Associations of Glomerular Number and Birth Weight With Clinicopathological Features of African Americans and Whites

ORIGINAL INVESTIGATIONS Pathogenesis and Treatment of Kidney Disease Associations of Glomerular Number and Birth Weight With Clinicopathological Features of African Americans and Whites Michael D. Hughson, MD,1 Glenda C. Gobe, PhD,2 Wendy E. Hoy, MD,3 R. Davis Manning Jr, PhD,4 Rebecca Douglas-Denton, BSc,5 and John F. Bertram,PhD5 Background: Hypertension and its cardiovascular complications affect African Americans more severely than whites, a disparity variously ascribed to low birth weight, low glomerular number, an exaggerated arteriolonephrosclerotic blood pressure response, and inflammation-induced oxidative stress. Study Design: Case series. Setting and Participants: Autopsy kidneys of 107 African Americans and 87 whites aged 18 to 65 years at a single medical center between 1998 and 2005. Excluded were persons with known premorbid kidney disease; pathological findings of severe arterioarteriolonephrosclerosis, nodular and diffuse diabetic glomerulosclerosis, or nonischemic cardiomyopathy. Predictors & Outcomes: Associations of: (1) race, age, sex, birth weight, obesity, and glomerular number (predictors) with hypertension and death from coronary artery (CAD) and cerebrovascular disease (CVD; outcomes); and (2) age, blood pressure, and race (predictors) with arteriolonephrosclerotic changes, including chronic tubulointerstitial inflammation (outcomes). Measurements: Hypertension ascertained from chart review and heart weight. Cause of death determined from chart review and autopsy findings. Birth weight obtained from birth records (115 persons). Total glomerular number (Nglom) estimated by using the dissector/fractionator technique. Arteriolosclerosis, glomerulosclerosis, cortical fibrosis, and chronic inflammation by using CD68 density were measured morphometrically. Results: 59 African Americans (55%) and 32 whites (37%) were classified as hypertensive. CAD and CVD were the cause of death in 64 (33%) and 18 persons (9%), respectively. By using multiple linear regression, birth weight (P ⬍ 0.001) and sex (P ⬍ 0.01), but not race (P ⫽ 0.3) or age (P ⫽ 0.2), predicted Nglom (P ⬍ 0.001; adjusted r2 ⫽ 0.176). Hypertension was associated with African American race (P ⫽ 0.04), older age (P ⬍ 0.001), and male sex (P ⫽ 0.01), but not with Nglom (P ⫽ 0.9), body mass index (P ⫽ 0.9), or birth weight (P ⫽ 0.4). Hypertension was the only significant factor associated with CAD and CVD (P ⬍ 0.001 for both). Interactions of age and blood pressure with race showed that although African Americans had more severe hypertension (P ⬍ 0.001) and arteriolosclerosis (P ⫽ 0.01) at a younger age than whites, there were no significant racial differences in degrees of arteriolosclerosis, glomerulosclerosis, cortical fibrosis, or CD68 density for any level of increased blood pressure. Limitations: The study is observational and descriptive. Conclusions: The more severe hypertension found in African Americans could not be attributed to racial differences in Nglom or birth weight. CAD and CVD death and increased arteriolonephrosclerosis, including CD68 density, were determined by using blood pressure without a significant interacting contribution from race. Am J Kidney Dis 52:18-28. © 2008 by the National Kidney Foundation, Inc. INDEX WORDS: Hypertension; race; arteriolonephrosclerosis; glomerular number; birth weight.

H

ypertension is a major risk factor for coronary artery disease (CAD) and cerebrovascular disease (CVD) and a common cause of chronic kidney disease.1,2 Recent clinical investigations have also shown that CAD strongly pre-

dicted current and future chronic kidney disease, establishing a bidirectional link between chronic diseases of both heart and kidney.3 Hypertension and arteriosclerosis have been identified as inflammatory states and their patho-

1 From the Department of Pathology, University of Mississippi Medical Center, Jackson, MS; 2Molecular and Cellular Pathology and 3Centre for Chronic Disease, University of Queensland, Brisbane, Australia; 4Department of Physiology and Biophysics, University of Mississippi Medical Center, Jackson, MS; and 5Department of Anatomy and Cell Biology, Monash University, Clayton, Victoria, Australia.

Received July 23, 2007. Accepted in revised form March 31, 2008. Originally published online as doi: 10.1053/j.ajkd.2008.03.023 on June 4, 2008. Address correspondence to Michael D. Hughson, MD, Department of Pathology, 2500 North State St, Jackson, MS 39216-4505. E-mail: [email protected] © 2008 by the National Kidney Foundation, Inc. 0272-6386/08/5201-0005$34.00/0 doi:10.1053/j.ajkd.2008.03.023

18

American Journal of Kidney Diseases, Vol 52, No 1 (July), 2008: pp 18-28

Glomerular Number in Human Hypertension

genesis was related to inflammation-induced oxidative stress.4-9 In experimental models and 1 clinical study, chronic inflammation has been linked to hypertension, and mycophenolate mofetil decreased blood pressure, presumably by suppressing inflammation.5-8,10-14 It also was proposed that essential hypertension might be the result of a nephron deficit.15-17 It was suggested that low nephron number prevents adequate excretion of dietary salt by pressure natriuresis at normal blood pressures, and increased renal perfusion pressures are needed to maintain salt and water homeostasis when salt intake is excessive. There is a direct correlation between birth weight and glomerular number, and populations with greater rates of low weight births have been identified as having greater rates of CVD and hypertension, the latter possibly on the basis of a congenital nephron deficit.18-21 The African American risk of developing hypertension and its cardiovascular complications is estimated to be 5 to 6 times greater than for whites.2,22 In renal biopsy studies, more severe arteriolosclerosis and arteriolonephrosclerosis have been described in African Americans compared with whites.23,24 The findings could not be attributed to differences in levels of hypertension and suggested that African Americans may have a more pronounced vascular response to high blood pressure. The prevalence of hypertension increases with obesity, and obesity has been shown to predict a greater likelihood for the development and progression of chronic kidney disease.25-27 The Southeast leads the United States in the prevalence of obesity, and the state of Mississippi has the largest percentage of citizens, both African American and white, classified as obese.28 The aims of this study are to investigate in an area of the United States in which obesity and hypertension are common the associations of: (1) age, sex, obesity, birth weight, glomerular number, and diagnosis of diabetes with hypertension status and death caused by CAD and CVD; and (2) the hypertension-related renal structural changes in arteriolosclerosis, glomerulosclerosis, and cortical fibrosis with age, blood pressure, and chronic tubulointerstitial inflammation, the latter assessed by using cortical CD68 density. In

19

both these aims, we are particularly interested in the effect of race.

METHODS This research used human autopsy tissue. The study was approved by the Institutional Review Board of the University of Mississippi Medical Center, Jackson, MS. Permission for autopsy was obtained from first of kin. Right kidneys were collected at autopsy at the University of Mississippi Medical Center from persons without known renal disease from 1998 through 2005. If both kidneys were grossly normal and equal or approximately equal in size, the right kidney was perfusion fixed with 10% buffered formalin. Kidneys showing arteriolonephrosclerosis were included in the study if the subcapsular cortex was slightly granular. To not have glomerular number unduly decreased by severe renal scarring and advanced glomerulosclerosis, kidneys were not collected if there were coarse pits or angular or depressed cortical scars. Kidneys from patients with diabetes were analyzed if they showed arteriolonephrosclerosis, but because of invariably advanced glomerular loss, kidneys were not analyzed if there was diffuse diabetic glomerulosclerosis with Kimmelstiel-Wilson nodules. Two cases of focal segmental glomerulosclerosis were excluded. Also excluded were cases of rheumatic heart disease, congenital heart disease (bicuspid aortic valves), hereditary hypertrophic cardiomyopathy, dilated cardiomyopathy, and cor pulmonale caused by primary pulmonary hypertension and pulmonary emphysema because normal cardiac function may not have been present, and heart weights would be increased independent of systemic blood pressure. The study population consisted of 107 African Americans (55 men, 52 women) and 87 whites (52 men, 35 women) 18 to 65 years of age. Causes of death were CAD, 33%; CVD, 9%; pulmonary embolus, 9%; neoplastic or infectious disease, 15%; pulmonary disease, 8%; accident, 14%; homicide, 7%; suicide, 1%; neurological disease not CVD, 2%; and unknown, 2%. Eleven subjects had a clinical diagnosis of type 2 diabetes, but did not have diabetic nephropathy diagnosable by using light microscopy. Blood pressures without medication were obtained from University of Mississippi Medical Center clinic, emergency department, or hospital records for 73 African Americans and 43 whites. Blood pressures from terminal hospital admissions were not used unless patients had a diagnosis of hypertension and blood pressures were increased. For patients classified as nonhypertensive, blood pressures were obtained within 3 years of the date of death. Mean arterial blood pressure (MAP) was calculated from an average of at least 3 blood pressure determinations as diastolic blood pressure plus one third of the difference between diastolic and systolic blood pressure. Patients were categorized on the basis of a history of hypertension, consistently increased blood pressure (ⱖ140/90 mm Hg), the presence of cardiomegaly, and severity of renal arteriolosclerosis as follows: (1) nonhypertensive: no history of hypertension, blood pressures not increased, and MAP of 106 mm Hg or less; (2) probably nonhypertensive: no history of hypertension, but no available blood pressure

20 record and 3 of 4 of the following: left ventricular wall thickness of 1.5 cm or less, heart weight of 450 g or less (ⱕ75% of nonhypertensive group 1 subjects), heart–body weight ratio of 5.09 g/kg or less (ⱕ75% of nonhypertensive group 1 subjects), and intimal thickening of interlobular arteries of 0.074 or less (ⱕ75% of nonhypertensive group 1 subjects); (3) probably hypertensive: history of hypertension, but no available blood pressure record and 3 of 4 of the following: left ventricular wall thickness greater than 1.5 cm, heart weight of 350 g or greater (ⱖ25% of hypertensive group 4 subjects), heart–body weight ratio of 4.53 g/kg or greater (ⱖ25% of hypertensive group 4 subjects), and intimal thickening of interlobular arteries of 0.02 or greater (ⱖ25% of hypertensive group 4 subjects); and (4) hypertensive: history of hypertension, repeatedly increased blood pressures, and MAP of 107 mm Hg or greater. For analysis, groups 1 and 2 were considered nonhypertensive and groups 3 and 4 were considered hypertensive. On this basis, 59 African Americans and 32 whites were classified as hypertensive and 48 African Americans and 55 whites were classified as nonhypertensive. Birth weights were obtained from the Public Health Statistics Division of the Mississippi Department of Health for 70 African Americans and 44 whites. The Department of Health did not record birth weights before 1951, making birth weights unavailable for persons older than 47 years in 1998 and 54 years in 2005, the inclusive years of kidney collection. Body mass index (BMI) and body surface area (BSA) were calculated from body length measured at autopsy and body weights obtained by using a full body scale. Perfused kidneys were bisected and immersed in 10% formalin. After 10 days, both halves of the kidney were cut into slices 4 mm thick, and every fourth slice of both halves was sampled for stereological examination, beginning with a first slice selected as a random number from 1 to 4. Selected slices were sent to Monash University, Clayton, Victoria, Australia, where they were processed for embedding in glycolmethacrylate for stereological estimation of total glomerular number (Nglom) and mean glomerular tuft volume by using the physical disector/fractionator combination. The methods have been described in detail previously.29-31 Representative kidney blocks from the upper pole and midportion of the kidney not sampled for stereological examination were paraffin embedded. These blocks were cut perpendicular to the cortical surface and contained the full thickness of the cortex and underlying medulla. Sections were cut 4 ␮m thick, stained with hematoxylin and eosin and periodic acid–Schiff-hematoxylin stains and with picrosirius red stains for fibrillar collagen. Counts of the percentage of obsolete glomeruli, proportion of hyalinized arterioles, and measurements of arterial intimal thickening were made by using periodic acid–Schiff-hematoxylin stained sections. Macrophages were manually stained by means of immunohistochemistry using Dako CD68 clone KP1 monoclonal antibody (Dako Corp, Carpinteria, CA) at a dilution of 1:200 at room temperature for 1 hour. Ten specimens from cases of hypertension were stained with prediluted antibodies for CD79a, CD3, CD8, and CD138 from Cell Marque Corp (Rocklin, CA) and CD4 from Vector Laboratories Corp (San Diego, CA) at a dilution of 1:80 using a Ventana automated stainer (Ventana Medical Systems, Inc, Tucson, AZ). Color

Hughson et al was generated with linking systems by means of diaminobenzidine. The severity of arterial intimal thickening was measured in interlobular arteries 90 to 250 ␮m in diameter as a ratio of the thickness of the intima to the outer wall diameter using the linear measurement function of Image-ProPlus morphometric software (Media Cybernetics Inc, Bethesda, MD). Cortical fibrosis was measured as the proportion of cortex staining red with the picrosirius stain using the automated Image-ProPlus area counting function. Macrophages were counted as nonglomerular CD68-positive cells within the renal cortex by using original magnification ⫻200 imported images into Image-ProPlus at a final magnification of ⫻400. The final counts were determined as the cortical density of CD68-positive cells per square millimeter. For cortical fibrosis and macrophage counts, a total of 36 nonoverlapping microscopic fields was examined moving from subcapsular to juxtamedullary cortex. Data were collected into Microsoft Excel (Microsoft Corp., Redmond, WA) and analyzed using Stata (StataCorp, College Station, TX) software. Pairwise relationships between variables were evaluated by means of Spearman rank-order correlation. Because of the relatively few cases of CVD, CAD and CVD deaths were grouped into 1 category of hypertension-associated illness as CAD-CVD. Logistic regression was used to evaluate the effects of the independent variables of age, sex, obesity, birth weight, glomerular number, and diagnosis of diabetes (predictors) on the categorical dependent variables of hypertension or CAD-CVD (outcomes). Linear regression was used to evaluate the influence of the interaction of race with age and MAP (predictors) on each of the continuous dependent variables of arterial intimal thickening, arteriolar hyalinization, cortical fibrosis, CD68 density, and glomerulosclerosis (outcomes). Differences between groups were analyzed by using t-test if data passed normality and equal variance tests and Mann-Whitney rank sum test if they did not. KolmogorovSmirnov tests were applied to continuous variables to determine whether they were normally distributed. Categorical variables were compared by using ␹2 or Fisher exact tests. For all statistical procedures, P less than 0.05 is considered significant.

RESULTS Clinical Characteristics Table 1 lists clinical features of hypertensive and nonhypertensive African Americans and whites. For both races, hypertensive subjects were older and had a greater proportion of deaths caused by CAD-CVD. The proportion of African Americans who were hypertensive was greater than for whites (P ⫽ 0.02), and the MAP of hypertensive African Americans was greater than for hypertensive whites (P ⫽ 0.02). The proportion of men for both races was greater in hypertensive subjects, but these sex differences were not statistically significant. Figure 1 shows the

Glomerular Number in Human Hypertension

21

Table 1. Clinical Features of the Study Cohort African American Nonhypertensive (n ⫽ 48)

Hypertensive (n ⫽ 59)

White

P

Nonhypertensive (n ⫽ 55)

⬍0.001

Hypertensive (n ⫽ 32)

P

Age (y)

34 (23-41)

46 (40-54)

41 (31-47)

49 (38-55)

Men

20 (41.7)

35 (59.3)

0.1

30 (54.6)

22 (68.8)

0.3

Body mass index (kg/m2) ⬍30 30-40 ⱖ40

30 ⫾ 6 30 (62.5) 15 (31.2) 3 (6.3)

30 ⫾ 8 35 (59.3) 15 (25.4) 9 (15.3)

0.6 0.9 0.8 0.3

29 ⫾ 7 35 (63.6) 15 (27.3) 5 (9.1)

31 ⫾ 10 18 (56.3) 9 (28.1) 5 (15.6)

0.3 0.9 0.9 0.5

3.29 ⫾ 0.55 n ⫽ 40

3.28 ⫾ 0.62 n ⫽ 30

0.9

3.28 ⫾ 0.59 n ⫽ 33

3.35 ⫾ 0.40 n ⫽ 11

0.7

94 (91-98) n ⫽ 34

120 (115-129)* n ⫽ 39

⬍0.001

95 (92-98) n ⫽ 26

113 (112.5-118) n ⫽ 17

⬍0.001

34 (57.6)

⬍0.01

12 (21.8)

9 (15.3)

0.01

Birth weight MAP (mm Hg) CAD or CVD death

8 (16.7)

Diabetes

0 (0)

1 (0)

19 (59.3) 1 (3)

0.02

0.03 0.7

Note: Values expressed as number (percent), mean ⫾ SD for normally distributed data, and median (25th to 75th percentiles) for non-normally distributed data. P reflects within-race comparisons of hypertensive and nonhypertensive subjects. Abbreviations: MAP, mean arterial pressure; CVD-CAD, death caused by cerebrovascular or coronary artery disease. *Significant between-race difference in MAP for hypertensive individuals (P ⫽ 0.02).

distribution of birth weights in African Americans and whites by hypertension status. Hypertensive African Americans had a greater range of birth weights than hypertensive whites, but there were no significant differences in average birth weights by race or hypertension status. For African Americans, 39% of subjects were obese and 11% were severely obese, with the proportions identical for whites. Although relationships of hypertension to obesity were not statistically significant, 57% of obese African Americans and 41% of obese whites were hypertensive. All African American patients with diabetes had hypertension.

Figure 1. Histograms show the distribution of birth weights for African Americans and whites by hypertension status.

Autopsy Findings and Correlates of CD68 Density Table 2 lists kidney pathological data, heart size, and Nglom by race and hypertension status. Hypertensive subjects had the stigmata of increased heart size and increased intrarenal arteriolosclerosis, glomerulosclerosis, cortical fibrosis, and cortical CD68 density, with no racial differences apparent. There were no statistical differences in Nglom between nonhypertensive and hypertensive subjects for either race. Distributions of Nglom by race and hypertension status are shown in Fig 2. There were 9 individuals who died of bacterial sepsis (2 African Americans and 7 whites) who had a mean cortical CD68 density of 44.6 cells/ mm2 (95% confidence interval, 26.4 to 62.8). No subject with a systemic infection had a history of hypertension, and average MAP was 93 mm Hg, with a range of 79 to 99 mm Hg. Hypertensive subjects had significantly greater cortical CD68 density than those who were nonhypertensive and did not have bacterial sepsis. Associations Among Glomerular Number, Birth Weight, and Body Size Spearman correlation for the 115 subjects who had birth weight available showed a significant direct correlation between birth weight and Nglom

Note: Values expressed as mean ⫾ SD for normally distributed data and median (25th to 75th percentiles) for non-normally distributed data. P reflects within-race comparisons of hypertensive and nonhypertensive subjects. Intimal thickening refers to interlobular arteries. Abbreviation: Nglom, total glomerular number of the right kidney. *Significant between-race difference in heart– body weight ratio (P ⫽ 0.02).

⬍0.001 ⬍0.001 0.4 ⬍0.01 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 490 (398-575) 5.60 (4.83-6.35) 841,069 ⫾ 291,903 2.07 (1.33-4.74) 0.043 (0.028-0.076) 0.79 (0.041-0.148) 0.098 (0.033-0.365) 15.0 (7.6-26.8) 350 (314-427) 4.71 (4.25-5.23)* 901,011 ⫾ 298,334 1.03 (0.54-1.98) 0.02 (0.009-0.044) 0.022 (0.00-0.052) 0.00 (0.00-0.058) 4.8 (2.3-12.9) Heart weight (g) Heart–body weight ratio (g/kg) Nglom Glomerulosclerosis (%) Cortical fibrosis Arterial intimal thickening Arteriolar hyalinization (%) CD68 (cells/mm2)

363 (300-430) 4.41 (3.68-4.95) 951,807 ⫾ 268,798 0.92 (0.30-1.98) 0.017 (0.006-0.051) 0.022 (0.00-0.070) 0.00 (0.00-0.045) 3.9 (2.0-8.3)

490 (400-559) 5.58 (4.97-6.30) 885,279 ⫾ 333,619 2.99 (1.29-6.57) 0.047 (0.024-0.107) 0.109 (0.059-0.171) 0.147 (0.031-0.302) 17.6 (9.9-31.5)

⬍0.001 ⬍0.001 0.3 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

P Hypertensive (n ⫽ 32) Nonhypertensive (n ⫽ 55) P Hypertensive (n ⫽ 59) Nonhypertensive (n ⫽ 48)

African American

White

Hughson et al

Table 2. Pathological Features of the Kidney and Cardiac Size for the Study Cohort

22

Figure 2. Histograms show the distribution of total right kidney glomerular number (Nglom) for African Americans and whites by hypertension status.

(r ⫽ 0.292; P ⬍ 0.01) in which low birth weight predicted a tendency to have fewer glomeruli. Birth weight also directly correlated with adult BSA (r ⫽ 0.342; P ⬍ 0.001) and, to a lesser degree, BMI (r ⫽ 0.208; P ⫽ 0.03). BSA directly incorporates height into the calculation and is a better measurement of hereditary size than BMI, which is more an index of acquired adiposity. BSA significantly and directly correlated with Nglom (r ⫽ 0.169; P ⫽ 0.02). Multiple linear regression showed that birth weight (P ⬍ 0.001) and sex (P ⬍ 0.01), but not race (P ⫽ 0.3) or age (P ⫽ 0.2), predicted Nglom (P ⬍ 0.001; adjusted r2 ⫽ 0.176). Table 3 lists sex differences in glomerular number and birth weight. Women, both African American and white, had approximately 15% fewer glomeruli than men, and African American women, although not white women, had significantly lower birth weights than men of the same race. As also listed in Table 3, Nglom was not statistically different between African Americans and whites for either sex. Nglom values adjusted for sex, age, and birth weight were 944,006 ⫾ 286,101 for African Americans and 891,130 ⫾ 287,151 for whites, also not statistically significant (P ⫽ 0.4). Because birth weights were not recorded by the Mississippi State Health Department until 1951, older subjects are underrepresented in the regression model for Nglom. Associations With Hypertension and Cardiovascular Death Between the ages of 18 and 47 years, 68 of 86 nonhypertensive (79%) and 35 of 47 hypertensive subjects (74%) had recorded birth weights,

Glomerular Number in Human Hypertension

23

Table 3. Nglom and Birth Weights of the Study Cohort Nglom

All subjects African American White P for race

Birth Weight

Women

Men

P for Sex

Women

Men

P for Sex

806,329 (n ⫽ 87) (660,278-967,976) 822,660 (n ⫽ 52) (682,664-973,943) 762,444 (n ⫽ 35) (651,793-967,976) 0.5

940,446 (n ⫽ 107) (734,049-1,174,939) 965,472 (n ⫽ 55) (726,091-1,198,257) 932,850 (n ⫽ 52) (748,378-1,146,986) 0.2

⬍0.001

3.14 (n ⫽ 55) (2.75-3.63) 3.10 (n ⫽ 36) (2.75-3.58) 3.30 ⫾ 0.45 (n ⫽ 19) 0.3

3.25 (n ⫽ 60) (3.18-3.75) 3.38 (n ⫽ 34) (3.18-3.77) 3.30 ⫾ 0.65 (n ⫽ 26) 0.2

0.03

0.02 0.02

0.02 0.9

Note: Values expressed as mean ⫾ SD for normally distributed data and median (25th to 75th percentiles) for non-normally distributed data. Abbreviation: Nglom, right kidney total number of glomeruli.

but between the ages of 48 and 65 years, only 5 of 17 nonhypertensive (29%) and 6 of 44 hypertensive subjects (14%) had recorded birth weights. In a model that included birth weight, logistic regression showed that hypertension was significantly predicted by African American race (P ⫽ 0.04), older age (P ⬍ 0.001), and male sex (P ⫽ 0.01), but not by BMI (P ⫽ 0.9), Nglom (P ⫽ 0.9), birth weight (P ⫽ 0.4; odds ratio, 0.686; 95% confidence interval, 0.291 to 1.619), or diabetes (P ⫽ 0.5) (P ⬍ 0.001, r2 ⫽ 0.170). This model is restricted to 114 subjects and does not include the large proportion of hypertensive subjects aged 48 to 65 years. The removal of birth weight analyzes all 194 subjects (Table 4). Results are similar to those including birth weight, but with African American race assuming greater significance (P ⫽ 0.001). Hypertension was not seen in these subjects until age 25 years for African Americans and age Table 4. Multivariate Analysis of Factors Associated With the Presence of Hypertension

30 years for whites. Adjustment of Nglom by sex, race, and age shows, as before adjustment, that hypertensive subjects have fewer glomeruli than those who were nonhypertensive, but that differences are not significant (Nglom hypertensive, 902,483 ⫾ 291,301; nonhypertensive, 852,088 ⫾ 290,808; P ⫽ 0.2). In logistic regression that includes birth weight, CAD-CVD was significantly predicted by hypertension (P ⬍ 0.001), but not by the other variables, with only age (P ⫽ 0.05) approaching significance (P ⬍ 0.001; r2 ⫽ 0.203). Race (P ⫽ 0.9), birth weight (P ⫽ 0.2; odds ratio, 0.575; 95% confidence interval, 0.233 to 1.416), sex (P ⫽ 0.2), BMI (P ⫽ 0.2), Nglom (P ⫽ 0.5), and diabetes (P ⫽ 0.7) were not significant predictors of CAD-CVD. The full model of 194 subjects that omits birth weight also showed that only hypertension predicted CAD-CVD (Table 5). Table 5. Multivariate Analysis of Factors Associated With Death From Coronary Artery or Cerebrovascular Disease Hypertension

Hypertension Odds Ratio

African American race Age (/1-y increase) Men Body mass index (/1-kg/m2 increase) Nglom (/10,000 glomeruli) Diabetes

P

95% Confidence Interval

3.351 0.001 1.087 ⬍0.001 2.133 0.03

1.660-6.678 1.052-1.222 1.060-4.291

1.020 0.999 5.565

0.977-1.064 0.999-1.000 0.589-52.546

0.4 0.3 0.1

Note: r2 for the model ⫽ 0.21. Abbreviation: Nglom, number of glomeruli per right kidney.

African American race Age (/1-y increase) Men Body mass index (/1-kg/m2 increase) Nglom (/10,000 glomeruli) Hypertension Diabetes

Odds Ratio

P

0.951 1.013 1.563

0.9 0.4 0.2

1.039 0.07 0.999 0.5 4.90 ⬍0.001 0.780 0.7

95% Confidence Interval

0.477-1.899 0.983-1.045 0.778-3.139 0.997-1.084 0.999-1.000 2.367-10.141 0.206-2.953

Note: r2 for the model ⫽ 0.146. Abbreviation: Nglom, number of glomeruli per right kidney.

24

Hughson et al

Table 6. Spearman Pairwise Correlations Between Age, Glomerular Number, Mean Arterial Blood Pressure, and Pathological Features of Arteriolonephrosclerosis

Age

Nglom

% GS

CF

It

Hyaline

CD68

MAP

⫺0.149 0.04

0.569 <0.001 ⴚ0.245 <0.001

0.434 <0.001 ⴚ0.174 <0.02 0.662 <0.001

0.517 <0.001 ⫺0.098 0.2 0.566 <0.001 0.525 <0.001

0.364 <0.001 ⫺0.025 0.7 0.455 <0.001 0.422 <0.001 0.541 <0.001

0.397 <0.001 ⫺0.119 0.1 0.522 <0.001 0.520 <0.001 0.479 <0.001 0.406 <0.001

0.417 <0.001 ⫺0.007 0.9 0.436 <0.001 0.446 <0.001 0.589 <0.001 0.498 <0.001 0.582 <0.001

Nglom % GS CF It Hyaline CD68

Note: Upper values are unadjusted correlation coefficients and lower values are probabilities. Significant relationships are in bold type. Cases of bacterial sepsis were omitted. Abbreviations: Nglom, number of glomeruli; % GS, percent glomerulosclerosis; CF, cortical fibrosis; It, intimal thickening of interlobular arteries; Hyaline, arteriolar hyalinization; CD68, CD68 staining density; MAP, mean arterial blood pressure.

Effects of Race, Age, and Blood Pressure on Pathological Findings Table 6 lists Spearman pairwise correlations between the pathological features of arteriolonephrosclerosis and CD68 density for 185 subjects (omitting those with bacterial sepsis) and MAP for 110 subjects. All variables directly correlated with each other and significantly increased with age. With increased age and glomerulosclerosis, there was a significant decrease in glomeruli. The interaction of race in linear regression with MAP showed that the pathological features of arteriolonephrosclerosis were strongly related to increased blood pressure (each feature, P ⬍

0.001) without significant contributions from race (glomerulosclerosis, P ⫽ 0.2; cortical fibrosis, P ⫽ 0.5; arterial intimal thickening, P ⫽ 0.9; arteriolar hyalinization, P ⫽ 0.1; and CD68 density, P ⫽ 0.3; Table 7). In models evaluating only race and age, but not MAP, both race and age contributed significantly to glomerulosclerosis, cortical fibrosis, and arterial intimal thickening. The contribution of race was most pronounced for arterial intimal thickening, with the intimal thickening ratio increasing at a rate of 0.36/y for whites (adjusted r2 ⫽ 0.133; P ⬍ 0.001) and 0.68/y for African Americans (adjusted r2 ⫽ 0.22; P ⬍ 0.001).

Table 7. Relationship of Race, Age, and Their Interaction and Race, MAP, and Their Interaction With Pathological Features of Arteriolonephrosclerosis

Age

Interaction P

Model r2

African American Race

0.138 (⬍0.001) 0.002 (⬍0.001)

0.04 0.06

0.18 0.15

0.035 (0.001) 0.003 (⬍0.001)

0.1

0.040 (0.09) 2.43 (0.4)

0.03 0.03

African American Race

Glomerulosclerosis Cortical fibrosis Arterial intimal thickening Arteriolar hyalinization CD68 density

1.519 (0.01) 0.020 (0.01)

0.004 (⬍0.001) 0.348 (0.01)

Interaction P

Model r2

0.950 (0.2) 0.106 (⬍0.001) 0.007 (0.5) 0.002 (⬍0.001)

0.8 0.9

0.15 0.19

0.21

⫺0.003 (0.9) 0.004 (⬍0.001)

0.7

0.36

0.11 0.05

⫺0.046 (0.1) 0.006 (⬍0.001) ⫺4.243 (0.3) 0.587 (⬍0.001)

0.3 0.9

0.29 0.12

MAP

Note: Coefficients from main-effects (noninteraction term) models are presented for the terms race and age or race and MAP (P). Coefficients for the statistically significant interaction terms were 0.093 for glomerulosclerosis, 0.004 for arteriolar hyalinization, and 0.532 for CD68 density. Model r2 is adjusted and reflects interactions. Abbreviation: MAP, mean arterial pressure.

Glomerular Number in Human Hypertension

Figure 3. Plot of fitted values in linear regression shows the correlation between age and mean arterial pressure (MAP) for African Americans (r ⫽ 0.558; P ⬍ 0.001) and whites (r ⫽ 0.153; P ⫽ 0.3). With ageing, African Americans had greater increases in blood pressure than whites.

There was a direct correlation between age and MAP that was highly significant for African Americans (r ⫽ 0.406; P ⬍ 0.001), but was weak and not significant for whites (r ⫽ 0.153; P ⫽ 0.3). Figure 3 shows these relationships in which linear regression predicted a gain in MAP of 0.85 mm Hg/y for African Americans (adjusted r2 ⫽ 0.302), but only 0.19 mm Hg/y for whites (adjusted r2 ⫽ 0.001). Nevertheless, the relationship between arterial intimal thickening and MAP was similar for both races (African Americans, r ⫽ 0.592; P ⬍ 0.001; whites, r ⫽ 0.645; P ⬍ 0.001). For each incremental increase in MAP, there was little racial difference in the severity of arteriolosclerosis (Fig 4).

25

hypertensive. Data also showed that African American race, male sex, and increased age, but not Nglom, birth weight, or BMI, were the significant predictors of hypertension. More severe and more prevalent hypertension are characteristic of community findings for African Americans within the Southeastern and other regions of the United States and have suggested there may be racial differences in the pathogenesis of hypertension or common mechanisms may be race sensitive.32-34 Progression of arteriolonephrosclerosis involves the interrelated processes of arteriolosclerosis, glomerulosclerosis, and cortical fibrosis with tubular and glomerular loss that increase with age and the increase in blood pressure. A greater degree of glomerulosclerosis and arteriolosclerosis was seen in hypertensive African Americans, but differences were not significant in the 2-way comparisons listed in Table 2. Multivariate analysis of the interactions of race, age, and MAP on the pathological features of arteriolonephrosclerosis showed that African Americans developed more severe arteriolonephrosclerosis at younger ages than whites, but this occurred in conjunction with greater levels of blood pressure. The interaction of race and MAP showed that for any level of increased blood pressure, there were no significant racial differences in any of the pathological features of arteriolonephrosclerosis. The relationship between MAP and arterial

Pathological Examination Examples of pathology-defining findings described in this report are available as online supplementary material at www.ajkd.org. These examples include findings associated with mild (Fig S1) and severe hypertension (Fig S2), as well as CD68 staining in septic individuals; nonseptic nonhypertensive individuals; and nonseptic hypertensive individuals (Figs S3A to C, respectively). Additional interstitial changes in hypertensive individuals are shown in Fig S4.

DISCUSSION In this study, hypertensive African Americans had a greater MAP than hypertensive whites, and a larger proportion of African Americans were

Figure 4. Linear regression shows the correlation between mean arterial pressure (MAP) and fitted values for intimal thickening of interlobular arteries for African Americans (r ⫽ 0.592; P ⬍ 0.001) and whites (r ⫽ 0.645; P ⬍ 0.001). Racial differences in the degree of intimal thickening for incremental increases in blood pressure were not significantly different.

26

intimal thickening was found to be very similar for each race and provided no indication that there was a difference between African Americans and whites in the structural response of small arteries to hypertension. Data also indicated that the most important determinant of death caused by CAD-CVD was hypertension, with race having little contribution to this outcome. In a previous study,18 a relationship between low nephron number and hypertension could not be shown for African Americans, but hypertensive whites appeared to have fewer glomeruli than nonhypertensive whites, a finding also observed by Keller et al16 in Europeans. With a larger number of cases, we now find that although hypertensive African Americans and whites had on average fewer glomeruli than nonhypertensive subjects of the same race, differences were not statistically significant. By analyzing more kidneys from women, we were able to show that they had approximately 15% fewer glomeruli than men, and Nglom made no significant contribution to hypertension for either sex or race. We previously reported a direct correlation between Nglom and birth weight in which linear regression estimated a gain of approximately 260,000 glomeruli for each 1-kg increase in birth weight.19 Nevertheless, there was wide variability for Nglom at all levels of birth weight.19 In the present study, a similar relationship was present that appeared in large part to be the result of smaller individuals having lighter birth weights and fewer glomeruli and larger individuals having heavier birth weights and more glomeruli. The significance of the correlation also was dependent on sex. Subjects in this study were born at or near term, and only 7 had birth weights less than 2.5 kg, with 5 of the 7 female. Data showed no significant relationship between low-term or near-term birth weights and hypertension, but it must be noted that the findings have little or no bearing on the effect that intrauterine growth retardation might have on glomerular number or adult blood pressure.35 The kidneys in this autopsy study were essentially normal for age, and our collection criteria excluded subjects who had chronic kidney diseases other than arteriolonephrosclerosis. The increases in MAP and severity of arteriolonephro-

Hughson et al

sclerosis emphasize that hypertension and hypertensive changes are strongly related to age. Essential hypertension is rare in young adults, but increases to 65% of the US population between 65 and 74 years of age.32 On the basis of a birth-weight–associated nephron deficit, the high prevalence of hypertension in older age groups would be difficult to explain by the comparatively lower rates of low birth weight of 13.0% for African Americans and 6.5% for whites.36 Some, but not all, clinical investigations have shown a relationship between low birth weight and the later increase in blood pressure in children and adults that is suggested to be the result of fetal growth retardation and low nephron number.20-22,33,34,37,38 Studies showing such a relationship are primarily from Europe, and there is little evidence that this is true for African Americans.16,17 Two recent investigations found that birth weight directly, rather than inversely, correlated with systolic blood pressure in African American children at 5 and 7 years of age, and in both studies, factors related to maternal poverty were important predictors of the increased blood pressure.33,34 In regard to inflammation, we found that renal CD68 density unadjusted for interactions with other features of arteriolonephrosclerosis was significantly increased in hypertensive kidneys, and CD68-positive cells were localized primarily to areas of cortical fibrosis and glomerular obsolescence. It also was found that systemic infection resulted in an accumulation of large numbers of macrophages within the kidneys, but cortical fibrosis was not present and patients were not hypertensive. These observations suggest that if inflammation has a role in hypertension, it is likely to be an integral part of arteriolonephrosclerosis and to augment, rather than initiate, the increase in blood pressure. Although we could not find a statistically significant relationship between obesity and hypertension in these autopsy subjects, 39% of both African Americans and whites were obese and 57% of obese African Americans and 41% of obese whites were hypertensive. This reflects both the 31.6% obesity rate reported in Mississippi and the high prevalence of coexisting hypertension.39,40 Rather than low nephron number, if a link exists between low birth weight and adult hyper-

Glomerular Number in Human Hypertension

tension, an alternate explanation may be that unusually high rates of low weight birth are a general marker of a population’s poor health. Beyond the neonatal period, these markers include recurrent childhood and adult infections, alcoholism, cigarette smoking, obesity, and poor diets high in refined carbohydrates, all common characteristics of groups at high risk of hypertension, renal disease, and CVD.41-44 It is postulated that subtle inflammatory damage precedes the development of hypertension and initiates a cycle of continuing renal damage.14 The high CD68 counts found in subjects with bacterial sepsis show that the kidney is affected by infections originating at other body sites. Seemingly localized conditions, such as dental carries, impetigo, scabies, and cutaneous abscesses, that are common in high-risk populations are low-grade systemic illnesses and may contribute to renal injury that promotes later hypertension.

ACKNOWLEDGEMENTS The findings were presented in part at the 2004 Annual Meeting of the American Society of Nephrology, October 23-November 1, St Louis, MO. Support: This research was funded by grants from the National Institutes of Health (NIH 1 R01 DK065970-01), NIH Center of Excellence in Minority Health (5P20M000534-02), The National Medical Research Council of Australia, Janssen-Cilag Australia Pty Ltd, and the American Heart Association (Southeastern Affiliate). Financial Disclosure: None.

SUPPLEMENTARY MATERIALS Figure S1: An interlobular artery and afferent arterioles visualized by using periodic acid–Schiff-hematoxylin staining of paraffin-embedded sections from a kidney from an individual with mild hypertension. Figure S2: An interlobular artery and afferent arterioles visualized by using periodic acid–Schiff-hematoxylin staining of paraffin-embedded sections from a kidney from an individual with severe hypertension. Figure S3: Immunohistochemical detection of CD68 in the renal cortex of a septic individual, a nonseptic nonhypertensive individual, and a nonseptic hypertensive individual. Figure S4: Interstitial changes in the renal cortex of a hypertensive individual. Note: The supplementary material accompanying this article (doi: 10.1053/j.ajkd.2008.03.023) is available at www.ajkd.org.

REFERENCES 1. Klag MJ, Whelton PK, Randall BL, Neaton JD, Brancati FL, Stamler J: End-stage renal disease in African Ameri-

27 can and white men. 16 Year MRFIT findings. JAMA 277:1293-1298, 1997 2. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY: Chronic kidney disease and the risk of death, cardiovascular events, and hospitalization. N Engl J Med 351:1296-1305, 2004 3. McClellan WM, Langston RD, Presley R: Medicare patients with cardiovascular disease have a high prevalence of chronic kidney disease and a high rate of progression to end-stage renal disease. J Am Soc Nephrol 15:1912-1919, 2004 4. Labarre CA, Zolaga GP: C-Reactive protein: From innocent bystander to pivotal mediator of atherosclerosis. Am J Med 117:499-507, 2004 5. Zoccali C, Maio R, Tripepi G, Mallamaci F, Perticone F: Inflammation as a mediator of the link between mild to moderate renal insufficiency and endothelial dysfunction in essential hypertension. J Am Soc Nephrol 17:S64-S68, 2006 (suppl 2) 6. McCormick MM, Rahimi F, Bobryshev YV, et al: S100A8 and S100A9 in human arterial wall. Implications for atherogenesis. J Biol Chem 280:41521-41529, 2005 7. Kluth DC, Lars-Peter E, Rees A: Multiple facets of macrophages in renal injury. Kidney Int 66:542-557, 2004 8. Eardly KS, Cockwell P: Macrophages and progressive tubulointerstitial disease. Kidney Int 68:437-455, 2005 9. Vaziri ND, Rodriquez-Iturbe B: Mechanisms of disease: Oxidative stress and inflammation in the pathogenesis of hypertension. Nat Clin Pract Nephrol 2:582-593, 2006 10. Rodriguez-Iturbe B, Pons H, Quiroz Y, et al: Mycophenolate mofetil prevents salt-sensitive hypertension resulting from angiotensin II exposure. Kidney Int 59:2222-2232, 2001 11. Tian N, Gu JW, Jordan S, Rose RA, Hughson MD, Manning RD Jr: Immune suppression prevents renal damage and dysfunction and reduces arterial pressure in saltsensitive hypertension. Am J Physiol Heart Circ Physiol 292:H1018-H1025, 2007 12. Rodriguez-Iturbe B, Quiroz Y, Shahkarami A, Li Z, Vaziri ND: Mycophenolate mofetil ameliorates nephropathy in the obese Zucker rat. Kidney Int 68:1041-1047, 2005 13. Johnson RJ, Herrera-Acosta J, Schreiner GF, Rodriguez-Iturbe B: Subtle acquired renal injury as a mechanism of salt-sensitive hypertension. N Engl J Med 346:913-923, 2002 14. Herrera J, Ferrebuz A, Macgregor EG, RodriguezIturbe B: Mycophenolate mofetil treatment improves hypertension in patients with psoriasis and rheumatoid arthritis. J Am Soc Nephrol 17:S218-S225, 2006 (suppl 3) 15. Brenner BM, Garcia DL, Anderson S: Glomeruli and blood pressure: Less of one, more the other? Am J Hypertens 1:335-347, 1988 16. Keller G, Zimmer G, Mall G, Ritz E, Amann K: Nephron number in patients with primary hypertension. N Engl J Med 348:101-108, 2003 17. Hughson MD, Douglass-Denton R, Bertram J, Hoy WE: Hypertension, glomerular number, and birth weight in African Americans and whites in the Southeastern United States. Kidney Int 69:671-678, 2006 18. Hughson M, Farris AB III, Douglass-Denton R, Hoy WE, Bertram JF: Glomerular number and size in autopsy

28 kidneys: The relationship to birth weight. Kidney Int 63:2113-2122, 2003 19. Barker DJP, Bagby SP, Hanson MA: Mechanisms of disease: In-utero programming in the pathogenesis of hypertension. Nat Clin Pract Nephrol 2:700-707, 2006 20. Barker DJP, Gluckman PD, Godfrey KM, Harding JE, Owens JA, Robinson JS: Fetal nutrition and cardiovascular disease in adult life. Lancet 341:938-941, 1993 21. Law CM, deSwiet M, Osmond C, et al: Initiation of hypertension in utero and its amplification throughout life. BMJ 306:24-27, 1993 22. Kiberd BA, Clase CM: Cummulative risk for developing end-stage renal disease in the US population. J Am Soc Nephrol 13:1635-1644, 2002 23. Marcantoni C, Ma LJ, Federspeil C, Fogo AB: Hypertensive nephrosclerosis in African Americans versus Caucasians. Kidney Int 62:172-180, 2002 24. Rostand SG, Cross SK, Kirk KA, Jee JY: Racial differences in renal arteriolar structure in children with minimal change nephropathy. Kidney Int 68:1154-1160, 2005 25. Kramer H, Luke A, Bidani A, Cao G, Cooper R, McGee D: Obesity and prevalent and incident CKD: The Hypertension and Detection Follow-up Program. Am J Kidney Dis 46:587-594, 2005 26. Gelber RP, Kurth T, Kausz AT, et al: Association between body mass index and CKD in apparently healthy men. Am J Kidney Dis 46:871-880, 2005 27. Ejerblad E, Fored CM, Linblad P, Fryzel J, McLauglin JK, Nyren O: Obesity and risk for chronic renal failure. J Am Soc Nephrol 17:1695-1706, 2006 28. Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS, Koplan KP: The continuing epidemic of obesity and diabetes in the United States. JAMA 286:1195-1200, 2001 29. Bertram JF: Analyzing renal glomeruli with the new stereology. Int Rev Cytol 161:111-172, 1995 30. Nyengaard JR: Stereologic methods and their application in kidney research. J Am Soc Nephrol 10:1100-1123, 1999 31. Johnson KJ, Wreford NG, Hoy WE, Bertram JF: Estimating total glomerular number in human kidneys with a physical disector/fractionator combination. Image Anal Stereol 19:105-108, 2000

Hughson et al 32. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P: The burden of adult hypertension in the United States. Hypertension 44:398-404, 2004 33. Rostand SG, Cliver SP, Goldenberg RL: Racial disparities in the association of foetal growth retardation to childhood blood pressure. Nephrol Dial Transplant 20:15921597, 2005 34. Hemachandra AH, Klebanoff MA, Furth SL: Racial disparities in the association between birth weight in the term infant and blood pressure at 7 years: Results from the Collaborative Perinatal Project. J Am Soc Nephrol 17:25762581, 2006 35. Ritz E: Is the renal risk of adults determined in-utero? Kidney Int 72:667-668, 2007 36. Iyasu S, Tomashek K, Barfield W: Infant mortality and low birth weight among black and white infants— United States, 1980-2000. MMWR Weekly. Available at: http://www.cdc.gov/mmwr/. Accessed March 1, 2008 37. Huxley R, Neil A, Collins R: Unravelling the fetal origins hypothesis: Is there really an inverse association between birthweight and subsequent blood pressure? Lancet 360:659-665, 2002 38. Falkner B, Hulman S, Kushner H: Effect of birth weight on blood pressure and body size in early adolescence. Hypertension 43:203-207, 2004 39. Trust for American Health Reports: Mississippi Adults 1st most obese in country; youth 8th overweight. Available at: http://healthyamericans.org/reports/obesity2007. Accessed March 1, 2008 40. Hall JE: The kidney, hypertension, and obesity. Hypertension 41:625-633, 2003 41. Hoy WE: Lessons in ethnonephrology. Kidney Int 70:251-257, 2005 42. Shoham DA, Vupputuri S, Kshirsagar AV: Chronic kidney disease and life course socioeconomic status: A review. Adv Chronic Kidney Dis 12:56-63, 2005 43. Hoy WE, Mathews JD, McCredie DA, et al: The multidimensional nature of renal disease: Rates and associations of albuminuria in an Australian Aboriginal community. Kidney Int 54:1296-1304, 1998 44. Taguchi A, Sanada M, Suei Y, et al: Tooth loss is associated with an increased risk of hypertension in postmenopausal women. Hypertension 43:1297-1300, 2004