Signs and symptoms of gastroparesis occurred with patients taking other types of immunosuppression. only in patients receiving allogeneic or matched unrelated Scleroderma is an autoimmune disease known to cause donor transplants in our study. gastroparesis and has many shared clinical features with Based on our review of the literature, clinical gastropar- chronic GVHD.1 Mild GVHD was documented in some esis is not a commonly recognized side-effect of either gastric biopsies, but was not a consistent ﬁnding and seems cyclosporine or tacrolimus. However, in patients who have unlikely to be a major causative factor. Additionally, the received renal transplants, gastric emptying is signiﬁcantly lower incidence of delayed gastric emptying in matched-

Bone Marrow Transplantation Gastroparesis following BMT DA Eagle et al 62 Table 3 Findings of upper endoscopy and biopsy

Patient Visual appearance Biopsy result

1 Nonspecificinflammation No GVHD 2 Gastritis Mild acute GVHD 4 Gastric erosions Possible mild GVHD; chronic active gastritis 6 Gastritis No GVHD; foveolar hyperplasia 7 Gastritis No GVHD; reactive gastropathy, mild fibrosis 9 Gastritis, ulcers and erosions Mild acute GVHD; focal neutrophilic gastritis 10 Gastritis and erosions Possible mild GVHD; superficial gastritis 12 Gastritis No GHVD 13 Esophagitis and duodenitis No GVHD

Table 4 Transplant characteristics of patients with and without gastro- immunosuppression other than tacrolimus may be contribu- paresis tory. The retrospective design of this study limits strong inferences from our data. Further prospective studies Factor Gastroparesis No gastroparesis P value = = designed to elucidate the cause of gastroparesis and assess (n 14) (n 40) the clinical response to prokinetic agents are needed. Age (median) 40 40 0.459 TBI 13/14 (93%) 33/40 (82%) 0.32 Acute clinical 10/14 (71%) 27/40 (68%) 0.53 References GVHD at any site 1 Quigley EM. Gastric and small intestinal motility in health CMV 4/14 (29%) 16/40 (40%) 0.33 antigenemia and disease. Gastroenterol Clin N Am 1996; 1: 113–145. 2 Jaehoon JO, Kim CH. Gastroparesis after a presumed viral illness: clinical and laboratory features and natural history. Mayo Clin Proc 1990; 65: 636–642. 3 Johnson DA, Winters C, Drane WE et al. Solid-phase gastric Table 5 Other transplant characteristics of patients with and without gastroparesis emptying in patients with Barrett’s esophagus. Dig Dis Sci 1986; 31: 1217–1220. 4 Fahey F, Zeissman HA, Collen M, Eggli D. Left anterior Factor Gastroparesis No gastroparesis P value (n = 14) (n = 40) oblique projection and peak-to-scatter ratio for attenuation compensation of gastric emptying studies. J Nucl Med 1989; 30: 233–239. Unrelated donor 2/14 (21%) 18/40 (45%) 0.038 (vs sibling donor) 5 Domstad P, Kim EE, Coupal JJ et al. Biologic gastric empty- Tacrolimus 3/14 (21%) 19/40 (47%) 0.08 ing time in diabetic patients, using Tc-99m-labeled resin-oat- meal with and without metoclopramide. J Nucl Med 1980; 21: 1098–1100. 6 Snover DC, Weisdorf SA, Vercellotis GM et al. An histopa- unrelated donor transplant patients (5% vs 38%) further thologic study of gastric and small intestinal graft-versus-host suggests that GVHD is not the major contributory factor to disease following allogeneic bone marrow transplantation. Hum Path 1985; 16: 387–392. the development of gastroparesis. Our institutional practice 7 Berkowitz N, Schulman L, McGregor C, Markowtiz D. Gas- is to favor tacrolimus over cyclosporine for patients receiv- troparesis after lung transplantation. Potential role in postoper- ing matched-unrelated donor transplants. The prokinetic ative respiratory complications. Chest 1995; 108: 1602–1607. properties of tacrolimus may explain the decreased inci- 8 Au J, Hawkins T, Venables C et al. Upper gastrointestinal dence of gastroparesis for patients receiving matched-unre- dysmotility in heart–lung transplant recipients. Ann Thorac lated donor transplants.11 The presence of CMV viremia Surg 1993; 55:94–97. did not appear to be related to the finding of delayed gastric 9 Brand R, DiBaise J, Quigley EM et al. Gastroparesis as a emptying on scintigraphic assessment and viral cytopathic cause of nausea and vomiting after high-dose chemotherapy changes were not seen in gastric biopsy specimens. and haemopoietic stem-cell transplantation. Lancet 1998; Our findings suggest that delayed gastric emptying is a 352: 1985. 10 Hecht JR, Lembo T, Chap L. Prolonged nausea and vomiting potential cause of persistent nausea, vomiting and other after high dose chemotherapy and autologous peripheral stem symptoms of dyspepsia following transplant. Patients usu- cell transplantation in the treatment of high risk breast carci- ally respond to prokinetic agents. For patients with signs noma. Cancer 1997; 79: 1698–1702. and symptoms of gastroparesis following allogeneic BMT, 11 Maes BD, Vanwalleghem J, Kuypers D et al. Differences in further evaluation with scintigraphic assessment or an gastric motor activity in renal transplant recipients treated with empiric trial of prokinetics would be reasonable. No sole FK-506 versus cyclosporin. Transplantation 1999; 68: causative factor could be identified, although the use of 1482–1485.

Bone Marrow Transplantation