Diabetologia (2003) 46:20–26 DOI 10.1007/s00125-002-0995-2 Articles Proinsulin and acute insulin response independently predict Type 2 diabetes mellitus in men – report from 27 years of follow-up study B. Zethelius1, L. Byberg1, C. N. Hales3, H. Lithell1, C. Berne2 1 Department of Public Health and Caring Sciences / Section of Geriatrics, Uppsala University, Uppsala, Sweden 2 Department of Medical Sciences, Uppsala University, Uppsala, Sweden 3 Department of Clinical Biochemistry, Addenbrooke’s Hospital, Cambridge, UK Abstract predictor variables, were analysed with logistic re- gression. Results are shown as odds ratios (ORs) with Aims/hypothesis. Defects in insulin secretion and insu- their 95% confidence intervals (CIs) for a one stan- lin action, resulting in compensatory hyperinsulinae- dard deviation increase in a predictor variable. mia, are the major abnormalities in the development Results. Cumulative incidence of Type 2 diabetes was of Type 2 diabetes mellitus (Type 2 diabetes). The 33% over 27 years of follow-up. Intact proinsulin most frequently used conventional immunoreactive (OR, 1.57, CI, 1.16–2.14), and 32–33 split proinsulin assays for insulin cross-react with proinsulin. In short- (OR, 1.70, CI, 1.20–2.39) were associated with devel- term studies (<5 years), proinsulin predicts the devel- opment of Type 2 diabetes, independent of AIR, ad- opment of Type 2 diabetes. We studied, with a 27-year justed for BMI and fasting glucose, whereas specific follow-up, the longitudinal relationships between in- insulin was not (OR, 1.31, CI, 0.98–1.77), nor was tact proinsulin, 32–33 split proinsulin, specific and IRI (OR, 1.25, CI, 0.96–1.63). Proinsulin and AIR immunoreactive insulin (IRI), acute insulin response interacted in the development of Type 2 diabetes (AIR) after an intravenous glucose load and the devel- (p<0.05). opment of Type 2 diabetes in a population-based co- Conclusion/interpretation. Proinsulin predicts the de- hort of 50-year-old men. velopment of Type 2 diabetes mellitus over a 27-year Methods. Fasting peptide concentrations were mea- period. [Diabetologia (2003) 46:20–26] sured in plasma samples, stored since 1970–73 using specific two-site immunometric assays. IRI was mea- Keywords Proinsulin, insulin, immunoreactive, body sured at baseline using radioimmunoassay. Associa- mass index, epidemiology, risk factors, prospective tions between development of Type 2 diabetes and study. Defects in insulin secretion and insulin action, result- ry dysfunction and insulin resistance, analysed by dif- ing in compensatory hyperinsulinaemia, are the major ferent clamp techniques, were independent predictors abnormalities in the development of Type 2 diabetes of worsening of glucose tolerance [3]. mellitus [1, 2]. It has been shown that insulin secreto- In several prospective studies, impaired insulin se- cretion assessed by a low acute insulin response (AIR) Received: 18 June 2002 / Revised: 20 August 2002 to a glucose load, was shown to be a predictor of the Published online: 8 January 2003 development of Type 2 diabetes [2, 4, 5, 6]. Further- © Springer-Verlag 2003 more, fasting hyperinsulinaemia indicative of insulin resistance is a predictor of Type 2 diabetes [6, 7, 8]. Corresponding author: B. Zethelius MD, Department of Pub- However, this association is less strong in the impaired lic Health and Caring Sciences / Section of Geriatrics, Uppsala University, Box 609, 75125 Uppsala, Sweden, E-mail: bjorn. than in the normal glucose tolerance state [9, 10]. A [email protected] complication in earlier studies is that conventional Abbreviations: AIR acute insulin response; IRI immunoreactive radioimmunoassays for insulin, measuring immunore- insulin. active insulin (IRI), cross-react with proinsulin. B. Zethelius et al.: Proinsulin and acute insulin response independently predict Type 2 diabetes mellitus in men 21 Proinsulin has been shown to predict the onset of and 1984 (1860 subjects participated) [21], at age 70 years Type 2 diabetes in subjects with normal or impaired (U-70) between 1991 and 1995 (1221 out of 1681 eligible men glucose tolerance in studies with follow-up time less participated) [22] and at age 77 years (U-77) between 1998 and 2001 (839 out of 1079 eligible men participated). than 7 years [11, 12, 13, 14, 15, 16]. Both proinsulin Type 2 diabetes was defined according to the World Health and specific insulin are inversely associated with insu- Organisation (WHO) criteria from 1999 [23] (using fasting lin sensitivity and positively associated with the AIR concentrations) and/or use of OHAs. At the first two investiga- in subjects with normal glucose tolerance [17] and in tions (U-50 and U-60), fasting blood glucose concentrations subjects with Type 2 diabetes [18]. The roles of proin- were analysed and Type 2 diabetes was thus defined as a sulin, insulin resistance and impaired insulin secretion fasting blood glucose concentration greater than or equal to in the development of Type 2 diabetes over longer 6.1 mmol/l and/or treatment with OHAs. At the third and fourth investigations (U-70 and U-77), fasting plasma glucose periods of time are not clear. concentrations were analysed and Type 2 diabetes was defined The primary aim of this study was to investigate as a fasting plasma glucose concentration greater than or equal the baseline fasting plasma concentrations of intact to 7.0 mmol/l and/or treatment with OHAs. The cumulative proinsulin, 32–33 split proinsulin, specific insulin and incidence of Type 2 diabetes was calculated, from age 50 years IRI, for the development of Type 2 diabetes and to es- to age 60 years as Type 2 diabetes identified at U-60, from age tablish the role of AIR at an IVGTT in these associa- 50 to age 70 as Type 2 diabetes identified at U-60 or U-70 and from age 50 years to age 77 years as Type 2 diabetes identified tions. Further, possible significant predictors by out- at U-60, U-70 or U-77. come associations were investigated in terms of differ- To avoid potential bias by using only fasting glucose con- ent length of follow-up, i.e. development of Type 2 centrations for the diagnosis of Type 2 diabetes, an additional diabetes 10, 20 and 27 years later. analysis was carried out using data from a 75-gram OGTT, which was carried out in all participants in U-70. Type 2 diabe- tes was defined as a fasting plasma glucose greater than or equal to 7.0 mmol/l or a 2-h glucose glucose greater than or Subjects and methods equal to 11.1 mmol/l, according to WHO criteria from 1999 [23] and/or treatment with OHAs. Study design. The Uppsala Longitudinal Study of Adult Men is a population-based cohort study of diabetes and cardiovascular Baseline characteristics and laboratory methods. The concen- disease in men [6, 19]. In 1970, all men born between 1920 trations of intact and 32–33 split proinsulin were analysed us- and 24 and residing in Uppsala were invited to a health survey ing the two-site immunometric assay technique [24], in plasma at age 50 years (U-50), between 1970 and 1973 (baseline). In samples that had been stored frozen (−70°C) since baseline. all, 2322 of 2841 invited men participated (82%). Plasma sam- Specific insulin concentrations were also analysed in these ples have been stored since that survey, allowing us to study samples using the Access Immunoassay System (Sanofi Pas- proinsulin, using highly specific assays, as a possible predictor teur Diagnostics, Marmes-la-Coquette, France), using a one- for the development of Type 2 diabetes in non-diabetic sub- step chemiluminescent immunoenzymatic assay. These proce- jects at baseline, using the follow-up data over 27 years. dures have been described in detail previously and the long- Due to a freezer failure causing the random loss of some term storage did not interfere with the assessments in plasma 1000 samples, concentrations of intact and 32–33 split proin- that had been stored for 25 years [20]. sulin and true insulin were analysed in baseline plasma sam- At baseline, the serum IRI concentrations were analysed ples from 1306 subjects. There were no differences in baseline with the Phadebas Insulin Test (Pharmacia AB, Uppsala, characteristics other than propeptides, between those with and Sweden) [19, 21]. those without their propeptide concentrations analysed, indi- An IVGTT, using a glucose dose of 0.5 g/kg body weight, cating that no selection bias had occurred due to the loss of was carried out in 1792 men at baseline [19]. The serum IRI samples [20]. concentrations during the IVGTT were measured in duplicate blood samples drawn before and at 4 and 6 minutes after the Subjects. This study was based on those subjects who were start of the glucose injection. The AIR was expressed as non-diabetic at baseline and had complete data from the base- the mean value of the serum insulin concentrations at 4 and line investigation and proinsulin and specific insulin assess- 6 minutes. ments and follow-up data from the 10-year follow-up (n=777), Blood glucose was analysed, at U-50 and U-60 using spec- from the 20-year follow-up (n=436), and from the 27-year fol- trophotometry and the glucose oxidase method [19, 21]. At U-70 low-up (n=348). In order to select non-diabetic men at base- and U-77 plasma glucose was analysed by the glucose dehydro- line, subjects with a fasting blood glucose greater than or equal genase method (Gluc-DH, Merck, Darmstadt, Germany). to 6.1 mmol/l or treatment with oral hypoglycaemic agents Body mass index (BMI) was calculated as weight (in un- (OHAs) were excluded. Subjects with insulin treatment only, derwear) / height squared (kg/m2).