Abstract
Background
Total kidney volume (TKV) is an imaging biomarker that may have diagnostic and prognostic utility. The relationships between kidney volume, renal function and cardiovascular disease (CVD) have not been characterized in a large community-dwelling population. This information is needed to advance the clinical application of TKV.
Methods
We measured TKV in 1852 Framingham Heart Study participants (mean age 64.1 ± 9.2 years, 53% women) using magnetic resonance imaging. A healthy sample was used to define reference values. The associations between TKV, renal function and CVD risk factors were determined using multivariable logistic regression analysis.
Results
Overall, mean TKV was 278 ± 54 cm3 for women and 365 ± 66 cm3 for men. Risk factors for high TKV (>90% healthy referent size) were body surface area (BSA), diabetes, smoking and albuminuria, while age, female and estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2 were protective. Participants with high TKV had higher odds of diabetes [odds ratio (OR) 2.15, P < 0.001] and lower odds of eGFR <60 mL/min/1.73 m2 (OR 0.32, P = 0.007). Risk factors for low TKV (<10% healthy referent size) were age, female and eGFR <60 mL/min/1.73 m2, while BSA and diabetes were protective. Participants with low TKV had higher odds of eGFR <60 mL/min/1.73 m2 (OR 6.12, P < 0.001) and albuminuria (OR 1.56, P = 0.03).
Conclusions
Low TKV is associated with markers of kidney damage including albuminuria and eGFR <60 mL/min/1.73 m2, while high TKV is associated with diabetes and decreased odds of eGFR <60 mL/min/1.73 m2. Prospective studies are needed to characterize the natural progression and clinical consequences of TKV.
Keywords: albuminuria, cardiovascular disease risk factors, chronic kidney disease, magnetic resonance imaging, total kidney volume
INTRODUCTION
Kidney size is often assessed clinically using a single linear length measurement. However, total kidney volume (TKV) is an emerging biomarker for investigating the structural and functional basis of disease. Kidney volume is thought to provide prognostic information regarding chronic kidney disease (CKD) risk, while renal volumetric analysis has been applied to the study of kidney transplantation outcomes and renal artery stenosis [1–3]. Support for TKV to be used as an end-point in clinical trials of polycystic kidney disease has been proposed and automatic image processing techniques are being developed to advance this goal [4, 5]. However, the true clinical utility of TKV has yet to be realized because the current understanding of normal kidney volume is incomplete.
Normative values of TKV have not been established in an unselected general population and it remains unclear what defines high and low TKV. Prior studies in adults that explored the associations between kidney volume, renal function and cardiovascular disease (CVD) risk factors were limited by imaging modality, selection bias or sample size [6, 7]. This information is necessary to further advance TKV as a research and diagnostic parameter.
Measurement of kidney volume using ultrasound is operator dependent and has poor reproducibility [8, 9]. Computed tomography (CT) exposes the patient to ionizing radiation. Using magnetic resonance imaging (MRI) to obtain accurate kidney volume measurements has been validated in animals, used selectively in humans and was proposed as the method of choice for volumetric analysis [10, 11]. The aims of the present study were to define sex-specific reference values of TKV using MRI in a longitudinally followed community-based cohort with robust risk factor analysis and to identify the clinical covariates associated with high and low TKV.
MATERIALS AND METHODS
Study population
The Framingham Heart Study began in 1948 and the design has been described previously [12]. Offspring of the original cohort were enrolled starting in 1971 and examined every 4–5 years at clinic visits [13]. Starting in 1994, the Omni cohort was recruited from residents of Framingham, Massachusetts, who identified themselves as members of minority groups. The current study includes 1763 participants from Offspring exam 7 and 160 participants from Omni exam 1 who were without contraindications to MRI. Fifty-four participants were excluded due to uninterpretable imaging data, 3 were excluded for missing covariates and 14 were excluded for having at least one kidney with fewer than seven MRI slices because a kidney of that length is generally considered atrophic. A final sample size of 1852 participants was available for analysis. The study protocol was approved by the institutional review boards at Boston University Medical Center and Beth Israel Deaconess Medical Center. All participants provided written informed consent and the authors adhered to the ethical principles set forth in the Declaration of Helsinki.
TKV measurements
Kidney volume data were derived from noncontrast MRI abdominal images acquired from 2002 to 2005 using a 1.5-Tesla whole-body MRI system (Gyroscan ACS-NT, Philips Healthcare, Best, The Netherlands). Thirty-six transverse images ranging from the aortic arch to the aorto-iliac bifurcation were acquired using an ECG-gated, fat-suppressed, T2-weighted black blood turbo spin-echo sequence [14]. Abdominal images had a slice thickness of 5 mm with a 5-mm interslice gap; in-plane spatial resolution was 1.03 × 0.64 mm. Each image was analyzed by a single expert reviewer (N.O.-M.) for the presence of kidney. Renal parenchymal contours were manually traced in each slice using commercial software (QMass 6.1, Medis, Leiden, The Netherlands). Renal sinus fat, cysts and large vessels were excluded. Left and right kidney volumes were calculated using slice summation whereby the parenchymal area of each kidney slice was multiplied by a section interval of 1 cm (5-mm slice thickness + 5-mm intersection gap) and then summing all values. This process is sometimes referred to as the voxel-count or disk summation method and is considered the reference standard [9, 15, 16]. The left and right kidney volumes were then summed to derive the TKV since clinical parameters reflect combined renal function.
Clinical measurements
Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m2). Body surface area (BSA) was calculated using the Mosteller formula [17]. Blood pressure was measured twice by a physician using a standardized protocol and the readings were averaged. Hypertension was defined as the use of antihypertensive medication, diastolic blood pressure ≥90 mmHg or systolic blood pressure ≥140 mmHg. Diabetes was defined as the use of anti-diabetes medication or a fasting glucose of ≥126 mg/dL. Smoking status was defined as never, current or former at the time of the clinic visit.
Serum creatinine was measured using the Jaffe assay on a Roche Hitachi 911 (Roche Diagnostics, Indianapolis, IN, USA) with an intra-assay coefficient of variation of 3.1%. Statistical calibration of serum creatinine was performed due to the variation in laboratory assays. Briefly, creatinine values from the National Health and Nutrition Examination Survey (NHANES) III were first calibrated in the Cleveland Clinic laboratory using a correction factor of 0.23 mg/dL. Mean creatinine values from the Framingham Heart Study were then matched with the corresponding age- and sex-specific NHANES III means [18]. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation [19]. Albuminuria was defined as a urinary albumin-to-creatinine ratio (UACR) of ≥17 mg/g for men and ≥25 mg/g for women to account for the sex-specific differences in urinary creatinine excretion [20].
Statistical analyses
Dichotomous values are presented as proportions, normally distributed continuous values are presented as sample means with standard deviations and non-normally distributed variables are shown as medians with interquartiles. A healthy referent group was identified from the overall study group. Referent-group participants were those with total cholesterol ≤240 mg/dL, high-density lipoprotein (HDL) cholesterol ≥45 mg/dL for men and ≥55 mg/dL for women, 18.5 kg/m2 < BMI < 30 kg/m2, eGFR ≥60 mL/min/1.73 m2 and free of hypertension, diabetes, albuminuria and CVD. From this group of 196 women and 112 men, the sex-specific lower 10th percentile and upper 90th percentile cut-off points of TKV were calculated to define low and high TKV, respectively. These thresholds were then applied to the overall group.
To study the relationships between volume and clinical characteristics, we applied linear regression models. Unadjusted univariate analyses used TKV as the dependent variable. The sex-specific standardized continuous or dichotomous independent variables were age, sex, BSA, hypertension, systolic and diastolic blood pressure, diabetes, glucose, eGFR, UACR, smoking status, total cholesterol, HDL cholesterol and triglycerides. Significant TKV predictors observed in the univariate models were subsequently used as independent variables for the multivariable-adjusted regression model.
Separately, we used stepwise logistic regression models to identify independent correlates for sex-specific low or high TKV compared with normal TKV. Candidate covariates included age, sex, height, weight, BSA, systolic blood pressure, use of antihypertensive medications, diabetes, fasting glucose, current smoking, eGFR <60 mL/min/1.73 m2, albuminuria, use of lipid lowering therapy, total cholesterol, HDL cholesterol and triglycerides. We did not identify any significant sex interaction. Finally, multivariable logistic regression models were used to determine the associations between either low or high TKV and characteristics including hypertension, diabetes, eGFR <60 mL/min/1.73 m2 and albuminuria. For the dependent variables of hypertension and diabetes, we adjusted for age, sex, BMI and smoking. For the dependent variables of eGFR <60 mL/min/1.73 m2 and albuminuria, we adjusted for age, sex, BMI, hypertension, diabetes and smoking. A two-tailed P-value of <0.05 was considered statistically significant. Statistical analyses were performed with SAS (version 9.2, SAS Institute, Inc., Cary, NC, USA).
RESULTS
Baseline characteristics
Sex-specific characteristics for all 1852 participants are detailed in Table 1. Overall, the mean age of our sample was 64.1 ± 9.2 years and 53% were women. The age range of participants was 37–89 years for women and 42–88 years for men. There were two participants <40 years old. Mean left kidney volumes for women and men were 141.0 ± 30.1 and 184.5 ± 34.9 cm3, respectively. Mean right kidney volumes for women and men were 137.1 ± 28.1 and 181.4 ± 34.8 cm3, respectively. Left and right kidney volumes and TKV for women were significantly less than volumes for men (all P< 0.001). The ratio of left to right kidney volume was 1.0 ± 0.2 in women and 1.0 ± 0.1 in men (P= 0.08). There was a positive correlation between TKV and the total number of renal slices (surrogate for length) of r = 0.57 for women and r = 0.54 for men (both P< 0.001). Mean TKV for women was 278.1 ± 54.4 and 365.8 ± 65.6 cm3 for men. Distribution of TKV by 10-year age groups for all participants is provided in Table 2 and shows a progressive decline in women starting after the <50 years age group. The reduction in TKV occurred later in men, starting after the 50–59-year age group.
Table 1.
Demographic and clinical characteristics for study population stratified by sex
| Characteristic | Women | Men |
|---|---|---|
| Number of participants | 981 (53%) | 871 (47%) |
| Age (years) | 63.9 ± 9.2 | 64.3 ± 9.2 |
| Height (cm) | 161.4 ± 6.2 | 174.9 ± 6.6 |
| Weight (kg) | 71.2 ± 15.1 | 86.9 ± 13.4 |
| Body mass index (kg/m2) | 27.4 ± 5.5 | 28.4 ± 4.1 |
| Body surface area (m2) | 1.78 ± 0.20 | 2.05 ± 0.18 |
| Non-Hispanic White | 891 (90.8%) | 805 (92.4%) |
| Hypertension | 290 (29.7%) | 326 (37.6%) |
| Diabetes | 65 (6.7%) | 97 (11.2%) |
| eGFR <60 mL/min/1.73 m2 | 58 (6.0%) | 62 (7.1%) |
| eGFR (mL/min/1.73 m2) | 86.0 ± 16.0 | 84.4 ± 15.7 |
| Albuminuria | 82 (8.6%) | 146 (17.1%) |
| Cardiovascular disease | 138 (14.1%) | 213 (24.5%) |
| Smoking status | ||
| Never | 525 (53.5%) | 411 (47.2%) |
| Current | 172 (17.5%) | 171 (19.6%) |
| Former | 284 (29.0%) | 289 (33.2%) |
| Total cholesterol (mg/dL) | 205.9 ± 36.3 | 193.0 ± 33.7 |
| HDL cholesterol (mg/dL) | 60.7 ± 16.3 | 45.3 ± 12.3 |
| Triglycerides (mg/dL) | 108 (77, 157) | 117 (80, 175) |
| Right kidney volume (cm3) | 137.1 ± 28.1 | 181.4 ± 34.8 |
| Left kidney volume (cm3) | 141.0 ± 30.1 | 184.5 ± 34.9 |
| TKV (cm3) | 278.1 ± 54.4 | 365.8 ± 65.6 |
Data are presented as mean ± standard deviation, number (%), or median and interquartiles (25%, 75%).
eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; TKV, total kidney volume.
Table 2.
Median (10%, 90%) TKV of women and men by 10-year age groups
| Age group | Women |
Men |
||
|---|---|---|---|---|
| n | TKV (cm3) | n | TKV (cm3) | |
| Overall | 981 | 271 (217, 352) | 871 | 360 (289, 447) |
| <50 years | 53 | 288 (239, 373) | 59 | 363 (301, 450) |
| 50–59 years | 308 | 284 (230, 368) | 241 | 375 (307, 461) |
| 60–69 years | 343 | 270 (216, 348) | 321 | 366 (297, 454) |
| ≥70 years | 277 | 256 (201, 322) | 250 | 339 (265, 424) |
TKV, total kidney volume.
Predictors of TKV
Univariate- and multivariable-adjusted linear regression models assessing predictors of TKV are illustrated in Table 3. Significant independent predictors of increased TKV included male, increased BSA, hypertension, diabetes, increased glucose, increased eGFR, current smoker and increased triglycerides. Younger age, increased diastolic blood pressure and increased HDL cholesterol were significant in the univariate models but nonsignificant in the multivariable-adjusted analysis.
Table 3.
Predictors of total kidney volume among all participants using linear regression where β indicates the increment in volume (cm3) per standard deviation for continuous variables or presence of dichotomous variables
| Characteristic | Univariatea |
Multivariableb |
||
|---|---|---|---|---|
| β | P-value | β | P-value | |
| Age (10-year) | −16.34 | <0.001 | −0.47 | 0.72 |
| Male | 87.71 | <0.001 | 86.48 | <0.001 |
| Body surface area (m2) | 33.70 | <0.001 | 31.12 | <0.001 |
| Hypertension | 18.29 | <0.001 | 7.26 | 0.002 |
| Systolic blood pressure (mmHg) | −1.20 | 0.49 | – | – |
| Diastolic blood pressure (mmHg) | 1.76 | <0.001 | −0.28 | 0.79 |
| Diabetes | 40.37 | <0.001 | 9.90 | 0.04 |
| Glucose (mg/dL) | 10.17 | <0.001 | 3.23 | 0.02 |
| eGFR (mL/min/1.73 m2) | 23.34 | <0.001 | 24.80 | <0.001 |
| UACR (mg/g) | 2.00 | 0.25 | – | – |
| Current smoker | 22.45 | <0.0001 | 11.61 | <0.001 |
| Former smoker | −0.84 | 0.82 | – | – |
| Total cholesterol (mg/dL) | −2.08 | 0.23 | – | – |
| HDL cholesterol (mg/dL) | −9.21 | <0.001 | 0.21 | 0.85 |
| Triglycerides (mg/dL) | 8.02 | <0.001 | 5.02 | <0.001 |
eGFR, estimated glomerular filtration rate; UACR, urinary albumin-to-creatinine ratio; HDL, high-density lipoprotein.
aUnadjusted single covariate analysis.
bMultivariable-adjusted using all significant predictors from the univarate analyses and where ‘–’ indicates that characteristic was not included in model.
Thresholds for low and high TKV
The healthy referent group of 196 women and 112 men was used to determine reference TKV values. In this group, mean TKV was 267.1 ± 43.7 cm3 for women and 354.8 ± 50.1 cm3 for men. Sex-specific TKV lower than the 10th percentile of this healthy group defined low TKV. Low TKV cut-off points were 219.9 cm3 for women and 301.0 cm3 for men. Similarly, TKV higher than the upper 90th percentile in the referent group defined high TKV. The thresholds for high TKV were 322.2 cm3 for women and 415.2 cm3 for men.
Participant characteristics and TKV
Sex-specific clinical characteristics of all participants with low and high TKV are presented in Table 4. Low TKV was more commonly observed with increased age, eGFR <60 mL/min/1.73 m2, albuminuria and CVD compared with both the middle and high TKV groups. Participants with high TKV had increased prevalence of hypertension, diabetes, lower HDL cholesterol, higher triglycerides and higher BSA compared with both the middle and low TKV groups.
Table 4.
Characteristics of women and men stratified by total kidney volume using the lower 10th percentile and upper 90th percentile thresholds derived from the healthy referent sample
| Characteristic | Women |
Men |
||||
|---|---|---|---|---|---|---|
| Low TKVa | Middle TKV group | High TKVb | Low TKVc | Middle TKV group | High TKVd | |
| Number of participants | 115 | 676 | 190 | 123 | 568 | 180 |
| Age (years) | 69.6 ± 8.1 | 63.9 ± 9.2 | 60.8 ± 8.1 | 69.4 ± 9.6 | 64.1 ± 9.0 | 61.3 ± 8.4 |
| Height (cm) | 156.6 ± 5.6 | 161.3 ± 5.8 | 164.6 ± 5.9 | 171.5 ± 6.6 | 174.7 ± 6.2 | 177.9 ± 6.7 |
| Weight (kg) | 62.0 ± 11.9 | 68.7 ± 11.7 | 85.6 ± 18.2 | 77.1 ± 11.7 | 85.3 ± 11.3 | 98.8 ± 12.4 |
| Body mass index (kg/m2) | 25.3 ± 4.4 | 26.5 ± 4.5 | 31.6 ± 6.9 | 26.2 ± 3.6 | 28.0 ± 3.6 | 31.3 ± 4.2 |
| Body surface area (m2) | 1.64 ± 0.17 | 1.75 ± 0.16 | 1.97 ± 0.22 | 1.91 ± 0.16 | 2.03 ± 0.15 | 2.21 ± 0.15 |
| Non-Hispanic White | 110 (95.7%) | 610 (90.2%) | 171 (90.0%) | 113 (91.9%) | 531 (93.5%) | 161 (89.4%) |
| Hypertension | 34 (29.6%) | 187 (27.7%) | 69 (36.5%) | 45 (36.6%) | 197 (34.9%) | 84 (46.7%) |
| Diabetes mellitus | 5 (4.4%) | 39 (5.8%) | 21 (11.1%) | 11 (8.9%) | 48 (8.5%) | 38 (21.1%) |
| eGFR <60 mL/min/1.73 m2 | 26 (22.8%) | 32 (4.8%) | 1 (0.53%) | 30 (24.4%) | 27 (4.8%) | 7 (3.9%) |
| eGFR (mL/min/1.73 m2) | 73.7 ± 17.5 | 85.2 ± 14.9 | 94.8 ± 17.7 | 73.4 ± 18.1 | 84.8 ± 14.0 | 91.5 ± 14.8 |
| Albuminuria | 16 (14.6%) | 46 (7.0%) | 20 (10.9%) | 29 (25.0%) | 81 (14.7%) | 36 (20.6%) |
| Cardiovascular disease | 20 (17.4%) | 95 (14.1%) | 23 (12.1%) | 34 (27.6%) | 136 (23.9%) | 43 (23.9%) |
| Smoking status | ||||||
| Never | 65 (56.6%) | 365 (54.0%) | 95 (50.0%) | 61 (49.6%) | 272 (47.9%) | 78 (43.3%) |
| Current | 18 (15.7%) | 104 (15.4%) | 50 (26.3%) | 19 (15.5%) | 102 (18.0%) | 50 (27.8%) |
| Former | 32 (27.8%) | 207 (30.6%) | 45 (23.7%) | 43 (35.0%) | 194 (34.2%) | 52 (28.9%) |
| Total cholesterol (mg/dL) | 208.4 ± 36.0 | 205.6 ± 34.6 | 205.2 ± 41.9 | 194.1 ± 35.4 | 193.8 ± 33.1 | 189.9 ± 24.3 |
| HDL cholesterol (mg/dL) | 62.6 ± 16.4 | 61.6 ± 16.5 | 56.3 ± 14.8 | 45.8 ± 10.9 | 46.2 ± 12.8 | 42.0 ± 11.0 |
| Triglycerides (mg/dL) | 107.0 (77, 156) | 105.0 (74, 152) | 124.0 (89, 165) | 110 (78, 143) | 115 (76, 174) | 137 (96, 205) |
| Right kidney volume (cm3) | 100.5 ± 13.5 | 131.8 ± 16.3 | 178.2 ± 20.6 | 135.0 ± 14.8 | 176.2 ± 19.9 | 229.4 ± 24.3 |
| Left kidney volume (cm3) | 99.0 ± 16.6 | 136.0 ± 17.3 | 184.3 ± 20.3 | 136.5 ± 19.0 | 180.0 ± 19.7 | 231.3 ± 24.2 |
| TKV (cm3) | 199.5 ± 19.2 | 267.8 ± 27.3 | 362.5 ± 34.4 | 271.5 ± 26.3 | 356.2 ± 31.5 | 460.8 ± 42.5 |
| Number of slices, right | 9.0 ± 1.0 | 9.9 ± 1.0 | 10.8 ± 1.1 | 10.1 ± 0.9 | 10.8 ± 1.1 | 11.9 ± 1.1 |
| Number of slices, left | 9.4 ± 1.2 | 10.4 ± 1.0 | 11.4 ± 1.0 | 10.5 ± 1.4 | 11.2 ± 1.0 | 12.1 ± 1.0 |
Data are presented as mean ± standard deviation, number (%), or median and interquartiles (25%, 75%).
eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; TKV, total kidney volume.
aTKV <219.9 cm3.
bTKV ≥322.2 cm3.
cTKV <301.0 cm3.
dTKV ≥415.2 cm3.
Risk factors for TKV
Stepwise logistic regression models were used to identify significant cross-sectional correlates of high and low TKV and are presented in Table 5. Risk factors for low TKV compared with those in the middle group included age per 10-years [odds ratio (OR) 1.67, P< 0.001], female (OR 1.67, P< 0.001) and eGFR <60 mL/min/1.73 m2 (OR 8.09, P< 0.001), while BSA (OR 0.24, P< 0.001) and diabetes (OR 0.47, P= 0.04) were protective. Risk factors of high TKV were BSA (OR 5.01, P< 0.001), diabetes (OR 2.01, P= 0.001), smoking (OR 1.94, P< 0.001) and albuminuria (OR 1.84, P= 0.005), while age per 10-years (OR 0.67, P< 0.001), female (OR 0.17, P< 0.001) and eGFR <60 mL/min/1.73 m2 (OR 0.19, P< 0.001) were protective.
Table 5.
Results of multivariable-adjusted stepwise logistic regression models for significant risk factors associated with high and low TKV in women and men
| Characteristic | High TKV |
Low TKV |
||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P-value | OR | (95% CI) | P-value | |
| Age (10-year) | 0.67 | (0.56–0.79) | <0.001 | 1.67 | (1.39–2.01) | <0.001 |
| Female | 0.17 | (0.12–0.24) | <0.001 | 6.96 | (4.39–11.05) | <0.001 |
| Body surface area (m2) | 5.01 | (4.07–6.17) | <0.001 | 0.24 | (0.19–0.32) | <0.001 |
| eGFR <60 mL/min/1.73 m2 | 0.19 | (0.07–0.48) | <0.001 | 8.09 | (4.85–13.50) | <0.001 |
| Diabetes | 2.01 | (1.33–3.32) | 0.001 | 0.47 | (0.23–0.95) | 0.04 |
| Smoking, current | 1.94 | (1.40–2.71) | <0.001 | Did not enter model | ||
| Albuminuria | 1.84 | (1.20–2.83) | 0.005 | Did not enter model | ||
TKV, total kidney volume; OR, odds ratio; 95% CI, 95% confidence interval, eGFR, estimated glomerular filtration rate.
TKV and chronic diseases
We examined the associations between low and high TKV and characteristics including hypertension, diabetes, eGFR <60 mL/min/1.73 m2 and albuminuria (Table 6). Compared with TKV in the middle group, low TKV had 6-fold higher odds of eGFR <60 mL/min/1.73 m2 (OR 6.12, P< 0.001) and 1.6-fold higher odds of albuminuria (OR 1.56, P= 0.03). There was no association between low TKV and either diabetes or hypertension. High TKV was associated with higher odds of diabetes (OR 2.15, P< 0.001) and lower odds of eGFR <60 mL/min/1.73 m2 (OR 0.32, P= 0.007). High TKV was not associated with either albuminuria or hypertension.
Table 6.
Results of multivariable-adjusted logistic regression models tested for significant associations between high and low TKV and hypertension, diabetes, chronic kidney disease and albuminuria
| Characteristic | High TKV |
Low TKV |
||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P-value | OR | (95% CI) | P-value | |
| Hypertensiona | 1.29 | (0.98–1.70) | 0.07 | 0.93 | (0.68–1.28) | 0.66 |
| Diabetesa | 2.15 | (1.43–3.24) | <0.001 | 0.75 | (0.42–1.34) | 0.33 |
| eGFR <60 mL/min/1.73 m2b | 0.32 | (0.14–0.74) | 0.007 | 6.12 | (3.86–9.69) | <0.001 |
| Albuminuriab | 1.45 | (0.97–2.17) | 0.07 | 1.56 | (1.03–2.37) | 0.03 |
TKV, total kidney volume; OR, odds ratio; 95% CI, 95% confidence interval.
aAdjusted for age, sex, body mass index and smoking.
bAdjusted for age, sex, body mass index, hypertension, diabetes and smoking.
DISCUSSION
In this cross-sectional study of community-dwelling adults, we first determined normative values for TKV from a healthy reference cohort and then identified thresholds for low and high TKV. We found that low TKV was associated with worse indices of renal function including albuminuria and eGFR <60 mL/min/1.73 m2 compared with normal TKV. There was also a progressive decline of TKV in women starting before the age of 50 and in men starting from the 50–59-year age group, suggesting a common pathway of volume loss. Risk factors for low TKV in both sexes included age and eGFR <60 mL/min/1.73 m2, consistent with the idea of reduced renal reserve. For high TKV, identified risk factors included diabetes, BSA and smoking but not eGFR <60 mL/min/1.73 m2. These findings coincided with the linear regression model showing that increased eGFR was a predictor of increased TKV as well as the logistic regression analysis demonstrating greater odds of diabetes and lower odds of eGFR <60 mL/min/1.73 m2 with high TKV. Neither low nor high TKV was associated with hypertension. Taken together, these findings establish a novel approach to renal structural analysis based on volume thresholds and demonstrate unique associations between TKV and CKD risk factors.
Volumetric study of kidney macroanatomy and associated CVD and CKD risk factors has been carried out previously [21]. A study of 539 selected patients (age 52 ± 18 years) who underwent contrast-enhanced CT showed that TKV was correlated with age, BSA and eGFR but not hypertension or diabetes [22]. The current study refines the relationship between TKV and diabetes to show an association with high TKV but not low TKV and confirms the lack of an association with hypertension.
Another study of 1344 potential kidney donors (age 44 ± 12 years) used contrast-enhanced CT to identify significant predictors of kidney parenchymal volume [7]. These included current smoking, BMI, GFR, serum glucose, albuminuria and HDL cholesterol. We found similar predictors of increased TKV with the notable exception of albuminuria. This could have been due to our use of UACR instead of 24-h urine collections used earlier. We also broadened the list of predictors to include hypertension, diabetes and increased triglycerides. Data from 493 elderly participants (age 79 ± 4 years) of the second AGES-Reykjavik Study showed an association between contrast-enhanced MRI-derived TKV and BMI, current smoking and diabetes [1]. Systolic and diastolic blood pressure did not enter the model and those observations are consistent with the present study. Overall, our findings extend the literature to show how clinically relevant covariates associate with well-defined high and low TKV in an unselected population.
Albuminuria, a marker of kidney damage, was previously reported to have a positive association with TKV. Those findings suggested that kidney enlargement was reflective of early injury as opposed to kidney volume reduction [1, 7]. In contrast to prior studies, we found an association between albuminuria and low TKV. There was no association between albuminuria and high TKV nor was UACR a significant predictor of increased TKV using linear regression analysis. However, albuminuria was observed in participants with high TKV (Table 4) and our results might be due to different populations, sex-specific definitions for albuminuria and collection methods.
eGFR is another marker used to evaluate kidney function that maintains a similar direction of association with TKV across most studies [1, 7, 22]. The current results showing an association between low TKV and eGFR <60 mL/min/1.73 m2 further explore this relationship at a more granular level not previously defined. Thus, an association between low TKV with albuminuria and eGFR <60 mL/min/1.73 m2 supports the concept that small kidneys are reflective of established kidney disease.
TKV is a function of the number and size of nephrons [23]. However, an increase in kidney size is not thought to be due to hyperplasia or glomerular neogenesis but rather due to hypertrophy [24]. Within the nephron, recent evidence showed that glomerular as well as tubular hypertrophy is associated with clinical characteristics including BMI and albuminuria [25]. At the same time, nephrosclerosis is a volume-losing lesion associated with age [26]. These microstructural changes of volume expansion and volume loss are reflected at the macrostructural level and are important to consider when interpreting TKV. For example, our observation that older age is protective for high TKV could be a consequence of increased nephrosclerosis rather than decreased nephron hypertrophy. The net effect of these microscopic changes on kidney volume is an area of ongoing investigation [27].
Increased kidney volume is also associated with glomerular hyperfiltration [7]. Various clinical conditions are characterized by both hyperfiltration and enlarged kidneys including diabetes, pregnancy, smoking and obesity. In the current study, diabetes, smoking and BSA were all risk factors for high TKV and there were increased odds of diabetes with high TKV. These findings are in accord with earlier observations and suggest that high TKV may be a macroscopic reflection of altered nephron function with attendant structural changes in glomerular and tubular volume. Identifying the unique drivers of nephron hypertrophy and variation among the different nephron segments may support our understanding of TKV.
Low TKV appears to be consistent with reduced renal functional reserve based on the associations with eGFR <60 mL/min/1.73 m2 and albuminuria. The idea of a smaller kidney and increased cortical echogenicity correlating with kidney disease is a well-accepted ultrasound finding [28]. CKD is generally considered to be injury through a common pathway of interstitial fibrosis, sclerotic glomeruli and tubular atrophy. The expansive nature of fibrosis and accumulation of scar tissue may progress to chronic and destructive pathological changes coupled with microvasculature rarefaction resulting in volume loss and renal dysfunction [29].
Our study posits that understanding kidney morphology may help identify individuals at risk for developing CVD and CKD. However, a central question when discussing TKV is the uncertainty of normal versus abnormal. Two earlier studies reported age- and sex-specific distributions of TKV across 10-year age groups from selected populations using CT [7, 22]. The current findings are consistent in showing that TKV in women starts to decrease at an earlier age compared with men, but our values are larger and this may be due to the differences in participant characteristics, image acquisition and analysis protocols.
It remains unclear whether high TKV represents an early adaptive and functional response with delayed maladaptive consequences [30]. A single longitudinal study measured kidney volume changes in the setting of weight loss therapy and this demonstrated the dynamic nature of TKV [6]. The study of 18 obese subjects had TKV measured by CT at baseline and then at 1 and 6 months after undergoing gastric bypass, gastric banding or diet initiation. All subjects lost weight and kidney volume was reduced in all but two patients. Measuring TKV and using set thresholds offers guidance to help interpret volumetric studies.
Strengths of our study include the large sample size from a community-dwelling population with a clinically relevant age range, robust risk factor ascertainment and use of noncontrast MRI coupled with manual segmentation. Limitations of the study include the cross-sectional study design and consequent lack of causal inference. Also, the study group was primarily non-Hispanic White Americans; our results may not be generalizable to other groups. Sex-specific thresholds for high and low TKV were not age-specific and this may be important in future studies given the observed reduction of TKV with age. A final limitation was our inability to distinguish cortical volume from medullary volume. This is particularly notable given findings that reduced TKV with age is attenuated by initial medullary volume increases.
In conclusion, this study demonstrates that low TKV is a structural marker of renal dysfunction and high TKV is associated with diabetes. Prospective studies are needed to characterize the natural progression of TKV with respect to CKD and CVD risk factors so as to confirm whether this structural biomarker portends future adverse outcomes.
CONFLICT OF INTEREST STATEMENT
None declared.
ACKNOWLEDGEMENTS
The Framingham Heart Study is conducted and supported in collaboration with Boston University by the National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH) project number N01HC25195; additional funding was provided by the NIH through the National Institute of Diabetes and Digestive and Kidney Diseases project number T32DK007053 to D.A.R. and NHLBI project number R01HL70279 to W.J.M., N.O.-M. and M.L.C.
REFERENCES
- 1. Woodard T, Sigurdsson S, Gotal JD. et al. Segmental kidney volumes measured by dynamic contrast-enhanced magnetic resonance imaging and their association with CKD in older people. Am J Kidney Dis 2015; 65: 41–48 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Cheung CM, Shurrab AE, Buckley DL. et al. MR-derived renal morphology and renal function in patients with atherosclerotic renovascular disease. Kidney Int 2006; 69: 715–722 [DOI] [PubMed] [Google Scholar]
- 3. Poggio ED, Hila S, Stephany B. et al. Donor kidney volume and outcomes following live donor kidney transplantation. Am J Transplant 2006; 6: 616–624 [DOI] [PubMed] [Google Scholar]
- 4. Alam A, Dahl NK, Lipschutz JH. et al. Total kidney volume in autosomal dominant polycystic kidney disease: a biomarker of disease progression and therapeutic efficacy. Am J Kidney Dis 2015; 66: 564–576 [DOI] [PubMed] [Google Scholar]
- 5. Kline TL, Korfiatis P, Edwards ME. et al. Automatic total kidney volume measurement on follow-up magnetic resonance images to facilitate monitoring of autosomal dominant polycystic kidney disease progression. Nephrol Dial Transplant 2016; 31: 241–248 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Johnson S, Rishi R, Andone A. et al. Determinants and functional significance of renal parenchymal volume in adults. Clin J Am Soc Nephrol 2011; 6: 70–76 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Wang X, Vrtiska TJ, Avula RT. et al. Age, kidney function, and risk factors associate differently with cortical and medullary volumes of the kidney. Kidney Int 2014; 85: 677–685 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Emamian SA, Nielsen MB, Pedersen JF. Intraobserver and interobserver variations in sonographic measurements of kidney size in adult volunteers. A comparison of linear measurements and volumetric estimates. Acta Radiol 1995; 36: 399–401 [PubMed] [Google Scholar]
- 9. Bakker J, Olree M, Kaatee R. et al. Renal volume measurements: accuracy and repeatability of US compared with that of MR imaging. Radiology 1999; 211: 623–628 [DOI] [PubMed] [Google Scholar]
- 10. Coulam CH, Bouley DM, Sommer FG. Measurement of renal volumes with contrast-enhanced MRI. J Magn Reson Imaging 2002; 15: 174–179 [DOI] [PubMed] [Google Scholar]
- 11. Pedersen M, Karstoft K, Lodrup A. et al. Advantages and controversies in the era of intrarenal volumetry. Am J Nephrol 2011; 33(Suppl 1): 40–45 [DOI] [PubMed] [Google Scholar]
- 12. Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann N Y Acad Sci 1963; 107: 539–556 [DOI] [PubMed] [Google Scholar]
- 13. Feinleib M, Kannel WB, Garrison RJ. et al. The Framingham Offspring Study. Design and preliminary data. Prev Med 1975; 4: 518–525 [DOI] [PubMed] [Google Scholar]
- 14. Oyama N, Gona P, Salton CJ. et al. Differential impact of age, sex, and hypertension on aortic atherosclerosis: the Framingham Heart Study. Arterioscler Thromb Vasc Biol 2008; 28: 155–159 [DOI] [PubMed] [Google Scholar]
- 15. Cohen EI, Kelly SA, Edye M. et al. MRI estimation of total renal volume demonstrates significant association with healthy donor weight. Eur J Radiol 2009; 71: 283–287 [DOI] [PubMed] [Google Scholar]
- 16. Zollner FG, Svarstad E, Munthe-Kaas AZ. et al. Assessment of kidney volumes from MRI: acquisition and segmentation techniques. AJR Am J Roentgenol 2012; 199: 1060–1069 [DOI] [PubMed] [Google Scholar]
- 17. Mosteller RD. Simplified calculation of body-surface area. N Engl J Med 1987; 317: 1098. [DOI] [PubMed] [Google Scholar]
- 18. Fox CS, Larson MG, Leip EP. et al. Predictors of new-onset kidney disease in a community-based population. JAMA 2004; 291: 844–850 [DOI] [PubMed] [Google Scholar]
- 19. Levey AS, Stevens LA, Schmid CH. et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009; 150: 604–612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Mattix HJ, Hsu CY, Shaykevich S. et al. Use of the albumin/creatinine ratio to detect microalbuminuria: implications of sex and race. J Am Soc Nephrol 2002; 13: 1034–1039 [DOI] [PubMed] [Google Scholar]
- 21. Emamian SA, Nielsen MB, Pedersen JF. et al. Kidney dimensions at sonography: correlation with age, sex, and habitus in 665 adult volunteers. AJR Am J Roentgenol 1993; 160: 83–86 [DOI] [PubMed] [Google Scholar]
- 22. Gong IH, Hwang J, Choi DK. et al. Relationship among total kidney volume, renal function and age. J Urol 2012; 187: 344–349 [DOI] [PubMed] [Google Scholar]
- 23. Nyengaard JR, Bendtsen TF. Glomerular number and size in relation to age, kidney weight, and body surface in normal man. Anat Rec 1992; 232: 194–201 [DOI] [PubMed] [Google Scholar]
- 24. Larsson L, Aperia A, Wilton P. Effect of normal development on compensatory renal growth. Kidney Int 1980; 18: 29–35 [DOI] [PubMed] [Google Scholar]
- 25. Elsherbiny HE, Alexander MP, Kremers WK. et al. Nephron hypertrophy and glomerulosclerosis and their association with kidney function and risk factors among living kidney donors. Clin J Am Soc Nephrol 2014; 9: 1892–1902 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Rule AD, Amer H, Cornell LD. et al. The association between age and nephrosclerosis on renal biopsy among healthy adults. Ann Intern Med 2010; 152: 561–567 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Denic A, Alexander MP, Kaushik V. et al. Detection and clinical patterns of nephron hypertrophy and nephrosclerosis among apparently healthy adults. Am J Kidney Dis 2016; (doi:10.1053/j.ajkd.2015.12.029, Epub ahead of print) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Hricak H, Cruz C, Romanski R. et al. Renal parenchymal disease: sonographic-histologic correlation. Radiology 1982; 144: 141–147 [DOI] [PubMed] [Google Scholar]
- 29. Ohashi R, Kitamura H, Yamanaka N. Peritubular capillary injury during the progression of experimental glomerulonephritis in rats. J Am Soc Nephrol 2000; 11: 47–56 [DOI] [PubMed] [Google Scholar]
- 30. Hostetter TH, Olson JL, Rennke HG. et al. Hyperfiltration in remnant nephrons: a potentially adverse response to renal ablation. Am J Physiol 1981; 241: F85–F [DOI] [PubMed] [Google Scholar]