Search
Close
Search
Advanced Search
Search Menu

Abstract

Background

Left ventricular hypertrophy is causally implicated in the high risk of death and heart failure (HF) in chronic kidney disease (CKD) patients. Whether the left ventricular mass index (LVMI) adds meaningful predictive power for mortality and de novo HF to simple risk models has not been tested in the CKD population.

Methods

We investigated this problem in 1352 CKD patients enrolled in the Chronic Renal Insufficiency Cohort (CRIC). LVMI was measured by echocardiography and the risks for death and HF were estimated by the Study of Heart and Renal Protection (SHARP) score, a well-validated risk score in CKD patients.

Results

During a median follow-up of 7.7 years, 326 patients died and 208 had de novo HF. The LVMI and the SHARP score and a cross-validated model for HF (CRIC model) were all significantly (P < 0.001) related to the risk of death and HF. LVMI showed a discriminatory power for death (Harrell’s C index 66%) inferior to that of the SHARP score (71%) and the same was true for the risk of HF both in the test (LVMI 72%, CRIC model 79%) and in the validation cohort (LVMI 71%, CRIC model 74%). LVMI increased very little the discriminatory (2–3%) and the risk reclassification power (3.0–4.8%) by the SHARP score and the CRIC model for HF for the same outcomes.

Conclusions

In CKD, measurement of LVMI solely for the stratification of risk of death and perhaps for the risk of HF does not provide evident prognostic values in this condition.

CKD, death, heart failure, LVH, left ventricular mass index, prognosis, SHARP risk score

INTRODUCTION

Alterations in left ventricular (LV) mass and structure are a hallmark in animal models of mild [1] and advanced [2] chronic kidney disease (CKD). In living kidney donors, LV mass shows a mild increase 1 year after kidney donation [3, 4], at the time when the glomerular filtration rate (GFR) measured by chromium-51-labeled ethylenediaminetetraacetic acid clearance [3] or estimated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation [4] is ∼30% lower than prior to donation. In CKD patients without heart failure (HF), the prevalence of LV hypertrophy (LVH) is increasingly greater comparing Stage G2–Stage G5 CKD, further supporting the importance of renal dysfunction in the development of LVH [5]. LVH in CKD also predicts incident HF and death even after adjustment for major cardiovascular risk factors and cardiac biomarkers such as brain natriuretic peptide and troponin T [6].

Current guidelines recommend considering echocardiography to refine cardiovascular (CV) risk assessment in hypertensive patients [7]. However, there are few studies in hypertensive patients examining the added value of echocardiography in addition to established risk scores based on classical risk factors like the Framingham score or SCORE [8]. Similarly, the presence of LVH is held as a useful measurement for risk stratification in CKD patients, and some investigators have suggested that the measurement of LV mass index (LVMI) should be periodically performed in these patients [9]. Although the causal role of LVH for the risk of HF and cardiovascular disease (CVD) in pre-dialysis CKD patients is beyond question, the prognostic value of this biomarker in the CKD population remains undefined because there is still no study investigating the prognostic value of LVM by applying risk discrimination [10] and re-classification analyses [11] and calibration [12]. A recent study in two European cohorts of end-stage kidney failure patients on haemodialysis [13] demonstrated that the measurement of LVMI fails to add predictive power to statistical models based on simple, easily available clinical parameters in this population. Since observational studies are population specific, the question of whether the measurement of LV mass may be useful for predicting clinical outcomes in the pre-dialysis CKD population needs to be specifically tested in the same population.

To be adopted in every day clinical practice, a given candidate prognostic biomarker, such as LVMI, should provide prognostic information beyond and above that provided by well-validated and simple risk prediction rules [14]. The Study of Heart and Renal Protection (SHARP) CKD-CVD risk calculator (available at http://dismod.ndph.ox.ac.uk/kidneymodel/app/) was developed in 9270 patients with moderate–severe CKD and validated in three external CKD cohorts [15], which provides reliable estimates of the death probability within 5 and 10 years by using simple and easily available risk factors.

In this study we investigated whether LVMI adds prognostic information to the risk estimates made by the SHARP risk calculator for predicting all-cause mortality and to a simple risk model for HF built within the Chronic Renal Insufficiency Cohort (CRIC) database, an ongoing multicentre cohort study including a large number of CKD patients with various degrees of renal dysfunction, who were enrolled between June 2003 and August 2008 from seven clinical centres in the USA [16].

MATERIALS AND METHODS

The study protocol of the CRIC study was approved by the institutional review boards at participating institutions and the research was conducted in accordance with the ethical principles of the Declaration of Helsinki.

Patients

The CRIC study was designed as a prospective cohort of approximately 3600 patients with mild–moderate CKD who were enrolled through seven clinical centres in the USA [16, 17]. Eligibility and exclusion criteria of the CRIC study are given in Supplementary data, Table S1.

Echocardiography

Echocardiograms were performed at 1 year after enrolment using standard techniques to acquire optimal views of cardiac function. Images were transferred to the core echocardiography laboratory (University of Pennsylvania), where they were read by a registered diagnostic cardiac sonographer according to the guidelines of the American Society of Echocardiography [18]. Readers were blinded to participants’ baseline estimated GFR (eGFR). LVMI was derived by the area length method and indexed to height2.7. According to the scope of this study, only CRIC patients of with available echocardiographic data were retained in the analysis (see Figure 1). In a sensitivity analysis, we also considered the additional prognostic value of LV ejection fraction and the E/A ratio as indicators of LV systolic and diastolic function, respectively.

FIGURE 1
Flow of patients across the study.
Open in new tabDownload slide

Flow of patients across the study.

Laboratory and blood pressure (BP) data

Samples for cystatin C were processed using a Siemens BNII nephelometer at the CRIC Central Laboratory, with an intra-assay coefficient of variation of 4.9%. eGFR was calculated from serum cystatin and creatinine using the CRIC study equation [19]. Biomarker assays for lipids, haemoglobin, phosphate, calcium, glucose, high sensitivity CRP and serum albumin were performed at the CRIC Central Laboratory at the University of Pennsylvania. BP was averaged over three measurements performed in a standardized fashion in a seated position at rest using a calibrated sphygmomanometer.

SHARP risk score calculator

The SHARP risk calculator permits the calculation of 5- and 10-year estimated death probability in Stages 3b–5 CKD patients >40 years of age [15]. The SHARP risk estimates are calculated conditional on age, gender, ethnicity, education, adult dependence, smoking, alcohol intake, body mass index (BMI), BP, high-density lipoprotein cholesterol, albumin, haemoglobin, phosphate, urinary albumin:creatinine ratio, background CV comorbidities, diabetes, CKD stage, CKD duration and type of nephropathy. For the purpose of this study, patients with missing values for at least one of these variables were excluded from the analysis (see Figure 1).

Follow-up study

The primary outcome was overall mortality, which was assessed over a median of 7.7 years of follow-up [interquartile range (IQR) 6.5–8.7 months]. HF (i.e. the secondary study outcome) was identified by asking study participants every 6 months if they were hospitalized, and selected hospitals or healthcare systems were queried for qualifying encounters. The first 30 discharge codes were identified for all hospitalizations, and codes relevant to HF resulted in the retrieval of medical records by study personnel for centralized adjudicated review. At least two study physicians reviewed all possible HF events using medical records and guidelines on clinical symptoms, radiographic evidence of pulmonary congestion, physical examination of the heart and lungs and, when available, central venous haemodynamic monitoring data and echocardiographic imaging. HF was confirmed when both reviewers agreed based on modified clinical Framingham criteria [20]. The follow-up time in the survival analysis started from the date of the echocardiographic assessment. Patients’ follow‐up was ended at the time of HF (for incident HF only), death, withdrawal, loss to follow‐up or the end of the follow‐up period, whichever occurred first. For both all-cause mortality and HF, withdrawal, loss to follow-up and the end of the follow-up period were censoring events. For HF, death prior to the occurrence of HF was a competing event.

Statistical analysis

Normally distributed continuous variables were expressed as mean ± SD, non-normally distributed continuous variables as median and IQR and binary variables as absolute frequency and percentage. The interrelationship between the SHARP risk scores (at 5 and 10 years) and LVMI was investigated by Spearman’s correlation coefficient (ρ), and the P-value and shared variance between these two variables was assessed by calculating R2.

To examine the additional prognostic values of LVMI for all-cause mortality and for HF, we randomly split the study cohort into a testing cohort and a validation cohort with a ratio of 1:1. The independent relationship between the incidence rate of all-cause mortality with the SHARP risk scores (i.e. 5-year and 10-year estimated death probabilities) and LVMI was investigated by univariable and multivariable Cox proportional hazards models in the testing cohort. In the Cox models, associations were expressed as hazard ratios (HRs) along with their 95% confidence intervals (CIs) and P-values.

The additional prognostic value of LVMI for all-cause mortality over that provided by the SHARP risk scores was investigated by discrimination [Harrell’s C (HC) index] [10], calibration [12] and risk reclassification [integrated discrimination improvement (IDI)] [11] metrics.

Because no specific risk calculator is available to predict the risk for HF in CKD patients, to estimate the additional prognostic value of LVMI to predict such an outcome, we developed a mechanistic model of HF within the CRIC in the testing cohort and evaluated the additional prognostic value of LVMI in the validation cohort. Because of the presence of the competitive risk of death prior to HF, the subdistribution hazards models (i.e. Fine and Gray’s models) were adopted [21]. We built a mechanistic model based on variables (listed in Table 1) related to the study outcome with P < 0.05 in univariate sub-distribution hazards models in the testing cohort; i.e. age, BMI, diabetes, systolic BP, haemoglobin, albumin, triglycerides, phosphate, eGFR, CV comorbidities and history of atrial fibrillation, and referred to this model as the CRIC model. Associations were expressed as subdistribution HRs (SHRs) along with their 95% CIs and P-values. We examined whether LVMI holds additional prognostic value for predicting HF using the CRIC model in the validation cohort. All the analyses were performed using Stata/IC 13.1 for Windows (StataCorp, College Station, TX, USA).

Table 1

Patients’ demographic, clinical and biochemical characteristics at the beginning of the follow-up

VariablesWhole cohortTest cohortValidation cohort
(n = 1352)(n = 676)(n = 676)
Age (years)61 ± 961 ± 962 ± 9
Male gender, %515151
Diabetic, %555357
Ex-smokers, %444346
Current smokers, %161714
Alcohol drinkers, %535254
BMI (kg/m2)32 ± 832 ± 832 ± 8
CKD duration (years)3 ± 23 ± 23 ± 2
Systolic BP (mmHg)128 ± 22129 ± 22128 ± 21
Diastolic BP (mmHg)69 ± 1269 ± 1368 ± 12
On antihypertensive treatment, %979797
On statins, %666765
Haemoglobin (g/dL)12.3 ± 1.712.4 ± 1.712.2 ± 1.7
Albumin (g/dL)4.0 ± 0.44.0 ± 0.44.0 ± 0.5
Total cholesterol (mg/dL)179 ± 44179 ± 43179 ± 45
High-density lipoprotein cholesterol (mg/dL)48 ± 1548 ± 1548 ± 16
Triglycerides (mg/dL)157 ± 103155 ± 99161 ± 108
Phosphate (mg/dL)3.8 ± 0.73.8 ± 0.73.8 ± 0.7
Calcium (mg/dL)9.3 ± 0.69.3 ± 0.69.3 ± 0.5
Glucose (mg/dL)118 ± 54116 ± 52120 ± 56
eGFR (mL/min/1.73 m2)31.6 ± 8.331.5 ± 8.431.7 ± 8.3
Urine albumin:creatinine ratio (mg/g), median (IQR)132 (17–679)147 (19–690)120 (16–673)
LVMI (g/m2.7)68.4 ± 24.668.6 ± 25.768.2 ± 23.5
CRP (mg/L), median (IQR)2.7 (1.1–6.4)2.8 (1.1–6.3)2.5 (1.1–6.6)
Diabetic nephropathy, %323034
Background CV comorbidities
Myocardial infarction/revascularization, %272629
Peripheral vascular disease, %10119
Congestive HF, %131412
Stroke, %131412
Atrial fibrillation/arrhythmia, %222222
VariablesWhole cohortTest cohortValidation cohort
(n = 1352)(n = 676)(n = 676)
Age (years)61 ± 961 ± 962 ± 9
Male gender, %515151
Diabetic, %555357
Ex-smokers, %444346
Current smokers, %161714
Alcohol drinkers, %535254
BMI (kg/m2)32 ± 832 ± 832 ± 8
CKD duration (years)3 ± 23 ± 23 ± 2
Systolic BP (mmHg)128 ± 22129 ± 22128 ± 21
Diastolic BP (mmHg)69 ± 1269 ± 1368 ± 12
On antihypertensive treatment, %979797
On statins, %666765
Haemoglobin (g/dL)12.3 ± 1.712.4 ± 1.712.2 ± 1.7
Albumin (g/dL)4.0 ± 0.44.0 ± 0.44.0 ± 0.5
Total cholesterol (mg/dL)179 ± 44179 ± 43179 ± 45
High-density lipoprotein cholesterol (mg/dL)48 ± 1548 ± 1548 ± 16
Triglycerides (mg/dL)157 ± 103155 ± 99161 ± 108
Phosphate (mg/dL)3.8 ± 0.73.8 ± 0.73.8 ± 0.7
Calcium (mg/dL)9.3 ± 0.69.3 ± 0.69.3 ± 0.5
Glucose (mg/dL)118 ± 54116 ± 52120 ± 56
eGFR (mL/min/1.73 m2)31.6 ± 8.331.5 ± 8.431.7 ± 8.3
Urine albumin:creatinine ratio (mg/g), median (IQR)132 (17–679)147 (19–690)120 (16–673)
LVMI (g/m2.7)68.4 ± 24.668.6 ± 25.768.2 ± 23.5
CRP (mg/L), median (IQR)2.7 (1.1–6.4)2.8 (1.1–6.3)2.5 (1.1–6.6)
Diabetic nephropathy, %323034
Background CV comorbidities
Myocardial infarction/revascularization, %272629
Peripheral vascular disease, %10119
Congestive HF, %131412
Stroke, %131412
Atrial fibrillation/arrhythmia, %222222

Data are expressed as mean ± SD unless stated otherwise.

Open in new tab
Table 1

Patients’ demographic, clinical and biochemical characteristics at the beginning of the follow-up

VariablesWhole cohortTest cohortValidation cohort
(n = 1352)(n = 676)(n = 676)
Age (years)61 ± 961 ± 962 ± 9
Male gender, %515151
Diabetic, %555357
Ex-smokers, %444346
Current smokers, %161714
Alcohol drinkers, %535254
BMI (kg/m2)32 ± 832 ± 832 ± 8
CKD duration (years)3 ± 23 ± 23 ± 2
Systolic BP (mmHg)128 ± 22129 ± 22128 ± 21
Diastolic BP (mmHg)69 ± 1269 ± 1368 ± 12
On antihypertensive treatment, %979797
On statins, %666765
Haemoglobin (g/dL)12.3 ± 1.712.4 ± 1.712.2 ± 1.7
Albumin (g/dL)4.0 ± 0.44.0 ± 0.44.0 ± 0.5
Total cholesterol (mg/dL)179 ± 44179 ± 43179 ± 45
High-density lipoprotein cholesterol (mg/dL)48 ± 1548 ± 1548 ± 16
Triglycerides (mg/dL)157 ± 103155 ± 99161 ± 108
Phosphate (mg/dL)3.8 ± 0.73.8 ± 0.73.8 ± 0.7
Calcium (mg/dL)9.3 ± 0.69.3 ± 0.69.3 ± 0.5
Glucose (mg/dL)118 ± 54116 ± 52120 ± 56
eGFR (mL/min/1.73 m2)31.6 ± 8.331.5 ± 8.431.7 ± 8.3
Urine albumin:creatinine ratio (mg/g), median (IQR)132 (17–679)147 (19–690)120 (16–673)
LVMI (g/m2.7)68.4 ± 24.668.6 ± 25.768.2 ± 23.5
CRP (mg/L), median (IQR)2.7 (1.1–6.4)2.8 (1.1–6.3)2.5 (1.1–6.6)
Diabetic nephropathy, %323034
Background CV comorbidities
Myocardial infarction/revascularization, %272629
Peripheral vascular disease, %10119
Congestive HF, %131412
Stroke, %131412
Atrial fibrillation/arrhythmia, %222222
VariablesWhole cohortTest cohortValidation cohort
(n = 1352)(n = 676)(n = 676)
Age (years)61 ± 961 ± 962 ± 9
Male gender, %515151
Diabetic, %555357
Ex-smokers, %444346
Current smokers, %161714
Alcohol drinkers, %535254
BMI (kg/m2)32 ± 832 ± 832 ± 8
CKD duration (years)3 ± 23 ± 23 ± 2
Systolic BP (mmHg)128 ± 22129 ± 22128 ± 21
Diastolic BP (mmHg)69 ± 1269 ± 1368 ± 12
On antihypertensive treatment, %979797
On statins, %666765
Haemoglobin (g/dL)12.3 ± 1.712.4 ± 1.712.2 ± 1.7
Albumin (g/dL)4.0 ± 0.44.0 ± 0.44.0 ± 0.5
Total cholesterol (mg/dL)179 ± 44179 ± 43179 ± 45
High-density lipoprotein cholesterol (mg/dL)48 ± 1548 ± 1548 ± 16
Triglycerides (mg/dL)157 ± 103155 ± 99161 ± 108
Phosphate (mg/dL)3.8 ± 0.73.8 ± 0.73.8 ± 0.7
Calcium (mg/dL)9.3 ± 0.69.3 ± 0.69.3 ± 0.5
Glucose (mg/dL)118 ± 54116 ± 52120 ± 56
eGFR (mL/min/1.73 m2)31.6 ± 8.331.5 ± 8.431.7 ± 8.3
Urine albumin:creatinine ratio (mg/g), median (IQR)132 (17–679)147 (19–690)120 (16–673)
LVMI (g/m2.7)68.4 ± 24.668.6 ± 25.768.2 ± 23.5
CRP (mg/L), median (IQR)2.7 (1.1–6.4)2.8 (1.1–6.3)2.5 (1.1–6.6)
Diabetic nephropathy, %323034
Background CV comorbidities
Myocardial infarction/revascularization, %272629
Peripheral vascular disease, %10119
Congestive HF, %131412
Stroke, %131412
Atrial fibrillation/arrhythmia, %222222

Data are expressed as mean ± SD unless stated otherwise.

Open in new tab

RESULTS

The source population of the CRIC study included 3520 patients (Figure 1). Among these, 223 patients were excluded because they were ≤40 years of age (i.e. they were outside the age range contemplated in the SHARP risk model), 314 patients because of a lack of information on CKD duration, 1220 because of non-applicability of the SHARP risk calculator (eGFR ≥45 mL/min/1.73 m2), 2 because of missing information on alcohol intake and 409 patients because they did not undergo echocardiography to measure LVMI. Thus 1352 patients were available for data analysis (Figure 1). The main demographic, clinical and biochemical data of the whole study population as well as of patients in the test and validation cohort (for the data analysis pertaining to HF) are given in Table 1. Their mean age was 61 years, 51% were males, 55% were diabetic, 60% were smokers and 53% were alcohol drinkers. The large majority of patients were on antihypertensive treatment (97%) and 66% were on statins. BMI was 32 kg/m2 on average and BP was 128/69 mmHg. The remaining biochemical and clinical data, including background CV comorbidities, are given in Table 1. A total of 1069 patients (79.1%) had LVH by echocardiography. As shown in Supplementary data, Figure S1, the LVMI had a positively skewed distribution with a median of 64 g/m2.7 (IQR 51–81).

The 5-year death probability estimated by the SHARP risk calculator had a positively skewed distribution, with a median value of 16.6% (IQR 10.5–27.5), while the distribution of the 10-year estimated death probability was quite normal, with a median value of 41.1% (IQR 26.1–57.7) (Supplementary data, Figure S2, upper panels). The distribution of the probability for HF by the CRIC model in both the test and the validation cohort was positively skewed (Supplementary data, Figure S2, lower panel).

LVMI was directly related to both 5- and 10-year death probabilities (both ρ = 0.31, P < 0.001) and to the risk for HF as estimated by the CRIC model in both the test (ρ = 0.45, P < 0.001) and the validation cohort (ρ = 0.42, P < 0.001) (Supplementary data, Figure S3).

Prognostic score, LVMI and all-cause mortality

During follow-up [median 7.7 years (IQR 6.5–8.7)], 327 patients died [3.4 deaths/100 persons-year (95% CI 3.0–3.7)]. On univariate Cox regression analyses (Table 2), both the SHARP scores [5-year HR (1% increase in the score) 1.044 (95% CI 1.038–1.050), P < 0.001; 10-year HR (1% increase in the score) 1.035 (95% CI 1.030–1.040), P < 0.001] and LVMI [HR (1 g/m2.7 increase) 1.019 (95% CI 1.015–1.023), P < 0.001] significantly related with the incidence rate of all-cause mortality. However, the strength of the association between the SHARP risk scores and mortality did not change when the LVMI was combined with the same scores (Table 2).

Table 2

Univariable and multivariable Cox proportional hazards models of all-cause mortality in the study population

Variables (units on increase)HR (95% CI)P-value
Univariable Cox proportional hazards models
 5-year estimated death probability (1%)1.044 (1.038–1.050) crude<0.001
 10-year estimated death probability (1%)1.035 (1.030–1.040)<0.001
 LVMI (1 g/m2.7)1.019 (1.015–1.023)<0.001
Multivariable Cox proportional hazards models
 5-year estimated death probability (1%)1.038 (1.032–1.045) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Multivariable Cox proportional hazards models
 10-year estimated death probability (1%)1.030 (1.025–1.035) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Variables (units on increase)HR (95% CI)P-value
Univariable Cox proportional hazards models
 5-year estimated death probability (1%)1.044 (1.038–1.050) crude<0.001
 10-year estimated death probability (1%)1.035 (1.030–1.040)<0.001
 LVMI (1 g/m2.7)1.019 (1.015–1.023)<0.001
Multivariable Cox proportional hazards models
 5-year estimated death probability (1%)1.038 (1.032–1.045) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Multivariable Cox proportional hazards models
 10-year estimated death probability (1%)1.030 (1.025–1.035) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Open in new tab
Table 2

Univariable and multivariable Cox proportional hazards models of all-cause mortality in the study population

Variables (units on increase)HR (95% CI)P-value
Univariable Cox proportional hazards models
 5-year estimated death probability (1%)1.044 (1.038–1.050) crude<0.001
 10-year estimated death probability (1%)1.035 (1.030–1.040)<0.001
 LVMI (1 g/m2.7)1.019 (1.015–1.023)<0.001
Multivariable Cox proportional hazards models
 5-year estimated death probability (1%)1.038 (1.032–1.045) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Multivariable Cox proportional hazards models
 10-year estimated death probability (1%)1.030 (1.025–1.035) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Variables (units on increase)HR (95% CI)P-value
Univariable Cox proportional hazards models
 5-year estimated death probability (1%)1.044 (1.038–1.050) crude<0.001
 10-year estimated death probability (1%)1.035 (1.030–1.040)<0.001
 LVMI (1 g/m2.7)1.019 (1.015–1.023)<0.001
Multivariable Cox proportional hazards models
 5-year estimated death probability (1%)1.038 (1.032–1.045) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Multivariable Cox proportional hazards models
 10-year estimated death probability (1%)1.030 (1.025–1.035) adjusted<0.001
 LVMI (1 g/m2.7)1.014 (1.010–1.018)<0.001
Open in new tab

The discriminatory powers of the 5- and 10-year SHARP risk scores for mortality (HC index) were identical, 71% and 71%, respectively, and higher than that provided by LVMI (66%) alone (Figure 2). The introduction of LVMI into the two Cox analyses including these two risk scores only provided a very modest increase (2%) in the discriminatory power of the same models (Figure 2).

FIGURE 2
Discrimination and risk calibration abilities of prognostic estimates for death derived from the SHARP calculator, LVMI and both.
Open in new tabDownload slide

Discrimination and risk calibration abilities of prognostic estimates for death derived from the SHARP calculator, LVMI and both.

Of note, by simultaneously including LVMI, LV ejection fraction and E/A ratio into the same prognostic model, this expanded model provided an HC (67.6%) for mortality that was quite similar to that of the model based on LVMI alone (66.0%), indicating that the inclusion of LV functional parameters does not improve the prognostic value of LVMI. Similarly, the inclusion of LV systolic and diastolic function in the LVMI-based model for predicting HF produced a modest increase in the discriminatory power both in the test (HC, from 72% to 75%) and in the validation cohort (HC, from 71% to 73%), i.e. figures not superior to those provided by the SHARP risk calculator.

The additional prognostic value of LVMI as assessed by the IDI confirmed that this biomarker provides just a modest increase in prognostic accuracy over the two SHARP risk scores (model based on 5-year estimated probability of death: IDI +3.1%, P < 0.001; model based on 10-year estimated probability of death: IDI +3.0%, P < 0.001).

In calibration analyses, the unadjusted Cox models based on 5- or 10-year estimated death probability (χ2 = 15.8, P = 0.001 and χ2 = 10.3, P = 0.02, respectively) or LVMI (χ2 = 9.5, P = 0.02) produced prognostic estimates for all-cause mortality that significantly differed from the observed probabilities of the same outcome denoting poor calibration (Figure 2) and the inclusion of LVMI into the two Cox analyses based on the two risk scores did not materially improve the calibration of the same models, which remained rather unsatisfactory (Figure 2).

LVMI and the risk for HF

During the follow-up period, 208 incident cases of HF were collected: 100 cases occurred in the test cohort and 108 cases in the validation cohort. The number of deaths was 107 in the test cohort and 122 in the validation cohort. On Cox regression analyses, taking into account the competitive risk of mortality (Fine and Gray approach), LVMI was significantly related to the incidence rate of HF both in the test and in the validation cohort (Tables 3 and 4). The predictive model developed in the test cohort (Table 3) and implemented in the validation cohort (Table 4) included age, diabetes, BMI, systolic BP, haemoglobin, albumin, triglycerides, phosphate, eGFR, CV comorbidities and history of atrial fibrillation (i.e. the variables that were significantly related with the risk of HF at univariate Cox regression analyses—CRIC model). Such a model provided a discriminatory power for predicting HF of 79% in the test cohort and of 74% in the validation cohort, figures higher than those provided by LVMI alone in the two cohorts (72 and 71%) (Figure 3). The addition of LVMI to the CRIC model provided just a modest increase in the discriminatory power of the same model both in the test (from 79% to 81%, +2%) and in the validation cohort (from 74% to 77%, +3%) (Figure 3). The additional prognostic value of LVMI as assessed by the IDI confirmed a modest increase in prognostic accuracy by this biomarker over the CRIC model in the test (+4.2%, P < 0.001) and in the validation cohort (+4.8%, P < 0.001). In the test cohort, the CRIC and the LVMI-based models and even more so the CRIC model + LVMI provided prognostic estimates that substantially differed (P ranging from 0.006 to 0.09) from the observed risk, indicating poor calibration. However, in the validation cohort, the three models were all calibrated fairly well (Figure 3).

FIGURE 3
Discrimination and risk calibration abilities of prognostic estimates for HF derived from the CRIC model, LVMI and both.
Open in new tabDownload slide

Discrimination and risk calibration abilities of prognostic estimates for HF derived from the CRIC model, LVMI and both.

Table 3

Subdistribution hazards models of incident HF (n = 100) in the test cohort (n = 676)

Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.020 (1.015–1.026);  < 0.0011.018 (1.010–1.026);  < 0.001
Age (1 year)1.02 (0.99–1.05); 0.07…1.03 (1.01–1.05); 0.04
Diabetes (0 = no, 1 = yes)2.32(1.30–4.16); 0.004…2.43 (1.33–4.44); 0.004
BMI (1 kg/m2)1.00 (0.98–1.03); 0.70…0.98 (0.95–1.01); 0.27
Systolic BP (1 mmHg)1.01 (1.00–1.02); 0.05…1.01 (0.99–1.02); 0.24
Haemoglobin (1 g/dL)0.90 (0.78–1.04); 0.17…0.93 (0.80–1.08); 0.33
Albumin (1 g/dL)1.05 (0.63–1.77); 0.84…1.12 (0.69–1.83); 0.65
Triglycerides (1 mg/dL)1.00 (0.99–1.01); 0.12…1.00 (0.99–1.01); 0.10
Phosphate (1 mg/dL)1.12 (0.89–1.40); 0.35…1.15 (0.89–1.49); 0.30
eGFR (1 mL/min/1.73 m2)0.97 (0.94–0.99); 0.01…0.97 (0.95–0.99); 0.04
CV comorbidities (0 = no, 1 = yes)2.31 (1.48–3.60); 0.001…2.08 (1.33–3.25); 0.001
Atrial fibrillation (0 = no, 1 = yes)1.71 (1.09–2.67); 0.02…1.60 (1.01–2.52); 0.04
Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.020 (1.015–1.026);  < 0.0011.018 (1.010–1.026);  < 0.001
Age (1 year)1.02 (0.99–1.05); 0.07…1.03 (1.01–1.05); 0.04
Diabetes (0 = no, 1 = yes)2.32(1.30–4.16); 0.004…2.43 (1.33–4.44); 0.004
BMI (1 kg/m2)1.00 (0.98–1.03); 0.70…0.98 (0.95–1.01); 0.27
Systolic BP (1 mmHg)1.01 (1.00–1.02); 0.05…1.01 (0.99–1.02); 0.24
Haemoglobin (1 g/dL)0.90 (0.78–1.04); 0.17…0.93 (0.80–1.08); 0.33
Albumin (1 g/dL)1.05 (0.63–1.77); 0.84…1.12 (0.69–1.83); 0.65
Triglycerides (1 mg/dL)1.00 (0.99–1.01); 0.12…1.00 (0.99–1.01); 0.10
Phosphate (1 mg/dL)1.12 (0.89–1.40); 0.35…1.15 (0.89–1.49); 0.30
eGFR (1 mL/min/1.73 m2)0.97 (0.94–0.99); 0.01…0.97 (0.95–0.99); 0.04
CV comorbidities (0 = no, 1 = yes)2.31 (1.48–3.60); 0.001…2.08 (1.33–3.25); 0.001
Atrial fibrillation (0 = no, 1 = yes)1.71 (1.09–2.67); 0.02…1.60 (1.01–2.52); 0.04

Second column: multiple Cox model including all univariate correlates of HF but LVMI; third column: univariate Cox model including LVMI; fourth column: multiple Cox model including all univariate correlates of HF and LVMI.

Open in new tab
Table 3

Subdistribution hazards models of incident HF (n = 100) in the test cohort (n = 676)

Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.020 (1.015–1.026);  < 0.0011.018 (1.010–1.026);  < 0.001
Age (1 year)1.02 (0.99–1.05); 0.07…1.03 (1.01–1.05); 0.04
Diabetes (0 = no, 1 = yes)2.32(1.30–4.16); 0.004…2.43 (1.33–4.44); 0.004
BMI (1 kg/m2)1.00 (0.98–1.03); 0.70…0.98 (0.95–1.01); 0.27
Systolic BP (1 mmHg)1.01 (1.00–1.02); 0.05…1.01 (0.99–1.02); 0.24
Haemoglobin (1 g/dL)0.90 (0.78–1.04); 0.17…0.93 (0.80–1.08); 0.33
Albumin (1 g/dL)1.05 (0.63–1.77); 0.84…1.12 (0.69–1.83); 0.65
Triglycerides (1 mg/dL)1.00 (0.99–1.01); 0.12…1.00 (0.99–1.01); 0.10
Phosphate (1 mg/dL)1.12 (0.89–1.40); 0.35…1.15 (0.89–1.49); 0.30
eGFR (1 mL/min/1.73 m2)0.97 (0.94–0.99); 0.01…0.97 (0.95–0.99); 0.04
CV comorbidities (0 = no, 1 = yes)2.31 (1.48–3.60); 0.001…2.08 (1.33–3.25); 0.001
Atrial fibrillation (0 = no, 1 = yes)1.71 (1.09–2.67); 0.02…1.60 (1.01–2.52); 0.04
Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.020 (1.015–1.026);  < 0.0011.018 (1.010–1.026);  < 0.001
Age (1 year)1.02 (0.99–1.05); 0.07…1.03 (1.01–1.05); 0.04
Diabetes (0 = no, 1 = yes)2.32(1.30–4.16); 0.004…2.43 (1.33–4.44); 0.004
BMI (1 kg/m2)1.00 (0.98–1.03); 0.70…0.98 (0.95–1.01); 0.27
Systolic BP (1 mmHg)1.01 (1.00–1.02); 0.05…1.01 (0.99–1.02); 0.24
Haemoglobin (1 g/dL)0.90 (0.78–1.04); 0.17…0.93 (0.80–1.08); 0.33
Albumin (1 g/dL)1.05 (0.63–1.77); 0.84…1.12 (0.69–1.83); 0.65
Triglycerides (1 mg/dL)1.00 (0.99–1.01); 0.12…1.00 (0.99–1.01); 0.10
Phosphate (1 mg/dL)1.12 (0.89–1.40); 0.35…1.15 (0.89–1.49); 0.30
eGFR (1 mL/min/1.73 m2)0.97 (0.94–0.99); 0.01…0.97 (0.95–0.99); 0.04
CV comorbidities (0 = no, 1 = yes)2.31 (1.48–3.60); 0.001…2.08 (1.33–3.25); 0.001
Atrial fibrillation (0 = no, 1 = yes)1.71 (1.09–2.67); 0.02…1.60 (1.01–2.52); 0.04

Second column: multiple Cox model including all univariate correlates of HF but LVMI; third column: univariate Cox model including LVMI; fourth column: multiple Cox model including all univariate correlates of HF and LVMI.

Open in new tab
Table 4

Subdistribution hazards models of incident HF (n = 108) in the validation cohort (n = 676)

Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.024 (1.017–1.031); < 0.0011.023 (1.014–1.032); < 0.001
Age (1 year)1.02 (0.99–1.04); 0.17…1.01 (0.99–1.04); 0.34
Diabetes (0 = no, 1 = yes)1.84 (1.12–3.03); 0.02…1.75 (1.05–2.91); 0.03
BMI (1 kg/m2)1.00 (0.98–1.04); 0.57…0.98 (0.95–1.01); 0.17
Systolic BP (1 mmHg)1.00 (0.99–1.01); 0.70…0.99 (0.98–1.00); 0.13
Haemoglobin (1 g/dL)1.04 (0.90–1.21); 0.61…1.00 (0.87–1.17); 0.95
Albumin (1 g/dL)0.76 (0.44–1.30); 0.31…0.76 (0.44–1.29); 0.30
Triglycerides (1 mg/dL)0.99 (0.98–0.99); 0.008…0.99 (0.98–0.99); 0.006
Phosphate (1 mg/dL)1.11 (0.83–1.48); 0.47…1.13 (0.84–1.52); 0.43
eGFR (1 mL/min/1.73 m2)0.96 (0.93–0.98); 0.001…0.96 (0.94–0.99); 0.006
CV comorbidities (0 = no, 1 = yes)1.94 (1.29–2.92); 0.001…1.72 (1.14–2.58); 0.01
Atrial fibrillation (0 = no, 1 = yes)1.99 (1.30–3.07); 0.002…1.75 (1.13–2.72); 0.01
Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.024 (1.017–1.031); < 0.0011.023 (1.014–1.032); < 0.001
Age (1 year)1.02 (0.99–1.04); 0.17…1.01 (0.99–1.04); 0.34
Diabetes (0 = no, 1 = yes)1.84 (1.12–3.03); 0.02…1.75 (1.05–2.91); 0.03
BMI (1 kg/m2)1.00 (0.98–1.04); 0.57…0.98 (0.95–1.01); 0.17
Systolic BP (1 mmHg)1.00 (0.99–1.01); 0.70…0.99 (0.98–1.00); 0.13
Haemoglobin (1 g/dL)1.04 (0.90–1.21); 0.61…1.00 (0.87–1.17); 0.95
Albumin (1 g/dL)0.76 (0.44–1.30); 0.31…0.76 (0.44–1.29); 0.30
Triglycerides (1 mg/dL)0.99 (0.98–0.99); 0.008…0.99 (0.98–0.99); 0.006
Phosphate (1 mg/dL)1.11 (0.83–1.48); 0.47…1.13 (0.84–1.52); 0.43
eGFR (1 mL/min/1.73 m2)0.96 (0.93–0.98); 0.001…0.96 (0.94–0.99); 0.006
CV comorbidities (0 = no, 1 = yes)1.94 (1.29–2.92); 0.001…1.72 (1.14–2.58); 0.01
Atrial fibrillation (0 = no, 1 = yes)1.99 (1.30–3.07); 0.002…1.75 (1.13–2.72); 0.01

Second column: multiple Cox model including all univariate correlates of HF but LVMI; third column: univariate Cox model including LVMI; fourth column: multiple Cox model including all univariate correlates of HF and LVMI.

Open in new tab
Table 4

Subdistribution hazards models of incident HF (n = 108) in the validation cohort (n = 676)

Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.024 (1.017–1.031); < 0.0011.023 (1.014–1.032); < 0.001
Age (1 year)1.02 (0.99–1.04); 0.17…1.01 (0.99–1.04); 0.34
Diabetes (0 = no, 1 = yes)1.84 (1.12–3.03); 0.02…1.75 (1.05–2.91); 0.03
BMI (1 kg/m2)1.00 (0.98–1.04); 0.57…0.98 (0.95–1.01); 0.17
Systolic BP (1 mmHg)1.00 (0.99–1.01); 0.70…0.99 (0.98–1.00); 0.13
Haemoglobin (1 g/dL)1.04 (0.90–1.21); 0.61…1.00 (0.87–1.17); 0.95
Albumin (1 g/dL)0.76 (0.44–1.30); 0.31…0.76 (0.44–1.29); 0.30
Triglycerides (1 mg/dL)0.99 (0.98–0.99); 0.008…0.99 (0.98–0.99); 0.006
Phosphate (1 mg/dL)1.11 (0.83–1.48); 0.47…1.13 (0.84–1.52); 0.43
eGFR (1 mL/min/1.73 m2)0.96 (0.93–0.98); 0.001…0.96 (0.94–0.99); 0.006
CV comorbidities (0 = no, 1 = yes)1.94 (1.29–2.92); 0.001…1.72 (1.14–2.58); 0.01
Atrial fibrillation (0 = no, 1 = yes)1.99 (1.30–3.07); 0.002…1.75 (1.13–2.72); 0.01
Variables (units of increase)SHR (95% CI); P-valueSHR (95% CI); P-valueSHR (95% CI); P-value
LVMI (1 g/m2.7)…1.024 (1.017–1.031); < 0.0011.023 (1.014–1.032); < 0.001
Age (1 year)1.02 (0.99–1.04); 0.17…1.01 (0.99–1.04); 0.34
Diabetes (0 = no, 1 = yes)1.84 (1.12–3.03); 0.02…1.75 (1.05–2.91); 0.03
BMI (1 kg/m2)1.00 (0.98–1.04); 0.57…0.98 (0.95–1.01); 0.17
Systolic BP (1 mmHg)1.00 (0.99–1.01); 0.70…0.99 (0.98–1.00); 0.13
Haemoglobin (1 g/dL)1.04 (0.90–1.21); 0.61…1.00 (0.87–1.17); 0.95
Albumin (1 g/dL)0.76 (0.44–1.30); 0.31…0.76 (0.44–1.29); 0.30
Triglycerides (1 mg/dL)0.99 (0.98–0.99); 0.008…0.99 (0.98–0.99); 0.006
Phosphate (1 mg/dL)1.11 (0.83–1.48); 0.47…1.13 (0.84–1.52); 0.43
eGFR (1 mL/min/1.73 m2)0.96 (0.93–0.98); 0.001…0.96 (0.94–0.99); 0.006
CV comorbidities (0 = no, 1 = yes)1.94 (1.29–2.92); 0.001…1.72 (1.14–2.58); 0.01
Atrial fibrillation (0 = no, 1 = yes)1.99 (1.30–3.07); 0.002…1.75 (1.13–2.72); 0.01

Second column: multiple Cox model including all univariate correlates of HF but LVMI; third column: univariate Cox model including LVMI; fourth column: multiple Cox model including all univariate correlates of HF and LVMI.

Open in new tab

DISCUSSION

In this study, in a large population of Stages G3b–G5 CKD patients of the CRIC, LVMI was independently associated with all-cause mortality and de novo HF but did not meaningfully improve risk prediction for these two conditions when added to clinical models. These findings suggest that measurement of LVMI by echocardiography provides a prognostic power for all-cause mortality inferior to that by a risk score based on easily available clinical information like the SHARP study risk calculator. When combined with the SHARP study score, LVMI only to a very modest degree improved the prediction of mortality by various prognostic tests, including calibration, risk reclassification and the integrated discrimination index. Overall, our analyses in the CRIC suggest that the measurement of LVMI by echocardiography solely for risk stratification in pre-dialysis CKD patients fails to provide meaningful additional prognostic value to the SHARP study risk score.

LVH is a common alteration in CKD and up to 79% of patients in the CRIC exhibited such an alteration. Previous studies in this cohort confirmed that LVMI is a powerful risk factor for all-cause mortality and progression to HF in CKD [6], supporting a causal role of LVH in the high CV risk in this population. Due to the causal role of LVH in the high risk for HF and other adverse CV outcomes, the value of LVMI for prognosis and risk stratification has long been taken for granted in studies in the general population where the analysis was based on Cox regression analysis only. These include the classic, time-honoured community study based on the Framingham cohort [22], studies in hypertensive patients [23–25], patients with coronary heart disease [26] or in the elderly population [27] and studies focusing on sudden death [28, 29]. Similarly, clinical studies focusing on LVMI in the pre-dialysis CKD population investigated the prognostic power of this parameter solely by Cox regression analysis [30, 31]. However, established causal risk factors like LVH or other major risk factors like hypercholesterolaemia are not necessarily valid prognostic factors. At the individual patient level, the prognostic power of hypercholesterolaemia (top quintile) for CV death is weak because only 15% of those who would later die of ischaemic heart disease would be identified on the basis of serum cholesterol levels [32]. By the same token, confirming previous analyses in the CRIC database [6], we found that LVMI is a highly significant predictor of all-cause death and incident HF by Cox regression analysis. However, we also found that LVMI holds very modest additional discriminant power for the identification of patients who die or develop HF during follow-up. Thus established risk factors for death and HF in the CKD population may have just weak or no predictive power for the same outcomes at the patient level. Studies aimed at establishing the value of prognostic biomarkers ought to be based on discriminant analysis and other prognostic analyses centred on patients rather than at a population level. In this study in a large CKD cohort, LVMI per se had a discriminatory power (HC index) for all-cause mortality and HF inferior to that provided by the SHARP risk score and the CRIC model and did not add clinically relevant discriminant power to that by the same risk models. Furthermore, LVMI failed to ameliorate the calibration of models for the risk of mortality and HF. Thus, in keeping with studies in the end-stage kidney disease population [13], LVMI conveys quite limited prognostic information in pre-dialysis CKD patients as well.

The low prognostic power of LVMI notwithstanding, it is important to stress that findings in this study are immaterial to the established value of echocardiographic studies in the CKD population. In the current guidelines, the presence of LVH and/or other alterations in LV mass and function represent a fundamental element for the management of patients with suspected HF [33]. However, as far as risk stratification is concerned, LVMI is outperformed by simple clinical risk scores for the risk of death like the SHARP study score and by the CRIC model for HF. Therefore nephrologists may better define the risk of death and HF by these scores rather than by measuring LVMI.

The main limitation of this study is the fact that more than half of the CRIC patients did not qualify for inclusion in the analysis and that we adopted a prognostic model for HF that we can only cross-validate within the CRIC. This model had no external validation. Thus the issue of whether LVMI adds prognostic model to the CRIC model for HF demands further testing in cohorts different from the CRIC. LVH in CKD has complex pathophysiology and needs to be interpreted in the functional scenario of the CV system rather than per se [34].

In conclusion, LVMI is inferior to well-validated clinical risk scores for the prediction of mortality in CKD. When combined with the SHARP score, LVMI improves the risk discrimination ability of this score quite modestly and fails to meaningfully improve risk calibration and reclassification. Thus it appears unlikely that the measurement of LVMI will improve stratification by risk models based on easily available clinical variables in CKD patients.

SUPPLEMENTARY DATA

Supplementary data are available at ndt online.

ACKNOWLEDGEMENTS

This study is part of the research activity of the European Cardiovascular and Renal Medicine (EURECA-m) working group of the ERA-EDTA.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES

1

Martin
FL
,
McKie
PM
,
Cataliotti
A.
Experimental mild renal insufficiency mediates early cardiac apoptosis, fibrosis, and diastolic dysfunction: a kidney-heart connection
.
Am J Physiol Integr Comp Physiol
2012
;
302
:
R292
–
R299

Google Scholar

Crossref
Search ADS

 
2

Amann
K
,
Wiest
G
,
Zimmer
G
 et al. 
Reduced capillary density in the myocardium of uremic rats—a stereological study
.
Kidney Int
1992
;
42
:
1079
–
1085

Google Scholar

Crossref
Search ADS
PubMed

 
3

Moody
WE
,
Ferro
CJ
,
Edwards
NC
 et al. 
Cardiovascular effects of unilateral nephrectomy in living kidney donors
.
Hypertension
2016
;
67
:
368
–
377

Google Scholar

Crossref
Search ADS
PubMed

 
4

Altmann
U
,
Böger
CA
,
Farkas
S
 et al. 
Effects of reduced kidney function because of living kidney donation on left ventricular mass novelty and significance
.
Hypertension
2017
;
69
:
297
–
303

Google Scholar

Crossref
Search ADS
PubMed

 
5

Park
M
,
Hsu
CY
,
Li
Y
 et al. 
Associations between kidney function and subclinical cardiac abnormalities in CKD
.
J Am Soc Nephrol
2012
;
23
:
1725
–
1734

Google Scholar

Crossref
Search ADS
PubMed

 
6

Dubin
RF
,
Deo
R
,
Bansal
N
 et al. 
Associations of conventional echocardiographic measures with incident heart failure and mortality: the chronic renal insufficiency cohort
.
Clin J Am Soc Nephrol
2017
;
12
:
60–68

Google Scholar

Crossref
Search ADS
PubMed

 
7

Bo
M.
Practice guidelines for the management of arterial hypertension of the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC): ESH/ESC task force for the management of arterial hypertension
.
J Hypertens
2013
;
31
:
1925
–
1938

Google Scholar

PubMed
OpenURL Placeholder Text

 
8

Girerd
N
,
Huttin
O
,
Zannad
F.
Echocardiography in hypertension in the era of risk stratification and personalized medicine
.
Am J Hypertens
2016
;
29
:
792
–
795

Google Scholar

Crossref
Search ADS
PubMed

 
9

Glassock
RJ
,
Pecoits-Filho
R
,
Barbareto
S.
Increased left ventricular mass in chronic kidney disease and end-stage renal disease: what are the implications?
Dial Transplant
2010
;
39
:
16
–
19

Google Scholar

Crossref
Search ADS

 
10

Harrell
FE
,
Lee
KL
,
Mark
DB.
Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors
.
Stat Med
1996
;
15
:
361
–
387

Google Scholar

Crossref
Search ADS
PubMed

 
11

Pencina
MJ
,
D'Agostino
RB
,
Steyerberg
EW
 et al. 
Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers
.
Stat Med
2011
;
30
:
11
–
21

Google Scholar

Crossref
Search ADS
PubMed

 
12

Crowson
CS
,
Atkinson
EJ
,
Therneau
TM.
Assessing calibration of prognostic risk scores
.
Stat Methods Med Res
2016
;
25
:
1692
–
1706

Google Scholar

Crossref
Search ADS
PubMed

 
13

Tripepi
G
,
Pannier
B
,
D’Arrigo
G
 et al. 
Reappraisal in two European cohorts of the prognostic power of left ventricular mass index in chronic kidney failure
.
Kidney Int
2017
;
91
:
704
–
710

Google Scholar

Crossref
Search ADS
PubMed

 
14

Cook
NR.
Use and misuse of the receiver operating characteristic curve in risk prediction
.
Circulation
2007
;
115
:
928
–
935

Google Scholar

Crossref
Search ADS
PubMed

 
15

Schlackow
I
,
Kent
S
,
Herrington
W
 et al. 
A policy model of cardiovascular disease in moderate-to-advanced chronic kidney disease
.
Heart
2017
;
103
:
1880
–
1890

Google Scholar

Crossref
Search ADS
PubMed

 
16

Feldman
HI.
The Chronic Renal Insufficiency Cohort (CRIC) study: design and methods
.
J Am Soc Nephrol
2003
;
14
(7 Suppl 2):
148S
–
1153

Google Scholar

Crossref
Search ADS

 
17

Lash
JP
,
Go
AS
,
Appel
LJ
 et al. 
Chronic renal insufficiency cohort (CRIC) study: baseline characteristics and associations with kidney function
.
Clin J Am Soc Nephrol
2009
;
4
:
1302
–
1311

Google Scholar

Crossref
Search ADS
PubMed

 
18

Lang
RM
,
Bierig
M
,
Devereux
RB
 et al. 
Recommendations for chamber quantification: a report from the American Society of Echocardiography’s guidelines and standards committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiograph
.
J Am Soc Echocardiogr
2005
;
18
:
1440
–
1463

Google Scholar

Crossref
Search ADS
PubMed

 
19

Anderson
AH
,
Yang
W
,
Hsu
C
 et al. 
Estimating GFR among participants in the Chronic Renal Insufficiency Cohort (CRIC) study
.
Am J Kidney Dis
2012
;
60
:
250
–
261

Google Scholar

Crossref
Search ADS
PubMed

 
20

He
J
,
Shlipak
M
,
Anderson
A
 et al. 
Risk factors for heart failure in patients with chronic kidney disease: the CRIC (Chronic Renal Insufficiency Cohort) study
.
J Am Heart Assoc
2017
;
6
:
e005336

Google Scholar

Crossref
Search ADS
PubMed

 
21

Fine
JP
,
Gray
RJ.
A proportional hazards model for the subdistribution of a competing risk
.
J Am Stat Assoc
1999
;
94
:
496
–
509

Google Scholar

Crossref
Search ADS

 
22

Levy
D
,
Garrison
RJ
,
Savage
DD
 et al. 
Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study
.
N Engl J Med
1990
;
322
:
1561
–
1566

Google Scholar

Crossref
Search ADS
PubMed

 
23

Koren
MJ
,
Devereux
RB
,
Casale
PN
 et al. 
Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension
.
Ann Intern Med
1991
;
114
:
345
–
352

Google Scholar

Crossref
Search ADS
PubMed

 
24

Bombelli
M
,
Facchetti
R
,
Carugo
S
 et al. 
Left ventricular hypertrophy increases cardiovascular risk independently of in-office and out-of-office blood pressure values
.
J. Hypertens
2009
;
27
:
2458
–
2464

Google Scholar

Crossref
Search ADS
PubMed

 
25

Schillaci
G
,
Verdecchia
P
,
Porcellati
C
 et al. 
Continuous relation between left ventricular mass and cardiovascular risk in essential hypertension
.
Hypertension
2000
;
35
:
580
–
586

Google Scholar

Crossref
Search ADS
PubMed

 
26

Ghali
JK
,
Liao
Y
,
Simmons
B
 et al. 
The prognostic role of left ventricular hypertrophy in patients with or without coronary artery disease
.
Ann Intern Med
1992
;
117
:
831
–
836

Google Scholar

Crossref
Search ADS
PubMed

 
27

Tsang
TSM
,
Barnes
ME
,
Gersh
BJ
 et al. 
Prediction of risk for first age-related cardiovascular events in an elderly population: the incremental value of echocardiography
.
J Am Coll Cardiol
2003
;
42
:
1199
–
1205

Google Scholar

Crossref
Search ADS
PubMed

 
28

Haider
AW
,
Larson
MG
,
Benjamin
EJ
 et al. 
Increased left ventricular mass and hypertrophy are associated with increased risk for sudden death
.
J Am Coll Cardiol
1998
;
32
:
1454
–
1459

Google Scholar

Crossref
Search ADS
PubMed

 
29

Reinier
K
,
Dervan
C
,
Singh
T
 et al. 
Increased left ventricular mass and decreased left ventricular systolic function have independent pathways to ventricular arrhythmogenesis in coronary artery disease
.
Hear Rhythm
2011
;
8
:
1177
–
1182

Google Scholar

Crossref
Search ADS

 
30

Eckardt
KU
,
Scherhag
A
,
Macdougall
IC
 et al. 
Left ventricular geometry predicts cardiovascular outcomes associated with anemia correction in CKD
.
J Am Soc Nephrol
2009
;
20
:
2651
–
2660

Google Scholar

Crossref
Search ADS
PubMed

 
31

Paoletti
E
,
De Nicola
L
,
Gabbai
FB
 et al. 
Associations of left ventricular hypertrophy and geometry with adverse outcomes in patients with CKD and hypertension
.
Clin J Am Soc Nephrol
2016
;
11
:
271
–
279

Google Scholar

Crossref
Search ADS
PubMed

 
32

Wald
N
,
Watt
H
,
Bailey
A
 et al. 
Apolipoproteins and ischaemic heart disease: implications for screening
.
Lancet
1994
;
343
:
75
–
79

Google Scholar

Crossref
Search ADS
PubMed

 
33

Yancy
CW
,
Jessup
M
,
Bozkurt
B
 et al. 
2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines
.
Circulation
2013
;
128
:
240
–
327

Google Scholar

OpenURL Placeholder Text

 
34

Zoccali
C
,
Bolignano
D
,
Mallamaci
F.
 Left ventricular hypertrophy in chronic kidney disease. In:
Turner
NN
,
Lameire
N
,
Goldsmith
DJ
 et al.  (eds).
Oxford Textbook of Nephrology
, 4th edn.
Oxford
:
Oxford University Press
,
2015
,
837
–
852

Google Scholar

Crossref
Search ADS

 
© The Author(s) 2019. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
Topic:
Issue Section:
Clinical Research
Download all slides
  • Supplementary data

    • Supplementary data

    gfz254_Supplementary_Data - zip file

    Comments

    0 Comments
    Comments (0)
    Submit a comment
    You have entered an invalid code
    Thank you for submitting a comment on this article. Your comment will be reviewed and published at the journal's discretion. Please check for further notifications by email.