Original Report: Patient-Oriented, Translational Research American
Journal of
Nephrology
Received: August 16, 2013 Accepted: December 27, 2013 Published online: February 6, 2014
Am J Nephrol 2014;39:122–129 DOI: 10.1159/000358335
Visceral Fat Level Is an Independent Risk Factor for Cardiovascular Mortality in Hemodialysis Patients Takayuki Okamoto a, b Satoshi Morimoto a Tatsuyoshi Ikenoue b Yoshiyuki Furumatsu c Atsuhiro Ichihara a
a
Department of Medicine II, Endocrinology and Hypertension, Tokyo Women’s Medical University, Tokyo, Department of Nephrology, Moriguchi Keijinkai Hospital, Moriguchi, and c Department of Health Economics and Industrial Policy, Osaka University Graduate School of Medicine, Suita, Japan
Key Words Visceral fat · Obesity · Computed tomography · Cardiovascular diseases · Atherosclerosis
Abstract Background: Obesity is an independent risk factor for morbidity and mortality in cardiovascular diseases not only in the general population, but also in hemodialysis (HD) patients. We previously reported that an increased visceral fat area (VFA) determined using computed tomography (CT) scans was associated with atherosclerosis in HD patients. However, whether a high VFA is associated with increased cardiovascular mortality in HD patients remains unknown. Therefore, we investigated the relationship between VFA and prognosis in HD patients. Methods: VFA was estimated in 126 patients on maintenance HD using CT scans. These patients were followed for 60 months. Results: Kaplan-Meier analysis revealed that the cardiovascular survival rate was significantly lower in the high-VFA group, with a VFA of 71.5 cm2 or greater, than in the low-VFA group, with a VFA of less than 71.5cm2. In univariate Cox proportional hazards analyses, age, albumin, low-density lipoprotein cholesterol, cardio-thoracic ratio and VFA above 71.5 cm2 were significantly correlated with cardiovascular deaths. In multivariate analyses testing these factors as dependent variables, VFA
© 2014 S. Karger AG, Basel 0250–8095/14/0392–0122$39.50/0 E-Mail [email protected] www.karger.com/ajn
above 71.5 cm2 was estimated to be an independent predictor of cardiovascular deaths. Conclusion: These results suggest that an increased VFA is a stronger risk factor for cardiovascular deaths in HD patients. Measuring VFA may be recommended for predicting the risk of cardiovascular diseases in HD patients. © 2014 S. Karger AG, Basel
Introduction
Patients undergoing maintenance hemodialysis (HD) are at an increased risk of death from cardiovascular diseases [1–3]. Risk factors include hypertension, diabetes mellitus (DM), vascular calcification and inflammation [4, 5]. Obesity is an independent risk factor for morbidity and mortality in cardiovascular diseases in the general population [6] and HD patients [7]. Interestingly, Odamaki et al. [8] reported that visceral fat area (VFA) estimated using computed tomography (CT) scans was significantly greater in HD patients than in control subjects of a similar body mass index (BMI). We previously reported that the incidence of high VFA was markedly higher than that of high BMI, high low-density lipoprotein cholesterol (LDL-C) or high triglycerides, and VFA was shown to be associated with both arterial stiffness and Satoshi Morimoto, MD, PhD Department of Medicine II, Endocrinology and Hypertension Tokyo Women’s Medical University 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666 (Japan) E-Mail smorimoto @ endm.twmu.ac.jp
Downloaded by: University of Tokyo 157.82.153.40 - 5/30/2015 5:14:47 AM
b
Subjects and Methods Study Subjects Subjects consisted of 126 outpatients on maintenance dialysis at Kadoma Keijinkai Clinic in Kadoma City, Japan, in January 2008. Each subject provided written informed consent to participate in the study, which was approved by the Review Board of Moriguchi Keijinkai Hospital, the mother hospital of Kadoma Keijinkai Clinic. Background Factors At enrollment, information was collected on age, gender, duration of HD, DM, BMI and waist circumference (WC). DM was diagnosed when patients were already receiving treatment with an oral hypoglycemic agent or insulin, or met the World Health Organization criteria [11]. WC was measured at the height of the umbilicus while the patient was standing and after exhaling. Predialysis systolic and diastolic blood pressures and the cardio-thoracic ratio from a chest x-ray obtained before HD was assessed on the first dialysis day of the week. Predialysis values of hemoglobin, blood sugar, total protein, albumin, total cholesterol, LDL-C, highdensity-lipoprotein cholesterol (HDL-C), triglycerides, albumincorrected calcium, inorganic phosphorus, blood urea nitrogen, creatinine, uric acid, ferritin, human atrial natriuretic polypeptide and intact parathyroid hormone of fasting blood on the first dialysis day of the week were measured by conventional methods at an external testing laboratory (Kishimoto Inc., Tomakomai City, Japan). Abdominal CT The distribution of body fat was determined using CT imaging. All CT scans were performed on a nondialysis day with the subject in a supine position, using an 8-slice multidetector CT scanner (Aquilion TSX-021A, Toshiba, Tokyo, Japan). Seven-millimeterthick slices (120 kV, 400 mA) were acquired. The subcutaneous and visceral fat layers were clearly defined with attenuation ranging from –60 to –160 HU. The line was drawn along the abdominal cavity and fat in the abdominal cavity was measured. Intraperitoneal fat was defined as the VFA. Extraperitoneal fat between the skin and muscle was defined as the subcutaneous fat area (SFA). SFA and VFA were measured at the level of the umbilicus. Study Protocol All patients were followed for 60 months, and the prognosis and cause of death were investigated. We explored possible cutoff points to distinguish the risk of cardiovascular deaths with the use of receiver operator characteristic (ROC) curves to determine the incremental prognostic value of VFA. Age, gender, dialysis history, DM, BMI, WC, hemoglobin, blood sugar, total protein, albumin,
VFA in HD Patients
total cholesterol, LDL-C, HDL-C, triglycerides, albumin-corrected calcium, inorganic phosphorus, blood urea nitrogen, creatinine, uric acid, ferritin, human atrial natriuretic polypeptide, intact parathyroid hormone, cardio-thoracic ratio, VFA and SFA were compared between high- and low-VFA groups. Multicolinearity was investigated for factors with a possible relationship with VFA. We evaluated the relationship between VFA and cardiovascular mortality using a spline after adjusting for age, gender, albumin and LDL-C as relevant confounding variables. Cardiovascular mortality was compared between the high- and low-VFA groups and predictors for cardiovascular deaths were assessed. The VFA or SFA of patients who died of cardiovascular diseases (CV+ group), those who died from noncardiovascular causes (CV– group) and those who survived (S group) were compared. All patients were divided into 5 groups according to VFA values to make the number of each group equal. The trend of mortality due to cardiovascular diseases, noncardiovascular diseases and all causes were compared among these 5 groups. Statistical Analysis To compare values in the high- and low-VFA groups, the nonpaired student t test and χ2 test were performed as appropriate. ROC curves and the corresponding area under the curve (AUC) were used to evaluate how the prediction model preformed on the test data. The cardiovascular mortality rate was compared between the high- and low-VFA groups using Kaplan-Meier survival curves. A comparison of probability curves was performed with the log-rank test. Considering a nonlinear association between VFA and CV mortality, we analyzed the associations between VFA and CV mortality as continuous variables after adjusting for age, gender, albumin and LDL-C, fitting a restricted cubic spline function with 4 knots (located at the 5th, 35th, 65th and 95th percentiles). We analyzed the correlation of WC and BMI with VFA, and checked for multicolinearity. Vif (variance inflation factor) was calculated using a regression model. We did univariate analyses to test the relationship between background factors and cardiovascular deaths. Subsequently, multivariate analysis with Cox proportional hazards model was done testing factors which were assessed to be related with cardiovascular deaths as dependent variables. Schoenfeld residuals were used to investigate the proportional hazards assumption. Among the 5 groups divided by VFA values, the trend of mortality was analyzed using the Cochran-Armitage trend test. Values were expressed as the mean ± SD and differences were considered significant at p < 0.05. Statistical analyses were performed using Stata version 11.2 (Stata Corp., College Station, Tex., USA) and JMP (version 8.0.2; SAS Institute, Cary, N.C., USA).
Results
A total of 126 patients with a mean age of 67 ± 12 years and VFA of 66.2 ± 47.0 cm2 were reviewed. All patients were followed for 60 months. Forty-six patients died over the 60 months and the average observation period was 49.8 ± 17.3 months (range 5–60). Fifteen patients died of cardiovascular diseases (CV+ group): 6 of acute myocardial infarction, 5 of congestive heart failure, 2 of cerebral Am J Nephrol 2014;39:122–129 DOI: 10.1159/000358335
123
Downloaded by: University of Tokyo 157.82.153.40 - 5/30/2015 5:14:47 AM
atherosclerosis in diabetic patients on maintenance HD [9]. Visceral obesity assessed by CT is a predictor of cardiovascular events in nondialysis chronic kidney disease patients [10]. However, whether high VFA is associated with increased cardiovascular mortality in dialysis patients remains unknown. Therefore, we investigated the relationship between VFA and prognosis in maintenance HD patients.
p value
66±12 31/26 50.1±46.3 33/24 22.8±2.6 87.8±6.4
69±12 41/28 52.9±49.5 33/36 19.9±2.5 78.3±8.2
0.155 0.323 0.739 0.266
Journal of
Nephrology
Received: August 16, 2013 Accepted: December 27, 2013 Published online: February 6, 2014
Am J Nephrol 2014;39:122–129 DOI: 10.1159/000358335
Visceral Fat Level Is an Independent Risk Factor for Cardiovascular Mortality in Hemodialysis Patients Takayuki Okamoto a, b Satoshi Morimoto a Tatsuyoshi Ikenoue b Yoshiyuki Furumatsu c Atsuhiro Ichihara a
a
Department of Medicine II, Endocrinology and Hypertension, Tokyo Women’s Medical University, Tokyo, Department of Nephrology, Moriguchi Keijinkai Hospital, Moriguchi, and c Department of Health Economics and Industrial Policy, Osaka University Graduate School of Medicine, Suita, Japan
Key Words Visceral fat · Obesity · Computed tomography · Cardiovascular diseases · Atherosclerosis
Abstract Background: Obesity is an independent risk factor for morbidity and mortality in cardiovascular diseases not only in the general population, but also in hemodialysis (HD) patients. We previously reported that an increased visceral fat area (VFA) determined using computed tomography (CT) scans was associated with atherosclerosis in HD patients. However, whether a high VFA is associated with increased cardiovascular mortality in HD patients remains unknown. Therefore, we investigated the relationship between VFA and prognosis in HD patients. Methods: VFA was estimated in 126 patients on maintenance HD using CT scans. These patients were followed for 60 months. Results: Kaplan-Meier analysis revealed that the cardiovascular survival rate was significantly lower in the high-VFA group, with a VFA of 71.5 cm2 or greater, than in the low-VFA group, with a VFA of less than 71.5cm2. In univariate Cox proportional hazards analyses, age, albumin, low-density lipoprotein cholesterol, cardio-thoracic ratio and VFA above 71.5 cm2 were significantly correlated with cardiovascular deaths. In multivariate analyses testing these factors as dependent variables, VFA
© 2014 S. Karger AG, Basel 0250–8095/14/0392–0122$39.50/0 E-Mail [email protected] www.karger.com/ajn
above 71.5 cm2 was estimated to be an independent predictor of cardiovascular deaths. Conclusion: These results suggest that an increased VFA is a stronger risk factor for cardiovascular deaths in HD patients. Measuring VFA may be recommended for predicting the risk of cardiovascular diseases in HD patients. © 2014 S. Karger AG, Basel
Introduction
Patients undergoing maintenance hemodialysis (HD) are at an increased risk of death from cardiovascular diseases [1–3]. Risk factors include hypertension, diabetes mellitus (DM), vascular calcification and inflammation [4, 5]. Obesity is an independent risk factor for morbidity and mortality in cardiovascular diseases in the general population [6] and HD patients [7]. Interestingly, Odamaki et al. [8] reported that visceral fat area (VFA) estimated using computed tomography (CT) scans was significantly greater in HD patients than in control subjects of a similar body mass index (BMI). We previously reported that the incidence of high VFA was markedly higher than that of high BMI, high low-density lipoprotein cholesterol (LDL-C) or high triglycerides, and VFA was shown to be associated with both arterial stiffness and Satoshi Morimoto, MD, PhD Department of Medicine II, Endocrinology and Hypertension Tokyo Women’s Medical University 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666 (Japan) E-Mail smorimoto @ endm.twmu.ac.jp
Downloaded by: University of Tokyo 157.82.153.40 - 5/30/2015 5:14:47 AM
b
Subjects and Methods Study Subjects Subjects consisted of 126 outpatients on maintenance dialysis at Kadoma Keijinkai Clinic in Kadoma City, Japan, in January 2008. Each subject provided written informed consent to participate in the study, which was approved by the Review Board of Moriguchi Keijinkai Hospital, the mother hospital of Kadoma Keijinkai Clinic. Background Factors At enrollment, information was collected on age, gender, duration of HD, DM, BMI and waist circumference (WC). DM was diagnosed when patients were already receiving treatment with an oral hypoglycemic agent or insulin, or met the World Health Organization criteria [11]. WC was measured at the height of the umbilicus while the patient was standing and after exhaling. Predialysis systolic and diastolic blood pressures and the cardio-thoracic ratio from a chest x-ray obtained before HD was assessed on the first dialysis day of the week. Predialysis values of hemoglobin, blood sugar, total protein, albumin, total cholesterol, LDL-C, highdensity-lipoprotein cholesterol (HDL-C), triglycerides, albumincorrected calcium, inorganic phosphorus, blood urea nitrogen, creatinine, uric acid, ferritin, human atrial natriuretic polypeptide and intact parathyroid hormone of fasting blood on the first dialysis day of the week were measured by conventional methods at an external testing laboratory (Kishimoto Inc., Tomakomai City, Japan). Abdominal CT The distribution of body fat was determined using CT imaging. All CT scans were performed on a nondialysis day with the subject in a supine position, using an 8-slice multidetector CT scanner (Aquilion TSX-021A, Toshiba, Tokyo, Japan). Seven-millimeterthick slices (120 kV, 400 mA) were acquired. The subcutaneous and visceral fat layers were clearly defined with attenuation ranging from –60 to –160 HU. The line was drawn along the abdominal cavity and fat in the abdominal cavity was measured. Intraperitoneal fat was defined as the VFA. Extraperitoneal fat between the skin and muscle was defined as the subcutaneous fat area (SFA). SFA and VFA were measured at the level of the umbilicus. Study Protocol All patients were followed for 60 months, and the prognosis and cause of death were investigated. We explored possible cutoff points to distinguish the risk of cardiovascular deaths with the use of receiver operator characteristic (ROC) curves to determine the incremental prognostic value of VFA. Age, gender, dialysis history, DM, BMI, WC, hemoglobin, blood sugar, total protein, albumin,
VFA in HD Patients
total cholesterol, LDL-C, HDL-C, triglycerides, albumin-corrected calcium, inorganic phosphorus, blood urea nitrogen, creatinine, uric acid, ferritin, human atrial natriuretic polypeptide, intact parathyroid hormone, cardio-thoracic ratio, VFA and SFA were compared between high- and low-VFA groups. Multicolinearity was investigated for factors with a possible relationship with VFA. We evaluated the relationship between VFA and cardiovascular mortality using a spline after adjusting for age, gender, albumin and LDL-C as relevant confounding variables. Cardiovascular mortality was compared between the high- and low-VFA groups and predictors for cardiovascular deaths were assessed. The VFA or SFA of patients who died of cardiovascular diseases (CV+ group), those who died from noncardiovascular causes (CV– group) and those who survived (S group) were compared. All patients were divided into 5 groups according to VFA values to make the number of each group equal. The trend of mortality due to cardiovascular diseases, noncardiovascular diseases and all causes were compared among these 5 groups. Statistical Analysis To compare values in the high- and low-VFA groups, the nonpaired student t test and χ2 test were performed as appropriate. ROC curves and the corresponding area under the curve (AUC) were used to evaluate how the prediction model preformed on the test data. The cardiovascular mortality rate was compared between the high- and low-VFA groups using Kaplan-Meier survival curves. A comparison of probability curves was performed with the log-rank test. Considering a nonlinear association between VFA and CV mortality, we analyzed the associations between VFA and CV mortality as continuous variables after adjusting for age, gender, albumin and LDL-C, fitting a restricted cubic spline function with 4 knots (located at the 5th, 35th, 65th and 95th percentiles). We analyzed the correlation of WC and BMI with VFA, and checked for multicolinearity. Vif (variance inflation factor) was calculated using a regression model. We did univariate analyses to test the relationship between background factors and cardiovascular deaths. Subsequently, multivariate analysis with Cox proportional hazards model was done testing factors which were assessed to be related with cardiovascular deaths as dependent variables. Schoenfeld residuals were used to investigate the proportional hazards assumption. Among the 5 groups divided by VFA values, the trend of mortality was analyzed using the Cochran-Armitage trend test. Values were expressed as the mean ± SD and differences were considered significant at p < 0.05. Statistical analyses were performed using Stata version 11.2 (Stata Corp., College Station, Tex., USA) and JMP (version 8.0.2; SAS Institute, Cary, N.C., USA).
Results
A total of 126 patients with a mean age of 67 ± 12 years and VFA of 66.2 ± 47.0 cm2 were reviewed. All patients were followed for 60 months. Forty-six patients died over the 60 months and the average observation period was 49.8 ± 17.3 months (range 5–60). Fifteen patients died of cardiovascular diseases (CV+ group): 6 of acute myocardial infarction, 5 of congestive heart failure, 2 of cerebral Am J Nephrol 2014;39:122–129 DOI: 10.1159/000358335
123
Downloaded by: University of Tokyo 157.82.153.40 - 5/30/2015 5:14:47 AM
atherosclerosis in diabetic patients on maintenance HD [9]. Visceral obesity assessed by CT is a predictor of cardiovascular events in nondialysis chronic kidney disease patients [10]. However, whether high VFA is associated with increased cardiovascular mortality in dialysis patients remains unknown. Therefore, we investigated the relationship between VFA and prognosis in maintenance HD patients.
p value
66±12 31/26 50.1±46.3 33/24 22.8±2.6 87.8±6.4
69±12 41/28 52.9±49.5 33/36 19.9±2.5 78.3±8.2
0.155 0.323 0.739 0.266
