Association of IGF-I and IGFBP-3 with health care costs and hospitalization: Results from a prospective observational study Sebastian E. Baumeister a, ⁎ ,1 , Nele Friedrich b,1 , Carsten Oliver Schmidt a , Henry Völzke a , Matthias Nauck b , Wolfgang Hoffmann a , Stefan Fleßa c , Paul Marschall c , Henri Wallaschofski b,1 a Institute for Community Medicine, University of Greifswald, Germany b Institute of Clinical Chemistry and Laboratory Medicine, University of Greifswald, Germany c Faculty of Law and Economics, University of Greifswald, Germany abstract article info Article history: Received 4 November 2010 Received in revised form 27 January 2011 Accepted 9 February 2011 Available online 9 March 2011 Keywords: IGF-I IGFBP-3 Health care cost Hospitalization Objective: Previous cohort studies found robust associations of serum insulin-like growth factor I (IGF-I) and its main binding protein IGFBP-3 with increased morbidity or mortality. This study investigates the relationships between IGF-I and IGFBP-3 with health care costs and hospitalization in a general population and whether adding IGF-I or IGFBP-3 to a model of established health care predictors improves prediction. Methods: Data from a population-based cohort study of 3139 men and women in Germany, aged 20 to 80 years at baseline were used (median follow-up time: 5.0 years). Self-reported physician visits, length of hospital stay were used to estimate annual costs. IGF-I and IGFBP-3 were categorized at the 10th and 90th percentile, to indicate ‘low’, ‘intermediate’, and ‘high’ concentrations, respectively. Results: Total annual health care costs, with the major component of inpatient costs, and risk of hospitalization at baseline and follow-up were higher in subjects with low compared to intermediate IGF-I or IGFBP-3, after multivariable-adjustment. Subjects with low in contrast to intermediate IGF-I exhibited 30.6% higher annual total costs 5 years after baseline examination, corresponding to a difference in adjusted costs of EUR 436.61. Conclusions: Low IGF-I and IGFBP-3 independently predict future health care costs and hospitalization. IGF-I or IGFBP-3 might be useful to identify subjects with excess health care use. The predictive performance of cross- sectional and longitudinal models of total and inpatient costs were slightly improved by adding IGF-I or IGFBP-3 but the cost-effectiveness of inclusion into prediction models needs to be examined. © 2011 Growth Hormone Research Society. Published by Elsevier Ltd. All rights reserved. 1. Introduction The insulin-like growth factor I (IGF-I), which is mostly carried by IGF binding protein 3 (IGFBP-3), is generally accepted as a key mediator of metabolic, endocrine, and finally anabolic effects of the growth hormone (GH) [1]. During the last decades, both disorders of the GH/IGF-I axis, GH deficiency (GHD) characterized by low serum IGF-I values, as well as acromegaly accompanied by IGF-I oversecre- tion, have been related to increased morbidity [2] and excess mortality risk [3,4] as well as impaired quality of life and functional status [5]. Two large population-based cohorts revealed that low- normal IGF-I concentrations are related to ischemic heart disease mortality and all-cause mortality [6,7]. The Study of Health in Pomerania (SHIP) also detected inverse associations between IGF-I or IGFBP-3 concentrations and mortality from all causes, cardiovas- cular disease or cancer in men and between IGFBP-3 and all-cause mortality in women [8]. Another study showed that elevated IGF-I levels were associated to higher all-cause mortality [7]. A recent study in elderly supports a U-shaped relationship for IGF-I concentrations and all-cause mortality and cardiovascular mortality [9]. Previous studies also identified associations between low IGF-I concentrations and increased risk of congestive heart failure and ischemic heart disease [10,11]. Furthermore, studies have shown that high concen- trations of IGF-I and low concentrations of IGFBP-3 were related to components of the metabolic syndrome [12–14]. In addition, high- normal levels of IGF-I have been reported to be associated with increased risk of cancer [15,16]. IGF-I or the IGF-I:IGFBP-3 ratio has therefore been suggested to be useful intermediate or surrogate markers of sub-clinical disease progression [17,18]. Growth Hormone & IGF Research 21 (2011) 89–95 Abbreviations: ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ATC, Anatomic therapeutic classification; AUC, Area under the curve; CI, Confidence interval; CKD, Chronic kidney disease; EUR, Euro; GGT, Gamma glutamyl transpepti- dase; GH, Growth hormone; GHD, GH deficiency; HDL-C, High-density lipoprotein cholesterol; IGF-I, Insulin-like growth factor I; IGFBP-3, IGF binding protein 3; MAPE, Mean absolute prediction error; R 2 , R-squared; RR, Relative risk; SD, Standard deviation; SHIP, Study of Health in Pomerania; TC, Total cholesterol; U.S., United States; WC, Waist circumference. ⁎ Corresponding author at: Institute for Community Medicine, University of Greifswald, Walther Rathenau Str. 48, D-17487 Greifswald, Germany. Tel.: +49 3834 8619573; fax: +49 3834 866684. E-mail address: sebastian.baumeister@uni-greifswald.de (S.E. Baumeister). 1 Contributed equally. 1096-6374/$ – see front matter © 2011 Growth Hormone Research Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ghir.2011.02.001 Contents lists available at ScienceDirect Growth Hormone & IGF Research journal homepage: www.elsevier.com/locate/ghir