Epidemiology
Mildly Decreased Renal Function and Its Relation to Left Ventricular Geometry Change
2019 Volume 83 Issue 11 Pages 2236-2241
Details
2019 Volume 83 Issue 11 Pages 2236-2241
Background: It is believed that left ventricular (LV) geometry change contributes to the elevated cardiovascular risk in patients with chronic kidney disease. However, data are less available on the association between LV geometry change and mildly decreased renal function within estimated glomerular filtration rate (eGFR) from 60 to 89 (mL/min/1.73 m2).
Methods and Results: In a cohort of 47,730 Koreans undergoing echocardiography as part of a health check-up, we evaluated the association of LV hypertrophy (LVH) and abnormal relative wall thickness (RWT) with 4 levels and 3 levels of eGFR in men (≥90, 89.99–80, 79.99–70, 69.99–60) and women (≥90, 89.99–80, 79.99–60), respectively. Multivariate logistic regression analysis was used to calculate the odds ratios (OR) and 95% confidence intervals (CI) for LVH and abnormal RWT, adjusting for conventional cardiovascular risk factors (adjusted OR [95% CI]). In the fully adjustment model, men did not show a significant association between LVH and levels of eGFR between 60 and 89. However, abnormal RWT was significantly associated with the levels of eGFR between 60 and 89. Women did not show a significant association of LVH and abnormal RWT with levels of eGFR between 60 and 89.
Conclusions: Men with mildly decreased renal function (eGFR between 60 and 89 mL/min/1.73 m2) had increased probability of LV geometry change represented by abnormal RWT.
Chronic kidney disease (CKD) is an independent risk factor for cardiovascular disease (CVD).1 The estimated glomerular filtration rate (eGFR) is commonly used to determine the stage of CKD as an index of renal function and it is evident that a severe decrease in eGFR <60 (mL/min/1.73 m2) accompanies an elevated risk of CVD.1 However, the association of cardiovascular risk with mildly decreased eGFR of 60–90 mL/min/1.73 m2 is equivocal. Some studies show increased risk of CVD in individuals with near normal eGFR,2,3 but others have failed to find a significant association.4,5
Abnormal left ventricular (LV) geometry change is frequently observed in CKD patients. The LV mass index (LVMI) and relative wall thickness (RWT) are commonly used to identify abnormal LV geometry and the guideline6 defines 3 forms: concentric remodeling (abnormal RWT and normal LVMI); concentric hypertrophy (abnormal RWT and LVMI); and eccentric LV hypertrophy ([LVH] abnormal LVMI and normal RWT). Increased LVMI characterizes LVH, which affects 15–21% of the general population, and 50–70% of patients with intermediate-stage CKD.7 LVH is associated with fatal cardiovascular outcomes in CKD patients.8 Although the LVMI is widely used to assess cardiac damage, including LVH, cardiac damage can be already present in patients with normal LV mass. One study has indicated that concentric remodeling detected by abnormal RWT but with normal LV mass is an early form of cardiac adaptation to high blood pressure.9 Concentric remodeling has a higher risk of cardiovascular death than normal geometry, and RWT is associated with increased risk of cardiovascular death independently of LVMI.10 Thus, it is expected that LVH and abnormal RWT are early pathologic changes in the heart of CKD patients. However, it is inconclusive when the LV geometry change occurs in the course of CKD progression. In particular, data are less available regarding the prevalence and risk of LVH and abnormal RWT in individuals with eGFR of 90–60 mL/min/1.73 m2.
Using data from a cohort of 47,730 Koreans undergoing echocardiography as part of a health check-up, we investigated the association of LVH and abnormal RWT with levels of eGFR ≥60 mL/min/1.73 m2.
Relevant clinical and echocardiographic data were obtained from the Kangbuk Samsung Health Study (KSHS), which is a cohort study to investigate the medical data of Koreans who have had a medical health check-up in Kangbuk Samsung Hospital. Korea’s Industrial Safety and Health law regulates that all Korean employees should receive a medical health check-up annually or biennially. According to this law, Korean companies contract with hospitals to ensure their employees and their spouses receive health check-ups. Thus, all the study participants were comprised of men and women undergoing health check-ups at Kangbuk Samsun Hospital, and most of them have undergone the health check-up without specific symptoms or medical problems.
Among the participants in the KSHS, we initially enrolled 55,214 subjects who had undergone echocardiography at least once between January 2013 and December 2014. Of these 55,214, we excluded subjects for the following reasons: 161 diagnosed with arrhythmia such as atrial fibrillation; 58 with systolic LV dysfunction (ejection fraction (EF) ≤50%); 1,448 with a history of cancer; 1,017 with serious medical conditions such as chronic obstructive pulmonary disease; 386 with a history of CVD including myocardial infarction or angina; 4,108 with missing data for LV dimensions, eGFR and other covariates (e.g., alcohol intake); 187 with eGFR <60 mL/min/1.73 m2 or microalbuminuria/albuminuria. Finally, the total number of eligible participants was 47,730. Ethics approvals for the study protocol and analysis of the data were obtained from the Review Board of Kangbuk Samsung Hospital. Informed consent was exempted because we only assessed retrospective data without personal identifying information.
Clinical and Laboratory MeasurementsStudy data included medical history assessed by self-administered questionnaire, anthropometric measurements and laboratory measurements. All study participants were asked to respond to a health-related behavior questionnaire, which included the topics of alcohol consumption, smoking and exercise. Diabetes mellitus (DM) was defined as fasting serum glucose level ≥126 mg/dL, or serum hemoglobin A1c (HbA1c) ≥6.5%, or a prior diagnosis of DM.11 Hypertension was defined as a prior diagnosis of hypertension or measured BP ≥140/90 mmHg at initial and follow-up examinations.
Blood samples were collected from an antecubital vein after more than 12 h of fasting. The fasting serum glucose was measured using the hexokinase method, and HbA1c was measured using an immunoturbidimetric assay with a Cobra Integra 800 automatic analyzer (Roche Diagnostics, Basel, Switzerland). Total cholesterol and triglyceride were measured using enzymatic colorimetric tests, low-density lipoprotein cholesterol (LDL-C) was measured using a homogeneous enzymatic colorimetric test, and high-density lipoprotein cholesterol (HDL-C) was measured using a selective inhibition method (Advia 1650 Autoanalyzer, Bayer Diagnostics; Leverkusen, Germany). The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was used to calculate eGFR.12 Normal renal function was defined as eGFR ≥90 mL/min/1.73 m2, and CKD stage 2 (eGFR 60–89 mL/min/1.73 m2) was used as a synonym of mildly decreased renal function.13
Echocardiographic Data CollectionThe LV function and structure of the study participants was evaluated by 2D transthoracic echocardiography with a 4-MHz, sector-type transducer probe (Vivid 7, GE, Milwaukee, USA; and E9, GE, USA). All echocardiography was performed by trained and registered sonographers following a standardized protocol.14 Images from standard parasternal long- and short-axis views were digitally stored and reviewed. LV end-diastolic diameter (LVEDD), LV endsystolic diameter (LVESD), interventricular septal thickness (IVST) and posterior LV wall thickness (PWT) was routinely measured in all participants. LV mass was calculated with the following formula: LV mass=0.8×{1.04[(LVEDD+IVST+PWT)3−(LVEDD)3]}+0.6 g,14 and indexed for body surface area (BSA); LV endsystolic volume (LVESV) and LV end-diastolic volume (LVEDV) were calculated by following formulae: 7.0/(2.4+LVESD)×LVESD3 and 7.0/(2.4+LVEDD)×LVEDD3. RWT was calculated as: (2×PWT)/LVEDD.14
LVH was defined as LVMI ≥115 in men and ≥95 in women, and RWT >0.42 was regarded as abnormal. EF was calculated as: (LVEDV−LVESV)/LVEDV×100. Detailed descriptions of the echocardiographic measurements and the characteristics of subjects are presented elsewhere.15
Statistical AnalysisData are presented as mean±standard deviation within BP groups for continuous variables and as proportions for categorical variables in men and women. We conducted the analyses in men and women because the cutoff point of LVH and the clinical and biochemical characteristics are different in men and women. The main clinical characteristics and echocardiographic parameters were compared between men and women using independent t-test for continuous variables and chi-square test for categorical variables.
The odds ratios (OR) of LVH and abnormal RWT in men and women were calculated using logistic regression analysis. Men were categorized into 4 groups of eGFR (≥90, 80–89.99, 70–79.99, 60–69.99 mL/min/1.73 m2) and women into 3 group (≥90, 80–89.99, 60–79.99 mL/min/1.73 m2) because of the small sample size for eGFR 60–70 mL/min/1.73 m2 in the women (n=86). The covariates of the adjusted model were selected from variables showing significant differences among eGFR groups and classic cardiovascular risk factors including body mass index (BMI), age, physical activity, alcohol intake, hypertension, fasting glucose, smoking, HDL-C and menopause only in women. Generalized additive model (GAM) analysis was performed to assess the association of eGFR (60–90) with LVH and abnormal RWT as a continuous variable. The All statistical analyses were performed using R 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria), and a value of P<0.05 was considered statistically significant in all analyses.
The clinical and echocardiographic characteristics of the study participants are listed in Table 1. Compared with women, men had the more unfavorable clinical characteristics, including low eGFR, high fasting glucose, low HDL-C, high BP, high BMI and high prevalence of DM and hypertension. With regard to the echocardiographic parameters, LV mass, LVMI and RWT were significantly higher in men than in women.
| Characteristics | Women (n=12,914) |
Men (n=34,816) |
P value |
|---|---|---|---|
| Age (years) | 38.2±8.5 | 40.4±7.2 | <0.001 |
| Fasting glucose (mg/dL) | 91.1±12.0 | 97.9±16.2 | <0.001 |
| HDL-C (mg/dL) | 65.5±15.3 | 52.6±12.8 | <0.001 |
| Serum creatinine (mg/dL) | 0.7±0.1 | 1.0±0.1 | <0.001 |
| eGFR (mL/min/1.73 m2) | 106.9±13.0 | 96.2±12.8 | <0.001 |
| BMI (kg/m2) | 21.9±3.2 | 24.8±3.0 | <0.001 |
| Systolic BP (mmHg) | 101.3±11.0 | 113.7±11.4 | <0.001 |
| Diastolic BP (mmHg) | 65.1±8.6 | 75.0±9.5 | <0.001 |
| Average alcohol use (g/day) | 4.7±11.8 | 17.7±23.3 | <0.001 |
| Current smoker (%) | 1.8 | 30.0 | <0.001 |
| High PA (%) | 12.8 | 13.9 | <0.001 |
| DM (%) | 2.2 | 5.8 | <0.001 |
| Antidiabetic medication (%) | 1.0 | 2.8 | <0.001 |
| Hypertension (%) | 5.5 | 17.9 | <0.001 |
| Antihypertensive medication (%) | 3.3 | 7.9 | <0.001 |
| LV mass (g) | 104.8±23.2 | 145.0±29.7 | <0.001 |
| LVMI | 66.1±12.6 | 76.7±14.0 | <0.001 |
| RWT | 0.304±0.048 | 0.332±0.049 | <0.001 |
| LVH (%) | 2.5 | 0.9 | <0.001 |
| Abnormal RWT (%) | 1.9 | 4.6 | <0.001 |
| EF (%) | 67.5±5.4 | 66.8±5.7 | 0.038 |
Continuous variables are expressed as mean (±SD), and categorical variables are expressed as number (percentage (%)). Abnormal RWT: >0.42; LVMI: >115 in men and >95 in women. BP, blood pressure; BMI, body mass index; EF, ejection fraction; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LVH, left ventricular hypertrophy; LVMI, left ventricular mass index; PA, physical activity; RWT, relative wall thickness; T2DM, type 2 diabetes mellitus.
Table 2 presents the unadjusted and adjusted ORs for LVH and abnormal RWT according to eGFR ≥60 in men. The unadjusted ORs showed a pattern of stepwise increase with declining eGFR, but adjustment for covariates negated the statistical significance for LVH. However, adjusted ORs for abnormal RWT significantly increased in inverse proportion to the level of eGFR. In women (Table 3), the unadjusted ORs for LVH and abnormal RWT showed a similar pattern of relationship as in the men. However, adjusted ORs did not show statistical significance for both LVH and abnormal RWT.
| Characteristics | Unadjusted OR | Adjusted OR | Case | Mean LVMI/RWT (±SD) |
|---|---|---|---|---|
| All participants | ||||
| LVH | ||||
| eGFR ≥90 (n=23,447) | 1.00 (Ref.) | 1.00 (Ref.) | 170 | 76.7±13.7 |
| eGFR 89.99–80 (n=7,365) | 1.29 (0.98–1.72) | 1.01 (0.75–1.34) | 69 | 76.3±14.0 |
| eGFR 79.99–70 (n=3,368) | 1.98 (1.43–2.73) | 1.11 (0.79–1.56) | 48 | 76.9±14.8 |
| eGFR 69.99–60 (n=636) | 4.91 (3.12–7.70) | 1.58 (0.97–2.57) | 22 | 79.7±17.0 |
| Abnormal RWT | ||||
| eGFR ≥90 (n=23,447) | 1.00 (Ref.) | 1.00 (Ref.) | 903 | 0.329±0.048 |
| eGFR 89.99–80 (n=7,365) | 1.43 (1.27–1.61) | 1.26 (1.12–1.43) | 399 | 0.335±0.051 |
| eGFR 79.99–70 (n=3,368) | 1.95 (1.68–2.26) | 1.44 (1.23–1.68) | 244 | 0.340±0.054 |
| eGFR 69.99–60 (n=636) | 3.19 (2.47–4.11) | 1.68 (1.28–2.21) | 72 | 0.354±0.057 |
Adjusted for BMI, age, PA, alcohol intake, fasting glucose, smoking, hypertension, and HDL-C. OR, odds ratio. Other abbreviations as in Table 1.
| Characteristics | Unadjusted OR | Adjusted OR | Case | Mean LVMI/RWT (±SD) |
|---|---|---|---|---|
| All participants | ||||
| LVH | ||||
| eGFR ≥90 (n=11,393) | 1.00 (Ref.) | 1.00 (Ref.) | 236 | 65.7±12.1 |
| eGFR 89.99–80 (n=1,034) | 2.45 (1.80–3.34) | 0.85 (0.57–1.26) | 51 | 67.6±14.5 |
| eGFR 79.99–60 (n=487) | 3.89 (2.71– 5.57) | 0.72 (0.45–1.14) | 37 | 71.7±16.3 |
| Abnormal RWT | ||||
| eGFR ≥90 (n=11,393) | 1.00 (Ref.) | 1.00 (Ref.) | 177 | 0.302±0.047 |
| eGFR 89.99–80 (n=1,034) | 2.42 (1.69–3.45) | 1.10 (0.71–1.69) | 38 | 0.315±0.052 |
| eGFR 79.99–60 (n=487) | 4.16 (2.79–6.19) | 1.37 (0.85–2.21) | 30 | 0.324±0.057 |
Adjusted for BMI, age, PA, alcohol intake, fasting glucose, smoking, hypertension, menopause and HDL-C. Abbreviations as in Tables 1,2.
The results of the GAM analyses are shown in Figure. In men, a linear correlation of LVH and abnormal RWT was observed below eGFR close to 70 and below eGFR close to 85, respectively. However, women did not show a specific pattern of relationship with LVH and abnormal RWT across all eGFR levels.
Generalized additive model (GAM) graphs for eGFR and (A) LVH and (B) abnormal RWT in men (adjusted for BMI, age, physical activity, alcohol intake, fasting glucose, smoking, hypertension, and HDL-cholesterol). GAM graphs for eGFR and (C) LVH and (D) abnormal RWT in women (adjusted for BMI, age, physical activity, alcohol intake, fasting glucose, smoking, hypertension, menopause, and HDL-C). BMI, body mass index; HDL-C, high-density lipoprotein cholesterol.
We conducted this study on the basis of a hypothesis that individuals with mildly decreased renal function are more strongly associated with LV geometry change, compared with individuals with normal renal function.
In the male study subjects, unadjusted ORs showed an inverse correlation of LVH and RWT with levels of eGFR across 60–89. This finding was supported by the GAM analysis, in which a linear correlation of LVH and abnormal RWT was observed below eGFR close to 70 and below eGFR close to 85, respectively. However, adjustment for covariates including hypertension markedly attenuated the statistical significance in the association. In the fully adjusted model, statistical significance in the association with LVH was nullified, but was maintained in the association with abnormal RWT. These findings may be explained by the role of hypertension in the pathogenesis of LVH. LVH commonly reflects a chronic adaptation to pressure and volume overload related to hypertension.16 It is established that hypertension has a crucial role in the development of LVH. Thus, adjustment for hypertension may negate the statistical significance in the association between LVH and eGFR of 60–89 mL/min/1.73 m2 in men. In contrast, abnormal RWT can be an early form of cardiac remodeling preceding increased LVMI.9 The presence of only abnormal RWT is defined as concentric cardiac remodeling,6 associated with adverse cardiovascular outcomes.10,17,18 Thus, it is inferred that statistical significance in the association between abnormal RWT and eGFR of 60–89 mL/min/1.73 m2 is maintained even after adjustment for hypertension.
Our results suggested that men with mildly decreased renal function have an accompanying increased risk of concentric remodeling reflected by abnormal RWT. Given that concentric remodeling is linked to fatal cardiovascular outcomes such as stroke, myocardial infarction and heart failure,10,17,18 it is postulated that men with mildly decreased renal function have worse cardiovascular prognosis compared with men with normal renal function. These results may provide additional information lacking in previous studies that have evaluated the association between renal function and LV geometry change. There have been studies demonstrating the increased probability of LV geometry change in mildly decreased renal function. Analysis of hypertensive patients presented a stepwise increase in the prevalence of LVH through CKD stages 2–4, confirming the inverse relationship between renal function and LV mass (β −0.287; P<0.0001).8 In the Hoorn study, mildly decreased eGFR from 60 to 90 mL/min/1.73 m2 was associated with an 8.3 g/m2 greater LVM in men after adjustment for age, glucose tolerance, hypertension and prior CVD.19 However, none of those studies analyzed the association of LV geometry change with eGFR stratified within CKD stage 2 (eGFR between 60 and 89 mL/min/1.73 m2). Although individuals with eGFR of 60–89 are regarded as having the same CKD stage 2 status, there is a wide range of eGFR. Thus, it is likely that there would be a difference in the risk of LV geometry change among individuals within CKD stage 2. Our study showed the probability of abnormal RWT significantly increased from eGFR <90 mL/min/1.73 m2. Considering the mean age of these male subjects (40.4±7.2 years), this finding indicated that LV geometry change can occur even in a relatively young population with preserved renal function. CKD is a form of age-related decline in renal function estimated as eGFR by approximately 10 per decade of life.3 Relatively young individuals with stage 2CKD are likely to progress into advanced CKD with age. In this population, the presence of LV geometry change can translate into adverse cardiovascular outcomes with the progression of CKD. Therefore, it is clinically important to provide proper management of other risk factors such DM, hypertension and dyslipidemia for this population in order to improve cardiovascular prognosis. Our results may be epidemiological evidence of the clinical significance of early identification of LV geometry change in early stage CKD.
In the present study, women did not show the significant association between LV geometry change and levels of eGFR within 60–89 mL/min/1.73 m2. The plausible explanations for this finding may be the sex difference in cardiovascular risk and the characteristics of the study participants. Female sex is generally regarded as protective against CVD. In an analysis of subjects pooled from 4 community-based studies, female sex had a protective effect for incident cardiac events (adjusted hazard ratio: 0.43 [0.39–0.48]) and stroke (adjusted hazard ratio: 0.81 [0.71–0.92]).20 Estrogen is thought to have a cardioprotective role in premenopausal women by promoting vasodilatation and decreasing production of reactive oxygen species, oxidative stress and fibrosis, which mediates a cardioprotective action through attenuation of cardiac remodeling.21 It has been demonstrated that the cardiovascular risk of women is greatly influenced by menopause, DM and dyslipidemia. Menopause is a female-specific cardiovascular risk factor that leads to a deficiency in estrogen and its cardioprotective effect and steep elevation of CVD.22 The presence of DM results in a 3–7-fold increase in the cardiovascular risk for women, compared with a 2–3-fold elevation of cardiovascular risk in men.23 Additionally, while LDL-C shows a linear relationship with cardiovascular risk in women below 65 years of age, low HDL-C is associated with greater cardiovascular risk in women over 65 years old than in equivalent aged men.24 Therefore, in the current female subjects, LV geometry change may be more frequently present in elderly menopausal women with, DM and low HDL-C than their counterparts. In this study population, adjustment for covariates including menopause, fasting glucose and HDL-C may eliminate the statistical significance in the association between LV geometry change and eGFR of 60–89 mL/min/1.73 m2. Moreover, it seems that these female subjects had favorable metabolic conditions. Their mean age was 38.2±8.5 years, and the proportion of menopausal women was 7.3% (n=948). Additionally, the women had more favorable metabolic profiles including lower prevalence of DM and hypertension, and lower levels in BMI and LDL-C than the men. Favorable risk profiles in the women may contributed to the lower probability of LV geometry change with mildly decreased renal function.
Study LimitationsFirst, our study participants were relatively young, which limits generalizing of the results to elderly women. Analysis of menopausal women may produce different results. Thus, it is necessary to conduct a risk assessment for women including both premenopausal and postmenopausal subjects.
Second, we evaluated renal function depending only on eGFR calculated by the CKD-EPI equation, which is based on serum creatinine that can be influenced by creatinine generation as a result of muscle mass and its turnover.25 However, this concern may be partly relieved by our study participants including relatively young and apparently healthy adults.
Third, our study subjects were only recruited from a cohort of the KSHS. Therefore, despite the large number of study subjects there might be a selection bias.
Abnormal RWT was significantly associated with eGFR <90 mL/min/1.73 m2 in men, which suggested that men with mildly decreased renal function have increased probability of LV geometry change. This association was not observed in women, which may be attributable to the favorable risk profiles of the women. Further studies should identify whether LV geometry change with mildly decreased renal function leads to poor cardiovascular prognosis.
This study was based on medical data collected and arranged by the Kangbuk Samsung Cohort Study (KSCS). Therefore, this study could only be done by virtue of the labor of all staff working in the KSCS and Total Healthcare Center, Kangbuk Samsung Hospital.
S.K.P. coordinated the study, analyzed the data and wrote the manuscript as a first author. J.G.K. verified the data on echocardiographic parameters as a cardiologist. P.-W.C. participated in conducting the study and writing the manuscript. J.-H.R. played role in analyzing the data and editing the manuscript. J.Y.J. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
The authors have nothing to disclose.
All authors had access to the data used in this study and participated in writing the manuscript.
