The Egyptian Journal of Internal Medicine

ORIGINAL ARTICLE
Year
: 2015  |  Volume : 27  |  Issue : 3  |  Page : 108--114

Urinary albumin excretion and progression of renal disease with impaired fibrinolytic activity in type 2 diabetes mellitus


Ebtessam Zakaria1, Maha Hossam Al-Din1, Nashwa S Ghanem1, Noha A Sadik1, Maha Assem1, Fatma Taha2,  
1 Department of Internal Medicine, Faculty of Medicine, Cairo University, Cairo, Egypt
2 Department of Clinical Biochemistry, Faculty of Medicine, Cairo University, Cairo, Egypt

Correspondence Address:
Noha A Sadik
7 Al Mahrossa Street, Ahmed Orabi, Al Mohandeseen, App 3, 694 Giza
Egypt

Abstract

Background Diabetic nephropathy is one of the major causes of end-stage renal disease. As impaired fibrinolysis can increase renal fibrosis, we investigated the relationship of impaired fibrinolysis, as assessed by the ratio of plasminogen activator inhibitor-1 (PAI-1) to tissue-type plasminogen activator (t-PA) with urinary albumin excretion (UAE) and renal disease progression in type 2 diabetes. Patients and methods A total of 50 patients with type 2 diabetes and 10 healthy control individuals were included in the study. Participants were subdivided according to UAE. Group 1 (G1) represented control individuals. Group 2 (G2) included patients without albuminuria. Group 3 (G3) included patients with microalbuminuria. Group 4 (G4) included patients with macroalbuminuria. Creatinine clearance and UAE were calculated. PAI-1 and t-PA were measured using an enzyme-linked immunosorbent assay kit, and the PAI-1/t-PA ratio was calculated as an index of impaired fibrinolysis. Results PAI-1 was highly elevated significantly in G4 when compared with the other groups, with a mean of 87.40 ± 17.03 IU/ml against 28.00 ± 6.98 IU/ml in G1, 46.4 ± 7.99 IU/ml in G2, and 64.10 ± 18.26 IU/ml in G3 (P < 0.001). Also, the serum level of t-PA in G4 was highly elevated significantly when compared with G1 and G2 with means of 16.85 ± 5.63 IU/ml against 7.95 ± 1.91 IU/ml and 10.45 ± 2.63 IU/ml, respectively (P < 0.001). The ratio of PAI-1/t-PA in G4 was significantly higher when compared with G1 (mean of 5.94 ± 2.81 against 3.54 ± 0.43; P = 0.01). PAI-1 and t-PA showed a significant positive correlation with UAE. Receiver operating characteristics curve analysis revealed that only PAI-1 and t-PA were significant discriminated factors for microalbuminuria and macroalbuminuria (P < 0.001). Conclusion Serum levels of PAI-1 and t-PA and the PAI-1/t-PA ratio were significantly increased in diabetic patients with higher UAE. Impaired fibrinolysis and increased UAE were associated with renal disease progression.



How to cite this article:
Zakaria E, Al-Din MH, Ghanem NS, Sadik NA, Assem M, Taha F. Urinary albumin excretion and progression of renal disease with impaired fibrinolytic activity in type 2 diabetes mellitus.Egypt J Intern Med 2015;27:108-114


How to cite this URL:
Zakaria E, Al-Din MH, Ghanem NS, Sadik NA, Assem M, Taha F. Urinary albumin excretion and progression of renal disease with impaired fibrinolytic activity in type 2 diabetes mellitus. Egypt J Intern Med [serial online] 2015 [cited 2024 Mar 28 ];27:108-114
Available from: http://www.esim.eg.net/text.asp?2015/27/3/108/165450


Full Text

 Introduction



It is well known that vascular injury and cardiovascular diseases are the main causes of death in diabetic patients [1] and diabetes is one of the most important causes of end-stage renal disease [2].

Fibrosis is the hallmark for the development of end-stage kidney disease. This process occurs as a result of abnormal deposition of extracellular matrix (ECM) in the basement membranes and interstitial tissues, associated with necrosis of the normal parenchymal cells and or apoptosis [3]. Replacement and degradation of the normal ECM by abnormal ECM is due to an increase in the proteolytic activity that is represented in the plasminogen activation system and matrix metalloproteinases [4].

Plasminogen activator inhibitor-1 (PAI-1) is a single-chain glycoprotein [5], the main inhibitor of the plasminogen activation system. It regulates intravascular fibrin deposition through the inhibition of the serine proteases tissue-type plasminogen activator (t-PA) and urokinase type PA and plasmin-mediated matrix metalloproteinase activation [6]. Also, PAI-1 is a component of ECM that binds to vitronectin (also known as protein S) [3]. This binding stabilizes PAI-1 in its active form and leads to the continuous inhibition of plasmin formation and matrix metalloproteinase activation [7]. In addition, PAI-1 has the ability to promote fibrosis through its effect on inflammatory cell recruitment, matrix turnover, fibroblast activation, and deposition [8].

In the nondiseased kidney, PAI-1 is almost absent; increased expression was found in many chronic kidney diseases leading to renal fibrosis and end-stage kidney failure [3]. Tumor growth factor-β and the renin-angiotensin-aldosterone system are the main inducers of PAI-1 expression, and their inhibition decreases PAI-1 expression in the glomerulus [9,10].

The resultant fibrosis and vascular injury that occurs in the kidney may be one of the most important factors that lead to glomerular capillary leakage of proteins and resultant albuminuria [11]. The increase in urinary albumin excretion (UAE) reflects the vascular damage all over the body [12], which may lead to the progression of diabetic nephropathy in type 2 diabetic patients [13].

The aim from our study was to detect the relation of impaired fibrinolysis, as assessed by the ratio of the PAI-1 to the t-PA to UAE and the progression of renal disease in type 2 diabetes.

 Patients and methods



Participants

A total of 50 patients with type 2 diabetes who attended the outpatient clinic of the Department of Internal Medicine, Kasr Al Aini Hospital, Cairo University, and 10 healthy age-matched control individuals were included in the study. All diabetic patients met the diagnostic criteria of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus [14]. The study was performed from December 2013 to August 2014. Participants were subdivided into four groups with regard to UAE. Group 1 represented the control group. Group 2 included patients with type 2 diabetes without albuminuria. Group 3 included patients with type 2 diabetes with microalbuminuria. Group 4 included patients with type 2 diabetes with macroalbuminuria.

All patients had diabetes of type 2 of 8-10 years' duration.

Exclusion criteria

Patients with end-stage renal failure, congestive heart failure, cerebral infarction, cerebral hemorrhage, and liver diseases were excluded.

Ethical aspects

Research protocols were approved by the medical ethics committee of Kasr al Aini Medical School, Cairo University. All participants provided a written informed consent after the research protocols were explained carefully to them. Informed consent was obtained from all the study participants and their approval was taken by signature.

Procedures

All participants underwent a complete screening panel, including history taking and physical examination. Laboratory investigation included the following: fasting blood sugar, 2-h postprandial blood sugar, hemoglobin A 1C (HbA1c), total cholesterol, triglycerides, and serum creatinine. All laboratory investigations were performed by an autoanalyzer on Roche Diagnostic Mannheim GmbH (Germany). Creatinine clearance and UAE were calculated. PAI-1 and t-PA were measured using the enzyme-linked immunosorbent assay kit, and the PAI-1/t-PA ratio was calculated as an index of impaired fibrinolysis.

Creatinine clearance was calculated using the Cockroft-Gault equation in mg/dl [15].

(140-age) [Wt (kg)]/72 (serum creatinine in mg/dl). If the patient is female, multiply the above by 0.85.

UAE was calculated as 24-h urine collection. No albuminuria was considered if the albumin excretion in the urine was less than 30 mg/24 h. Microalbuminuria was considered if albumin excretion in the urine was in the range between 30 and 300 mg/24 h. Macroalbuminuria was considered if the albumin excretion was more than 300 mg/24 h.

PAI-1 (IU/ml) and t-PA (IU/ml) kits were supplied from Hyphen BioMed (USA). The complete enzyme-linked immunosorbent assay kit was used for tissue-PAI-1 and for t-PA antigen.

Specimen collection

Blood was collected on EDTA through a clean venipuncture avoiding any blood activation; it was then decanted after a 20-min centrifugation at 2500 g to avoid platelet activation, stored frozen at -20°C or below for up to 6 months, and thawed for 15 min at 37°C just before use. To avoid diurnal variation, PAI-1 and t-PA were measured on fasting samples collected at morning.

The Zymutest PAI-1 antigen kit is a one-step, two-site immunoassay for measuring human tissue-PAI-1 in the plasma. The measurement of PAI-1 was based on the method described by Declerck et al. [16]. The Zymutest t-PA kit is a two-site immunoassay for measuring human t-PA in the plasma. The measurement of t-PA was based on the method described by Juhan-Vague et al. [17].

Statistical analysis

Data was analyzed through statistical package of social science software program, version 18 (SPSS).Statistical Package for the Social Science; SPSS Inc., Chicago, Illinois, USA.

Data were summarized using mean and SD for quantitative variables and frequency and percentage for qualitative variables.

A comparison between groups was performed using the one-way analysis of variance test for quantitative variables with the post-hoc Tukey's test for pairwise comparisons and the χ2 -test for qualitative variables. The Spearman correlation coefficient was calculated to test the association between quantitative variables. Receiver operating characteristics (ROC) analysis was used to test the discriminatory ability of certain measures to predict microalbuminuria and macroalbuminuria and to derive the most suitable cutoff point with considerable sensitivity and specificity.

P values equal to or less than 0.05 were considered statistically significant.

 Results



Our patients have type 2 diabetes of 8-10 years' duration. The demographic and laboratory data of our diabetic patients are shown in [Table 1]. Diabetic patients with a higher UAE had significantly higher serum creatinine and lower creatinine clearance levels. The serum creatinine level and the 24-h UAE were significantly higher in patients with macroalbuminuria compared with the other groups [Table 1].{Table 1}

In [Table 2], PAI-1 was statistically significantly different between all groups (P < 0.001), and with the post-hoc Tukey's test for pairwise comparisons, it was significantly higher in diabetic patients in comparison with the control group and significantly higher in diabetic patients with macroalbuminuria in comparison with the other groups. Patients with macroalbuminuria showed significant elevation in the serum level of t-PA when compared with individuals without albuminuria (P < 0.001) and it was higher but of no statistical significance when compared with diabetic patients with microalbuminuria. The ratio of PAI-1/t-PA was statistically significantly higher in diabetic patients, with a higher UAE when compared with the control group (P = 0.01).{Table 2}

PAI-1 was strongly correlated with serum creatinine and UAE (P < 0.001) and negatively correlated with creatinine clearance (P = 0.001); t-PA also showed a significantly moderate positive correlation with UAE [Table 3] and [Figure 1],[Figure 2] and [Figure 3].{Figure 1}{Figure 2}{Figure 3}

ROC curve analysis revealed that only PAI-1 and t-PA were significant discriminators for macroalbuminuria and microalbuminuria (P < 0.001) ([Table 4] and [Table 5], [Figure 4] and [Figure 5].{Figure 4}{Figure 5}{Table 3}{Table 4}{Table 5}

 Discussion



Our study showed that the serum level of PAI-1 was statistically significantly higher in diabetic patient subgroups compared with the control group and there was a significant increase in the serum level of PAI-1 in the diabetic patient subgroup in comparison with each other according to the increased UAE, that is, the four groups were statistically significantly different from each other. Also, we found that the cutoff point of 73.5 IU/ml was the level of PAI-1 to discriminate macroalbuminuria, and the cutoff point of 43.0 IU/ml was the level of PAI-1 to discriminate microalbuminuria using ROC curve analysis.

These findings were supported by Chudy P et al. [18], who found that PAI-1 levels were elevated in both diabetic patients with normoalbuminuria and diabetic patients with macroalbuminuria in comparison with the control group. Collins et al. [19] found that an elevated plasma level of PAI-1 mediates diabetic vascular complications and suggested that diabetic nephropathy was the major implication of high PAI-1 levels. Also, Szelag et al. [20] reported that the elevated PAI-1 was associated with diabetic complications, such as microalbuminuria, neuropathy, retinopathy, and lower-extremity arterial diseases, suggesting that circulating PAI-1 may correlate with increased vascular injury.

Bastard et al. [21] had shown that the plasma level of PAI-1 increased in diabetics. Also, Yener et al. [22] observed that the PAI-1 level was significantly elevated in 61 normotensive, normoalbuminuric type 2 diabetic patients without diabetes-related complications. Festa et al. [23] showed that the progression of PAI-1 over time, in addition to high baseline PAI-1 levels, was associated with incident diabetes.

However, Nagi et al. [24] found that the PAI-1 activity was similar between nondiabetic and diabetic patients.

Our study showed a significant increase in the serum level of t-PA in diabetic patients with albuminuria (either microalbuminuria or macroalbuminuria) and that t-PA was a significant discriminator for macroalbuminuria and microalbuminuria and the minimum values of t-PA to discriminate microalbuminuria and macroalbuminuria were 10.25 and 12.5 IU/ml, respectively.

This was supported by Beer et al. [25], who documented increased t-PA and PAI-1 in diabetic patients with and without vascular complications. Also, the prospective study conducted by Wannamethee et al. [26] documented higher levels of t-PA as a risk factor for the development of diabetes in the healthy individuals.

However, Chudy P et al. [18] found that plasma levels of t-PA were not significantly increased in either normoalbuminuric or microalbuminuric diabetic groups compared with controls.

Umpaichitra et al. [27] studied hemostatic factors in 12 type 2 diabetic patients and 17 nondiabetic obese adolescents; they found that PAI-1 activities were significantly greater in diabetic patients than in control individuals, whereas t-PA activities were significantly lower. They concluded that the elevated PAI-1 and lower t-PA activities suppress fibrinolysis in adolescents with type 2 diabetes, adding a risk factor for cardiovascular disease.

These contradictory results concerning t-PA were explained by Sobel et al. [28], who found that both PAI-1 and total t-PA antigen (free t-PA plus t-PA complexed with PAI-1) concentrations were increased in association with diabetes, but the concentration of free t-PA and hence the t-PA activity were not increased. Hence, the increased total t-PA is masking the impaired fibrinolytic system activity associated with type 2 diabetes.

We found that the ratio of PAI-1/t-PA in diabetic patients with macroalbuminuria was statistically significantly higher when compared with the control group, that is, diabetic patients with macroalbuminuria had significant elevation in PAI-1/t-PA than the control individuals. Thus, impaired fibrinolysis was higher in diabetic patients with a higher UAE.

Impaired fibrinolysis was related to the activated renin-angiotensin-aldosterone system in type 2 diabetic patients as angiotensin two directly induces the expression of PAI-1 and indirectly through the secretion of transforming growth factor-b, a potent profibrotic molecule that also induces the expression of PAI-1 and stimulates ECM synthesis and fibrosis [12]. Previous studies found that high doses of angiotensin 2 receptor inhibitor decrease glomerulosclerosis through the inhibition of PAI-1 [10]. The combination between decreased ECM degradation due to impaired fibrinolysis and proteolysis with the increased ECM synthesis and fibrosis induced by transforming growth factor-b predispose one to tissue fibrosis, and so increase renal nephropathy and renal mortality [12].

Our study showed a significantly higher serum creatinine concentration in patients with a higher UAE and impaired fibrinolysis. This was supported by the study conducted by Kamgar et al. [12], who examined the changes that occur in the serum creatinine concentration as an index of kidney function over a 7-year follow-up period in relationship to the fibrinolytic ratio. A marked increase in serum creatinine was observed in patients with a greater UAE and higher PAI-1/t-PA, and they related the association between high PAI-1 and UAE to the endothelial dysfunction that occurs in these diabetic patients.

We found a significant positive correlation between PAI-1 with serum creatinine and the 24-h urinary albumin and a significant negative correlation with creatinine clearance. Regarding t-PA, it showed a significant positive correlation with the 24-h urinary albumin. Hence, the deterioration in renal function was associated with elevations in PAI-1 and t-PA.

We did not find any correlations between PAI-1 and fasting serum lipids. This was supported by Ho and Jap [29] In contrast, Lira et al. [30] found a positive correlation between PAI-1 and serum lipids, suggesting that an increased risk of cardiovascular disease associated with a sedentary lifestyle is linked to an elevated PAI-1, and Peter et al. [18] found a positive correlation between PAI-1 levels and triglycerides, which proved that dyslipidemia and hypofibrinolysis are closely linked.

Also, we did not find any correlation between PAI-1 and fasting blood sugar; this was also supported by Ho and Jap [29] who did not find any correlation between PAI-1 activity and the increase in glucose and fasting insulin, but Heldgaard et al. [31] found increased levels of PAI-1 and t-PA to be associated with increased fasting blood sugar.

From this study, we concluded that serum levels of PAI-1 and t-PA and the PAI/t-PA ratio were significantly elevated in type 2 diabetic patients with higher UAE. PAI-1 and t-PA were significantly positively correlated with deteriorations in renal functions. Hence, our study suggested that impaired fibrinolytic activity in type 2 diabetic patients and increased UAE are correlated with the progression of renal disease through deteriorations in renal functions. These results may indicate that in addition to hyperglycemia, hypertension, and dyslipidemia that occur in association with type 2 diabetes, impaired fibrinolysis may be one of the important contributors that predispose one to renal complications.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1Morrish NJ, Wang SL, Stevens LK, Fuller JH, Keen H. Mortality and causes of death in the WHO Multinational Study of Vascular Disease in Diabetes. Diabetologia 2001; 44: (Suppl 2):S14-S21.
2 U.S. Renal Data System: 2008 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. Available online at http://www.usrds.org/2008 [Last accessed on 2009 Sep 22].
3 Rerolle JP, Hertig A, Nguyen G, Sraer JD, Rondeau EP. Plasminogen activator inhibitor type 1 is a potential target in renal fibrogenesis. Kidney Int 2000; 58:1841-1850.
4 Rondeau E, Sraer JD, Schleuning WD. The renal plasminogen activating system, in molecular nephrology: kidney function in health and diseases. edited by Schlondorff D, Bonventre JV. New York: Marcel Dekker Inc. 1995; 699-715.
5 Eddy AA, Fogo AB. Plasminogen activator inhibitor-1 in chronic kidney disease: evidence and mechanisms of action. J Am Soc Nephrol 2006; 17:2999-3012.
6 Ichimura A, Matsumoto S, Suzuki S, Dan T, Yamaki S, Sato Y, et al. A small molecule inhibitor to plasminogen activator inhibitor 1 inhibits macrophage migration. Arterioscler Thromb Vasc Biol 2013; 33:935-942.
7 Blasi F. Proteolysis, cell adhesion, chemotaxis, and invasiveness are regulated by the u-PA-u-PAR-PAI-1 system. Thromb Haemost 1999; 82:298-304.
8 Genovese F, Manresa AA, Leeming DJ, Karsdal MA, Boor P. The extracellular matrix in the kidney: a source of novel non-invasive biomarkers of kidney fibrosis?. Fibrogenesis Tissue Repair 2014; 7:4.
9 Aldigier JC, Kanjanbuch T, Ma LJ, Brown NJ, Fogo AB. Regression of existing glomerulosclerosis by inhibition of aldosterone. J Am Soc Nephrol 2005; 16:3306-3314.
10Ma LJ, Nakamura S, Aldigier JC, Rossini M, Yang H, Liang X, et al. Regression of glomerulosclerosis with high-dose angiotensin inhibition is linked to decreased plasminogen activator inhibitor-1. J Am Soc Nephrol 2005; 16:966-976.
11Rezaie AR. Vitronectin functions as a cofactor for rapid inhibition of activated protein C by plasminogen activator inhibitor-1. Implications for the mechanism of profibrinolytic action of activated protein C. J Biol Chem 2001; 276:15567-15570.
12Kamgar M, Nobakhthaghighi N, Shamshirsaz AA, Estacio RO, McFann KK, Schrier RW. Impaired fibrinolytic activity in type II diabetes: correlation with urinary albumin excretion and progression of renal disease. Kidney Int 2006; 69:1899-1903.
13Karnib HH, Ziyadeh FN. The cardiorenal syndrome in diabetes mellitus. Diabetes Res Clin Pract 2010; 89:201-208.
14Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Follow-up Report on the Diagnosis of Diabetes Mellitus. Diabetes Care 2003; 26:5-20.
15Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16:31-41.
16Declerck PJ, Alessi MC, Verstreken M, et al. Measurement of plasminogen activator inhibitor-1 in biological fluids with a murine monoclonal antibody based enzyme-linked-immunosorbent assay. Blood 1998; 71:220-225.
17Juhan-Vague I, Pyke SD, Alessi MC, Jespersen J, Haverkate F, Thompson SG. Fibrinolytic factors and the risk of myocardial infarction or sudden death in patients with angina pectoris. ECAT Study Group. European Concerted Action on Thrombosis and Disabilities. Circulation 1996; 94:2057-2063.
18Chudy P, Kotulicova D, Stasko J, et al. The relationship among TAFI, t-PA, PAI-1 and F1R2 in type 2 diabetic patients with normoalbuminuria and microalbuminuria. Blood Coagul Fibrinolysis 2011; 22:493-498.
19Collins SJ, Alexander SL, Lopez-Guisa JM, Cai X, Maruvada R, Chua SC, et al. Plasminogen activator inhibitor-1 deficiency has renal benefits but some adverse systemic consequences in diabetic mice. Nephron Exp Nephrol 2006; 104:e23-e34.
20Szelag B, Wroblewski M, Castenfors J, Henricsson M, Berntorp K, Fernlund P, Sundkvist G. Obesity, microalbuminuria, hyperinsulinemia, and increased plasminogen activator inhibitor 1 activity associated with parasympathetic neuropathy in type 2 diabetes. Diabetes Care 1999; 22:1907-1908.
21Bastard JP, Pieroni L, Hainque B Relationship between plasma plasminogen activator inhibitor 1 and insulin resistance. Diabetes Metab Res Rev 2000; 16:192-201.
22Yener S, Comlekci A, Akinci B, Demir T, Yuksel F, Ozcan MA, et al. Soluble CD40 ligand, plasminogen activator inhibitor-1 and thrombin-activatable fibrinolysis inhibitor-1-antigen in normotensive type 2 diabetic subjects without diabetic complications. Effects of metformin and rosiglitazone. Med Princ Pract 2009; 18:266-271.
23Festa A, Williams K, Tracy RP, Wagenknecht LE, Haffner SM. Progression of plasminogen activator inhibitor-1 and fibrinogen levels in relation to incident type 2 diabetes. Circulation 2006; 113:1753-1759.
24Nagi DK, Knowler WC, Hanson RL, Ali VM, Yudkin JS. Plasminogen activator inhibitor (PAI-1) and non-insulin-dependent diabetes in Pima Indians, south Asians and Europeans. Populations at varying risk of NIDDM and coronary artery disease. Thromb Haemost 1996; 75:921-927.
25Beer S, Feihl F, Ruiz J, Juhan-Vague I, Aillaud MF, Wetzel SG, et al. Comparison of skin microvascular reactivity with hemostatic markers of endothelial dysfunction and damage in type 2 diabetes. Vasc Health Risk Manag 2008; 4:1449-1458.
26Wannamethee SG, Sattar N, Rumley A, Whincup PH, Lennon L, Lowe GD. Tissue plasminogen activator, von Willebrand factor, and risk of type 2 diabetes in older men. Diabetes Care 2008; 31:995-1000.
27Umpaichitra V, Hussain MM, Castells S. Plasminogen activator inhibitor-1 and tissue-plasminogen activator in minority adolescents with type 2 diabetes and obesity. Pediatr Res 2005; 58:483-487.
28Sobel BE, Tilton L, Neimane D, Schnure J Increased tissue-type plasminogen activator: a facade in the fibrinolytic system in type 2 diabetes. Coron Artery Dis 2005; 16:31-35.
29Ho CH, Jap TS. Relationship of plasminogen activator inhibitor-1 with plasma insulin, glucose, triglyceride and cholesterol in Chinese patients with diabetes. Thromb Res1993; 69:271-277.
30Lira FS, Rosa JC, Lima-Silva AE, Souza HA, Caperuto EC, Seelaender MC, et al. Sedentary subjects have higher PAI-1 and lipoproteins levels than highly trained athletes. Diabetol Metab Syndr 2010; 2:7.
31Heldgaard PE, Sidelmann JJ, Hindsberger C, Olivarius Nde F, Henriksen JE, Gram J. Relationship of glucose concentrations with PAI-1 and t-PA in subjects with normal glucose tolerance. Diabet Med 2006; 23:887-893.