• Acute cellular rejection;
  • impact of rejection;
  • prognostic factors;
  • small intestine transplantation


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Intestinal transplantation has evolved over the years with major improvements in patient and graft survival. Acute cellular rejection of the intestine, however, still remains one of the most challenging aspects of postoperative management. We analyzed retrospectively collected data from 209 recipients of primary intestinal grafts at our institution over the past 11 years. A total of 290 episodes of biopsy-proven rejection requiring clinical treatment were analyzed. Rejection episodes doubled in length, on average, with each increasing grade (mild, moderate, severe). We observed increased incidence of overall rejection and particularly severe rejection in recipients of isolated intestinal and liver-intestine grafts in comparison with multivisceral grafts. Two rejection history variables had a significant negative impact on graft survival: the occurrence of a severe rejection episode and a rejection episode lasting ≥21 days. The lower incidence rate of severe rejection in recipients of multivisceral grafts might be due to a combination of increased donor lymphatic tissue and larger load of donor-derived immune competent cells present in the graft. The development of more effective monitoring and treatment protocols to prevent the occurrence of severe and/or lengthy rejection episodes is of critical importance for intestinal graft survival.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Transplantation of the intestine for the treatment of irreversible intestinal failure has evolved over the past 10 years from a procedure performed only in a few patients by a few centers to a real therapeutic option offered by many institutions and in multiple countries. The latest statistics from the International Bowel Transplant Registry show that 19 countries and 61 centers now perform intestinal transplantation (1). In the United States alone, 302 intestinal transplants were performed between 1995 and 1999; in the following 5 years (2000–2004) the number rose to 564 cases (2). Advancements in the management of intestinal grafts both in terms of surveillance and treatment of complications, along with improved immunosuppression regimens, have improved graft and patient survival at 1 year to well above 70% (3,4). An ‘Achilles' heel’ of intestinal transplantation has always been acute cellular rejection. Prevention, early recognition and prompt treatment of rejection are closely interconnected with the immunosuppressive regimens used (induction, maintenance therapy, choice of immunosuppression), types of organs transplanted (with or without liver graft, or as part of a multivisceral graft) and the existence (or lack thereof) of concurrent infectious complications (viral infections, posttransplant lymphoproliferative disorders, etc.). At our institution, we have performed over 200 intestinal transplants (combining adult and pediatric cases) during the past 11 years. We elected to retrospectively review and analyze each episode of acute cellular rejection in recipients of primary intestinal grafts with the goal of identifying any patterns that could predict the clinical course of the rejection episode. First, we defined what constituted a rejection episode, based on clinical, endoscopic and histopathological data. We then attempted to determine the overall severity of each rejection episode, length of time with rejection and treatment(s) given for each episode. Finally, in addition to determining the influence of factors such as type of graft and recipient age (i.e. pediatric vs. adult) on the incidence rate of developing rejection, we analyzed the impact of the patient's rejection history on graft survival and its cause-specific components.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References


During the period taken into consideration (August 1994 through February 2006), all recipients of primary intestinal grafts were included, with at least 3 months follow-up. Our analysis focused on rejections occurring in primary grafts. Types of transplant included isolated intestinal grafts (I), combined liver-intestinal grafts (LI), multivisceral grafts (MVT, defined as en block replacement of liver, stomach, pancreas and intestine) or modified multivisceral grafts (MMVT, defined as en block transplantation of stomach, pancreas and intestine) (5,6). Large intestine was also included in some patients as well as donor spleen in some of the more recent MVT and MMVT grafts. The techniques for all the abovementioned transplants have been previously described (5,7).


Induction immunosuppressive treatment changed over the years (6,8–10). We utilized either no induction or induction with muronomab-CD3 (OKT3, Ortho-Biotech, Raritan, NJ) from 1994 through 1997. This was followed starting in 1998 by the use of daclizumab (Zenapax®, Roche Pharmaceuticals, Nutley, NJ) administered intravenously (2 mg/kg/dose) at the time of transplant and then intermittently (2 mg/kg/dose on day 7, 14, then every other week through 3 months, followed by 1 mg/kg i.v. every other week for 3 additional months) during the first 6 months after transplant. Starting in 2001, alemtuzumab (Campath 1H®, Berlex Laboratories, Montville, NJ) was utilized as induction treatment (0.3 mg/kg/dose × 2 intravenously on day of transplant and on days 3 and 7) as previously described. The initial experience with Campath 1H in the pediatric population was halted when a significant increase in morbidity and mortality was observed (6,10). Currently, Campath 1H is utilized only for patients 4 years of age and older, while daclizumab is utilized for the younger recipients. Baseline immunosuppression consisted of tacrolimus (Prograf®, Fujisawa Pharmaceuticals, Deerfield, IL), with trough levels maintained between 15 and 20 ng/mL during the first 3 months in daclizumab induction recipients, and 10–15 ng/mL in Campath 1H recipients. Subsequently, tacrolimus trough levels were kept at 8–12 ng/mL. Maintenance steroids (methylprednisolone, Solumedrol®, Pharmacia and Upjohn Co., Kalamazoo, MI) were not used in patients who received Campath-1H induction, but were utilized in all other recipients, being weaned off at approximately 6–9 months after transplant.

Monitoring of graft

Starting in 1998, protocol evaluation of the intestinal graft with endoscopies was performed on all transplant recipients, with biopsies sent for histological evaluation. In patients older than 2 years of age, zoom magnification endoscopy was utilized in order to better assess the intestinal mucosa (11,12). Endoscopies were performed 2–3 times weekly during the first month, followed by once weekly during months 2–3 and once monthly during months 4–6 postoperatively. When clinical signs and symptoms of rejection were present, and during rejection episodes, endoscopies were performed as often as every other day, then spaced longer as the rejection episode resolved.

Histopathologic grading of rejection

All specimens were sent to the Division of Immunopathology at our institution, including samples received from outside hospitals. Experienced transplant pathologists evaluated the samples by hematoxylin/eosin stain and, if needed, performed additional studies to rule out infectious diseases. The current nomenclature used by the transplant pathologist community for grading of acute cellular rejection was followed (13). Readings of ‘no rejection’ or ‘indeterminate’ were called negative. Mild, moderate and severe rejections were graded as positive biopsies and assigned accordingly for subsequent analysis. In those cases where a biopsy was read as ‘mild to moderate’ or ‘moderate to severe’, the higher grade category was assigned.

Definition of rejection episodes

We defined rejection episodes as those intervals of time when all of the following criteria were fulfilled: clinical signs/symptoms of rejection present (including endoscopic appearance of intestinal mucosa), positive histological findings and a corresponding increase in immunosuppressive therapy. The first day of rejection was assigned as the first positive biopsy on histology; the last day of the episode was assigned as the day of the first of two negative consecutive biopsies at the end of treatment cycle. The duration of the rejection episode was calculated as the number of days between the two dates. Alternatively, the end of the rejection episode was assigned when the patient was discharged from the hospital (for a resolved episode), had the graft explanted or expired.

The overall grade of a rejection episode was defined by the highest reading of all biopsies performed during that episode (i.e. the maximum observed grade).

There were a few cases where patients presented with clinical signs of rejection and were treated empirically with a bolus of steroids, but the histology revealed no rejection. These episodes were not considered as rejections. Similarly, if a patient had a finding of acute cellular rejection on histological reading of the biopsy but clinically was asymptomatic with endoscopy showing normal intestinal mucosa, then the patient was considered to have no rejection, received no therapy and was followed clinically with serial endoscopies.

Treatment of rejection

Mild rejection episodes were treated by a bolus of methylprednisolone, followed by a weaning cycle of steroids plus an increase in tacrolimus levels to early posttransplant values. We utilized antilymphocyte agents early in mild rejection episodes if clinical and histological results did not improve within 48 h of steroid treatment. Moderate or severe rejection episodes were mostly treated with OKT3 (2.5–10 mg i.v. daily) for 7–14 days. Removal of the graft with or without retransplantation was at times necessary. The strongest type of immunosuppression needed to treat each episode of rejection was considered for analysis.

Posttransplant lymphoproliferative disease

Patients were suspected to have posttransplant lymphoproliferative disease (PTLD) by the following criteria: presence of clinical signs or symptoms like weight loss, hypoalbuminemia, lymphadenopathy, palpable masses and diarrhea. PTLD was diagnosed when histology findings such as gastrointestinal or lymphatic tissue biopsies (hematoxylin-eosin, CD20 and CD3 immunohistochemistry, EBER in situ hybridization), positive gene rearrangements by polymerase chain reaction (PCR) for the immunoglobulin kappa light chain and the gamma chain genes of the T-cell receptor or positive PCR in blood or plasma for Epstein-Barr virus were detected.

Statistical methods

Rejection rates within a given time period were estimated as the observed number of rejection episodes divided by the total patient-months of follow-up during that period (and prior to any graft failure that may have occurred). Standard errors (SE) of these rates were estimated assuming exponential rates of failure within the time period. Means and SE of patient characteristics were reported, and tests of association among patient characteristics were performed using Pearson (uncorrected) chi-squared and t-tests. For skewed distributions, comparisons were performed on the log scale, and in these cases the geometric (rather than arithmetic) mean was reported. Cox stepwise regression (multivariable) analyses were performed to determine prognostic factors, considering all available baseline characteristics, for the following hazard rates: development of a first rejection (freedom-from-rejection), development of severe rejection and intestinal graft failure due to rejection. In order to avoid length-time bias (i.e. mistakenly assigning the patient's status at baseline), indicator time-dependent covariates (14–16) were defined in summarizing each patient's rejection history and considered in Cox stepwise regressions of the following hazard rates of failure: graft survival, graft failure due to rejection, death due to infection not triggered by graft failure and graft failure or death due to causes other than rejection or infection (17). The time-dependent expression ‘by t months posttransplant’ designates that at every time ‘t’ the hazard rate is compared between patients that have versus have not experienced the rejection episode of interest. Thus, as more episodes occur over time the designation of patients into the two subgroups ‘event occurred’ versus ‘event not occurred’ will also change (thus, the term ‘time-dependent covariate’). Cox stepwise regression was also performed for the hazard rate of developing PTLD. For each of these analyses the score chi-squared test criterion was used, and only those variables with a univariable p-value of 0.10 or less were considered in the multivariable analysis.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References


A total of 209 patients received primary intestinal grafts, 128 (61%) pediatric and 81 (39%) adults. Of these patients, 106 (51%) were male and 103 (49%) female. Age at transplant ranged from 96 days to 59 years, with a median of 6.3 years. Fifty-one patients (24% of total, 40% of pediatric) were aged less than 1 year. A total of 147 patients were Caucasian, 31 African American, 28 Hispanic and 3 Asian. The most common diagnosis in the pediatric population was gastroschisis, followed by necrotizing enterocolitis and intestinal atresia; in the adult population, mesenteric thrombosis, trauma and Gardner's syndrome were the most common indications for transplant. Of the four possible types of transplant, 62 patients received I, 32 LI, 20 MMVT and 95 MVT transplants. Median follow-up among graft survivors (n = 85) was 50.6 months (range 1–131 months).

Incidence and time of development of rejection

A total of 290 episodes of rejection were observed. Not all patients experienced rejection. The observed percentage of patients who did not experience rejection was 32.1% (67/209), while the observed percentages of patients who experienced one, two and three or more rejection episodes were 34.9% (73/209), 17.7% (37/209) and 15.3% (32/209), respectively. Overall, there was a trend for increased freedom-from-rejection in the MMVT and MVT groups in comparison with I and LI groups (p = 0.09). Among pediatric recipients, freedom-from-rejection was significantly more favorable (p = 0.0008) in MMVT and MVT recipients, whereas no differences were observed among adults (p = 0.27).

Among the 142 patients experiencing rejection, an average of 2.04 ± 0.13 episodes per patient was observed. Median time to development of the first rejection episode was 18 days posttransplant (range 3 days to 6.73 years). First rejection episodes occurring within the first year were distributed as follows: by 1 month posttransplant, 63.4% (90/142); by 3 months posttransplant, 82.4% (117/142); and by 12 months posttransplant, 90.1% (128/142). Additionally, most of the 290 episodes of rejection, whether they were first or subsequent episodes, occurred early in the posttransplant period, with 33.1% (96/290) by 1 month, 51.4% (149/290) by 3 months and 73.4% (213/290) by 12 months posttransplant (Table 1). The observed rate of first rejection (calculated using all patients) declined sharply after the first month posttransplant; similarly the rate of any rejection also declined sharply over time. The furthest rejection, a mild episode, was observed 10 years posttransplant.

Table 1.  Observed rates of first1 and any2 rejection by time after transplant
First posttransplantFirst rejection ± SE1Any rejection ± SE2
  1. Number expressed in episodes per patient-month of follow-up (up to a first rejection episode1, graft failure1,2 or death1,2).

0–1 Month0.626 ± 0.0660.493 ± 0.050
2–3 Months0.170 ± 0.0330.164 ± 0.022
4–6 Months0.028 ± 0.0120.069 ± 0.013
7–12 Months0.020 ± 0.0080.048 ± 0.008
13–24 Months0.015 ± 0.0060.029 ± 0.005
25–36 Months0.020 ± 0.0090.023 ± 0.005
>36 Months0.007 ± 0.0040.012 ± 0.002

Prognostic factors for freedom-from-rejection

For the whole cohort, five characteristics were associated with a higher incidence rate of developing any type of rejection on univariable analysis (Table 2): transplant prior to 2001 (p < 0.00001); no induction with Campath-1H (p = 0.009); not receiving spleen as part of the graft (p = 0.01); longer cold ischemia time (p = 0.02) and no induction at all (p = 0.03). However, due to the strong association of the latter four characteristics (Campath-1H, spleen, ischemia time and no induction) with date of transplant, multivariable analysis yielded only transplant date before 2001 as a significant factor. Campath-1H induction showed no association with freedom-from-rejection in pediatric recipients (p = 0.23), but was significant in adults (p = 0.003).

Table 2.  Results of the Cox stepwise (multivariable) regression to determine prognosticators for the rate of developing a first acute rejection episode (n = 209, 142 failures)
Baseline characteristic% with characteristicp-Value (univariable)p-Value to enter Cox model (multivariable)
  1. 1Median values are shown for continuous variables.

Recipient gender (male)50.7% (106/209)0.29 
Recipient race (non-Caucasian)29.7% (62/209)0.31 
Recipient age1Median = 6.3 years0.66 
Recipient age <1 year24.4% (51/209)0.35 
Recipient age >18 years (adult)38.8% (81/209)0.97 
Donor gender (male)56.9% (119/209)0.87 
Donor age1Median = 5 years0.17 
Cold ischemia time1Median = 423 min (n = 171)0.020.11
Warm ischemia time1Median = 38 min (n = 198)0.99 
Nonidentical blood type 2.9% (6/209)0.69 
No. of HLA matches1Median = 1 (n = 103)0.36 
Positive T- or B-cell cross-match 6.7% (14/209)0.070.40
In-hospital pretransplant status43.1% (90/209)0.50 
Liver included in graft60.8% (127/209)0.76 
Spleen not included in graft79.4% (166/209)0.010.46
Received I or LI45.0% (94/209)0.090.67
Received no induction15.3% (32/209)0.030.36
Received no Campath-1H70.8% (148/209)0.0090.91
Date of transplant <199819.6% (41/209)0.010.29
Date of transplant <200142.6% (89/209)<0.00001<0.00001 (selected)

No difference was seen in freedom-from-rejection between male and female recipients (Table 2); in addition, adult female patients who had previously been pregnant or had children at the time of transplant were not more prone to rejection as compared to adult female recipients who had not had children before transplant (no female patient became pregnant or had children after transplant).

A positive cross-match did not influence the development of rejection. Only 14 patients had T- or B-cell positive pretransplant cross-match (eight had both T- and B-cell positive, six only B-cell positive), and their rate of development of rejection did not differ significantly from the rest of the cohort (Table 2). Panel of reactive antibodies (PRA) was not available for study in all patients; however, as previously described (18), positive pretransplant PRA was not associated with an increase in the development of rejection.

Histopathologic grades of rejection

Among a total of 290 observed episodes of rejection, 44.8% were mild (130/290), 38.3% moderate (111/290) and 16.9% severe (49/290). First episode of rejection was mild in 61 cases, moderate in 56 cases and severe in 25 cases. The grade of the first rejection episode differed between pediatric and adult recipients, with mild rejection more frequent in pediatric and moderate rejection in adults. First episode in pediatric patients was 52.9% mild (45/85), 30.6% moderate (26/85) and 16.5% severe (14/85), while in adults it was 28.1% mild (16/57), 52.6% moderate (30/57) and 19.3% severe (11/57) (p = 0.01). Distribution of the 290 rejection episodes according to severity is shown for each transplant type in Table 3.

Table 3.  Distribution of rejection episodes according to severity (histopathological grade), by type of transplant
Type of graftTotal rejectionsMildModerateSevere
Intestine 94 39 2827
Liver-intestine 51 21 2010
Modified multivisceral 26  9 13 4
Multivisceral119 61 50 8

When we analyzed freedom from the development of severe rejection, three factors were identified by univariable analysis as significant: receiving an I or LI graft (p = 0.00002), receiving no induction (p = 0.02) and transplantation prior to 2001 (p = 0.02). However, since a significantly greater percentage of MVT/MMVT have been performed since 2001, transplant type and receiving no induction were the only two factors selected into the multivariable Cox model, with resulting p-values of 0.00003 for I/LI graft and 0.05 for receiving no induction. Severe rejection was therefore overall significantly more likely to occur in recipients of I and LI, as compared to MMVT and MVT recipients (Figure 1). Specifically, 41 of 209 patients developed a severe rejection episode, 25 pediatric and 16 adults. Among pediatric recipients, severe rejection was observed in 36.4% (20/55) of I and LI recipients combined in comparison with only 6.8% (5/73) in MMVT and MVT recipients combined (p = 0.00002, log rank test). In adults, there was a trend toward similar results, with severe rejection occurring in 28.2% (11/39) of I and LI recipients combined versus 11.9% (5/42) in MMVT and MVT recipients combined (p = 0.11).


Figure 1. Freedom from severe rejection in recipients of intestinal transplants. Hazard rate of developing severe rejection is significantly higher for recipients of I and LI (n = 94, 31 failures) as compared to recipients of MVT and MMVT (n = 115, 10 failures) (p = 0.00001, log-rank test).

Download figure to PowerPoint

The pattern for increased risk of severe rejection for recipients of I/LI was consistent throughout different eras of immunosuppression, whether patients received no induction at all, received their transplant before 2001 with induction or received their transplant after 2001 (Campath or daclizumab induction) (Table 4). In addition, for all patients who received their graft after 2001, no difference was observed between Campath and daclizumab induction for the development of any type of acute rejection (Figure 2A) or severe rejection (Figure 2B).

Table 4.  Observed percentages of patients who developed severe rejection, by type of transplant, date of transplant and whether an induction agent was given
Type of graftNo inductionTransplant date < 2001 and induction therapy1Transplant date ≥ 2001 and induction therapy2
  1. 1Of the 57 patients, 48 received daclizumab, 6 OKT3 and 3 Cytoxan as induction agent.

  2. 2Of the 120 patients, 61 received Campath-1H and 59 received daclizumab as induction agent.

Intestine50% (4/8)42% (8/19)29% (10/35)
Liver-intestine50% (5/10)22% (4/18) 0% (0/4)
Modified multivisceral 0% (0/0) 0% (0/4)19% (3/16)
Multivisceral14% (2/14) 6% (1/16) 6% (4/65)
Total34% (11/32)23% (13/57)14% (17/120)

Figure 2. Freedom from any type of rejection (A) and freedom from severe rejection (B) in recipients of intestinal transplants after 2001, according to induction treatment given (Campath-1H, n = 61 versus daclizumab, n = 59). (A) n = 120, 37 failures each in the daclizumab and Campath-1H groups, p = 0.83, log-rank test. (B) n = 120, nine failures each in the daclizumab and Campath-1H groups, p = 0.72, log-rank test.

Download figure to PowerPoint

Length of rejection episodes

Table 5 summarizes the duration of rejection episodes in days, stratified by severity. The geometric mean and median length of rejection episodes in days was strongly associated with the grade of rejection (p < 0.00001). For each increasing level of rejection severity, we observed roughly a doubling of the total time with rejection. On average, mild rejection episodes lasted about 1 week, moderate rejection episodes about 2 weeks and severe rejection episodes close to 4 weeks. It should also be noted that in 37 of the 290 rejection episodes (27 severe, 7 moderate and 3 mild) the episode length was curtailed due to graft explant, retransplant or patient death.

Table 5.  Length of rejection episodes (in days) according to severity (SE = standard error)
  1. 1Comparison of the mean length of rejection episodes on the log scale among the three grades of rejection was highly significant (p < 0.00001).

Geometric mean*/SE16.48*/1.0813.32*/1.0824.57*/1.13

Seventeen patients (nine pediatric, eight adults) experienced a rejection episode that lasted 45 days or longer (15 lasted between 45 and 65 days, 1 lasted 99 days and 1 lasted 157 days). Most of such ‘prolonged’ episodes were severe (13/17), and the remaining episodes were moderate (4/17). In addition, most of these episodes were observed in I recipients (12/17). Of these patients, 10 ultimately experienced graft loss due to rejection (of which 4 were retransplanted and 2 are still alive), 3 died of infection, 1 died of complications related to hemolytic anemia and 3 are currently alive with a functioning graft.

Treatment of rejection

Analysis of the treatment given for rejection showed that steroids alone were used overall in 63.8% of the cases (185/290). Antilymphocyte agents were required in 34.8% of the cases overall (101/290): OKT3 in 83 episodes, Campath 1H in 8, Campath 1H plus OKT3 in 3, ALG/ATG in 4, ATG plus OKT3 in 1. Cytoxan was used in two additional cases. Mild rejection episodes were treated with steroids alone in 85.4% (111/130) of cases, while moderate rejection episodes were treated with steroids alone in 56.8% of cases (63/111) (Figure 3). Severe rejection episodes were treated with antilymphocyte agents in 69.4% of cases (34/49). A greater percentage of rejection episodes were treated with antilymphocyte agents since 2001. Specifically, in the treatment of moderate rejection episodes 53.7% of patients received antilymphocyte treatment after 2001 versus 27.3% before 2001 (p = 0.006), while in severe rejection episodes we observed no change in treatment strategy over time. Removal of the graft and/or retransplantation was necessary in 16 patients, with 2 having recurrent episodes of moderate rejection and 14 having severe rejection unresponsive to treatment.


Figure 3. Treatment of rejection episodes in intestinal transplant recipients.

Download figure to PowerPoint

Analysis of the impact of rejection on graft survival and its cause-specific components

Among the 209 patients, a total of 124 graft failures/deaths were observed: 36 due to rejection, 36 due to infection and 52 due to other causes. Among the 36 patients who experienced graft failure due to rejection, 28 were due to severe rejection, 5 were due to chronic rejection, 2 were due to refractory moderate rejection and 1 was due to vascular rejection. Cox stepwise regression (multivariable) analysis of the patient's rejection history (timing, frequency, severity and length) found two independent factors associated with a significantly higher hazard rate of graft failure due to rejection (Table 6): (i) the development of a severe rejection episode by t months posttransplant (p < 0.00001; Cox model coefficient ± SE: 3.331 ± 0.442) and (ii) the development of a rejection episode that lasted for at least 21 days by t months posttransplant (p = 0.00001; Cox model coefficient ± SE: 2.307 ± 0.524). Development of a severe rejection episode had by far the strongest association; however, the length of rejection episode variable still contained independent prognostic value. These results were comparable in pediatric and adult recipients. The development of a severe rejection episode had no association with the hazard rates of death due to infection and graft failure or death due to other causes. The effect of severity of rejection and time on rejection became even more evident if we stratified patients into four categories: no severe rejection and no rejection ≥21 days; rejection ≥21 days but no severe rejection; severe rejection but no rejection lasting ≥21 days; having both a severe rejection and at least one rejection lasting ≥21 days. In these four groups the observed percentage of patients having graft failure due to rejection was 1.5% (2/136), 15.6% (5/32), 75.0% (9/12) and 70.0% (20/29), respectively. Finally, six of seven patients who lost their grafts due to chronic or refractory moderate rejection had rejection episodes lasting 21 days or longer.

Table 6.  Rejection history prognosticators for the hazard rate of graft failure due to rejection (n = 209, 36 failures)
Rejection history variables1Score test statistic (univariable)p-Value to enter the Cox model (multivariable)
  1. 1Each of the rejection history variables is a time-dependent covariate, representing the status of an intervening event (with value 0 prior to the occurrence of the event and value 1, if applicable, following the event).

  2. 2A score test statistic of 19.52 or above yields a statistically significant p-value of 0.00001 or less.

  3. 3The multivariable chi-squared test statistics for ‘severe rejection’ and ‘rejection ≥ 21 days’ were 56.8 and 19.4, respectively.

Any rejection 28.120.99
Two or more rejections 25.120.57
Three or more rejections 20.020.73
Mild rejection  0.00.16
Moderate rejection  0.10.92
Mild or moderate rejection  0.30.40
Severe rejection272.82<0.000013 (selected into model)
Rejection lasting ≥ 14 days 56.120.42
Rejection lasting ≥ 21 days 95.920.000013 (selected into model)
Rejection lasting ≥ 28 days119.420.49

Chronic rejection

Five patients, all recipients of isolated intestinal graft, went on to develop features of chronic rejection, and eventually required their grafts to be removed. Retransplant was performed in all of these patients, with two of them still alive. The median time for removal of the intestinal graft in these patients was 4.4 years from the original transplant (range 4.3–10.2 years). Analysis of the rejection history in these patients revealed a median of four episodes of acute rejection per patient (range 2–6 episodes). Although only one of five patients had sustained a severe episode of rejection, the remaining four patients all experienced at least one rejection episode lasting longer than 21 days.

Rejection and development of PTLD

We observed a total of 21 cases of documented PTLD (14 pediatric, 7 adults). Median time to development was 21 months posttransplant. Statistical analysis revealed that the occurrence of a rejection episode (as a time-dependent covariate) was associated with a significantly higher incidence rate of developing PTLD (p = 0.01). The use of OKT3 to treat a rejection episode had an even stronger association with a higher rate (p = 0.00003) of developing PTLD. In fact, the observed incidence of PTLD among those patients who received OKT3 for treating rejection was 17.6% (12/68) versus 6.4% (9/141) in those who never received OKT3. An additional analysis was performed to determine if the added risk of PTLD due to OKT3 use would diminish over time following its initial use. The Cox model coefficients ± SE for use of OKT3 indicated that the increased rate of PTLD, while highly significant during the first 12 months following use of OKT3 (2.66 ± 0.54, p < 0.00001), decreased to the point of no significance beyond the first year that followed OKT3 use (0.20 ± 0.79, p = 0.80).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Intestinal transplantation is now recognized by the gastroenterology and surgical communities as the only potentially life-saving therapy available for irreversible intestinal failure with life-threatening complications (19–21). Patient and graft survival rates for recipients of intestinal transplants have been steadily improving in recent years, and with such improvements, intestinal transplantation has become cost-effective as compared to long-term parenteral nutrition (21). Improved immunosuppression agents, aggressive graft surveillance protocols and new diagnostic and treatment modalities for PTLD are some of the factors that have contributed to the more favorable outcomes. Acute cellular rejection still poses a formidable threat to the management of such patients, especially in the early postoperative period. A reduction in the incidence and severity of rejection episodes could translate into earlier reduction of immunosuppressants and a corresponding reduction in their long-term complication rates (infections, nephrotoxicity, diabetes, PTLD).

Our data are retrospective in nature, but encompass a period of 11 years, with a cohort of over 200 patients. The large number of rejection episodes (n = 290) allowed us the capability to accurately describe their occurrence rates over time along with good power for determining prognostic factors. Even though we performed 27 first retransplants in our cohort we elected to study only those rejection episodes that occurred following primary transplantation, since the risk of graft loss following retransplantation is known to be significantly higher. We also elected to exclude from the analysis those findings of chronic rejection, mostly because a well-defined histopathological grading system has yet to be established for chronic rejection in intestinal transplantation. In addition, a thorough analysis of chronic rejection would require full thickness intestinal biopsies to analyze changes over time in the submucosa and vasculature of the graft. Our focus was therefore concentrated on analyzing episodes of acute cellular rejection.

We observed that not all patients develop acute rejection; in fact, about one-third of recipients did not experience acute rejection. Pediatric recipients of MMVT or MVT are significantly more likely to remain rejection-free as compared to I or LI pediatric recipients, while this is not observed in adult recipients. It is still uncertain as to why there is a difference in freedom-from-rejection between adult and pediatric recipients.

There were a number of patients in whom protocol biopsies were read by the pathologist as consistent with rejection. However, because of the lack of clinical symptoms and a normal appearance on endoscopic exam, no treatment was given to the patients. We excluded such group of subclinical rejection episodes from our analysis. A separate analysis of these subclinical rejection episodes is currently underway at our center. Preliminary data show that only a minority of patients with subclinical rejection go on to develop clinically significant rejection requiring treatment.

Among patients who do experience acute rejection, the first episode occurs at a median time of 2½ weeks posttransplantation. This points out to the utmost importance of performing frequent endoscopies and histopathological evaluations of the graft throughout the first few months posttransplant. The frequency of graft monitoring can then be decreased thereafter, based on the observation that most first rejections and overall number of rejection episodes decrease over time and are mostly clustered within the first year postoperatively (although a first episode of rejection can occur even as far out as seven years after transplant). The use of zoom video endoscopy (ZVE) has been described by our group as one strategy to improve monitoring of the intestinal grafts (11,12). We were not able to perform a randomized study comparing outcome of patients with and without the use of ZVE; therefore, we could not determine whether its use could diagnose rejection earlier or shorten time of treatment. In addition, ZVE can be utilized only in recipients with grafts from donors 2 years of age or older, a criterion that would therefore exclude a majority of the pediatric patients included in our analysis.

In-depth analysis of freedom-from-rejection also points out the effect of time and, possibly, a learning curve: before the year 2001, patients were significantly more likely to experience rejection than after 2001. Among the many factors, routine use of induction therapy contributed to an increased rate of freedom-from-rejection.

The average length of rejection episodes correlated with the histopathologic severity and appeared to double with each increasing grade level. While a mild rejection episode would last approximately a week, a moderate episode took about 2 weeks to resolve, and a severe episode lasted for about a month.

Although we observed that the occurrence of a mild or moderate rejection had no significant effect on graft survival, our analysis identified two critical variables that affect graft survival: progression to a severe rejection and a lengthy (≥21 days) rejection. If left alone, acute cellular rejection can quickly progress from mild to moderate and severe, often over the course of just a few days, at times even within 24 h. It is therefore essential that treatment of any confirmed rejection episode be implemented early and aggressively, coupled with frequent graft monitoring with endoscopies to assess response to therapy. As an example, mild rejection episodes can be treated with a bolus of steroids followed by weaning cycle (plus an increase in baseline immunosuppression); however, endoscopic or histological unresponsiveness to steroids within 2–3 days should dictate early use of antilymphocyte agents. There are, therefore, two factors that one needs to consider in the early stages of a rejection episode: first, try to avoid progression toward a higher grade; second, try to avoid a prolonged course.

We also observed that chronic rejection may not entirely depend on the development of severe rejection. Indeed, most of the patients who developed chronic rejection had experienced only moderate rejections, pointing to the fact that severe rejection leads more often to rapid loss of the graft, while repeated episodes of rejection of a lesser degree (mild and/or moderate) can trigger chronic damage to the bowel graft with loss of function over a long time.

There have been new agents that have been found to be useful in the treatment of rejection refractory to agents like OKT3 or thymoglobulin. We utilized alemtuzumab (6,8,9) during acute cellular rejection episodes in eight patients, with preliminary results that appear to be comparable to the use of OKT3. One caveat in the use of alemtuzumab to treat refractory rejection is that there might be a higher incidence of opportunistic infections (22). An alternative agent for the treatment of refractory rejection is tumor necrosis factor alpha antibody. The use of such agents has been successfully described in case reports (23,24). More controlled studies need be performed, however, before such agents can be routinely utilized in clinical practice.

It has been previously postulated that grafting of the liver together with the intestine might confer some degree of ‘protection’ from rejection in the recipient (25–27). Our findings suggest that there is increased freedom-from-rejection but more significantly a decreased incidence of severe rejection when a multivisceral graft is used, even in recipients of MMVT, whose graft does not include the liver. We observed these results especially in the pediatric population, and it suggests that some additional factor (or factors) might be responsible for a ‘protective’ effect.

One potential hypothesis is that susceptibility to rejection might be linked to the relative proportion of donor lymphatic tissue transferred with the graft versus native lymphatic tissue left in the recipient. In MVT recipients, removal of the native organs includes a near-total exenteration of the gastrointestinal tract (save for a segment of large intestine), including the native spleen. This leads to a much more complete lymphadenectomy than in recipients of I or LI grafts, where splenic, celiac and gastric lymph nodes are left in situ. In addition, transplantation of MMVT and MVT leads to a physiological transplant of a larger amount of lymphatic tissue (again the celiac, pancreatic and gastric lymphatic stations), as compared to I and LI recipients. One of the current hypotheses for organ engraftment and functional tolerance of solid organ grafts is the theory of clonal exhaustion-deletion, whereby bidirectional trafficking of host- and graft-derived immunocompetent cells leads to the interaction and apoptosis of those cell lineages potentially harmful to the graft (28–30). It is then possible that when a MVT graft is transplanted, the relative total mass of immunocompetent cells derived from donor and recipient may be more balanced than in I or LI recipients, where there is a preponderance of the recipient's lymphatic tissue. In addition to that, the observation that this effect is much more marked in the pediatric population might be due to the relative immaturity of the pediatric immune system, with a higher potential for ‘immunologic education’ at the level of central (i.e. thymus) and/or peripheral lymphoid tissues. Again, these are clearly only potential explanations. Hopefully, future investigations into the mechanisms of rejection of intestinal grafts will help to clarify these issues as well as further the improvement that clinical intestinal transplantation has achieved in the past 10 years.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References
  • 1
    Intestinal Transplant Registry Data. Current Results.
  • 2
    Organ Procurement and Transplantation Network. National Data Reports.
  • 3
    Kato T, Gaynor JJ, Selvaggi G et al. Intestinal transplantation in children: A summary of clinical outcomes and prognostic factors in 108 patients from a single center. J Gastrointest Surg 2005; 9: 7589.
  • 4
    Grant D, Abu-Elmagd K, Reyes J et al. 2003 report of the Intestine Transplant Registry: A new era has dawned. Ann Surg 2005; 241: 607613.
  • 5
    Kato T, Ruiz P, Thompson JF et al. Intestinal and multivisceral transplantation. World J Surg 2002; 26: 226237.
  • 6
    Tzakis AG, Kato T, Levi DM et al. One hundred multivisceral transplants at a single center. Ann Surg 2005; 242: 480493.
  • 7
    Kato T, Nishida S, Tzakis A. Small bowel and multivisceral transplantations. Transplant Rev 2005; 19: 4754.
  • 8
    Tzakis AG, Kato T, Nishida S et al. Preliminary experience with Campath 1H (C1H) in intestinal and liver transplantation. Transplantation 2003; 75: 12271231.
  • 9
    Tzakis AG, Kato T, Nishida S et al. Alemtuzumab (Campath-1H) combined with tacrolimus in intestinal and multivisceral transplantation. Transplantation 2003; 75: 15121517.
  • 10
    Kato T, Tzakis AG, Selvaggi G et al. Intestinal and mulitvisceral transplantation in children. Ann Surg 2006; 243: 756766.
  • 11
    Kato T, O'Brien CB, Nishida S et al. The first case report of the use of a zoom video endoscope for the evaluation of small bowel graft mucosa in a human after intestinal transplantation. Gastrointest Endosc 1999; 50: 257262.
  • 12
    Kato T, Gaynor JJ, Nishida S et al. Zoom endoscopic monitoring of small bowel allograft rejection. Surg Endosc 2006; 20: 773782.
  • 13
    Ruiz P, Bagni A, Brown R et al. Histological criteria for the identification of acute cellular rejection in human small bowel allografts: Results of the pathology workshop at the VIII International Small Bowel Transplant Symposium. Transplant Proc 2004; 36: 335337.
  • 14
    Mantel N, Byar DP. Evaluation of response time data involving transient states—An illustration using heart transplant data. J Am Stat Assoc 1974; 69: 8186.
  • 15
    Crowley J, Hu M. Covariance analysis of heart transplant survival data. J Am Stat Assoc 1977; 72: 2736.
  • 16
    Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York : John Wiley and Sons, Inc., 1980.
  • 17
    Gaynor JJ, Kato T, Selvaggi G et al. The importance of analyzing graft and patient survival by cause of failure: An example using pediatric small intestine transplant data. Transplantation 2006; 81: 11331140.
  • 18
    Kato T, Mizutani K, Terasaki P et al. Association of emergence of HLA antibody and acute rejection in intestinal transplant recipients: A possible evidence of acute humoral sensitization. Transplant Proc 2006; 38: 17351737.
  • 19
    Kaufmann S, Atkinson JB, Bianchi A et al. Indications for pediatric intestinal transplantation: A position paper of the American Society of Transplantation. Pediatr Transpl 2001; 5: 8087.
  • 20
    Goulet O, Lacaille F, Jan D, Ricour C. Intestinal transplantation: Indications, results and strategy. Curr Opin Clin Nutr Metab Care 2000; 3: 329338.
  • 21
    Fishbein TM, Kaufman SS, Florman SS et al. Isolated intestinal transplantation: Proof of clinical efficacy. Transplantation 2003; 76: 636640.
  • 22
    Peleg AY, Husain S, Kwak EJ et al. Opportunistic infections in 547 organ transplant recipients receiving alemtuzumab, a humanized monoclonal CD-52 antibody. Clin Infect Dis 2007; 44: 204212.
  • 23
    Pascher A, Klupp J, Langrehr JM, Neuhaus P. Anti-TNF-alpha therapy for acute rejection in intestinal transplantation. Transplant Proc 2005; 37: 16351636.
  • 24
    Pascher A, Radke C, Dignass A et al. Successful infliximab treatment of steroid and OKT3 refractory acute cellular rejection in two patients after intestinal transplantation. Transplantation 2003; 76: 615618.
  • 25
    Abu-Elmagd K, Reyes J, Bond G et al. Clinical intestinal transplantation: A decade of experience at a single center. Ann Surg 2001; 234: 404417.
  • 26
    Murase N, Demetris AJ, Matsuzaki T et al. Long survival in rats after multivisceral versus isolated small-bowel allotransplantation under FK 506. Surgery 1991; 110: 8798.
  • 27
    Calne R, Davies H. Organ graft tolerance: The liver effect. Lancet 1994; 343: 6768.
  • 28
    Starzl TE, Demetris AJ, Murase N, Ildstad S, Ricordi C, Trucco M. Cell migration, chimerism, and graft acceptance. Lancet 1992; 339: 15791582.
  • 29
    Starzl TE, Demetris AJ. Transplantation milestones: Viewed with one- and two-way paradigms of tolerance. JAMA 1995; 273: 876879.
  • 30
    Starzl TE, Zinkernagel RM. Antigen localization and migration in immunity and tolerance. N Engl J Med 1998; 339: 19051913.