The effect of epinephrine on immunoreactive insulin levels in man

(1)

The effect of epinephrine on immunoreactive

insulin levels in man.

D Porte Jr, … , T Kuzuya, R H Williams

J Clin Invest. 1966;45(2):228-236. https://doi.org/10.1172/JCI105335.

Research Article

Find the latest version:

(2)

Journal of Clinical Investigation Vol. 45, No. 2, 1966

The

Effect of

Epinephrine

on

Immunoreactive

Insulin

Levels

in

Man

*

DANIEL PORTE,JR.,tALANL.GRABER,TAKESHI

KUZUYA,4

AND

ROBERTH. WILLIAMS §

(From the University of Washington Department of Medicine, Seattle, Wash.)

The mechanisms by which epinephrine elevates

the blood sugar have been of considerable interest for more than 25 years. In an extensive review in 1956, Ellis suggested that the two major in-fluences on glucose metabolism were increased

glycogenolysis in the liver and peripheral tissues, and direct inhibition of glucose uptake by muscle (1). There is, however, the additional possibility that epinephrine might decrease insulin levels.

The introduction of insulin immunoassay (2, 3) has allowed usto examine this third hypothesis in

man. Our data show that epinephrine infusion is

notassociated withariseinserumimmunoreactive

insulin (IRI) despite significant hyperglycemia. In addition, epinephrine was found to inhibit the

expected rise in IRI after administration of

ex-ogenous glucose, glucagon, and tolbutamide.

Methods

Volunteers selected were apparently normal, ambula-tory, nonhospitalized, nonobese, young adult men and women. There were 13 men of ages 21 to 29, weight

range 61 to 95 kg, and height 148 to 197 cm, and 6 women, of ages 21 to 31, weight range 51 to 65 kg, and

height 160 to 179 cm. None had any history of major

*_{Submitted for} _publication _{July 19,} _1965; _accepted

November 2, 1965.

Supported in part by grants AM-02456 and

T1-AM-5020 from the National Institute of Arthritis and Meta-bolic Diseases. A portion of this work was conducted

through the Clinical Research Center facility of the

Uni-versity of Washington (National Institutes of Health

grantFR-37).

Presented in part before the American Society for Clinical Investigation, May 3, 1965, Atlantic City, N.J.

tThis work was carried out during the tenure of an

Advanced Research Fellowship of the American Heart Association and supported in part by the Idaho Heart Association.

tPresent address: Third Department of Internal

Medicine, University of Tokyo, Hongo, Tokyo, Japan.

§Address requests for reprints to Dr. Robert H.

Williams, Dept. of Medicine, University of Washington,

Seattle, Wash. 98105.

medical illness. None had any close relative who was known to be diabetic. All were within 15% of their ideal bodyweight.' Approximately 20% had been previ-ously tested at random with the 100-g oral glucose tol-erance test andwere normal by the criteria of Mosenthal and Barry (4). All had fasting plasma glucose less than 105 mg per 100 ml. For the study, no solid food wastaken after the previous evening meal at 6 to 7 p.m., and no liquids except water were taken after midnight. No smoking was allowed on the day of the test. Upon arriving at the Clinical Research Center for study the volunteer was put to bed, and a slow intravenous 0.85%o

NaCl drip was begun through each of two indwelling venous plastic catheters. Blood samples were taken through one catheter, and drugs or glucose was infused through the other with a constant drive syringe infusion pump. No anticoagulants were given to the patient or were present in the sampling syringe.

Blood samples were obtained every 15 minutes and sampled for measurements of glucose, IRI, free fatty acids (FFA), and glycerol. Blood for insulin assay was allowed to clot for I hour at room temperature, and the serum was frozen at - 190 C until analyzed by the double antibody immunoprecipitation technique (5). Neither epinephrine (1 ,ug per ml) nor tolbutamide (50 mg per ml) had any measurable effects when added to

the immunoassay system in vitro. The glucagon

prepa-ration assayed 5 to 7 AsU of immunoreactive insulin per

microgram of glucagon. This amount of insulin was

considered to have no significant effect upon serum IRI

with the infusion rate of glucagon used (5 sAg per

min-ute). The humaninsulin standard was supplied.2 Blood

samples for FFA, glucose, and glycerol were

anticoagu-lated with heparin and kept chilled at40 C. The plasma

was separated by centrifugation in the cold and frozen

at -19° C for future analysis. FFA were titrated

against standard baseby amodified Dole procedure (6).

Plasma glucose was measured by a Technicon

auto-analyzer with aferricyanidereagent. Glycerolwas

meas-uredbythe enzymatic method of Wieland (7) with

com-mercially available glycerokinase and glycerophosphate

dehydrogenase.8

'As calculated from statistical tables of the

Metro-politan LifeInsurance Co.

2_Lot_no. _10-1862, ₂₄_{U per mg,} _{supplied by}_{courtesy of}

Dr. J. Schlichtkrull, Novo Therapeutic Laboratories,

Co-penhagen.

3C. F. Boehringer and Sons, Mannheim, Germany.

(3)

EPINEPHRINE AND INSULIN

00 tr Ir 00 u) u) u

c0 0 o000000oO 00

_ON 0 0000 0- 00

0) 00

W"-41I%.0Z4

000o o m 00 00

d4MU) e) U) -q _0C4 -- en

00"LO 00 ON 0% Ci 00b t- In ) U')(4e 0 -'U) -4 \

U O0U)Ot t-' C4 00 U) Ci 0 i 0%o 0% _0 0 W-4-- V-- -

-0% 0% - V U) V C4ci 0% -4 0 00 0 0000 0% 0 0% - 00 00_ O -.1 b) C11 00 m 0 0% 0 00 000%r-00 U ₀ Ci U) VoU) em)0 -4

0 0% 0 0% 00 0% 00 0% 00 00

ci --0 V0 0% 4 d4 0

00 0 0% 00 0% 00 0% 00

00 V0 U) +4 c _00

-0%

C 0--U -.Ci4eq'en

u) r 0 N IttoV

- 00 0 0-0 V -0 \0

ci U)04 0' U V" 000

_) c V 0000 _ 0 o U) 00

ci Ci - ci ci c

V0 0 ci V 0o0V o-'000 V - i - _ ci -d4 " _U) _d4 Ci 00 cq

00 -R4 00 1-t It~ dmR ar-. _4 _ . ---q0 I-- U)M cinU 0-- _.d

_ 0 Vo eU) X m 4 0%

_" _ "

C4 o 0 0

*

C:

0

I

ao

- _ZL

=

CLo

U ) 00 %0 ci

oo

ci )U)c

_ _

00 Z0 U) U) 4U )

'4 - 4 00 -4

U) 0 U) V Vo

NO NO co Xo s V V 00VW

t- U) 00U

00 on U) 14 Ci

U) U) 0% U) V

No in _0 O 00 0% 0 00 O% _) U) _~e4

-0 U) -ooo)

0% 00 0% 00 0% 0% 0% +4 ci U

0% 00 0% 0% 0% *

*z

5

NO

5

0 0

U)

00 0C)

0

.'.

-a) .d0

--<E

0_

0

to

U)

0~

. _

00

co

if)

0 Un

if)

0

-4 U)

0

0-%

U)

0

Ul)

0

U)

0)

0

in U) Va

0

Go

U)

0

U) 0

0

U)

0

U)

() :: U)

$)

U)

0

.

U) $

0

U)

a)

$

U)

0$

EN

000 0% 0% 0%

,-I

0% 0 U)0"- Ci

0% 0 ON 00O

ci-'

229

*ci

t

C:

V)

co Cl

0) 4)J

CO

o

r.

a)

*0 0o 00

0-Vo 0- 0 Ci Ci c - U) Ci Ci U) co

U) o ieU)

U) 0-C)Ci 0oU) UC U

U) Voo 0- 0

Ci4 00 00 0% Cq 000000(

V

.Ei

.--o

04

(L)

*

E.

bo

5

a;

o

00 0 C)

VI4 00 ci U) U 00 U) 0% U)

Ul) I-- --t 0 U) U) 0- 0 Cq U)

_- "

(4)

PORTE, GRABER, KUZUYA, AND WILLIAMS

Glucose 300mg /minI.V.

PLASMA0.~c~

GLUCOSAE10| a o , h

0

Mean±95%

5001 _{~~~~~~~~~~~Conf}_idence

TT

I T. interval

IMMUNO-400-]

REACTIVE

INSULIN 300::

200-100 7

60 120 180 240 0 60 120 20240 MINUTES

FIG. 1. A COMPARISON OF THE EFFECTS OF

EPINEPH-RINE ANDGLUCOSE. Fivesubjects were given epinephrine; 10 subjects were given glucose. Each pair of brackets represents the 95% confidence limit of the mean value. Brackets that do not overlap are significant at least at the p<0.05 level. For each study the mean of four con-trol samples taken during the first hour is set at 100%, and all values are compared with this mean control value.

All drugs given to the subjects were commercially

available and included tolbutamide, 1 g per 20 ml;4

glu-cagon, 1 mg per ml;5 and epinephrine, 1,000,ug per ml.6

When epinephrine was infused, it was protected against

oxidation by ascorbic acid (2.5 mg per ml of solution).

Results

The infusion of epinephrine, 6 MAg per minute,

for 1 hour in five subjects resulted in a rise in plasma glucoseto ameanof 168 mg per 100ml,or

180%o above the basal level. Despite this rise in

glucose, serum IRI remained unchanged until

af-ter the infusion was stopped. Within 15 minutes of stopping the

epinephrine

infusion, serum IRI increased to amean of 41 uU perml from amean

basal level of 12 MAU per ml (Table I, Figure

1).

This rapid rise was maximal between 15 and 30

4Upjohn, Kalamazoo, Mich.

5_{Crystalline glucagon,} Eli Lilly Co., Indianapolis, Ind. 6_{Parke, Davis} _& _Co., _Detroit, Mich.

minutes. In a few selected patients the serum IRI was always lower at 5 or 10 minutes than at

15 minutes after epinephrine was stopped. This IRI response was compared to that obtained in 10 subjects with an infusion of glucose (300 mg per minute) in an attempt to duplicate the plasma glucose values found with epinephrine. Although the mean plasma glucose in these 10 subjects was lower at the end of a 1-hour glucose infusion than

attheend of the epinephrine infusion, there was a

significant (p <0.05) rise in mean serum IRI for all four samples taken during the glucose infusion. This increase progressed to a maximum of 250%

above a meancontrol IRI of 7.8 MAU per ml (Ta-ble I). By inspection of Figure 1 it can be seen

that the mean IRI values during glucose infusion are all significantly different from those during epinephrine infusion. Thus, the hyperglycemia associated with an infusion of 6

Mg

per minute of epinephrine should have been sufficient to cause a

significant rise in serum IRI. Two other

sub-jects given prolonged infusions of epinephrine (6

,ug per

minute)

for 4 and 7.5 hours showed similar results. One of these is shown in Figure 2.

De-spite persistent hyperglycemia there was a

mini-mal, if any, rise in IRI until the infusion was

stopped, when the usual postepinephrine rebound

wasfound. Inthese studies, FFA clearly reached

an early peak and then declined during the

infu-sion to basal levels despite the stable insulin val-ues, whereas plasma glycerol remained elevated

throughout theepinephrine infusion.

To determine the effect ofepinephrineupon IRI

when other means of elevating the plasma glucose

.l0n PLASMA GLYCEROL.05

05-FREE 5-FATTY OH

ACID

5_j

1751 1504 PLASMA

I25-GLUCOSE l00-I

7

5-50j

IMMUNO- 2OI REACTIVE

INSULIN 0

MLU/ml. 5

EPINEPHRINE 6,5/imgin.I.V. .i

or_..

2

..~~~~-__~

3

HOURS

4 5 6

FIG. 2. THEEFFECT OF PROLONGED INFUSION OF

EPINEPH-RINEONSUBJECTD.B.

(5)

EPINEPHRINE AND INSULIN or stimulating insulin secretion were used, we

ad-ministered glucagon and tolbutamide. Each

com-pound was given to each subject both with and without simultaneous-epinephrine infusion.

Glucagon, 5 ,ug per minute, resulted in plasma glucose levels similar to those achieved with epi-nephrine, 6 pg per minute, but somewhat higher than thoseobtainedwith glucose, 300 mg per min-ute (Figure 3, TableII). In contrast to

epineph-rine infusion, the serum IRI was elevated during

the glucagon infusion and declined after the infu-sion without a rebound. In contrast to glucose

infusion, the serum IRI levels were much higher

during the glucagon infusion, although both in-fusions were followed by a prompt return of IRI

to basal levels. Despite the rapid decline in IRI

after glucagon infusion each patient experienced

atypical "hypoglycemic reaction" within 40 to 50 minutes. These reactions were characterized

by

subjective symptoms of nervousness and

appre-hension and accompanied by increased

perspira-tion and tachycardia. They were coincident with

the lowest plasma glucose and were followed bya

prompt rise in plasma glucose and FFA

(Figure

3, Table II). This type of reactionwas notfound

after glucose or epinephrine infusions. When

1. V. Glucagon 5jug1min 300-FREE 200T FATTY ACID 100-0i 915 PLASMA 14 GLUCOSE q.) ~15 QZ b)4 a) 0 0 m a) 0)4 $) 0) a

Meont 95 %

Conf idence intervol

60 120 110 240 0 60 120 180 Minutes

FIG. 3. A COMPARISON OF GLUCAGON INFUSION WITH

GLtJCAGON AND EPINEPHRINE INFUSION. The same four

subjects weregivenboth infusions. See Figure 1.

06 4.0 04 bed ECu CO ._ ._ 0 0 vC E 4; > _ 00 0. E. -E ._

HE

0 tIo -0% 0 W) 0 0 0 O -0 0 0 0 m 0 0% _ 0 00 U) 40 0 _4 0 0 _ Ua U) t-0 0 o u: _I tQ

0 t-e _ 0o oU)00

00t'0O(LN NU),

r-In00

(f)%000

%d Ln0t~ji

*

V) 00 o 0 000- noCj_~

-

-_\0t- CON 00\O

CM4 in)00

*_o) 00_ 0 Oo0%0o

44~~~~~~~b^1a C,

-'--- bOD N-0%0

C.)

NO ' l0o

ZL00 f%-+ to )0

-i _in) --4-.-4 ~CM4CM4C

bo 00\%000Cf) 0 dg

_, _

,

_ \_

-,,

_

\_

-_4

C1 mene,m C- enC N----N CL

C-*-C

-~~~~0%O--- - -- CM

-4)

-cmaNf~E.dN ,0CM-f)

t-o00 .00 CM 00%0\0

0 mCK00 \000%It

C0N000%00 0\00 0%00

~~CU'0%~4 t~00.v

LO en \C

-if)00#) 0+ * _ ._ :-if) 0 co 0 -4

if)%o-414

1--4 0000 )nm

CKeCM 000

if)_mE _mE_in)

if)

WIf)

0_%0

I--d4CM_CM4

d410-ON

__

CM40cM-I-4

Ui) 00

V-O 0C-r -CMCMi)

enr- ON

r-CE) O

--f

mIC o

r-_4vE

* _-cq+CN U.) 00 ba

C0C0--4 m' Or~

_ _2e eq

CU

C.. +

-oI,.00 -i

.1000C14

(6)

0r co 00

-0iu) '0 U) o%

0%tr20 000U%0 '0

N~~~ ~c_com N

r(t4 - '0t_ o0%

co(- 00 (C4

U) )tnt- inco m C-0' 000-.0-'0 0) U).-2.0)t-C7

.2_2_ _

C4 Q~CD"W)DC4 in U) 2-.

o0%in eoe 20 U NO) I on

.0

26i 4 U) (4 in~ M

0

obeu bv - 0

ui 00

(-4 (- U U 0 _t-coo o- W

b .-oU)0 00W

'0

(-)40% *w m e

0%0.0000 0

0Pe

0t0UU U)U)000o

00c0 0%00 '0U)%OU)'0s

0U)nU)Io 0 Co 0 0sC> 0

00 In 0o r-0% -.C% 0%

2-0% 2- 0% 0%00 0% t- c0 0% 00 m 000NNo 0% t-0% 0% 00

0% 0-0o00000 O %%02-r _

'0(-0% _-4

-0(-4~ ~ - C

0 2U = U) '

_ _ _ *E s 0 > uI) 0o ) 00 U u) r-. -4 N00

20

m wIr m _Ntv cd C)3m _

0-0

ceomtn)

.0

140o 0 U 4

N NNt- b

I-t-b 000

00 -40

0 \00U U) - N

0N co t0 - 000. 2-o 0000

O2s -0'

00rO b :\

-v

-0

0

)X

0%0r0.(-40 0- U)%00r

0%0-_0 00)%

glucagon, 5 ,ug per minute, and epinephrine, 6 ,ug

per minute, were infused simultaneously into the same four subjects, the serum IRI response was

significantly less during the infusion (p < 0.05 by paired comparisons for all four samples). When the infusion was stopped, the serum IRI levels rose in a pattern similar to that when epinephrine was given alone, and they were significantly higher than after a 1-hour glucagon infusion (Figure 3, TableIII).

In an analogous fashion sodium tolbutamide, 1 g in a single iv injection, was given to five sub-jects on 2 different days. On 1 day epinephrine, 6 ug per minute, was infused for 15 minutes be-fore and 45 minutes after tolbutamide, whereas on the other day tolbutamide alone was given. The insulin response was significantly diminished by

the epinephrine (p <0.05 for the 10- and

20-minute post-tolbutamide samples), and the usual pattern of rebound of IRI after termination of the epinephrine infusion was observed (Figure 4, Ta-bleIII).

- g Tolbutomide IN. _Epinephrine

& ephr I% FREE FATTY ACID ztj @ PLASMA GLUCOSE 1 1 ztl 2H~ 0; E 0) 20 -._4 .0₀ V '0 20 0 0 ._ 0; 0 0 10001 800- IMMUNO-REACTIVE 600-INSULIN 400-200

6 0 12 io 0b 6o 1 0 160 2io

MINUTES

FIG. 4. A COMPARISON OF THE EFFECTS OF 1 G OF

TOLBUTAMIDE GIVEN ALONE OR DURING AN EPINEPHRINE

INFUSION. The same four subjects were tested twice.

(7)

An attempt was then made to compare the effect of exogenous glucose upon the IRI response. This study was complicated by the much higher glucose levels when glucose, 300 mg per minute,

and epinephrine, 6 ,ug per minute, were infused

simultaneously into four different subjects. Com-plete inhibition of insulin rise was noted during the infusions in two subjects, but only partial in-hibition was observed in the other two (Figure 5,

Table IV). The partial inhibition was suggested

by the observation that all subjects showed a large rebound of serum IRI after the infusions were

stopped, with a peak between 15 and 30 minutes.

In these latter studies, when both substances were infused, glycerol was measured tohelp distinguish between afall in FFAcausedby decreased

lipoly-sis (decreasing glycerol) and that caused by

in-creased uptake, re-esterification of FFA (stable

or elevated glycerol), or both. This is based on

the evidence that insulin lowers plasma glycerol,

reflecting a direct inhibition of adipose tissue

lipolysis (8-10). In the two individuals (D.K.,

M.O.) who showed no IRI rise during the

com-bined infusions, there was a relative plateau of

0 0) U 4)0 'I. 00 0 U 4)0 0) 00 - 0) 0) 0. .4) 0) 4) 0) 0 4) 0 00 4) *00 00 0 0) 4) 4) 00 0 U 0)

--I20 180 240 0 60 120 180 240 MINUTES C) 06 0 0 02 Cd C Cd 0

0s

0 0 WI 0 U, 0 0z U) rn 0 00 0 C4 W) O 00 U,) '0 0 V) _ 00 0 U) 0 0 04 0 '0 '4. 0 OS 00 0 U) :6 Un

0n0 '.0

c;

.%_m_-I- !00 U))

_ _4 _4

-O0Go0 m .; . 0000O

0 -_

_ _ *

0 0

o C)C4Co

0 _~~~~0

u v

00 in0W

. ,~~~~~.

+ m N -4 c0o N e

C4_ 4_ 4_ _ t ,

_ x + .= +c= o C

0(e00

C-2 m

-2t

00,0 =

-0 OCh

00000C7

U) t- Q\

-000000

000 t-O 0 - 0

0 0

t

t-000 0

co 0co O~

4

- co0~ W) vo U,)

' m0(4 N0

U.)_'tC4 0 in

_00t-4; 4;

00''00) )00co00

+ m ov -blz< + U)covoo'

-G 0 - N _

00_ 0_

;o0 co 00' ;+0000'0

00

o o cK *

r-- I -ON

r00 00 in0c 0

4- -Co= Co= :>U000'0 233 FREE FATTY ACID I-i R PLASMA GLUCOSE t-14 C.~ MMU N-REACTIVE INSULIN 0 C-(-4C Cd Cd C: E 00 4.0 C C 0 Gq 0 Cd v .0 Cd I*

FIG. 5. A COMPARISON OF THE EFFECT OFEPINEPHRINE

WITH EPINEPHRINE AND GLUCOSE INFUSION. Five

sub-jects were given epinephrine; four different subjects

weregivenboth infusions. SeeFigure 1.

-1

(8)

GLUCOSE GLUCAGON

±95% CONFIDENCE T ~$T> INTERVAL

Ti=120min. T 33mln.

0 30 60 0 1530 60

MINUTES POST INFUSION

FIG. 6. A COMPARISON OF THE MEAN RATE OF FALL OF

THE PLASMA GLUCOSE AFTER A GLUCAGON INFUSION WITH

THE RATE OF FALL AFTERA GLUCOSE INFUSION. ti=

half-time.

plasma glycerol, whereas in the two subjects with

a modest but definite rise in IRI during the

com-bined infusions (C.H., D.M.) there was an early

rise and then a decrease of about 50%o in glycerol

levels by the end of 1 hour, despite the continuing

infusion of epinephrine (Table II).

Discussion

Epinephrine administration was associated with

stable basal insulin levels despite hyperglycemia

when given alone, and it inhibited the usual

in-creases in serum IRI produced byexogenous

glu-cose, glucagon, and tolbutamide. Those subjects

with the greatest insulin responses to tolbutamide,

glucose, or glucagon alone tended to be the same

subjects in whom a partial response was still

ob-served when such compounds were infused

simul-taneously with epinephrine. This indicates that

the inhibition can bepartially overcome by a

suffi-ciently strongstimulus. Because of the variability

in the magnitude of insulin responses in different

subjects to any specific stimulus tested, a high

degree of statistical significance was not always

found in these studies. Therefore, caution must

be exercised in the quantitative interpretation of

results from one study. However, considered

to-gether, the pattern of responses was _always the

same when epinephrine was given, i.e., there was

alwaysalesserincrease inthe IRI levelduringthe

insulinogenic stimulus plus epinephrine infusion than was observed for the insulinogenic stimulus alone. Furthermore, a rapid postinfusion in-crease in IRI occurred in every experiment in which epinephrine was used either alone or with other compounds. This constant pattern, al-though not susceptible to statistical analysis, tends

to reinforce the conclusion that epinephrine does indeed prevent or diminish an insulin response. The magnitude of the rebound after epinephrine was variable but appeared to be greater in those subjects with larger IRI responses to glucose, glu-cagon, or tolbutamide alone. The variation in in-dividual response was not related to obesity (11) since none of the patients was obese, nor was it found to correlate well with sex, body surface area, height, or weight. No measurements of plasma epinephrine levels were made, and it is possible these might have correlated better with the degree of epinephrine inhibition. Such variations in the magnitude of a normal IRI response make it difficult to delineate a quantitatively pathologic response without a much larger experience with apparently normal population groups. However, the glucagon study was unusual in several

re-spects. The IRI levelswere higher than observed with glucose alone, all values for IRI during glu-cagon infusion were higher, and the 1-hour value

was statistically different at the p <0.05 level;

more striking was the observation that each

in-fusion was followed by clinically obvious

hypogly-cemia. This late hypoglycemic phase after

glu-cagon infusion has been observed by others (12,

13) and, together withourIRI data,suggests that glucagon may raise insulin levels by some mecha-nism in addition to its

hyperglycemic

effects.

Such aninference is supported by noting the rate

offall of the blood sugar after these infusionswere

stopped. The higher insulin values during a

glu-cagon infusion were followed by a faster rate of

glucose disappearance, whereas the lower insulin

values during a glucose infusionwere followed

by

a slower rate of glucose disappearance

[half-time

(ti)

= 33 minutes, glucagon; t4 = 120

minutes,

glucose; Figure

6].

These observations are

con-sistentwith the hypothesisthateither moreinsulin is secretedorlessinsulin is

degraded during

a

glu-cagon infusion thanduringa

glucose

infusion with

comparableblood sugar levels.

200-PLASMA GLUCOSE

(%OF

MEAN\IOO-\CONTROL /80-

60-401

(9)

The more prolonged epinephrine infusions

indi-cate that epinephrine's suppressive effect on IRI levels can be maintained for hours. In addition we noted, as have others (14, 15), that after an initial rise FFA levels decline during continuous epinephrine infusions. This fall in FFA was at-tributed to the effects of insulin and glucose to re-esterify fatty acids in adipose tissue. The ob-servation in our studies that glycerol levels re-mained elevated while FFA declined is consistent with low levels of insulin during epinephrine in-fusion, because it has been shown that raising insulin levels by exogenous insulin administration lowers glycerol levels (8, 10). If low insulin lev-els were present, declining FFA levlev-els could still result from either increased re-esterification of FFA in adipose tissue by the high glucose levels alone, by increased peripheral uptake of FFA secondaryto thecardiovascular effects of epineph-rine, or by both.

The hypothesis that epinephrine inhibits the secretion ofinsulin was suggested as early as 1930 by Colwell and Bright (16) when they noted a low respiratory quotient during epinephrine infusions in cats. These early studies have been recently strengthened by studies in vivo in dogs, when pan-creatic venous blood was biologically assayed for insulin after an epinephrine infusion (17), and by in vitro studies in rabbit pancreatic slices, when epinephrine was found to prevent glucose-stimu-lated release of IRI (18). Although this latter study suggests that the inhibition is a direct effect upon pancreatic beta cell function, several other possible mechanisms can be suggested. It is pos-sible that circulatory changes associated with epi-nephrine may have been responsible for theeffects upon insulin levels. In particular a decreased splanchnic or liver blood flow could result in less insulin release into the general circulation without any change in pancreatic response. Studies of hepatic blood flow in man with doses of epineph-rine comparable to those used in our study have shown, however, that increased rather than

de-creased splanchnic blood flow would be expected (19, 20). We have estimated splanchnic blood flow with indocyanine green (21, 22) in two sub-jects by using a central injection site and

periph-eral blood sampling site (23). In both subjects clearance of indicator was unchanged by an

epi-nephrine infusion. With the limitations of this

technique, these results are compatible with either

no change or an increase insplanchnic blood flow, but not a decrease. Therefore, a reduction in

splanchnic flow as an explanation for the effect of

epinephrine on blood insulin levels would appear

unlikely. Alternative possibilities, which can be

summarized briefly, include a) adirect or indirect

effect of epinephrine on the pancreatic

microcir-culation; b) production of a metabolically active

substance by epinephrine, which in turn decreases

pancreatic insulin release; and c) acceleration of

insulin removal that prevents the usual rise in

insulin level., None of these other alternatives were investigated and, therefore, further studies inman will be required to evaluate them.

Summary

Epinephrine, 6

jug

per minute, was infused

into normal subjects. Despite marked hypergly-cemia, immunoreactive insulin remained at basal

levels until the infusion was stopped. This inhi-bition of an expected increase in immunoreactive insulin was maintained for up to 7 hours by

con-tinuous infusion of epinephrine. Immunoreac-tive insulin levels were lower thanexpected when epinephrine was infused simultaneously with glu-cose, glucagon, and tolbutamide. These data are

considered to be consistent with an inhibitory ef-fect of epinephrine uponpancreatic insulin release.

Acknowledgments

We wish to thank Susan Page, Marlene Zuti, Alison

Skeel, and Ron Harbeck for their skillful technical

as-sistance. Dr. Francis C. Wood, Jr., kindly reviewed the manuscript, and Dr. Ruth Kirk assisted with the statistical interpretations.

References

1. Ellis, S. The metabolic effects of epinephrine and

related amines. Pharmacol. Rev. 1956, 8, 485.

2. Yalow, R. S., and S. A. Berson. Immunoassay of

endogenous plasma insulin in man. J. clin. Invest.

1960, 39, 1157.

3. Morgan, C. R., and A. Lazarow. Immunoassay of insulin: two antibody system. Plasma insulin levels of normal, subdiabetic and diabetic rats.

Diabetes 1963, 12, 115.

4. Mosenthal, H. O., and E. Barry. Criteria for and

interpretation of normal glucose tolerance tests.

Ann. intern. Med. 1950, 33, 1175.

(10)

5. Samols, E., and D. Bilkus. A comparison of in-sulinimmunoassays. Proc. Soc. exp. Biol. (N. Y.)

1964, 115, 79.

6. Trout, D. L., E. H. Estes, Jr., and S. J. Friedberg.

Titration of free fatty acids in plasma: a study of current methods and a new modification. J.

Lipid Res. 1960, 1, 199.

7. Wieland, 0. Eine enzymatishe Methode zur

Bestim-mung von Glycerin. Biochem. Z. 1957, 329, 313. 8. Hagen, J. H. The effect of insulin on

concentra-tion of plasma glycerol. J. Lipid Res. 1963, 4, 46.

9. Jungas, R. L., and E. G. Ball. Studies on the

me-tabolism of adipose tissue. XII. The effects of

in-sulin and epinephrine on free fatty acid and

glyc-erol production in the presence and absence of

glucose. Biochemistry 1963, 2, 383.

10. Mueller, P. S., and W. H. Evans. Responses of plasma glycerol concentrations to epinephrine,

norepinephrine, glucose, insulin, and prolonged

fasting in man. J. Lab. clin. Med. 1963,61, 953. 11. Karam, J. H., G. M. Grodsky, F. Ch. Pavlatos, and

P. H. Forsham. Critical factors in excessive

se-rum-insulin response to glucose. Obesity in

ma-turity-onset diabetes and growth hormone in

acromegaly. Lancet 1965, 1, 286.

12. Roth,J., S. M.Glick, R. S.Yalow, and S.A. Berson.

Secretion of human growth hormone: physiologic

and experimental modification. Metabolism 1963,

12, 577.

13. Bondy, P. K., and L. R.Cardillo. The effect of

glu-cagon on carbohydrate metabolism in normal

hu-man beings. J. clin. Invest. 1956, 35, 494.

14. Havel, R. J., and A. Goldfien. The role of the

sym-pathetic nervous system in the metabolism of free

fatty acids. J. Lipid Res. 1959, 1, 102.

15. Shafrir, E., K. E. Sussman, and D. Steinberg. The nature of the epinephrine-induced hyperlipidemia

in dogs and its modification by glucose. J. Lipid Res. 1959, 1, 109.

16. Colwell, A. R., and E. M. Bright. The use of con-stant glucose injections for the study of induced

variations in carbohydrate metabolism. IV.

Sup-pression of glucose combustion by continuous

pro-longed epinephrine administration. Amer. J. Physiol. 1930, 92, 555.

17. Kosaka, K., T. Ide, T. Kuzuya, E. Miki, N. Kuzuya, and S. Okinaka. Insulin-like activity inpancreatic

vein blood after glucose loading and epinephrine

hyperglycemia. Endocrinology 1964, 75, 9.

18. Coore, H. G., and P. J. Randle. Regulation of

in-sulin secretion with pieces of rabbit pancreas

in-cubated in vitro. Biochem. J. 1964, 93, 66.

19. Bradley, S. E. Liver injury in Transactions of the Fifth Conference, Josiah Macy, Jr., Foundation,

F. W. Hoffbauer, Ed. 1946, p. 38.

20. Kibler, R. F., W. J. Taylor, and J. D. Myers. The effect of glucagon on net splanchnic balances of

glucose, amino acid nitrogen, urea, ketones, and oxygen in man. J. clin. Invest. 1964, 43, 904.

21. Caesar, J., S. Shaldon, L. Chiandussi, L. Guevara,

and S. Sherlock. The use of indocyanine green in

the measurement of hepatic blood flow as a test

of hepatic function. Clin. Sci. 1961, 21, 43.

22. Wiegand, B. D., S. G. Ketterer, and E. Rapaport.

The use of indocyanine greenfor the evaluation of

hepatic function and blood flow in man. Amer. J.

dig. Dis. 1960, 5, 427.

23. Porte, D., Jr., A. L. Graber, T. Kuzuya, and R. H.

Williams. Unpublished data.