Body Mass, Blood Pressure, Glucose, and Lipids
Does Plasma Insulin Explain Their Relationships?
Frangois Cambien, Jean-Michel Warnet, Evelyne Eschwege, Alain Jacqueson,
Jacques Lucien Richard, and Gabriel Rosselin
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
Obesity, hypertension, a high plasma level of glucose, and some llpid abnormalities
(high plasma levels of cholesterol and triglycerides) often occur in the same individuals. Some authors have postulated that the elevated levels of plasma insulin in obese
individuals may explain this association. To explore this hypothesis further, the relationships between body mass index, fasting plasma glucose and insulin, blood pressure, serum lipids, and apoproteins were Investigated in a group of 2144 healthy
middle-aged men. Analysis of the data show that the associations between body mass
index and blood pressure or lipid variables are largely independent of plasma glucose
and Insulin. Plasma glucose is strongly related to blood pressure in nonobese subjects. Plasma Insulin is not associated with blood pressure independently of body
mass index and plasma glucose; however, the simultaneous elevation of body mass
index, plasma glucose, and insulin is strongly associated with blood pressure. The
results also confirm that plasma insulin is positively related to triglycerides and
negatively related to high density Iipoprotein cholesterol independently of plasma
glucose and body mass index. (Arteriosclerosis 7:197-202, March/April 1987)
I
tein lipase and on VLDL synthesis. These observations are
of clinical importance because if insulin is a determinant of
either blood pressure or lipoproteins levels, it might at least
partly explain the frequent association of obesity, hypertension, elevated plasma glucose, and lipids or Iipoprotein
abnormalities in many individuals, particularly those from
industrialized countries and would point to insulin's role in
the determinism of coronary heart disease and suggest
preventive measures. 1011
The purpose of this analysis is to investigate the independent and interactive "effects" of body mass index
(BMI), plasma glucose, and insulin on systolic (SBP) and
diastolic (DBP) blood pressure and on some lipid parameters, after adjustment for the possibly confounding factors
of age, alcohol consumption, and cigarette smoking. The
data used were obtained at the initial examination of the
Paris Prospective Study II (PPS II).
n a previous work using the data of the Paris Prospective
Study I (PPS I),1 we studied the relationship between
blood pressure, heart rate, and plasma glucose and insulin
levels obtained after an oral load of glucose. Analysis of
the results demonstrated that in an active population of
middle-aged men, glucose level was independently associated with blood pressure, whereas insulin was not. From
that analysis, we concluded that insulin is unlikely to affect
the blood pressure level. However, the results of recent
clinical2'3 and epidemiologic4 studies have emphasized
the possible role of insulin as a determinant of blood pressure level and these observations are supported by physiopathological hypotheses stressing the effect of insulin on
sodium transports.5 On the other hand, other investigators
have demonstrated an independent relationship between
plasma insulin and some plasma lipids and lipoproteins6"9
which could be explained by the effect of insulin on lipopro-
Methods
F. Cambien and J.L. Richard are from Unite d'Epidemiologie
Cardio-vasculaire, Institut National de la Sante et de la Recherche
Medicale, Unite 258, Paris. J.M. Warnet and A. Jacqueson are
from the Laboratoire de Recherche sur les Maladies Cardio-vasculaires de la Direction des Affaires Sanitaires et Sociales de
Paris, Paris. E. Eschwege is from Unite de Recherche Statistique,
Institut National de la Sante et de la Recherche Medicale, Unite
21, Villejuif. G. Rosselin is from Unite de Diabetologie et Etude
Radio-immunologique des Hormones Proteiques, Institut National de la Sante et de la Recherche Medicale, Unite 55, Paris,
France.
The Paris Prospective Study II is organized by the Groupe
d'Etude sur I'Epidemiologie de I'Atherosclerose. We thank the
Fondation Francaise de la Nutrition and the Haut Comite d'Etude
et d'lnformation sur I'Alcoolisme for their financial support.
Address for reprints: Dr. F. Cambien, Unite d'Epidemiologie
Cardio-vasculaire, INSERM U 258, Hopital Broussais, 96 rue Didot, 75674 Paris Cedex 14, France.
Received July 7,1986; revision accepted November 24,1986.
Study Population and Testing
The reported results are taken from the cross-sectional
data obtained from a subgroup of 2144 healthy middleaged men at entry into the PPS II who agreed to participate
in the study. These men were not under treatment for
hypertension or diabetes. This investigation on cardiovascular risk factors began in 1982, but follow-up data are not
yet available. The study population was the male employees of a large public organization in Paris. Each subject
answered a series of questionnaires, including one on
smoking habits and one on alcohol consumption.12 A fasting blood sample was drawn and extensive biochemical
measurements were performed, including plasma glucose, insulin, cholesterol, triglycerides, high density lipo-
197
198
ARTERIOSCLEROSIS V O L 7, No 2, MARCH/APRIL 1987
protein (HDL) cholesterol, apo A-1, and apo B. Blood pressure was measured with subjects resting in a recumbent
position by two different examiners using a mercury sphygmomanometer. The same cuff size (12 cm) was used for
every subject. The mean of both measurements is used in
this analysis; DBP is phase V of Korotkoff; BMI is weight
divided by height squared (kg/m2).
A total of 143 subjects who were being treated for either
hypertension or diabetes were excluded from the analysis.
Biochemical Methods
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
Plasma glucose, plasma triglycerides, total cholesterol,
and HDL cholesterol after precipitation of very low density
lipoprotein (VLDL) and low density lipoprotein (LDL)13
were determined by an automatic enzymatic procedure
(Technicon SMA system). Insulin was measured by the
radioimmunologic method with separation on talc.14 The
apoproteins A-1 and B were simultaneously assayed by
immunoelectrodiffusion.15
Statistical Methods
The analysis was divided into two parts. First, the dependent variables (blood pressure and lipid parameters) and
the independent variables (BMI, plasma glucose, and insulin) were considered as quantitative variables. Standard
methods were used to compute the means, standard deviations, and Pearson correlation coefficients.16 The computer program BMDP 1R16 was used to perform the multiple regression analysis shown in Tables 3 and 4. Second,
the independent variables were divided according to their
approximate upper deciles. This threshold was chosen a
priori to get groups of subjects with elevated levels of these
variables. To avoid any biased choice no other grouping
was attempted. The generated dummy variables are conventionally called overweight, high plasma glucose, and
high plasma insulin. The relationships between high plasma glucose and high plasma insulin on the one hand, and
the dependent variables on the other hand in nonoverweight and overweight subjects were tested by a two-way
analysis of variance after adjustment on covariates, by use
of the BMDP 2V program. In the regression analyses and
analyses of variance, plasma triglycerides and Insulin
were log-transformed to get more normal distributions.
Table 1. Mean Values of Variables In This Analysis
Variables
Age
Alcohol consumption (ml/day)
Number of cigarettes (/day)
Body mass Index (kg/m2)
Plasma glucose (mg/100 ml)
Plasma Insulin (^U/ml)
Systolic blood pressure (mm Hg)
Diastollc blood pressure (mm Hg)
Cholesterol (mg/100 ml)
HDL cholesterol (mg/100 ml)
LDL + VLDL cholesterol (mg/100 ml)
Trtglycerldes (mg/100 ml)
Apo A-1 (mg/100 ml)
Apo B (mg/100 ml)
The mean values of the variables obtained in 2144 subjects are shown in Table 1. The correlations between the
variables are given in Table 2. BMI was slightly correlated
with the confounding factors: positively with age and alcohol consumption and negatively with cigarette consumption. It was correlated with SBP, DBP, and with all lipid
variables: positively with total cholesterol, LDL + VLDL
cholesterol, triglycerides, and apo B; and negatively with
HDL cholesterol and apo A-1. Plasma Insulin was not correlated with the confounding factors, whereas plasma glucose was positively correlated with age and alcohol consumption. SBP and DBP were more strongly correlated
with plasma glucose than with plasma insulin. Plasma glu-
43.0 + 4.8
42.0 ±32.0
5.9±9.6
25.3 ±2.9
95.6 ±10.6
8.8 ±8.3
134.2±13.7
83.4 ±9.8
214.3 ±39.8
54.0 ±13.2
160.3±41.2
121.2 ±82.3
155.3 ±22.4
94.9 ±25.2
Data are means ± so; n = 2144 subjects.
cose was slightly positively correlated with total cholesterol, VLDL + LDL cholesterol, triglycerides, and apo B, but
was not correlated with HDL cholesterol or apo A-1. Plasma insulin was slightly negatively correlated with HDL cholesterol and apo A-1, and positively correlated with triglycerides, but was not correlated with total cholesterol, LDL +
VLDL cholesterol, or apo B.
Multiple Regression Analysis
To look for independent associations, a series of multiple regression analyses were performed with blood pressure or the lipid parameters as dependent variables and
BMI, plasma glucose, plasma insulin, and the confounding
factors as independent variables.
Blood Pressure
According to the results of the multiple regression analyses given in Table 3, BMI, plasma glucose, age, and alcohol consumption were independently, positively associated with SBP and DBP, whereas plasma insulin was not.
According to the t values, plasma glucose appears more
strongly associated with DBP than with SBP.
Table 2. Correlations of Body Mass Index, Plasma
Glucose, and Insulin with All Other Variables
Body mass
index
Results
Mean Values and Correlations
Means
Body mass index
Plasma glucose
Plasma insulin
Age
Alcohol consumption
Cigarettes per day
Systolic blood pressure
Diastollc blood pressure
Cholesterol
HDL cholesterol
LDL + VLDL cholesterol
Triglycerides
Apo A-1
ApoB
—
0.100
0.110
-0.069
0.283
0.312
0.159
-0.238
0.230
0.262
-0.112
0.213
Plasma
glucose
Plasma
insulin
0.239
—
0.267
0.168
—
0.150
0.154
-0.002
0.218
0.267
0.135
-0.011
0.134
0.152
-0.012
0.125
-0.048
0.022
-0.020
0.089
0.089
-0.005
-0.140
0.040
0.145
-0.104
0.004
INSULIN, BLOOD PRESSURE AND LIPIDS
and diastolic blood pressure; the significant interaction
term for DBP reflects the fact that the relationship between
high plasma glucose and diastolic blood pressure was
stronger when the plasma insulin level was not elevated
(Table 6).
In overweight subjects, high plasma glucose and high
plasma insulin were positively associated with blood pressure. However, the mean values suggest that blood pressure is elevated when both glucose and Insulin levels are
elevated but not when only one is Increased. This is further
supported by the significant interaction terms in the analysis of variance.
Table 3. Multiple Regression Analysis of Blood
Pressure
Dependent variable
Independent variable
Body mass index
Plasma glucose
Plasma insulin (log)
Age
Alcohol consumption
Cigarette smoking
Systolic blood
pressure
Diastolic blood
pressure
0.228 (10.3)
0.132 (6.0)
0.009 (0.4)
0.083 (3.9)
0.116 (5.4)
-0.015 (0.7)
R 2 = 0.124
0.233 (10.7)
0.177 (8.3)
0.024 (1.1)
0.121 (5.8)
0.085 (4.0)
-0.046 (2.2)
fl2 = 0 .157
199
Cambien et al.
Lipid Variables
Values are standardized regression coefficients (f values in
parentheses). When r > 1.96, p < 0.05; when t > 3.29, p < 0.001;
when f > 4.42, p < 0.00001. flz values are significant at p <
0.00001.
In nonoverweight subjects, plasma triglycerides were
the only lipid variable that was related to high plasma glucose level (Table 7). On the other hand, high plasma insulin was positively associated with triglycerides and negatively associated with HDL cholesterol and apo A-1. Total
cholesterol, LDL + VLDL cholesterol, and apo B were not
increased in the presence of a high level of plasma insulin.
In overweight subjects, high plasma glucose was not
significantly related to the lipid variables, whereas high
plasma insulin was significantly associated with all lipid
variables except apo A-1. Apparently, total cholesterol and
apo B are Increased only when both plasma glucose and
insulin are elevated and not when only one of them is
increased. This is supported by the significant interaction
terms in the analysis of variance.
Lipid Variables
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
As shown in Table 4, alcohol consumption was positively associated with all lipid variables. BMI and cigarette
consumption were positively associated with triglycerides,
LDL + VLDL cholesterol and apo B and were negatively
associated with HDL cholesterol and apo A-1. Plasma glucose was positively associated with total cholesterol,
VLDL + LDL cholesterol, triglycerides, and apo B but not
with HDL cholesterol or apo A-1. On the other hand, plasma insulin was positively associated with triglycerides and
negatively associated with total cholesterol, HDL cholesterol, apo A-1, and apo B.
Discussion
Analysis of Variance
BMI, plasma glucose, and plasma insulin were then divided according to their approximate upper deciles to get
groups of subjects with elevated levels of these variables;
the new dummy variables obtained (overweight, high plasma glucose, and high plasma Insulin) are described in
Table 5. The mean levels of blood pressure and the lipid
variables in the subgroups of the cross classification of
these variables are shown in Tables 6 and 7. The analyses
of variance were performed separately for nonoverweight
and overweight subjects.
In the PPS II examination, the fasting values of plasma
glucose and insulin were obtained, whereas in the PPS I
examination, the fasting and the 2-hour post-load plasma
glucose and insulin values were measured. The previously
reported1 results of the PPS I concerned post-load values;
however, the results using fasting values were very similar
(unpublished results). Furthermore, we consider that fasting values might be more appropriate for the study of the
interrelationships of glucose and insulin with coronary risk
factors since the predictive values of the latter were established in the fasting state.
Blood Pressure
Blood Pressure
In nonoverweight subjects, high plasma glucose was
significantly associated with an increase in both systolic
In this study, the blood pressure levels are likely to be
slightly overestimated in overweight subjects, since the
Table 4.
Multiple Regression Analysis of Lipid Variables
Dependent variable
Independent variable
Total cholesterol
Triglycerides
(log)
HDL cholesterol
LDL + VLDL
cholesterol
Apo A-1
ApoB
Body mass index
Plasma glucose
Plasma insulin (log)
Age
Alcohol consumption
Cigarette smoking
0.125 (5.5)
0.074 (3.3)
-0.044 (2.0)
0.144 (6.7)
0.124 (5.6)
0.045 (2.0)
fl2 = 0.076
0.277 (12.7)
0.081 (3.8)
0.069 (3.3)
0.069 (3.3)
0.061 (2.9)
0.183 (8.7)
H 2 = 0.159
-0.278 (12.7)
0.016 (0.7)
-0.051 (2.4)
0.160 (7.6)
0.209 (9.8)
-0.156 (7.4)
O 2 = 0.144
0.210 (9.2)
0.066 (3.0)
-0.026 (1.2)
0.088 (4.1)
0.053 (2.4)
0.093 (4.2)
R2 = 0.079
-0.120 (5.3)
-0.006 (0.3)
-0.085 (3.9)
-0.010 (0.5)
0.267 (12.1)
-0.078 (3.6)
fl2 = 0.086
0.211 (9.3)
0.068 (3.1)
-0.080 (3.6)
0.039 (1.8)
0.068 (3.1)
0.109 (5.0)
fl2 = 0.077
Values are standardized regression coefficients (rvalues in parentheses). See footnote of Table 3 for p values. R2 values are significant
at p< 0.00001.
200
ARTERIOSCLEROSIS V O L 7, No 2, MARCH/APRIL 1987
Table 5. Body Mass Index, Plasma Glucose, and Insulin in Subgroups of Cross-Classification of Overweight,
High Plasma Glucose, and High Plasma Insulin
Nonoverweight
Overweight
Low glucose
Low insulin High insulin
Number of subjects
1588
24.4±2.1
Body mass index
Plasma glucose
92.9 ±7.4
6.3 ± 3.8
Plasma insulin
143
25.3±2.1
92.7 ±7.8
25.1 ± 10.7
Low glucose
High glucose
High glucose
Low insulin
High insulin
Low insulin High insulin
Low insulin
High insulin
155
25.2 ±2.2
114.0±6.1
6.8 ±4.2
30
25.8 ±2.2
119.9±12.7
29.9 ±13.6
50
126
30.6 ±2.0 30.8 ±2.1
95.1 ±7.1 96.7±6.6
8.5 ±4.3 26.2 ±10.8
33
30.2 ±1.7
113.8 ±7.8
8.6 ±5.3
19
32.7 + 3.7
122.3 ±26.2
33.9 ±15.3
Values are means ± SD. Body mass index (BMI), plasma glucose, and insulin are divided according to their upper deciles (the dummy variables generated are called overweight (BMI > 28.7 kg/m2), high plasma glucose (glucose > 108 mg/100 ml), and high plasma
insulin (insulin > 17 ^U/ml) in the text and tables.
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
same cuff size was used to measure blood pressure in all
subjects despite their weights. However, since the analysis was performed separately in nonoverweight and overweight subjects, the results are unlikely to be affected by
this error of measurement. The results corroborate those
obtained in the previous analysis.1 Both analyses demonstrate that in the entire population blood pressure is more
strongly correlated with plasma glucose than with plasma
insulin, and that there is no independent relationship between plasma insulin and blood pressure after adjustment
for BMI and plasma glucose. However, this overall linear
independence in a general male population, does not exclude the possibility of a specific pattern of relationship at
the upper extremes of the distributions of BMI, plasma
glucose, or insulin. The analysis shows that in nonoverweight subjects, a high plasma glucose level is associated
with a mean increase in SBP and DBP of approximately 9
and 7 mm Hg, respectively, whereas high plasma insulin is
not independently associated with blood pressure. In overweight subjects, the simultaneous elevation of plasma glucose and insulin is associated with a mean increase in
SBP and DBP of approximately 15 and 7 mm Hg, respectively (see Table 6).
The strong relationship between BMI and blood pressure is well known 17 ' 18 and its causal nature is generally
acknowledged since many experimental and clinical studies have shown that weight reduction is associated with a
mean decrease of blood pressure.19'20 However, although
many hypotheses have been proposed, the mechanism of
this relationship is not yet understood. One of the most
recurrent hypotheses points to the increased mean level of
insulin observed in obese subjects as a consequence of
insulin resistance. Indeed, an elevation of plasma insulin
could affect blood pressure by two mechanisms: 1) by
acting on sodium transports, it might induce a positive
sodium balance and favor an extravascular shift of fluid
and sodium;5 2) by acting on the sympathetic nervous
system, it might influence norepinephrine release in the
circulation.21 An independent association between plasma
insulin and hypertension in nonoverweight subjects has
been reported in two studies. Berglund et al.22 found a
higher insulin level in a group of lean hypertensives unselected for glucose tolerance, than in a group of normotensives of similar weight. In a subgroup of the study sample
of the Israel Study of Glucose Intolerance, Obesity and
Hypertension, Modan et al. 4 found a relationship between
serum insulin and hypertension independent of glucose
intolerance and body mass index. In subjects who are not
glucose intolerant the relationship between serum insulin
and blood pressure was almost identical whether or not the
subjects were obese, whereas in glucose intolerant and
diabetic subjects this relationship was less obvious (see
Figure 2 in reference 4). On the other hand, Christlieb et
al. 3 found no relationship between insulin and hypertension in nonobese glucose intolerant subjects, and in the
Paris Prospective Study II, there is no relationship between plasma insulin and blood pressure in nonoverweight
subjects whether they have a normal or an elevated plasma glucose level.
In overweight subjects, the results also differ among
studies. An independent association between insulin and
blood pressure was found in the Israel Study,4 in the study
of Christlieb et al. 3 and in a group of very obese women
studied by Lucas et al.2 No association, however, was
Table 6. Analysis of Variance of Blood Pressure according to Levels of Plasma Glucose and Insulin in Nonoverweight and Overweight Subjects
Two-way analysis of variance*
Blood pressure (mm Hg)
Nonoverweight subjects
Systolic blood pressure
Diastolic blood pressure
Overweight subjects
Systolic blood pressure
Diastolic blood pressure
Low glucose
Low insulin
High insulin
High glucose
Low insulin
High insulin
Glucose
Insulin
Interaction
132.2 ±12.5
81.9 ±9.0
135.2 + 17.3
84.3 ±8.5
141.6±17.3
88.9 + 10.9
142.0 ±16.9
86.1 ±12.1
<0.0001
0.0004
0.24
0.86
0.31
0.007
140.7 ±14.4
88.2 ±10.7
139.2 ±14.8
87.6 ±10.4
139.3 ±13.3
89.8 ±12.0
155.2 ±19.7
96.9±11.7
0.013
0.011
0.001
0.026
0.0007
0.045
Values in the first 4 columns are means ± SD; values in the last 3 columns are probability levels.
'Adjusted on the covariates age, alcohol consumption, and cigarette smoking.
INSULIN, BLOOD PRESSURE AND LIPIDS
Cambien et al.
201
Table 7. Analysis of Variance of Llpld Variables according to Levels of Plasma Glucose and Insulin In Nonoverwelght and Overweight Subjects
Two-way analysis of variance*
Variables
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
Nonoverweight subjects
Total cholesterol
Trtglycerides
HDL cholesterol
VLDL + LDL cholesterol
Apo A-1
ApoB
Overweight subjects
Total cholesterol
Triglycerldes
HDL cholesterol
VLDL + LDL cholesterol
Apo A-1
ApoB
Low glucose
High Insulin
Low Insulin
High glucose
Low Insulin
High Insulin
Glucose
Insulin
Interaction
212.5 ±38.5
111.5±73.9
54.8±13.1
157.8 ±39.9
156.0 ±22.0
93.8 ±24.9
208.9 ±38.8
130.7 ±84.4
50.9±11.5
157.9±41.1
150.1 ±20.9
91.0 ±24.3
222.8±38.8
136.3 ±82.0
57.1 ±15.0
165.7 ±42.3
158.9 ±25.7
98.6 ±26.8
221.2 ±37.3
160.7 ±128.7
53.2 ±14.0
168.0 ±36.9
155.8±22.1
97.8 ±25.3
0.10
0.004t
0.60
0.17
0.33
0.13
0.61
0.004
0.004
0.66
0.025
0.57
0.83
0.83
0.86
0.80
0.77
0.74
220.1 ±42.4
163.2 ±102.6
50.0 ±12.5
170.0 ±42.7
153.8±22.2
102.1 ±26.0
224.6 ±47.1
164.9 ±89.4
47.0 ±10.9
177.5±45.2
147.2 ±20.0
103.1 ±25.4
220.9 ±59.4
151.2 ±99.4
48.8±10.1
172.1 ±55.9
144.6 ±26.0
95.0 ±22.6
246.3 ±42.6
234.0 ±157.2
43.6 ±12.1
202.7 ±45.2
150.2 ±24.7
117.2 ±23.1
0.17
0.11
0.084
0.071
0.36
0.34
0.045
0.009
0.050
0.012
0.58
0.012
0.20
0.036
0.39
0.13
0.15
0.013
Data In the first 4 columns are means ± so; values are mg/100 ml. Data in the last 3 columns are probability levels.
'Adjusted on the covariables age, alcohol consumption, and cigarette smoking.
tT analysis of variance was performed on the log of triglycerldes.
found In the study of Berglund et al. 22 and In the study of
Welnsier et al.23 after adjustment for a different Index of fat
distribution. In the Paris Prospective Study II, we note a
strong interactive association between plasma glucose
and insulin and blood pressure In overweight subjects,
suggesting that these subjects are at a high risk of hypertension only when they are insulin-resistant. It is hard to
explain these different results; the study groups, however,
were very different and factors such as the sex-ratio
and the degree of obesity could have accounted for the
differences.
The strong relationship between plasma glucose and
blood pressure independent of insullnemla and BMI observed In the PPS II and in the PPS I could be the consequence either of the direct effect of glucose on blood pressure or of the influence of a third factor affecting both these
variables. There is not much experimental evidence for a
direct action of glucose on factors affecting blood pressure
Independently of plasma insulin level. However, the high
vascular reactivity observed in diabetics 24 ' a could exist in
nondiabetic subjects who have a mild elevation of plasma
glucose. As we discussed in our former paper,1 the activity
of the sympathetic nervous system could be the link between blood pressure and plasma glucose, and we hypothesize that the pressor effect of the sympathetic
nervous system on blood pressure might increase in hyperglycemic subjects if the passive intracellular diffusion of
glucose Increases the responsiveness of the smooth vascular cells.
Llpld Variables
The results of the regression analyses demonstrate an
independent relationship between BMI and all the lipid
variables. The association was positive with cholesterol,
LDL + VLDL cholesterol, triglycerides, and apo B, and
negative with HDL cholesterol and apo A-1. Plasma glucose was significantly related to cholesterol, LDL + VLDL
cholesterol, triglycerldes, and apo B, but not to HDL cholesterol and apo A-1. Plasma insulin was associated with
all the llpld variables except LDL + VLDL cholesterol. The
analyses of variance show that in nonoverwelght subjects,
high plasma Insulin was more strongly related to the lipid
variables than hyperglycemia; this association was positive with the lipid variables usually linked to the risk of
cardiovascular diseases (LDL + VLDL cholesterol, triglycerides, and apo B) and negative with the so-called protective factors (HDL cholesterol and apo A-1). In overweight
subjects, the patterns of relationship suggest that the state
of insulin resistance that corresponds to a simultaneous
elevation of plasma glucose and insulin is associated with
elevated levels of triglycerides and apo B. However, the
relatively small number of subjects in this particular group
must be kept in mind.
A positive correlation between insulin level and triglycerides has been reported previously.8'8i 9 Some studies have
also shown a negative correlation between insulin level
and HDL cholesterol6'8>9 especially HDLj. 7 Experimental
data indicate that insulin activates llpoproteln llpase activity which Increases chylomicrons and VLDL catabolism in
the circulation;29 furthermore, insulin stimulates VLDL synthesis in the liver.27 These two effects probably play a role
in the association observed in clinical and epidemiologic
studies between insulin and triglycerides. We are not
aware of any direct effect of insulin on the synthesis of apo
A-1 (the main protein constituent of HDL cholesterol) in the
liver. However, the inverse relationship between triglycerides and HDL cholesterol (r = - 0.403 in this study) is well
known and suggests that the effect of insulin on the catabolism of VLDL could partly explain the inverse relationship between insulin and HDL cholesterol.
Conclusions
The results of this epidemiologic study in nondiabetic,
middle-aged men do not confirm the association between
202
A R T E R I O S C L E R O S I S V O L 7, No 2, MARCH/APRIL
blood pressure and plasma insulin Independent of obesity
and plasma glucose observed in other studies. However,
insulin resistance, characterized by high corpulence and
high plasma glucose and insulin levels is a strong correlate
of blood pressure. On the other hand, these results stress
the strong relationship between plasma glucose and blood
pressure in nonoverweight subjects. As already shown in
other studies, plasma insulin is independently positively
associated with triglycerides and negatively associated
with HDL cholesterol.
References
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
1. Camblen F, Jacqueson A, Richard JL, Rosselin G, Duclmetlere P. Associations between systolic blood pressure,
heart rate, and post-load plasma glucose and insulin. In:
Eschwege E, ed. Advances in diabetes epidemiology. Amsterdam: Elsevier Biomedical Press. 1982:189-196
2. Lucas CP, Estigarrlbla JA, Darga LL, Reaven GM. Insulin
and blood pressure in obesity. Hypertension 1985;7:702-706
3. Chrlstlleb AR, Krolewskl AS, Warram JH, Soeldner JS. Is
insulin the link between hypertension and obesity? Hypertension 1985;7(suppl ll):ll-54-ll-57
4. Modan M, Halkln H, Almog S, et al. A link between hypertension obesity and glucose intolerance. J Clin Invest 1985;
75:809-817
5. DeFronzo RA. The effect of insulin on renal sodium metabolism: a review with clinical implications. Diabetologia 1981;
21:165-171
6. Orchard TJ, Becker DJ, Bates M, Kuller LH, Drash AL.
Plasma insulin and lipoprotein concentrations: an atherogenic association? Am J Epidemiol 1983;118:326-337
7. Slemenda C, Orchard TJ, Laporte R, Becker OJ, Russell
R, Kuller LH. Insulin, obesity, and high density lipoprotein
cholesterol (abstr). Am J Epidemiol 1985;122:510
8. Zavaronl I, DaU'Agllo E, Alpl O, et al. Evidence for an independent relationship between plasma insulin and concentration of high density lipoprotein cholesterol and triglyceride.
Atherosclerosis 1985;55:259-266
9. Capaldo B, Rlvellese A, Santoro D, Farlnaro E, Marotta G,
Riccardl G. Relationship between insulin response to intravenous glucose and plasma lipoproteins in healthy men. Artery 1985;13:108-126
10. Duclmetlere P, Eschwege E, Papoz L, Richard JL, Claude
JR, Rosselin GE. 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
11. Pyoraia K, Savulalnen E, Kaukola S, Haapakoskl J. High
plasma insulin as coronary heart disease risk factor. In: Eschwege E, ed. Advances in diabetes epidemiology. INSERM
Symposium 22. Amsterdam: Elsevier Biomedical Press,
1982:143-148
Index Terms:
obesity
insulin
• glucose
• blood pressure
1987
12. Warnet JM, Camblen F, Vernier V, et al. Relation between
consumption of alcohol and fatty acids esterifying serum cholesterol in healthy men. Br Med J 1985;290:1859-1861
13. Jacqueson A, Orosco M, Rakotovao R, Duclmetlere P,
Warnet JM. Statistical evaluation of two routine precipitating
assays for HDL-cholesterol quantitation. Med Lab Sci 1981;
38:57-59
14. Rosselin EG, Assan R, Yalow R, Berson SA. Separation of
antibody bound and unbound peptide hormone labeled with
iodined 131 by talcum powder and precipitated silica. Nature
1966;212:355
15. Fruchart JC, Kora I, Cachera C, Clavey V, Duthllleul P,
Moschetto Y. Simultaneous measurement of plasma apolipoproteins A1 and B by electroimmuno-assay. Clin Chem
1982:28:59-62
16. Dixon WJ. BMDP statistical software. Berkeley, CA: University of California Press, 1983:237-249
17. Chiang BN, Perlman LV, Epstein FH. Overweight and hypertension. Circulation 1969:39:403-421
18. Stamier R, Stamler J, Rledllnger WF, Algera G, Roberts
RH. Weight and blood pressure. Findings in hypertension
screening of 1 million Americans. JAMA 1978;240:16071610
19. Relsln E, Abe R, Modan M, Sllverberg OS, Ellahou HE,
Modan B. Effect of weight loss without salt restriction on the
reduction of blood pressure in overweight hypertensive patients. N Engl J Med 1978;298:1-6
20. Tuck ML, Sowers J, Dornfeld L, Kledzlk G, Maxwell M. The
effect of weight reduction on blood pressure, plasma renin
activity, and plasma aldosterone levels in obese patients. N
Engl J Med 1981;304:930-933
21. Chrlstensen NJ. Catecholamines and diabetes mellitus. Diabetologia 1979;16:211-224
22. Berglund G, Larsson B, Andersson O, et al. Body composition and glucose metabolism in hypertensive middle-aged
males. Acta Med Scand 1976;200:163-169
23. Welnsler RL, Norrls DJ, Birch R. Serum insulin and blood
pressure in an obese population. Int J Obesity 1986;10:
11-17
24. Weldmann P, Beretta-Plccoll C, Keusch G, et al. Sodiumvolume factor, cardiovascular reactivity and hypertensive
mechanism of diuretic therapy in mild hypertension associated with diabetes mellitus. Am J Med 1979;67:779-784
25. Weldmann P, Beretta-Plccoll C, Trost BN. Pressor factors
and responsiveness in hypertension accompanying diabetes
mellitus. Hypertension 1985;7(suppl ll):ll-33-ll-42
26. Nikkila EA, Huttunen JK, Ehnholm C. Post-heparin plasma
lipoprotein lipase and hepatic lipase in diabetes mellitus. Relationship to plasma triglyceride metabolism. Diabetes 1977;
26:11-21
27. Reaven GM, Greefleld MS. Diabetic hypertriglyceridemia.
Evidence for three clinical syndromes. Diabetes 1981;30:
66-75
• body mass index
• lipoproteins
• apoproteins
Downloaded from http://atvb.ahajournals.org/ by guest on June 17, 2017
Body mass, blood pressure, glucose, and lipids. Does plasma insulin explain their
relationships?
F Cambien, J M Warnet, E Eschwege, A Jacqueson, J L Richard and G Rosselin
Arterioscler Thromb Vasc Biol. 1987;7:197-202
doi: 10.1161/01.ATV.7.2.197
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville
Avenue, Dallas, TX 75231
Copyright © 1987 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/7/2/197
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright
Clearance Center, not the Editorial Office. Once the online version of the published article for which permission
is being requested is located, click Request Permissions in the middle column of the Web page under Services.
Further information about this process is available in the Permissions and Rights Question and Answerdocument.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/