Six-Minute Walk Distance Predicts Outcomes in Liver Transplant Candidates

Abstract

The 6-minute walk test (6MWT) is a simple tool for assessing submaximal exercise capacity. We sought to determine whether 6-minute walk distance (6MWD) predicts outcomes in patients with cirrhosis. The Pulmonary Vascular Complications of Liver Disease 2 (PVCLD2) study is a multicenter, prospective cohort study that enrolled adults with portal hypertension during liver transplantation evaluation. We excluded subjects with incident or prevalent portopulmonary hypertension. 6MWT was performed using standardized methods. Cox proportional hazards modeling and multivariable linear regression analysis were performed to determine the relationship between baseline 6MWD and outcomes. The study sample included 352 subjects. The mean 6MWD was 391 meters ± 101 meters. For each 50-meter decrease in 6MWD there was a 25% increase in the risk of death (hazard ratio 1.25, 95% CI [1.11,1.41], p<0.001) after adjustment for age, gender, body mass index, MELD-Na, and liver transplant as a time-varying covariate. In a multistate model, each 50-meter decrease in 6MWD was associated with an increased risk of death prior to liver transplant (p<0.001) but not after transplant. 6MWD was similar to MELD-Na in discriminating mortality. Each 50-meter decrease in 6MWD was associated with an increase in all-cause (p<0.001) and transplant-free hospitalizations (p<0.001) in multivariable models for time-to-recurrent events. Shorter 6MWD was associated with worse Short Form-36 physical (p<0.001) and mental component scores (p=0.05). In conclusion, shorter 6MWD is associated with increased risk of death, hospitalizations, and worse quality of life in patients evaluated for liver transplantation. 6MWD may be a useful adjunct for risk-assessment in patients undergoing liver transplant evaluation.

Hepatic cirrhosis is a highly morbid condition, responsible for 1.3 million deaths annually.¹ The only definitive treatment is liver transplantation. Liver transplant centers primarily determine organ allocation by utilizing the Model for End-Stage Liver Disease (MELD-Na) score. The use of MELD-Na has been associated with a lower risk of death for individuals on transplant wait lists but pre-transplant mortality remains high.²

The 6-minute walk test (6MWT) is a simple, cost-effective submaximal test that is routinely used to assess exercise capacity in several heart and lung conditions.³ Both absolute and change in 6-minute walk distance (6MWD) predict mortality in patients with a variety of cardiopulmonary diseases, including those awaiting lung transplant.⁴ The American Society of Transplantation suggests that 6MWT be used to stage frailty in liver transplant patients primarily on the basis of two single-center cohort studies (one retrospective and one prospective, both with 6MWT performed in a standardized manner using published guidelines) which demonstrated a significant relationship between lower 6MWD and worse clinical outcomes.^5–7 These studies have not assessed the relationship between 6MWD and the risk of hospitalizations or compared performance to MELD-Na. A more comprehensive understanding of the relationship between 6MWD and outcomes is essential to galvanize the standardized inclusion of 6MWT into routine evaluation of patients with end-stage liver disease (ESLD) and to justify interventions such as structured rehabilitation.

In this study, we aimed to assess the relationship between 6MWD and mortality, hospitalizations, and quality of life in the Pulmonary Vascular Complications of Liver Disease (PVCLD2) study. We hypothesized that shorter 6MWD would be associated with increased pre-transplant mortality, increased hospitalizations, and decreased quality of life as assessed by the Short Form-36. We also compared the performance of 6MWD to MELD-Na in predicting mortality and explored the relationship between levels of circulating angiogenic biomarkers and 6MWD to gain insight into potential mechanisms of alterations in cardiopulmonary reserve in this population.

Methods

Sample Cohort and Study Sample

The PVCLD2 study was a multicenter, prospective cohort study that enrolled patients being evaluated for liver transplantation at the University of Pennsylvania, Mayo Clinic and the University of Texas at Houston from 2013–2017. Patients evaluated for liver transplantation during the enrollment period with portal hypertension with or without intrinsic liver disease who did not meet exclusion criteria were approached for enrollment. Patients with active infection, recent (less than two weeks) gastrointestinal bleeding, or those who underwent prior liver or lung transplant were excluded. Patients provided written informed consent. All research was conducted in accordance with both the Declarations of Helsinki and Istanbul. No donor organs were obtained from executed prisoners or other institutionalized persons at the transplant centers involved in this study. The Institutional Review Board at the respective centers approved the study (University of Pennsylvania, Office of Regulatory Affairs Protocol number 816361; Mayo Clinic institutional review board protocol 12-007715; and University of Texas Committee for the Protection of Human Subjects protocol HSC-MS-12-0481).

The study sample included subjects who underwent initial liver transplant evaluation and 6MWT, which was performed as part of the study protocol during the baseline visit. We excluded subjects with missing 6MWT data and those with portopulmonary hypertension.

Clinical and Laboratory Sampling

Data collected included history and exam from the day of study procedures and information from the medical record. Phlebotomy was performed after overnight fasting with the exception of water and both serum and plasma were banked at −80°C. Plasma angiopoietin-2 was measured using the Meso Scale Discovery Human Angiopoietin-2 Kit (#K151KCD, Meso Scale Diagnostics, Rockville MD). Plasma vascular cell adhesion molecule-1 (VECAM-1) (#DVC00) was quantified using ELISA (R&D Systems, Minneapolis, MN). von Willebrand Factor (vWF) antigen was quantified using immunoturbidimetrics from Stago (Cat #00518; Parsippany, NJ). Additional measured biomarkers included serum angiostatin, soluble c-Kit, soluble E-selectin, soluble epithelial growth factor receptor, tenascin-C, soluble Tie-2, soluble VEGF receptors-1, -2, and -3, platelet derived growth factor AB/BB, platelet endothelial cell adhesion molecule-1, plasma fractalkine, endostatin and VEGF-A. MELD-Na score was derived from laboratory values measured as part of the study protocol and calculated using the following formulas: MELD=10*((0.957*ln(creatinine))+(0.378*ln(Bilirubin))+(1.12*ln(INR)))+6.43 and MELD-Na=MELD+(1.32*(137-Na))−(0.033*MELD*(137-Na)).^8,9

All 6MWTs were performed using 2002 American Thoracic Society standards.³ The test was administered by trained technicians in the same corridor for all participants at each site, without a “warm-up” test. Study staff trained on-site personnel to conduct the test without encouragement or accompaniment. Training included a lecture and demonstration of the test and each technician received certification after performing an acceptable test on a mock subject. All subjects completed the Short Form-36 (SF-36) during their baseline study visit. All study visits and procedures occurred on an outpatient basis. The research team contacted subjects every six months until the end of the study in 2017. Hospitalization dates, liver transplantation, and death were obtained from patients, the medical record and the subjects’ physicians. Subjects alive at the end of follow-up were censored on May 25, 2017.

Statistical Analysis

We summarized data using the mean (standard deviation), median (interquartile range) or counts (proportions) as appropriate. Cox proportional hazards models were used to assess the association between baseline 6MWD and death. Sensitivity analyses considered liver transplant as a time-varying covariate, as a competing risk, and in a multi-state model. All models were adjusted for age, gender, body mass index (BMI), and MELD-Na based on biologic plausibility. A Kaplan-Meier plot was used to summarize mortality by baseline 6MWD quartiles. Discrimination for transplant-free mortality was measured using the C-index and area under time-dependent receiver operating characteristic (ROC) curves. Hospitalizations were summarized using a mean cumulative function plot and modeled using relevant approaches for recurrent events.¹⁰ Bivariate and multivariable linear regression models adjusted for age, gender, body mass index, and MELD-Na score were used to assess the association between 6MWD and SF-36 physical component (PCS) and mental component (MCS) scores and circulating angiogenesis biomarkers, respectively. Additional analyses included tests for interaction by liver disease etiology, the addition of liver disease etiology as a covariate to our multivariable models, exclusion of subjects with hepatopulmonary syndrome (HPS) and hepatocellular carcinoma (HCC), and testing for evidence of effect modification between HPS and HCC respectively and angiogenesis markers and the relationship with 6MWD. Alpha was set at 0.05. All analyses were conducted with R Software (version 3.6.1).

Results

A total of 454 subjects were enrolled in PVCLD2. Forty-three (9%) were excluded for prevalent or incident portopulmonary hypertension and 59 (13%) were excluded for missing 6MWD data. The final study sample included 352 subjects (Figure 1). The mean age was 57 years ± 9 years, 66% were men, and 16% were current smokers. The most common etiologies of cirrhosis were alcohol consumption (n=140, 40%) and hepatitis C infection (n=140, 40%) (Table 1). Seventy-three (37%) of subjects with data from necessary testing at baseline had HPS. The mean 6MWD at baseline was 391 meters ± 101 meters. A total of 50 (14%) subjects were lost to follow-up before the end of study follow-up, 132 (38%) underwent liver transplant, and 75 (21%) died during the study period (Table E1).

Figure 1.

Selection of study sample.

Table 1.

Demographics, liver disease characteristics, and past medical history

Variable	N	Overall
Age (years), mean ± SD	352	57 ± 9
Female gender, n (%)	352	118 (34)
Body mass index (kg/m2), mean ± SD	351	30 ± 7
Race/Ethnicity, n (%)	352
Non-Hispanic white		257 (73)
Hispanic white		52 (15)
Non-Hispanic black		34 (10)
Other		9 (3)
Born in the United States/Puerto Rico, n (%)	351	331 (94)
Language spoken in the household, n (%)	351
English		334 (95)
Spanish		3 (1)
Other		14 (4)
Education, n (%)	351
No schooling or Grades 1-11		34 (10)
High school or GED		112 (32)
Some college education or technical/vocational certificate		81 (23)
Associate or Bachelor’s degree		96 (27)
Professional or Graduate degree		28 (8)
Family income for past 12 months, n (%)	351
$19,999 and below		90 (26)
$20,000 – $49,999		91 (26)
$50,000 – $99,999		66 (19)
$100,000 and above		66 (19)
Unknown		38 (11)
Etiology of liver disease, n (%)
Alcohol	352	140 (40)
Hepatitis C infection	352	140 (40)
Autoimmune hepatitis	352	15 (4)
Non-alcoholic fatty liver disease	352	78 (22)
Hepatitis B infection	352	7 (2)
Primary sclerosing cholangitis	352	16 (5)
Primary biliary cirrhosis	352	25 (7)
Cryptogenic cirrhosis	352	16 (5)
Other	352	17 (5)
History of liver disease complications, n (%)
Ascites	352	242 (69)
Varices	352	251 (71)
Variceal bleeding	352	118 (34)
Encephalopathy	352	202 (57)
Multiple paracenteses	352	110 (31)
Spontaneous bacterial peritonitis	352	22 (6)
Hepatocellular carcinoma	352	106 (30)
Hepatic hydrothorax	352	46 (13)
Transjugular intrahepatic porto-systemic shunt	352	31 (9)
Past medical history, n (%)
Chronic obstructive pulmonary disease	352	34 (10)
Chronic bronchitis	352	26 (7)
Asthma	352	34 (10)
Venous thromboembolism	352	14 (4)
Diabetes mellitus	352	123 (35)
Hypertension	352	171 (49)
Hypercholesterolemia	352	75 (21)
Congestive heart failure	352	19 (5)
Hepatopulmonary syndrome	197*	73 (37)
Smoked at least 100 cigarettes in lifetime, n (%)	351	218 (62)
Pack-years for ever-smokers, median [IQR]	162	13 [5, 31]
Smoked in the last 30 days, n (%)	352	55 (16)
Consumed alcohol, n (%)	351	323 (92)
Duration of alcohol consumption (years), median [IQR]	323	30 [20, 40]
Current alcohol use, n (%)	352	19 (5)
Medications, n (%)
Beta-blockers	350	183 (52)
SBP prophylaxis/Antibiotics	350	148 (42)
Bile acid resins	350	39 (11)
MELD-Na score, median [IQR]	352	14.0 [10, 18]
Functional class, n (%)	351
I		122 (35)
II		153 (44)
III		75 (21)
IV		1 (0)
Six-minute walk distance (meters), mean ± SD	352	391 ± 101

GED: General Educational Development;

^*Individuals with data from necessary testing for HPS at baseline visit

Shorter 6MWD was associated with higher mortality (Table 2, Figure 2). For each 50-meter decrement in 6MWD, there was a 29% increase in the risk of death (unadjusted hazard ratio [HR] 1.29, 95% confidence interval [CI] [1.16, 1.44], p<0.001). This was unchanged after adjustment for age, gender, BMI and MELD-Na score and when liver transplant was considered as a time-varying covariate (adjusted HR 1.25, 95% CI [1.11,1.41], p<0.001). Similarly, each 50-meter decrease in 6MWD was associated with a 27% increase in the risk of transplant-free death (adjusted HR 1.27, 95% CI [1.13, 1.44], p<0.001). Findings persisted when subjects with HPS were excluded (adjusted HR 1.34, 95% CI [1.08, 1.66], p=0.01). Sensitivity analyses adjusting for etiology of liver disease and excluding those with HCC did not significantly change these findings (Data not shown).

Table 2.

Models for six-minute walk distance and the risk of death

Models	HR (95% CI)	p value	1 aHR (95% CI)	p value
Overall mortality	1.29 (1.16, 1.44)	<0.001	1.30 (1.15, 1.46)	<0.001
Overall mortality with transplant as a time-varying covariate	1.30 (1.16, 1.44)	<0.001	1.25 (1.11, 1.41)	<0.001
Transplant-free mortality	1.36 (1.21, 1.52)	<0.001	1.27 (1.13, 1.44)	<0.001
2 Mortality with transplant as competing risk (Fine-Gray model)	1.36 (1.24, 1.49)	<0.001	1.35 (1.23, 1.49)	<0.001
3 Multistate model
Transition from study baseline to liver transplant	0.96 (0.88, 1.04)	0.38	0.90 (0.81, 0.98)	0.02
Transition from study baseline to death without liver transplant	1.36 (1.21, 1.52)	<0.001	1.27 (1.13, 1.44)	<0.001
Transition from liver transplant to death after transplant	0.90 (0.65, 1.26)	0.56	0.99 (0.67, 1.45)	0.95

¹Adjusted for age, gender, body mass index, MELD-Na score

²Subdistributional hazards ratio

³Schema in Figure E1

All hazard ratios are for a 50-meter decrease in 6MWD

Figure 2.

Kaplan Meier plot of survival by 6-minute walk distance quartiles. Tft:log rank test for trend. 1^st quartile: 79 meters to <335 meters, 2^nd quartile: 335 meters to <397 meters, 3^rd quartile 397 meters to <459 meters, 4^th quartile 459 meters to 687 meters.

In a multistate model (Figure E1), each 50-meter decrease in 6MWD was associated with an increased risk of death prior to or without liver transplant (adjusted HR 1.27, 95% CI [1.13, 1.44], p<0.001) but not after liver transplant (adjusted HR 0.99, 95% CI [0.67, 1.45], p=0.95) (Table 2). After adjusting for age, gender, BMI and MELD-Na score, a 50-meter decrease in 6MWD was associated with a 10% lower likelihood of having a liver transplant (adjusted HR 0.90, 95% CI [0.81, 0.98], p=0.02). In a sensitivity analysis excluding individuals with HPS, this finding persisted but was no longer statistically significant (adjusted HR 0.88, 95% CI [0.74, 1.03], p=0.12). The probability of survival for various 6MWD values at various time points is shown in Figure E2.

The C-indices of 6MWD and MELD-Na to discriminate transplant-free mortality were 0.69 and 0.66, respectively. Using the area under time-dependent ROC curves for 6MWD and MELD-Na to discriminate transplant-free mortality, values were similar between the two at the various time points up to 48 months (Figure 3). Adding 6MWD to a model with MELD-Na alone statistically improved the fit of the model (likelihood ratio test p<0.001, c-index=0.72), suggesting that mortality prediction from MELD-Na alone could be marginally improved by including 6MWD. There was no significant interaction between MELD-Na and 6MWD (p for interaction=0.42).

Figure 3.

Time-dependent ROC curves for transplant-free mortality. A. Time-dependent ROC curves at 12, 24, 36, and 48 months; B. Area under the ROC curve by time.

Each 50-meter decrease in 6MWD was associated with an 11% higher likelihood of all-cause hospitalizations (unadjusted HR 1.11, 95% CI [1.07, 1.16], p<0.001) (Table E3, Figure 4). Effect estimates remained largely unchanged after adjustment for age, gender, BMI and MELD-Na and when transplant was modeled as a time varying covariate (adjusted HR 1.09, 95% CI [1.04, 1.14], p<0.001). Results were unchanged in recurrent event models (adjusted HR 1.12, 95% CI [1.03, 1.22], p=0.02) (Table E3). Results were unchanged when subjects with HPS were excluded except for adjusted transplant free hospitalizations, which no longer was significant (adjusted HR 1.08, 95% CI [0.99, 1.19]). Hospitalization for liver transplantation was not included in these analyses.

Figure 4.

Mean cumulative function plot of hospitalizations by six-minute walk quartiles. 1^st quartile: 79 meters to <335 meters, 2^nd quartile: 335 meters to <397 meters, 3^rd quartile 397 meters to <459 meters, 4^th quartile 459 meters to 687 meters.

There was a linear relationship observed between 6MWD and quality of life as assessed by the SF-36. A shorter 6MWD was associated with a significantly lower PCS and MCS, reflecting worse health-related quality of life. For example, for each 50-meter decrease in 6MWD, after adjusting for age, gender, BMI and MELD-Na score, PCS decreased by 2.5 units (95% CI [−3.0, −2.0], p < 0.001) and MCS decreased by 0.6 units (95% CI [−1.3, −0.02], p = 0.045) (Table E4). The PCS result may reach the minimally important difference whereas the MCS does not. Pairwise correlations between SF-36 domains and 6MWD are shown in Figure 5.

Figure 5.

Correlation between six-minute walk distance and quality of life via Medical Outcome Study Short Form (SF-36) physical component score (PCS) and mental component score (MCS).

Finally, we examined whether a panel of angiogenic biomarkers in peripheral blood were associated with 6MWD. After adjusting for age, gender, BMI, and MELD-Na, higher angiopoietin-2, VCAM-1 and vWF collagen binding activity levels were associated with shorter 6MWD (for every 10 ng/mL increase in ang-2 or VCAM-1 or 10 IU/mL increase in vWF collagen binding activity, 6MWD decreased by −15.4 meters (95% CI [−23.5, −7.2], p <0.001); −0.2 meters (95% CI [−0.4, −0.1], p=0.01); −155.2 meters (95% CI [−301.9, −8.6], p=0.04), respectively) (Table E5, Figure 6). While there was evidence of significant pairwise correlations between additional biomarkers and 6MWD, we failed to detect significant associations with these markers using linear regression (Figure E3a and E3b). Both HPS and HCC have been independently linked to angiogenesis biomarkers,^11,12 however, additional analyses performed excluding subjects with HPS and HCC did not substantially change the findings. In the models excluding participants with HCC, vWF collagen binding activity was no longer significant but significant relationships emerged in the remaining cohort with VEGFR-3 and PDGF-AB/BB (data not shown).

Figure 6.

Correlation between angiogenic biomarkers angiopoietin-2 (ng/mL), VCAM-1 (ng/mL), vWF-Collagen Binding (IU/mL) and 6MWD six-minute walk distance.

Discussion

In this large prospective, multicenter observational cohort of ESLD patients undergoing evaluation for liver transplantation, we have demonstrated that 6MWD predicts clinically relevant outcomes and performs at least at well as (if not better than) MELD-Na. The association between 6MWD and outcomes was particularly strong in those who had not undergone liver transplantation where shorter 6MWD was associated with increased mortality and with decreased likelihood of receiving a liver transplant. Shorter 6MWD was also associated with increased hospitalizations, worse health-related quality of life, and increased circulating angiogenic biomarkers.

Two smaller single center studies (n = 121 and 213) have previously demonstrated a significant relationship between shorter 6MWD and worse survival.^6,13 A larger retrospective single center study including 694 individuals at the time of liver transplant evaluation found that reduced 6MWD was an independent predictor of mortality.¹⁴ A smaller prospective cohort study including 106 cirrhotic patients listed for liver transplant from two centers also demonstrated 6MWD was an independent predictor of mortality and inversely related to Child scores.¹⁵ In 250 subjects with variable stages of cirrhosis, shorter 6MWD was associated with worse survival outside of the transplant referral setting.¹⁶ To our knowledge, ours is the largest multicenter prospective cohort to date and the first study to include data on hospitalizations and angiogenic biomarkers as well as comparison to performance of MELD-Na.

The only effective treatment for ESLD is liver transplantation but, as a consequence of organ scarcity, transplant waitlist mortality remains high.¹⁷ The existing system for organ allocation incorporates geographic location and medical urgency defined primarily by MELD-Na as a surrogate marker for illness severity.¹⁸ MELD-Na has been widely accepted, as it is a relatively objective tool able to identify the “sickest first.” However, MELD-Na only incorporates laboratory results directly related to hepatic function and fails to capture any metric of cardiopulmonary reserve, despite their established associations with worse outcomes in patients with ESLD.^5,19–21 To date, tools used to assess frailty have not been incorporated into the prioritization of patients for liver transplant as many are considered too subjective.²²

The 6MWD is objective and reproducible with good test-retest reliability and inter-rater reliability.^23,24 It is inexpensive, quick to perform, and requires little interpretation. While cardiopulmonary exercise test may also be useful to capture cardiopulmonary reserve it is more burdensome and difficult to scale.²⁵ In addition to adequately predicting important patient centered outcomes and the simplicity with which it can be implemented, 6MWD is a potentially modifiable target as compared with risk assessment tools like MELD-Na. For example, structured rehabilitation is not a part of current standard liver transplant recommendations but may improve waitlist and peri-operative outcomes for liver transplantation. A recent large prospective cohort study of liver transplant candidates who received an individualized home-based exercise program demonstrated improvements in liver frailty index scoring, 6MWD, and survival.²⁶ Several small pilot studies that enrolled cirrhotic patients awaiting liver transplantation to structured exercise programs have demonstrated feasibility, safety and modest improvements in metrics of cardiopulmonary fitness.^27,28,29,30 Mobility and walking is an integral part of rehabilitation for cirrhotic patients and 6MWD results may have the added psychological benefit of providing actionable feedback to patients. Together with our results, these studies suggest that rehabilitation to improve 6MWD in this patient population may warrant further study in randomized clinical trials. Future work should also address the minimal clinically important difference in 6MWD with rehabilitation interventions.

While mortality is the ultimate endpoint, hospitalizations in patients with ESLD represent a significant burden on the healthcare system, costing an estimated $18.8 billion annually.³¹ Our finding that shorter 6MWD was associated with increased hospitalizations has important public health implications. Health-related quality of life is an important patient reported outcome especially in patients with chronic diseases. In a previous study, the PCS of the SF-36 was moderately correlated with 6MWD but not the MCS.⁶ In our cohort, there were significant (albeit moderate) correlations between quality of life as measured by both the PCS and MCS and shorter 6MWD, indicating quality of life may improve in this fragile patient population if we target functional status.

In an effort to elucidate potential mechanistic links between liver disease and cardiopulmonary reserve, we examined the relationship between circulating angiogenic biomarkers and 6MWD. Higher angiopoietin-2, VCAM-1 and vWF collagen binding activity were associated with shorter 6MWD. This suggests that dysregulation of angiogenic signaling and inflammation leads to either cardiopulmonary or systemic vascular remodeling, which may contribute to reductions in exercise tolerance. While liver sinusoidal endothelial cells are thought to be the source of angiopoietin-2 and it is speculated to be involved in hepatic regeneration³², higher angiopoietin-2 levels are also associated with worse non-alcoholic fatty liver disease in children and greater liver fibrosis stage in patients with chronic hepatitis C.^33,34 Higher angiopoietin-2 and vWF levels have been correlated with worse outcomes in patients with heart failure, acute myocardial infarction and pulmonary arterial hypertension, and in some studies track with therapy and changes in 6MWD.^35–39 It is conceivable that these elevated angiogenic biomarkers contribute to worsened cardiopulmonary reserve and subsequently lower 6MWD. We recently published that in the PVCLD2 cohort, subjects with HPS had significantly elevated levels of angiogenesis markers when compared with non-HPS subjects with advanced liver disease.¹¹ In the current study, there was minimal evidence of effect modification between circulating angiogenesis markers and HPS diagnosis and the impact on 6MWD, with the exception of c-KIT (data not shown). Angiogenesis has been implicated in the pathogenesis of HCC and skeletal muscle function during exercise including in heart failure patients.^12,40,41 However, excluding subjects with HCC from the current study did not substantially alter our findings. While it remains to be seen whether these angiogenic derangements are a driver or consequence of disease in ESLD patients, these findings are hypothesis generating and further work in this arena may identify novel therapeutic targets.

Our study has limitations. As we enrolled subjects from transplant centers and excluded a small number of subjects with missing 6MWD (n=59, 13%), our results are subject to selection bias. While our results persisted despite adjustment for covariates, confounding is still certainly possible. While predictive of important outcomes, 6MWD may not be a true surrogate for ESLD disease severity to the same degree as MELD-Na as it is not in the direct causal pathway and impacted by multiple organ systems. The 6MWT may be useful to further risk stratify patients awaiting liver transplantation and to identify patients who may benefit from early interventions like exercise rehabilitation. While addition of 6MWD to MELD-Na may marginally improve discrimination for survival, overall predictive ability remained moderate and we failed to identify an optimal 6MWD threshold, which could be due to cohort heterogeneity or the complex and dynamic nature of ESLD. Future studies should incorporate 6MWD in prospective candidate selection as the next step toward validation. Taken together, these results suggest the 6MWT provides complementary and valuable predictive information in ESLD patients and warrants further investigation.

In summary, 6MWD is a simple tool commonly used in clinical practice that could be leveraged to both predict clinical outcomes and provide a modifiable target in patients with ESLD awaiting liver transplantation.

Abbreviations:

6MWT

6-minute walk test

6MWD

6-minute walk distance

PVCLD2

The Pulmonary Vascular Complications of Liver Disease 2

MELD-Na

Model for End-Stage Liver Disease

ESLD

end-stage liver disease

BMI

body mass index

ROC

receiver operating characteristic

PCS

physical component score

MCS

mental component score

HPS

hepatopulmonary syndrome

HCC

hepatocellular carcinoma