Biomarkers may help clinicians predict cardiovascular risk. We aimed to determine if the addition of endocrine, metabolic, and obesity-associated biomarkers to conventional risk factors improves the prediction of cardiovascular and all-cause mortality.
In a population-based cohort study (the Study of Health in Pomerania) of 3,967 subjects (age 20–80 years) free of cardiovascular disease with a median follow-up of 10.0 years (38,638 person-years), we assessed the predictive value of conventional cardiovascular risk factors and the biomarkers thyrotropin; testosterone (in men only); insulin-like growth factor-1 (IGF-1); hemoglobin A1c (HbA1c); creatinine; high-sensitive C-reactive protein (hsCRP); fibrinogen; urinary albumin-to-creatinine ratio; and waist-to-height ratio (WHtR) on cardiovascular and all-cause death.
During follow-up, we observed 339 all-cause including 103 cardiovascular deaths. In Cox regression models with conventional risk factors, the following biomarkers were retained as significant predictors of cardiovascular death after backward elimination: HbA1c, IGF-1, and hsCRP. IGF-1 and hsCRP were retained as significant predictors of all-cause death.
For cardiovascular death, adding these biomarkers to the conventional risk factors changed the C-statistic from 0.898 to 0.910 (p = 0.02). The net reclassification improvement was 10.6%. For all-cause death, the C-statistic changed from 0.849 to 0.853 (P = 0.09).
Citation: Schneider HJ, Wallaschofski H, Völzke H, Markus MRP, Doerr M, Felix SB, et al. (2012) Incremental Effects of Endocrine and Metabolic Biomarkers and Abdominal Obesity on Cardiovascular Mortality Prediction. PLoS ONE 7(3): e33084. https://doi.org/10.1371/journal.pone.0033084
Editor: Christian Schulz, Heart Center Munich, Germany
Received: June 16, 2011; Accepted: February 9, 2012; Published: March 16, 2012
Copyright: © 2012 Schneider et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: SHIP is part of the Community Medicine Net (http://www.medizin.uni-greifswald.de/icm) of the University of Greifswald, which is funded by grants from the German Federal Ministry of Education and Research (BMBF, grant 01ZZ96030, 01ZZ0701); the Ministry for Education, Research, and Cultural Affairs; and the Ministry for Social Affairs of the Federal State of Mecklenburg–West Pomerania. The contributions to data collection made by field workers, study physicians, ultrasound technicians, interviewers, and computer assistants are gratefully acknowledged. Pfizer and Novo Nordisc provided partial grant support for the determination of serum samples (IGF-I or testosterone) and data analysis. Statistical analyses were partly supported by the German Research Foundation (DFG Vo 955/5-2). The funding sources were not involved in design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Competing interests: Harald J. Schneider received research grants from Pfizer, travel grants from Novartis, Pfizer, and Lilly, speaker fees from Novo Nordisk and Pfizer, and is a member of the German KIMS (Pfizer International Metabolic Survey) board, a scientific advisory board evaluating the effects of growth-hormone replacement in hypopituitarism sponsored by Pfizer. Henri Wallaschofski received research grants from Pfizer and Novo Nordisc and is member of the German KIMS (Pfizer International Metabolic Survey) board, a scientific advisory board evaluating the effects of growth-hormone replacement in hypopituitarism sponsored by Pfizer. Henry Völzke, Marcus Dörr, Marcelo Markus, Stephan B. Felix, Mathias Nauck, and Nele Friedrich report no conflict of interest. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.
Scoring systems based on classic risk factors, including sex, age, hypertension, dyslipidemia, and smoking, predict the future risk of cardiovascular events or death –; however, these risk factors explain only part of cardiovascular risk.
Thus, attempts have been made to improve the prediction of cardiovascular risk by adding multiple novel biomarkers to the classic risk factors. These biomarkers have included markers of inflammation, kidney function, cardiac damage, endothelial function, metabolism, and oxidative stress, among others –.
However, these novel markers' abilities to improve prediction were mostly disappointing. Many studies failed to detect a clinically relevant improvement in risk prediction , , . Biomarkers that appeared to improve prediction in some studies ,  did not perform well in other cohorts .
Differences in hormone levels are associated with cardiovascular risk. Both elevated and suppressed thyrotropin concentrations have been associated with increased cardiovascular mortality, though not consistently . Low levels of insulin-like growth factor-1 (IGF-1), a metabolic and anabolic hormone as well as a mediator of growth hormone action, have been associated with increased cardiovascular mortality in some studies , , but not in all . In men, low testosterone levels predict cardiovascular and all-cause mortality –. Hemoglobin A1c (HbA1c) predicted cardiovascular events in several studies, independently of diabetes –, although these results were not confirmed by another study .
In addition, recent studies suggest that measures of abdominal obesity, such as the waist-to-height ratio (WHtR), are associated with cardiovascular risk, independent of classic cardiovascular risk factors , . To our knowledge, no one has studied whether a multimarker approach that includes endocrine and metabolic parameters along with abdominal obesity will improve the predictive value of classic cardiovascular risk factors.
We hypothesized that a comprehensive set of easily assessed biomarkers that reflect different potential pathways of cardiovascular risk, including hormonal imbalance, glucose metabolism, abdominal obesity, inflammation, and kidney damage, adds incrementally to the use of conventional risk factors to predict cardiovascular and all-cause death.
Materials and Methods
The Study of Health in Pomerania (SHIP) is a longitudinal, representative, population-based cohort study in West Pomerania, a region in northeast Germany. Baseline data was collected from 1997 through 2001. A total of 4,308 subjects (response proportion: 69%) participated , . All participants gave written informed consent. The study conformed to the principles of the Declaration of Helsinki, and was approved by the Ethics Committee of the University of Greifswald.
Of the 4,308 participants, 18 pregnant women, 63 with incomplete conventional cardiovascular risk factor information and 260 with history of major cardiovascular events (stroke, myocardial infarction or heart surgery) at baseline were excluded, resulting in a study population of 3,967 subjects. Vital status information was acquired at annual intervals from the time of enrollment through December 2009. Subjects were censored at either death or failure to follow-up. The median duration of follow-up was 10.0 years (25th percentile 9.3; 75th percentile 10.0). Death certificates were coded by a certified nosologist according to the International Classification of Diseases, 10th revision (ICD10). Additionally, two internists (H.W. and M.D.) independently validated the underlying causes of death and performed a joint reading together with a third internist (H.V.) in cases of disagreement. Cardiovascular death included Codes I10 to I79.
Information on age, sex, and medical history was collected with computer-aided personal interviews. Smoking status was assessed by self-report, and subjects were categorized as either current smokers (at least one cigarette per day within the past year) or non-smokers. Anthropometric characteristics were measured according to written, standardized instructions in accordance with World Health Organization standards (WHO 1987). Waist circumference was measured to the nearest 0.1 cm midway between the lower rib margin and the iliac crest in the horizontal plane, using an inelastic tape measure. Blood pressure was measured three times with an appropriate-sized cuff after five minutes of rest in a sitting position, and the mean of the second and third measurement was recorded. The definition of diabetes mellitus was based on self-reported physician's diagnosis or self-reported use of antidiabetic medication in the last seven days. The history of cardiovascular disease (CVD) was based on a self-reported history of myocardial infarction, stroke or cardiac surgery.
Non-fasting blood samples were drawn from the cubital vein, with the patient in the supine position. A urine sample was collected. The samples were taken between 07:00 a.m. and 04:00 p.m. and analyzed immediately or stored at −80°C until biomarkers could be measured. In addition, internal quality controls were performed at least daily.
Urine samples were stored for a maximum of 2 days until measurement. Serum creatinine levels were determined with the Jaffé method (Hitachi 717; Roche Diagnostics, Germany). The urinary albumin concentration was determined with a Behring Nephelometer (Siemens BN albumin; Siemens Healthcare, Marburg, Germany). Total and high-density lipoprotein (HDL) cholesterol were measured photometrically (Hitachi 704; Roche, Mannheim, Germany).
Serum IGF-1 was determined with automated two-site chemiluminescence immunoassays (Nichols Advantage; Nichols Institute Diagnostica GmbH, Bad Vilbel, Germany). Total testosterone levels were measured with competitive chemiluminescent enzyme immunoassays on an Immulite 2500 analyzer (Siemens Immulite 2500 Total Testosterone, ref. L5KTW, Lot 110; Siemens Healthcare Medical Diagnostics, Bad Nauheim, Germany). Serum thyrotropin levels were measured with immunochemiluminescence (Byk Sangtec Diagnostica GmbH, Frankfurt, Germany). HbA1c levels were determined with high-performance liquid chromatography (Bio-Rad Diamat, Munich, Germany). Plasma fibrinogen concentrations were assayed as described by Clauss (19), using an Electra 1600 analyzer (Instrumentation Laboratory, Barcelona, Spain). HsCRP was determined immunologically on a Behring Nephelometer II with commercially available reagents from Dade Behring (Dade Behring, Eschborn, Germany).
Categorical data were expressed as percentages; continuous data were expressed as medians (25th percentile; 75th percentile). Continuous variables were truncated at the 1st and 99th percentile. Univariate analysis was performed, with χ2 testing for categorical variables and Mann–Whitney U-tests for continuous distributions. For regression analyses, skewed variables were log-transformed.
We included the following parameters as conventional risk factors: age (continuous); sex (binary); systolic blood pressure (continuous); antihypertensive medication (binary); HDL cholesterol (continuous); total cholesterol (continuous); diabetes (binary); and current smoking (binary). We assessed the following biomarkers as continuous variables, analyzing the effects of an increase of one standard deviation (SD) from the mean: thyrotropin, IGF-1, testosterone, hsCRP, fibrinogen, HbA1c, creatinine, urinary albumine-to-creatinine ratio (UACR), and WHtR. Additionally, we analyzed cutoffs that were considered clinically useful and were derived from the literature for the following variables: thyrotropin below and above the reference range versus the reference range (0.25–2.12 mIU/l) ; IGF-1 below the 10th sex- and age-specific percentile ; and testosterone below 10.4 nmol/l . Testosterone was measured in men only, and all analyses that included testosterone were performed in men only.
We proceeded in four steps to assess the prediction of cardiovascular death. First, we performed Cox proportional hazards regression analyses for single biomarkers, unadjusted and adjusted for conventional risk factors, to predict cardiovascular death. Second, we included all biomarkers that were significant after adjustment into a prediction model, using backward elimination with conventional risk factors forced into the model. Third, we used the C-statistic, as described by Pencina et al. , to compare the predictive values of conventional risk factors with single, novel risk factors and the models built in the second step . Finally, we classified subjects into groups with low (<2%), moderate (2% to 9%) and high (>9%) risk, based on the recommendations of the European Society of Cardiology , and calculated the net reclassification improvement (NRI) for cardiovascular death. The bias-corrected accelerated bootstrap resampling procedure of Efron and Tibshirani was used to obtain 95% confidence intervals for NRIs. Observed risk at 10-years estimated from Kaplan-Meier curve were used to estimate the expected number of subjects who died and who did not die at 10-year follow-up.
In secondary analyses, we repeated the first three steps with all-cause death as outcome. A two-sided p-value of <0.05 was considered statistically significant. Statistical analyses were performed with SAS 9.1 (SAS Institute Inc., Cary, NC, USA).
Baseline characteristics are shown in table 1. During the 38,638 person-years of follow-up, we observed 339 (8.6%) all-cause deaths (rate per 1,000 person years: 8.8), including 103 (2.6%) deaths due to cardiovascular disease (rate per 1,000 person years: 2.7).
Prediction of events by single biomarkers
The results of the Cox regression analyses for predicting cardiovascular death with single biomarkers are shown in table 2. WHtR, HBA1c, testosterone below 10.4 nmol/l, hsCRP, and IGF-1 levels below the 10th sex- and age-specific percentile remained significant predictors after adjustment for conventional risk factors.
For all-cause death, WHtR, hsCRP, fibrinogen, IGF-1 levels below the 10th percentile, and testosterone below 10.4 nmol/l remained significant predictors after adjustment for conventional risk factors (table 3).
Risk prediction with combined biomarkers and risk factors
Table 4 shows the results of backward elimination models that included all significant biomarkers from the previous step with conventional risk factors forced into the model for cardiovascular death. HsCRP, HbA1c, and low IGF-1 remained significant predictors after backward elimination. In men, low testosterone was not retained in the model after backward elimination. Therefore, we did not continue with separate analyses by sex. HsCRP and low IGF-1 remained significant predictors for all-cause death (table 5).
Table 6 shows the C-statistics for cardiovascular death. Adding biomarkers to all conventional risk factors lead to a significant (p = 0.02) but slight change in the C-statistic for cardiovascular death from 0.898 (95%-CI 0.873–0.923) to 0.910 (95%-CI 0.886–0.934). For all-cause death, the C-statistic changed (p = 0.09) from 0.849 (95%-CI 0.830–0.868) to 0.853 (95%-CI 0.833–0.872) after biomarkers were included (table 7). The reclassification for cardiovascular death was 10.61% (95%-CI −0.28–24.86) as shown in table 8.
In this population-based cohort study, we analyzed whether combining endocrine and metabolic parameters with a marker of abdominal obesity improved the prediction of cardiovascular death with conventional risk factors. We identified a set of additional biomarkers, including low IGF-1, HbA1c, and hsCRP that caused a moderate but significant improvement in the prediction of cardiovascular deaths and correctly reclassified 11% of the population at risk. The predictive value was stronger for cardiovascular than for all-cause death. This result is not surprising because the risk factors selected were aimed at cardiovascular risk and not at other causes of death.
Our results confirm the findings of previous studies showing that hsCRP is a significant predictor of cardiovascular risk in a multimarker approach –. Our results extend the findings of previous reports by assessing the associations of HbA1c, and low IGF-1 in this context.
HbA1c was included in one multimarker study, but only in subjects with diabetes . Several studies focusing on HbA1c found a predictive role independent of diabetes [19–219]. We confirmed this role in combination with other biomarkers.
Testosterone in men and thyrotropin did not improve prediction. Although low levels of testosterone have been shown to predict mortality even after adjustment for cardiovascular risk factors , the fact that the associations became weaker after adjustment in our study suggests that the effects of testosterone are mediated by cardiovascular risk factors. This result is consistent with previous studies showing that testosterone is associated with cardiovascular risk factors , . Similar associations may play a role in the results found for thyrotropin .
Low IGF-1 was a significant predictor. Experimental data show IGF-1's protective effects against systemic inflammation, insulin resistance, and free fatty acid production, all of which are consequences of obesity . IGF-1 remained a significant predictor among parameters that reflect these different pathways. This finding suggests that additional parameters are involved in the association between IGF-1 and cardiovascular death. Possibly, the effects of IGF-1 on endothelial function and cardiac growth and function ,  or other, unknown factors play a part.
Previous studies have shown that WHtR is a strong and linear predictor of cardiovascular risk, compared with other measures of abdominal and overall obesity , ,  whereas BMI shows a U-shaped association with cardiovascular risk and mortality –. WHtR was not significant in this multimarker approach. Most likely, this is due to the fact that abdominal obesity, as measured by the WHtR, increases risk by promoting conventional cardiovascular risk factors.
The identification of biomarkers to improve the ability to predict cardiovascular risk has been a focus of interest in recent years. Our results show a modest improvement in prediction compared with other studies –. The inclusion of IGF-1 and HbA1c may have contributed to this effect. Other studies found higher incremental effects of biomarkers than our study , . Selection of high-risk patients or different outcomes might play a part in these findings.
It has been shown that biomarkers can result in relevant reclassification, even in the absence of significant increases in C-statistics . In our study, 11% of subjects were correctly net reclassified by the new biomarkers. Generally, one can expect that the lower the costs for a biomarker are the greater is the willingness to accept small increments in discrimination.
We selected biomarkers that can be easily obtained in every-day clinical practice. Therefore, the increment in our study, though only moderate, might be relevant if validated, given the low cost and effort associated with biomarker assessment – HbA1c is a routine laboratory parameter and hsCRP is increasingly used in risk assessment. IGF-1 is mainly used as marker of pathological changes in growth hormone secretion in an endocrine setting. All of the biomarkers we measured can be assessed clinically or in blood and do not require fasting.
Several limitations need to be addressed. First, because we aimed to explore new biomarkers, our findings need validation in additional cohorts. We did not test other potentially relevant markers, such as troponin or B-type natriuretic peptide. We expect that these factors are more relevant in populations with higher baseline risks of cardiac damage than our sample had. As in any study of risk prediction, multiple testing is an issue, although we tried to minimize the number of tests we administered. In addition, we only assessed cardiovascular and all-cause mortality. We do not know if our findings are generalizable to other cardiovascular events.
In summary, we found a set of biomarkers that improve the prediction of cardiovascular death. The effort required to assess these parameters is low. Thus, the increase in the C-statistic, though only moderate, and the net reclassification of 11% could have a clinical benefit, once validated.
Conceived and designed the experiments: HJS NF HW. Performed the experiments: HV HW MD MN. Analyzed the data: HJS HW HV MM MD SBF MN NF. Contributed reagents/materials/analysis tools: HV HW. Wrote the paper: HJS. Critically revised manuscript for important intellectual content: HJS HW HV MM MD SBF MN NF.
- 1. Anderson KM, Wilson PW, Odell PM, Kannel WB (1991) An updated coronary risk profile. A statement for health professionals. Circulation 356–362.
- 2. Expert Panel on DetectionEvaluationTreatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) (2001) Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP). JAMA 2486–2497.
- 3. Assmann G, Cullen P, Schulte H (1998) The Munster Heart Study (PROCAM). Results of follow-up at 8 years. Eur Heart J 19: Suppl AA2–11.
- 4. Conroy RM, Pyorala K, Fitzgerald AP, Sans S, Menotti A, et al. (2003) Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur Heart 987–1003.
- 5. Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, et al. (1998) Prediction of coronary heart disease using risk factor categories. Circulation 1837–1847.
- 6. Blankenberg S, McQueen MJ, Smieja M, Pogue J, Balion C, et al. (2006) Comparative impact of multiple biomarkers and N-Terminal pro-brain natriuretic peptide in the context of conventional risk factors for the prediction of recurrent cardiovascular events in the Heart Outcomes Prevention Evaluation (HOPE) Study. Circulation 201–208.
- 7. Blankenberg S, Zeller T, Saarela O, Havulinna AS, Kee F, et al. (2010) Contribution of 30 biomarkers to 10-year cardiovascular risk estimation in 2 population cohorts: the MONICA, risk, genetics, archiving, and monograph (MORGAM) biomarker project. Circulation 2388–2397.
- 8. Melander O, Newton-Cheh C, Almgren P, Hedblad B, Berglund G, et al. (2009) Novel and conventional biomarkers for prediction of incident cardiovascular events in the community. JAMA 49–57.
- 9. Ridker PM, Buring JE, Rifai N, Cook NR (2007) Development and validation of improved algorithms for the assessment of global cardiovascular risk in women: the Reynolds Risk Score. JAMA 611–619.
- 10. Wang TJ, Gona P, Larson MG, Tofler GH, Levy D, et al. (2006) Multiple biomarkers for the prediction of first major cardiovascular events and death. N Engl J Med 2631–2639.
- 11. Zethelius B, Berglund L, Sundstrom J, Ingelsson E, Basu S, et al. (2008) Use of multiple biomarkers to improve the prediction of death from cardiovascular causes. N Engl J Med 2107–2116.
- 12. Ochs N, Auer R, Bauer DC, Nanchen D, Gussekloo J, et al. (2008) Meta-analysis: subclinical thyroid dysfunction and the risk for coronary heart disease and mortality. Ann Intern Med 832–845.
- 13. Laughlin GA, Barrett-Connor E, Criqui MH, Kritz-Silverstein D (2004) The prospective association of serum insulin-like growth factor I (IGF-I) and IGF-binding protein-1 levels with all cause and cardiovascular disease mortality in older adults: the Rancho Bernardo Study. J Clin Endocrinol Metab 114–120.
- 14. Friedrich N, Haring R, Nauck M, Lüdemann J, Rosskopf D, et al. (2009) Mortality and serum insulin-like growth factor (IGF)-I and IGF binding protein 3 concentrations. J Clin Endocrinol Metab 1732–1739.
- 15. Saydah S, Graubard B, Ballard-Barbash R, Berrigan D (2007) Insulin-like growth factors and subsequent risk of mortality in the United States. Am J Epidemiol 518–526.
- 16. Haring R, Völzke H, Steveling A, Krebs A, Felix SB, et al. (2010) Low serum testosterone levels are associated with increased risk of mortality in a population-based cohort of men aged 20–79. Eur Heart J 1494–1501.
- 17. Shores MM, Matsumoto AM, Sloan KL, Kivlahan DR (2006) Low serum testosterone and mortality in male veterans. Arch Intern Med 1660–1665.
- 18. Laughlin GA, Barrett-Connor E, Bergstrom J (2008) Low serum testosterone and mortality in older men. J Clin Endocrinol Metab 68–75.
- 19. Adams RJ, Appleton SL, Hill CL, Wilson DH, Taylor AW, et al. (2009) Independent association of HbA(1c) and incident cardiovascular disease in people without diabetes. Obesity 559–563.
- 20. Gerstein HC, Swedberg K, Carlsson J, McMurray JJ, Michelson EL, et al. (2008) The hemoglobin A1c level as a progressive risk factor for cardiovascular death, hospitalization for heart failure, or death in patients with chronic heart failure: an analysis of the Candesartan in Heart failure: Assessment of Reduction in Mortality and Morbidity (CHARM) program. Arch Intern Med 1699–1704.
- 21. Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, et al. (2001) Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of european prospective investigation of cancer and nutrition (EPIC-Norfolk). BMJ 15–18.
- 22. Chonchol M, Katz R, Fried LF, Sarnak MJ, Siscovick DS, et al. (2010) Glycosylated hemoglobin and the risk of death and cardiovascular mortality in the elderly. Nutr Metab Cardiovasc Dis 15–21.
- 23. Gelber RP, Gaziano JM, Orav EJ, Manson JE, Buring JE, et al. (2008) Measures of obesity and cardiovascular risk among men and women. J Am Coll Cardiol 2008; 52: 605–615.
- 24. Schneider HJ, Friedrich N, Klotsche J, Pieper L, Nauck M, et al. (2010) The predictive value of different measures of obesity for incident cardiovascular events and mortality. J Clin Endocrinol Metab 1777–1785.
- 25. John U, Greiner B, Hensel E, Ludemann J, Piek M, et al. (2001) Study of Health In Pomerania (SHIP): a health examination survey in an east German region: objectives and design. Soz Praventivmed 186–194.
- 26. Völzke H, Alte D, Schmidt CO, Radke D, Lorbeer R, et al. (2011) Cohort Profile: The Study of Health in Pomerania. Int J Epidemiol 294–307.
- 27. Völzke H, Alte D, Kohlmann T, Ludemann J, Nauck M, et al. (2005) Reference intervals of serum thyroid function tests in a previously iodine-deficient area. Thyroid 279–285.
- 28. Pencina MJ, D'Agostino RB (2004) Overall C as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation. Stat Med 2109–2123.
- 29. Newson RB (2010) Comparing the predictive power of survival models using Harrell's C or Somers' D. The Stata Journal 339–358.
- 30. Haring R, Volzke H, Felix SB, Schipf S, Dorr M, et al. (2009) Prediction of metabolic syndrome by low serum testosterone levels in men: Results from the Study of Health in Pomerania. Diabetes 2027–2031.
- 31. Schneider HJ, Sievers C, Klotsche J, Bohler S, Pittrow D, et al. (200) Prevalence of low male testosterone levels in primary care in Germany: cross-sectional results from the DETECT study. Clin Endocrinol 446–454.
- 32. Volzke H, Schwahn C, Wallaschofski H, Doerr M (2007) Review: The association of thyroid dysfunction with all-cause and circulatory mortality: is there a causal relationship? J Clin Endocrinol Metab 2421–2429.
- 33. Rajpathak SN, Gunter MJ, Wylie-Rosett J, Ho GY, Kaplan RC, et al. (200) The role of insulin-like growth factor-I and its binding proteins in glucose homeostasis and type 2 diabetes. Diabetes Metab Res Rev 3–12.
- 34. Juul A (2003) Serum levels of insulin-like growth factor I and its binding proteins in health and disease. Growth Horm IGF Res 113–170.
- 35. Empen K, Lorbeer R, Völzke H, Robinson DM, Friedrich N, et al. (2010) Association of serum insulin-like growth factor I with endothelial function: Results from the population-based Study of Health in Pomerania (SHIP). Eur J Endocrinol 617–623.
- 36. Schneider HJ, Klotsche J, Silber S, Stalla GK, Wittchen HU (2011) Measuring abdominal obesity: effects of height on distribution of cardiometabolic risk factors risk using waist circumference and waist-to-height ratio. Diabetes Care 34: e7.
- 37. Flegal KM, Graubard BI, Williamson DF, Gail MH (2007) Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2028–2037.
- 38. Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, et al. (2006) Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet 666–678.
- 39. Gruberg L, Weissman NJ, Waksman R, Fuchs S, Deible R, et al. (2002) The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity paradox? J Am Coll Cardiol 578–584.