AMERICAN JOURNAL OF EPIDEMIOLOGY
Vol. 104, No. 2
Copyright © 1976 by The Johns Hopkins University School of Hygiene and Public Health
Printed in U.S.A.
Original Contributions FAMILIAL AGGREGATION OF URINARY KALLIKREIN CONCENTRATION IN CHILDHOOD: RELATION TO BLOOD PRESSURE, RACE AND URINARY ELECTROLYTES 1 STEPHEN H. ZINNER,2 HARRY S. MARGOLIUS,3 BERNARD ROSNER, HARRY R. REISER AND EDWARD H. KASS Zinner, S. H. (Charming Laboratory, Boston City Hospital, Boston, MA 02118), H. S. Margolius, B. Rosner, H. R. Keiser and E. H. Kass. Familial aggregation of urinary kallikrein concentration in childhood: Relation to blood pressure, race and urinary electrolytes. Am J Epidemiol 104:124-132, 1976. Biochemical alterations that have been correlated with elevated blood pressure have not received extensive epidemiologic study because of the technical difficulties involved. Because the excretion of urinary kallikrein is reduced significantly in adult hypertensives, we have studied urinary kallikrein in a cohort of children in whom familial aggregation of blood pressure has been demonstrated. Casual specimens of urine were obtained in household surveys, and urinary concentration of kallikrein was determined in 601 children aged 5-18 years. The children were from 163 families, whose members also had their blood pressures measured. Familial aggregation of urinary kallikrein concentration was found by analysis of variance (F = 3.45, p < 0.001) even in these casual specimens and was demonstrable for black and white children analyzed separately. Urinary kallikrein concentration was significantly lower in black children than in white children (p < 0.001) and was positively correlated with urinary creatinine and urinary potassium and inversely related to urinary sodium concentrations. Urinary kallikrein concentration also was being altered by - season (being lowest in the summer) and by time of day (being highest in the morning). Families with the lowest mean kallikrein concentrations tended to have higher blood pressures than did families with the highest mean kallikrein concentrations, although the effect is small and subject to many variables. hypertension; kallikrein; urine
Urinary kallikrein excretion has been studied extensively in relation to blood pressure in normal and hypertensive adults and in laboratory animals (1-4). Urinary
kallikrein is a renal enzyme which attacks an alpha globulin substrate, kininogen, to produce kallidin, a potent vasodilator kinin (5). Margolius et al. (1, 4) have
Received for publication October 20, 1975, and in final form January 9, 1976. Abbreviations: EU, esterase units; SDU, standard deviation units. 'From the Channing Laboratory, Department of Medicine, Harvard Medical School, and the Departments of Medicine and Medical Microbiology, Boston City Hospital, Boston, MA; the Section on Experimental Therapeutics, National Heart and Lung Institute, National Institutes of Health, Bethesda, MD; and the Department of Medicine, Roger Williams
General Hospital, and the Division of Biological and Medical Sciences, Brown University, Providence, RI. 2 Reprint requests to Dr. Zinner, Channing Laboratory, Boston City Hospital, Boston, MA 02118. 3 Present address: Departments of Pharmacology and Medicine, Medical University of South Carolina, Charleston, SC. This work was supported by a Grant-in-Aid from the American Heart Association and with funds contributed in part by the Greater Boston Heart Association, and by Research Grant #HD03693 from
124
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
125
URINARY KALLIKREIN IN CHILDHOOD
demonstrated that urinary kallikrein excretion is decreased or absent in untreated adults with essential hypertension. The present study afforded the opportunity to investigate the distribution of the concentration of this enzyme in the urines of children whose blood pressures were being taken in a standardized manner in their homes. Extensive epidemiologic studies of biochemical factors that have been associated with blood pressure have not been undertaken in the past, perhaps due to the many technical problems involved. Urinary kallikrein determinations therefore seemed suitable for epidemiologic study. Study of urinary kallikrein in children is also important because familial aggregation of blood pressure is repeatedly demonstrable in childhood (6, 7), and there is a significant, positive relationship between initial blood pressure in children and a measurement made four years later (7). These data suggest that stratification of blood pressures within peer groups might begin in childhood. They also raise the possibility that essential hypertension may have its origins in childhood and be potentially detectable in childhood. This hypothesis, based entirely on epidemiologic analyses, could be further strengthened if biochemical findings related to hypertension could be demonstrated in childhood. Urinary kallikrein studies were a first step in this direction.
study of women who were or were not bacteriuric in pregnancy in 1955-1959 (8). All families were visited in their homes and blood pressures were measured with the Kass-Mollo-Christensen portable automated blood pressure recorder, which gives a direct print-out of the blood pressure reading, minimizing observer error (9). Blood pressures were taken three times in succession after the child had been resting quietly in a chair for at least five minutes. To adjust for age and sex, all blood pressures were expressed in standard deviation units (SDU), which are defined as the observed blood pressure reading minus the mean pressure for the two-year age and sex group divided by the standard deviation of the pressures in the group. A casual, clean voided urine specimen was obtained from each child and from each mother. The urine was refrigerated immediately and stored at 4 C under toluene until analyzed for kallikrein, sodium, potassium and creatinine concentrations. Urinary kallikrein was measured by the radiochemical esterolytic assay of Beaven et al. (10) as modified by Margolius et al. (3), and values are expressed as esterase units (EU) per ml. This measurement is associated with an interassay variation of ±10 per cent (3). Sodium and potassium concentrations were measured by flame photometry, and creatinine concentration was measured with a Technicon autoanalyzer.
MATERIALS AND METHODS
A total of 601 children aged 5-18 years in 163 families were observed. These families were studied four years earlier (6) and initially were part of a long-term follow-up the National Institute of Child Health and Human Development. Presented in part at the annual meeting of the American Society for Clinical Investigation, Atlantic City, NJ, May, 1974, and at the Second International Symposium on the Epidemiology of Hypertension, Chicago, IL, Sept., 1974. The authors are indebted to Mrs. Olga Ulchak, R.N., and to Louis Martin and Frank Sacks for their assistance.
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
RESULTS
The distribution of urinary kallikrein in EU/ml is presented in figure 1. In the esterolytic method, 0.001 EU/ml is the lowest urinary kallikrein concentration detectable. As seen in the figure, 17 children (2.8 per cent) had values below this level. Because of the extreme skewness of the distribution shown here, a logarithmic transformation was prepared. A Chi square goodness of fit test (xio2 = 16.09, NS) indicated that the transformed distribu-
126
STEPHEN H. ZINNER ET AL.
200
180-
20 •
0.0010.0049
0.0050.0099
0.0100.0149
0.0150.0190
0.0200.0249
0.0250.0290
0.0300.0349
0.0350.0390
0.0400.0449
0.0450.0490
0.0500.0549
Urinary Kallikrein Concentration (E.U./ml)
FIGURE 1. Distribution of urinary kallikrein concentration in a-N-p-tosyl-L arginine-['H]-methyl esterase units (EU/ml) in 601 children.
tion (shown in figure 2) is well approximated by a normal curve. The logarithmic transformation for kallikrein concentration was employed for the remainder of the analyses. The mean urinary kallikrein ±1 standard deviation was 4.236 ± 0.94. As seen in table 1, analysis of variance for urinary kallikrein concentration revealed that the variance within families is significantly less than that among all children studied (p < 0.001). Similar results were seen when the data for black and white children were analyzed separately. These data indicate that there is significant familial clustering of urinary kallikrein concentration in children of both races. Familial aggregation also was seen when urinary kallikrein concentration was adjusted for urinary creatinine concentration in the multiple regression analysis (F
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
= 4.86, p < 0.001). Data demonstrating familial aggregation of blood pressure in these children have been published elsewhere (6, 7). With single regression analysis, several variables were found to affect urinary kallikrein concentration in the children (table 2). Urinary kallikrein concentration increased with increasing urinary potassium and creatinine concentrations and was related positively to the mother's urinary kallikrein concentration. Moreover, black children had significantly lower mean urinary kallikrein concentrations than did white children, and this was found at all levels of urinary potassium concentration (table 3). Urinary kallikrein concentrations were slightly higher in female than male children, higher in urine obtained in the morning as opposed to other times in the
URINARY KALLIKREIN IN CHILDHOOD
day and lower in the summer than in other seasons. Also, as seen in table 4, the racial difference in urinary kallikrein concentration persisted at every level of urinary creatinine concentration. Urinary kallikrein concentrations were lower in black than in white children whether children were studied in the morning or afternoon, summer or winter. In addition, the urinary kallikrein concentrations of the children were inversely related to the diastolic blood pressures of mothers and children (table 2). The variables affecting urinary kallikrein concentration were studied using multiple regression analysis, a procedure which considers the effect of several independent variables simultaneously (table 5). In order of decreasing importance, urinary potassium and urinary creatinine concentrations, race, weight, season, sex, time of day and urinary sodium concentrations
127
were significantly related to urinary kallikrein concentration. These eight variables accounted for 43 per cent of the variability of urinary kallikrein. The contribution of each of the individual variables to the overall R2 is also seen in table 5. The direction of the correlations is the same as for the single regression analyses (table 2). Although there was no relation of urinary kallikrein to urinary sodium concentrations by single regression analysis, there was an inverse relation of these variables after the effects of other variables had been removed in the multiple regression analysis (b = -0.148, p < 0.001). Urinary kallikrein concentration was inversely related to the ratio of urinary sodium to urinary potassium concentrations (b = —0.66, p < 0.001). To further examine the relation of urinary kallikrein to blood pressure in child-
130120IIOIOO90-
0>
80 •
70 "o |
60-
z 50 40 30 20 10 1.0-1.49 150-1.99 2.0-2.49 2.5-2.99 3.0-3.49 3.5-399 4.0-4.49 4.5-499 5.0-549 55-599 60-649 65-6.99
Log Urinary Kallikrein Concentration (EU/ml) FIGURE 2. Distribution of log urinary kallikrein concentration in 601 children.
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
70-749
128
STEPHEN H. ZINNER ET AL.
hood, those families with the highest 10 per cent and lowest 10 per cent of mean log urinary concentrations were studied. Mean log urinary concentrations for siblings in the family were selected and mean family blood pressures compared. The data for mothers were excluded, and mean family blood pressure in SDU for children was multiplied by x/rcT where n equals the number of children studied in each family. Mean blood pressure scores for the "low kallikrein" families were significantly higher than for the "high kallikrein" families (for systolic pressure: +0.21 SDU ± 1.4 versus -0.78 SDU ± 0.8, p < 0.05; for diastolic pressure: +0.54 SDU ± 1.5 ver-
sus -0.62 SDU ± 0.8, p < 0.02). In addition, 15 of the 16 "low kallikrein" families were black compared to only 5 of 16 "high kallikrein" families (x 2 = 10.8, p < 0.005). The inverse relation of urinary kallikrein concentration and blood pressure was more striking in the 364 children studied in the summer and autumn (regression coefficient = -0.12, p < 0.01) than in the 234 children studied in the winter and spring (regression coefficient = -0.06, NS). Similarly, when selected by blood pressures, the families (12 black, 4 white) with the highest 10 per cent of diastolic pressures had significantly lower mean kallikrein concentration than the families (8
TABLE 1
Analysis of variance log urinary kallikrein concentration in 601 children from 163 families Mean square (variance)
Degrees of freedom, D.F.
Ratio of mean squares, F
All families Among families Within families
1.851 0.536
162 438
3.451
< 0.001
White families • Among families Within families
0.953 0.462
73 172
2.061
< 0.001
Black families Among families Within families
1.688 0.591
92 259
2.859
Vol. 104, No. 2
Copyright © 1976 by The Johns Hopkins University School of Hygiene and Public Health
Printed in U.S.A.
Original Contributions FAMILIAL AGGREGATION OF URINARY KALLIKREIN CONCENTRATION IN CHILDHOOD: RELATION TO BLOOD PRESSURE, RACE AND URINARY ELECTROLYTES 1 STEPHEN H. ZINNER,2 HARRY S. MARGOLIUS,3 BERNARD ROSNER, HARRY R. REISER AND EDWARD H. KASS Zinner, S. H. (Charming Laboratory, Boston City Hospital, Boston, MA 02118), H. S. Margolius, B. Rosner, H. R. Keiser and E. H. Kass. Familial aggregation of urinary kallikrein concentration in childhood: Relation to blood pressure, race and urinary electrolytes. Am J Epidemiol 104:124-132, 1976. Biochemical alterations that have been correlated with elevated blood pressure have not received extensive epidemiologic study because of the technical difficulties involved. Because the excretion of urinary kallikrein is reduced significantly in adult hypertensives, we have studied urinary kallikrein in a cohort of children in whom familial aggregation of blood pressure has been demonstrated. Casual specimens of urine were obtained in household surveys, and urinary concentration of kallikrein was determined in 601 children aged 5-18 years. The children were from 163 families, whose members also had their blood pressures measured. Familial aggregation of urinary kallikrein concentration was found by analysis of variance (F = 3.45, p < 0.001) even in these casual specimens and was demonstrable for black and white children analyzed separately. Urinary kallikrein concentration was significantly lower in black children than in white children (p < 0.001) and was positively correlated with urinary creatinine and urinary potassium and inversely related to urinary sodium concentrations. Urinary kallikrein concentration also was being altered by - season (being lowest in the summer) and by time of day (being highest in the morning). Families with the lowest mean kallikrein concentrations tended to have higher blood pressures than did families with the highest mean kallikrein concentrations, although the effect is small and subject to many variables. hypertension; kallikrein; urine
Urinary kallikrein excretion has been studied extensively in relation to blood pressure in normal and hypertensive adults and in laboratory animals (1-4). Urinary
kallikrein is a renal enzyme which attacks an alpha globulin substrate, kininogen, to produce kallidin, a potent vasodilator kinin (5). Margolius et al. (1, 4) have
Received for publication October 20, 1975, and in final form January 9, 1976. Abbreviations: EU, esterase units; SDU, standard deviation units. 'From the Channing Laboratory, Department of Medicine, Harvard Medical School, and the Departments of Medicine and Medical Microbiology, Boston City Hospital, Boston, MA; the Section on Experimental Therapeutics, National Heart and Lung Institute, National Institutes of Health, Bethesda, MD; and the Department of Medicine, Roger Williams
General Hospital, and the Division of Biological and Medical Sciences, Brown University, Providence, RI. 2 Reprint requests to Dr. Zinner, Channing Laboratory, Boston City Hospital, Boston, MA 02118. 3 Present address: Departments of Pharmacology and Medicine, Medical University of South Carolina, Charleston, SC. This work was supported by a Grant-in-Aid from the American Heart Association and with funds contributed in part by the Greater Boston Heart Association, and by Research Grant #HD03693 from
124
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
125
URINARY KALLIKREIN IN CHILDHOOD
demonstrated that urinary kallikrein excretion is decreased or absent in untreated adults with essential hypertension. The present study afforded the opportunity to investigate the distribution of the concentration of this enzyme in the urines of children whose blood pressures were being taken in a standardized manner in their homes. Extensive epidemiologic studies of biochemical factors that have been associated with blood pressure have not been undertaken in the past, perhaps due to the many technical problems involved. Urinary kallikrein determinations therefore seemed suitable for epidemiologic study. Study of urinary kallikrein in children is also important because familial aggregation of blood pressure is repeatedly demonstrable in childhood (6, 7), and there is a significant, positive relationship between initial blood pressure in children and a measurement made four years later (7). These data suggest that stratification of blood pressures within peer groups might begin in childhood. They also raise the possibility that essential hypertension may have its origins in childhood and be potentially detectable in childhood. This hypothesis, based entirely on epidemiologic analyses, could be further strengthened if biochemical findings related to hypertension could be demonstrated in childhood. Urinary kallikrein studies were a first step in this direction.
study of women who were or were not bacteriuric in pregnancy in 1955-1959 (8). All families were visited in their homes and blood pressures were measured with the Kass-Mollo-Christensen portable automated blood pressure recorder, which gives a direct print-out of the blood pressure reading, minimizing observer error (9). Blood pressures were taken three times in succession after the child had been resting quietly in a chair for at least five minutes. To adjust for age and sex, all blood pressures were expressed in standard deviation units (SDU), which are defined as the observed blood pressure reading minus the mean pressure for the two-year age and sex group divided by the standard deviation of the pressures in the group. A casual, clean voided urine specimen was obtained from each child and from each mother. The urine was refrigerated immediately and stored at 4 C under toluene until analyzed for kallikrein, sodium, potassium and creatinine concentrations. Urinary kallikrein was measured by the radiochemical esterolytic assay of Beaven et al. (10) as modified by Margolius et al. (3), and values are expressed as esterase units (EU) per ml. This measurement is associated with an interassay variation of ±10 per cent (3). Sodium and potassium concentrations were measured by flame photometry, and creatinine concentration was measured with a Technicon autoanalyzer.
MATERIALS AND METHODS
A total of 601 children aged 5-18 years in 163 families were observed. These families were studied four years earlier (6) and initially were part of a long-term follow-up the National Institute of Child Health and Human Development. Presented in part at the annual meeting of the American Society for Clinical Investigation, Atlantic City, NJ, May, 1974, and at the Second International Symposium on the Epidemiology of Hypertension, Chicago, IL, Sept., 1974. The authors are indebted to Mrs. Olga Ulchak, R.N., and to Louis Martin and Frank Sacks for their assistance.
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
RESULTS
The distribution of urinary kallikrein in EU/ml is presented in figure 1. In the esterolytic method, 0.001 EU/ml is the lowest urinary kallikrein concentration detectable. As seen in the figure, 17 children (2.8 per cent) had values below this level. Because of the extreme skewness of the distribution shown here, a logarithmic transformation was prepared. A Chi square goodness of fit test (xio2 = 16.09, NS) indicated that the transformed distribu-
126
STEPHEN H. ZINNER ET AL.
200
180-
20 •
0.0010.0049
0.0050.0099
0.0100.0149
0.0150.0190
0.0200.0249
0.0250.0290
0.0300.0349
0.0350.0390
0.0400.0449
0.0450.0490
0.0500.0549
Urinary Kallikrein Concentration (E.U./ml)
FIGURE 1. Distribution of urinary kallikrein concentration in a-N-p-tosyl-L arginine-['H]-methyl esterase units (EU/ml) in 601 children.
tion (shown in figure 2) is well approximated by a normal curve. The logarithmic transformation for kallikrein concentration was employed for the remainder of the analyses. The mean urinary kallikrein ±1 standard deviation was 4.236 ± 0.94. As seen in table 1, analysis of variance for urinary kallikrein concentration revealed that the variance within families is significantly less than that among all children studied (p < 0.001). Similar results were seen when the data for black and white children were analyzed separately. These data indicate that there is significant familial clustering of urinary kallikrein concentration in children of both races. Familial aggregation also was seen when urinary kallikrein concentration was adjusted for urinary creatinine concentration in the multiple regression analysis (F
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
= 4.86, p < 0.001). Data demonstrating familial aggregation of blood pressure in these children have been published elsewhere (6, 7). With single regression analysis, several variables were found to affect urinary kallikrein concentration in the children (table 2). Urinary kallikrein concentration increased with increasing urinary potassium and creatinine concentrations and was related positively to the mother's urinary kallikrein concentration. Moreover, black children had significantly lower mean urinary kallikrein concentrations than did white children, and this was found at all levels of urinary potassium concentration (table 3). Urinary kallikrein concentrations were slightly higher in female than male children, higher in urine obtained in the morning as opposed to other times in the
URINARY KALLIKREIN IN CHILDHOOD
day and lower in the summer than in other seasons. Also, as seen in table 4, the racial difference in urinary kallikrein concentration persisted at every level of urinary creatinine concentration. Urinary kallikrein concentrations were lower in black than in white children whether children were studied in the morning or afternoon, summer or winter. In addition, the urinary kallikrein concentrations of the children were inversely related to the diastolic blood pressures of mothers and children (table 2). The variables affecting urinary kallikrein concentration were studied using multiple regression analysis, a procedure which considers the effect of several independent variables simultaneously (table 5). In order of decreasing importance, urinary potassium and urinary creatinine concentrations, race, weight, season, sex, time of day and urinary sodium concentrations
127
were significantly related to urinary kallikrein concentration. These eight variables accounted for 43 per cent of the variability of urinary kallikrein. The contribution of each of the individual variables to the overall R2 is also seen in table 5. The direction of the correlations is the same as for the single regression analyses (table 2). Although there was no relation of urinary kallikrein to urinary sodium concentrations by single regression analysis, there was an inverse relation of these variables after the effects of other variables had been removed in the multiple regression analysis (b = -0.148, p < 0.001). Urinary kallikrein concentration was inversely related to the ratio of urinary sodium to urinary potassium concentrations (b = —0.66, p < 0.001). To further examine the relation of urinary kallikrein to blood pressure in child-
130120IIOIOO90-
0>
80 •
70 "o |
60-
z 50 40 30 20 10 1.0-1.49 150-1.99 2.0-2.49 2.5-2.99 3.0-3.49 3.5-399 4.0-4.49 4.5-499 5.0-549 55-599 60-649 65-6.99
Log Urinary Kallikrein Concentration (EU/ml) FIGURE 2. Distribution of log urinary kallikrein concentration in 601 children.
Downloaded from https://academic.oup.com/aje/article-abstract/104/2/124/101654 by INSEAD user on 25 July 2018
70-749
128
STEPHEN H. ZINNER ET AL.
hood, those families with the highest 10 per cent and lowest 10 per cent of mean log urinary concentrations were studied. Mean log urinary concentrations for siblings in the family were selected and mean family blood pressures compared. The data for mothers were excluded, and mean family blood pressure in SDU for children was multiplied by x/rcT where n equals the number of children studied in each family. Mean blood pressure scores for the "low kallikrein" families were significantly higher than for the "high kallikrein" families (for systolic pressure: +0.21 SDU ± 1.4 versus -0.78 SDU ± 0.8, p < 0.05; for diastolic pressure: +0.54 SDU ± 1.5 ver-
sus -0.62 SDU ± 0.8, p < 0.02). In addition, 15 of the 16 "low kallikrein" families were black compared to only 5 of 16 "high kallikrein" families (x 2 = 10.8, p < 0.005). The inverse relation of urinary kallikrein concentration and blood pressure was more striking in the 364 children studied in the summer and autumn (regression coefficient = -0.12, p < 0.01) than in the 234 children studied in the winter and spring (regression coefficient = -0.06, NS). Similarly, when selected by blood pressures, the families (12 black, 4 white) with the highest 10 per cent of diastolic pressures had significantly lower mean kallikrein concentration than the families (8
TABLE 1
Analysis of variance log urinary kallikrein concentration in 601 children from 163 families Mean square (variance)
Degrees of freedom, D.F.
Ratio of mean squares, F
All families Among families Within families
1.851 0.536
162 438
3.451
< 0.001
White families • Among families Within families
0.953 0.462
73 172
2.061
< 0.001
Black families Among families Within families
1.688 0.591
92 259
2.859
