Cardiorespiratory Fitness (Peak Oxygen Uptake): Safe and Effective Measure for Cardiovascular Screening Before Kidney Transplant

Abstract

Background

Significant heterogeneity exists in practice patterns and algorithms used for cardiac screening before kidney transplant. Cardiorespiratory fitness, as measured by peak oxygen uptake (VO_2peak), is an established validated predictor of future cardiovascular morbidity and mortality in both healthy and diseased populations. The literature supports its use among asymptomatic patients in abrogating the need for further cardiac testing.

Methods and Results

We outlined a pre–renal transplant screening algorithm to incorporate VO _2peak testing among a population of asymptomatic high‐risk patients (with diabetes mellitus and/or >50 years of age). Only those with VO _2peak <17 mL/kg per minute (equivalent to <5 metabolic equivalents) underwent further noninvasive cardiac screening tests. We conducted a retrospective study of the a priori dichotomization of the VO _2peak <17 versus ≥17 mL/kg per minute to determine negative and positive predictive value of future cardiac events and all‐cause mortality. We report a high (>90%) negative predictive value, indicating that VO _2peak ≥17 mL/kg per minute is effective to rule out future cardiac events and all‐cause mortality. However, lower VO _2peak had low positive predictive value and should not be used as a reliable metric to predict future cardiac events and/or mortality. In addition, a simple mathematical calculation documented a cost savings of ≈$272 600 in the cardiac screening among our study cohort of 637 patients undergoing evaluation for kidney and/or pancreas transplant.

Conclusions

We conclude that incorporating an objective measure of cardiorespiratory fitness with VO _2peak is safe and allows for a cost savings in the cardiovascular screening protocol among higher‐risk phenotype (with diabetes mellitus and >50 years of age) being evaluated for kidney transplant.

Clinical Perspective

What Is New?

• There is significant heterogeneity in the cardiac screening process for patients with end‐stage renal disease being evaluated for kidney and/or kidney/pancreas transplant.

• There is little evidence to suggest that current conservative strategies are effective in mitigating future cardiovascular disease events.

What Are the Clinical Implications?

• We demonstrate that use of an objective metric of cardiorespiratory fitness, such as peak oxygen uptake in the pretransplant cardiovascular screening algorithm, had a high negative predictive value and resulted in minimizing additional cardiac testing.

• We advocate for the incorporation of cardiorespiratory fitness assessment (peak oxygen uptake) into the pretransplant cardiac screening algorithm.

Introduction

Cardiovascular disease (CVD) is the leading cause of mortality in patients with end‐stage renal disease and in those individuals receiving a kidney transplant. Despite several consensus statements, there is wide variability among centers relating to the cardiac evaluation of a patient before kidney transplant (Tables S1 and S2).1 The variability in practice patterns would indicate that no algorithm is uniformly considered ideal. In addition, none of these algorithms, to our knowledge, include an objective measure of cardiorespiratory fitness.

Peak oxygen uptake (VO_2peak) is an established objective measure of cardiorespiratory fitness and functional capacity and is a validated predictor of future cardiovascular morbidity and mortality in both healthy and diseased populations. VO_2peak <15 and <12 mL/kg per minute for men and women with diagnosed coronary heart disease, respectively, has been associated with the highest risk of death.2 In addition, the measurement of VO_2peak is low cost and noninvasive. It outperforms more expensive and invasive tests, such as right‐sided heart catheterization and echocardiogram, as an independent predictor of major adverse cardiovascular events and all‐cause mortality.2, 3 VO_2peak is the product of maximal cardiac output and maximal arterial‐venous oxygen difference. It represents the ability of the cardiopulmonary system to deliver oxygen to exercising tissue and the ability of peripheral tissues to extract and use oxygen.4 Individuals experiencing CVD have decreased tissue perfusion, thus necessitating higher tissue oxygen extraction and resulting in decreased VO_2peak. The typical resting metabolism of a human (resting VO₂) is 3.5 mL oxygen/kg per minute or by definition 1 metabolic equivalent (MET). Therefore, a 3 MET activity, such as walking at 2.5 mph, expends 3 times the energy used at rest. The MET values of typical activities of daily living have been extensively reported.5 The measured VO_2peak of an individual can be translated to their MET capacity by dividing their VO_2peak by 3.5.

The American College of Cardiology/American Heart Association (ACC/AHA) recommends no cardiac testing before intermediate risk surgery, such as a kidney transplant, in an asymptomatic patient with functional status ≥4 METs (ie, VO_2peak=14 mL/kg per minute), even among those with known cardiac risk factors. However, most US transplant centers do not adhere to the ACC/AHA recommendations and perform, at minimum, a noninvasive cardiac stress test among asymptomatic patients who fit the cardiac high‐risk criteria. The cardiac high‐risk criteria include any of the following: older patients (>50 years of age), presence of diabetes mellitus, and/or presence of coronary artery disease (CAD). Some centers perform coronary angiography on all cardiac high‐risk patients.

Despite this overcautious and conservative approach, there is little evidence to suggest that these strategies are effective in mitigating future CVD events, but they could potentially introduce more risk and cost with additional testing. None of the current pretransplant cardiac screening strategies objectively quantify and include cardiorespiratory fitness of the patient in the screening algorithm. We hypothesized that incorporating an objective measure of cardiorespiratory fitness, such as VO_2peak, allows for a safe and cost‐effective pretransplant cardiovascular screening process.

Methods

Cardiovascular screening for patients undergoing kidney or kidney/pancreas transplant evaluation at Mayo Clinic (Phoenix, AZ) has evolved over time. In October 2011, after extensive discussion with all stake holders, including cardiologists, transplant nephrologists, transplant surgeons, and anesthesiologists, the decision was made to incorporate VO_2peak, obtained by performing cardiopulmonary exercise test (CPET), into the standard‐of‐care pretransplant cardiac screening algorithm (Figure). All patients ≥50 years of age and/or with diabetes mellitus underwent CPET. The VO_2peak provided an objective measure of cardiorespiratory fitness and was a branch point metric in the algorithm. Only those with VO_2peak <17 mL/kg per minute underwent further cardiac screening with pharmacologic sestamibi stress test. Among the cohort with known ischemic heart disease with or without prior cardiac intervention, our protocol includes the CPET being performed along with pharmacologic sestamibi stress test, irrespective of the result of VO_2peak.

Figure 1.

Pretransplant cardiac screening algorithm for kidney and/or pancreas transplant. CABG indicates coronary artery bypass grafting; CAD, coronary artery disease; CPET, cardiopulmonary exercise test; DM, diabetes mellitus; MI, myocardial infarction; TTE, transthoracic echocardiogram; VO _2peak, peak oxygen uptake.

Because the ACC/AHA guidelines were not written specifically for patients with end‐stage renal disease with respect to the recommendations of no further cardiac testing among patients with functional capacity of >4 METs (VO_2peak=14 mL/kg per minute), we chose to adopt a more conservative measure of VO_2peak, <17 mL/kg per minute, to determine if noninvasive stress test should be performed. This is equivalent to an exertion of >5 METs. This more conservative cutoff for VO_2peak at <17 mL/kg per minute was chosen with the goal of optimizing the negative predictive value (NPV) of the test (ie, to identify those who did not require further testing). Thus, the threshold for VO_2peak at <17 mL/kg per minute was established a priori as a branch point metric in the screening algorithm to rule out the need for further cardiac testing.

Subjects underwent CPET testing on a treadmill/bicycle ergometer using a ramp protocol with the goal of achieving 6 to 9 minutes until voluntary exhaustion occurred. Gas exchange parameters were measured using a computerized breath‐by‐breath analyzer (Medgraphics Corp, St Paul, MN), which was calibrated before each test. A 12‐lead ECG and oxygen saturations were monitored throughout the study, along with periodic blood pressure measurements. The test was conducted by a respiratory technician/exercise specialist, along with a cardiology registered nurse, and interpreted by a cardiologist with specific training in exercise testing.

Study Cohort

We conducted a retrospective study of all patients with documented VO_2peak who were evaluated for kidney/pancreas transplant between November 2011 and September 2014. The study was reviewed and approved by the Institutional Review Board at Mayo Clinic. The informed consent requirement was waived. Patients had an average follow‐up of 4.04±1.11 years after baseline VO_2peak testing. During the follow‐up, study outcome events, including CVD event, all‐cause mortality, and a composite outcome of CVD event and all‐cause mortality, were documented. CVD event was defined as cardiac ischemic event, myocardial infarction, percutaneous transluminal coronary angioplasty, coronary artery bypass grafting, or cerebrovascular accident.

Statistical Analyses

The data, analytic methods, and study materials will not be made available to other researchers because the methods are purely simple statistical methods and transparent and raw data are protected information.

Patient demographic characteristics are displayed comparing those with VO_2peak <17 mL/kg per minute versus those with VO_2peak ≥17 mL/kg per minute using Pearson χ² test for categorical variables or 2‐sample t test for continuous variables. Log‐rank test was used to compare the outcomes among the 2 groups. The positive predictive value (PPV), NPV, sensitivity, and specificity of VO_2peak at end of study period were calculated. Patients with known CAD with VO_2peak ≥17 mL/kg per minute were excluded from the analyses calculating PPV and NPV because the algorithm requires a pharmacologic sestamibi stress test in all patients with history of CAD, irrespective of the value of VO_2peak. P<0.05 was considered statistically significant. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc, Cary, NC).

Results

Descriptive Analyses of Study Cohort

The study cohort included 637 recipients undergoing cardiac evaluation for kidney and/or pancreas transplant. Mean age of study cohort was 56.6 years, 74% were >50 years of age, 61% were men, 52% were diabetic, and 21% had a history of CAD (9% had cerebrovascular disease and 12% had known peripheral artery disease). During the average follow‐up of 1476.0±408.5 days, 292 individuals received transplant, in which 288 received a kidney transplant. Of the 288 who received a kidney transplant, 71 received living donor transplants and 217 received deceased donor transplants.

Mean VO_2peak of study cohort was 15.1±4.4 mL/kg per minute. Among the study cohort of 637 patients receiving CPET, 183 (29%) had a VO_2peak ≥17 mL/kg per minute, and 23 of the 183 had history of known CAD. As outlined in the screening algorithm, all patients with CAD had pharmacologic sestamibi stress test performed, irrespective of VO_2peak; as such, these 23 patients with VO_2peak ≥17 mL/kg per minute were excluded from the analyses. Interestingly, among these 23 patients, 16 (70%) had normal pharmacologic sestamibi stress test.

Thus, our study cohort for statistical analyses included a total of 160 patients with VO_2peak ≥17 mL/kg per minute who did not undergo further CVD testing, and 454 patients had VO_2peak <17 mL/kg per minute and underwent noninvasive stress test with pharmacologic sestamibi stress test.

Table 1 compares the demographics and baseline characteristics of the patients with VO_2peak <17 versus ≥17 mL/kg per minute. Those with VO_2peak <17 mL/kg per minute were older; were more likely to be women, have history of diabetes mellitus, have history of CVD, and have history of peripheral vascular disease; and were more likely to be past or present smoker and to be taking aspirin.

Table 1.

Comparison of Demographics and Comorbidities for VO_2peak <17 vs ≥17 mL/kg per Minute

Variables	VO2peak <17 mL/kg per min (N=454)	VO2peak ≥17 mL/kg per min (N=160)	P Value
Age, mean (SD), y	58.3 (11.9)	51.7 (14.1)	<0.0001
Aged ≥50 y, N (%)	354 (78)	98 (61)	<0.0001
Male sex, N (%)	260 (57)	110 (69)	0.011
Race, N (%)			0.58
Black	52 (11)	17 (11)
White	301 (66)	113 (71)
Others	101 (22)	33 (18)
History of comorbidities, N (%)
Hypertension	430 (95)	141 (88)	0.005
Diabetes mellitus	276 (61)	42 (26)	<0.001
Cerebrovascular disease	45 (10)	6 (4)	0.02
Peripheral vascular disease	72 (16)	3 (2)	<0.001
Hyperlipidemia	269 (59)	72 (45)	0.002
History of smoking, N (%)			0.002
Never	218 (48)	103 (64)
Past	177 (39)	42 (26)
Current	59 (13)	15 (9)
Pretransplant ASA, N (%)	186 (41)	49 (27)	<0.001
Type of transplant			0.003
Kidney alone	182 (40)	84 (53)
Simultaneous kidney and pancreas	9 (2)	3 (2)
Pancreas alone	0 (0.0)	2 (1)

ASA indicates acetylsalicylic acid; VO_2peak, peak oxygen uptake.

Outcomes: Comparison of Patients With VO_2peak <17 mL/kg per Minute Versus Those With VO_2peak ≥17 mL/kg per Minute

Cardiovascular events during study period

A total of 454 patients had VO_2peak <17 mL/kg per minute and underwent pharmacologic sestamibi stress test. As previously described, 160 patients with VO_2peak ≥17 mL/kg per minute did not undergo further CVD testing. The PPV and NPV of VO_2peak ≥17 mL/kg per minute and future CVD were calculated. We observed a high NPV. We observed this among the entire cohort and subgroups of those who underwent transplant during the follow‐up and those who remained on the wait list (Table 2).

Table 2.

Outcomes: Comparing VO_2peak <17 vs ≥17 mL/kg per Minute

Variable	VO2peak <17 mL/kg per min (N=Total Number), Event Number	VO2peak ≥17 mL/kg per min (N=Total Number), Event Number	Log‐Rank P Value	Sensitivity, %	Specificity, %	Positive Predictive Value, %	Negative Predictive Value, %
Cardiovascular event (ischemia/CABG/MI/CVA)
Entire cohort	(N=454) 45	(N=160) 8	0.0481	84.9	27.1	9.9	95.0
Wait‐listed cohort	(N=261) 30	(N=71) 5	0.2341	85.7	22.2	11.5	93.0
Transplanted cohort	(N=193) 9	(N=89) 2	0.323	81.8	32.1	4.7	97.8
All‐cause mortality
Entire cohort	(N=454) 79	(N=160) 17	0.0496	79.0	30.2	17.4	88.5
Wait‐listed cohort	(N=261) 63	(N=71) 16	0.7334	79.7	21.7	24.1	77.5
Transplanted cohort	(N=193) 16	(N=89) 1	0.0194	94.1	33.2	8.3	98.9
Composite outcome (includes all‐cause mortality and cardiovascular event)
Entire cohort	(N=454) 110	(N=160) 30	0.0043	84.0	28.8	24.2	86.9
Wait‐listed cohort	(N=261) 81	(N=71) 17	0.2204	82.7	23.1	31.0	76.1
Transplanted cohort	(N=193) 24	(N=89) 3	0.0184	88.9	33.7	12.4	96.6

CABG indicates coronary artery bypass grafting; CVA, cerebrovascular accident; MI, myocardial infarction; VO_2peak, peak oxygen uptake.

Among these 454 patients, with VO_2peak <17 mL/kg per minute, 28 had abnormal pharmacologic sestamibi stress test (presence of ischemia, infarction, and/or ejection fraction <40) requiring a referral to cardiologist for further recommendations. Cardiac (myocardial infarction, ischemia, percutaneous transluminal coronary angioplasty, coronary artery bypass grafting, and/or cerebrovascular accident) events among those patient with VO_2peak <17 versus ≥17 mL/kg per minute were 45 (10%) versus 13 (8%) events (log‐rank P=0.23). The detailed description of the 13 events among the patients with VO_2peak ≥17 mL/kg per minute is included in Table S3. None of these 13 patients had any significant electrocardiographic changes noted during the CPET test. There was no specific pattern with respect to timing, description of event, and patient characteristics among those with positive event versus those without cardiac event among patients with VO_2peak ≥17 mL/kg per minute.

Mortality during study period

The PPV and NPV of VO_2peak ≥17 mL/kg per minute and all‐cause mortality were calculated. We observed a high NPV. We observed this among the entire cohort and subgroup who underwent transplant during the follow‐up and those who remained on the wait list (Table 2). Total number of deaths during study period was 100 (15.7%). The death rate among those patients with VO_2peak <17 versus ≥17 mL/kg per minute at the end of study period was 17.4% versus 11.5% (log‐rank P=0.074).

Discussion

Cardiac testing before kidney transplant is highly debated, with no universally adopted pretransplant cardiac screening algorithm. Our proposed algorithm suggests that use of an objective metric of cardiorespiratory fitness, such as VO_2peak, in the pretransplant cardiovascular screening algorithm had a high NPV and resulted in minimizing additional cardiac testing. In our cohort, ≈25% (160 of 454) of high‐risk patients safely averted further cardiac testing, which resulted in substantial cost savings (≈$272 600) during the pretransplant cardiac screening.

The literature, including a recent ACC/AHA suggests there is wide variation and heterogeneity in the cardiac screening pretransplant (Tables S1 and S2).1 None of these recommendations use objective measure of cardiorespiratory fitness in the algorithm. The Kidney Disease Outcome Quality Initiative guidelines advocate a more liberal approach in the use of noninvasive stress tests for all high‐risk patients, which is at odds with the ACC/AHA.1 This could potentially raise concern from insurance companies and payers, questioning adoption of best practices that are not cost‐effective and not congruent with the major cardiac societies' recommendations with respect to cardiac screening of patients before kidney and/or pancreas transplant. We demonstrated that even in “higher‐risk” patient cohort, the high NPV of VO_2peak ≥17 mL/kg per minute supports ACC/AHA guidelines in this group and, thus, prevents unnecessary cardiac testing and results in significant cost savings.

The reimbursement charge of performing the CPET to obtain VO_2peak is ≈$200, and that of pharmacologic sestamibi stress test is >$2500. So, in this example, during study time period, 637 “high‐risk” cohorts were evaluated. High risk was defined as anyone with one of the following criteria: ≥50 years of age, presence of diabetes mellitus, and/or history of CAD. All the 637 individuals had the CPET, incurring cost of $127 400 ($200×637=$127 400). A total of 183 (29%) of the study cohort had a VO_2peak ≥17 mL/kg per minute; of these individuals, 23 had history of CAD and, thus, had pharmacologic sestamibi stress test, per the protocol. However, 160 patients who otherwise would have had the pharmacologic sestamibi stress test did not have it done [savings of $400 000 (160×$2500)]. Thus, total cost saving of cardiac screening tests in this example was $272 600 ($400 000–$127 400). Our simple arithmetic calculation alludes to significant cost savings in using this approach; however, a robust formal economic model with cost‐effectiveness analyses will need to be performed to validate the true cost‐effectiveness of this approach.

Cardiorespiratory fitness, as measured by VO_2peak as a marker of cardiorespiratory fitness, is a strong predictor of future cardiac events and all‐cause mortality6 in otherwise healthy individuals and in those with advanced CVD, such as congestive heart failure.7 The results of this present study are in keeping with the literature that supports the growing body of evidence that cardiorespiratory fitness, as measured by VO_2peak, can be safely used to rule out future cardiac events and mortality. Thus, VO_2peak testing could abrogate the need for more invasive cardiac testing in a cohort of patients with end‐stage renal disease being evaluated for kidney transplant. In addition, event rates in our study were consistent with reported national rates, which enhance the external validity of this study. External validation using our protocol by other transplant programs on their high‐risk phenotype will further consolidate the generalizability of our conclusions.

The goal of cardiac screening and testing before kidney transplant is to identify, intervene, and reduce future cardiovascular morbidity and mortality. The optimal screening test ideally will be cost‐effective, with a high true positive rate and high true negative rate or proportion. However, given the relatively low event rate, a high true PPV may be difficult to achieve. However, one can strive to have high confidence in a negative test. We report a high NPV >90%, indicating that the high VO_2peak (as defined by VO_2peak ≥17 mL/kg per minute in this study) among those patients who were able to perform the CPET is sufficient to rule out future cardiac events and all‐cause mortality. However, a low value cannot be used as a reliable predictive measure to determine development of future cardiac events or mortality. We conclude that the current cardiac screening algorithm before kidney and/or pancreas transplant needs to be refined. Incorporation of an objective measure of functional capacity using VO_2peak into the pretransplant cardiac screening algorithm is a safe “rule out” test for the requirement of further cardiac testing.

Disclosures

None.

Supporting information

Table S1. Published Recommendations for Testing for CAD in Asymptomatic Kidney Transplantation Candidates¹

Table S2. Summary of Survey and Registry Data Demonstrating Variation in Cardiac Evaluation Practices Across US Transplantation Centers¹

Table S3. Characteristics of patients with cardiovascular event within study period among cohort with VO_2peak >17 mL/kg per minute (N=183)

(J Am Heart Assoc. 2018;7:e008662 DOI: 10.1161/JAHA.118.008662.)29853444