Clin Transplant 2001: 15: 269–275Printed in Ireland. All rights reser�ed

Pre-transplant cardiac testing forkidney–pancreas transplant candidates andassociation with cardiac outcomes

Lin K, Stewart D, Cooper S, Davis CL. Pre-transplant cardiac testingfor kidney–pancreas transplant candidates and association with cardiacoutcomes.Clin Transplant 2001: 15: 269–275. © Munksgaard 2001

Abstract: Background: Coronary artery disease is a major cause ofmortality following renal transplantation, especially in those patientswith diabetes. The accurate prediction of cardiac risk is therefore amajor focus of the pre-transplant evaluation. The objective of thisstudy was to retrospectively evaluate the ability of non-invasive cardiactesting (standard echocardiography, stress echocardiography, exercisetolerance testing, and nuclear myocardial perfusion) performed within 1yr of kidney–pancreas transplant to predict post-transplant myocardialinfarction.Methods: Clinical history and pre-transplant cardiac testing performedwithin 1 yr prior to transplantation were reviewed in a non-blindedfashion for 165 kidney–pancreas transplantation patients receiving allo-grafts between June 1990 and May 1998. The predictive values of clini-cal symptoms and cardiac testing for cardiac events (fatal and non-fatalmyocardial infarctions) up to 1 yr post-transplant were calculated.Results: Clinical history had a negative predictive value of 98% forcardiac events occurring within 1 yr following testing and 97% within 1yr post-transplant. Collectively, non-invasive testing had a negative pre-dictive value of 97% for 1 yr post-testing and 1 yr post-transplant.Conclusion: Clinical history alone is highly suggestive but not sufficientfor the prediction of post-kidney–pancreas transplant myocardial in-farction. Although a useful supplement, cardiac testing does not predictall cardiac events out to 1 yr post-testing. In this high-risk patientpopulation with diabetes and renal failure, other methods of risk as-sessment are needed to more accurately predict long-term cardiac out-come for patients awaiting transplantation.

Karen Lina,b, Doug Stewartc,Stephanie Cooperc and ConnieL Davisa,b

Divisions of a Nephrology and c CardiologyDepartment of Medicine, b Division ofTransplantation, Department of Surgery,University of Washington Medical Center,Seattle, WA, USA

Key words: coronary disease –kidney–pancreas transplantation – riskassessment – risk factors

Corresponding author: Karen Lin, c/o Con-nie L. Davis, M.D., Box 356174, 1959 NEPacific Street, Seattle, WA 98195, USA.Tel: +1 206 5986079; fax: +1 2065986706; e-mail cdavis@u.washington.edu

Accepted for publication 16 April 2001

Coronary artery disease is responsible for approxi-mately 50% of peri-operative deaths following re-nal transplantation, especially in diabetic patients(1–6). The detection of coronary artery diseaseprior to transplantation has therefore been a majorgoal of transplant programs. Identification of is-chemic disease is a challenge, however, particularlyfor diabetic transplant candidates as often thosewith coronary disease are asymptomatic (2, 4, 6).Thus, identification of pre-transplant coronary dis-ease frequently depends upon cardiac testing (1–4,7). The testing options are numerous and carryvariable predictive values in different patient popu-

lations according to such factors as underlyingdiseases, age, sex, and body habitus (1, 7). TheAmerican Society of Transplantation reportedguidelines for the work-up of renal transplant can-didates in 1995 that included the use of non-inva-sive cardiac stress testing (7). However, theguidelines did not include a precise algorithm forcardiac testing due to the lack of firm support fora single test.

Over time, because of growing waiting lists andprolonged waiting times, the goal of pre-transplantcardiac testing has changed. The current goal isnot only to predict cardiac risk at the time of

269

Lin et al.

listing, but also to predict the risk for an extendedperiod of time on the waiting list. Furthermore,given the scarcity of organs, the early death of arecipient from an avoidable illness must be pre-vented. Thus, to better understand how effectiveour current program of combining clinical historyand non-invasive cardiac stress testing was at theprediction of myocardial infarction following kid-ney–pancreas transplantation, we undertook a ret-rospective analysis of all kidney–pancreastransplant recipients through May of 1998.

Patients and methods

The goal of the study was to evaluate the accuracyof clinical symptoms and non-invasive cardiac testsas screening strategies for pre-transplant cardiacdisease. The records of all University of Washing-ton kidney–pancreas recipients transplanted be-tween June 1990 and May 1998 (n=165) wereretrospectively analyzed from 1 yr prior to trans-plant onward to 1 yr post-transplant. Eight pa-tients received kidneys from living donors, and 152received cadaveric kidneys. Records were reviewedfor pre-transplant clinical symptoms, non-invasivecardiac testing, and cardiac events occurring within1 yr post-testing and 1 yr post-transplant. Allpatients underwent non-invasive cardiac testing us-ing a method that was determined by their historyof exercise tolerance and the medical expertisewhere their cardiac testing was performed. Giventhe distance to Seattle (from Alaska, Idaho, Mon-tana and around the state) and referring physicianpreference, many tests were performed in the pa-tient’s hometown. Positive or non-diagnostic non-invasive tests and patients of clinical concern werefurther investigated by angiography. Two cardiolo-gists reviewed all testing reports (from UWMCand referring centers) to standardize the interpreta-tion of findings.

Classification methodClinical symptoms of coronary disease

Clinical indicators of coronary disease such as my-ocardial infarction, angina, and congestive heartfailure within 1 yr prior to transplant wererecorded if clearly documented by the primaryphysician. Non-fatal myocardial infarctions weredefined by a doubling of the upper limits of normalCK-MB fraction and electrocardiogram (EKG)changes, including ST segment elevation. A cardiacevent occurring after transplant evaluation wasdefined as the development of a fatal or non-fatalmyocardial infarction.

Cardiac testing

All cardiac testing was re-evaluated by two cardiol-ogists (DS, SC) in a non-blinded fashion andclassified as normal, positive, or non-diagnostic fordisease based on standardized criteria.

Standard EKGs were labeled as ‘positive’ if wallmotion abnormalities or severe hypokinesis werepresent. Dobutamine stress EKGs were ‘positive’ ifinducible wall motion abnormalities, severe hy-pokinesis, a change in ventricular size, or EKGchanges were noted.

Exercise tolerance tests were ‘positive’ if two ormore of the following were present: inability toexercise longer than 6 min, ST depression of 2 mmor greater at peak exercise, an abnormal bloodpressure response, three or more exertional PVCs,or ventricular tachycardia. A test was ‘inconclu-sive’ if only one positive criterion was met: STdepression between 1 and 2 mm, or flat ordownsloping ST segments. A test was ‘non-diag-nostic’ if a sub-maximal heart rate was achieved.

Nuclear myocardial perfusion tests were ‘positive’if stress images demonstrated a large area of de-creased perfusion compared with resting images. Atest was ‘intermediate risk’ for coronary disease ifthere was a fixed defect on both rest and stressimages. Stress was induced by exercise, dipyri-damole, persantine, dobutamine, or adenosine.

Angiography was ‘positive’ if there was 70% orgreater stenosis in a major artery (LAD, RCA, leftcircumflex). It was ‘inconclusive’ if there wereglobal or regional ventricular wall motion abnor-malities or a 50–69% stenosis of a major coronaryartery or one of its branches.

ResultsPre-transplant clinical symptoms and post-transplantcardiac events

Thirty-two of the 165 patients had pre-transplantcoronary disease as determined by the presence ofangina, congestive heart failure, or myocardial in-farction within 1 yr prior to transplant (Table 1).Of the 28 patients with angina, 2 experienced amyocardial infarction within 1 yr post-transplant,1 during transplant surgery and the other 1 monthpost-transplant. In all, seven post-transplant car-diac events occurred at an average of 38 monthsfollowing transplant in patients with pre-transplantangina (range 0–72 months). One of the 4 patientswith congestive heart failure (CHF) had a fatalmyocardial infarction at 48 months post-trans-plant. No events occurred in the 5 patients withpre-transplant myocardial infarction. The averagetime for all post-transplant events in all patients

270

Cardiac testing in kidney–pancreas transplant

with pre-transplant clinical symptoms was 34months; the average patient follow-up in thosewith events was 47 months and in those withoutevents was 38 months. For clinical history, thenegative predictive value was 98% for events occur-ring within 1 yr post-evaluation and 97% for eventsoccurring within 1 yr post-transplant (Table 2).

Pre-transplant cardiac testing and post-transplant cardiacevents

The average length of time between cardiac work-up and transplantation was 11 months (range 1–32months), but only testing within 1 yr prior totransplant was included in the study.

Standard echocardiogram

Eighty-seven patients had standard, non-stressechocardiogram (ECHO) within 1 yr prior totransplantation. Three tests were positive for coro-nary disease. Two patients underwent intervention,one angioplasty and one coronary artery bypassgrafting (CABG). None of these 3 patients hadpost-transplant events. One patient with a normaltest had a post-transplant myocardial infarction 5months after transplantation. The negative predic-tive value of the echocardiogram for cardiac eventswas 93% for events occurring within 1 yr post-test-ing and 89% for events within 1 yr post-transplant(Table 2).

Stress echocardiogram

Stress echocardiography was performed in 45 pa-tients within 1 yr prior to transplantation. Stresswas induced by exercise according to the Bruceprotocol in 20 tests and by dobutamine in 25 tests.The two positive tests were investigated by angiog-raphy. Angiography was normal in one patientand revealed coronary disease requiring angio-plasty in the other. Neither patient experienced acardiac event. Five of the 43 patients (12%) withnegative tests had a myocardial infarction at anaverage of 8 months post-transplant, including twointra-operative events. The negative predictive

value was 93% for events within 1 yr post-testingand 88% for events within 1 yr post-transplant(Table 2).

Exercise tolerance test

Nineteen exercise tolerance tests were performed.No test was positive however, 2 patients had un-dergone coronary artery bypass prior to the exer-cise test. Neither of these patients hadpost-transplant events. One patient with a negativetest had an intra-operative myocardial infarction.The negative predictive value was 95% for both 1yr post-testing and 1 yr post-transplant (Table 2).

Myocardial perfusion test

Ninety-five myocardial perfusion studies (78 thal-lium, 17 sestamibi) were performed. Of the thal-lium scans, there were 51 exercise, 12dipyridamole, 8 persantine, 5 dobutamine, and 2adenosine. The sestamibi tests consisted of 8 exer-cise, 3 dipyridamole, 3 persantine, 2 adenosine,and 1 dobutamine. Twelve patients had positivestudies; 10 underwent angiography, and 2 were notfurther evaluated. Of the 10 angiograms, therewere six negative studies while four positive studieslead to CABG. Six of the 12 positive studies werepreceded by angina; two with angina also had hada pre-test myocardial infarction. Of the 12 patientswith positive tests, two deaths occurred within 1 yrpost-transplant, one from myocardial infarctionand one from pulmonary embolism. The infarctoccurred in one of the 2 patients experiencingpre-transplant angina who did not undergo furthercardiac evaluation or treatment. Among the 83patients with negative tests, five post-transplantmyocardial infarctions occurred within 1 yr post-transplant at an average of 4 months, with twointra-operative events. The negative predictivevalue was 94% for events occurring within 1 yrpost-testing and 93% for events occurring within 1yr post-transplant (Table 2).

In summary, cardiac work-ups utilizing non-in-vasive testing had a negative predictive value of97% with a 3% risk of missing coronary disease.

Table 1. Pre-transplant clinical symptoms and cardiac events within 1 yr of transplantation

Pre-transplant symptoms Post-operative events No post-operative events Total cases(n=165)(n=159)(n=6)

2826Angina 20 5Pre-transplant MI 5

440CHF32Any of above 302

No symptoms 1331294

271

Lin et al.

Table 2. Predictive value of non-invasive cardiac testing up to 1 yr post-transplant

Sensitivity Specificity PPVTest NPV NPV(%) (%) (%) 1 yr post-transplant (%) 1 yr post-testing (%)

33 81 6History 97 9817 88All cardiac testing 5 97 970 97 0Echocardiogram 89 93

Stress echocardiogram 0 95 0 88 93Exercise tolerance N/A 100 N/A 95 95

20 88 8Myocardial perfusion 93 94

Accuracy declined as the length of follow-up in-creased from 1 yr post-test to 1 yr post-transplant(Table 2). Of the 20 patients with positive tests, 10had an intervention and 1 patient undergoing an-gioplasty had an event (Table 3, Fig. 1). In patientswith both a positive clinical history and positivenon-invasive testing, only 25% (2 of 8) had acardiac event within 1 yr of transplant. Cardiacevents occurred, on average, at 3 months withpositive testing and at 5 months in those withnegative testing.

Angiography

Fifty-one angiograms were performed followingpositive or non-diagnostic non-invasive testing orin patients of clinical concern. Eighteen an-giograms followed positive non-invasive testing.Eight of the 9 patients with positive angiogramshad angina; 2 had experienced a pre-transplantmyocardial infarction; and 1 had CHF. All 9 pa-tients were treated. Treatment included medica-tion, angioplasty, stent, or CABG (6 patients). Thepatient who underwent angioplasty had a non-fatalmyocardial infarction at 5 months post-transplant;no other events occurred. Four patients (10%) withnegative angiograms had post-transplant myocar-dial infarctions on average at 48 months (range:0–72 months, including one intra-operative event).

Pre-transplant CABG

Seven patients underwent pre-transplant CABGfollowing positive studies. One patient experiencedpost-operative renal failure and infection. Noneexperienced post-transplant cardiac events (Table3).

Post-transplant outcomes

There were seven cardiac events (4.2%) within 1 yrpost-transplant, including five non-fatal and twofatal myocardial infarctions. The five non-fatalmyocardial infarctions occurred within 1 yr follow-ing negative cardiac testing. Of the 7 patients, 2

had pre-transplant angina that preceded negativecardiac testing. There were two events predicted bypositive testing, and five events were missed bypre-transplant testing. The location of the eventcould not be predicted from testing. There were atotal of 14 post-transplant myocardial infarctions(seven after the first post-transplant year) and 24deaths. During the entire period of follow-up therewere eight deaths attributed to cardiac events andthe remainder to hepatic disease, stroke, sepsis,and pulmonary embolism (range: intra-operativeto 72 months).

Discussion

Coronary artery disease is a major cause of mor-bidity and mortality among transplant recipients,accounting for 15–50% of deaths following renaltransplantation (1–6). The disease risk is higher inthe transplant population than in the general pop-ulation due to an increased prevalence of estab-lished risk factors (hypertension, diabetes,hyperlipidemia, hyperhomocysteinemia) as well asfactors unique to transplantation (acute rejections,steroid use, prevalence of hyperlipidemia, cy-tomegalovirus (CMV) activation) (8–10). Identifi-cation of coronary disease has become a majorfocus of the pre-transplant evaluation in order toimprove patient and allograft survival by initiatingappropriate treatment strategies prior to transplan-tation (7, 11).

Table 3. Medical interventions as a result of positive non-invasive cardiactesting or angiography*

Number of patients Cardiac eventsTreatment modality

01MedicationAngioplasty 1 1Stent and angioplasty 1 0Coronary artery bypass 07

graft – CABGAny of the above 110No treatment 10 1

*From 20 patients with positive non-invasive cardiac testing.

272

Cardiac testing in kidney–pancreas transplant

Fig. 1. Diagram of testing and interventions up to 1 yr post-transplant in kidney–pancreas transplant candidates. Interventionincludes medication treatment, angioplasty, stent placement, or CABG. No intervention means no further testing following thepositive testing or no further testing following negative angiogram.

There is no generally accepted algorithm forpre-transplant cardiac risk stratification for kid-ney–pancreas transplant programs. The variabilityof cardiac testing accuracy has resulted in greatdiversity of the required pre-transplant cardiacevaluation among transplant centers, ranging fromminimal (clinical history only) to extensive (an-giography in all candidates) (2, 3). Likewise, aconsensus has not been reached about re-cardiactesting of patients while on the transplant waitinglist (4, 5).

Although there are no accepted pre-transplantcardiac evaluation algorithms, the most commonapproach utilizes a combination of clinical criteria,non-invasive testing, and, when indicated by test-ing, angiography (1–6, 12). Screening by clinicalrisk factors has been reported to potentially elimi-nate the need for cardiac testing in 50% of renaltransplant candidates (13). However, because pan-creas transplantation involves increased stressescompared with kidney-only transplantation andbecause one-third of diabetic transplant candidatesmay have clinically silent coronary disease, screen-ing the candidates using clinical symptoms alonehas been felt to be inadequate (2, 6).

Exercise-based screening tests are a conventionalway of assessing coronary artery disease; however,diabetics with severe neuromyopathy and poorconditioning may not be able to adequately per-form such tests. The sensitivity and specificity, 84

and 87% respectfully, decrease to 67 and 62% inthe presence of diabetes (1). For patients unable tosatisfactorily complete exercise testing, stress nu-clear scans provide a sensitivity of 86%, specificityof 79%, and negative predictive value of 83% (4).However, the value of nuclear perfusion scans re-mains uncertain in diabetic patients due to the highprevalence of both left ventricular hypertrophy(LVH) (resulting in false positive tests) and bal-anced ischemia secondary to diffuse small vesselcoronary atherosclerosis (causing false negative re-sults) (2). Another testing option is the dobutamineechocardiogram which has been reported to be aneffective screening test for the prediction of post-operative and long-term cardiac events in renaltransplant recipients (5). It has a reported sensitiv-ity of 95%, specificity of 86%, and a negativepredictive value of 97% in some studies (6). How-ever, this test as the others, is dependent upon theexpertise of the individuals interpreting the imageswith false positive tests not uncommon in thosewith pre-test wall motion abnormalities or car-diomyopathies and false negative results in thosewho fail to attain an adequate heart rate response(1, 5, 6). In fact, the use of both nuclear perfusionand echocardiography may provide more meaning-ful information than either alone by correlatingmuscle perfusion with muscle function (1, 14).

Currently, there are 48000 patients awaiting akidney transplant. The average waiting time is

273

Lin et al.

between 1 and 5 years depending on geographiclocation, blood type, and sensitization to humanleukocyte (HLA) antigens (www.unos.org). Thereare 2500 patients awaiting kidney–pancreas trans-plantation and 1100 awaiting a pancreas alonetransplant (www.unos.org). As a consequence ofthe prolonged waiting times, not only is there aneed to identify cardiac risk at the initial trans-plant evaluation, but also the patient’s risk whilewaiting on the list. In order to establish the mostaccurate and cost-effective initial evaluation strat-egy, and provide guidelines for cardiac re-evalua-tion while awaiting transplantation, the predictivevalues of each test and the duration of predictabil-ity needs to be more firmly established.

Our study demonstrates that clinical history andcardiac testing performed within 1 yr prior totransplant appear to have resulted in a decrease incardiac events by identifying patients with signifi-cant coronary disease pre-transplant. However, be-cause not all tests were performed in all patientsand because testing results were used to alter ther-apy, the exact predictive values of this approachcannot be calculated. Yet again, this approachlikely resulted in an improved negative predictivevalue for testing in this study compared to others.Clinical history alone appears to be highly sugges-tive of the future risk, although not foolproof, witha negative predictive value of 98% within 1 yr oftesting.

Overall, non-invasive cardiac stress testing inour patients with diabetes and renal failure had anegative predictive value of 97% within 1 yr post-testing and 1 yr post-transplant. It also resulted ina change in clinical therapy that very likely im-proved the outcome in patients undergoing CABGor stenting. Non-invasive testing correlated withclinical history in only 25% of patients. However,although non-invasive testing was a useful supple-ment to clinical history, it also missed coronarydisease. The least effective screening test in ourarea was stress echocardiography with a negativepredictive value of 93% for events occurring within1 yr post-testing and 88% for events occurringwithin 1 yr post-transplant. There were no compli-cations associated with cardiac testing. In all, 5patients with a negative history and negative test-ing had cardiac events within 1 yr of transplan-tation.

It is clear that the occurrence of cardiac eventsin patients without symptoms or findings on non-invasive testing or even angiography indicate thatanother approach to cardiac evaluation is war-ranted. Transplantation is an elective surgical pro-cedure meant to provide an increased quality oflife and longevity for patients with end-stage renal

disease. Furthermore, cardiac events may result inthe loss of the patient and the allograft, an increas-ingly rare gift. Therefore, the goal of the pre-trans-plant cardiac evaluation remains more exactingthan it is prior to other procedures. In order toaccomplish this goal, perhaps testing intervalsshould be changed to every 6 months while pa-tients are listed and a means to more accuratelypredict thrombogenic and inflammatory potentialshould be developed (15, 16). Although data fromthis and prior studies do not allow the outline of abest test algorithm, pre-transplant care should fo-cus on limiting known risk factors, smoking, ho-mocysteine levels, oxidant stress, hyperlipidemia,hypertension, the development of LVH, and pro-thrombotic states (1, 8, 17–22).

AcknowledgementsThis work was supported in part by the Student Scholars inCardiovascular Disease and Stroke Award from the Amer-ican Heart Association, Dallas, Texas.

References

1. BRAUN WE, MARWICK TH. Coronary artery disease inrenal transplant recipients. Cleve Clin J Med 1992: 61: 370.

2. MANSKE CL, THOMAS W, WANG Y, WILSON RF. Screen-ing diabetic transplant candidates for coronary artery dis-ease: identification of a low risk subgroup. Kidney Int1993: 44: 617.

3. LE A, WILSON R, DOUEK K et al. Prospective risk stratifi-cation in renal transplant candidates for cardiac death. AmJ Kidney Dis 1994: 24: 65.

4. MISTRY BM, BASTANI B, SOLOMON H et al. Prognosticvalue of dipyridamole thallium-201 screening to minimizeperioperative cardiac complications in diabetics undergo-ing kidney or kidney–pancreas transplantation. ClinTransplant 1998: 12: 130.

5. BRENNAN DC, VEDALA G, ANSTEY ME et al. Pretrans-plant dobutamine stress echocardiography is useful andcost-effective in renal transplant candidates. TransplantProc 1997: 29: 233.

6. BATES JR, SAWADA SG, SEGAR DS et al. Evaluation usingdobutamine stress echocardiography in patients with in-sulin-dependent diabetes mellitus before kidney and/orpancreas transplantation. Am J Cardiol 1996: 77: 175.

7. KASISKE BL, RAMOS EL, GASTON RS et al. The evalua-tion of renal transplant candidates: clinical practice guide-lines. J Am Soc Nephrol 1995: 6: 1.

8. SILKENSEN JR. Long-term complications in renal trans-plantation. J Am Soc Nephrol 2000: 11: 582.

9. KASISKE BL, CHAKKERA HA, ROEL J. Explained andunexplained ischemic heart disease risk after renal trans-plantation. J Am Soc Nephrol 2000: 11 (9): 1735.

10. KASISKE BL. Risk factors for accelerated atherosclerosis inrenal transplant recipients. Am J Med 1988: 84 (6): 985.

11. KASISKE BL. Risk factors for cardiovascular disease afterrenal transplantation. Miner Electrolyte Metab 1993: 19(3): 186.

12. WILSON RA, NORMAN DJ, BARRY JM et al. Noninvasivecardiac testing in the end-stage renal disease patient. BloodPurif 1994: 12: 78.

274

Cardiac testing in kidney–pancreas transplant

13. MARWICK TH, STEINMULLER DR, UNDERWOOD DA etal. Ineffectiveness of dipyridamole spect thallium imagingas a screening technique for coronary artery disease inpatients with end-stage renal failure. Transplantation 1990:49: 100.

14. AMANULLAH AM, CHAUDHRY FA, HEO J et al. Compari-son of dobutamine echocardiography, dobutamine ses-tamibi, and rest-redistribution thallium-201 single-photonemission computed tomography for determining contrac-tile reserve and myocardial ischemia in ischemic cardiomy-opathy. Am J Cardiol 1999: 84 (6): 626.

15. SHAH PK. Plaque disruption and thrombosis: potentialrole of inflammation and infection. Cardiol Rev 2000: 8(1): 31.

16. PAHOR M, ELAM MB, GARRISON RJ, KRITCHEVSKY SB,APPLEGATE WB. Emerging noninvasive biochemical mea-sures to predict cardiovascular risk. Arch Intern Med 1999:159 (3): 237.

17. LAIGHT DW, CARRIER MJ, ANGGARD EE. Antioxidants,diabetes and endothelial dysfunction. Cardiovasc Res2000: 47 (3): 457.

18. HOOGWERF BJ, YOUNG JB. The HOPE study. Ramiprillowered cardiovascular risk, but vitamin E did not. CleveClin J Med 2000: 67 (4): 287.

19. LUFT FC. Renal disease as a risk factor for cardiovasculardisease. Basic Res Cardiol 2000: 95 (Suppl 1): 172.

20. BROWN DW, GILES WH, CROFT JB. Left ventricularhypertrophy as a predictor of coronary heart disease mor-tality and the effect of hypertension. Am Heart J 2000: 140(6): 848.

21. BOAZ M, SMETANA S, WEINSTEIN T et al. Secondaryprevention with antioxidants of cardiovascular disease inendstage renal disease (SPACE): randomized placebo-con-trolled trial. Lancet 2000: 356 (9237): 1213.

22. KASISKE BL. Cardiovascular disease after renal transplan-tation. Semin Nephrol 2000: 20 (2): 176.

275