The cardiovascular risk factor plasminogen activator inhibitor type 1 is related to insulin resistance

The Cardiovascular

Risk Factor Plasminogen Activator Is Related to Insulin Resistance

B.J. Potter van Loon, C. Kluft, J.K. Radder,

M.A.

Blankenstein,

Inhibitor

Type 1

and A.E. Meinders

The cardiovascular risk factor plasminogen activator inhibitor type 1 (PAI-1) has been associated with abdominal obesity, hypertension, hypertriglyceridemia, hyperinsulinemia, glucose intolerance, and type II diabetes, conditions known to be linked with insulin resistance. To determine whether PAI-1 is related to insulin resistance, we studied nine obese nondiabetics and 10 obese type II diabetics by means of a sequential hyperinsulinemic euglycemic clamp study. Plasma PAI-1 antigen (Ag) correlated significantly with peripheral insulin resistance, represented by the insulin level at which peripheral glucose uptake (PGU) is half-maximal ([ED5,,PGU] r = .87, P < ,001). Multiple regression analysis including indices of hepatic and peripheral insulin action, fasting plasma insulin levels, triglyceride levels, blood pressure (BP), waist to hip ratio (WHR), and body mass index (&MI) disclosed EDSOPGU to account for 76% of the variance of PAL1 Ag. We suggest that PAL1 contributes to the increased cardiovascular risk encountered with insulin resistance. Copyright iilm 1993 by W.B. Saunders Company

A

CONSTELLATION OF cardiovascular risk markers referred to as syndrome X1 has been identified, consisting of abdominal obesity, hypertension, elevated fasting plasma levels of very-low-density lipoprotein triglyccridcs, reduced levels of high-density lipoprotein cholesterol. hyperinsulincmia, glucose intolerance, and type II diabetes mellitus. The common denominator of this syndromc is insulin resistance.’ Plasminogcn activator inhibitor type 1 (PAI-1) is a potent inhibitor of fibrinolysis by binding to and rapidly inactivating both tissue-type plasminogen activator and urokinasetype plasminogen activator.? Increased levels of PAI- and decreased plasma fibrinolytic activity have been demonstrated Im survivors of acute myocardial infarction”-’ and in patients with coronary artery stenosis.h-x Furthermore, PAIis of prognostic value in predicting recurrence of myocardial infarction.” A positive association between fasting plasma insulin levels and PAI- has been demonstrated by several investigators.“‘-” In addition, PAI- has been linked to separate feature:, of syndrome X including abdominal obesity,‘” hypcrtc nsion, 14-lhhypertriglyceridemia,3~7~‘z~‘4~‘6 glucose intolerance, and type II diabetes melIitus.‘l~lJ The associations of PAX-1 with features of syndrome X suggest a link between PAI-I and insulin resistance. To elucidate whether such a relationship exists, we determined both plasma levels of PAIL1 antigen (Ag) and of insulin and directly measured insulin action by means of the hypcrinsulinemic euglycemic glucose clamp technique. PATIENTS

AND METHODS

Nine obese nondiabetic (group I) and 10 obese type 11 diabetic (group II) subjects participated in the study after giving informed consent. The protocol had been approved by the Ethical Committee of the University Hospital Leiden. Clinical characteristics are presented in Table 1. Nondiabetic subjects had no family history of diabetes and a normal glucose tolerance test.” Diabetic subjects were treated with diet only: none had clinical macrovascular disease. proteinuria, or (pre)proliferative retinopathy. All subjects had a systolic blood pressure (BP) of 160 mm Hg or less and .a diastolic BP of 95 mm Hg or less. and none were treated with antihypertensive medication. BP was measured in with subjects in the sitting position after 5 minutes of rest: Korotkoff phase V was recorded as diastolic BP. Body mass index (BMI) was calculated as weight divided by height squared (kg/m?); waist to hip

Metabo//:;m, Vol 42, No 8 (August), 1993: pp 945-949

ratio (WHR) was calculated kiewski et al.‘”

according

to the method

of Krot-

C’ltrrnp Studies Insulin action was measured using the sequential hyperinsulinemit euglycemic glucose clamp technique’” in combination with a 3-?H-glucose labeled tracer infusion.:” Subjects with diabetes were rendered normoglycemic with an overnight variable insulin infusion before the clamp study. Glucose was clamped at S mmol/L by infusion of 20% dextrose at a variable rate. which was adjusted every 5 minutes to the corresponding plasma glucose level (Glucose Analyzer. Beckman Instruments. Palo Alto, CA). Insulin (Humuline Regular, Eli Lilly, Indianapolis. IN) was infused during four consecutive Z-hour periods at the following rates: basal. 0.75. 1.5. and IO mU/kg/min following bolus injections of 0. 4,s. and 100 mu/kg, respectively, at the start of each Z-hour period. Basal equals zero in nondiabetics and 0.31 * 0.06 mU/kg/min in subjects with diabetes. ie. the dose necessary to obtain a fasting plasma glucose level of 5 mmol/Lwithout the need for concurrent dextrose infusion. EDTA-anticoagulated plasma was examined half-hourly and during the last 30 minutes of each l-hour period at IO-minute intervals for determination of the momentary insulin level and plasma 3-iH-glucose specific activity.?” Peripheral glucose uptake (PGU) was calculated as either the quotient of 3-jH-glucose infusion rate and plasma 3--‘H-glucose specific activity or as the rate of infusion of exogenous glucose, whichever was highest. Hepatic glucose production (HGP) was calculated as the difference between PGU and exogenous infusion rate. A dose-response curve was constructed relating the momentary insulin level to PGU. The dose-response curve can be characterized by the maximal PGU (%‘,,,PGU). representing peripheral insulin responsiveness, and by the insulin level at which PGU is half-maximal (EDx,,PGU). representing peripheral insulin sensitivity.” In the case of a decrease in insulin sensitivity, ie, insulin resistance, the curve is shifted to the right. which is represented by an increase in

From the Depatiments of Elldocrinolo~lMetaholism and Internal Medicine, lJnivrrsi& Hospital. Leiden; the Gaubius Laborator?: IVVO-TNO, Leiden: and the Lahorato? for Endocrinology, Uniwrsir\ Hospital, Utrecht, The Netherlands. Submitted March 23, 1992; accepted October 14. 1992 Address reprint requests to B.J. Potter \‘an Loon. MD, PhD, Sint Lucas Ziekmhuis, Department of Internal Medicine, Jan Tooropstraat 164, 1061 AE Amsterdam, The Netherlands. Copyright 10 1993 by W.B. Saunders Company OO26-049.5193/4208-0005$03.0010 945

PO-ER

946

Table 2. Results of Metabolic Studies and Plasma Levels of PAI-1,

Table 1. Clinical Characteristics of Participating Subjects Nondiabetic Age

47 5 2

W)

Sex (M/F)

Insulin, Proinsulin, and Triglycerides in Obese Nondiabetic (group I)

Diabetic

P

52 + 2

NS

514

614

and Type II Diabetic Subjects (group II)

Range

NS

WHR

0.96 + 0.02

0.98 + 0.02

NS

ED,,PGU (mu/L)

BMI (kg/mz)

32.6 t 1.1 138181

33.8 2 1.0

NS

i,,,PGU

144184

NS

Basal HGP (bmol

FBG (mmol/L)

4.7 -t 0.2

9.4 +- 1.4

-. .OOl

ED,,HGP (mU/L)t

HBA,, (%)

5.3 * 0.1

7.7 * 0.4

< .OOl

Insulin (mu/L)

Triglycerides (mmol/L)

2.5 + 0.84

3.5 5 1.1

NS

BP (mean, mm Hg)

VAN LOON ET AL

Abbreviations: FBG,fasting blood glucose; HbA,,, glyosylated hemoglobin (reference values, 4.4% to 6.7%).

(pmol

kg

’

min-I)*

kg~ 1 min-I)

Group

21

II

25.8-81.6

59.2

IT 3.5

42.3 ? 3.6

7.8-12.2

9.2

z? 0.4

10.3 2 0.4

14.114

28

14 *

107.607

160

t

0.9-l 1.8

2.1

?I 0.5

34

66i

2

Proinsulin (ng/L)

15-91

206 t 34

2 4

Triglycerides (mmol/L) PAI- (ng/mL)*

Group

127

7-44

t

1

62-394

11

11

22 + 3 354 ? 54 3.3 + 1.0

2 8

44 + 7

*Group I v group II, P < .Ol.

EDsoPGU. the

Similarly.

insulin

(ED&GP).

level ?I

hepatic

at which

insulin sensitivity is represented

HGP

is half-maximally

hy

suppressed

PAI- 1Ag, Insulin, Proinsuh. and Triglyerides

PAI- I Ag PAI-

Ag follows

tGroup I v group II, P i .05.

a diurnal

morning hours.2z.Z3 Therefore

rhythm

with

a peak

blood was collected

during

at X:30

AM

the just

the clamp study in 3.8% sodium citrate. To validate the in diabetics on a day without prior overnight insulin infusion, blood was collected on a control day within h weeks of the clamp at 8:30 AM for determination of PAI-I Ag. before

results

The collected blood was centrifuged immediately 20 minutes at 4°C to obtain platelet-free

at 2.000 x g for

plasma for determination

of PAI-

Ag (TintElize lot number 210220. Biopool, Umela. To ensure that platelet activation was absent in the aliquots, P-thromboglobulin level was measured: values were 15 to 148 ng/mL (mean, 81 ng/mL).14 All assays were performed in duplicate; the interassay coefficient of variation for PAI- I Ag was 8% and the intraassay coefficient 8%. Sweden).

Fasting Plasma Insulin, Proinsulin, and Trig!vcerides Because diabetic subjects had received insulin during the night before the clamp, plasma levels of insulin, proinsulin, and triglycerides in diabetics were suppressed on the day of the clamp. Serum and plasma therefore were collected in both diabetics and nondiabetics after an overnight fast on two different occasions within h weeks of the clamp study for determination of fasting plasma immunoreactive insulin,‘c fasting plasma proinsulin,‘h and triglyceride levels.27 The mean of each pair of values obtained in every

Values of PAI- Ag on the day of the clamp at 8:30 AM arc reported in Table 2. PAI- Ag level on the control day in diabetic subjects was 49 ‘_ 7 ng/mL. Fasting plasma Icvcls of insulin, proinsulin, and triglycerides are reported in Table 2.

A significant correlation was found between PAI- I Ag on the clamp day and ED5,,PGU in both group I (r = .9X. P < .OOl) and group II (r = .X7, P < .Ol). Further correlation studies were performed on the combined groups after excluding significant differences in regression lines of the individual groups. Figure 1 dcmonstrates the correlation between PAI- Ag and EDS,,PGU (r = .X7, P < ,001). In addition, PAI- Ag correlated with V,,,PGU (r = -.51, P < .05) and with EDX,,HGP (r = .67. P < .Ol). As expected, PAI- Ag correlated with fasting plasma insulin levels (r = .6S, P < .Ol). Furthermore, the previously described correlations between PAT-I on the one hand and BMI, WHR, triglyccridcs, and systolic and diastolic BP on the other wcrc not significant in our small number of subjects.

person was taken.

Statistics Results Statistical regression distribution, (two-tailed

are expressed as means + standard error of the mean. analysis was performed by bivariate and multivariate analysis and by t testing. In cases of nonnormal group nonparametric tests were used. A probability level test) below .OS was considered significant. RESULTS

Clamp Studies glucose

I

01

mmol/L (coefficient of variation, 3.3%) during the clamp. Insulin levels were 16 2 3 and 38 k 8 mU/L during the first 2-hour period in groups I and II, respectively. Subsequent periods yielded insulin levels of 96 k 5, 199 f 11, and 4,548 t 226 mU/L and. were similar in both groups. Results of EDSOPGU, V,,,PGU, basal HGP, and EDS,,HGP assays are reported in Table 2. Plasma

level

averaged

5.06

?

0.04

50

loo

150

200 ED,pgu

250

300

350

400

WJ/lI

Fig 1. Relation between peripheral insulin sensitivity and plasma PAI- Ag. Plasma levels of PAI- Ag correlate strongly (r = .87, P < .OOl) with the insulin level at which PGU is half-maximal (ED,PGU), which represents the degree of resistance to the action of insulin on peripheral tissues. In both nondiabetic (0; r = .98. P < ,001) and type II diabetic (*) subjects. (r = .87,P < .Ol), this correlation is significant.

PAI-

947

REL.ATION TO INSULIN RESISTANCE

PAI-I Ag on the day of the clamp was not significantly related to fastingplasma proinsulin levels (r = 37). Multiple regression analysis including the parameters of insulin action (EDSOPGU. V,,,PGU, basal HGP, EDSo HGP) and the factors known to be associated with PAI1 “‘~‘?-ih(fasting plasma insulin, BMI. WHR, triglycerides, systolic and diastolic BP) identified ED,,PGU, V,,,,,PGU, diastolic BP, BMI, and WHR to be the only variables significantly related to PAIAg (F = 24.4, P < .OOl), explaining 90% of the variance of PAIAg (Table 3). Seventy-six percent of the variance of PAI- Ag could be explained by EDSoPGU, with each of the remaining four parameters accounting for 1% to 9% of the variance. Fasting plasma insulin level was not independently associated with PAI- Ag. Substitution of the diabetics’ PAIAg values on the clamp day with values obtained on the control day without prior insulin infusion essentially resulted in a similar outcome of the regression analysis (F = 12.4, P < .OOl); EDs,lPGU accounted for 54% of the variance, V,,,PGU for 4%, and1 BMI for 13%, whereas fasting plasma insulin level was not independently related to PAI- Ag. DISCUSSION

In this study, we have shown a strong correlation between peripheral insulin resistance, as expressed by EDsOPGU derived from euglycemic clamp data, and morning plasma levels of PAIAg (Fig 1). The correlation between EDSOPGU and levels of PAI- Ag appeared to be dominant in a multivariate regression analysis including parameters of insulin action and parameters known to be associated with PAI-I. To validate the relationship between insulin resistance and PAIAg, we also determined PAIAg levels in diabetics on a control day on which they had not received prior overnight insulin. Substitution of the original value of PAI- Ag in the multivariate regression analysis with this alternative confirmed the results. Our results seem to contrast with previously published data, which emphasize the relationship between levels of PAI- and insulin.iri-r~JsJh N evertheless, like these investigators, we also found a significant correlation between fasting plasma insulin levels and PAI-1. However, as we directly measured peripheral and hepatic insulin action and included the results of these measurements in the multiple regressilon analysis, the relationship between fasting plasma Table 3. Multivariate

Regression Analysis of PAL1 Ag

in Nondiabetics

and Type II Diabetics

p ? SE

f Value

SequentialRz

EDS,PGU

.34 + .04

8.9

76%

c,,,PGU

.65 + 22

2.9

79%

Diastolic BP BMI WHR

-.80

+ .32

1.72 k .67 -145

+ 42

-2.5

80%

2.6

81%

-3.5

90%

NOTE. llndependent variables included in the regression model were EDSOPGU, i,,.PGU,

HGP, EDS,HGP, fasting plasma insulin level, BMI,

WHR, fasting plasma triglyceride levels, and systolic and diastolic BP. p denotes the regression coefficient. adjusted R2 = 0.87.

F = 24.4; P

<

,001;

R2

=

.90;

insulin levels and PAI- Ag appeared to be secondary to the association of PAI- Ag and peripheral insulin resistance. Apparently, the link is not between insulin and PAI-1, but between peripheral insulin resistance and PAI-1. To our knowledge, only Landin et alijJh have directly assessed the relation between PAI- and insulin action by clamp studies. Their results are in line with ours, demonstrating an inverse relationship between PAI-I activity and glucose disposal in normotensive and hypertensive meni and in obese and lean women.‘” These investigators performed clamp studies with a single insulin level and measured PAI- activity rather than PAI- Ag levels. The association between glucose disposal and PAIin their studies seems weaker than the associations between PAIand insulin; moreover, they did not perform a multivariate analysis. This, in combination with differences in techniques, may explain the apparent contradictions between their studies and ours. It therefore appears that PAI- is associated with insulin resistance, hypertension, 14.16hypertrig]yceridemia,3,7.x,'2.14.'6 abdominal obesity, I3 hyperinsulinemia,i’1-i3~rs~i6 and type II diabetes mellitus.iiJ4~zx As all the above features are corelated and all are associated with insulin resistance, this suggests that high PAI- Ag levels should be included in the syndrome of insulin resistance. However, caution is warranted, since we have only studied obese middle-aged normotensive nondiabetic and type II diabetic subjects. Plasma fibrinolytic activity is dependent on the balance between activators and inhibitors. The principal activator, tissue-type plasminogen activator (t-PA), is normally bound to an excess of PAI-1, the principal inhibitor. Therefore plasma t-PA activity correlates more strongly (inversely) with plasma levels of PAI-12Y,3n than with levels of t-PA Ag in plasma. High levels of PAI- Ag, by inhibiting t-PA and thereby inhibiting fibrinolysis, may contribute to the increased incidence of cardiovascular disease in subjects with syndrome X. It is not clear what the mediator is between insulin resistance and PAI-1; it has been suggested that insulin may be such a mediator.“’ In vitro, both insulin and proinsulin promote the secretion of PAI- by cultured hepatocytes,32x33 but experimental hyperinsulinemia does not elicit an increase in PAI- activity in humans.34,3s Recent findings of increased fasting plasma levels of proinsulin in insulinresistant states”6-3s as well as the occurrence of cardiovascular disease following proinsulin administration3” have prompted us to investigate whether fasting plasma proinsulin levels are more closely associated with PAIthan is insulin sensitivity. However, the correlation between proinsulin and PAI- Ag appeared to be insignificant. In addition, had proinsulin been included in the multiple regression analysis (data not shown), it would have decreased PAIAg. Therefore it is unlikely that proinsulin is a mediator between insulin resistance and elevated PAI- Ag levels. In conclusion, we have demonstrated that plasma levels of PAI- Ag are related to peripheral insulin resistance in middle-aged obese nondiabetic and obese type II diabetic subjects. The mechanisms of this relationship require

POTTER VAN LOON ET AL

948

further elucidation. We suggest that the increased cardiovascular risk in insulin-resistant conditions may be mediated in part by impaired fibrinolysis secondary to elevated levels of PAI-I.

ACKNOWLEDGMENT The authors wish to thank A.C.W. de Bart for technical assistance and A.H. Zwinderman advice.

and P. Meyer for statistical

REFERENCES I. Reaven GM: Role of insulin Diabetes37:lSYS1607. lY88 2. Sprengers ED, Kluft Blood 69:3X1-387. 1987

resistance

C: Plasminogen

in human activator

disease. inhibitors.

3. Hamsten A, Wiman B, de Faire U. et al: Increased plasma levels of a rapid inhibitor of tissue plasminogen activator in young survivors of myocardial infarction. N Engl J Med 313:1557-1563. 1985 4. Nilsson TK, Johnsson 0: The extrinsic tibrinolytic system in survivors of myocardial infarction. Thromb Res 4X:621-630. lYX7 5. Almer LO. Ohlin H: Elevated levels of the rapid inhibitor of plasminogen activator (t-PAI) in acute myocardial infarction. Thromb Res 37:33S-329. 19X7 6. Francis RB. Kawanishi D. Baruch T. et al: Impaired librinolysis in coronary artery disease. Am Heart J 115:776-7130, 1988 7. Aznar .I, Estelles A. Tormo G. et al: Plasminogen activator inhibitor activity and other fibrinolytic variables in patients with coronary artery disease. Br Heart J 59:535-54 1, lY88 8. Olofsson BO, Dahlen G. Nilsson TK: Evidence for increased levels of plasminogen activator inhibitor and tissue plasminogen activator in plasma of patients with angiographically verified coronary artery disease. Eur Heart J 10:77-X2. IYXY 9. Hamsten A, de Faire U. Walldius G. et al: Plasminogen activator inhibitor in plasma: Risk factor for recurrent myocardial infarction. Lancet 2:3-9, 19X7 IO. Vague P. Juhan-Vague I, Aillaud MF. et al: Correlation between blood fibrinolytic activity. plasminogen activator inhibitor level. plasma insulin level, and relative body weight in normal and obese subjects. Metabolism 35:250-253. 19X6 Il. Juhan-Vague I. Roul C. Alessi MC, et al: Increased ogen activator inhibitor activity in non insulin dependent patients-Relationship with plasma insulin. Thromb 61:370-373. 19X9

plasmindiabetic Haemost

12. Juhan-Vague I, Vague P. Alessi MC. et al: Relationship between plasma insulin. triglyceride, body mass index and plasminogen activator inhibitor 1. Diabete Metab 13:331-336, 1987 13. Landin K. Stigendahl L. Eriksson E. et al: Abdominal obesity is associated with an impaired fibrinolytic activity and elevated plasminogen activator inhibitor-l. Metaholism 39:1044104X, 1990 14. Heinrich J, Nguyen T, Kokott R, et al: Interrelationships between coronary risk factors and the tibrinolytic system. Fibrinolysis 4:37. 1990 (suppl I. abstr) 15. Auwerx J, Bouillon R. Cohen D, et al: Tissue-type plasminogen activator antigen and plasminogen activator inhibitor in diabetes mellitus. Arteriosclerosis X68-72, 1988 16. Landin K, Tengborn L, Smith U: Elevated fibrinogen and plasminogen activator inhibitor (PAI-1) in hypertension are related to metabolic risk factors for cardiovascular disease. J Int Med ‘27:273-27X, 1990 17. National Diabetes Data Group: Classitication and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1039-1057. 1979 18. Krotkiewski M. Bjorntorp P, Sjdstrom L, et al: Impact of obesity on metabolism in men and women. Importance of regional adipose tissue distribution. J Clin Invest 72:1150-l 162. 1983 19. DeFronzo

RA. Tobin JD, Andres

R: Glucose

clamp

tech-

nique: A method for quantifying insulin secretion and resistance. Am J Physiol237:E214-E223. 1979 20. Nanketvis A. Proietto J. Aitken P. et al: Dilferential effects of insulin therapy on hepatic and peripheral insulin sensitivity in type 11 (Non-insulin-dependent) diabetes. Diahrtologia 23:32032s. 19x7 21. Kahn CR: Insulin resistance, insulin insensitivity, and insulin unresponsiveness: A necessary distinction. Metabolism 27: 180% 1902. 197X (suppl 2) 77. Kluft C. Jie AFH, Rijken DC. et al: Daytime fluctuations in blood of tissue-type plasminogen activator (t-PA) and its fastacting inhibitor (PAI-I). Thromb Haemost S9:32Y-332, 19X9 23. Andreotti F. Davies GJ, Hackett D, et al: Circadian variation of tibrinolytic factors in normal human plasma. Fihrinolysis 2:YO-02. 19XX(suppl 1) 23. Van Hulsteijn LH: Clinical applications of the alrsav ot plasma Iibrinopeptide A and beta-thromboglobulin. Doctoralthesis. University of Leiden. The Netherlands. 19X2 75. Kuzuya H. Blix PM, Horwitz DL. et al: Determination elf free and total insulin and C-peptide in insulin-treated diabetics. Diabetes %:22-29, 1976 26. Huisman J. Oolbekkink M. Erkelens DW. ct al: Radioimmunoassay for human proinsulin. in Huisman J: Proinsulin: Influence of exogenous administration of rDNA proinsulin on carbohydrate and lipid metabolism. Doctoral thesis, University of Utrecht. The Netherlands. 1990. pp 47-65 77. Rantela GS. Hall RG, Bekie SZ, et al: A kinetic method fat the rapid and automatic determination of triglycerides in biological fluids. Clin Chem 20:476-4X1. 1976 2X. Small M. Kluft C. MacCuish AC. et al: Tissue plasminogrn activator inhibition in diabetes mellitus. Diabetes Care 1?:65S-65X. 19x9 29. Rijken DC. Sprengers ED. v’an der Hoeven RJ. et al: Plasminogen activator inhibitor 1 (PAI-I) antigen levels in human samples. Fihrinolysis 2:79-X2, 198X (suppl 2) 30. Gram J, Jespersen J, Kluft C. et al: On the usefulness of tihrinolysis variables in the characterization of a risk group for myocardial infarction. Acta Med Stand 221: 149.lS3. 1987 31. Juhan-Vague 1. Alessi MC. Vague P: Increased plasma plasminogen activator inhibitor 1 levels. A possible link hetwern insulin resistance and atherothrombosis. Diahetologia 34:JS7-462. 1991 37. Alessi MC. Juhan-Vague I. Kooistra T. et al: Insulin stimulates the synthesis of plasminogen activator inhibitor 1 hy the human hepatocellular cell line Hep G?. Thromb Haemost 60:491494, 1988 33. Kooistra T. Bosma PJ. Tons HAM. et al: Plasminogen activator inhibitor 1: Biosynthesis and mRNA level are increased by insulin in cultured human hepatocytes. Thromh Haemost 62:723-72X. 19x9 i-l. Potter van Loon BJ. de Bart ACW. Radder JK. et al: Acute exogenous hyper-insulinaemia does not result in elevation of plasma plasminogen activator inhibitor-t (PAI- ) in humans. Fibrinolysis 4:93-94, 1990 (suppl 2) 35. Landin K, Tenghorn L. Chmielewska J. et al: The acute effect of insulin on tissue plasminogen activator and plasminogen activator inhibitor in man. Thromb Haemost 65: 130-133. 1991 36. Temple RC. Carrington CA, Luzio SD. et al: Insulin

PAI-

949

REILATION TO INSULIN RESISTANCE

deficiency

in non-insulin-dependent

diabetes.

Lancet

1:293-295.

198’) 37. Porte D, Kahn SE: Hyperproinsulinaemia and amyioid in NIDDM. Clues to etiology of islet beta-cell dysfunction. Diabetes 38: 1333- 1336. 1989 38. Madsbad S, Tronier B. Dejgaard A. et al: Plasma proinsulin

in obesity and type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia 33:A212, 1990 (suppl. abstr) 39. Spradlin CT, Galloway JA, Anderson JH: Apparent increase in coronary heart disease with human proinsulin (HPI)Randomization failure or HP1 effect? Diabetologia 33:AhO. 1990 (suppl. abstr)