(Hypertension. 1997;30:1554-1559.)
© 1997 American Heart Association, Inc.
Articles |
From the Department of Pediatrics, University of Minnesota Medical School (Minneapolis) (A.R.S.); and the Department of Epidemiology and Public Health, University of Miami Medical School, Miami, Fla (O.G.-M., R.J.P.).
Correspondence to Alan R. Sinaiko, MD, University of Minnesota Medical School, 420 Delaware St., SE, Box 491 UMHC, Minneapolis, MN 55455. E-mail Sinai001{at}maroon.tc.umn.edu.
| Abstract |
|---|
|
|
|---|
Key Words: adolescence blood pressure insulin lipids
| Introduction |
|---|
|
|
|---|
Previous studies have shown a relation between fasting insulin and both blood pressure and lipids in obese adolescents19 20 21 and young adult offspring of hypertensive parents.22 The Bogalusa study reported a significant correlation between fasting insulin and blood pressure for all age groups between 5 and 30 years, with the exception of the 13 to 17-year-old group23 and varying correlations between fasting insulin and lipids that are dependent on age, sex, and race.24 A prospective analysis in a population of Finnish children showed a significant relation between fasting insulin and systolic blood pressure.25
The study presented herein was conducted in a group of 14 to 18-year-old adolescents and their parents participating in a long-term intervention study. The data are intended to further clarify the relation among fasting insulin, lipids, and blood pressure in adolescents before the onset of hypertension and to examine the association of these data obtained in children with similar data obtained in their parents.
| Methods |
|---|
|
|
|---|
The participants were completing a 4-year sodium/potassium clinical intervention trial previously described in detail.26 They had been recruited from the upper 15 percentiles of systolic blood pressure distribution after a school blood pressure screening and randomly assigned to one of the following groups: (1) a low-sodium group instructed on a daily dietary sodium intake of 1600 mg (70 mEq of sodium per day); (2) a potassium chloride supplementation group (capsules of KCl, 1 mEq/kg body wt/24 h); (3) a placebo group (capsules identical to the KCl group); (4) a control group that did not receive any intervention. The three intervention groups were seen every 3 months for 4 years; the control group was seen twice, once at entry into the study and again at the end of the 4 years.
The results of the intervention trial have been published.27 A total of 243 children entered the trial and were assigned to the four groups with stratification by sex. At the final study visit, a fasting early morning blood sample was obtained from 183 participants (87 boys, 96 girls), and this cohort forms the basis for this report. There were no significant differences between group lipid or fasting insulin levels, and therefore, the data from the 183 participants have been combined for this analysis. In addition to the data obtained in the children, fasting blood samples were obtained from 286 nondiabetic parents (164 mothers, 122 fathers). Anthropometric and blood pressure measurements required a separate clinic visit; therefore, it was possible to obtain these data in only 241 parents (143 mothers, 98 fathers), accounting for the lower number when analyses included adjustment for body mass index (BMI).
Blood pressure was measured twice on the right arm, with subjects in the seated position using a random-zero sphygmomanometer, and the average of the two measurements was used in the analyses. Fasting blood samples were drawn into EDTA tubes for the lipid analyses and into serum tubes for the insulin analyses. The tubes were centrifuged immediately, and the plasma and serum samples were kept refrigerated at 4°C. Total cholesterol, HDL cholesterol (HDL-C), and triglyceride levels were measured enzymatically by the Lipid Research Clinic Laboratory in the Division of Epidemiology, School of Public Health, University of Minnesota, as previously described.28 Lipid and lipoprotein analyses were monitored externally every 3 months by the CDC/NHLBI Lipid Standardization Program (Atlanta, Ga) and were monitored internally using the Lipid Research Clinic protocol.28 LDL cholesterol (LDL-C) was calculated with the Friedwald equation.29 Serum insulin was measured by the University of Minnesota clinical laboratory using a radioimmunoassay kit (Equate RIA kit, Binax Corp).
All data are expressed as mean±SE. Comparisons of means were conducted
using ANOVA. Pearson correlation analysis and multiple linear
regression were used for other analyses. A value of
P
.05 was considered to be statistically significant.
| Results |
|---|
|
|
|---|
|
Values for fasting insulin and lipids for children and parents are
listed in Table 2
. Comparisons between
boys and girls and between fathers and mothers presented in
Table 2
were made after adjustment for BMI, although similar results
were obtained in analyses performed before adjustment. Mean
fasting insulin levels were slightly higher in the children than
parents, but these differences were not significant (ANOVA;
F=1.92; P=.13). The lipid pattern was different
between children and adults. Total cholesterol and HDL-C
were significantly higher in girls than boys, but total
cholesterol, triglycerides, and LDL-C were
significantly higher and HDL-C was significantly lower in fathers than
mothers. Fasting insulin was significantly correlated with body size
(weight, BMI, waist-hip ratio, and triceps skinfold thickness) in
children and parents (Table 3
).
|
|
Fasting insulin was significantly correlated with systolic
blood pressure in the children (r=.29, P=.0001)
and remained significantly correlated after adjustment for BMI
(r=.20, P=.0076). This also was true for the
relation between parent fasting insulin and systolic blood
pressure (r=.2, P=.0016 before and
r=.14, P=.0326 after adjustment for BMI). The
correlation between fasting insulin and lipids differed in children and
parents (Table 4
). There was a
significant correlation in children between fasting insulin and total
cholesterol, triglycerides, LDL-C, and HDL-C;
in contrast, fasting insulin in fathers was significantly correlated
only with triglycerides, LDL-C, and HDL-C; in mothers,
fasting insulin was significantly correlated only with
triglycerides and HDL-C. After adjustment for BMI, the
correlation with fasting insulin was no longer significant in children
for any of the lipids but remained significant in the fathers
for LDL-C and HDL-C and in the mothers for
triglycerides.
|
Familial relationships were examined by developing correlations between
parent and child fasting insulin, lipids, and systolic blood
pressure (Table 5
). Before adjustment for
BMI, the relation between fathers and their children was significant
for fasting insulin, triglycerides, and HDL-C, whereas a
significant correlation was found for fasting insulin and all lipids
between mothers and their children. After adjustment for BMI, these
correlations continued to be significant with the exception of the
father-child triglycerides. Systolic blood pressure
was not significantly correlated between parents and children, although
this relation was of borderline significance in mothers
(r=.15, P=.0749). By using multiple linear
regression analysis with the children's systolic blood
pressure as the dependent variable and independent variables of
age, BMI, children's fasting insulin, father and mother
systolic blood pressure, and father and mother fasting insulin,
a significant relationship was shown for mothers' systolic
blood pressure (P=.0251) and mothers' fasting insulin
(P=.0252). Similar results were obtained for all
relationships between parents and children when either triceps or waist
measurements were substituted for BMI in the analyses.
|
| Discussion |
|---|
|
|
|---|
The adolescents in this study were participating in a sodium/potassium intervention trial, raising a question about the effect of low-sodium or high-potassium intake on the measurements included in these analyses. There were no differences among the intervention groups in mean blood pressure, although there was a modest decrease over 3 years in the rate of increase in systolic blood pressure in the girls on the high-potassium and low-sodium diets.27 However, there were no significant differences among the intervention groups with regard to serum sodium or potassium, fasting insulin, lipid levels, or measures of body size. The effect of sodium and potassium on insulin has previously been reported. Fasting insulin in normal, healthy young adults is not significantly different between periods of low (10 mEq/d) and high dietary sodium intake.30 Similarly, a more moderate reduction in dietary sodium (80 mEq/d) did not result in a significant change in either fasting insulin or total cholesterol, triglycerides, HDL-C, or LDL-C compared with a period of high-sodium intake.31 Animal studies have shown a stimulatory effect of potassium on fasting insulin but only following acute intravenous administration, resulting in increases in serum potassium levels of 1 mEq/L or greater; mild elevations of serum potassium (0.3-0.7 mEq/L) did not result in an increase in fasting insulin.32 On the basis of those previous studies and the electrolyte, lipid, and anthropometric data from these adolescents, it is unlikely that the sodium/potassium intervention had a significant influence on the insulin and lipid data in the present study.
The finding that a relation between insulin and blood pressure can be identified in nonhypertensive adults,10 33 in normotensive young-adult children of hypertensives,34 in children,23 25 and in the present group of nonhypertensive adolescents suggests that this association predates the onset of overt hypertension. However, few studies have addressed comprehensively this association in children or adolescents. The Bogalusa study has reported a significant relation between fasting insulin and blood pressure for all age groups between ages 5 and 26 years, with the exception of adolescents 13 to 17 years.23 Although it has been shown in diabetic35 and nondiabetic35 36 children that insulin sensitivity decreases with puberty coincident with important changes in growth, it is not apparent why a gap in this relation should exist between childhood and adulthood or why the Bogalusa data (23) in adolescents should differ from the results of the present study. The study of cardiovascular risk in young Finns conducted in children 3 to 18 years old showed not only significant correlations between fasting insulin and systolic blood pressure independent of body weight but also that fasting insulin at baseline predicts blood pressure after a 3-year period of observation.25
A significant relation between fasting insulin and blood pressure in adolescents was seen in the study described here. After adjustment for BMI, this relation decreased but remained significant. Fasting insulin also was significantly correlated with total cholesterol, triglycerides, LDL-C, and HDL-C, but these correlations were not significant after adjustment for BMI. Correlations between fasting insulin and lipids in the CARDIA study of 18 to 30-year-olds were similar to the results in this study, but they remained significant after adjustment for BMI.37 In the Bogalusa study, fasting insulin was significantly related to both VLDL-C and HDL-C in white adolescents and to VLDL-C in black adolescents, but it was not related significantly to LDL-C in either race.24
Obesity is an important confounding factor in the relation between insulin and blood pressure and lipids,1 but its influence on these risk factors during the first two decades is not clearly defined. Insulin resistance has been investigated more completely in obese adolescents,19 and the relation to blood pressure and other factors has been found to be similar to that in adults. It is associated with chronic sodium retention,38 increased sodium sensitivity,39 and increased forearm vascular resistance,40 and these are reversible with exercise and weight loss.40 41 The data from this study suggest that weight is an important mediator of the insulin effect in nonobese adolescents, as well, because the significant relation to blood pressure decreased and the significant relation to lipids was negated after adjustment for BMI. This is in contrast to the effect of fasting insulin independent of weight noted in young adults in the CARDIA study37 and in older adults, in whom insulin is more commonly correlated with triglycerides and HDL-C.33 Thus, obesity may play a primary causative role in initiating insulin-mediated cardiovascular risk during the first two decades of life, with this association subsequently becoming dependent on other factors beginning in early adulthood.
Fasting insulin was slightly but not significantly higher in the children than parents, but blood pressure, weight, and BMI were significantly higher in the adults. Because of the larger body size and blood pressure measurements in the parents, it might have been expected that insulin resistance and fasting insulin would also be higher in the parents. An explanation for this finding is not readily apparent. However, the higher fasting insulin levels in the children may be related to their age because, as noted above, it is well known that insulin resistance increases significantly during puberty before returning to prepubertal levels in young adulthood.35 36
Fasting insulin and lipid levels in the adolescents were significantly related to levels in their parents, independent of BMI or other measures of obesity. Correlations between the mothers and children were significant for all lipid and insulin measurements, and significant correlations were found for fasting insulin, triglycerides, and HDL-C between fathers and children. The aggregation of lipid levels within families has been previously recognized42 and forms the basis for current lipid screening recommendations in children.42 However, we are not aware of any previous data on direct correlations between offspring and parents or the influence of body size on these correlations. The fact that the correlations between parents and children remained significant after adjustment for BMI suggests that genetic and/or familial environmental factors other than those mediated via weight may control insulin metabolism within families. The stronger relation between the mothers and children does not clarify the degree of separation between genetic and environmental contributions to offspring lipid metabolism and insulin resistance. However, studies of blood pressure in families have shown a greater relation between offspring and mothers beginning at birth43 and extending through grade school,44 45 46 consistent with the significant relation between children's systolic blood pressure and mothers' systolic blood pressure and fasting insulin in this study.
Confirmation that the relation of insulin with blood pressure and lipids is present before adulthood and that levels in children are related to levels in their parents has potentially important implications. First, it lends additional strength to the contention that the roots of essential hypertension and other cardiovascular risk extend back into childhood, and it reinforces the need to investigate these factors before adulthood. If insulin resistance is related etiologically to hypertension, it should predate the appearance of elevated blood pressure or it should increase coincident with a rise in blood pressure in individuals tracking toward hypertensive blood pressure levels. Second, it suggests a potential for the identification of at-risk individuals before adulthood and for introduction of intervention strategies to reduce the incidence of cardiovascular disease. Third, it supports a focus of investigation and intervention on the family in an attempt to detect genetic and/or environmental influences that exert an early developmental effect on cardiovascular health.
Fasting insulin in adults is significantly and positively correlated with subsequent change in lipids and blood pressure during aging.47 It seems reasonable to predict that the same relation may be present earlier in life, and this is supported by the documented abnormalities in insulin resistance that are present in normotensive offspring of hypertensive parents.34 Studies initiated at the earliest stages of the disease will be required to address these issues.
| Acknowledgments |
|---|
Received June 30, 1997; first decision July 23, 1997; accepted July 23, 1997.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. R. Sinaiko, J. Steinberger, A. Moran, R. J. Prineas, B. Vessby, S. Basu, R. Tracy, and D. R. Jacobs Jr Relation of Body Mass Index and Insulin Resistance to Cardiovascular Risk Factors, Inflammatory Factors, and Oxidative Stress During Adolescence Circulation, April 19, 2005; 111(15): 1985 - 1991. [Abstract] [Full Text] [PDF] |
||||
![]() |
National High Blood Pressure Education Program Wor The Fourth Report on the Diagnosis, Evaluation, and Treatment of High Blood Pressure in Children and Adolescents Pediatrics, August 1, 2004; 114(2/S2): 555 - 576. [Full Text] [PDF] |
||||
![]() |
A. R. Sinaiko, R. P. Donahue, D. R. Jacobs Jr, and R. J. Prineas Relation of Weight and Rate of Increase in Weight During Childhood and Adolescence to Body Size, Blood Pressure, Fasting Insulin, and Lipids in Young Adults : The Minneapolis Children's Blood Pressure Study Circulation, March 23, 1999; 99(11): 1471 - 1476. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. R. Sinaiko and R. J. Prineas Reduction of Cardiovascular Disease: What Is the Role of the Pediatrician? Pediatrics, November 1, 1998; 102(5): 61e - 61. [Abstract] [Full Text] |
||||
![]() |
K. Masuo, H. Mikami, T. Ogihara, and M. L. Tuck Familial Hypertension, Insulin, Sympathetic Activity, and Blood Pressure Elevation Hypertension, July 1, 1998; 32(1): 96 - 100. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hypertension Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 1997 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |