Cardiac Risk Assessment in Liver Transplant Candidates: A Survey of National Practice Patterns

TO THE EDITOR:

In liver transplantation (LT), improvements in surgical techniques, immunosuppression strategies, and infection management have extended the lifespan of both the graft and the patient. In 2021, cardiac disease is the leading contributor to health care use and morbidity after LT.⁽¹⁾ This observation is a result of performing transplants in patients of older age, the rise of nonalcoholic steatohepatitis (NASH) as a leading indication for transplantation, and the relationship between the unique physiology of cirrhosis and pathologic changes in myocardial structure and function.⁽²⁾ In this context, the burden of pretransplant cardiac risk is rising, and accurate estimation of this risk is paramount for appropriate resource allocation, optimization of clinical outcomes, and ultimate posttransplant success. However, there has been inadequate inclusion of patients with cirrhosis in clinical practice guidelines for preoperative cardiac evaluation and the resultant lack of clear evidence-based guidance for cardiac risk estimation in LT. How transplant centers are evaluating cardiac risk prior to LT in practice is unknown.

To address this gap in knowledge, we sought to characterize current practice patterns by surveying adult LT programs across the United States using Research Electronic Database Capture (Vanderbilt University, Nashville, TN). The survey was distributed to transplant hepatologists, transplant surgeons, and anesthesiologists between July 2020 and October 2020 and addressed personnel involved, protocols used, and specific modalities used in the risk assessment and diagnosis of coronary artery disease (CAD) and cardiac function (Supplement A). We hypothesize that there is substantial variability between transplant programs’ protocols for screening cardiac disease and managing identified cardiac pathology.

Results

Of 116 LT programs surveyed, 61 (53%) programs responded across all 11 United Network for Organ Sharing (UNOS) regions. Respondents from the programs included the following: 2 transplant surgeons, 18 anesthesiologists, and 41 transplant hepatologists. Of the 61 respondent programs, the mean number of total transplants performed in 2019 was 83 (range, 12–169) with an average of 56 (67%) performed in patients with Model for End-Stage Liver Disease (MELD) scores >25; the median MELD score at transplant was 27.7. Of the 55 nonresponding programs, the mean number of total transplants performed was 44 (range, 3–203) with an average of 27 (63%) performed in patients with MELD scores >25; the median MELD score at transplant was 28.3. There was no significant difference in the number of transplants performed in patients with MELD scores >25 (Fisher’s exact test; P = 0.56) or median MELD at transplant (t test; P = 0.65). Data are presented in Table 1.

TABLE 1.

Descriptive Statistics of the Responding and Nonresponding Programs

Variable	Responding Programs	Nonresponding Programs	P Value
Number of programs	61	55
Mean transplants performed (range)	83 (12–169)	44 (3–203)
LT patients with MELD scores >25, n (%)	56 (67)	27 (63)	0.56 (Fisher's exact test)
Median MELD score at transplant	27.7	28.3	0.65 (t test)

Results are graphically summarized in Fig. 1. The majority (n = 44; 72%) of programs followed written protocols for cardiac risk assessment, with 18 (41%) programs’ protocols in place for at least 5 years. With respect to personnel involved, a dedicated cardiologist was used for risk stratification in 36 programs (59%), with 50% of these involved in selection. Most programs (70%) involved an anesthesiologist in risk stratification; 73% of these were involved in selection.

FIG. 1.

The results of a national survey assessing practice patterns of cardiac risk assessment in LT candidates. This figure proposes a comprehensive algorithm (gray boxes, light shading) for assessing cardiac risk and the corresponding results of the survey (dark gray shading). Additional response details for a specific part of the algorithm are provided in the colored boxes. Red text highlights key unanswered questions that future research can focus on.

For the assessment of cardiac risk, 92% of programs used a risk-based approach with the most common risk factors being prior myocardial infarction/stroke (89%), prior revascularization (89%), diabetes mellitus (87%), age (84%), and active/past tobacco use (80%). Only 34 programs (56%) used NASH as a risk factor. For CAD detection, most programs used a 12-lead electrocardiogram (EKG; 84%) and dobutamine stress echocardiogram (DSE; 80%). Less than half (49%) used single-photon emission computed tomography (SPECT), and only 30% routinely used coronary computed tomography angiography (CCTA). There was a split in programs choosing to defer invasive coronary angiography to before listing for LT (52%); however, only 8 (13%) had a specific management protocol for identified CAD. For the assessment of cardiac function, nearly all programs (93%) obtained a transthoracic echocardiogram (TTE), although only 8 (13%) programs routinely used myocardial deformation imaging, and 56% specifically used DSE to assess cardiac function. Finally, most programs repeated assessments for both CAD and cardiac function (85% and 89%, respectively), although the time interval varied from <6 months to >2 years.

Discussion

Our nationally representative survey demonstrates a heterogeneous approach to cardiac risk assessment and testing across US transplant centers that does not reflect current practice-based recommendations.^(3,4) Reported approaches at the majority of centers likely reflect reliance on older American Heart Association/American College of Cardiology and American Association for the Study of Liver Diseases guidelines last updated in 2012.^(5,6) For example, these support routine use of DSE for cardiac risk estimation, which has subsequently been demonstrated to have poor accuracy and predictive value in patients with cirrhosis.⁽²⁾ As a transplant community, we are lagging behind new evidence-based practice guidance for cardiac risk assessment in this population.⁽³⁾ First, only half of the programs use a dedicated cardiologist in evaluation, with a further minority involved in candidate selection despite universal recommendations across multiple guidance documents⁽⁶⁾ on the importance of “routine involvement” of a dedicated cardiology team for cardiac risk evaluation. This variation may stem from a lack of clarity around what constitutes “routine involvement” and the impact that level of involvement may have on cardiac outcomes. Second, there is emerging guidance on the evaluation of subclinical myocardial dysfunction using advanced echocardiography including tissue Doppler and myocardial deformation imaging.⁽⁴⁾ These diagnostic studies have been shown to predict posttransplant cardiovascular disease and are included in the new proposed criteria for defining cirrhotic cardiomyopathy,⁽⁷⁾ yet very few programs routinely perform them. Finally, there is a large body of literature in the general population, with early evidence for safety and efficacy in LT candidates, on the role of anatomic assessment of CAD using noninvasive computed tomography imaging.^(2,8) There have also been substantial advancements in the approach to asymptomatic CAD in kidney transplant recipients⁽⁹⁾ favoring a less invasive approach. There is a critical need for well-designed clinical studies in LT candidates to determine the optimal approach to the detection and management of asymptomatic CAD.

An additional area of uncertainty revolves around the timing and implications of subsequent assessment of cardiac risk. As our data show, most programs repeat cardiac assessments. However, how changes in cardiac risk influence transplant status (ie, waitlist removal, internal holds) are unknown. The currently available data on cardiac risk in LT do not provide guidance on what relative or absolute change in cardiac risk is prohibitive for transplantation. It is plausible that repeat cardiac testing does not in fact alter transplant decisions (or outcomes) once a patient is listed.

The principal limitation of this article is the possibility of response bias as a result of the potentially important differences in the types of programs that tended to respond to the survey. Specifically, the programs that responded to the survey performed more transplants per year than the programs that did not respond. However, the number of high MELD (MELD >25) transplants were similar between the responding and nonresponding programs.

However, this survey suggests that there is a heterogenous approach to cardiac evaluation for LT in the United States. In addition, it suggests that, despite the presence of contemporary evidenced-based guidance statements for cardiac risk assessment in LT, many centers continue to base their cardiac evaluations on older, outdated recommendations. The best quality assessment of cardiac risk is critically important given the scarcity of this lifesaving resource. However, whether this practice variability impacts patient-level and center-level outcomes remains unknown as currently UNOS/Organ Procurement and Transplantation Network does not capture cardiac outcomes after LT. This represents an area ripe for standardization to improve the identification of cardiac disease and optimize risk for cardiac events after transplant. Given the current variability in practice approaches, it may be of great benefit for programs to freely share and compare risk assessment protocols to be able to critically analyze algorithms and resultant cardiac outcomes. Although this approach has been initiated internationally,^(10,11) there are no such current efforts within the United States, hampering our ability to arrive at a consensus regarding the optimal approach to cardiac risk evaluation in pre-LT candidates.

Abbreviations:

CAD

coronary artery disease

CCTA

coronary computed tomography angiography

DSE

dobutamine stress echocardiogram

EKG

electrocardiogram

FRAILT

Liver Frailty Index

Hx

history of

LT

liver transplantation

MELD

Model for End-Stage Liver Disease

MR

magnetic resonance

NASH

nonalcoholic steatohepatitis

SPECT

single-photon emission computed tomography

TTE

transthoracic echocardiogram

UNOS

United Network for Organ Sharing