LIVER TRANSPLANTATION 20:54–62, 2014
ORIGINAL ARTICLE
Aerobic Capacity During Cardiopulmonary Exercise Testing and Survival With and Without Liver Transplantation for Patients With Chronic Liver Disease s Lipcsey,1 Caroline Tallis,1 Kyne Woodsford,1 William Bernal,1 Rosa Martin-Mateos,1 Miklo 1 1 Mark J. Mcphail, Christopher Willars, Georg Auzinger,1 Elizabeth Sizer,1 Michael Heneghan,1 Simon Cottam,2 Nigel Heaton,3* and Julia Wendon1* 1 Liver Intensive Therapy Unit, 2Department of Anaesthetics, and 3Liver Transplant Surgical Service, Institute of Liver Studies, King’s College Hospital, King’s College London, London, United Kingdom
Chronic liver disease (CLD) is associated with muscle wasting, reduced exercise tolerance and aerobic capacity (AC). Measures of AC determined with cardiopulmonary exercise testing (CPET) may predict survival after liver transplantation (LT), but the relationship with nontransplant outcomes is uncertain. In patients assessed for LT, we examined the relationship of CPET AC parameters with the severity of liver disease, nutritional state, and survival with and without LT. Patients assessed for elective first LT who underwent CPET and an anthropometric assessment at a single center were studied. CPET-derived measures of AC that were evaluated included the peak oxygen consumption (VO2 peak) and the anaerobic threshold (AT). Three hundred ninety-nine patients underwent CPET, and 223 underwent LT; 45% of the patients had a VO2 peak < 50% of the predicted value, and 31% had an AT < 9 mL/kg/minute. The VO2 peak and AT values correlated with the Model for End-Stage Liver Disease score, but they more closely correlated with serum sodium and albumin levels. The handgrip strength correlated strongly with the VO2 peak. Patients with impaired AC had prolonged hospitalization after LT, and nonsurvivors had lower AT values than survivors 1 year after transplantation (P < 0.05); this was significant in a multivariate analysis. One hundred seventy-six patients did not undergo LT; the 1-year mortality rate was 34.6%. The AT (P < 0.05) and VO2 peak values (P < 0.001) were lower for nonsurvivors. In a multivariate analysis, AT was independently associated with nonsurvival. In conclusion, AC was markedly impaired in many patients with CLD. In patients who did not undergo transplantation, impaired AT was predictive of mortality, and in patients undergoing LT, it was related to postoperative hospitalization and survival. AC should be evaluated as a modifiable factor for improving patient survival whether or not C 2013 AASLD. LT is anticipated. Liver Transpl 20:54-62, 2014. V Received June 4, 2013; accepted September 30, 2013. Many patients with chronic liver disease (CLD) develop progressive impairment of their fitness and exercise capacity.1 The clinical significance of these
changes is reflected by the close association of altered musculature and exercise capacity with the risk for major disease-related complications and mortality.2-8
Abbreviations: AC, aerobic capacity; AT, anaerobic threshold; BMI, body mass index; BMIc, corrected body mass index; CLD, chronic liver disease; CPET, cardiopulmonary exercise testing; HGS, handgrip strength; HR, hazard ratio; LT, liver transplantation; MAC, midarm circumference; MAMC, midarm muscle circumference; MELD, Model for End-Stage Liver Disease; TSF, triceps skinfold; VO2 peak, peak oxygen consumption. The authors have no conflicts of interest or financial support to report. *These authors contributed equally to this work. Address reprint requests to William Bernal, M.D., Liver Intensive Therapy Unit, Institute of Liver Studies, King’s College Hospital, King’s College London, Denmark Hill, London, United Kingdom SE5 9RS. Telephone: 44 203 299 4488; FAX: 44 203 299 3367; E-mail: [email protected] DOI 10.1002/lt.23766 View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
C 2013 American Association for the Study of Liver Diseases. V
LIVER TRANSPLANTATION, Vol. 20, No. 1, 2014
Cardiopulmonary exercise testing (CPET) represents the gold standard for the determination of exercise capacity and cardiorespiratory reserve, and it generates measures of a number of aspects of aerobic capacity (AC), which is the ability of the body to consume and use oxygen during exercise.8,9 The determined measures include the peak oxygen consumption (VO2 peak) achieved during testing and the point at which an apparent shift occurs from aerobic to anaerobic metabolism [ie, the anaerobic threshold (AT)]. The former is a measure related to the peak workload achieved and is dependent on volition, whereas the latter is a submaximal measure that is independent of a patient’s volitional effort.1,8,9 Together, these broadly indicate the ability of the cardiopulmonary system to deliver oxygen to the peripheral tissues and the ability of the tissues to use that oxygen. Studies have shown that both of these measures are significantly lower in patients with CLD in comparison with both predicted values and healthy controls, and they are related to survival during wait listing for transplantation and after surgery.3,5,8,10,11 These findings are potentially important because AC can be improved though physical training interventions. In patients with other chronic diseases in whom AC is similarly impaired, such interventions clearly improve the quality of life, reduce resource use, and may prolong survival.12-14 These therapies have yet to be tested in patients with CLD and could represent an important but untapped clinical resource. However, although CPET testing is apparently safe and effective in patients with CLD, it is expensive and is not uniformly available. Furthermore, studies reporting on its use in such patients have been mostly small and focused on the relationship between preoperatively determined measures of AC and short-term posttransplant outcomes. In these studies, the relationship between CPET measures of AC and standard measures of CLD severity has not been consistent.1 Muscle bulk has also been recently shown to be related closely to survival for patients with CLD; other more straightforward measures of muscle size and function might usefully identify patients with CLD and impaired AC who might benefit from interventions or aid in decision making concerning suitability for transplantation.2,15-17 Therefore, in a large cohort of patients from a single center, we examined the relationship between findings from CPET performed during the evaluation for potential liver transplantation (LT) and patients’ clinical features. Our specific objectives were (1) to clarify the relationship between parameters of AC determined with CPET and those assessing the severity of liver disease and anthropomorphic measures of muscle bulk, function, and overall nutritional state and (2) to confirm the previously demonstrated relationship between CPET measures of AC and post-LT survival, with the latter extended from short-term survival to longer term survival after LT. Finally, we wanted to investigate the relationship between AC and survival
BERNAL ET AL. 55
in patients who did not undergo transplantation. In doing so, we hoped to identify the clinical features that characterize the patients at greatest risk for impaired AC and the patient subgroup most likely to benefit from future interventions.
PATIENTS AND METHODS This retrospective, observational cohort study examined 399 consecutive patients undergoing a protocolized LT assessment at the Institute of Liver Studies (King’s College Hospital, London, United Kingdom) between February 2007 and June 2011. Only patients with cirrhotic liver disease who were assessed for elective first LT were studied, and the exclusion criteria included a patient age less than 16 years and standard exclusions for performing CPET.18 The decision to accept a patient for LT was made by a multidisciplinary team of clinicians and was not based on CPET performance or a nutritional assessment alone.
CPET Patients underwent CPET at the time of the LT assessment for elective LT. Any b-adrenoceptorblocking medication was discontinued at least 12 hours before the test, and ascites was drained by paracentesis. CPET was conducted upright with a cycle ergometer (Ergoline, Bitz, Germany), an ergospirometer, and a continuous 12-lead electrocardiograph (Ultima CardiO2, MGC Diagnostics, St. Paul, MN, or Schiller AG, Baar, Switzerland) by a physiology team that was dedicated to performing the test and was independent of the LT assessment team. The equipment was calibrated daily. CPET was executed according to a fixed protocol with respect to the monitoring, startup, test procedure, and termination. In brief, a ramp protocol was used with the load increasing in 5- to 10-W increments, and it was terminated when maximal exertion was reached or when safety criteria for the termination of the test were fulfilled. Maximal exertion was classified as a failure to maintain the ergometer speed for more than 30 seconds despite encouragement or the development of symptoms of fatigue, pain, or lightheadedness. Work rate increments were predetermined with a prediction of the expected work capacity and a test duration of less than 10 minutes. The primary assessed measures of AC were the VO2 peak and the ventilatory AT, which were determined with the V-slope method.9 Predicted values used for comparison were calculated with the default equations of the CPET unit.9
Nutritional Assessment Nutritional and anthropometric assessment was performed by a specialist dietician according to standard methods within 3 days of the CPET evaluation. Heights and weights were obtained with precisions of 1 cm and 0.1 kg, respectively. The anthropometric
56 BERNAL ET AL.
parameters that were primarily assessed were the uncorrected body mass index (BMI), corrected body mass index (BMIc; corrected for fluid retention), midarm muscle circumference (MAMC), and handgrip strength (HGS). BMI was calculated as the weight divided by the height squared (kg/m2), and BMIc was estimated as the dry weight divided by the height squared (kg/m2). The midarm circumference (MAC) and the triceps skinfold (TSF) thickness were first measured to the nearest millimeter in the dominant arm with a measurement tape and a skinfold caliper with a pressure of 10 g/mm2 of contact surface (Holtain, Ltd., London, United Kingdom). Measurements were taken midway between the tip of the acromion and the olecranon process with the patient standing in a relaxed position. MAMC was then calculated from MAC and TSF with the following formula: MAMC ¼ MAC-ðp3TSFÞ The average of 3 measurements was used. MAC, TSF, and MAMC were also expressed as centiles of normal age- and sex-specific population values.19 Voluntary HGS was measured in the dominant hand with a calibrated dynamometer (Takei, Tokyo, Japan). The better of 2 consecutive measurements was recorded (with a 1-minute recovery time between attempts), and HGS was presented as both an absolute value and as a percentage of a sex-specific normal value.19,20
Data Set and Statistical Analysis The etiology of CLD and its severity according to the Model for End-Stage Liver Disease (MELD) score at assessment and on the day of surgery (for patients undergoing LT) were determined.21 The donor risk index was used to assess graft quality.22 The primary outcome measures that were considered included mortality 1 year after CPET for patients not undergoing transplantation during this time period and mortality 1 year after surgery for patients undergoing LT. The secondary outcomes that were considered included the length of stay in the intensive care unit and in the hospital after LT. Statistical comparisons with SPSS 19 used nonparametric univariate testing, and correlations between variables were assessed with Spearman correlations. Cox regression analysis was used to assess the relationships of variables and survival, and a multivariate analysis using backward variable selection included those variables showing significance at the P < 0.2 level in a univariate analysis. Risk factors were separately analyzed with the Kaplan-Meier method and log-rank testing to test for survival differences between groups. In an analysis of nontransplant survival, patients were censored if they underwent transplantation. Data are presented as medians and interquartile ranges or as numbers and percentages. All data were de-identified before the analysis, and the study was approved by the research ethics com-
LIVER TRANSPLANTATION, January 2014
mittee of King’s College Hospital National Health Service Foundation Trust.
RESULTS The study cohort comprised 399 patients; 130 (33%) were female, and the median age was 56 years (interquartile range 5 48-61 years). Patient characteristics are shown in Table 1. One hundred twenty-five patients (31%) had CLD resulting from alcohol, 121 (30%) had a viral etiology, and 153 (38%) had other etiologies; in all, 145 patients (36%) of any etiology had hepatocellular carcinoma. The median MELD score at the time of CPET was 14 (interquartile range 5 11-18), and the overall duration of follow-up after testing was 680 days (interquartile range 5 434-1026 days). The median peak workload achieved during CPET was 86 W (interquartile range 5 63-110 W), and the maximum heart rate was 143 bpm (interquartile range 5 121-157 bpm). The median AT value was 10.9 mL/kg/minute (8.3-13.3 mL/kg/minute), and the median VO2 peak was 15.9 mL/kg/minute (1319.5 mL/kg/minute); 45% of the cases had a VO2 peak < 50% of the predicted value, and 31% had an AT value < 9 mL/kg/minute.
Clinical Correlates of AC Correlates of AC are shown in Table 2. For the cohort as a whole and split by sex, both VO2 peak and AT values correlated closely with the MELD score and its components (except for the serum bilirubin level) and more closely with the serum sodium and albumin levels. The VO2 peak but not AT correlated closely with age and BMI, and weak or no correlations were seen with anthropometric parameters (including BMIc) with the exception of HGS. This correlated with both measures of AC and more strongly with the VO2 peak. This correlation remained strong when HGS was assessed as a percentage of sex-specific control values (r 5 0.287, P < 0.001) and when the VO2 peak was considered as a percentage of predicted values [HGS (kg), r 5 0.179, P < 0.002; HGS (% of control values), r 5 0.299, P < 0.001], and in a sex-specific analysis, it was stronger for males [HGS (% of control values), r 5 0.368, P < 0.001]. The VO2 peak and AT values were lower for patients with an alcohol-related etiology versus patients with other etiologies, but this did not reach statistical significance [VO2 peak: 15.4 mL/kg/minute (IQR 5 13.1-17.9 mL/kg/minute) versus 16.5 mL/kg/minute (interquartile range 5 13.2-20.2 mL/kg/minute), P 5 0.054; AT: 10.4 mL/kg/minute (interquartile range 5 8.2-12.5 mL/kg/minute) versus 11 mL/kg/minute (interquartile range 5 8.7-13.6 mL/kg/minute), P 5 0.08]. Those with ascites at assessment had significantly lower VO2 peak values [14.3 mL/kg/minute (interquartile range 5 12-16.6 mL/kg/minute) versus 17.5 mL/kg/minute (interquartile range 5 14.5-20.8 mL/kg/minute), P < 0.001] and AT values [9.5 mL/
LIVER TRANSPLANTATION, Vol. 20, No. 1, 2014
BERNAL ET AL. 57
TABLE 1. Clinical Features of the Study Cohort at the Time of Assessment
Clinical and laboratory data Age (years)* MELD score* Sodium (mmol/L)* Albumin (g/L)* Etiology [n (%)] Alcohol Viral Other Hepatocellular carcinoma [n (%)] Ascites [n (%)] Anthropometry data BMI (kg/m2)* BMIc (kg/m2)* MAC (cm)* MAC Centile† TSF (cm)* TSF Centile† MAMC (cm)* MAMC centile† HGS (kg)* HGS (% of normal value)* CPET data Maximum heart rate (bpm)* Peak workload (W)* AT (mL/kg/minute)* VO2 peak (mL/kg/minute)* Predicted VO2 peak (%)* VE/VCO2*
Males (n 5 268)
Females (n 5 131)
All (n 5 399)
56 (49-61) 14 (11-18) 136 (133-139) 32 (28-36)
56 (48-61) 14 (11-17) 137 (134-139) 33 (28-38)
56 (48-61) 14 (11-18) 136 (133-139) 32 (28-36)
98 92 79 98 103
(36) (34) (29) (36) (38)
27 29 74 47 38
(21) (22) (57) (36) (29)
125 121 153 145 141
(31) (30) (38) (36) (35)
26.7 (23.5-29.9) 25.0 (22.2-29.2) 30.3 (27-34) 5-10, 75-90 10.8 (7.0-16) 10-25, 75-90 26.7 (24.2-29.4) 5-10, 75-90 33.0 (26.2-38.7) 83 (66-97)
24.1 (21.9-29.5) 23.8 (20.3-29.7) 27.3 (24.6-32.3) 10-25, 50-75 12.8 (9.6-18.8)
ORIGINAL ARTICLE
Aerobic Capacity During Cardiopulmonary Exercise Testing and Survival With and Without Liver Transplantation for Patients With Chronic Liver Disease s Lipcsey,1 Caroline Tallis,1 Kyne Woodsford,1 William Bernal,1 Rosa Martin-Mateos,1 Miklo 1 1 Mark J. Mcphail, Christopher Willars, Georg Auzinger,1 Elizabeth Sizer,1 Michael Heneghan,1 Simon Cottam,2 Nigel Heaton,3* and Julia Wendon1* 1 Liver Intensive Therapy Unit, 2Department of Anaesthetics, and 3Liver Transplant Surgical Service, Institute of Liver Studies, King’s College Hospital, King’s College London, London, United Kingdom
Chronic liver disease (CLD) is associated with muscle wasting, reduced exercise tolerance and aerobic capacity (AC). Measures of AC determined with cardiopulmonary exercise testing (CPET) may predict survival after liver transplantation (LT), but the relationship with nontransplant outcomes is uncertain. In patients assessed for LT, we examined the relationship of CPET AC parameters with the severity of liver disease, nutritional state, and survival with and without LT. Patients assessed for elective first LT who underwent CPET and an anthropometric assessment at a single center were studied. CPET-derived measures of AC that were evaluated included the peak oxygen consumption (VO2 peak) and the anaerobic threshold (AT). Three hundred ninety-nine patients underwent CPET, and 223 underwent LT; 45% of the patients had a VO2 peak < 50% of the predicted value, and 31% had an AT < 9 mL/kg/minute. The VO2 peak and AT values correlated with the Model for End-Stage Liver Disease score, but they more closely correlated with serum sodium and albumin levels. The handgrip strength correlated strongly with the VO2 peak. Patients with impaired AC had prolonged hospitalization after LT, and nonsurvivors had lower AT values than survivors 1 year after transplantation (P < 0.05); this was significant in a multivariate analysis. One hundred seventy-six patients did not undergo LT; the 1-year mortality rate was 34.6%. The AT (P < 0.05) and VO2 peak values (P < 0.001) were lower for nonsurvivors. In a multivariate analysis, AT was independently associated with nonsurvival. In conclusion, AC was markedly impaired in many patients with CLD. In patients who did not undergo transplantation, impaired AT was predictive of mortality, and in patients undergoing LT, it was related to postoperative hospitalization and survival. AC should be evaluated as a modifiable factor for improving patient survival whether or not C 2013 AASLD. LT is anticipated. Liver Transpl 20:54-62, 2014. V Received June 4, 2013; accepted September 30, 2013. Many patients with chronic liver disease (CLD) develop progressive impairment of their fitness and exercise capacity.1 The clinical significance of these
changes is reflected by the close association of altered musculature and exercise capacity with the risk for major disease-related complications and mortality.2-8
Abbreviations: AC, aerobic capacity; AT, anaerobic threshold; BMI, body mass index; BMIc, corrected body mass index; CLD, chronic liver disease; CPET, cardiopulmonary exercise testing; HGS, handgrip strength; HR, hazard ratio; LT, liver transplantation; MAC, midarm circumference; MAMC, midarm muscle circumference; MELD, Model for End-Stage Liver Disease; TSF, triceps skinfold; VO2 peak, peak oxygen consumption. The authors have no conflicts of interest or financial support to report. *These authors contributed equally to this work. Address reprint requests to William Bernal, M.D., Liver Intensive Therapy Unit, Institute of Liver Studies, King’s College Hospital, King’s College London, Denmark Hill, London, United Kingdom SE5 9RS. Telephone: 44 203 299 4488; FAX: 44 203 299 3367; E-mail: [email protected] DOI 10.1002/lt.23766 View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
C 2013 American Association for the Study of Liver Diseases. V
LIVER TRANSPLANTATION, Vol. 20, No. 1, 2014
Cardiopulmonary exercise testing (CPET) represents the gold standard for the determination of exercise capacity and cardiorespiratory reserve, and it generates measures of a number of aspects of aerobic capacity (AC), which is the ability of the body to consume and use oxygen during exercise.8,9 The determined measures include the peak oxygen consumption (VO2 peak) achieved during testing and the point at which an apparent shift occurs from aerobic to anaerobic metabolism [ie, the anaerobic threshold (AT)]. The former is a measure related to the peak workload achieved and is dependent on volition, whereas the latter is a submaximal measure that is independent of a patient’s volitional effort.1,8,9 Together, these broadly indicate the ability of the cardiopulmonary system to deliver oxygen to the peripheral tissues and the ability of the tissues to use that oxygen. Studies have shown that both of these measures are significantly lower in patients with CLD in comparison with both predicted values and healthy controls, and they are related to survival during wait listing for transplantation and after surgery.3,5,8,10,11 These findings are potentially important because AC can be improved though physical training interventions. In patients with other chronic diseases in whom AC is similarly impaired, such interventions clearly improve the quality of life, reduce resource use, and may prolong survival.12-14 These therapies have yet to be tested in patients with CLD and could represent an important but untapped clinical resource. However, although CPET testing is apparently safe and effective in patients with CLD, it is expensive and is not uniformly available. Furthermore, studies reporting on its use in such patients have been mostly small and focused on the relationship between preoperatively determined measures of AC and short-term posttransplant outcomes. In these studies, the relationship between CPET measures of AC and standard measures of CLD severity has not been consistent.1 Muscle bulk has also been recently shown to be related closely to survival for patients with CLD; other more straightforward measures of muscle size and function might usefully identify patients with CLD and impaired AC who might benefit from interventions or aid in decision making concerning suitability for transplantation.2,15-17 Therefore, in a large cohort of patients from a single center, we examined the relationship between findings from CPET performed during the evaluation for potential liver transplantation (LT) and patients’ clinical features. Our specific objectives were (1) to clarify the relationship between parameters of AC determined with CPET and those assessing the severity of liver disease and anthropomorphic measures of muscle bulk, function, and overall nutritional state and (2) to confirm the previously demonstrated relationship between CPET measures of AC and post-LT survival, with the latter extended from short-term survival to longer term survival after LT. Finally, we wanted to investigate the relationship between AC and survival
BERNAL ET AL. 55
in patients who did not undergo transplantation. In doing so, we hoped to identify the clinical features that characterize the patients at greatest risk for impaired AC and the patient subgroup most likely to benefit from future interventions.
PATIENTS AND METHODS This retrospective, observational cohort study examined 399 consecutive patients undergoing a protocolized LT assessment at the Institute of Liver Studies (King’s College Hospital, London, United Kingdom) between February 2007 and June 2011. Only patients with cirrhotic liver disease who were assessed for elective first LT were studied, and the exclusion criteria included a patient age less than 16 years and standard exclusions for performing CPET.18 The decision to accept a patient for LT was made by a multidisciplinary team of clinicians and was not based on CPET performance or a nutritional assessment alone.
CPET Patients underwent CPET at the time of the LT assessment for elective LT. Any b-adrenoceptorblocking medication was discontinued at least 12 hours before the test, and ascites was drained by paracentesis. CPET was conducted upright with a cycle ergometer (Ergoline, Bitz, Germany), an ergospirometer, and a continuous 12-lead electrocardiograph (Ultima CardiO2, MGC Diagnostics, St. Paul, MN, or Schiller AG, Baar, Switzerland) by a physiology team that was dedicated to performing the test and was independent of the LT assessment team. The equipment was calibrated daily. CPET was executed according to a fixed protocol with respect to the monitoring, startup, test procedure, and termination. In brief, a ramp protocol was used with the load increasing in 5- to 10-W increments, and it was terminated when maximal exertion was reached or when safety criteria for the termination of the test were fulfilled. Maximal exertion was classified as a failure to maintain the ergometer speed for more than 30 seconds despite encouragement or the development of symptoms of fatigue, pain, or lightheadedness. Work rate increments were predetermined with a prediction of the expected work capacity and a test duration of less than 10 minutes. The primary assessed measures of AC were the VO2 peak and the ventilatory AT, which were determined with the V-slope method.9 Predicted values used for comparison were calculated with the default equations of the CPET unit.9
Nutritional Assessment Nutritional and anthropometric assessment was performed by a specialist dietician according to standard methods within 3 days of the CPET evaluation. Heights and weights were obtained with precisions of 1 cm and 0.1 kg, respectively. The anthropometric
56 BERNAL ET AL.
parameters that were primarily assessed were the uncorrected body mass index (BMI), corrected body mass index (BMIc; corrected for fluid retention), midarm muscle circumference (MAMC), and handgrip strength (HGS). BMI was calculated as the weight divided by the height squared (kg/m2), and BMIc was estimated as the dry weight divided by the height squared (kg/m2). The midarm circumference (MAC) and the triceps skinfold (TSF) thickness were first measured to the nearest millimeter in the dominant arm with a measurement tape and a skinfold caliper with a pressure of 10 g/mm2 of contact surface (Holtain, Ltd., London, United Kingdom). Measurements were taken midway between the tip of the acromion and the olecranon process with the patient standing in a relaxed position. MAMC was then calculated from MAC and TSF with the following formula: MAMC ¼ MAC-ðp3TSFÞ The average of 3 measurements was used. MAC, TSF, and MAMC were also expressed as centiles of normal age- and sex-specific population values.19 Voluntary HGS was measured in the dominant hand with a calibrated dynamometer (Takei, Tokyo, Japan). The better of 2 consecutive measurements was recorded (with a 1-minute recovery time between attempts), and HGS was presented as both an absolute value and as a percentage of a sex-specific normal value.19,20
Data Set and Statistical Analysis The etiology of CLD and its severity according to the Model for End-Stage Liver Disease (MELD) score at assessment and on the day of surgery (for patients undergoing LT) were determined.21 The donor risk index was used to assess graft quality.22 The primary outcome measures that were considered included mortality 1 year after CPET for patients not undergoing transplantation during this time period and mortality 1 year after surgery for patients undergoing LT. The secondary outcomes that were considered included the length of stay in the intensive care unit and in the hospital after LT. Statistical comparisons with SPSS 19 used nonparametric univariate testing, and correlations between variables were assessed with Spearman correlations. Cox regression analysis was used to assess the relationships of variables and survival, and a multivariate analysis using backward variable selection included those variables showing significance at the P < 0.2 level in a univariate analysis. Risk factors were separately analyzed with the Kaplan-Meier method and log-rank testing to test for survival differences between groups. In an analysis of nontransplant survival, patients were censored if they underwent transplantation. Data are presented as medians and interquartile ranges or as numbers and percentages. All data were de-identified before the analysis, and the study was approved by the research ethics com-
LIVER TRANSPLANTATION, January 2014
mittee of King’s College Hospital National Health Service Foundation Trust.
RESULTS The study cohort comprised 399 patients; 130 (33%) were female, and the median age was 56 years (interquartile range 5 48-61 years). Patient characteristics are shown in Table 1. One hundred twenty-five patients (31%) had CLD resulting from alcohol, 121 (30%) had a viral etiology, and 153 (38%) had other etiologies; in all, 145 patients (36%) of any etiology had hepatocellular carcinoma. The median MELD score at the time of CPET was 14 (interquartile range 5 11-18), and the overall duration of follow-up after testing was 680 days (interquartile range 5 434-1026 days). The median peak workload achieved during CPET was 86 W (interquartile range 5 63-110 W), and the maximum heart rate was 143 bpm (interquartile range 5 121-157 bpm). The median AT value was 10.9 mL/kg/minute (8.3-13.3 mL/kg/minute), and the median VO2 peak was 15.9 mL/kg/minute (1319.5 mL/kg/minute); 45% of the cases had a VO2 peak < 50% of the predicted value, and 31% had an AT value < 9 mL/kg/minute.
Clinical Correlates of AC Correlates of AC are shown in Table 2. For the cohort as a whole and split by sex, both VO2 peak and AT values correlated closely with the MELD score and its components (except for the serum bilirubin level) and more closely with the serum sodium and albumin levels. The VO2 peak but not AT correlated closely with age and BMI, and weak or no correlations were seen with anthropometric parameters (including BMIc) with the exception of HGS. This correlated with both measures of AC and more strongly with the VO2 peak. This correlation remained strong when HGS was assessed as a percentage of sex-specific control values (r 5 0.287, P < 0.001) and when the VO2 peak was considered as a percentage of predicted values [HGS (kg), r 5 0.179, P < 0.002; HGS (% of control values), r 5 0.299, P < 0.001], and in a sex-specific analysis, it was stronger for males [HGS (% of control values), r 5 0.368, P < 0.001]. The VO2 peak and AT values were lower for patients with an alcohol-related etiology versus patients with other etiologies, but this did not reach statistical significance [VO2 peak: 15.4 mL/kg/minute (IQR 5 13.1-17.9 mL/kg/minute) versus 16.5 mL/kg/minute (interquartile range 5 13.2-20.2 mL/kg/minute), P 5 0.054; AT: 10.4 mL/kg/minute (interquartile range 5 8.2-12.5 mL/kg/minute) versus 11 mL/kg/minute (interquartile range 5 8.7-13.6 mL/kg/minute), P 5 0.08]. Those with ascites at assessment had significantly lower VO2 peak values [14.3 mL/kg/minute (interquartile range 5 12-16.6 mL/kg/minute) versus 17.5 mL/kg/minute (interquartile range 5 14.5-20.8 mL/kg/minute), P < 0.001] and AT values [9.5 mL/
LIVER TRANSPLANTATION, Vol. 20, No. 1, 2014
BERNAL ET AL. 57
TABLE 1. Clinical Features of the Study Cohort at the Time of Assessment
Clinical and laboratory data Age (years)* MELD score* Sodium (mmol/L)* Albumin (g/L)* Etiology [n (%)] Alcohol Viral Other Hepatocellular carcinoma [n (%)] Ascites [n (%)] Anthropometry data BMI (kg/m2)* BMIc (kg/m2)* MAC (cm)* MAC Centile† TSF (cm)* TSF Centile† MAMC (cm)* MAMC centile† HGS (kg)* HGS (% of normal value)* CPET data Maximum heart rate (bpm)* Peak workload (W)* AT (mL/kg/minute)* VO2 peak (mL/kg/minute)* Predicted VO2 peak (%)* VE/VCO2*
Males (n 5 268)
Females (n 5 131)
All (n 5 399)
56 (49-61) 14 (11-18) 136 (133-139) 32 (28-36)
56 (48-61) 14 (11-17) 137 (134-139) 33 (28-38)
56 (48-61) 14 (11-18) 136 (133-139) 32 (28-36)
98 92 79 98 103
(36) (34) (29) (36) (38)
27 29 74 47 38
(21) (22) (57) (36) (29)
125 121 153 145 141
(31) (30) (38) (36) (35)
26.7 (23.5-29.9) 25.0 (22.2-29.2) 30.3 (27-34) 5-10, 75-90 10.8 (7.0-16) 10-25, 75-90 26.7 (24.2-29.4) 5-10, 75-90 33.0 (26.2-38.7) 83 (66-97)
24.1 (21.9-29.5) 23.8 (20.3-29.7) 27.3 (24.6-32.3) 10-25, 50-75 12.8 (9.6-18.8)
