The diagnosis and treatment of cardiac rejection is

ORIGINAL ARTICLES: CARDIOVASCULAR Routine Surveillance Endomyocardial Biopsy: Late Rejection After Heart Transplantation David A. Heimansohn, MD, Robert J. Robison, MD, John M. Paris III, MD, Robert G. Matheny, MD, JoAnne Bogdon, BN, and Carl J. Shaar, PhD Department of Cardiovascular/Thoracic Surgery, and Transplant Department, St. Vincent Hospital and Health Care Center, Indianapolis, Indiana Background. Transplant programs use routine surveillance endomyocardial biopsies (RSEMB), which are performed at preset intervals to diagnose cardiac rejection. This retrospective study determined the incidence of graft rejection detected by RSEMB. Methods. The records of 95 patients who underwent heart transplantation between 1987 and 1995 were reviewed. Rejection incidence was recorded for 80 patients who survived at least 30 days, with a mean follow-up of 35 months. Results. One thousand five hundred sixteen total biopsies were performed; 1,170 were RSEMB. Four hundred seventy-five total rejection episodes occurred and 269 (56%) were diagnosed by RSEMB. Two distinct patient groups were identified. The majority (70 patients), had a decline in the incidence of rejection and no rejection episodes were identified by RSEMB after 36 months. In contrast, the high rejection group (10 patients) had a significantly higher ongoing rejection rate (p < 0.04 to p < 0.001) throughout their postoperative course up to 72 months. Conclusions. The majority of our transplant patients demonstrate a decrease in rejection with time and do not require RSEMB beyond 30 months. We identified a group of patients who exhibited a higher rate of rejection and need continued RSEMB. (Ann Thorac Surg 1997;64:1231 6) 1997 by The Society of Thoracic Surgeons The diagnosis and treatment of cardiac rejection is important for long-term survival after cardiac transplantation. Percutaneous transvenous endomyocardial biopsy (EMB), first introduced to heart transplantation at Stanford University School of Medicine in 1972, remains the standard for diagnosing cardiac rejection [1]. The protocol involves the use of routine surveillance endomyocardial biopsies (RSEMB) early after transplantation and at increasing intervals indefinitely. Controversy exists as to the necessity of RSEMB at more than the first year after transplantation. In addition to significant costs, EMBs have the potential morbidity of For editorial comment, see page 1230. tricuspid regurgitation, cardiac perforation, and entrance site complications [2,3]. Although some clinicians support indefinite use of RSEMB after transplantation [4], others believe RSEMB exhibits low yield for diagnosing rejection after 1 year and suggest that continued RSEMB does not improve patient survival [5,6]. The retrospective observations reported here led us to question the value of indefinite RSEMB and their ability to diagnose cardiac Presented at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb. 3 5, 1997. Address reprint requests to Dr Heimansohn, Department of Cardiovascular/Thoracic Surgery, St. Vincent Hospital and Health Care Center, 8333 Naab Road, Suite 300, Indianapolis, IN 46260. late rejection episodes at more than 3 years in a heart transplant population. Material and Methods Patient Population This retrospective study encompassed our cardiac transplant population from March 1987 to June 1995. Follow-up was completed on 80 of 95 patients who survived at least 30 days after transplantation. Follow-up of the 80 patients was complete through June 30, 1995, and represented a mean follow-up of 35 26 months (range, 3 to 99 months). A total of 1,516 EMBs were performed in 80 patients postoperatively during the study period and included 1,170 RSEMB. Definitions Rejection was defined as a clinical event that required specific augmentation of immunosuppressive therapy as determined by the clinician and was associated with a biopsy sample demonstrating at least mild to moderate rejection. Biopsy-documented rejection was defined by standard Billingham criteria or at least grade 3a or above by the International Society of Heart and Lung Transplantation classification [7,8]. The beginning of a rejection episode was identified with a positive biopsy sample and the simultaneous initiation of augmented immunosuppression and was concluded with the first subsequent negative biopsy sample. A biopsy performed per protocol without associated symptoms or rejection was defined as 1997 by The Society of Thoracic Surgeons 0003-4975/97/$17.00 Published by Elsevier Science Inc PII S0003-4975(97)00732-7

1232 HEIMANSOHN ET AL Ann Thorac Surg ENDOMYOCARDIAL BIOPSY 1997;64:1231 6 Table 1. Endomyocardial Biopsy Protocol After Heart Transplantation Interval First month: once a week Second and third month: every other week Fourth month to sixth months: every other month Sixth month to 18th months: every 3 months Eighteenth month to 3 years: every 6 months Three years and beyond: yearly a Yearly Total 1st year 12 biopsies 2nd year 3 biopsies 3rd year 2 biopsies a On the basis of this study, we no longer perform routine surveillance endomyocardial biopsy at more than 30 months postoperatively. RSEMB. A biopsy performed because of symptoms, or as follow-up to a positive biopsy sample, was defined as an indicated biopsy. Routine surveillance endomyocardial biopsies were performed according to the protocol described in Table 1. Patients who continued to have rejections up to 72 months after transplantation were classified as high rejecters whereas those who had no rejections diagnosed at more than 36 months were identified as low rejecters. Patient Surveillance Care In addition to endomyocardial biopsies, all patients attended transplant clinics after the operation at 3-month intervals for their entire postoperative experience. Routine clinical diagnostic testing included: a chest roentgenogram, a blood chemistry profile, measurement of circulating cyclosporine and magnesium concentrations, and a complete blood count with a differential white cell count. At annual intervals (at the transplant birth date ) each patient, regardless of how long after transplantation, received a follow-up coronary angiogram, a firstpass ejection fraction evaluation, and an echocardiogram. Patients with recognized coronary artery disease received regularly scheduled cardiolite stress tests. Patients who presented with symptoms suggestive of rejection underwent indicated biopsies. Those symptoms included one or more of the following: palpitations or flutter, congestive heart failure, lethargy, exertional fatigue, or loss of appetite. Data Collection The hospital records of the study patients were reviewed for the following data: date of transplantation, age at transplantation, gender, race, immunosuppression regimen, routine surveillance biopsy dates and results, indicated biopsy dates and results, and medication changes instituted as a result of a rejection-positive EMB. Inclusive dates of review included time from heart transplant to death or to June 30, 1995. Immunosuppression Therapy Most patients received a 7-day induction course of antithymocyte globulin and a triple-drug regimen of cyclosporine, azathioprine, and prednisone. Augmentation of immunosuppression or a 3-day pulse of Solu-Medrol (Upjohn, Kalamazoo, MI), or both, was used to treat most episodes of graft rejection. OKT3 was used to treat severe or refractory rejection. In addition, methotrexate was used in 5 patients for continued graft rejection refractory to standard treatment. Data Analysis All data analysis was done using JMP Statistical Discovery Software (SAS Institute Inc, Cary, NC). Group biopsy rejection rates during each time interval were compared using the 2 values with and without continuity correction, Fisher s exact probability, Phi coefficient, and Mc- Nemers 2. Mean number of rejections per time interval were compared using a t-test for uneven groups. A p value of 0.05 or less indicated a significant difference between parameters. Results Eighty patients included in this study underwent heart transplantation between March 1987 and June 1995. Actuarial survival rates at 1, 5, and 8 years for our heart transplant patient population were 83%, 76%, and 71%, respectively. This retrospective study encompassed our cardiac transplant population from March 1987 to June 1995. Ninety-five patients underwent heart transplantation for the following conditions: dilated cardiomyopathy, 40% (38/95); ischemic cardiomyopathy, 55.8% (53/95); and other causes, 4.2% (4/95). Fifteen (16%) patients died within 30 days of transplantation and were not included in the data analyses. Late deaths were the result of rejection in 4 (3 vascular, 1 cellular) patients and other causes in 5 (pneumonia, cerebrovascular accident, and myocardial infarction caused by coronary artery disease). A total of 1,516 EMBs were performed during the study period. Major complications were defined as those that resulted in admission to the hospital, and minor ones were defined as those that did not prolong observation time in the catheterization laboratory. Two (0.1%) patients experienced deep venous thrombosis and 2 (0.1%) experienced pseudoaneurysms that were considered to be major and minor complications, respectively. As seen in Figure 1, 1,170 (77%) of the 1,516 total biopsies performed were RSEMB. Two hundred sixtynine graft rejection episodes were identified by these RSEMB, which included more than 56% of all rejection episodes. The trend shows a high rate of rejection diagnosed by RSEMB during the first 3 months after transplantation (38%), with an early steep, then gradual, decline in rejection with time out to 72 months. A similar curve is seen with rejection episodes when all biopsy samples are reviewed, and the curve also demonstrated a decreasing rate of rejection over time. This decreasing rate of rejection has been demonstrated in the literature and explains why all RSEMB protocols are performed less frequently with time after cardiac transplantation [10]. Figure 2 summarizes graft rejection rates observed

Ann Thorac Surg HEIMANSOHN ET AL 1997;64:1231 6 ENDOMYOCARDIAL BIOPSY 1233 Fig 1. Total number of routine surveillance endomyocardial biopsies performed and the total number of graft rejections detected during routine surveillance endomyocardial biopsies in 80 transplant patients. during all (routine and indicated) EMBs. A careful review of the data identified a significant trend in a subgroup of 10 patients who had significantly (p 0.05 to p 0.001) higher rejection rates than the remaining 70 patients at most time intervals throughout the study period. Whereas the 70 low-rejection patients had no rejection at more than 42 months, the 10 high-rejection patients continued to have ongoing rejections up to 72 months. The identification of these two separate groups of patients led us to examine the records of all patients to see if this trend continued when only those rejection episodes identified by RSEMB were included. Figure 3 summarizes graft rejection rates observed during RSEMB only. The 10 high-rejection patients demonstrated a similar trend with higher rates of rejection at most intervals compared to the 70 low-rejection patients. Both groups had a similar rate of rejection at 3 and 12 months, but the high-rejection group had a much higher rejection rate at most other times. The persistence of rejection in this high-rejection group continued up to 72 months whereas the low-rejection group had no rejections diagnosed by RSEMB at more than 36 months. Figure 4 summarizes the mean frequency of diagnosed graft rejections per patient during RSEMB performed during the entire observation period after transplantation and compares the two groups of patients. A similar declining rejection rate with a plateau after 12 months, as seen in Figures 2 and 3, was apparent when the frequency of graft rejections per patient was analyzed. Rejections per patient within the 70-patient group plateaued at 12 months after transplantation and dropped to zero at 36 months. Comparison of the two patient groups revealed that the mean graft rejection frequency per patient was significantly (p 0.05) higher in the highrejection group compared with the low-rejection group at most time intervals up to 72 months. This higher rate of rejection continued at most intervals throughout the follow-up period, and continued beyond the point at Fig 2. Graft rejection rates of the high-rejection and low-rejection patient groups during follow-up indicated and routine surveillance endomyocardial biopsies.

1234 HEIMANSOHN ET AL Ann Thorac Surg ENDOMYOCARDIAL BIOPSY 1997;64:1231 6 Fig 3. Graft rejection rates of the high-rejection and low-rejection patient groups during follow-up routine surveillance endomyocardial biopsies. which the low-risk group was showing no further rejection episodes. This discovery led us to identify a predictive characteristic of high-rejection patients as compared with the low-rejection group. The high-rejection group had 11.1 cumulative rejections per patient in the first 24 months compared with 4.7 rejections per patient in the low-rejection group. The ranges of rejection rates per interval were 0.2 to 4.1 and 0.1 to 3.1 for the high- and low-rejection groups, respectively. The trend toward reduced rejection rates, although lower in the low-rejection group, was not significantly different than that observed in the high-rejection group. Comment Cardiac transplantation became widespread in the early 1980s primarily because of greater success in the prevention of cardiac rejection as the result of cyclosporine treatment. The use of EMB early in the transplant experience improved the treatment of cardiac rejection and increased first-year survival significantly. The Stanford program introduced EMB clinically in 1972 to diagnose rejection, and their first-year survival increased from 40% to 65% [6]. Techniques and instrumentation used to perform endomyocardial biopsy have improved significantly compared with those used during the early experience and now EMB can be done with low morbidity [2]. The majority of biopsies performed in the transplant population are done indefinitely, but at lengthening intervals with time, as part of a preset protocol. As the heart transplant experience grew, programs noticed that the rate of rejection decreased with time and began questioning the necessity of continuing RSEMB indefinitely. Spratt and colleagues [6] observed a decline in graft rejection identified by RSEMB, after 18 months saw no rejection in asymptomatic patients, and concluded that RSEMB was unnecessary at more than 9 months. In addition, Sethi and coworkers [5] found a similar decline Fig 4. Frequency of graft rejection per patient in the high-rejection and low-rejection patient groups during follow-up routine surveillance endomyocardial biopsies.

Ann Thorac Surg HEIMANSOHN ET AL 1997;64:1231 6 ENDOMYOCARDIAL BIOPSY 1235 Table 2. Patient Demographics Characteristic Total Population (n 80) Low-Rejection Group (n 70) High-Rejection Group (n 10) p Value Age (y) 52.3 9.3 53.1 8.2 44.6 12.7 NS b (19 65) a (31 65) (19 59) Sex (female) 20 (25%) 15 (27.3%) 5 (50%) NS c Rejections per patient d 4.7 11.7 NS b a Mean standard deviation (range). b Probability determined by a t test for uneven groups. c Probability determined by 2 analysis. d Accumulative rejection episodes/patient 0 to 24 months postoperatively. NS not significant. in the detection of graft rejection in asymptomatic patients using RSEMB and concluded that routine biopsy was not helpful at more than 12 months. In contrast, Van Trigt and associates [4] reviewed their results in 120 cardiac transplant patients and found a rejection incidence of 15% to 20% at more than 3 years after transplantation and recommended continued use of RSEMB in the later interval after cardiac transplantation. In the pediatric population Zales and colleagues [9] noted that the noninvasive methods used to diagnose rejection were less sensitive and specific than needed and occasionally overestimated rejection. They recommended continued use of RSEMB in the pediatric population, not only to diagnose episodes of rejection, but to identify when treatment for rejection was unwarranted. They also demonstrated a low morbidity associated with performing endomyocardial biopsy and reported one complication in more than 330 biopsies [9]. In view of the findings of others, as well as the consideration of morbidity and cost, we evaluated our own transplant population experience with RSEMB. We reviewed all biopsy results, specifically to determine the importance of RSEMB in our population. Our study led to the recognition of a subpopulation of transplant patients who continued to experience graft rejection at more than 36 months after transplantation and who accounted for all the graft rejections identified by RSEMB after that time in our entire population. Our data confirm previously published reports that the rate of rejection decreases with time except for a slight increase during the 18- to 30-month intervals postoperatively. Only a few patients in our series continued to experience rejection at more than 36 months. On close examination, these patients with late rejection helped us identify a subgroup of patients the high-rejection group who experienced the majority of rejection episodes at more than 1 year after transplantation. It is important to be able to predict those patients at higher risk for rejection, so that they can be monitored more closely and treatment can be instituted early, leading to better long-term results. The high-rejection group consisted of 10 patients who had an accumulative number of RSEMB-identified rejection episodes of 11.7 per patient during the first 24 months after transplantation; this compared with 4.7 cumulative episodes of rejection per patient during the same period in the remaining 70 patients. In addition, further analysis of the high-rejection group revealed that they continued to demonstrate a much higher rate of rejection throughout the study period (Fig 4). At more than 36 months the low-rejection group had no rejection diagnosed by RSEMB, whereas the high-rejection group continued with 1.4 rejection episodes per patient at 42 months, with a decreasing but ongoing rate thereafter. Kirklin and associates [10] identified several risk factors predictive of subsequent rejection. One of the most significant factors was an increased number of previous rejection episodes. The conclusion that more episodes of early rejection will predict a higher rate of late rejection was supported by our high-rejection group. Kirklin and associates identified other factors, such as younger age and female sex, that were predictive of patients at risk for rejection. Table 2 shows our groups to have similar trends regarding identifiers of rejection, although they are not statistically significant because of the low number of patients in both groups. In addition, our high-rejection group had ongoing rejection that occurred frequently without significant symptoms. Rejection would not have been identified in this group if RSEMB had not been performed, because the patients were asymptomatic and an indicated biopsy would not have been warranted. These observations led us to conclude that high-rejection patients needed ongoing RSEMBs for diagnosis and appropriate treatment of rejection. The 70 patients in the low-rejection group, who were the majority of patients in our transplant population, have a predictable decline in the rate of rejection. In this group, no rejection occurred after 36 months. The few episodes of late rejection that did occur were identified using indicated biopsies because of clinical symptoms. Regardless of time after transplantation, patients underwent routine quarterly, or as needed, clinical evaluations. During each clinic visit a transplant cardiologist reviewed each patient s journal, performed a physical examination that included a thorough evaluation of heart sounds, complete blood count, and blood chemistry laboratory tests, and ordered any diagnostic tests clinically necessary if warranted. All transplant patients underwent an annual (transplantation anniversary) coronary angiogram, and the clinic visit just before that angiography included an echocardiogram and a first-pass ejection fraction evaluation. We feel that these precautions help

1236 HEIMANSOHN ET AL Ann Thorac Surg ENDOMYOCARDIAL BIOPSY 1997;64:1231 6 to preclude the surprise acute rejection that could be missed in the absence of RSEMB. In conclusion, RSEMB are indicated up to 30 months postoperatively in a majority of heart transplant patients. Thereafter, most episodes of rejection are associated with symptoms and can be diagnosed with an indicated biopsy. A subgroup of patients can be identified who need ongoing RSEMB as a result of high rates of ongoing rejection. References 1. Caves PK, Stinson EB, Billingham M, Shumway NE. Percutaneous transvenous endomyocardial biopsy in human heart recipients. Ann Thorac Surg 1973;16:325 6. 2. Bhat G, Burwid S, Walsh R. Morbidity of endomyocardial biopsy in cardiac transplant recipients. Am Heart J 1993;125: 1180 1. 3. Hausen B, Albes JM, Rohde R, Demertzis S, Mugge A, Schafers HJ. Tricuspid valve regurgitation attributable to endomyocardial biopsies and rejection in heart transplantation. Ann Thorac Surg 1995;59:1134 40. 4. Van Trigt P, Davis RD, Shaeffer GS, et al. Survival benefits of heart and lung transplantation. Ann Surg 1996;223:576 84. 5. Sethi GK, Rosado LJ, McCarthy M, Butman SS, Copeland JG. Futility of yearly heart biopsies in patients undergoing heart transplantation. J Thorac Cardiovasc Surg 1992;104:90 3. 6. Spratt P, Sivathasan C, Macdonald P, Keogh A, Chang V. Role of routine endomyocardial biopsy to monitor late rejection after heart transplantation. J Heart Lung Transplant 1991;10:912 4. 7. Billingham ME. Dilemma of variety of histopathologic grading systems for acute cardiac allograft rejection by endomyocardial biopsy. J Heart Transplant 1990;9:272 6. 8. Billingham ME, Cary NRB, Path MRC, Hammond ME, Kemnitz J. A working formulation for the standarization of nomenclature in the diagnosis of heart and lung rejection: heart rejection study group. J Heart Transplant 1990;9: 587 92. 9. Zales VR, Crawford DD, Backer CL. Role of endomyocardial biopsy in rejection surveillance after heart transplantation in neonates and children. J Am Coll Cardiol 1994;23:766 71. 10. Kirklin JK, Naftel DC, Bourge RC, et al. Rejection after cardiac transplantation. A time-related risk factor analysis. Circulation 1992;86(Suppl 2):236 41. DISCUSSION DR JACK J. CURTIS (Columbia, MO): I greatly enjoyed this paper and was excited to see it on the program. At the University of Missouri we have followed a similar but slightly different trend. We stop performing biopsies in patients who have not had any rejection by 2 years, and that is about 50% of our patients. I wonder if you have analyzed your groups that way, the nonrejectors, to see what your outcome was. And also, there are those people who have pesky rejections during the first year or 2 but then go 2 years without having a subsequent rejection. We also stop doing biopsies in those people. I am wondering if we might be missing something based on your data as you have presented it. I enjoyed your paper. DR HEIMANSOHN: Thank you, Dr Curtis, for your kind comments. Within our study population of 80 patients, 10 patients never exhibited a myocardial rejection episode. These patients would fall into the low-rejection group, and our cutoff of 2.5 years would go along with your finding. We would certainly consider the patients who have had a lot of episodes of rejection in the first year to be in the higher risk group, although we would then evaluate them at 2.5 years. DR JAMES K. KIRKLIN (Birmingham, AL): Could you be more specific about your current protocol for discontinuation of biopsies? Second, all of us have had the experience of occasional tragic patients who go years without a rejection episode and then come in with profound low cardiac output to die of acute rejection. Do you believe there is any role for substituting, let us say, at 2 years in that low-rejection group some surveillance echocardiographic assessment periodically so that you might get a clue of diminished ventricular function perhaps weeks before the fatal event? DR HEIMANSOHN: Although patients do not receive routine surveillance endomyocardial biopsies after 30 months, all patients do present to a transplant patient clinic every 3 months postoperatively and are seen by a transplant cardiologist at those clinics. Diagnostics for each clinic include a chest roentgenogram, blood chemistry profile, and complete blood count with differential. In addition, each patient undergoes an annual angiogram, echocardiogram, and ejection fraction determination.