Introduction

Recently, basic and clinical research have improved our understanding of cardio–renal damage and suggested that an interaction between pathophysiological (that is, hemodynamic, inflammatory and neuro-hormonal) mechanisms amplifies structural and functional cardio–renal derangement.1 The presence of both renal dysfunction and left ventricular hypertrophy (LVH) indicates cardiovascular risk and a poorer prognosis compare with the isolated cardiac or renal damage. This interaction is evident even in patients with mild renal insufficiency.2 According to the Multi-Ethnic Study of Atherosclerosis, concentrations of cystatin C, a superior marker of renal function, were inversely associated with left ventricular end-diastolic volumes and directly associated with concentricity independent of traditional cardiovascular risk factors in 4910 subjects with normal to mild renal dysfunction.3 In addition, Fesler et al.4 reported that increased left ventricular mass predicts subsequent renal dysfunction in both chronic kidney disease (CKD) patients and healthy subjects. However, it remains unclear whether subclinical renal damage, such as that detectable only by histological examination, is associated with altered left ventricular geometry in healthy subjects.

Although in general living kidney donors are healthy individuals, their pretransplant biopsy (0-h biopsy) specimens, obtained immediately after nephrectomy, often reveal varying degrees of global glomerulosclerosis (GGS) and arteriolosclerosis. In the general population, GGS is reportedly present in 1–3% of total glomeruli in adult subjects older than 50 years of age: the number of subjects with GGS increases with increasing age, reaching up to 30% by 80 years of age.5, 6 However, the precise causes are still unknown. A study analyzing biopsy specimens from 1208 living kidney donors reported that GGS is not completely attributable to age-related decline in glomerular filtration rate (GFR).7 On the other hand, we have noted that preoperative echocardiography in these subjects occasionally reveals LVH or left ventricular remodeling. LVH and left ventricular remodeling are common in hypertensive subjects and the elderly;8 however, some of our subjects with LVH have been normotensive and not elderly.

Therefore, we examined the cardio–renal interaction relationship between GGS at 0-h biopsies and LVH in near healthy subjects.

Materials and Methods

Patients

Three hundred and nine consecutive living kidney donors from March 2006 to October 2012 were initially included. Twenty-two of these donors were excluded because a 0-h biopsy or preoperative echocardiography was not performed. Thirty-nine donors were excluded because of inadequate biopsy specimens that contained fewer than 10 glomeruli or included simple cysts. Another 10 donors were excluded because some relevant laboratory findings were missing. The remaining 238 subjects were enrolled in this study. No subject had a history of stroke, coronary heart disease or heart failure. This study was part of a retrospective cohort study conducted by our department, which was approved by the Clinical Research Ethical Committee of the Kyushu University of Japan (Practice #24–54). Written informed consent was obtained from each subject.

Evaluation of echocardiography

Echocardiographic studies had been performed within the 6 months before donor nephrectomy using a iE33 echography system (Philips Health Care, Andover, MA, USA), Vivid 7 Dimension cardiovascular ultrasound system (GE Healthcare, Chalfont St Giles, Bucks, UK) or Artida 4D (Toshiba Medical Systems, Tochigi, Japan). All tracings were evaluated by several laboratory technicians and cardiologists. Overall, one-dimensional left ventricular measurements and two-dimensional (apical 4- and 2-chamber) views were obtained, in accordance with the recommendations of the American Society of Echocardiography. Relative wall thickness at end diastole was calculated as the ratio of the sum of left ventricular septal and posterior wall thickness to left ventricular internal end-diastolic diameter. Left ventricular mass was calculated using the Penn-cube method as described by Devereux et al.9 LVH was defined as left ventricular mass index greater than 110 g m−2 in female and 125 g m−2 in male subjects.10 Left ventricular remodeling was defined as increased relative wall thickness (more than 0.45) without hypertrophy. Concentric and eccentric LVH were defined as the presence or absence of increased relative wall thickness (more than 0.45), respectively.

Definitions of renal function and other relevant clinical variables before surgery

Hypertension was considered to be present if at least two consecutive resting blood pressure measurements before surgery met the criteria of the Seventh Joint National Committee Guidelines on Prevention, Detection, Evaluation and Treatment of High Blood Pressure11 or the subject had an established history of hypertension that had necessitated medical treatment. GFR was determined based on urinary clearance of 99mtechnetium-labeled diethylene triamine pentaacetic acid using the constant infusion technique with urine collection and normalized according to body surface area. Diabetes mellitus was defined as a fasting blood glucose concentration of more than 126 mg dl−1 and hemoglobin A1c more than 6.5% during preoperative evaluation. Hemoglobin A1c was measured according to Japanese Diabetes Society/Japanese Society of Clinical Chemistry guidelines, which were standardized by adding 0.4% to the estimate as the National Glycohemoglobin Standardization Program equivalent value.12 All subjects who were defined as having impaired glucose tolerance or diabetes mellitus received strict dietary therapy or hypoglycemic agents or both until normal glucose tolerance had been achieved. By the time of kidney donation, glucose tolerance had normalized in all donors. Obesity was defined as body mass index ⩾25 kg m−2. Dyslipidemia was defined as fasting total cholesterol concentration ⩾220 mg dl−1, triglyceride concentration ⩾150 mg dl−1 or use of lipid-modifying agents.

Data collection

Patient history, physical examination findings and laboratory values were collected. The assessed clinical variables were age, sex, comorbid medical conditions, use of medication (for example, antihypertensive drugs), body mass index and systolic and diastolic blood pressure. Body surface area was calculated by the du Bois method.13 Baseline laboratory data, such as blood hemoglobin and urea nitrogen, and serum creatinine, uric acid, total cholesterol, triglyceride and glucose concentrations, were obtained within a few days before surgery, along with the evaluation of a spot urine sample to assess the urinary protein/creatinine ratio and a dipstick urinary test.

Histological evaluation of 0-h renal biopsy

A wedge biopsy had been obtained from the outer cortex of each donor kidney immediately before transplantation. The sample was fixed in a paraformaldehyde–glutaraldehyde solution and then embedded in paraffin. Next, the sample was stained with hematoxylin-eosin, periodic acid-Schiff, methenamine silver and Masson trichrome stains. %GGS was determined using light microscopy. To quantify other characteristics of nephrosclerosis, such as renal arteriosclerosis, arteriolar hyalinosis and interstitial fibrosis, a semi-quantitative grading score based on the Banff criteria was used.14 Arteriolosclerosis was defined as follows:+1, fibro-intimal thickening<intimal diameter;+2, fibro-intimal thickening=intimal diameter; and+3, fibro-intimal thickening > intimal diameter. Each biopsy specimen was evaluated by at least two specialists in renal pathology (AT and KM) who were blinded to the clinical data.

Statistical analysis

Values are expressed as mean (s.d.) or number (%). The subjects were divided into tertiles according to %GGS (low, 0–3.45%; moderate, 3.46–11.76%; and high, ⩾11.77%). Trends in continuous and categorical values across the tertiles were examined by the Jonckheere–Terpstra and Cochran–Armitage tests, respectively.15, 16 The prevalence of left ventricular morphological changes and the various renal histology grades in each tertile of %GGS were evaluated using the linear regression analysis.17 After adjusting for the potential confounding risk factors (except triglyceride) that had been found to be statistically significant by the trend analyses (two-tailed P-value<0.05), the odds ratios and 95% confidence intervals for LVH were calculated by logistic regression. P<0.05 was considered statistically significant. All statistical analyses were conducted using JMP version 9.0.2, or SAS version 9.2 (SAS Institute, Cary, NC, USA) for trend testing of continuous values.

Results

Trend analysis of baseline characteristics

Table 1 shows the characteristics of all subjects according to each tertile of %GGS. Overall subjects, the mean number of glomeruli per biopsy was 26±13. Trace proteinuria (±) was detected by the dipstick urinary test in only two subjects; however, their urinary protein/creatinine ratios were less than 0.10 g per gCr. The other 236 subjects had negative dipstick urinary tests and urinary protein/creatinine ratios. The mean GFR was 117±27 ml min−1 per 1.73 m2. On admission, 28 subjects were taking antihypertensive drugs, which were angiotensin-converting enzyme inhibitors or angiotensin receptor blockers in 17 cases. Eighteen subjects with dyslipidemia were receiving statins, bezafibrate or ezetimibe. Sixteen subjects had been diagnosed with diabetes mellitus during preoperative evaluation. Some of them had been receiving diet therapy and short-term antidiabetic medication or both to normalize their fasting blood glucose concentrations by the time of kidney donation. There were significant trends toward correlations between %GGS tertiles and age, hypertension, history of diabetes mellitus, dyslipidemia, systolic/diastolic blood pressure, pulse pressure, GFR, total serum cholesterol and triglyceride.

Table 1 Mean values or frequencies of potential risk factors and laboratory variables for overall baseline characteristics and tertiles according to %GGS
Full size table

Distribution of left ventricular geometric changes according to tertiles of %GGS

Figure 1 shows the left ventricular morphological changes for each tertile of %GGS. The number of patients with left ventricular morphological changes was proportional to these tertiles (P<0.01). In addition, a substantial trend was observed between tertiles of %GGS and left ventricular mass index (P for this trend <0.01).

Figure 1
figure 1

The distribution of left ventricular geometric changes and left ventricular mass index (LVMI). The overall prevalence of left ventricular remodeling and left ventricular hypertrophy (concentric and eccentric) is increased along with the tertiles of %GGS (P<0.01). A substantial trend is also observed between the tertiles of %GGS and LVMI (P for trend <0.01). %GGS; percentage of global glomerulosclerosis, LVMI; left ventricular mass index.

Full size image

Histological analysis of the relationship between %GGS and other nephrosclerotic changes

Table 2 shows proportion of other histological findings characteristic of nephrosclerosis categorized according to Banff’s criteria. Individuals in the higher two categories of tertiles had a significantly higher proportion of advanced lesions than those in the lowest tertile.

Table 2 Analysis of relationship between tertiles according to %GGS and various histological changes of nephrosclerosis
Full size table

Multivariate logistic regression models for LVH

The results of multivariate logistic regression analyses are shown in Table 3. In Model 1, unadjusted analysis showed that the OR for the highest tertile of %GGS was significantly greater than that for the lowest tertile. In Model 2, the OR for the highest tertile of %GGS remained significantly increased after adjusting for age and sex. Finally, even after adjusting for age, sex, systolic blood pressure, history of diabetes mellitus, total serum cholesterol and GFR, the subjects in the highest tertile of %GGS had a sevenfold greater risk of LVH than did those in the lowest tertile of %GGS (in Model 3, P<0.05).

Table 3 Multivariate logistic regression analysis of LVH
Full size table

Discussion

In the present study, we identified a significant association between GGS at 0-h biopsies and LVH as evaluated by preoperative echocardiography. This is the first study that has demonstrated a subclinical cardio–renal interaction in a near healthy population.

Nephrosclerosis is defined by the presence of glomerulosclerosis, arteriosclerosis, tubular atrophy and interstitial fibrosis on renal biopsy. Age or variations in blood pressure cannot completely explain the prevalence of GGS at 0-h biopsies from living kidney donors.5 In experimental models of nephrosclerosis, the renin–angiotensin–aldosterone system, sympathetic nerve overactivity, inflammatory reactions induced by reactive oxygen species and impaired bioavailability of nitric oxygen are considered to have important roles in the development of nephrosclerosis.18, 19, 20, 21

On the other hand, LVH is an adaptive response to increased cardiac work, which is the product of left ventricular pressure and stroke volume.22 The causes for this compensatory mechanism are the same as those listed above for nephrosclerosis.23, 24, 25

In patients with essential hypertension, LVH is more frequent in those with mild to moderate CKD than in those with normal renal function. Nardi et al.26 reported an increasingly higher prevalence of LVH with declining renal function in their study of 293 CKD patients. Moreover, after adjusting for age, sex, diastolic blood pressure and hemoglobin, they found that CKD is an independent predictor of left ventricular mass index. In the present study, we also determined that even in patients with subclinical kidney dysfunction, the relationship between LVH and renal histological damage is independent of other established factors for atherosclerosis. These findings suggest that nonclassical atherogenic factors such as the renin–angiotensin–aldosterone system in the peripheral organs, sympathetic nervous system overactivity, local inflammation and oxidative stress contribute to both left ventricular remodeling and chronic renal injury, even in subjects without CKD.27

Recently, an in vivo study showed that a decrease in nephron numbers is a significant determinant of increased blood pressure and cardio–renal damage generally, and of renal structural damage and LVH in particular.28, 29 Genetic or environmental factor may affect the degree of cardio–renal injury in healthy individuals.

Age is the most important factor in the relationship between nephrosclerosis and LVH. Age-related renal change is associated with the declining renal blood flow as shown by the xenon washout technique.30 Its blunted vasodilatory response to intrarenal acetylcholine administration suggests that age-related decline in renal blood flow has a vascular cause. Conversely, left ventricular dysfunction is sensed by baroreceptors, which activate the sympathetic nervous system. Impaired cardiovascular hemodynamic variables, such as increased end-diastolic pressure, increased peripheral vessel resistance and decreased stroke volume, potentially impair the peripheral blood flow in the kidneys.1 Although aging is the strongest risk factor for LVH, we speculate that left ventricular remodeling and arterial stiffness accelerate the age-related decline in renal blood flow and that this interaction produces a vicious cycle. Age-dependent GFR decreases are reportedly more pronounced in individuals with concentric LVH than in those with hypertension and normal left ventricular mass.4, 31 In addition, we identified by univariate analysis that pulse pressure increases along with tertiles of %GGS and found a weak correlation between left ventricular mass index and pulse pressure (Supplementary Figure 1). On the other hand, the association between decreased numbers of intact glomeruli and LVH may indicate that sodium retention contributes to these pathophysiological phenomena.32

This study has several limitations. First, there is a selection bias toward subjects with a family history of end-stage renal disease. The causes of renal diseases include hypertension and diabetes mellitus. Moreover, the subjects were self-selected by their agreement to serve as living kidney donors. Second, the biopsy specimens included only the superficial renal cortex. In the present study, the frequency of GGS was relatively higher than that reported by an autopsy study.5 However, according to Ito33, in the early stages of hypertension, renal injury occurs predominantly in the juxtamedullary nephrons, the majority of other nephrons remaining relatively intact. On the other hand, in an autopsy study, Kasiske34 reported that GGS occurs uniformly across all zones of the cortex. We could not confirm that the %GGS in our biopsy specimens is representative of the whole kidney. Finally, because this study is cross-sectional in nature, we cannot reach cause-and-effect conclusions. Further prospective studies are needed to clarify the long-term prognostic significance of the relationship between LVH and GGS.

Conclusion

We have demonstrated a correlation between the GGS and LVH in subjects with normal renal function. This relationship is independent of age, sex, blood pressure, renal function and other conventionally recognized atherogenic factors. A cardio–renal interaction may exist before there is an obvious decline in renal function.