|Home | About | Journals | Submit | Contact Us | Français|
It is unclear whether benefits outweigh harms for routine screening and prophylactic revascularization to prevent coronary artery disease (CAD) in asymptomatic kidney transplant candidates.
Pilot feasibility study with prospective observational data collection and patient interviews.
Consecutive patients referred for kidney and/or pancreas transplant at 26 major transplant centers in the US.
Older age, diabetes, prior cardiovascular disease and multiple traditional CAD risk factors.
Eligibility and willingness to participate in a randomized controlled trial (RCT) to study the effect of CAD screening on major adverse cardiac events.
Patients who would be candidates for a hypothetical RCT of CAD screening were interviewed and asked if they would participate in such a trial. Sample size for the trial was estimated using data on Medicare patients in the United States Renal Data System with major adverse cardiac events as the primary endpoint.
Among consecutive eligible patients, CAD evaluation was not indicated in 398 (24%), already completed before referral in 602 (36%), and pending (and hence eligible for a RCT) in 665 (40%). Of 241 interviewed, 73% indicated they would be willing to participate in a RCT. We estimated that approximately 4000 would need to be enrolled to detect a 20% reduction in major adverse cardiac events at greater than 80% power at P<0.05.
Willingness to participate in an actual clinical trial may be different than indicated in an interview.
A RCT to compare the effects of routine screening for CAD versus no screening on major adverse cardiac events is feasible.
Cardiovascular disease is the leading cause of death after kidney transplant, and the incidence of myocardial infarction is highest in the period immediately after transplant.(1) Therefore, a compelling argument can be made to screen transplant candidates for occlusive coronary artery disease (CAD) in order to identify lesions for preemptive coronary revascularization and thereby reduce the risk of kidney transplant. Indeed, the standard practice in the US has been to screen high-risk asymptomatic patients for CAD with non-invasive testing and/or coronary angiography.(2;3) Patients who are found to have critical lesions (generally >70% occlusion) undergo preemptive revascularization. Opinion-based, consensus guidelines have generally recommended this approach as part of the routine pre-transplant evaluation.(4-6) However, recent evidence from the general population suggests that routine screening for CAD in asymptomatic patients prior to major non-cardiac surgery may not improve outcomes that are important to patients.(7-9) The American College of Cardiology (ACC) / American Heart Association (AHA) does not recommend routinely screening asymptomatic patients facing intermediate-to-high risk surgery, if their functional status allows them to perform 4 or more metabolic equivalents.(10) Thus, the ACC/AHA guidelines for the general population are in direct conflict with those for CKD patients facing kidney transplant.
There are a number of potential benefits and harms associated with CAD screening in asymptomatic patients referred for transplantation (Box 1). A RCT would be appropriate to determine whether routine CAD screening effectively reduces adverse outcomes important to patients, e.g. mortality, acute myocardial infarction, and acute coronary syndrome, and that the benefits of screening would outweigh the harms. This study was carried out to determine whether such a trial would be feasible.
Most transplant centers in the US screen asymptomatic patients who are considered high risk for CAD, as described in the American Society of Transplantation guidelines.(4) Screening is usually carried out for patients who have one or more of the following 4 characteristics:
Subsequent screening is usually repeated after a prescribed (but center-specific) interval.
We assumed that a RCT would include a “standard practice” control group that would call for screening asymptomatic high-risk patients with non-invasive stress testing, followed by coronary angiography in patients with an abnormal stress test, and prophylactic revascularization of critical coronary lesions found on angiography. The intervention group in the proposed RCT would call for patients to receive care as described in the ACC/AHA guidelines for perioperative management of patients undergoing non-cardiac surgery.(10) These guidelines indicate that asymptomatic patients undergoing kidney transplant would generally not be screened as described above, unless their functional status was restricted to the point that they could not perform 4 metabolic equivalents.
In the proposed RCT, all patients with signs and symptoms suggesting unstable coronary disease, e.g. an acute coronary syndrome or unstable angina, would be excluded. These symptomatic patients would receive all necessary testing and intervention that is required to relieve their symptoms. In the proposed RCT, all patients would receive maximal medical care per the transplant center's usual practice. The primary endpoint would be major adverse cardiac events (MACE), and would include: myocardial infarction, acute coronary revascularization (to relieve symptoms), and cardiac death. The analysis would be by intention-to-treat and time-to-event.
This pilot study was designed to determine whether it would be feasible to conduct a RCT of routine screening versus not screening for CAD prior to kidney transplant. The questions that we set out to answer in this pilot study included:
We assumed that the proposed RCT would randomly allocate patients referred for kidney or simultaneous kidney and pancreas transplant to follow either the current standard of practice for CAD screening at the center, or to follow the 2007 ACC/AHA guidelines for perioperative management of non-cardiac surgery.(10).
We contacted kidney transplant centers in the US that were in the top 50th percentile based on the numbers of adult living donor transplants in the previous 2 years (2006-07). The numbers of living donor transplants were determined from data reported to the Organ Procurement and Transplantation Network (OPTN) (12). We used living donor transplants as the criteria for selecting centers, because we initially anticipated that the trial would enroll only patients who were likely to receive a living donor transplant. The goal was to have 30 centers participate.
The pilot study was approved at the Institutional Review Board (IRB) at each study site. Study coordinators at each site were asked to fill out case report forms that included information on each patient referred for transplant during the pilot study, i.e. from the time the site first began looking for patients to interview until the time the 10th and final interview at their site was completed. We set out to recruit 30 sites, a number which was based on the anticipated number of sites that would be needed to recruit patients for the proposed RCT. There were 35 centers contacted before we were able to meet the goal of 30 participating centers. Although 5 centers declined to participate in this pilot study, 2 of the declining centers indicated they would likely participate in the proposed RCT if asked.
The coordinator at each site was asked assess all consecutive patients referred for transplant evaluation and to interview those who qualified for the hypothetical study (and agreed to be interviewed). Thus, interviewees were chosen in the same manner that study subjects would be chosen and asked to participate if the study were actually being conducted. We asked the coordinator at each site to continue screening for potential participants and conducting interviews until 10 interviews had been completed. Thus, the sample of patients at each site included all consecutive patients who were referred for evaluation from the start of the pilot study to the time that 10 interviews were completed. At some sites, many patients were screened to get 10 interviews, while in others very few were screened to complete the 10 interviews. For those who were screened, but not interviewed, the reasons they were not interviewed (unable to give consent, etc) were tabulated.
The following information was collected for each patient referred for transplant evaluation:
In addition, each participant who was interviewed was asked whether or not he/she could perform activities at a level of 4 metabolic equivalents, e.g. light housework, or climbing a flight of stairs. Interviewees were also asked whether they would participate in a RCT (after it was described) if the study were being held today: 1) definitely yes, 2) probably yes, 3) probably not, 4) definitely not, or 5) other, and indicate the reason.
The characteristics of patients and the results of the survey questionnaire were tested for differences between study sites using a chi-square test. As part of this pilot feasibility study, we also performed a sample size calculation that was not based on the results of the pilot study per se, but made assumptions on enrollment and event rates for a primary MACE endpoint using a superiority design. Rates were estimated from the United States Renal Data System incident End-Stage Renal Disease population 2000-06. We included high risk patients with a live donor kidney transplant or registered on the deceased donor kidney transplant waiting list and at least one of: 1) age 50 or older, 2) diabetes, or 3) prior cardiovascular disease. The rate of deceased donor transplant among patients wait-listed without a living donor (N=77,834) was 25.29 per 100 patient-years. The rate of living donor transplants among patients with a living donor (N=34,802) was 133.77 per 100 patient-years. The percent with a living donor would be 31%. We also assumed that 30 transplant centers would participate.
Thirty centers agreed to participate, however 4 were not able to conduct interviews or to submit data, and these 4 centers were excluded from further analysis. Six centers completed fewer than 10 interviews (N=4, 5, 6, 7, 9, and 9 interviews, respectively), and 20 centers completed all 10 interviews. Thus, the total number of interviews completed was 241. Of the 241 interviews conducted, 10 were in patients who had a low functional capacity (i.e., could perform less than 4 metabolic equivalents) and may therefore not have been candidates to be randomized to cardiac screening, but were nevertheless interviewed.
During the period of study, there were 1752 patients referred to the 26 participating centers for transplant evaluation. Of these, 72 could not give informed consent due to: language barriers (N=60; 83%), cognitive limitations (N=13, 14%), or other reasons not specified (N=2, 3%). Of the remaining 1680 participants, 10 had severely limited functional capacity (could perform less than 4 metabolic equivalents) thereby making cardiac stress testing something to be considered under the ACC/AHA guidelines.(10) This left 1670 who would have been able to participate in a RCT of cardiac evaluations.
There were 5 patients in whom it was not known whether cardiac evaluation was already completed, or if not completed whether it was indicated, and these 5 were excluded from further analysis. Of the remaining 1665 patients (Table 1), 602 (36%) had already had a cardiac evaluation performed before they were referred to the transplant center. This practice of prereferral cardiac evaluation varied substantially from center to center (Fig. 1). In 3 of the 26 participating centers, over half of the patients had already undergone a cardiac evaluation before referral. However, in 12 centers this occurred in 25% or less of the referred transplant candidates.
Of the 1665 transplant candidates that were referred to the 26 centers during the period of this study, 1267 (76%) would have been asked to undergo a cardiac evaluation (Fig. 2). However, this varied between 100% at 5 centers, to less than 75% at 5 centers, with the other 16 centers between 75-100% (Fig. 2).
Reasons that were given for requiring cardiac evaluations included diabetes, prior cardiovascular disease, older age (variably defined), multiple cardiovascular disease risk factors, and other or unknown (Table 2). The most common reason for “other” was a single cardiovascular risk factor, e.g. hypertension, obesity, or cigarette smoking.
In this pilot feasibility study, we asked centers to estimate the number of transplant candidates referred for evaluation who would likely have a suitable living donor. Of the 1665 candidates referred to the participating 26 centers, 1110 (67%) were thought to have a potential living donor-(Fig. 3). Six centers thought more than 95% of their candidate referrals had a potential living donor, while 3 centers thought less than 50% had potential living donors (Fig. 3).
There were 241 interviews conducted at 26 centers. Of these, 10 were obtained from patients who may not have been candidates for a RCT, because the could not perform 4 or more metabolic equivalents, and may have required testing under the ACC/AHA guidelines.(10) Of the remaining 231 who were interviewed, 18 had already had a cardiac evaluation performed and 1 was judged not to need a cardiac evaluation; In theory, these 19 patients may not have been eligible for a RCT (and should not have been interviewed). Although we tabulated the results for all 241 interviews, the results were similar for the subsets of interviews that excluded the patients that may not have been candidates for a RCT (Table 2). Among the 241 interviewed, 73.0% said they would participate. However, there was substantial center-to-center variability in responses (Fig. 4). In a regression analysis, there were 5 centers where the response rate was significantly different from the others (Fig. 4).
Rates of MACE were determined from USRDS data (Table 3). We estimated that the number of transplant candidates that would need to be enrolled to detect a 20% reduction in MACE at greater than 80% power would be approximately 4000 (Table 4).
The principle findings in this study were: 1) most transplant centers were willing to participate in the propose RCT, 2) very few patients would not be able to give informed consent or would not have adequate functional status (4 metabolic equivalents), 3) a number of centers reported that the cardiac evaluations were already completed in a high proportion of patients even before they were referred for transplant evaluation, 4) 76% of patients referred would have needed a cardiac evaluation by criteria currently employed by the evaluating kidney transplant centers, if such an evaluation had not already been obtained, 5) many patients would be excluded if the study required transplant to occur soon after evaluation (restricting the RCT to patients with a living donor), and 6) about 75% of the patients interviewed indicated they would be willing to participate in the proposed RCT.
Virtually all of the centers that participated in this pilot feasibility study indicated they would also participate in the proposed RCT. Of 35 centers that were initially asked to participate in this pilot study, only 5/35 (14%) declined. Another 4 indicated that they wanted to participate, but ultimately they were not able to conduct the data collection and interviews. This high participation rate indicates a high level of interest and likelihood that there would be enough centers to conduct the proposed RCT.
Of the 1752 patients referred for transplant evaluation during the study period, only 72 (4.1%) could not give informed consent (Fig. 5). Most of these were due to language barriers that could be overcome if the RCT accommodated Spanish-speaking candidates. Similarly, only 10 (0.6%) had functional capacity reduced to 4 METS or less (Fig. 5), thus making them candidates for cardiac evaluation according to the ACC/AHA guidelines,(10) and ineligible to participate in a RCT.
A major barrier to conducting the proposed RCT was encountered in centers that indicated that cardiac evaluations were often performed in patients before they were referred to the transplant center. However, this practice varied greatly from center-to-center, and it would be possible to exclude centers from a RCT where most of the patients referred already had cardiac evaluations. There were only 3 of the 26 participating centers where more than half had already had a cardiac evaluation before referral (Fig. 1). It is also possible that some centers could persuade their referring physicians to defer evaluations until patients were asked if they wished to participate in the RCT.
It was remarkable that 76% of patients required a cardiac evaluation as part of the transplant work up (Fig. 2). This was a consistent finding across all participating transplant centers. The most common reasons given for needing a cardiac evaluation were older age, diabetes, and multiple cardiac risk factors (Table 2). Although study coordinators were allowed to record more than one reason for the cardiac evaluation, none of them did so, and the results therefore likely represent the “primary reason” for the evaluation.
A RCT that used MACE after transplant as the primary endpoint would likely need to limit enrollment to patients that would be transplanted soon after evaluation (and randomization). This requirement would mean that most participants would need to have a living donor. Unfortunately, in this pilot study only about two-thirds had a potential living donor at the time of evaluation (Fig. 3). Therefore, restricting enrollment to patients having a potential living donor would substantially restrict enrollment in a RCT (Fig. 5). In addition, the relevance of the study results could be questioned if enrollment were restricted to mostly candidates with living donors. Similarly, with a primary MACE endpoint restricted to events occurring after transplant, patients would be censored if MACE occurred while they were waiting for transplant. This would reduce the power of the study and increase the number of patients that would need to be recruited. Therefore, it is likely that a RCT would not require that patients be transplanted soon after enrollment, and would use as the primary endpoint MACE occurring after randomization, either before or after transplant.
It was encouraging that about 75% of potential RCT participants that were interviewed indicated “if the study were held today” they would be willing to participate (Table 3). Interestingly, there was remarkable variability in the responses between centers (Fig. 4). This variability could be due to the way in which the study was presented to patients. It is possible that if coordinators were adequately trained and coached in presenting the study to eligible participants, that more would participate.
The sample size calculation indicated that approximately 4000 participants would be needed to achieve enough statistical power (>80%) to detect a 20% difference in MACE between 2 study groups. Of course, this calculation is based on a number of assumptions. However, this number is same as the sample size calculated for the Folic Acid for Vascular Outcome Reduction In Transplantation (FAVORIT) trial. The FAVORIT trial randomly allocated stable kidney transplant recipients with elevated homocysteine levels to placebo or vitamin therapy to lower homocysteine in an effort to reduce MACE.(13) Similarly, the Assessment of LEscol in Renal Transplantation (ALERT) study randomly allocated 2012 stable kidney transplant recipients to placebo versus treatment with fluvastatin to reduce MACE.(14) Certainly, both the FAVORIT and the ALERT trials differ in many ways from the RCT of cardiac evaluations proposed here. However, it is encouraging that the sample size for our proposed RCT is similar to those of 2 trials that enrolled transplant patients and used a similar MACE endpoint.
There are many reasons that the results of the present study should be interpreted with caution. The patients studied may differ from patients who might be asked to participate in the proposed RCT. Whether patients indicating that they would participate “if the study were held today” would indeed participate in an actual RCT can only be determined by actually conducting a trial. Nevertheless, the information provided in this pilot feasibility study will be valuable for designing the RCT, and indicates that indeed such a trial is feasible. The sample size calculation that we conducted also indicates that a trial using the combined MACE endpoint would have enough statistical power to detect a clinically significant difference between 2 groups (P<0.05). We assumed that the only compelling reason to perform a cardiac evaluation prior to transplant is to improve patient outcomes. If centers use the results of screening to exclude patients from transplant, then the results of the proposed RCT may not influence their decisions for doing cardiac evaluations. However, it seems unlikely that many centers exclude patients on the basis of tests that are not proven to accurately predict patient outcomes. Finally, we restricted our study to stress tests and coronary angiography used to detect CAD. We did not examine tests used to screen for left ventricular dysfunction. We are not aware of evidence that screening for left ventricular dysfunction improves patient outcomes, but there are data that LV dysfunction often improves as a result of transplant per se.(15)
In summary, trials in the general population have failed to provide evidence that screening asymptomatic uremic patients for CAD before major non-cardiac vascular surgery is indicated. However, screening candidates for kidney transplantation is the standard of care in the US. The results of the current study suggest that it would be feasible to randomly allocate patients referred for kidney and/or pancreas transplant to either 1) following the ACC/AHA guidelines (no routine screening if asymptomatic) or 2) screening high-risk individuals according to standard practice. If the primary endpoint were MACE, we estimate that approximately 4000 patients would be needed to have at least 80% power to detect a 20% difference in MACE (P<0.05).
The COST Investigators (in alphabetical order) are Nada Alachkar (The Johns Hopkins University, Baltimore, MD), Roy D. Bloom, MD (University of Pennsylvania, Philadelphia, PA), Daniel C. Brennan, MD (Barnes-Jewish Hospital, Washington University, St. Louis, MO), Jonathan S. Bromberg, MD (Mount Sinai Medical Center, New York City, NY), Harini A. Chakkera, MD (Mayo Clinic, Phoenix, AZ), Laurence K.F. Chan, MD, PhD (University Colorado Health Sciences Center, Aurora, CO), David J. Cohen, MD (Columbia University Medical Center, New York City, NY), Lorenzo G. Gallon, MD (Northwestern University Hospital, Chicago, IL), William C. Goggins, MD (Indiana University, Clarian Transplant Institute, Indianapolis, IN), J. Harold Helderman, MD (Vanderbilt University School of Medicine, Nashville, TN), Donald E. Hricik, MD (Case Western Reserve University, Cleveland, OH), Ajay A.K.Israni, MD, MS (Hennepin County Medical Center, Minneapolis, MN), Steven Katznelson, MD (California Pacific Medical Center, San Francisco, CA), Alan B. Leichtman, MD (University of Michigan, Ann Arbor, MI), Jimmy A. Light, MD (Washington Hospital Center, Washington, DC), Connie C.L. Manske, MD (University of Minnesota, Minneapolis, MN), Thomas R. McCune, MD (Sentara Norfolk General Hospital, Norfolk, VA), Joseph M. Nogueira, MD (University of Maryland Medical System, Baltimore, MD), Todd E. Pesavento, MD (The Ohio State University, Columbus, OH), Mary Prendergast, MD (University of Alabama at Birmingham, Birmingham, AL), David Serur, MD (New York Presbyterian Hospital, Cornell University, New York, NY), Steve L. Tomlanovich, MD (University of California at San Francisco, San Francisco, CA), Rafael Villicana, MD (Cedars Sinai Medical Center, Los Angeles, CA), Francis L. Weng, MD, MSCE (St. Barnabas Medical Center, Livingston, NJ), Patricia West-Thielke, PharmD (University of Illinois, Chicago, IL), Carlos F. Zayas, MD (Piedmont Hospital, Atlanta, GA).
Support: This work was supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases,National Institutes of Health (No. 1R21DK080315-01; Project Period: 09/30/2007 – 08/31/2009).
Financial Disclosure: The authors declare that they have no relevant financial interests.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.