Genetic Factors that Predispose the Child to Develop Hypertension

CHILDHOOD HYPERTENSION

0031-3955/93 $0.00 + .20

GENETIC FACTORS THAT PREDISPOSE THE CHILD TO DEVELOP HYPERTENSION Richard M. Schieken, MD

GENETICS OF HYPERTENSION

Two prevalent theories exist for the genetic relationship to hypertension. The first is that hypertension is a specific disease entity caused by a single gene. This results in a distribution of blood pressure in the adult population that is skewed to the right. The population contains two distributions, the normal distribution and the hypertensive population, which has only high blood pressures. The second explanation is that the disease hypertension results from one or more abnormalities within a complex array of systems such as electrolyte transport systems and sympathetic control mechanisms, each of which may have genetic abnormalities. When the critical genes or combinations of genes occur, the polygenic expression of the disease hypertension occurs. This article summarizes the data suggesting that in children, high blood pressure or a prehypertensive state is controlled either by a major gene effect or by the polygenic expression of genes. Many genes acting independently on other biologic systems may indirectly control the level of blood pressure by influencing systems such as ion transport or catecholamine responses to stress. Additionally, information is provided to explain the global genetic and environmental forces that influence the level of blood pressure in healthy young adolescent children. From the Division of Pediatric Cardiology, Department of Pediatrics, The Children's Medical Center; Medical College of Virginia, Richmond, Virginia

PEDIATRIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • FEBRUARY 1993

1

2

SCHIEKEN

TRANSPORT ABNORMALITIES

For many years, a positive relationship between salt intake and the prevalence of hypertension in countries throughout the world has been cited as evidence for the etiologic role of salt in the development of human hypertension. 10 The Intersalt Study failed to demonstrate that moderate differences in sodium consumption contribute to the variation in blood pressure levels among populations. 9 They did observe, however, a possible relationship between blood pressure and urinary sodium. Because of the hotly debated relationship between salt ingestion and blood pressure, investigators turned to electrolyte transport to search for abnormalities that might link salt and hypertension. In children, lithium countertransport (LCT) has been found to relate to blood pressure levels. 24 LCT is measured as the maximal rate (Vmax) of lithium efflux by sodium-stimulated LCT in lithium-loaded cells. 2 In an epidemiologic study of transport mechanisms of children and adults, LCT was positively associated with systolic blood pressure in black children but not white children. 12 Girls of both races had lower values than boys. In both races, LCT was positively associated with body mass index. A family history of hypertension did not correlate with higher levels of LCT. In contrast to the children (mean age, 12.7 years), adolescents' (mean age, 16 years) LCT was correlated to diastolic blood pressure in those with one hypertensive parent. Hunt et alB investigated the relationships of three cation transport systems with blood pressure and plasma lipids in children and adults. In hypertensive adults the levels of the cation transport measures were significantly higher than in normotensive subjects. In children, however, only lithium efflux correlated with mean blood pressure. This appears to be another piece of evidence that the genes that regulate blood pressure in adults and children are different.

SODIUM SENSITIVITY

Falkner et al 4 studied the interactive effects of sodium sensitivity and the family history of hypertension in young black adults. Young adults who had been followed in the Philadelphia Childhood Blood Pressure Study during adolescence (11-18 years) participated in a follow-up study to investigate the interactive effects of family history of hypertension, sodium sensitivity, and cardiovascular reactivity. At follow-up the subjects' ages ranged from 18 to 23 years. The sodiumsensitive response was defined as an increase in mean arterial pressure of equal to or more than 5 mm Hg after a daily dietary supplement of 10 g NaCl. A greater incidence of sodium sensitivity (53% vs 47%) was found for these black subjects as compared with a comparable white sample. No differences in blood pressure were found either by race or sodium sensitivity status before beginning the sodium supplementation. The family history of hypertension was greater in the blacks (67%

GENETIC FACTORS THAT PREDISPOSE THE CHILD TO DEVELOP HYPERTENSION

3

vs 38%) compared with the white subjects. In these black subjects, no differences in blood pressure were found in response to mental stress in the form of mental arithmetic. When grouped according to family history, however, the subjects with a positive family history had a greater response to mental stress after sodium loading. The investigators concluded that in the sodium loaded condition, sodium sensitivity interacted with a positive family history of hypertension. In young black subjects this sensitivity resulted in a significantly greater resting blood pressure and blood pressure under the stimulus of mental stress. One possible mechanism underlying an excessive reactivity after sodium loading could be abnormal regulation of plasma renin activity. An epidemiologic study of the relationship between plasma renin activity and a positive family history of hypertension was conducted in Japan. 21 More than 600 children, aged 10 to 14 years, participated. A positive family history was defined as one or more parents or grandparents with blood pressures in excess of 160/95 mm Hg. In both boys and girls, children from hypertensive families had higher systolic blood pressures. The levels of plasma renin activity were inversely related to systolic blood pressure in the family history negative group but did not relate to the blood pressure of the positive family history group. The investigators concluded that the feedback mechanism of plasma renin activity was defective in the positive family history group.

FAMILIAL AGGREGATION

Familial aggregation of blood pressure was studied in a population study in Belgium. lO Both systolic and diastolic blood pressure were significantly correlated in siblings. The largest parent-offspring correlation was between father and son. Of importance, no blood pressure concordance was found for fathers and their daughters. A longitudinal cohort study of children's blood pressure in Minneapolis divided the population sample into groups by family history.14 Blood pressures in the children with a family history of hypertension had higher systolic blood pressures both at the initial screening and at nine subsequent measurements. The parents in the positive family history group had lower income, greater body weight, and were less well educated. The correlations of mothers and their children were higher than the correlations of fathers and their children. In a study of children living in London, Holland and Beresford7 found no difference in the correlation coefficients of systolic blood pressure between either same sex or opposite sex siblings. To our knowledge, as yet, no convincing evidence for sex differences in the genetic regulation of blood pressure has been described in either adults or children. Patterson et aP6 compared the familial aggregation found in white American families with that of Mexican-American families. Significant spousal correlations were observed for blood pressures in the Mexican-

4

SCHIEKEN

American families. Within these families, the fathers' blood pressure correlated with children of both sexes. For mothers, weak correlations were found with all family members. These relationships persisted after adjustment for environmental confounders such as diet, activity level, and body mass index. Brandao et aP investigated the familial relationship between weight and blood pressure. In families with a parent whose systolic or diastolic blood pressure exceeded the 90th percentile, the children's systolic blood pressure was higher than a group whose parents' blood pressures were below the 50th percentile. Diastolic blood pressure did not differ between the groups. In general, weight related to an increase in systolic but not diastolic blood pressure. Palti et aP5 studied the aggregation of blood pressure among sibling pairs of the same sex. The height, weight, and blood pressure measurements were performed as each child entered the first grade. Siblingsibling correlations were significant for systolic blood pressure (r = 0.15). Diastolic blood pressure correlations were not significant. The investigators were aware that their correlations were lower than other reported values and suggested that the use of casual blood pressure measurement had introduced a large amount of variability. A sample of children from the Muscatine Study were divided according to their systolic blood pressure into three groups to investigate familial aggregation. 3 The blood pressure groups were the following: low « fifth percentile); middle (> fifth < ninety-fifth percentile); and high labile (> 95th when first sampled but < 95th percentile when resampled 4 to 6 months later). Both the systolic and diastolic blood pressures aggregated more strongly in the families of children with labile high blood pressure than in the families with low or middle blood pressure. Correlations were tabulated for the paired parent and sibling relationships both after adjustment for age and further adjustment for age and body size. Little change occurred in the correlations of diastolic blood pressure for the sibling-sibling pairs after body size was added to age. Investigators from the Framingham Heart Study studied the familial aggregation of blood pressure among first-degree relatives. 6 Familial blood pressure associations were evaluated using blood pressure measurements obtained from adult offspring and these were compared with those measurements taken on their parents 25 years ago. At that time, the parents were 5 years older, on average, than their children's current age. Both paternal and maternal blood pressure correlated significantly with that of the offspring. These relationships persisted after adjustment for correlates such as weight, alcohol consumption, oral contraceptive use, and other covariates known to influence blood pressure. GENETIC EPIDEMIOLOGY

The influence of genes and household environment was studied in children enrolled in the Montreal Adoption Study.13 The population

GENETIC FACTORS THAT PREDISPOSE THE CHILD TO DEVELOP HYPERTENSION

5

sample included 1176 parents, 756 adopted children, and 445 children with their natural parents. Based on maximum likelihood estimates presented in models for parents and offspring, shared genes accounted for 61 % of the variance, whereas environment shared by parents and children accounted for the remaining 38%. Shared genes accounted for 58% of the variance of diastolic blood pressure, with 42% caused by environment shared across the generations. The distributions of systolic, diastolic, and mean blood pressure were analyzed for evidence of commingling and segregation in a large French-Canadian study.I8 This study included parents and their singleton children and twin and adopted offspring. There was no evidence to suggest commingling of distributions for diastolic or mean blood pressure. For systolic blood pressure, evidence for commingling, in the form of two skewed distributions, was found both in the parent and the parent-offspring data sets. Statistical evidence supported the presence of a major effect for systolic blood pressure. Genetic analysis of systolic blood pressure did not support a major gene effect in both generations, however. As an alternative explanation for these confusing data, the investigators suggested that a true major gene effect was expressed in the parents but not yet expressed in the offspring. Perrusse et aP7 used complex segregation analysis to determine the best genetic models explaining the variance of systolic blood pressure in a population based sample. They studied more than 1200 school aged children who represented the probands of 278 pedigrees enrolled in schools in Rochester, MN. They suggested that the variability in systolic blood pressure may be influenced by major effects of allelic variation at a single gene, which are sex and age dependent. There may be particular genotypes determined by a single gene that are associated with a steeper increase of systolic blood pressure with age. The impact that the single gene had on the variance of blood pressure increased with age. The major effects of this single gene accounted for about 5% of the variability at age 5 years and 60% of the variability at age older than 50 years. These rates of increase differed in males and females.

TWIN STUDIES

Twin studies offer a unique opportunity to study individuals who have all of their genes in common. These monozygotic twins can be contrasted with dizygotic twins, who on average share one half of their genes. In addition, not only can male-male and female-female pairs of monozygotic and dizygotic twins be compared, but unlike-sex dizygotic twins provide insight into differential effects in males and females. Although twin studies often report higher estimates of heritability than either population or large family studies, the identical age of the siblings in twins reduces confounding temporal variables. Beginning in 1969, the National Heart and Lung Institute of the

6

SCHIEKEN

National Institutes of Health sponsored the physical examination of 250 monozygotic and 264 dizygotic twin pairs of adult white males. 5 The subjects were aged 42 to 55 years when enrolled in the study. Both systolic and diastolic blood pressures were measured. The distributions of both the systolic and diastolic blood pressures, which included both members of the pairs of twins, were similar to those distributions that were made up of individuals who were not twins. For both systolic and diastolic blood pressure there was evidence for a substantial genetic influence on the population variability. The genetic influence on the variability of systolic blood pressure was estimated to be as much as 82% of the total, whereas for diastolic blood pressure the estimate was 64% genetic. There was no evidence for different blood pressure levels between monozygotic and dizygotic twins. The risk for the brother of a hypertensive monozygotic twin developing hypertension himself was significantly higher than the risk for the brother of a dizygotic hypertensive twin. The univariate genetic analysis of blood pressure was studied in ll-year-old boys and girls who participated in The Medical College of Virginia Twin Study.19 The study population included 251 white twin pairs. For analysis the twins were divided into five groups according to both sex and zygosity. Zygosity was determined by blood grouping, enzyme tests, and DNA analysis. Sexual staging was performed based on the classification of Tanner. Model fitting methods were used for the genetic analysis of the data. Path models were specified and fitted using the LISREL VI program. Figure 1 shows the form of the model for both like-sex and unlike-sex twins. The latent (unobserved and causal) variables are the (additive) genetic effects on each twin (G1 and G2); the environmental effects shared by members of twin pairs (C); and the environmental effects unique to each individual twin within a pair (El and E2). These latent variables are assumed to have unit standard deviation. The correlation between the genetic effects ('Y) is unity for monozygotic twins. For dizygotic twins, the correlation is one half when mating is random. The residual environmental effects, El and E2, are uncorrelated. The regression of the phenotypes of the twins, Tl and T2, on the latent variables is specified by the path coefficients h (path from genotype to phenotype); c (from common environment to phenotype); and e (from unique environment to phenotype). Figure IB shows the path model for unlike-sex twin pairs which are, de facto, dizygotic. The model allows for two kinds of sex differences in genetic and environmental effects. The relative magnitudes of the effects may differ between sexes, represented in the model by using h, c, and e for path coefficients in males and h', c', and e' for the corresponding values in females. In addition, we allow for the possibility that different genes and qualitatively different shared environments may contribute to each sex by allowing the correlation between the genetic influences on males and on females (rg) to be less than unity and hence the overall genetic correlation for unlike-sex dizygotic twins (rg) to be less than 0.5. Similarly, the correlation between the

GENETIC FACTORS THAT PREDISPOSE THE CHILD TO DEVELOP HYPERTENSION

1

'(

"2 r h

~

G1

~

G2

1"

Gm

"1

Gf

1"

"i

I i, ,i\ Ii,

E1

T1

T2

C1

C2

~ A

1.0

7

Tf

Tm

E2

'l\

Em

Cm

B

~ r

Cf

Ef

c

Figure 1. A, Path diagram for like-sex twin pairs. The phenotypes of the twins (T" T2) are modeled as being determined by additive genetic effects (G" G2), environmental effects common to both twins (C), and environmental effects specific to each twin (E" E2). 'I is the correlation of genetic effects on twin 1 with those on twin 2. For monozygotic twins, 'I = 1. For dizygotic twins, 'I = 0.5. B, Path diagram for unlike-sex twin pairs. Additive genetic, common environmental, and specific environmental effects are allowed to exert different effects on the male twin and female twin. The parameter rh represents the correlation between the genetic effects in the male and female twins. The parameter rc represents the correlation of the effects of the common (shared) environment on the male and female twins. (From Schieken RM, Eaves LJ, Hewitt JK, et al: The univariate genetic analysis of blood pressure in children (The Medical College of Virginia Twin Study). Am J Cardiol 64:1333-1337, 1989; with permission.)

shared environmental influences on males and on females (rJ is allowed to be less than unity.16 The model-fitting results are summarized in Table 1. Maximum likelihood parameter estimates are presented for the most parsimonious model that fits the data. The simple univariate model that allowed for the additive effects of genes (h) and the environmental effects unique to the individual (e) fit for the variable heart rate. For none of the variables did the model detect shared environmental effects. For all of the variables the additive effects of genes accounted for about half or more of the variance. The genes controlling systolic blood pressure were different in boys and girls but accounted for a similar portion of the variance. For diastolic blood pressure, the genes accounted for differing portions of the variance in boys and girls. In the Medical College of Virginia Twin Study,19 striking evidence for sex differences in the regulation of blood pressure was first detected by observing that the Pearson correlations of male-female twin pairs

8

SCHIEKEN

Table 1. ANALYSIS RESULTS SystOliC Blood Pressure "Best" model Parameter Estimates (%) h2 e2 h'2 e'2 Goodness ot Fit Tests Chi-square

dt P-value Shared Environment Tests Chi-square

dt P-value Sex Differences Tests rh = 1 tests Chi-square

dt P-value h

= h'tests Chi-square dt P-value

Herh = 0

66 34 66 34

± ± ± ±

8 4 8 4

Diastolic Blood Pressure hh'erh

64 36 51 49

± ± ± ±

Heart Rate

=0

Herh = 1

11 4 10 6

61 39 61 39

± ± ± ±

8 4 8 4

11.1 13 0.61

11.1 12 0.53

20.3 13 0.09

0.0 1 >0.90

0.0 1 0.84

0.0 1 >0.90

7.9 1 0.01

2.2 1 0.14

1.3 1 0.25

0.7 1 0.40

4.8 1 0.03

2.0 1 0.15

Parameter estimates are ± standard error. From Schieken RM, Eaves LJ, Hewitt JK, et al: The univariate genetic analysis of blood pressure in children (The Medical College of Virginia Twin Study). Am J Cardiol 64:1333-1337, 1989; with permission.)

for systolic and diastolic blood pressures approached zero. 19 The sex difference in the genetic control of blood pressure may be a consequence of the more advanced sexual maturation of the girls. 23 The estimates of heritability in the Medical College of Virginia Study were similar for both systolic and diastolic blood pressure to those reported by the Montreal Adoption Study. Nonetheless, the univariate analyses reported here leave unanswered a range of important questions that can only be addressed by a more refined multivariate genetic analysis of the kind developed by Martin et a1. 11 Multivariate genetic analyses have been performed to investigate the genetic relationships of systolic blood pressure, diastolic blood pressure, and body mass index. 20 The analyses (Fig. 2) suggest that there are genetic paths that are shared between body mass index and systolic blood pressure. In addition, there are shared genetic paths between systolic and diastolic blood pressure. No paths were detected between body mass index and diastolic blood pressure. The genetic path linking systolic and diastolic blood pressure appears to be independent of shared genetic effects with body mass index. This multivariate genetic analysis allowed an initial understanding of the epidemiologic association of body mass index and blood pressure.

Males

Females

Figure 2. Best fitting model for a bivariate genetic analysis of systolic blood pressure (SBP), diastolic blood pressure (DBP), and body mass index (BMI) that allows the variance of SBP to be partitioned. The relative magnitude of a given effect on an observed variable is indicated by the width of the arrow connecting them. Dotted arrows represent causal paths that were fixed at zero (dropped from the model) during the model fitting process. Depicted here is the model for the unlike-sex dizygotic twin pairs; however, the results for like-sex male pairs are the same as those for the males shown here. Likewise, the results for like-sex females are the same as those for the females shown here. \C

10

SCHIEKEN

It allowed the consideration of the effects of systolic blood pressure, diastolic blood pressure, and body mass index simultaneously and the characterization of the genetic and environmental effects separately. Using a related analysis, the correlation coefficients for these variables were separated into genetic and environmental components. Virtually all of the systolic blood pressure-body mass index and the systolicdiastolic blood pressure correlations were explained by genetic effects.

SUMMARY

Familial aggregation, population, and twin studies all point to important genetic influences on the level of blood pressure in childhood and adolescence. Whether a major gene effect operates during childhood has not been determined. The investigation of polygenic paths leads to the study of variables such as ion transport and reactivity paths that appear to be under strong genetic influences. The evidence suggests that abnormalities in these paths might be linked to a prehypertensive state. Univariate genetic analyses of systolic and diastolic blood pressure show that a significant portion of the variability of these variables is under genetic control. Moreover, during early adolescence, boys differ from girls in the regulation of their resting blood pressure. Multivariate genetic analyses show that in these young adolescents, genetic paths shared with body mass index appear to influence systolic but not diastolic blood pressure. The genetic relationship between systolic and diastolic blood pressure appears largely to be independent of body mass index. Genetic studies can partition the genetic and environmental influences on blood pressure and identify shared paths with variables previously believed to be linked epidemiologically. This information may have the capacity to be the framework for public health guidelines developed to lower the incidence of adult hypertension. This article was supported by the National Institutes of Health, and National Heart, Lung, and Blood Institute (R01 HL 31010).

References 1. Brandao AP, Brandao AA, Araujo EM: The significance of physical development on the blood pressure curve of children between 6 and 9 years of age and its relationship with familial aggregation. J Hypertens 7:S37, 1989 2. Canessa M, Adragna W, Solomon HS, et al: Increased sodium-lithium countertransport in red cells of patients with essential hypertension. N Engl J Med 302:772, 1980 3. Clarke WR, Schrott HG, Burns TL, et al: Aggregation of blood pressure in the families of children with labile high systolic blood pressure. The Muscatine Study. Am J Epidemiol 123:67, 1986

GENETIC FACTORS THAT PREDISPOSE THE CHILD TO DEVELOP HYPERTENSION

11

4. Falkner B, Kushner H, Khalsa DK, et al: Sodium sensitivity, growth and family history of hypertension in young blacks. J Hypertens 4(suppl):S38, 1986 5. Feinlieb M, Garrison R, Borhani N, et al: Studies of hypertension in twins. In Paul o (ed): Epidemiology and Control of Hypertension, Symposia Specialists, Miami, 1975, p 3 6. Havlik RJ, Garrison RJ, Feinleib M, et al: Blood pressure aggregation in families. Am J Epidemiol 110:304, 1979 7. Holland WW, Beresford SAA: Factors influencing a blood pressure in children. In PaulO (ed): Epidemiology and Control of Hypertension. New York, Grune & Stratton, 1975, p 375 8. Hunt SC, Williams RR, Smith JB, et al: Associations of three erythrocyte cation transport systems with plasma lipids in Utah subjects. Hypertension 8:30, 1986 9. Intersalt Co-operative Research Group: Sodium, potassium, body mass, alcohol and blood pressure: The INTERSALT Study. J Hypertens 6(suppI4):S584, 1988 10. Joossens JV: Stroke, stomach cancer, and salt: A possible clue to the prevention of hypertension. In Kesteloot H, Joossens JV (eds): Epidemiology of Arterial Blood Pressure. The Hague, Martinus Nijhoff, 1980, p 489 11. Martin NG, Eaves LJ, Kearsey MJ, et al: The power of the classical twin study. Heredity 40:97, 1978 12. McDonald A, Trevisan M, Cooper T, et al: Epidemological studies of sodium transport and hypertension. Hypertension 10 (5 Pt 2):142, 1987 13. Mongeau JG, Biron P, Sing CF, et al: The influence of genetics and household environment upon the variability of normal blood pressure: The Montreal Adoption Survey. Clin Exp Hypertens [A] 8:653, 1986 14. Munger RG, Prineas RJ, Gomez-Marin 0, et al: P~rsistent elevation of blood pressure among children with a family history of hypertension: The Minneapolis Children's Blood Pressure Study. J Hypertens 6:647, 1988 15. Palti H, Ramlal A, Adler B, et al: Familial aggregation of blood pressure, weight and height among sibling pairs at school entry. Journal of Chronic Diseases 38:575, 1985 16. Patterson TL, Kaplan RM, Sallis JF, et al: Aggregation of blood pressure in AngloAmerican and Mexican-American families. Prev Med 16:616, 1987 17. Perrusse L, Moll PP, Sing CF, et al: Evidence that a single gene with gender-and age-dependent effects influences systolic blood pressure determination in a population-based sample. Am J Hum Genet 49:94, 1991 18. Rice T, Bouchard C, Borecki IB, et al: Commingling and segregation analysis of blood pressure in a French-Canadian population. Am J Hum Genet 46:37, 1990 19. Schieken RM, Eaves LJ, Hewitt JK, et al: The univariate genetic analysis of blood pressure in children: The MCV Twin Study. Am J Cardiol64:1333-1337, 1989 20. Schieken RM, Mosteller M, Goble MM, et al: Multivariate genetic analysis of blood pressure and body size: The Medical College of Virginia Twin Study. Circulation, 1992, in press 21. Shibutani Y, Sakamoto K, Katsuno S, et al: An epidemiological study of plasma renin activity in schoolchildren in Japan: Distribution and its relation with family history of hypertension. J Hypertens 6:489, 1988 22. Staessen J, Bulpitt q, Fagard R, et al: Familial aggregation of blood pressure, anthropometric characteristics and urinary excretion of sodium and potassium. A population study in two Belgian towns. Anthro, 1984 23. Tell GS: Cardiovascular disease risk factors related to sexual maturity: The Oslo youth study. Journal of Chronic Diseases 38:633, 1985 24. Turner ST, Johnson M, Boerwinkle E, et al: Sodium-lithium countertransport and blood pressure in healthy blood donors. Hypertension 7:955, 1985

Address reprint requests to Richard M. Schieken, MD Pediatric Cardiology Box 26 MCV Station Medical College of Virginia Richmond, VA 23298-0026