Blood pressure, insulin, and glycemia in nondiabetic subjects

Blood Pressure, Insulin, and Glycemia in Nondiabetic Subjects

ARTHUR M. FOURNIER, M.D. MARIA T. GADIA, M.D. DENISE B. KUBRUSLY, M.D. JAY S. SKYLER, M.D. JAY M. SOSENKO, M.D. Miami, Florida

The relation of blood pressure to fasting (basal) insulin and glycosylated hemoglobin (hemoglobin A,) was examined in 248 nondiabetic subjects (137 women and 111 men). None of the subjects was taking antihypertensive medication. There were statistically significant associations of systolic and diastolic blood pressure with insulin levels (r = 0.24, p
From the Divisions of General Medicine and Endocrinology, Department of Medicine, University of Miami School of Medicine, Miami Florida. Requests for reprints should be addressed to Dr. Arthur M. Fournier, University of Miami School of Medicine, Department of Medicine (R-103), P.O. Box 016960, Miami, Florida 33101. Manuscript accepted February 4, 1986.

Subjects. The original study group included 143 women and 114 men. All subjects were employees of Jackson Memorial Hospital or the University of Miami School of Medicine. None of the subjects had a history of diabetes mellitus, and none was taking antihypertensive medication. Sixty-eight percent of the women and 88 percent of the men were white. Procedure. Subjects fasted for 12 hours. Following venipuncture, measurements of height and weight and a medical history were obtained. Blood pressure was measured in the right arm after the subject had been supine in a quiet room for three minutes. The measurement was repeated two minutes later and the average of the two readings was recorded. Measurements

were

made

with

a standard

mercury

manometer

according

to the

guidelines of American Heart Association. Large cuffs were used when

May

1986

The American

Journal

of Medicine

Volume

80

861

BLOOD

PRESSURE

TABLE

I

AND

INSULIN

Characteristics Subjects

IN NONDIABETIC

(mean

Women (n = 137)

SUBJECTS-FOURNIER

ET AL

f SD) of Study

Men (n = 111)

tolerance, and they were also excluded. Thus, 137 women and 111 men were include in the analyses to be described. Characteristics of subjects are presented in Table I. There was little difference between men and women in both systolic and diastolic blood pressures. Women were slightly older and more obese than men. Hemoglobin A+ values were significantly (p
p Value

Systolicblood pressure 119% 15 122 f 12 NS (mm W Diastolicblood pressure 78f 11 77f 10 NS (mm Hg) Age (years) 44.2 f 8.9 41.9 f 11.9 NS Quetelet’s index 2.64 f 0.61 2.59 f 0.33 NS (kg/cm2)X IO3 HemoglobinA+ 6.5 f 0.6 6.2 f 0.6
TABLE II

Correlation Coefficients for Associations Blood Pressure with Insulin” and Hemoglobin A, Levels

Women(n = 137) Insulin Hemoglobin A1 0.28+++ Systolic 0.24++ 0.22+ Diastolic 0.30+++ Log-fastinginsulin. +p <0.05; ++p
of

Men(n = 111) A+ Insulin Hemoglobin 0.23+ 0.20+

0.11 0.07

l

appropriate. Hemoglobin A+ was assayed by column chromatography (Isoiab inc., Akron, Ohio) after prewashing of the hemoiysate with the .temperature maintained at 23OC. insulin was measured by radioimmunoassay [5]. Data Analysis. A logarithmic (base 10) transformation was made to normalize the fasting insulin distribution. All regression procedures and statistical testing were performed on these transformed values. Blood pressure was adjusted to age for each gender on the basis of coefficients derived from multiple regression analyses. Statistical significance for the comparison of means was assessed with the Student t test. Calculations were performed through the Statistical Package for the Social Sciences. Ail reported p values are two-tailed. ion-exchange chromatography underestimates the meausrement of hemoglobin A+ in persons heterozygous for hemoglobin S. Thus, black subjects were screened for this abnormality, and for the eight subjects in whom the sickle trait was detected, the hemoglobin A+ value was adjusted as previously described [6]. RESULTS

Three women and three men who had hemoglobin A+ values greater than 9.0 percent were excluded from the analyses. Two-hour glucose tolerance tests (75 g of carbohydrate) were performed in the remaining subjects who were in the top 10 percent of the hemoglobin A+ distribution. Qf these, three women had impaired glucose

862

May

1986

The

American Journal

of Medicine

Volume

COMMENTS

The study described herein shows a strong association between blood pressure and factors associated with giucase metabolism in nondiabetic subjects and supports a

80

BLOOD

TABLE

III

Systolic (mm Hg) Actual Adjusted Diastolic (mm Hg) Actual Adjusted

IV

AND

INSULIN

IN NONDIABETIC

SUBJECTS-FOURNIER

Actual and Adjusted (to mean age) Blood Pressure Values (mean f SEM) According Highest Tertiles of insulin and Hemoglobin AI Distributions in Women (n = 137) Insulin Highest

Lowest

TABLE

PRESSURE

116f2 117f2 75 f 76f

1 1

p Value

122f2 121 f

2

0.046 0.151

81 f 80 f

2 2

0.007 0.021

Lowest

to Lowest

Insulin Highest

p Value

125 f 122f2

2

0.061 0.023

81 f2 79 f 2

0.005 0.119

114f2 116f2 74f 1 76 f. 1

Systolic (mm Hg) Actual Adjusted Diastolic (mm Hg) Actual Adjusted

119f2 120f 76 f 76 f

2 2 2

123f 123f

p Value

Lowest

and

Hemoglobin At Highest

Actual and Adjusted (to mean age) Blood Pressure Values (mean f SEM) According Highest Tertiles of Insulin and Hemoglobin A, Distributions in Men (n = 111) Lowest

ET AL

to Lowest

an

Hemoglobin At Highest

p Value

2 2

0.253 0.254

122f2 123 f

2

125f2 123 f

2

0.453 0.963

79 f 2 79 f 2

0.218 0.220

77 f 78 f

2 2

78 f 77 f

2 2

0.963 0.719

physiologic role for insulin in blood pressure regulation. Multiple regression analyses were utilized to allow for the effect of adiposity in the analyses. With this allowance, the associations between blood pressure and insulin level were markedly reduced, although still statistically significant. This could be interpreted to mean that the blood pressure-insulin association is largely confounded by obesity and that insulin has little, if any, biologic effect upon blood pressure. An alternative interpretation is that insulin contributes to the known association between blood pressure and adiposity. Indirect evidence supports the validity of the latter interpretation. Insulin plays a physiologic role in sodium regulation in animal models. DeFronzo [4], using a closed-clamp technique in dogs, demonstrated that small increments in insulin levels stimulated increased tubular reabsorption of sodium in the distal nephron. This effect was shown to occur within the physiologic range of insulin values. If insulin has a similar action in humans, this mechanism may explain the observed relation between blood pressure and insulin level. Decreased responsiveness to insulin, mediated by obesity or any other cause, would lead to higher insulin levels in order to maintain blood gluose levels at or near normal. The price paid for this control of blood glucose at higher insulin levels would be increased tubular reabsorption of sodium and a volume-dependent increase in blood pressure. How might the physiologic effect of insulin on blood pressure relate to the development of hypertension? Al-

TABLE V

Correlation Coefficients with Quetelet’s Index

for Associations

Women In = 1371 Systolic blood pressure Diastolic blood pressure Insulin* Hemoglobin A, * Log-fasting insulin. + p <0.05; ++ p
TABLE

VI

Insulin”

(pU/ml)

Insulin + Age (years) Insulin + Age + Quetelet’s (kg/cm*)

0.32t+t 0.32ttt 0.47ttt 0.17+

(* SE) for Multiple Analysis Blood Pressure (mm Hg) Systolic Diastolic

-

13.2 f

3.2++t

9.8 + 2.4+tt

12.4 f

3.lttt

9.2 f 2.3ttt

0.39

O.OB+tt

f

8.6 f 0.37 index X lo3

0.231 0.24t 0.26tt 0.00

tt+ p
Coefficients Regression

Variables Entered

Men In = 1111

3.3+

0.30

i O.OG++t

6.3 f

2.5t

f

0.08Ttt

0.29

f

O.QGttt

5.07 f

1.76t+

3.88

f

1.32+t

of Medicine

Volume

Log-fasting insulin. + p <0.05; ++ p
ttt p
1986

Journal

l

May

The American

80

863

BLOOD

PRESSURE

AND

INSULIN

IN NONDIABETIC

SUBJECTS-FOURNIER

ET AL

though associations of blood pressure with hemoglobin A, and insulin have not specifically been examined in the nondiabetic state, there is evidence for an association between hypertension and glucose metabolism. In one report, systolic blood pressure was found to correlate with glycemic control, as indicated by hemoglobin A, values, especially in women [I]. In the Framingham Study, workers noted an association between systolic hypertension and diabetes [7]. Interestingly, diabetic patients were found to have greater frequencies of hypertension and obesity, and higher lipid levels even prior to the diagnosis of diabetes. There are no data concerning the relation of blood pressure to insulin levels in these studies. Our study complements and extends the work of Modan and her co-workers [B]. They reported an association of hypertension and hyperinsulinemia that held even in a subset of their study population with normal glucose tolerance, as assessed by standard glucose tolerance testing. Our study demonstrates an insulin effect not in hypertension, but in physiologic blood pressure variation. Our study population was much more “normal”; the subjects were younger and more normotensive. Only 4 percent of our subjects met the Modan study criteria for hypertension. Subjects with elevated glycosylated hemoglobin and abnormal glucose tolerance were excluded. The significantly higher hemoglobin At values and lower insulin levels observed in women compared with men point to sex differences in glucose metabolism that are poorly understood. The basis for the association of blood pressure and hemoglobin Al levels that we observed in women but not in men is unclear. lt could, in part, result from the relation between hemoglobin Al and insulin levels (r = 0.28, p
but not the hemoglobin A, level correlated with blood pressure in our male cohort supports the hypothesis that it is insulin per se, and not the osmotic effect of increased serum glucose levels, that is responsible for the blood pressure effect of insulin. Previous arguments concerning the pathogenesis of hypertension in diabetes need to be m-evaluated. It was argued that glucose metabolism cannot play a pathogenetic role in the development of hypertension in diabetic patients, since hypertension often precedes the diagnosis of diabetes or is discovered shortly after the diagnosis [3]. This argument is weakened, however, if variation in glucose metabolism affects blood pressure in the nondiabetic state, as our data suggest. Insulin resistance leading to hyperinsulinemia rather than hyperglycemia is probably the common link between hypertension, obesity, and diabetes. Elevated plasma insulin levels (associated with relative insulin resistance) have been proposed as a risk factor for cardiovascular disease [g-12]. The basis for this increased risk is thought to be mediated by the effect of insulin on lipid metabolism [ 13,141. High-density lipoprotein levels are inversely related to insulin levels in subjects without diabetes [ 15,16]. Our data suggest that another major risk factor-blood pressure-may also be modulated by glucose metabolism, and in particular by insulin, in nondiabetic subjects. ACKNOWLEDGMENT We wish to express our gratitude to Ms. Fanny Fernandez for her secretarial assistance and to Ms. Martha S. Krikeles, Ms. Melanie H. Ashby, Ms. Catherine W. Murphy, and Ms. Marcella Cruz for their technical assistance.

REFERENCES 1.

2. 3. 4. 5. 6.

7.

8. 9.

664

Klein BE, Kline R, Moss SE: Blood pressure in a population of diabetic persons diagnosed after 30 years of age. Am J Public Health 1984; 74: 336-339. Drury PL: Diabetes and arterial hypertension. Diabetologia 1983; 24: 1-9. Christlieb R: The hypertensions of diabetes. Diabetes Care 1982; 5: 50-58. DeFronzo RA: The effect of insulin on renal sodium metabolism. Diabetologia 1981; 21: 165-171. Heding LG: Determination of total serum insulin in insulin treated diabetic patients. Diabetologia 1972; 8: 260-266. Sosenko JM, Fluckinger R, Platt OS, Gabbay KH: Glycosylation of variant hemoglobins in normal and diabetic patients. Diabetes Care 1980; 3: 590-593. Garcia MJ, McNamara RM, Gordon T, Kannell WB: Morbidity and mortality in diabetics in the Framingham population. Diabetes 1974; 3: 105-l 11. Modan M, Halkin H, Shiomo A, et al: Hyperinsulinemia-a link between hypertension, obesity and glucose intolerance. J Clin Invest 1985; 75: 809-817. Garcia-Webb P, Bonzer AM, Whitney D, Masarei JRL: Insulin resistance-a risk factor for coronary heart dfsease? Stand J Clin Lab Invest 1983; 43: 667-685.

May

1966

The American

Journal

of Medicine

Volume

10.

11.

12.

13. 14. 15.

16.

60

Ducimetiere P, Eschevege E, Papoz L, Richard L, Claude JR, Rosselin G: Relationship of plasma insulin levels to the incidence of myocardial infarction and coronary heart disease mortality in a middle-aged population. Diabetologia 1980; 19: 205-210. Enger SC, Erikssen J: High density lipoproteins, insulin secretion and coronary risk factors in latent coronary insufficiency. Stand J Clin Lab Invest 1979; 39: 627-634. Pyorala K: Relationship of glucose tolerance and plasma insulin to the incidence of coronary heart disease: results from two population studies in Finland. Diabetes Care 1979; 2: 131-141. Stout RW: The relationship of abnormal circulating insulin to atherosclerosis. Atherosclerosis 1977; 27: 1-13. Stout RW: Diabetes and atherosclerosis-the role of insulin. Diabetologia 1979; 16: 141-150. Orchand TJ, Becker DJ, Bates M, Kuller LH, Drash AL: Plasma insulin and lipoprotein concentrations: an atherogenie association. Ann J Epidemiol 1983; 118: 326-337. Kubrusly DB, Goldberg RB, Fournier AM, Skyler JS, Hsia SL, Sosenko JM: Associations between plasma lipoprotein components and glycemia in normal women (abstr). Clin Res 1984; 32: 225A.