Graft assessment for acute rejection after intestinal transplantation: current status and future perspective

Abstract

Intestinal transplantation has since its inception evolved as a lifesaving treatment option for patients with irreversible intestinal failure who can no longer be sustained on parenteral nutrition. Improvement in short-term survival after transplantation has also justified the expansion of treatment indications. Unfortunately, success is somewhat limited by a plateau observed in long-term survival. The reason for this sub-optimal long-term result experienced in this cohort may in part be attributed to the intestinal graft with the lymphoid content it carries inflicting the host with multiple complications where acute cellular rejection is one of the most common causes for graft loss. Graft monitoring is for this reason of paramount importance and detection of rejection at an early stage essential to enable early instigation of treatment and successful reversal of the pathology. Due to the challenges in diagnosing acute rejection with a noninvasive marker we are still limited to a surveillance protocol using endoscopy and biopsies for the diagnosis of rejection. The purpose of our paper is to review the adequacy of different methods in monitoring the graft for acute rejection using biomarkers, endoscopy and imaging. In conclusion, the evidence base continues to support the use of histology for the diagnosis of acute rejection. The role of biomarkers are still debatable, although markers such as calprotectin might be beneficial in excluding an ongoing process.

Keywords:

• Intestinal transplantation
• multivisceral transplantation
• surveillance
• rejection
• biomarkers
• endoscopy
• imaging

Introduction

Intestinal transplantation is defined as a process of grafting an intestinal component (jejunoileum), the development of which was essential to address the issue of intestinal failure and its corresponding complications [1]. Intestinal failure is characterized by an inability of the gastrointestinal tract to meet nutritional demands and thus requiring either temporary or permanent parenteral nutrition to maintain health and/or growth [2]. Most often this condition is caused by short bowel syndrome due to surgical resection, although motility diseases, mucosal defects, obstruction and fistula may also progress to intestinal failure [3]. Overall, the long term survival for patients with non-malignant intestinal failure on PN is excellent (5 year survival 64–91% [4,5]). Therefore, intestinal transplantation is only reserved for patients who have failed parenteral nutrition and developed life-threatening complications [6]. Other indications include lifesaving therapy in portomesenteric thrombosis, frozen abdomen and benign or slow-growing tumour involving the hilum of the liver [7–9]. The three main intestinal transplant procedures that are currently being performed are isolated small bowel, multivisceral (graft including the stomach) and combined liver-intestine grafts. Although, many variations of these procedures exist [1,10].

The short term survival following intestinal transplantation has improved over the last decade. However, the rates of graft failure have not improved beyond the first year and the long-term survival remains essentially the same: 76, 56, and 43% at 1, 5, and 10 years respectively [11]. The reason for the high mortality observed in this group is due to the immunologic complications (e.g., rejection, sepsis post-transplant lymphoproliferative disorders) that these patients succumb to where acute cellular rejection and sepsis remains the most frequent causes for graft loss [1,11]. The proportion of patients contracting acute cellular rejection during the first year varies from 31% among adult patients with a liver containing graft to 60% among children with an isolated intestine [12]. The detection of rejection at an early stage is therefore crucial for early instigation of treatment and successful reversal of the pathology. Likewise, delayed detection may result in severe exfoliative rejection and/or patient death [10]. Unfortunately, due to the challenges in diagnosing acute rejection with a noninvasive marker we are still limited to a surveillance protocol using endoscopy and biopsies for the diagnosis of rejection. However, it is important for clinicians to know the potential value of different diagnostic modalities. We herein review data on different methods in monitoring the graft for acute rejection using biomarkers, endoscopy, and imaging (Table 1).

Table 1. The different methods used for detecting acute cellular rejection.

Methods for detecting acute cellular rejection
Modality	Usefulness
Biomarkers
Citrulline	Limited value. Excludes advanced rejection but not necessarily early stages.
Calprotectin	Potential value in excluding an ongoing process.
Granzym B & Perforin	Not used in clinical practice Variable expression during the first months and immunosuppression decreases the expression of the molecules.
Microbiota	Insufficient data.
DSA	Insufficient data. Role for detection and daily monitoring is unclear.
Endoscopy
Standardized videoendoscopy	Histology is the mainstay for diagnosis. Endoscopy alone has shown to be inadequate especially in symptomatic patients.
Zoom endoscopy	Insufficient data. Conflicting results regarding its usefulness.
Video capsule endoscopy	Useful in certain scenarios, for example, rejection is suspected despite normal findings from standardized endoscopies. The procedure is limited by the risk of capsule retention and the inability of obtaining biopsies.
Imaging
CT & MRI	Insufficient data. MRI may not be feasible in compromised patients
Ultrasound	Insufficient data. Unlikely to reveal early signs of rejection.

Biomarkers

Citrulline

Citrulline is an amino acid that is derived predominantly from enterocytes in the small bowel. This marker has been studied in several conditions, but primarily in short bowel syndrome and is considered a reflection of intestinal function or mass and a predictor of nutritional outcome [13,14]. Therefore, the hypothesis was to predict if decreased measurements of citrulline could detect the enterocyte damage seen in rejection and if so which levels were considered relevant.

The importance to define the normal variation in citrulline after transplantation was addressed by a group from New York, they showed a significant reduction of citrulline within the first months after transplantation and with a gradual increase during the first 3 months. The values then leveled off from the third to sixth month. Thereafter, the levels seemed to be similar to that in healthy individuals. This period is also the most critical time period for developing rejection and the results from citrulline need to be interpreted with caution during the first months [15,16]. After the occurrence of rejection in the earlier months, a delay to normalize citrulline was noted and in certain cases, this normalization never occurred [16]. With regard to rejection, the Miami group has shown that the citrulline levels were inversely proportional to the degree of rejection. High sensitivity and specificity were also seen in adults and children (96 and 69%) for moderate and severe rejection using a cut-off of 13 µmoles/L. Similarly, a higher negative predictive value was seen (95 and 99%, respectively) using the cut-off of 20 for all types of rejection and 10 for moderate/severe rejection but with a lower positive predictive value (14% in children and 29%), a lower sensitivity was also seen for mild rejection (70%). The group concluded that citrulline is excellent for excluding rejection [17,18]. However, the New York group revealed the difficulties in differentiating rejection from other conditions such as viral enteritis [19]. Adjusting the cutoff for renal function has also been shown to be of importance [20]. In summary, multiple factors need to be considered when measuring citrullines such as time elapsed after surgery, renal function and infection. Furthermore, the benefit of this test is limited due to the difficulties in depicting earlier stages of rejection and in obtaining a rapid result. Nevertheless, it seems to be useful to exclude advanced rejection.

Calprotectin

Calprotectin is a protein that is found in the neutrophilic cytosol and its fecal concentration increases in situations that cause mucosal infiltration of leukocytes. The use of fecal calprotectin is predominantly for the monitoring of inflammatory bowel disease [21]. Therefore, the evaluation of calprotectin as a biomarker for graft rejection was suggested [22] but potential difficulties in different cut-off values for different manufacturers exist. Furthermore, calprotectin is better in detecting alterations in the colon than in the small bowel [23].

In an early pilot study from Omaha, calprotectin was collected from the ileostomy of 68 patients and 35 controls with correlation to histologic findings. A receiver–operator characteristic analysis was utilized and a fecal calprotectin of 92 was found to predict rejection with 77% sensitivity and 83% specificity [22]. On the follow-up study by the same group, 732 samples were measured from 72 patients at frequent time points. These results showed significant interpatient variability with difficulties in proposing generalized cut-off values [24]. Similarly, studies from Buenos Aires (11 patients, 137 measurements) and Miami (29 patients, 122 measurements) have shown high sensitivity of 76–100% but a low specificity of 47% for rejection thus highlighting the difficulty in identifying rejection. Nevertheless, low levels of Calprotectin might be used as an assurance of not having intestinal pathology [25,26]. Similarly, our centre uses calprotectin for surveillance purposes in asymptomatic individuals where endoscopy is troublesome. Furthermore, routine evaluation with calprotectin may have an additional benefit in conjunction with lower gastrointestinal endoscopy which is to guide further management in the advent of high calprotectin and normal endoscopy.

Granzyme B and perforin

Allograft rejection is believed to be caused by cytotoxic T cells that recognize its target cells through mediators such as perforin and granzyme B [27]. The relevance of identifying an increased expression of these molecules as an early biomarker for intestinal rejection was therefore investigated [28–30].

The Bologna group examined these markers with the aim to correlate the markers to common transplant-related complications such as rejection, virus (cytomegalo virus, Epsein-Barr virus) and post-transplant lymphoproliferative disorders. A significant difference was noted between the different conditions with the highest accuracy seen with rejection (Sensitivity and specificity of granzyme 80%, 79% and perforin 70%, 79% respectively). Although almost 1000 samples were collected during the study half of them were discarded since these patients had a combination of conditions. Furthermore, due to the variable expression of the molecules during the first month, the test was only useful after this period. The test was not either useful in following disease progression during an episode of rejection since high doses of immunosuppressive drugs decreased the concentration of both molecules [28]. Therefore, the relevance of these tests is questionable and has not yet reached clinical practice.

Microbiota

The gut microbiota provides important effects in metabolism, nutrition, and immune function, it has also been associated with inflammatory bowel disease, obesity, and a range of other conditions [31,32]. Due to the complex interactions were seen between the microbiome and the immune system this field poses an interesting target in the field of intestinal transplantation. However, insufficient data exist describing the association between microbiota composition and transplantation outcomes [33,34].

The Omaha group aimed to characterize the ileal microbiota composition in patients with acute rejection, pre-rejection and no rejection using deep sequencing (16 s ribosomal RNA gene tags) from ileal effluent. In total, 19 patients were examined using 35 samples. In the advent of rejection, the proportion of phylum Firmicutes and Lactobacillales were decreased while the phylum Proteobacteria significantly increased. Furthermore, the ROC analysis showed that the proportion of bacteria such as Firmicutes may differ between rejection and no rejection [35].

Similarly, shifts in microbiota have also been observed during chronic rejection in animal models [36]. However, experience from kidney transplantation has revealed a high inter-individual expression and a longitudinal measurement for each individual was therefore suggested [33]. Although these results are interesting, further studies using a larger cohort are needed and the question of whether the alterations of the microbiota are causal or consequential remains unanswered [35].

Donor-specific antibodies

Donor specific antibodies are defined as antibodies that are produced against the donor’s antigen (DSA). In transplantation, this definition often refers to antibodies directed towards donor HLA but DSA does also exist against donor ABO antigens and other non-HLA antigens too [37]. These HLA-binding antibodies may exist before the transplantation or develop after the procedure and may persist even after immunosuppression. In solid organ transplantation, DSA has been associated with decreased graft survival and is linked to both acute rejection and chronic rejection after intestinal transplantation [38–41]. In contrast, grafts containing a liver have shown to have a reduced risk for DSA since they are more likely to clear preformed DSAs [39]. The monitoring of DSA is therefore frequently used. However, the optimal management of DSAs is still under debate and improved treatment regimens are emerging [41,42]. The presence of DSA may therefore be viewed as an independent predictor of acute rejection but not as an indicator of rejection itself.

Non-HLA antibodies (antiangiotensin II type I receptor antibodies and antiendothelin type A receptor antibodies) have also been described as potential facilitators of rejection according to a study from Berlin. Formation of these antibodies may actually precede HLA antibodies (DSA) and thus trigger a more forceful immunologic response. Furthermore, patients with non-HLA antibodies also had a higher frequency of rejection than controls (80 vs 55%) [43]. These markers might be useful in clinical practice but needs to be investigated in larger studies priorly.

Endoscopy

Endoscopy with mucosal biopsies remains the golden standard for diagnosing rejection. Therefore, centers perform frequent endoscopies in patients with signs of graft rejection as well as for surveillance [1]. The endoscopic features (erythema, friability, granularity and ulcerations) alone have not been sufficient for diagnosis but instead rely heavily on histology [44]. Histologic features consist of architectural distortion, inflammation in the lamina propria, crypt apoptosis and epithelial injury [45,46]. The given treatment is thereafter decided based on the clinical situation, endoscopic findings and histology. Furthermore, the features are sequentially followed through repeated endoscopies to evaluate the result of antirejection treatment [47]. However, this strategy is flawed especially in smaller units with less access to experienced transplant pathologists and endoscopists.

The endoscopic features seen during rejection are from either patchy to widespread and will progress as the condition worsens. In patients who are symptomatic with a normal ileoscopy the importance to obtain proximal histology from the jejunum also needs to be mentioned even if isolated rejection here is rare [48,49]. Furthermore, experience from the Pittsburgh group has shown that visual inspection alone will miss 6% (n = 32) of rejections from asymptomatic patients and 25% (n = 68) from symptomatic patients, thereby highlighting the importance of obtaining histology [49]. Over the last decade, the quality of the endoscopic equipments have further improved especially after implementing high definition endoscopy systems resulting in an improved correlation with histology [47]. Nevertheless, the overall sensitivity is still inadequate using endoscopy alone. To address this issue the Miami group investigated the use of zoom endoscopy. Zoom endoscopy had an improved specificity (98%) in adult patients but similar sensitivity (45%) [50] when compared to other studies (sensitivity 45–69%, specificity 79–83% [47,49,51]) Interestingly they chose only to treat rejections visible to the eye and this resulted in a spontaneous recovery in most cases (98%,58 episodes). The advantage of using zoom endoscopy over standardized video endoscopy was not experienced by the Pittsburgh group [49]. Yet, another hurdle in the interpretation of the endoscopic findings persists as the community lacks a universal grading system for the severity of acute cellular rejection. Therefore, the Gothenburg group constructed a five-stage endoscopy score [47] and its usefulness is currently being evaluated through a prospective multicentre trial. Moreover, the value of performing regular surveillance endoscopies has recently been questioned by a group in New York. They compared a protocol where patients (n = 25) only received endoscopies while having clinical symptoms (increased stoma output, fever, bacteraemia) with a historical cohort (n = 28) on a surveillance protocol. No difference in the incidence of rejection was seen and similar graft and patient survival between the two cohorts were seen. In total, a reduction of the number of endoscopies from 226 (historical cohort) to 57 (symptomatic cohort) was observed. Although, there were limitations regarding the sample size and the follow-up time (2014–2016) the concept is intriguing [52]. Nevertheless, the absence of a reliable biomarker and the fact that rejection might occur in asymptomatic patients makes this strategy somewhat challenging to adopt especially for smaller centres. Furthermore, detecting rejection in a later phase is a probable risk with this strategy which may result in an inability to treat rejection effectively.

Videocapsule endoscopy

The use of VCE was evaluated in a case series in Gothenburg. Video capsule endoscopy was used in 12 instances in 7 patients following intestinal transplantation. Although the authors deemed the examination useful in most examination (83%), the complication with retention was present in two cases and agreement between VCE and histology for inflammation/rejection was moderate (Kappa 0.79) [53]. Similarly, an abstract from a multicenter study including 33 VCE in 23 patients showed the usefulness of 73% (23 cases) with a change in management in 19% (6 cases) and a retention rate of 9% (2 cases) [54]. Although these data are interesting they also reveal the limitations with regards to the risk of retention and inability to obtain histology which makes the case for graft surveillance difficult.

Imaging

The surgical anatomy might be challenging following transplantation due to the variation in surgical approaches and a review with the surgical team is essential if abnormal findings are suspected. The appearance of the transplanted bowel is similar to that of a normal gut, although the bowel is often centrally placed within the abdomen [55]. A reperfusion injury alongside a graft that lacks innervation with diminished lymphatic drainage is often seen as a slight dilation of the intestinal segments with luminal dilation (up to 4 cm) or bowel wall thickening. Furthermore, mesenteric edema (91%), small volume mesenteric nodes (59%), and mild fold/wall thickening (65%) of 3–5mm are also commonly seen. These findings are important to differentiate from a pathologic process and are seen during the first 3 months [55–57]. Albeit, these alterations in the small bowel and mesentery may also subside within days after the procedure [58].

Computed tomography (CT) and magnetic resonance imaging (MRI)

Acute rejection is a histologic diagnosis using endoscopic biopsies with a distribution that might be patchy or outspread. However, frequent imaging for other causes may raise the question of rejection. Findings from multidetector CT with intravenous contrast have shown that non-specific diffuse bowel wall thickening (95%), mucosal hyperenhancement (71%) along with moderate volume ascites (95%) is present in varying frequencies during rejection [56,59]. These findings are suggestive but not diagnostic for rejection. Hence, the importance of a clinical picture to guide further management. Furthermore in severe rejection complications such as intestinal necrosis and bowel infarction or bowel ischaemia may follow. Although there are a few reports describing their experience with MRI its usefulness still remains unclear and similar findings to CT have been reported [55,59,60]. The use of positron-emission tomography and CT (PET-CT) in the presence of post-transplant lymphoproliferative disorders (PTLD) have shown benefits. However, its accuracy in detecting acute cellular rejection remains unknown [58].

Ultrasound (US) and laser Doppler

Immediately after the operation, the US serves as an important investigation for documenting the patency of graft vasculature and anastomosis in children [61]. However, its role in the surveillance and for detection of rejection is not established. The advantage of visualizing the bowel in real-time sonography is appealing but according to early results, this has been hampered by impaired visibility due to intraluminal intestinal gas and abdominal dressings [62,63]. Nonetheless, recently an abstract from Cambridge did reveal some advantages with this investigation showing that in 7 patients’ sonographic features with mural thickening, decreased peristalsis and mesenteric vascularity had correlation with rejection [64]. Laser Doppler is another technique that involves inserting a fiberoptic probe into the ostomy and mucosal perfusion was measured using laser Doppler flowmetric monitoring. In a pilot study, 5 patients underwent 75 separate recordings after intestinal transplantation. A sudden decrease in the mucosal blood flow was seen prior to two instances of rejection, although other pathologies showed similar reductions. Larger studies are needed to evaluate its usefulness in the setting of graft monitoring [65].

Discussion

Biomarkers pose as the ideal strategy for monitoring the immune system but have so far proven insufficient. With markers such as calprotectin and citrulline being unable to show sufficient results, they can be viewed at best as reasonably helpful as exclusionary tests only. Over the last years, much attention has been drawn towards DSA but the role for daily monitoring and detection of rejection are not fully known. Possibly a combination of markers might be more useful with the aim of detecting rejection at an early phase. With the current popularity seen in metabolomics, identification of new markers is possible although larger studies are needed [66,67] Therefore, serial mucosal biopsies are the only route for diagnosing acute cellular rejection. Since visual impression has failed to reach an adequate detection rate potentially this could improve if artificial intelligence (AI) using deep learning would be implemented in future practice. AI will likely revolutionize the field of gastroenterology with results showing improved detection rate of neoplastic lesions and the capacity to characterize alterations [68]. A further improvement might be to switch over to endocytoscopy that enables us to visualize images with >500-fold magnification [69], along with AI systems that subsequently characterize the lesion itself making histology unnecessary. However, this is only but the beginning of AI and as of now no systems are adapted for detecting rejection but focuses solely on colorectal polyps and malignancies. Furthermore, imaging with radiological methods (CT, MRI, US) have failed to undergo systematic studies with statistical methods but the possibility to detect early signs of rejection with these methods is unlikely but not improbable as detection with newer MR cameras is yet to be reported. Therefore, at present these methods may have an additive effect in detecting the consequences of rejection but not confirmatory for rejection per se.

Disclosure statement

The author report no conflict of interest.

Additional information

Funding

Funding for this study was provided by grants from the Swedish state under the agreement between the Swedish government and the country councils [ALFGBG-875431].