Fasting Intact Proinsulin Is a Highly Specific Predictor of Insulin

Epidemiology/Health Services/Psychosocial Research ORIGINAL ARTICLE

Fasting Intact Proinsulin Is a Highly Specific Predictor of Insulin Resistance in Type 2 Diabetes

1,2 1 ANDREAS PFUTZNER¨ , MD, PHD SABINE PAHLER, MSC sulin is a commonly secreted by-product 3 4 THOMAS KUNT, MD THOMAS KONRAD, MD of ␤-cell secretion in later stages of type 2 1 5 CLOTH HOHBERG, MD GEORG L¨UBBEN, MD 1 1 diabetes and is involved in the develop- AGNES MONDOK, MD THOMAS FORST, MD ment of macrovascular disease, whereas des64,65 is usually not present in the cir- culation (11,12). Assays that try to differentiate be- OBJECTIVE — In later stages of type 2 diabetes, proinsulin and proinsulin-like molecules are tween intact proinsulin and the specific secreted in increasing amounts with insulin. A recently introduced chemiluminescence assay is and unspecific proinsulin derivatives able to detect the uncleaved “intact” proinsulin and differentiate it from proinsulin-like mole- were used in cross-sectional and prospec- cules. This investigation explored the predictive value of intact proinsulin as an insulin resistance marker. tive follow-up studies to investigate the real relation of immunoreactive insulin RESEARCH DESIGN AND METHODS — In total, 48 patients with type 2 diabetes (20 with coronary risk and to explore and women and 28 men, aged 60 Ϯ 9 years [means Ϯ SD], diabetes duration 5.1 Ϯ 3.8 years, BMI confirm the important role of intact pro- Ϯ 2 Ϯ 31.2 4.8 kg/m , and HbA1c 6.9 1.2%) were studied by means of an intravenous glucose insulin in the prediction and diagnosis of tolerance test and determination of fasting values of intact proinsulin, insulin, resistin, adiponec- type 2 diabetes (13,14). Recent presenta- tin, and glucose. Insulin resistance was determined by means of minimal model analysis (MMA) tions indicate a potential value for fasting (as the gold standard) and homeostatis model assessment (HOMA). intact proinsulin to become a marker of insulin resistance, e.g., to identify patients RESULTS — There was a significant correlation between intact proinsulin values and insulin eligible for an insulin sensitizer therapy resistance (MMA P Ͻ 0.05 and HOMA P Ͻ 0.01). Elevation of intact proinsulin values above the reference range (Ͼ10 pmmol/l) showed a very high specificity (MMA 100% and HOMA 92.9%) (15), or to monitor treatment success dur- and a moderate sensitivity (MMA 48.6% and HOMA 47.1%) as marker for insulin resistance. ing sensitizer treatment (16). Summariz- Adiponectin values were slightly lower in the insulin resistant group, but no correlation to ing all scientific literature, there is a great insulin resistance could be detected for resisitin in the cross-sectional design. likelihood that intact and/or total proin- sulin may become an important diagnos- CONCLUSIONS — Elevated intact proinsulin seems to indicate an advanced stage of ␤-cell tic tool in type 2 diabetes. exhaustion and is a highly specific marker for insulin resistance. It might be used as arbitrary The following cross-sectional study marker for the therapeutic decision between secretagogue, sensitizer, or insulin therapy in type describes the relationship between intact 2 diabetes. proinsulin as measured by means of a re- cently introduced, very specific chemilu- Diabetes Care 27:682–687, 2004 minescence method (17) and several other clinical and biochemical markers of insulin resistance in a well-characterized n the past decades, insulin resistance proinsulin, which was present in many group of 48 patients with dietary and and cardiovascular disease in type 2 di- assays for insulin measurement. Rela- orally treated type 2 diabetes. I abetic patients have been linked to hy- tively few and contradicting data are perinsulinemia as an independent risk available on the relation of proinsulin to factor (1–3). However, many studies were insulin resistance, cardiovascular disease, RESEARCH DESIGN AND not able to confirm the negative impact of and atherosclerosis (6–8). Prospective METHODS — The study was con- insulin on cardiovascular risk (4–6). One studies have demonstrated that the deter- ducted at the Institute for Clinical Re- explanation for this overall confusing data mination of the proinsulin-to-insulin ra- search and Development (Mainz, may be found in the cross-reactivity be- tio may be used to predict deterioration in Germany) in accordance with Good Clin- tween insulin and its precursor molecule, glucose tolerance (9,10). Des31,32 proin- ical Practice Guidelines and was approved ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● by the responsible local ethics review committee. In total, 48 patients on oral From the 1Institute for Clinical Research and Development, Mainz, Germany; the 2University of Applied Sciences, Rheinbach, Germany; the 3Department of Internal Medicine, Al Zayed Hospital Abu Dhabi, United anti-diabetic treatment were enrolled in Arab Emirates; the 4Institute for Metabolic Research Frankfurt, Frankfurt, Germany; and 5Takeda Pharma the investigation (20 women and 28 men, Aachen, Germany. aged 60 Ϯ 9 years [means Ϯ SD], diabetes Address correspondence and reprint requests to Andreas Pfu¨ tzner, MD, PhD, Institute for Clinical Re- duration 5.1 Ϯ 3.8 years, BMI 31.2 Ϯ 4.8 search and Development (IKFE) Parcusstr. 8, D-55116 Mainz, Germany. E-mail: [email protected]. 2 Ϯ Received for publication 5 June 2003 and accepted in revised form 11 December 2003. kg/m , and HbA1c 6.9 1.2%). Abbreviations: HOMA, homeostatis model assessment; MMA, minimal model analysis. The patients arrived in the morning of © 2004 by the American Diabetes Association. the study day after an overnight fast, and

682 DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 Pfu¨tzner and Associates

Table 1—Patient characteristics for insulin-resistant and insulin-sensitive patients according estimate was expressed as fractional SD to MMA and HOMA analysis (the ratio between the SD of the estimate and the estimated value expressed as a Insulin resistant Insulin sensitive P percentage). Estimation of Si with frac- MMA tional SD Ͼ25% were excluded from fur- n 35 13 ther evaluation. Insulin resistance was Ϫ4 Ϫ1 Age (years) 59 Ϯ 960Ϯ 6NSdefined as values Ͻ1.5 ϫ 10 min Ϫ1 Ϫ1 Disease duration (months) 54 Ϯ 42 65 Ϯ 46 NS ␮U ml (20). Body weight (kg) 94 Ϯ 19 85 Ϯ 13 NS, 0.118 Homeostatis model assessment BMI (kg/m2) 31.8 Ϯ 5.2 29.5 Ϯ 3.3 NS, 0.145 (HOMA) score calculation was applied as Hypertension (number, %) 28 (80%) 10 (77%) NS a second measure for insulin resistance Ϯ Ϯ Ͻ HbA1c (%) 7.1 1.2 6.2 0.6 0.05 analysis as previously described (21). The Fasting glucose (mg/dl) 119 Ϯ 33 102 Ϯ 22 NS, 0.093 estimate of insulin resistance by HOMA Fasting insulin (␮U/ml) 13.1 Ϯ 6.6 9.4 Ϯ 3.9 NS, 0.063 score was calculated with the following Fasting intact proinsulin (pmol/l) 19.9 Ϯ 11.7 6.0 Ϯ 2.6 Ͻ0.01 formula: fasting serum insulin (␮U/ml) ϫ Proinsulin/insulin ratio (pmol/mU) 1.21 Ϯ 0.70 0.71 Ϯ 0.36 Ͻ0.05 fasting plasma glucose (mmol/l)/22.5. As Fasting resistin (ng/l) 0.07 Ϯ 0.12 0.11 Ϯ 0.18 NS described by Hedblad et al. (22), patients Fasting adiponectin (␮/ml) 7.05 Ϯ 4.43 9.70 Ϯ 5.66 NS, 0.095 with HOMA score values exceeding the Fibrinogen (mg/dl) 2.9 Ϯ 0.9 2.6 Ϯ 0.6 NS 75th percentile (i.e., 2.0) were considered HOMA analysis to have insulin resistance. n 34 14 Glucose levels were determined using Age (years) 58 Ϯ 962Ϯ 6 NS, 0.92 a standard laboratory reference method Disease duration (months) 54 Ϯ 34 63 Ϯ 59 NS (glucose oxidase method) (Super GL RLT, Body weight (kg) 96 Ϯ 17 81 Ϯ 14 Ͻ0.01 Mo¨hnesee, Germany). Lipid values (tri- BMI (kg/m2) 32.3 Ϯ 4.9 28.5 Ϯ 3.5 Ͻ0.05 glycerides, total cholesterol, and HDL Hypertension (%) 27 (79%) 11 (79%) NS cholesterol) were determined using stan- Ϯ Ϯ Ͻ HbA1c (%) 7.2 1.2 6.2 0.8 0.05 dard dry chemistry methods. Fasting glucose (mg/dl) 126 Ϯ 29 86 Ϯ 13 Ͻ0.001 Insulin and proinsulin were deter- Fasting insulin (␮U/ml) 14.6 Ϯ 5.5 6.0 Ϯ 1.7 Ͻ0.001 mined by means of specific chemilumi- Fasting intact proinsulin (pmol/l) 15.4 Ϯ 11.6 5.3 Ϯ 2.1 Ͻ0.005 nescence tests (MLT insulin [intra- and Proinsulin/insulin ratio (pmol/mU) 1.58 Ϯ 0.66 0.66 Ϯ 0.36 Ͻ0.001 interassay coefficient of variation {CV} 3.8 Fasting resistin (ng/l) 0.08 Ϯ 0.15 0.08 Ϯ 0.11 NS and 2.3%, reference range Ͻ30 ␮U/ml] Fasting adiponectin (␮g/ml) 6.95 Ϯ 4.33 9.77 Ϯ 5.68 NS, 0.068 and MLT Intact Proinsulin [intra- and in- Fibrinogen (mg/dl) 2.8 Ϯ 0.9 2.7 Ϯ 0.6 NS terassay CV 5.2 and 8.6%, reference range Ͻ Data are means Ϯ SD unless otherwise indicated. 10 pmol/l]; Sciema Diagnostics, Mainz, Germany). The tests were carried out as blood samples were taken for the deter- cording to Bergmann by means of the previously described (17). mination of insulin, intact proinsulin, SAAM 1.2 software package (SAAM Insti- Resistin concentrations were deter- glucose, resisitin, adiponectin, and lipids. tute, Seattle, WA). Precision parameters mined by means of a clinical research Thereafter, a standardized modified fre- quently sampled intravenous glucose tol- erance test was performed for the determination of insulin resistance by means of the minimal model analysis (MMA) according to Bergman et al. (18). On the investigational day, all anti- diabetic medications were discontinued in the morning. Before starting the proce- dure, capillary blood glucose had to be Ͻ180 mg/dl. Glucose (300 mg/kg body wt, glucose 40%) was injected in a bolus over a maximum time period of 2 min at baseline. Twenty minutes later, insulin (0.07 units/kg body wt) (Huminsulin, Eli Lilly) was injected in a bolus as previously described (19). Frequent blood samples for the determination of blood glucose and serum insulin were taken from Ϫ15 to 180 min. Calculation of insulin sensi- Figure 1—Analysis of fasting intact proinsulin values in relation to the insulin sensitivity status tivity was performed using the MMA ac- as assessed by MMA and HOMA analysis. ***P Ͻ 0.001

DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 683 Intact proinsulin and insulin resistance

ues above the reference range showed a very high specificity (MMA 100% and HOMA 92.9%) and a moderate sensitivity (MMA 48.6% and HOMA 47.1%) as a marker for insulin resistance. The hyper- bolic correlation between intact proinsu- lin values and the MMA value and the linear correlation between the HOMA score values and the intact proinsulin concentrations are given in Figs. 2 and 3. The clinical characteristics of the pa- tients with elevated or normal intact pro- insulin levels in consideration of their insulin resistance status as assessed by MMA is given in Table 2. Although pa- tients with elevated intact proinsulin val- ues were significantly different from the combined other two groups with regard to body weight, BMI, HbA1c, fasting glu- cose, fasting insulin, fasting intact proin- Figure 2—Correlation between fasting intact proinsulin values and the insulin sensitivity index sulin, and proinsulin-to-insulin ratio, according to the MMA. there were no significant differences be- tween the insulin-resistant or insulin- sensitive patients with normal fasting enzyme-linked immunosorbent assay significantly correlated in the contin- morning intact proinsulin concentra- provided by Immundiagnostik, Ben- gency table analysis (P Ͻ 0.001). The de- tions. When comparing the patients with sheim, Germany (intra- and interassay mographic data and the laboratory results elevated intact proinsulin levels individu- CV 4.2 and 16.2%, reference range not of the respective groups are given in Table ally to the other two groups of resistant defined). 1. There were significant differences be- and sensitive patients with normal proin- Adiponectin measurements were per- tween insulin-sensitive and insulin- sulin levels, the differences from the insu- formed by means of the Human Adi- resistant patients for both methods for a lin-sensitive group were much more ponectin Radioimmunoassay (Linco, St. variety of clinical and laboratory parame- pronounced than the differences from Charles,MO).Asgivenbythemanufactur- ters, including HbA1c, fasting intact pro- the insulin-resistant group (Fig. 3 and er, this assay has intra- and interassay insulin (Fig. 1), and the proinsulin-to- Table 2). CVs of 1.8–6.2% and 6.9–9.3%, respec- insulin ratios. It could be demonstrated No significant differences between tively (reference range not defined). that the elevation of intact proinsulin val- the groups were observed with regard to Statistical analysis was performed us- ing descriptive statistics and appropriate parametrical and nonparametrical tests. A Student’s t test and Mann-Whitney’s U test were used to compare the means of the variables measured. All tests were two- sided. Regression analyses were per- formed by means of the nonparametric method of Passing and Bablok (23). Re- sults with P values Ͻ0.05 were considered statistically significant. All calculations were made with the SPSS version 9.0 sta- tistical package.

RESULTS — MMA resulted in 35 insu- lin-resistant and 13 insulin-sensitive pa- tients, including 49% (17 of 35) and 0 (0 of 13) of patients with elevated intact pro- insulin levels, respectively (P Ͻ 0.05). HOMA analysis of the same group re- vealed 34 resistant (16 with elevated in- tact proinsulin, 47%) and 14 sensitive patients (1 with elevated proinsulin, 7%, Figure 3—Correlation between fasting intact proinsulin values and the Si according to the HOMA P Ͻ 0.01). HOMA and MMA proved to be analysis.

684 DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 Pfu¨tzner and Associates

Table 2—Results for patients with different intact proinsulin secretion status

Elevated intact Normal intact Insulin resistant Insulin sensitive proinsulin proinsulin ϩ normal intact ϩ normal intact (Ͼ10 pmol/l) (Ͻ10 pmol/l) proinsulin proinsulin n 17 31 18 13 Age (years) 58 Ϯ 10 60 Ϯ 860Ϯ 960Ϯ 6 Disease duration (months) 60 Ϯ 45 63 Ϯ 46 65 Ϯ 54 61 Ϯ 34 Body weight (kg) 99 Ϯ 18 88 Ϯ 17* 90 Ϯ 19 85 Ϯ 13* BMI (kg/m2) 33.2 Ϯ 5.2 30.1 Ϯ 4.3* 30.5 Ϯ 4.9 29.5 Ϯ 3.3* Ϯ Ϯ Ϯ Ϯ HbA1c (%) 7.5 1.3 6.5 1.0† 6.8 1.1 6.2 0.6† Fasting glucose (mg/dl) 127 Ϯ 32 107 Ϯ 29* 112 Ϯ 33 102 Ϯ 22* Fasting insulin (␮U/ml) 15.8 Ϯ 5.6 10.1 Ϯ 5.6† 10.6 Ϯ 6.6* 9.4 Ϯ 3.9† Fasting intact proinsulin (pmol/l) 23.7 Ϯ 11.3 6.3 Ϯ 2.2‡ 6.6 Ϯ 1.9‡ 6.0 Ϯ 2.6‡ Proinsulin/insulin ratio (pmol/mU) 1.58 Ϯ 0.64 0.79 Ϯ 0.50‡ 0.86 Ϯ 0.58† 0.71 Ϯ 0.35‡ Resistin (ng/l) 0.07 Ϯ 0.13 0.08 Ϯ 0.15 0.07 Ϯ 0.12 0.11 Ϯ 0.18 Adiponectin (␮g/ml) 6.46 Ϯ 3.94 8.49 Ϯ 5.24 9.70 Ϯ 5.66 7.61 Ϯ 4.86 Triglycerides (mg/dl) 197 Ϯ 123 161 Ϯ 83 151 Ϯ 59 176 Ϯ 110 Total cholesterol (mg/dl) 208 Ϯ 37 201 Ϯ 48 194 Ϯ 39 212 Ϯ 59 HDL cholesterol (mg/dl) 44 Ϯ 747Ϯ 10 47 Ϯ 947Ϯ 10 Fibrinogen (mg/dl) 3.0 Ϯ 1.0 2.7 Ϯ 0.7 2.7 Ϯ 0.8 2.6 Ϯ 0.7 Hip/waist circumference ratio 1.03 Ϯ 0.07 1.03 Ϯ 0.07 1.04 Ϯ 0.07 1.02 Ϯ 0.06 Hypertension 14 (82%) 24 (77%) 14 (78%) 10 (77%) Complications Macrovascular 4 (24%) 6 (19%) 5 (28%) 1 (8%) Microvascular 6 (35%) 6 (19%) 3 (17%) 3 (23%) Therapy No drug 4 (24%) 6 (19%) 4 (22%) 2 (15%) Sulfonylurea 12 (71%) 11 (35%) 6 (33%) 5 (38%) Metformine 8 (47%) 17 (55%) 9 (50%) 8 (62%) Data are means Ϯ SD or n (%). *P Ͻ 0.05, †P Ͻ 0.001, and ‡P Ͻ 0.001 vs. elevated proinsulin group. the resistin or lipid values. There was a ten- epidemiological studies (37). Like the use intact proinsulin levels, with cutoff values dency toward lower adiponectin concen- of insulin alone, the determination of in- of Ͼ10 pmol/l for intact proinsulin and tration in the insulin resistant patients. sulin resistance by HOMA is limited to the Ͻ1.5 for the MMA. early stage of type 2 diabetes because Proinsulin concentrations are related CONCLUSIONS — Type 2 diabetes ␤-cell secretion in later stages is also com- to atherosclerosis and cardiovascular dis- is characterized by impaired insulin sen- prised of intact proinsulin and split prod- ease (6–9,25,26). A therapeutic trial for sitivity and deficient ␤-cells, both occur- ucts. Thus, insulin levels and HOMA the treatment of type 2 diabetes with ex- ing with different degrees of severity. score can be low or normal even in pro- ternally administered biosynthetic proin- Understanding of the metabolic situation gressed stages of insulin resistance. sulin was stopped prematurely because of by the physician has, however, a major im- The presented study was performed an increase in the incidence of cardiovas- pact on the proper selection of an effective to investigate the suitability of fasting cular events in the proinsulin-treated therapy. The two reference methods for intact proinsulin as measured by a re- study population (27). In our cross- measurement of insulin resistance, the cently introduced chemiluminescence sectional study, 24% of the patients with hyperinsulinemic-euglycemic clamp test assay to assess ␤-cell function and to elevated intact proinsulin levels and 19% and the frequently sampled intravenous predict insulin resistance. Significant of the patients with normal intact proin- glucose test, cannot be used in daily prac- differences were seen and had to be ex- sulin levels suffered from macrovascular tice (24). Therefore, simple clinical scores pected with HOMA for the glucose and disease. and biochemical markers are under inves- insulin values due to the calculation pro- Our findings suggest that even after a tigation in order to provide the clinical cedure of the HOMA score. Proinsulin short disease duration, patients with type diagnosis of insulin resistance and secre- and insulin secretion are closely related 2 diabetes may belong to any of the fol- tion disorder with sufficient specificity in later stages of type 2 diabetes, which lowing three groups: 1) patients with sub- and sensitivity to be suitable in general is confirmed by the regression analysis clinical insulin resistance (possibly practice. for the intact proinsulin levels and the in- representing an earlier pathophysiologi- HOMA is accepted with some limi- sulin secretion–dependent HOMA score cal stage), 2) patients with pronounced tations as a standard method for the de- values. A hyperbolic correlation was seen clinical insulin resistance without signifi- termination of insulin resistance in between the MMA values and the fasting cant deterioration of ␤-cell secretion

DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 685 Intact proinsulin and insulin resistance

(possibly an intermediate stage), and 3) glitazones, or combinations thereof) or SL: Hyperinsulinemia does not increase patients with pronounced clinical insulin insulin therapy to protect ␤-cell function. the risk of fatal cardiovascular disease in resistance with significant deterioration of In contrast to intact proinsulin, there elderly men or women without diabetes: ␤-cell secretion (possibly a later stage) as was only a tendency to decreased adi- the Rancho Bernardo Study, 1984–1991. ponectin levels in the insulin-resistant Am J Epidemiol 140:857–869, 1994 indicated by an increase in fasting intact 6. Katz RJ, Ratner RE, Cohen RM, Eisen- proinsulin secretion. It is remarkable that group. This finding is in line with other hower E, Werme D: Are insulin and pro- patients with impairment of ␤-cell func- recent investigations that have also shown insulin independent risk markers for tion were not different from the others lower adiponectin plasma levels in sub- premature coronary artery disease? Diabe- with regard to their disease duration or jects with insulin resistance (32). As re- tes 45:736–741, 1996 other clinical features. This may be ex- ported by other groups, there was also no 7. Båvenholm P, Proudler A, Tornvall P, plained by the well-known delay of dia- difference in resistin values between sen- Godsland I, Landau C, De Faire U, Ham- betes diagnostics. Already 20–30% of sitive and resistant patients (33). The clin- sten A: Insulin, intact and split proinsulin, ical implications of resistin measurements and coronary artery disease in young patients at the time of their first diagnosis men. Circulation 92:1422–1429, 1995 of diabetes are in a progressed stage of in humans are still a matter of debate, but our findings indicate that resistin may not 8. Yudkin JS, Denver AE, Mohamed-Ali V, type 2 diabetes with micro- and macro- Ramaiya KL, Nagi DK, Goubet S, McLarty vascular complications (28). be a powerful predictive marker for insu- DG, Swai A: The relationship of concen- Such a potential classification might lin resistance in a cross-sectional study trations of insulin and proinsulin-like have substantial therapeutic conse- design. molecules with coronary heart disease quences. It is common practice to treat Further larger cross-sectional and prevalence and incidence: a study of two patients initially with diet and exercise prospective studies are required to con- ethnic groups. Diabetes Care 20:1093– and continue with oral treatment (most firm our findings and hypotheses that 1100, 1997 9. Haffner SM, Gonzalez C, Mykka¨nnen L, commonly with secretagogues or met- fasting intact proinsulin, as measured by means of the new chemiluminescence as- Stern M: Total immunoreactive proinsu- formin), before the introduction of sensi- lin, immunoreactive insulin and specific tizer and/or insulin treatment. Although say, may become an interesting diagnostic tool for both selection of an appropriate insulin in relation to conversion to type 2 diet and metformin therapy have both diabetes: the Mexico City Diabetes Study. been reported to decrease insulin resis- therapy and evaluation of treatment suc- Diabetologia 40:830–837, 1997 cess in patients with type 2 diabetes. In tance and proinsulin levels in patients 10. Nijpels G, Popp-Snijders C, Kostense PJ, conclusion, however, elevated proinsulin with type 2 diabetes (29,30), blood pro- Bouter LM, Heiner RJ: Fasting proinsulin is a highly specific marker for clinically insulin levels showed no change or were and 2-h post-load glucose levels predict relevant insulin resistance. the conversion to type 2 diabetes in sub- increased in patients treated with sufony- jects with impaired glucose tolerance: the lurea (31). In our cross-sectional study Hoorn study. Diabetologia 39:113–118, population, 71% of the insulin-resistant Acknowledgments— This study was sup- 1996 patients with elevated proinsulin concen- ported by an unrestricted grant from Takeda 11. Wareham NJ, Byrne CD, Hales CN: Role tration were treated with sulfonylurea Pharma, Germany. of insulin and proinsulin in diabetic vas- alone or in combination with metformin, We greatfully acknowledge the excellent cular disease. Metabolism 44 (Suppl. 4): whereas only 33% of the insulin-resistant laboratory work performed by Mirjam Lo¨big, S76–S82, 1995 patients with normal intact proinsulin Maja Knesovic, and Erika Budo-Guetaifi at the 12. Yudkin JS: Circulating proinsulin-like Institute for Clinical Research and Develop- had this kind of therapy. However, pro- molecules. J Diabetes Complications 7:113– ment Laboratory (IKFE) and the assistance of 123, 1993 spective studies will be a more appropri- the clinical staff of the IKFE clinical unit. 13. Kahn SE, Halban PA: Release of incom- ate measure to investigate the influence of pletely processed proinsulin is the cause sulfonylurea therapy on ␤-cell deteriora- of the disproportionate proinsulinemia of tion as measured by the fasting intact pro- References NIDDM. Diabetes 46:1725–1732, 1997 insulin chemiluminescence. It has already 1. Welborn TA, Wearne K: Coronary heart 14. Røder ME, Porte D, Schwartz RS, Kahn been demonstrated with a less specific disease incidence and cardiovascular SE: Disproportionately elevated proinsu- mortality in Busselton with reference to lin levels reflect the degree of impaired proinsulin assay in Japanese patients that ␤ a 12-week treatment with pioglitazone glucose and insulin concentrations. Dia- -cell secretory capacity in patients with betes Care 2:154–160, 1979 non-insulin dependent diabetes mellitus. alone or in combination with a sulfonyl- 2. Despre´s JP, Lamarche B, Maurie`ge P, Can- J Clin Endocrinol Metab 83:604–608, urea leads to a decrease in fasting proin- tin B, Dagenais GR, Moorjani S, Lupien PJ: 1998 sulin levels, whereas treatment with Hyperinsulinemia as an independent risk 15. Festa A, D’Agostino R, Mykka¨nen L, Tracy sulfonylurea alone did not decrease pro- factor for ischiemic heart disease. N Engl RP, Hales CN, Howard BV, Haffner SM: insulin in the third comparator group J Med 334:952–957, 1996 LDL particle size in relation to insulin, (16). Considering our findings, it can be 3. Arrants J: Hyperinsulinemia and cardio- proinsulin, and insulin sensitivity: the In- speculated (with all limitations of our vascular risk. Heart Lung 23:118–122, sulin Resistance Atherosclerosis Study. study design and power) that even after a 1994 Diabetes Care 22:1688–1693, 1999 4. Welin L, Eriksson H, Larsson B, Ohlson 16. Kubo K: Effect of Pioglitazone on blood short disease duration half of the insulin- LO, Svardsudd K, Tibblin G: Hyperinsu- proinsulin levels in patients with type 2 resistant patients would most likely not linemia is not a major coronary risk factor diabetes. Endocr J 49:323–328, 2002 benefit from a secretagogue treatment, in elderly men: the study of men born in 17. Pfu¨ tzner A, Kunt T, Lo¨big M, Knesovic M, but would better be treated with a sensi- 1913. Diabetologia 35:766–770, 1992 Forst T: Clinical and laboratory evalua- tizer (i.e., exercise plus diet, metformin, 5. Ferrara A, Barett-Connor EL, Edelstein tion characteristics of two new chemilu-

686 DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 Pfu¨tzner and Associates

minescence assays for intact and total sample sizes: application of linear regres- subjects: implications for preventing proinsulin. Clin Chem Lab Med 41:1234– sion procedures for method comparison coronary heart disease during the pre- 1238, 2003 studies in clinical chemistry, part II. J Clin diabetic state. Circulation 101:975–980, 18. Bergman RN: Towards physiological un- Chem Clin Biochem 22:431–445, 1984 2000 derstanding of glucose tolerance: mini- 24. Saad MF, Anderson RL, Laws A, Wa- 29. Davies MJ, Metcalfe J, Day JL, Grenfall A, mal-model approach. Diabetes 38:1512– tanabe RM, Kades WW, Chen YD, Sands Hales CN, Gray IP: Improved beta-cell 1527, 1989 RE, Pei D, Savage PJ, Bergman RN: A com- function, with reduction in secretion in 19. Konrad T, Vicini P, Kusterer K, Hoflich A, parison between the minimal model and intact and 32/33 split proinsulin, after di- Assadkhani A, Bohles HJ, Sewell A, the glucose clamp in the assessment of etary intervention in subjects with type 2 ␣ Tritschler HJ, Cobelli C, Usadel KH: -Li- insulin sensitivity across the spectrum of diabetes mellitus. Diabet Med 11:71–78, poic acid treatment decreases serum lac- glucose tolerance: Insulin Resistance Ath- 1994 tate and pyruvate concentrations and erosclerosis Study. Diabetes 43:1114– 30. Nagi DK, Ali VM, Yudkin JS: Effect of met- improves glucose effectiveness in lean and 1121, 1994 formin on intact proinsulin and des 31,32 obese patients with type 2 diabetes. Dia- 25. Haffner SM, Mykka¨nen L, Stern MP, Val- proinsulin concentrations in subjects betes Care 22:280–287, 1999 dez RA, Heisserman JA, Bowsher RR: Re- with non-insulin-dependent (type 2) dia- 20. Pacini G, Bergman RN: MINIMOD: a lationship of proinsulin and insulin to betes mellitus. Diabet Med 13:753–757, computer program to calculate insulin cardiovascular risk factors in nondiabetic 1996 sensitivity and pancreatic responsibility subjects. Diabetes 42:1297–1302, 1993 31. Davies MJ, Metcalfe J, Day JL, Grenfell A, from the frequently sampled intravenous 26. Yudkin JS, May M, Elwood P, Yarnell JW, Hales CN, Gray IP: Effect of sulphonylu- glucose tolerance test. Comput Methods Greenwood R, Davey SG: Concentrations rea therapy on plasma insulin, intact pro- Programs Biomed 23:113–122, 1986 of proinsulin like molecules predict coro- 21. Matthews DR, Hosker JP, Rudenski AS, nary heart disease risk independently of insulin and 32/33 split proinsulin in Naylor BA, Treacher DF, Turner RC: Ho- insulin: prospective data from the Caer- subjects with diabetes mellitus. Diabet meostasis model assessment: insulin re- philly study. Diabetologia 45:327–336, Med 10:293–298, 1993 sistance and ␤-cell function from fasting 2002 32. Maeda N, Takahashi M, Funahashi T, plasma glucose and insulin concentra- 27. Galloway JA, Hooper SA, Spradlin CT, Kihara S, Nishizawa H, Kishida K, Naga- tions in man. Diabetologia 28:412–419, Howey DC, Frank BH, Bowsher RR, retani H, Matsuda M, Komuro R, Ouchi 1985 Anderson JH: Biosynthetic human proin- N, Kuriyama H, Hotta K, Nakamura T, ␥ 22. Hedblad B, Nilsson P, Janzon L, Berglund sulin: review of chemistry, in vitro and in Shimomura I, Matsuzawa Y: PPAR li- G: Relation between insulin resistance vivo receptor binding, animal and human gands increase expression and plasma and carotid intima-media thickness and pharmacology studies, and clinical trial concentration of adiponectin, an adipose- stenosis in non-diabetic subjects: results experience. Diabetes Care 15:666–692, derived protein. Diabetes 50:2094–2099, from a cross-sectional study in Malmo, 1992 2001 Sweden. Diabet Med 17:299–307, 2000 28. Haffner SM, Mykka¨nen L, Festa A, Burke 33. Ukkola O: Resistin: a mediator of obesity- 23. Passing H, Bablok W: Comparison of sev- JP, Stern MP: Insulin-resistant prediabetic associated insulin resistance or an inno- eral regression procedures for method subjects have more atherogenic risk fac- cent bystander? Eur J Endocrinol 147:571– comparison studies and determination of tors than insulin-sensitive prediabetic 574, 2002

DIABETES CARE, VOLUME 27, NUMBER 3, MARCH 2004 687