However, the mean percentage of insulin unac- counted for was similar ... Data are presented as microunits rounded to the nearest tenth. t Microunits of insulin ...
s
tudies on the Relationship Between Insulin Concentrations and Insulin Action S. EDWIN FINEBERG AND STEPHEN H. SCHNEIDER
The relationship between insulin concentrations and insulin effects was assessed using a 30-min steadystate perfusion of insulin across the human forearm and a 90'inin recovery period in 17 normal men. During perfusion, calculated insulin increments in forearm arterial plasma insulin were 87 ± 4 (group I), 161 ± 17 (group II), and 333 ± 55 /xU/ml (group III), respectively. Measured venous insulin increments were 33 ± 4, 66 ± 6, and 231 ± 2 7 /xU/ml. During perfusion, venous and arterial increments were linearly related (r = 0.88, P < 0.001). With discontinuation of perfusion, venous increments of insulin became undetectable after 15 min in groups I and II, and after 30 min in group III. Of the total microunits of insulin perfused, 46.0 ± 11.0%, 45.3 ± 9.4%, and 36.5 ± 4.8%, respectively, remained unaccounted for 90 min after perfusion. Effects of insulin on arteriovenous differences of FFA and potassium persisted throughout the recovery period, with peak effects occurring after perfusion for all groups. Estimated interstitial insulin levels in the three groups fell below 10 ju,U/ml by 45, 60, and 90 min after perfusion, respectively. Although peripheral tissues had a significant capacity to sequester insulin, the persistence of biologic effects was not consistent with increased concentrations within the interstitial spaces. Effects of insulin upon glucose waned first, followed by effects upon potassium and t h e n Hpolysis. DIABETES CARE 5: 292-299, MAY-JUNE 1982.
T
he introduction of continuous subcutaneous insulin infusion (CSII) and low-dose intravenous insulin therapy in the treatment of patients with insulin-dependent diabetes and diabetic ketoacidosis1"3 has increased the need to better understand the relationship between measured hormone concentrations and the biologic effects of insulin. Studies in type I diabetes using CSII suggest that the relationship between venous insulin concentrations and glycemic control is not easily defined.4-5 In addition to such factors as age, nutritional state, etc., the concentration of biologically active insulin immediately surrounding the cell should determine its effects. Modeling studies by Sherwin et al. demonstrate that the effects of insulin are proportional to its concentration within a slowly equilibrating tissue compartment.6 In these studies, venous plasma insulin concentrations were not measured nor were the effects of insulin on metabolites other than glucose assessed. Zierler and Rabinowitz demonstrated that the effects of increments in insulin concentrations (calculated mean arterial increment of 38 /aU/ml) persisted for up to 40 min after terminating the perfusion of insulin across forearm tissues.7
292
When insulin concentrations in the artery were increased by 300- 700 jitU/ml, effects on glucose uptake (not seen during low-dose studies) persisted for up to 60 min after insulin perfusions were discontinued.8 Venous hormone concentrations were not measured during these studies. Subsequently, it has been shown that the effects of insulin persist for several hours after the termination of CSII and intravenous insulin infusion.4*9'10 These studies suggest that measured venous concentrations reflect only indirectly the concentration of insulin reaching the tissues. In circumstances when the insulin concentratfons decrease rapidly, venous insulin concentrations appear to be a relatively poor indicator of the effect of insulin. This is understandable since insulin action probably results from postreceptor cellular factors, as well as the net impact of binding, degradation, and sequestration of insulin within a given tissue.11 The purpose of these studies was to assess the relationship between insulin concentrations and insulin effects in human peripheral tissues. Use of the perfused human forearm enabled us to calculate increases in arterial insulin concentrations and to directly monitor insulin sequestration, return of insulin to the circulation, and various effects of insulin upon
DIABETES CARE, VOL. 5 NO. 3, MAY-JUNE 1982
RELATIONSHIP BETWEEN INSULIN CONCENTRATIONS AND ACTION/S. E. FINEBERG AND S. H. SCHNEIDER
physiologic saline were perfused. During insulin perfusions, 1.3 ng/kg/min of insulin was perfused in 10 subjects and 4 ng/kg/min was perfused in seven individuals. Separate syringes and tubing were used for baseline, insulin perfusion, and recovery periods. Seven subjects achieved high insulin increments, with perfusion of the higher dose (group III). Insulin concentrations achieved differed between two groups of individuals (low, group I and intermediate, group II) receiving the lower dose of insulin. This is explainable on the basis of differences in forearm plasma flow (see RESULTS). Single component pork insulin (courtesy of Dr. John Galloway, Eli MATERIALS AND METHODS Subjects. Seventeen healthy, nonobese men were studied. Lilly Co., lot xx0240AMX/C2Ct-30164A, Indianapolis, InThese individuals had no family history of diabetes and regu- diana) was used for all studies. Recovery period perfusates larly ingested at least 150 g of carbohydrate and 1 g of protein were identical to those used during baseline. Sample handling. Seven milliliters of blood was obtained per kg daily. Informed consent was obtained before each experiment. Procedures conformed to the principles of the Hel- from each site at each sampling time. One milliliter was sinki Declaration and were performed in the General Clini- placed in a disposable Wintrobe tube for determination of cal Research Centers of Boston University Medical Center plasmacrit. The remainder of the sample was double spun at and Indiana University School of Medicine. Perfusions were 4°C. This plasma was used for determination of concentrations of dye, FFA, glucose, potassium, and insulin. All conbegun after an overnight fast of 12-14 h. 12 Technique. A minor modification of the forearm perfusion centrations were corrected for heparin-filled dead space. Analytic methods. FFA, glucose, and insulin concentratechnique of Andres and Zierler was used.12-13 Subjects lay tions were determined in duplicate, and potassium was detersupine on a padded platform with the arm to be studied supported on an armboard. After skin anesthesia was obtained mined in triplicate. Glucose was analyzed by a glucose-oxiwith 2% Xylocaine, deep and superficial venous catheters dase colorimetric method. Potassium concentrations were were inserted in a retrograde manner. Differentiation of deep measured by internal standard flame photometry. FFA were 15 and superficial venous beds was confirmed at the time of in- measured by a modification of the method of Novak. sertion by noting differences in oxygen saturation between Plasma insulin concentrations were determined by radioimcharcoal the beds and palpation of the catheter tips. The brachial ar- munoassay using Wright antibody lot no. 496 and a 13 tery was cannulated in the antecubital fossa with an 18-gauge separation technique we have previously described. Data handling. Data are expressed as mean ± standard thin-walled Cournand needle to which a double-lumen LiHenthal adaptor was fitted. This adaptor allowed the sam- error of the difference (SED) or standard error of the mean pling of incoming blood through a central cannula that ex- (SEM).16 A positive arteriovenous difference reflects net uptended beyond the needle tip and perfusion of insulin and take of a substance, while a negative difference reflects net dye through a side arm. Hand circulation was occluded with output from that venous compartment. A positive change a pediatric blood pressure cuff at the wrist 5 min before and with respect to baseline indicates decreased output and/or an during each sampling period to prevent venous admixture.14 increased uptake, and a negative change indicates increased The hand circulation was occluded throughout the insulin net output and/or decreased uptake. Arteriovenous differperfusion period. Arterial and venous specimens were ob- ences were averaged for the 30-min insulin perfusion period tained simultaneously. Venous catheters and the central ar- and for 30-60, 60-90, and 90-120 min after beginning the terial cannula were kept patent by a slow infusion of physio- hormone perfusion. Data were analyzed in terms of differlogic saline. An Evans Blue dye solution was perfused at 0.10 ence from mean baseline by repeated measures analysis of ml/min throughout. Blood flow was calculated using a modi- variance and paired t tests adjusted for multiple comparification of the Fick principle.12 At the end of each study, sons. 16,17 forearm volume was measured in a large glass cylinder by Increment in arterial concentrations of insulin were calcuwater displacement from the proximal antecubital crease to lated by the following: the upper edge of the wrist cuff. Average forearm volumes in groups I, II, and III were 980, 920, and 943 ml, respectively. Total /u,U insulin perfused/min Plasma flow in ml/min Experimental protocol. The protocol consisted of a 1-h baseline period, 30 min of insulin perfusion, and a 90-min recovery period. Samples were obtained at 15-min intervals Sequestration of insulin with forearm tissues during the perthroughout the baseline period, at 10-m:n intervals during fusion period was calculated by (calculated arterial increthe perfusion of insulin, and at 15-min intervals for the first ment in /u.U/ml — venous increment in /xU/ml) X plasma 60 min of recovery. A final sample was obtained 2 h after the flow X 30 min. Insulin returned to the circulation during the recovery pestart of insulin perfusion. During baseline and recovery periods only 150 /Ag/ml riod was calculated from the sum of areas under a plot of time Evans Blue dye, 2500 /xg/ml human serum albumin, and versus venous plasma insulin increment. The fraction of insuperficial (mostly fat and skin) and deep (mostly muscle) tissues. Venous and systemic arterial plasma insulin concentrations were measured directly as were concentrations of glucose, free fatty acids (FFA), and potassium during 30-min insulin perfusion and 90-min recovery periods. Increases in arterial and venous insulin concentrations during these experiments were comparable to those observed during insulin therapy or after feeding.
DIABETES CARE, VOL. 5 NO. 3, MAY-JUNE 1982
293
RELATIONSHIP BETWEEN INSULIN CONCENTRATIONS AND ACTION/S. E. FINEBERG AND S. H. SCHNEIDER
TABLE 1 Plasma flow rates and arterial plasma concentrations* of glucose potassium and free fatty acids (FFA)
sulin unaccounted for was: Insulin sequestered — insulin returned Total insulin perfused
Fbwt
In groups I, II, and III, venous plasma insulin concentrations fell at rates of 6.27 ± 0.15, 6.30 ± 0.21, and 6.30 ± 0.10%/min between the end of perfusion and the first sampling period 15 min later. During this period the decay rate for each study was used to calculate venous return. Following this sampling period, venous insulin increments could be measured only in group III for an additional 15 min. The rate of venous insulin decay determined from group III data during the latter period (3.4 ± 1.0%/min) was applied to the remainder of the recovery period for all groups to estimate the late phase of venous insulin return. This seemed justified based on the similarity in first phase decay rates for all groups. Calculation of compartmental concentrations of insulin was based on a published model for insulin kinetics.6 The assumptions used were (1) The density of forearm tissues was equivalent to water. (2) After 30 min of perfusion, insulin levels in the rapidly equilibrating compartment were equivalent to 100% of the arterial concentrations, and the interstitial fluid space levels were 70.5% of arterial. (3) The total insulin space was 15.7% of body wt (6.2% and 9.5%, respectively).
Group I (5) Baseline 10'-30' 3O'-6O' 60'-90' 90'-120' Group II (5) Baseline 10'-30' 30'-60' 60'-90' 90'-120' Group III (7) Baseline 10'-30' 3O'-6O' 60'-90' 90'-120'
Glucoset
Potassium^
FFAt
2.83 3.04 2.93 2.59 2.39
0.48 0.65 0.61 0.46 0.47
5.13 5.19 5.19 5.23 5.13
±0.26 ± 0.24 ± 0.24 ± 0.24 ±0.23
4.29 4.26 4.25 4.24 4.25
± ± ± ± ±
0.21 0.20 0.19 0.23 0.24
0.803 0.839 0.811 0.843 0.942
±0.158 ±0.159 ± 0.166 ±0.174 ±0.140
1.38 1.62 1.65 1.57 1.46
0.19 0.23 0.22 0.18 0.16
5.14 5.16 5.30 5.28 5.37
± 0.11 ± 0.18 ± 0.19 ± 0.17 ±0.18
4.03 4.01 4.02 4.01 4.01
± ± ± ± ±
0.07 0.04 0.05 0.07 0.08
0.564 0.624 0.680 0.678 0.651
± ± ± ± ±
0.054 0.045 0.064 0.081 0.118
2.93 2.93 2.83 2.98 2.45
0.38 0.36 0.35 0.41 0.21
5.18 5.10 5.20 5.30 5.15
± 0.20 ± 0.23 ± 0.23 ± 0.21 ±0.20
3.94 3.97 3.94 3.89 3.91
± 0.04 ± 0.03 ± 0.05 ±0.11 ± 0.11
0.912 0.852 0.838 0.929 0.927
± ± ± ± ±
0.062 0.089 0.104 0.104 0.098
Time period is indicated in the left-hand column. The number of experiments is indicated by parentheses. The number of pooled observations for each experiment is baseline:4, 1O'-3O':3, 3 0 ' - 6 0 ' : 3 , 6 0 ' - 9 0 ' : 3 , and 90'-120': 2, mean ± SEM are presented. t Flow is presented as ml/min/dl of forearm tissue. $ Concentrations are presented as /xmol/ml of plasma.
RESULTS
Plasma flow and arterial concentrations (Table I). Flow re-
mained stable thoughout all 17 experiments. Flow was lowest in the subjects of group II and about the same in the other groups. Arterial glucose, potassium, and FFA concentrations did not change significantly during either insulin perfusion or recovery. Basal arteriovenous differences of the deep (A-DV) and super-
ficial (A-SV) beds (Table 2). Mean arteriovenous differences of glucose, potassium, and FFA across forearm tissues during the control period are shown in Table 2. Net positive arteriovenous differences of glucose were found for all groups in both venous beds. No differentiation between beds with regard to glucose uptake was noted. Mean deep potassium ar-
TABLE 2 Baseline arteriovenous differences* in /nmol/ml
Group I (5) Glucose A-DV A-SV Deep vs. superficial Potassium A-DV A-SV Deep vs. superficial FFA A-DV A-SV Deep vs. superficial
Pt
Group II (5)
Pt
Group III (7)
Pt
0.42 ± 0.08 0.38 ± 0.05 0.05 ± 0.05
