Transplant patients on immunosuppression represent a risk group for opportunistic infections, including adenoviral infections. The clinical and histopathologic findings of posttransplant adenoviral enteropathy in 3 adult intestinal transplant patients are described. The histopathologic pitfalls for the differential diagnosis between intestinal adenoviral enteropathy and acute rejection are discussed. Adenoviral enteropathy is an opportunistic infection that may follow aggressive treatment for small bowel allograft rejection, but which may require no specific treatment. It is associated with mild mixed inflammation in the lamina propria and slight increase of crypt apoptosis, resembling low-grade acute rejection. The identification of characteristic viral inclusions in the surface or crypt epithelium points to the diagnosis of adenoviral enteropathy, in spite of the increased crypt apoptosis.

Adenovirus, a common pathogen that would normally cause mild, self-limited respiratory, gastrointestinal, and/or ocular disease in the immunocompetent host, has been identified as a potential cause of increased morbidity and mortality in some transplant recipients.^1,2 The diseases associated with adenoviral infection in this group of patients range from local mild disease to widely disseminated/systemic disease.^2–6 The prevalence of adenoviral enteritis (AVE) could be as high as 20.8% in the pediatric population,⁷ where it is usually, but not always, a mild disease.³ Acute cellular rejection (ACR), on the other hand, if left untreated, has the potential to result in graft loss. Unfortunately, morphologically differentiating AVE from ACR can be difficult.^8,9 Also, the “natural” history of AVE in small bowel allograft is unclear, and while many reports in children suggest a self-limited infection in many cases, AVE-related deaths have been reported in the adult population.⁹ What is generally accepted, however, is that misinterpreting AVE as rejection, as with some other opportunistic infections, could portend grave outcomes. In this study, therefore, we report the findings in 3 of our patients with confirmed AVE who were followed up for 311 to 1092 days. The follow-up management and outcomes are described. The importance of and the helpful features for differentiating AVE from ACR are emphasized.

Adult small bowel transplant recipients who had their engraftments and follow-up at the University of Pittsburgh Medical Center, and were followed up clinically and with biopsies as indicated or as per protocol, constitute the baseline population for this report. A total of 101 small bowel transplants were performed at the University of Pittsburgh Medical Center between 2001 and 2004 (47 males and 54 females; M:F = 2:2.3). Patients' ages ranged from 17.6 to 67.8 years (mean and median of 44.7 years and 46.3 years, respectively). Most of these patients had received either a Campath-1H or Thymoglobulin conditioning regimen at the time of transplantation. Of these, 3 patients were identified with a confirmed diagnosis of adenoviral enteropathy by stool culture and electron microscopy, and were followed up for 311 to 1092 days following the diagnosis of AVE. Electron microscopy in each case was performed on formalin-fixed, paraffin-embedded tissue material as follows: The area of interest was removed from the paraffin block and placed into xylene for 9 hours to remove the paraffin; during this period, the xylene solution was changed 3 times. The tissue was then rehydrated by placing for 15 minutes each in ethanol of different concentrations (2 of 100%, 1 each of 95%, 70%, and 50%). The sample was then placed into three 15-minute changes of 0.1M sodium cacodylate buffer. The tissue was postfixed in 1% osmium tetroxide for 1 hour, dehydrated through graded alcohols, infiltrated using propylene oxide and an EMbed 812-Araldite mixture, and embedded in EMbed 812-Araldite. Blocks were polymerized overnight at 60°C and sectioned on a Reichart UltraCut S. Sections were stained with uranyl acetate and lead citrate and viewed on a FEI Philips CM12 electron microscope with an AMT (Advanced Microscopy Techniques, Boston, Mass) imaging system.

A 47-year-old man had a second small bowel transplant for short-gut syndrome, following intestinal volvulus. His initial graft had experienced multiple ACRs and had been lost to chronic rejection, having lasted for 341 days. He received a transplant in 2003 but developed a mild ACR 10 days posttransplant. Given his history of multiple severe ACRs in the first graft, he received 5 days of OKT-3 with steroid pulse. The rejection initially resolved, only to recur on day 57, and this time was treated with Campath-1H. The specimen from a follow-up biopsy on day 64 showed resolution of the rejection. However, material from another biopsy on day 66 showed the presence of adenovirus-type nuclear inclusions (Figures 1 and 2 ) with mild lymphoplasmacytic inflammation and up to 5 apoptotic bodies per 10 crypts. The inclusions stained negatively by immunohistochemistry for adenovirus antigens, using a mouse monoclonal adenovirus antibody (Adenovirus Ab-5, clone M58 + M73; Lab Vision Corp, Fremont, Calif). However, stool culture and electron microscopy on formalin-fixed, paraffin-embedded tissue (Figure 3, A and B ) confirmed adenoviruses. The increased number of apoptoses was interpreted as relating to adenoviral infection, and so the patient received no therapeutic intervention. Subsequent biopsy specimens revealed persistent infection for about 2 weeks, but this later cleared. The patient remained well for several weeks. However, he lost his graft again on day 311 to chronic rejection. The explanted failed graft revealed no adenoviral inclusions.

A 55-year-old woman with superior mesenteric vein thrombosis received a small bowel transplant in 2002. She developed ACR on day 24, for which she was initially treated with a steroid bolus, and later, with 6 days of OKT-3 due to nonresponse. The rejection initially resolved but then recurred on day 41, and the OKT-3 therapy was repeated. This time there was no resolution, as she remained symptomatic, and the results of subsequent biopsies revealed persistent rejection. She received 3 doses of Campath-1H on day 56, followed by resolution of rejection. Material from a follow-up biopsy on day 73, however, revealed adenovirus-type nuclear inclusions, with mild lymphoplasmacytic inflammation and up to 10 apoptotic bodies per 10 crypts. The inclusions, as in the first case, also stained negatively by immunohistochemistry for adenovirus antigens, using the same antibody. However, follow-up electron microscopy confirmed adenoviruses. The patient continued on her regular maintenance immunosuppressants. Specimens from several repeat biopsies showed persistent inclusions but with no evidence of worsening injury, inflammation, or graft dysfunction. By day 150, however, the biopsy specimens from both the loop and chimney revealed much fewer apoptotic bodies (2–3 apoptotic bodies per 10 crypts) and no evidence of viral inclusions. At the last follow-up (3 years after these events), the patient is still doing well with a functioning graft.

A 55-year-old woman with short-gut syndrome following Crohn disease received an allogenic small bowel transplant in November 2004. She did well, and the specimen from a surveillance biopsy on day 175 posttransplantation showed an increase in lamina propria lymphoplasmacytic infiltrates, associated with increased crypt apoptoses and typical adenovirus nuclear inclusions. These were subsequently confirmed as adenoviruses on electron microscopy, in spite of negative immunohistochemistry, using the same antibody as in the previous patients. She received no additional immunosuppression, and repeated follow-up biopsy specimens revealed disappearance of these inclusions after several weeks. She continues to do well and is free of symptoms as of day 497 of follow-up (ie, 322 days of postadenovirus follow-up).

This report presents 3 patients with adenoviral enteropathy. All of our patients had been maintained on tacrolimus-based immunosuppression. Two of these patients showed features of adenovirus enteropathy at posttransplant days 66 and 73, respectively, following aggressive treatment for rejection, including Campath-1H and/or OKT-3. The third patient was asymptomatic, and her diagnosis was made from the results of a surveillance biopsy on day 175 posttransplantation. She had received Campath-1H pretreatment at the time of transplantation. Histopathologic features associated with this virus include increased lymphoplasmacytic mucosal inflammation and increased crypt apoptosis (Figures 1 and 2 ). The number of apoptoses on 1 occasion was within the range one would consider sufficient for ACR. While ACR was considered as a possibility, the understanding of the clinical context of these biopsy findings, and the acknowledgment of the potential overlap between rejection and AVE prompted a decision not to further increase immunosuppression, which in retrospect proved to be right. The morphologic findings supporting AVE, rather than rejection, include (generally) fewer apoptoses; nuclear enlargement with “dusky” inclusions; and supporting culture and/or electron microscopic findings. Other authors have found more neutrophils in AVE patients,⁹ but neutrophils did not constitute a significant population of cells in any of our patients. The most specific of all these are the nuclear inclusions, which incidentally may not always be present in every tissue sample. The other features are not only subjective but also defy a sharp cut-off line between the extent of inflammation and apoptoses that would reliably separate rejection from AVE.

The importance of correctly differentiating AVE from ACR stems from the likelihood of systemic dissemination of adenovirus infection in (excessively) immunosuppressed individuals.^5,10 In this light, therefore, the policy for evaluation of small bowel allografts should encourage obtaining multiple tissue samples by the transplant physicians, to be complemented by examination of multiple levels, so as to optimize the chances of detecting viral inclusions in the appropriate situations. It is apparent from our findings, as well as those of others, that crypt apoptoses do not always imply cellular rejection. Therefore, the understanding of the clinical context of every biopsy specimen cannot be overemphasized. Adenoviral infections tend to follow an increase of immunosuppression beyond the baseline. Therefore, the persistence of inflammation and crypt apoptoses in patients who have otherwise been sufficiently treated for ACR should warrant a careful search for other causes of crypt injury, such as AVE. In difficult cases, electron microscopy and/or other ancillary methods, including molecular techniques, should also be considered.

As with other opportunistic viral infections, the risk for AVE is greater with increasing immunosuppression, and may particularly complicate aggressive treatment of rejection episodes. Unlike other viruses, however, adenoviral infections appear to produce milder disease and are potentially self-limited; the greater challenge is their recognition and differentiation from cellular rejection. Also, adenovirus does not appear to be a significant risk factor for cellular rejection, as is the case with some other viral infections. Pathologists interpreting small bowel allograft biopsy material should be familiar with the appearance of adenoviral inclusions and the associated pattern of inflammation with increased crypt apoptoses, and should be aware that negative immunohistochemistry does not necessarily exclude this diagnosis. In our laboratory, the antibody appears to have low sensitivity, as diagnostic inclusions were present on the sections stained. However, other reasons for immunohistochemistry failure may include focal distribution of diagnostic inclusions and/or suboptimum tissue preservation, among others. Further studies may hopefully uncover more sensitive antibody clones in the future.

The authors acknowledge Ms Ardith Ries of the Department of Pathology, University of Pittsburgh Medical Center, for her assistance with electron micrograph production.

Click on thumbnail for full-sized image.