Cardiovascular Disease Outcomes Related to Early Stage Renal Impairment After Liver Transplantation

Abstract

Background

In the general population, even mild renal disease is associated with increased cardiovascular (CV) complications. Whether this is true in liver transplant recipients (LTR) is unknown.

Methods

This was a retrospective cohort study of 671 LTR (2002-2012) from a large urban tertiary care center and 37 322 LTR using Vizient hospitalization data linked to the United Network for Organ Sharing. The 4-variable Modification of Diet in Renal Disease equation estimated glomerular filtration rate (eGFR). Outcomes were 1-year CV complications (death/hospitalization from myocardial infarction, heart failure, atrial fibrillation, cardiac arrest, pulmonary embolism, or stroke) and mortality. Latent mixture modeling identified trajectories in eGFR in the first liver transplantation (LT) year in the 671 patients.

Results

Mean (SD) eGFR was 72.1 (45.7) mL/min per 1.73 m². Six distinct eGFR trajectories were identified in the local cohort (n = 671): qualitatively normal-slow decrease (4% of cohort), normal-rapid decrease (4%), mild-stable (18%), mild-slow decrease (35%), moderate-stable (30%), and severe-stable (9%). In multivariable analyses adjusted for confounders and baseline eGFR, the greatest odds of 1-year CV complications were in the normal-rapid decrease group (odds ratio, 10.6; 95% confidence interval, 3.0-36.9). Among the national cohort, each 5-unit lower eGFR at LT was associated with a 2% and 5% higher hazard of all-cause and CV-mortality, respectively (P < 0.0001), independent of multiple confounders.

Conclusions

Even mild renal disease at the time of LT is a risk factor for posttransplant all-cause and CV mortality. More rapid declines in eGFR soon after LT correlate with risk of adverse CV outcomes, highlighting the need to study whether early renal preservation interventions also reduce CV complications.

Chronic kidney disease (CKD) is a major complication after liver transplantation (LT). This is in part due to the use of nephrotoxic calcineurin inhibitors (CNI), cyclosporine and tacrolimus, the standard immunosuppressive agents given after LT.1,2 More than half of LT recipients have stage 3 CKD (glomerular filtration rate [GFR], 30-59 mL/min per 1.73 m²) by 2 years post-LT and up to 20% progress to ≥ stage 4 (GFR, <30 mL/min per 1.73 m²) within 5 years.2–6 Recipients with advanced CKD have poor long-term survival and may eventually require kidney transplantation, requiring additional costs and organ resources.

In the general population, it is well known that advanced CKD is associated with significant cardiovascular (CV) risks.7 It has also been shown that even less advanced CKD (GFR, 60-89; stage 2) is associated with an increased risk of both all-cause and CV-specific mortality: each 10-point lower GFR below 90 mL/min per 1.73 m² is associated with a 10% to 15% higher relative risk of death or nonfatal CV complications.8–11 Cardiovascular risk increases over time after LT12–14 and is a leading cause of morbidity and mortality, particularly within the first year of LT.12,15–17 Although it may be assumed that there is an inverse association between GFR and all-cause and CV-specific mortality in a LT population, even in those with less advanced CKD, this has never been proven. If such an association was clearly made, LT practice patterns, such as CNI minimization or withdrawal, might potentially change to preserve GFR even at these higher levels.18,19

Current risk prediction models only take into account GFR level at baseline, usually at the time of LT and do not consider the impact of dynamic changes in GFR over time. Glomerular filtration rate trajectory patterns during the first year after LT and their impact on early (1 year) CV, and other complications have not been described. We hypothesized that temporal patterns of GFR change (eg, trajectories) within the first year of LT might better estimate risk of posttransplant CV complications compared with single time points, and that even milder forms of CKD (eg, GFR 60-90 mL/min per 1.73m²) will be associated with increased all-cause and CV-specific mortality after LT. To test these hypotheses, the aims of this study were (1) to characterize subgroups of individuals with similar trajectories in GFR within the first year of LT, (2) to determine the association of these GFR trajectories with CV complications compared with single time point GFR estimates, and (3) to determine the impact of early CKD on all-cause and CV-specific mortality among a national LT cohort.

PATIENTS AND METHODS

Protocol, Design, Data Sources, and Inclusion Criteria

This was a retrospective cohort study comprised of a local LT cohort (Northwestern Medicine Enterprise Data Warehouse [NMEDW]) for initial hypothesis testing followed by validation using data from a national LT cohort (Vizient). The NMEDW currently stores 35 terabytes of data on ~6.4 million unique patients, including 30 million outpatient encounters, 4 million inpatient hospitalizations, and 272 million clinical and laboratory diagnostic results since approximately 1998. The national cohort included hospitalization-level data from Vizient that was linked to the Organ Procurement and Transplantation Network (OPTN) LT database. Vizient is currently the largest network of healthcare systems in the United States and includes the University HealthSystem Consortium (UHC), which is an alliance of over 117 academic hospitals and 300 affiliates representing 90% of the nation’s nonprofit academic medical centers. Vizient’s database included self-reported billing claims data from 68 US transplant centers. The linking procedure for these databases has been described previously.16

For the current study, analyses were restricted to all adult (age ≥ 18 years) patients who underwent first liver-alone transplant from February 1, 2002, to December 31, 2011, with follow-up data through December 31, 2012. The Northwestern University Institutional Review Board and the Health Resources and Services Administration of the United States Department of Health and Human Services through the OPTN approved the study protocol.

Assessment of Renal Function

Patients were classified according to the estimated glomerular filtration rate (eGFR) with the use of the 4-variable Modification of Diet in Renal Disease equation incorporating age, race, sex, and serum creatinine level: eGFR = 175 × SerumCr − 1.154 × age − 0.203.20 For women and blacks, the product of this equation was multiplied by a correction factor of 0.742 and 1.21, respectively.20,21 We also divided eGFR into 7 ordinal categories (ie, eGFR <15, 15-29, 30-44, 45-59, 60-74.9, 75.0-89.9, and at least 90.0 mL/min per 1.73 m²) incorporating the guidelines of the National Kidney Foundation and the Kidney Disease: Improving Global Outcomes CKD Work Group.21,22

Covariates

Prevalent comorbid conditions at the time of transplant were assessed by admission International Classification of Diseases Ninth Revision codes present on admission at the initial transplant admission and/or Medicare severity diagnosis-related groups. Transplantation for nonalcoholic steatohepatitis (NASH) was defined as a primary listing diagnosis for NASH or cryptogenic cirrhosis with at least 1 risk factor for the metabolic syndrome (pretransplant obesity, diabetes, hypertension, and/or dyslipidemia). Smoking status was assessed using natural language processing, which is a validated technique that converts free text information relevant to smoking status that is commonly found in clinical documentation into a structured form that can be readily processed and analyzed.23 Potential risk factors for CV-related morbidity and mortality after LT were examined based on a comprehensive set of a priori clinical hypotheses. These risk variables included known traditional (eg, diabetes, smoking status) CV disease risk markers as well as transplant-specific critical illness indicators known to contribute to competing mortality risk (eg, Model for End-Stage Liver Disease [MELD] score). The full range of recipient, donor, operative and perioperative potential risk markers for CV complications that were evaluated in this study is listed in Table S1, SDC (http://links.lww.com/TP/B551). Standard immunosuppression at our center during the study time period (2002-2012) included induction with steroids alone followed by early CNI initiation. For patients with renal injury, mycophenolate mofetil was added at the discretion of the treating physician with the goal of reducing target CNI levels as a renal protective strategy.

Study Outcomes

The primary study endpoint was a major CV complication occurring within 1 year of LT within the local cohort. The 1 year timeframe was chosen as this focuses on those complications most likely to affect 1-year patient and graft survival, which is an important quality metric in orthotopic liver transplant.24 Cardiovascular complications were defined as death from a CV cause or hospitalization for a major CV event as defined by a discharge International Classification of Diseases Ninth Revision or common procedural terminology code within the first 3 diagnosis positions that indicated myocardial infarction, revascularization (eg, percutaneous coronary intervention or coronary artery bypass grafting), heart failure, atrial fibrillation, cardiac arrest, pulmonary embolism, transient ischemic attack, and/or stroke during the initial transplant admission, or subsequent hospitalization (Table S2, SDC, http://links.lww.com/TP/B551).15–17 The discharge diagnosis was used to capture in-hospital and rehospitalization events. Documentation of a major CV complication was based on the first documented event in the medical record after LT. Cardiovascular complications within a year of LT are not available in the national cohort as OPTN does not capture CV outcomes and Vizient only reliably captures hospitalizations within 90 days of the index hospital stay.16 Thus, the secondary endpoints evaluated in the national cohort were restricted to all-cause and CV-related mortality. Mortality was determined from the OPTN database which is matched to the social security death master file located within the Social Security Death Index. Recipient cause of death was determined by a physician’s review (L.B.V.) of primary underlying and contributory causes of death (including all free text inputs) listed in the OPTN database. Any potential case with death due to CV disease, defined as primary cause of death from arrhythmia, heart failure, myocardial infarction, primary cardiac arrest, thromboembolism, and/or stroke, was then reviewed by an independent panel of 3 physicians (2 cardiologists: D.L.J., J.W.; 1 surgeon, A.S.).15 Each patient is represented only once in the cohort. Patients were censored at time of death, date of last follow-up, or time of retransplantation.

Statistical Analysis

Definition of Baseline eGFR

Renal function may rapidly fluctuate during the perioperative period due to multiple factors, including alterations in effective circulating volume and introduction of CNIs. Thus, in the local cohort, we used unadjusted Cox proportional hazard models to test the hypothesis that mean eGFR recorded between 7 and 30 days posttransplant would be more predictive of post-LT outcomes than a single measure of eGFR at the time of transplant. Based on these analyses, baseline eGFR was defined as the mean eGFR recorded between 7 and 30 days posttransplant for primary analyses in the local cohort. For secondary analyses in the national cohort, baseline eGFR was defined as the eGFR at transplant because OPTN only captures eGFR at transplant, 6 months and 1-year post-LT. Transplant eGFR and mean eGFR between 7 and 30 days posttransplant were mildly correlated (Pearson r = 0.438; P < 0.0001).

Local Cohort eGFR Trajectory Analysis

Identification of distinct subpopulations that have unique profiles over time can be accomplished using a contemporary statistical technique called group-based trajectory modeling. Group-based trajectory modeling is a statistical approach designed to identify clusters of individuals after a similar progression of some behavior (in this case, eGFR) over time. The methodology assumes that the population is composed of a finite number of distinct groups that can be defined by their trajectories.25 Trajectories in eGFR were modeled among all patients (n = 671) with 3 or more repeated values of eGFR within the first year of LT. We used latent class models to identify subgroups that share a similar underlying trajectory in eGFR.26 These models were fit using SAS Proc Traj.27 The optimal number of trajectory classes was determined using the Bayesian information criterion such that no group included less than 3.5% of the population. As in other studies, to account for the uncertainty in eGFR trajectory group assignment, we calculated the posterior predicted probability for each individual of being a member in each of the classes.27 Participants were assigned to the trajectory group for which they had the greatest posterior predictive probability. Baseline characteristics and unadjusted frequency of outcomes were compared between the trajectories with the use of analysis of variance and chi-squared test (or Fisher exact test, as appropriate) for continuous and categorical variables, respectively. To estimate the cross-sectional association of trajectory group with 1-year major CV complications, trajectory group membership was included as an independent variable in a logistic regression model examining predictors of 1-year major CV complications. Covariates in multivariable models were chosen a priori based on prior study of predictors of 1-year CV complications in LT recipients17 and included baseline recipient age, race, sex, education, working status at time of transplant, pretransplant comorbidities (smoking, diabetes, hypertension, renal disease, complications of cirrhosis), pretransplant CV disease (eg, myocardial infarction, atrial fibrillation, heart failure, cardiac arrest and/or stroke), degree of liver dysfunction at time of transplant (bilirubin and international normalized ratio [INR]), and baseline eGFR. We also performed a priori sensitivity analyses excluding patients who had received any form of renal replacement therapy (RRT) within 1 year after LT (n = 60).

National Cohort Mortality Analysis

Kaplan-Meier estimates, stratified according to the eGFR categories, for death from any cause and CV-specific death are presented as event curves. Cox proportional hazards models were used to examine the association of either continuous eGFR or eGFR category with all-cause and CV-specific mortality. Covariates in multivariable models were chosen a priori based on prior studies of predictors of CV-specific mortality in LT recipients15–17 and included baseline recipient age, race, sex, transplant center, education, complications of end-stage liver disease present at transplant, degree of pretransplant liver dysfunction (bilirubin, INR) and pretransplant CV disease, diabetes, obesity, renal disease, chronic pulmonary disease, and pulmonary hypertension status. We then tested for interactions of eGFR by race and by sex in all models. Sensitivity analyses were performed excluding patients who died of a cardiac arrest after LT (n = 519).

SAS software version 9.4 and R version 3.3.2 were used to complete all statistical analyses. R package ggplot2 was used to generate all figures.28 All P values are 2-sided, and a P value of less than 0.05 is considered to indicate statistical significance.

RESULTS

Local Cohort Trajectories of eGFR Within the First Year After LT

Six distinct trajectories of eGFR during the first year posttransplant were identified (Figure 1): 3.9% (n = 26) of patients started at a normal eGFR and experienced a slow decrease in eGFR throughout follow-up (referred to as the “Normal-Slow Decrease” group), 3.7% (n = 25) started at a normal eGFR but experienced a rapid decrease in eGFR starting around 6 weeks posttransplant (normal-rapid decrease), 17.6% (n = 118) of patients experienced mildly diminished, but stable eGFR levels (mild stable), 35.0% (n = 235) started at mildly diminished eGFR levels and experienced a slow decrease in eGFR over time (mild-slow decrease), 30.4% (n = 204) of patients experienced moderately diminished, but stable eGFR levels (moderate-stable), and 9.4% (n = 63) had severely diminished but stable egfr levels (severe-stable).

FIGURE 1.

eGFR trajectories by months from liver transplant in the NMEDW. Total, N = 671.

Patients in the severe-stable eGFR group were more likely to be male, older, have received a transplant for NASH and have hypertension, diabetes, obesity, and a preexisting diagnosis of CKD (Table 1). Patients in the normal-rapid decrease group were predominantly male, non-Hispanic white and had received a transplant for hepatitis C with an average MELD score of 23.8 (9.7). These patients had a high prevalence of pretransplant hyponatremia (20.0%) and spontaneous bacterial peritonitis (42.9%), though only 12.0% had documented hepatorenal syndrome and only 8.0% had received some form of RRT before LT. Of note, there were no differences in the burden of pretransplant CV disease comorbidity between groups.

TABLE 1.

Characteristics among liver transplant recipients stratified by eGFR trajectory group, OPTN/NMEDW data, 2002-2012

eGFR Trajectories and 1-year CV Complications

The prevalence of a major CV complication within 1 year of LT varied from 19.6% (n = 46) in the mild-stable eGFR trajectory group up to 56.0% (n = 14) in the normal-rapid decrease eGFR group (Figure 2, Table 2). Even among groups that started at a similar eGFR level, patients who experienced steeper declines in eGFR had a higher occurrence of CV morbidity, 56.0%(n = 14) versus 26.9%(n = 7), respectively (Figure 2, P <0.0001). In comparison to individuals in the mild-stable group, those in trajectory groups with patterns of lower eGFR had increasingly greater odds of having a major CV complication (Table 3, P <0.0001). Notably, the greatest odds of CV morbidity were seen in the normal-rapid decrease group (odds ratio [OR], 7.07; 95 % confidence interval [95% CI], 2.77-18.02). Associations were moderately attenuated when adjusted for demographics and pretransplant comorbidities (Table 3). Associations in the reduced eGFR groups were more significantly attenuated when adjusted for baseline eGFR; however, associations remained significant for patients in the normal-rapid and normal-slow decrease groups even when adjusted for baseline eGFR (OR, 10.55; 95% CI, 3.02-36.85 and OR, 4.90; 95% CI, 1.42-16.78, respectively). In sensitivity analyses, the eGFR trajectory patterns were similar, and their association with CV morbidity was essentially unchanged when individuals who were recorded as having received any form of RRT within 1 year of transplant were excluded from analysis (n = 60).

FIGURE 2.

All-cause mortality and 1-year cardiovascular morbidity by eGFR trajectory group, OPTN/NMEDW Data, 2002-2012. Pairwise comparison with the mild stable group as the reference; P values are denoted as * 0.05-0.01, ** 0.01-0.001, *** 0.001-0.0001, **** <0.0001.

TABLE 2.

Major CVD complications within 1 year of LT stratified by eGFR trajectory group, NMEDW, 2002-2011

TABLE 3.

ORs for 1 y major CV complications by eGFR trajectory group, OPTN/NMEDW data, 2002-2012

National Cohort Characteristics

During the period of the study, we identified 37 322 first liver-alone recipients within the Vizient database with an OPTN LT record between February 1, 2002, and December 31, 2011. Clinical characteristics stratified by baseline eGFR category are shown in Table 4. At LT, 28.3% of patients had normal/high eGFR (≥90 mL/min per 1.73 m²), 30.0% mildly decreased (eGFR, 60-89), 14.2% mildly to moderately decreased (eGFR, 45-59.9), 12.6% moderately to severely decreased (eGFR, 30-44.9), 10.6% severely decreased (eGFR, 15-29.9), and 4.2% had kidney failure (eGFR, < 15). Median pretransplant laboratory MELD score in the sample population was 19.5 (11). As expected, laboratory MELD score was inversely related to eGFR category.

TABLE 4.

Characteristics among 37,322 first liver alone recipients stratified by baseline eGFR category at the time of LT, OPTN/UHC data, 2002-2012

Relationship Between eGFR Category at Transplant and Post-Liver Transplant Mortality

In unadjusted analyses, lower eGFR category at the time of LT was associated with decreased all-cause and CV-specific survival (Figure 3). Each 5-unit lower eGFR at transplant was associated with a 2% higher hazard of all-cause mortality and a 5% higher hazard of CV-specific mortality (P = <0.0001, Table 5). In multivariable analyses adjusted for recipient age, race, sex, transplant center, education, severity of end-stage liver disease at transplant and pretransplant CV disease, diabetes, obesity, renal disease, chronic pulmonary disease, and pulmonary hypertension status, lower eGFR category at transplant was associated with increased all-cause and CV-specific mortality in a dose-dependent manner (P < 0.0001, Table 5). Compared with patients with normal/high eGFR (≥90) at the time of transplant, even recipients with mildly decreased eGFR (60-74.4) had higher hazard of death after LT (HR, 1.11; 95% CI, 1.03-1.20). There was no significant interaction by race or sex for either all-cause or CV-specific mortality in the fully adjusted models. Associations were unchanged in sensitivity analyses excluding patients who died of cardiac arrest, which may not always be considered specifically CV-related (n = 519, data not shown). Because creatinine is part of both eGFR and MELD calculations, we assessed the relationship between lab MELD at transplant and all-cause and CV-related mortality. Laboratory MELD (per 5 unit increase) was not associated with all-cause (HR, 0.99; 95% CI, 0.88-1.11) or cardiovascular disease-specific (HR, 1.04; 95% CI, 0.89-1.34) mortality.

FIGURE 3.

Unadjusted (A) all-cause and (B) CV-specific mortality by baseline eGFR at the time of LT, OPTN/Vizient data, 2002-2012. Color key used is the rainbow to show the trend by baseline eGFR category, from red representing the 0 to 14.9 baseline eGFR category to violet representing the 90+ baseline eGFR category.

TABLE 5.

Hazard ratios for all-cause and CV-specific mortality by baseline^a eGFR or eGFR category at the time of LT, OPTN/UHC data, 2002-2012

DISCUSSION

In a large single-center cohort, we have demonstrated that there are distinct trajectories of eGFR associated with differences in early major CV complications after LT. Membership in a trajectory group with more rapidly decreasing eGFR within the first year, even if normal at the time of LT, represented the most significant independent predictor of CV complications within that year. Concurrently, in the national cohort, lower eGFR category at the time of LT was associated with higher long-term all-cause and CV-specific mortality in a dose-dependent manner. As in our single-center data, even categories with less advanced CKD (stage 2) were associated with a higher risk of all-cause and CV-specific mortality independent of pretransplant comorbidities.9 These corroborated observations should now be tested in clinically relevant trials of proactive interventions early after LT to not only preserve renal function but also decrease CV risk.

Although the pathogenesis of CKD after LT is multifactorial, therapeutic strategies aimed at minimizing renal injury have focused on reducing the exposure of patients to nephrotoxic CNIs. The results of these studies can be summarized along 3 main themes: (1) Immediate CNI minimization: renal function may benefit from delayed CNI dosing29 but this can be transient,30–32 and antibody induction therapy often used may lead to risks inherent to more global immunosuppression;33 (2) late (ie, >6 months) CNI minimization or withdrawal: not proven to adequately preserve renal function presumably because CKD is often already irreversible;34–43 (3) delayed (ie, 1-6 months) early CNI minimization: a recent study demonstrated significant renal benefit post-LT when tacrolimus therapy was minimized in combination with a mammalian target of rapamycin inhibitor (mTOR-I), everolimus, starting at 1-month post-LT.18,19 Importantly, the vast majority of patients in this trial had GFR greater than 60. However, most of the current research and clinical interest has involved liver recipients with GFR less than 60 where clinicians routinely direct management strategies (eg, CNI reduction) to improve GFR.2,6 In fact, most LT clinicians are satisfied if their patients have GFR of 60 or greater or CKD stage 2. This perception that stage 2 CKD is acceptable is not based on available clinical data or outcomes.

Previous studies have demonstrated that pretransplant renal dysfunction is associated with increased posttransplant all-cause and CV mortality, even when controlled for degree of pretransplant liver dysfunction.44,45 In the current study, we extend these findings by demonstrating that ongoing post-LT renal dysfunction is independently associated with fatal and nonfatal adverse CV outcomes, even among patients with stage 2 CKD and even when controlled for pretransplant renal dysfunction. In particular, patients with minimal or no CKD that have a more rapid trajectory of decline in GFR early after LT are particularly at risk for CV events. These findings highlight a particularly vulnerable population who may especially benefit from early GFR preservation strategies. Notably, these patients had a fairly similar pretransplant CV risk profile to those in trajectory groups with diminished eGFR. Future studies are needed to determine whether early identification and intervention in this high-risk population alters eGFR trajectories and improves CV outcomes after LT.

There are multiple reasons by which all stages of renal dysfunction increase CV risk, now shown in a LT population. First, factors associated with renal decline, including anemia, oxidative stress, derangements in calcium-phosphate homeostasis, elevated levels of fibroblast growth factor 23,46 inflammation, and conditions promoting coagulation, are all associated with accelerated atherosclerosis and endothelial dysfunction.47 Other markers that progressively increase with renal decline include albuminuria, proteinuria, homocysteinemia, and hyperuricemia, all of which are prevalent among patients with cirrhosis.11,47 Synergism exists when conventional CV risk factors, such as diabetes and hypertension, perpetuate renal disease and progressive renal decline increases the potency of such risk factors.7 The mechanisms by which rapidly changing eGFR, even when starting at normal levels post-LT, require further study but may be due to more accelerated pathogenic pathways or fluid retention from renal impairment, all exacerbating ischemic and nonischemic CV events.

The findings from our study also highlight the fact that, within a LT population, there are heterogenous patterns in eGFR trajectories. Latent class modeling, as used in these current analyses, identifies different patterns of eGFR change as separate trajectory groups, thus providing a more realistic assessment of cumulative risk compared with single time point or mean levels. Latent class modeling is novel with important potential applications in transplantation. For one, in transplant, we have multiple repeated data points that are routinely collected over time. Thus, through current electronic medical record graphics, we can easily plot and analyze change in risk markers over time to help us better understand dynamic changes. It is also important to note that risk relationships are not always linear, and latent class modeling allows for observation of nonlinear relationships. Thus, the statistical methods presented here provide insight into the LT population heterogeneity as compared with more traditional methods, which only consider risk factors at the time of disease detection.

Strengths of the present study include its large sample size from independent patient cohorts, rigorous assessment of multiple recipient and donor-related factors including harmonized data from several sources, and the use of novel statistical methods to evaluate the influence of eGFR trajectories. There are limitations that warrant mention. First, we estimated GFR using the MDRD-4 formula, which may be inaccurate in patients with cirrhosis due to protein-calorie malnutrition.48 This can lead to either an overestimate or underestimate of true GFR in patients receiving LT.49 However, in a meta-analysis of organ recipients (35% liver), the MDRD-4 equation, although imperfect, was the most accurate and has the advantage of serial assessments (trajectories) compared with measured GFRs (iohexol, iothalamate) that are costly and not feasible in practice.50 Second, we did not measure the rates of urinary albumin or protein excretion, factors that may drive the independent effect of eGFR on CV outcomes. Other renal-specific factors that were not included in our model may also affect the influence of GFR on CV outcomes. We also cannot account for clinical changes in practice, including alterations in immunosuppression, which may account for some of the drivers of renal change in these populations. We also emphasize that the relationship between eGFR trajectory and CV complications is cross-sectional and thus, causation cannot be determined. Baseline eGFR also differed between the 2 cohorts due to the limitations of the data structure within the national cohort and thus, we cannot reliably differentiate between the effects of acute kidney injury in the immediate perioperative period versus ongoing chronic injury on post-LT mortality. Finally, there is potential for reverse causation, particularly in the normal-rapid decrease eGFR group, wherein a patient may have had a high mortality associated event (eg, sepsis) that then resulted in a rapid decline in renal function immediately before death.

In conclusion, pattern of change in eGFR within the first year of LT may provide additional information about the risk for adverse CV outcomes. We have also demonstrated that all-cause and CV-specific mortality is increased even among patients with milder forms of renal injury after LT. These findings represent a unique opportunity for early identification and treatment of renal dysfunction, even when mild. Additional research is needed to examine the utility of specific eGFR trajectories in other populations of LT recipients, the mechanisms underlying why rapid GFR change associates more highly with adverse CV outcomes, and the timing and impact of early, proactive interventions on these clinical outcomes.