International Journal of Cardiology 171 (2014) e24–e25
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Letter to the Editor
Serum bilirubin levels and risk of prediabetes in young and healthy adults☆ Matthias Bossard a,b,1,2, Stefanie Aeschbacher b,1,2, Tobias Schoen b,2, Thomas Hochgruber b,2, Mirco von Rotz b,2, Jonas Blum b,2, Martin Risch c,d,2, Lorenz Risch c,e,f,2, David Conen b,⁎,2 a
Cardiology Division, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland Department of Medicine, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland Labormedizinisches Zentrum Dr. Risch, Schaan, Principality of Liechtenstein d Division of Laboratory Medicine, Kantonsspital Graubünden, Chur, Switzerland e Division of Clinical Biochemistry, Medical University, Innsbruck, Austria f Private University, Triesen, Principality of Liechtenstein b c
a r t i c l e
i n f o
Article history: Received 10 September 2013 Accepted 30 November 2013 Available online 6 December 2013 Keywords: Prediabetes Total serum bilirubin Oxidative stress Reactive oxygen species Epidemiology Biomarker
Inflammation and oxidative stress have been implicated in the development of diabetes mellitus, which is one of the leading risk factors for cardiovascular diseases [1–4]. In this context, the antioxidant and anti-inflammatory effects of bilirubin, a breakdown product of the hemoglobin catabolism, have been increasingly recognized in recent years [5,6]. Serum bilirubin concentrations appear to be inversely related to cardiovascular risk factors [7,8], and have been associated with improved insulin sensitivity in animal models [8]. In this analysis, we addressed several unanswered questions regarding these relationships. First, it is relatively unclear whether inflammation and oxidative stress are an early or late phenomenon in
☆ Funding: The Liechtenstein Government, the Swiss Heart Foundation, the Swiss Society of Hypertension, the University of Basel, the University Hospital Basel, the Hanela Foundation, Schiller AG and Novartis supported the GAPP study. David Conen was supported by a grant from the Swiss National Science Foundation (PP00P3_133681). ⁎ Corresponding author at: Department of Medicine, University Hospital, Petersgraben 4, 4031 Basel, Switzerland. Tel.: +41 61 328 66 96; fax: +41 61 265 57 34. E-mail address: [email protected] (D. Conen). 1 M.B. and S.A. contributed equally to this manuscript. 2 All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. 0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2013.11.125
diabetogenesis. Second, biomarkers associated with early changes in glucose homeostasis may improve diabetes risk prediction in the general population. Third, prior studies have been mainly performed in non-European populations, and the generalizability to Western populations remains unclear [7,8]. Finally, possible gender differences for this relationship have not been thoroughly evaluated [8]. We performed a cross-sectional study of healthy adults aged 25–41 years, participating in the ‘Genetic and phenotypic determinants of blood pressure and other cardiovascular risk factors’ (GAPP) study in the Principality of Liechtenstein [9]. Individuals with known diabetes, body mass index (BMI) N 35 kg/m2 and prevalent cardiovascular disease were excluded [9]. Informed consent was obtained from every participant. The local ethics committee approved the study protocol. Prediabetes was defined as an HbA1c level between 5.7 and 6.4% [1,9]. Information about personal, lifestyle (including diet, smoking and physical activity) and medical factors were obtained by questionnaire. Height, weight, waist and hip circumference and blood pressure were measured using standardized methodology. All biomarkers used in this study were assayed as described in detail previously [9]. Multivariable logistic regression models were constructed to compare the odds of prediabetes across sex-specific quartiles of total bilirubin and to adjust for potential confounders. Linear trends were calculated using the sex-stratified, quartile-specific median bilirubin values. For subgroup analyses, logistic regression models were stratified by age, sex, BMI, hypertension and smoking status. Differences across subgroups were tested by including bilirubin by subgroup interaction terms in the fully adjusted non-stratified models. Baseline characteristics of the 1758 participants stratified by the presence or absence of prediabetes are shown in Table 1. Serum bilirubin levels were 8.6 (6.8; 12.0) and 10.3 (6.8; 13.7) umol/l among those with and without prediabetes (p b 0.0001, Table 1). Multivariable logistic regression models of prediabetes risk across sex-specific quartiles of total serum bilirubin levels confirmed an inverse relationship between bilirubin and prediabetes, as shown in Table 2. Sensitivity analysis using waist to hip ratio instead of BMI in the fully adjusted model provided very similar results with ORs (95%CI) of 1.0 (reference), 0.83 (0.59; 1.17); 0.57 (0.41; 0.78) and 0.52 (0.37; 0.72) across increasing quartiles of bilirubin. Similar findings were also obtained when hemoglobin was added as a potential confounder to the fully adjusted
M. Bossard et al. / International Journal of Cardiology 171 (2014) e24–e25 Table 1 Baseline characteristics stratified by normoglycemics and prediabetics. n = 1758
Normoglycemia (n = 1322)
Prediabetes (n = 436)
pValue⁎
Male sex, % Age, years Body mass index, kg/m2 Smoking, % Current Past Never Highest education level achieved, % High school College University degree Regular consumption of fruits/ vegetables, % Alcohol consumption, g/day Vigorous physical activity, % Fat mass, % Muscle mass, % Body water, % Hypertension, %a Bilirubin, µmol/l LDL cholesterol, mmol/l HDL cholesterol, mmol/l High sensitive-CRP, mg/l Alanine transaminase, U/l Aspartate transaminase, U/l γ-glutamyl transferase, U/l Uric acid, µmol/l
599 (45.3) 37 (32; 40) 24.4 ± 3.7
237 (54.4) 39 (34; 41) 25.4 ± 3.8
0.001 b0.0001 b0.0001 0.0002
252 (19.1) 328 (24.9) 738 (56.0)
124 (28.4) 88 (20.2) 224 (51.4) 0.74
475 (37.0) 320 (25.0) 388 (30.2) 247 (18.7)
162 (37.8) 107 (25.0) 120 (28.0) 92 (21.1)
0.64 (0.00; 1.71) 631 (47.7) 25 ± 7 35 ± 4 55 ± 5 173 (13.1) 10.3 (6.8; 13.7) 2.92 ± 0.80 1.55 ± 0.41 0.9 (0.5; 1.8) 20.0 (15.0; 29.0) 22.0 (18.0; 26.0) 14.0 (11.0; 22.0) 286.1 ± 76.0
0.64 (0.00; 2.05) 239 (54.8) 25 ± 7 36 ± 4 54 ± 5 74 (17.0) 8.6 (6.8; 12.0) 3.26 ± 0.95 1.47 ± 0.43 1.0 (0.5; 2.0) 23.0 (16.0; 32.0) 22.0 (19.0; 27.0) 17.0 (12.0; 27.0) 299.4 ± 79.0
0.27 0.21 0.01 0.35 0.16 0.87 0.04 b0.0001 b0.0001 0.0004 0.01 0.001 0.34 b0.0001 0.02
Data are presented as mean ± SD, median (interquartile range) or number (percentage). ⁎ p values were based on t-tests, Wilcoxon rank sum tests or Chi-square tests, as appropriate. a Hypertension was defined as a systolic blood pressure ≥140 mm Hg or a diastolic blood pressure ≥90 mm Hg or hypertensive treatment.
model (OR (95%CI) 1.0 (reference), 0.86 (0.60; 1.21); 0.60 (0.44; 0.83) and 0.58 (0.41; 0.81) across increasing quartiles of bilirubin). Women had significantly lower bilirubin levels compared with men (median (interquartile range) 8.55 (6.84; 11.97) versus 10.26 (7.35; 13.68) umol/L, p b 0.0001). In sex-stratified models, we found an OR (95% CI) for prediabetes of 0.34 (0.23; 0.50), p b 0.0001 per one unit increase in log-transformed bilirubin levels among men and 0.72 (0.51; 1.02), p = 0.07 among women (p for interaction = 0.0006). Quartilespecific ORs confirmed these differential relationships (data not shown). In addition, the inverse relationship between bilirubin levels and risk of prediabetes tended to be stronger in overweight/obese individuals (p for interaction 0.07) and in hypertensive participants (p for interaction 0.08). None of the other subgroups by bilirubin interaction tests were statistically significant. In this study, we found a strong inverse relationship between bilirubin levels and risk of prediabetes, even after adjustment for multiple confounders. These data are in agreement with prior studies and extend prior findings to a well characterized sample of young and healthy European individuals [7,8]. Our analysis therefore suggests that oxidative stress and inflammation are implicated early in the development
e25
of diabetes mellitus [3]. In addition, we found that the strong inverse relationship between bilirubin and prediabetes was mainly confined to men. Gender dependent differences in diabetogenesis and associated risk for consecutive complications are a topic of interest and possible influencing factors remain controversial [10]. It seems that lifestyle, diet, fat and muscle distribution and hormonal difference may play a pivotal role [1,8,10]. Higher oxidative stress burden among men may make the protective effect of elevated bilirubin levels particularly important. It is also important to consider that females have lower hemoglobin levels, which is a precursor of bilirubin. However, adjusting for hemoglobin levels in this study did not influence our results. A trend to a stronger inverse association was also observed among overweight/ obese individuals and among those with hypertension. Inflammation and oxidative stress have been associated with obesity and hypertension. Thus, these data again suggest that the protective effect of bilirubin may be particularly important among those with an elevated oxidative stress burden. These data are in agreement with experimental studies showing that reactive oxygen species (ROS) and reactive nitrogen species (RNS) may particularly damage pancreatic β-cells [3]. A strength of this study is the population based design and the large sample of well-characterized participants. Potential limitations that should be taken into account are the following: First, we performed a cross-sectional analysis which does not allow for causal inferences. Second, the study population mainly consists of white adults and the generalizability to other populations is uncertain. Third, we specifically assessed the relationship between bilirubin and prediabetes. It is unclear whether the same associations would be found for overt type 2 diabetes. In conclusion, we found a highly significant inverse association between total bilirubin levels and prediabetes. These data support the concept that oxidative stress may be an important factor in the early course of diabetes development. This effect may be stronger in men. References [1] Nolan CJ, Damm P, Prentki M. Type 2 diabetes across generations: from pathophysiology to prevention and management. Lancet Jul 9 2011;378(9786):169–81. [2] Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. J Clin Invest Jul 2006;116(7):1793–801. [3] Houstis N, Rosen ED, Lander ES. Reactive oxygen species have a causal role in multiple forms of insulin resistance. Nature Apr 13 2006;440(7086):944–8. [4] Meigs JB, Larson MG, Fox CS, Keaney Jr JF, Vasan RS, Benjamin EJ. Association of oxidative stress, insulin resistance, and diabetes risk phenotypes: the Framingham Offspring Study. Diabetes Care Oct 2007;30(10):2529–35. [5] Stocker R, Yamamoto Y, Mcdonagh AF, Glazer AN, Ames BN. Bilirubin is an antioxidant of possible physiological importance. Science Feb 27 1987;235(4792):1043–6. [6] Tomaro ML, Batlle AM. Bilirubin: its role in cytoprotection against oxidative stress. Int J Biochem Cell Biol Mar 2002;34(3):216–20. [7] Oda E. Bilirubin is negatively associated with A1C independently of fasting plasma glucose, age, obesity, inflammation, hemoglobin, and iron in apparently healthy Japanese men and women. Diabetes Care Oct 2010;33(10):e131. [8] Choi SH, Yun KE, Choi HJ. Relationships between serum total bilirubin levels and metabolic syndrome in Korean adults. Nutr Metab Cardiovasc Dis Jan 2013;23(1):31–7. [9] Conen D, Schoen T, Aeschbacher S, et al. Genetic and phenotypic determinants of blood pressure and other cardiovascular risk factors (GAPP). Swiss Med Wkly 2013;143:w13728. [10] Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA Mar 15 2006;295(11):1288–99.
Table 2 Quartiles of serum bilirubin levels and risk of prediabetes.
Unadjusted model Age- and sex adjusted model Fully adjusted model
n
Quartile1
Quartile 2
Quartile 3
Quartile 4
p for linear trend
1758 1758 1652
Ref. Ref. Ref.
0.83 (0.61; 1.14) 0.82 (0.59; 1.13) 0.83 (0.59; 1.17)
0.59 (0.44; 0.79) 0.55 (0.41; 0.74) 0.56 (0.41; 0.78)
0.53 (0.39; 0.71) 0.50 (0.37; 0.68) 0.51 (0.36; 0.71)
b0.0001 b0.0001 b0.0001
Fully adjusted model is adjusted for age, sex, body mass index, hypertension, smoking status, alcohol consumption, low density lipoprotein cholesterol, high density lipoprotein cholesterol, education, physical activity, fruit/vegetable consumption, body composition, C-reactive protein, uric acid, alanine transaminase, aspartate transaminase and gammaglutamyl transferase. The lower sample size in the fully adjusted model is due to missing covariate data.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Letter to the Editor
Serum bilirubin levels and risk of prediabetes in young and healthy adults☆ Matthias Bossard a,b,1,2, Stefanie Aeschbacher b,1,2, Tobias Schoen b,2, Thomas Hochgruber b,2, Mirco von Rotz b,2, Jonas Blum b,2, Martin Risch c,d,2, Lorenz Risch c,e,f,2, David Conen b,⁎,2 a
Cardiology Division, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland Department of Medicine, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland Labormedizinisches Zentrum Dr. Risch, Schaan, Principality of Liechtenstein d Division of Laboratory Medicine, Kantonsspital Graubünden, Chur, Switzerland e Division of Clinical Biochemistry, Medical University, Innsbruck, Austria f Private University, Triesen, Principality of Liechtenstein b c
a r t i c l e
i n f o
Article history: Received 10 September 2013 Accepted 30 November 2013 Available online 6 December 2013 Keywords: Prediabetes Total serum bilirubin Oxidative stress Reactive oxygen species Epidemiology Biomarker
Inflammation and oxidative stress have been implicated in the development of diabetes mellitus, which is one of the leading risk factors for cardiovascular diseases [1–4]. In this context, the antioxidant and anti-inflammatory effects of bilirubin, a breakdown product of the hemoglobin catabolism, have been increasingly recognized in recent years [5,6]. Serum bilirubin concentrations appear to be inversely related to cardiovascular risk factors [7,8], and have been associated with improved insulin sensitivity in animal models [8]. In this analysis, we addressed several unanswered questions regarding these relationships. First, it is relatively unclear whether inflammation and oxidative stress are an early or late phenomenon in
☆ Funding: The Liechtenstein Government, the Swiss Heart Foundation, the Swiss Society of Hypertension, the University of Basel, the University Hospital Basel, the Hanela Foundation, Schiller AG and Novartis supported the GAPP study. David Conen was supported by a grant from the Swiss National Science Foundation (PP00P3_133681). ⁎ Corresponding author at: Department of Medicine, University Hospital, Petersgraben 4, 4031 Basel, Switzerland. Tel.: +41 61 328 66 96; fax: +41 61 265 57 34. E-mail address: [email protected] (D. Conen). 1 M.B. and S.A. contributed equally to this manuscript. 2 All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. 0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2013.11.125
diabetogenesis. Second, biomarkers associated with early changes in glucose homeostasis may improve diabetes risk prediction in the general population. Third, prior studies have been mainly performed in non-European populations, and the generalizability to Western populations remains unclear [7,8]. Finally, possible gender differences for this relationship have not been thoroughly evaluated [8]. We performed a cross-sectional study of healthy adults aged 25–41 years, participating in the ‘Genetic and phenotypic determinants of blood pressure and other cardiovascular risk factors’ (GAPP) study in the Principality of Liechtenstein [9]. Individuals with known diabetes, body mass index (BMI) N 35 kg/m2 and prevalent cardiovascular disease were excluded [9]. Informed consent was obtained from every participant. The local ethics committee approved the study protocol. Prediabetes was defined as an HbA1c level between 5.7 and 6.4% [1,9]. Information about personal, lifestyle (including diet, smoking and physical activity) and medical factors were obtained by questionnaire. Height, weight, waist and hip circumference and blood pressure were measured using standardized methodology. All biomarkers used in this study were assayed as described in detail previously [9]. Multivariable logistic regression models were constructed to compare the odds of prediabetes across sex-specific quartiles of total bilirubin and to adjust for potential confounders. Linear trends were calculated using the sex-stratified, quartile-specific median bilirubin values. For subgroup analyses, logistic regression models were stratified by age, sex, BMI, hypertension and smoking status. Differences across subgroups were tested by including bilirubin by subgroup interaction terms in the fully adjusted non-stratified models. Baseline characteristics of the 1758 participants stratified by the presence or absence of prediabetes are shown in Table 1. Serum bilirubin levels were 8.6 (6.8; 12.0) and 10.3 (6.8; 13.7) umol/l among those with and without prediabetes (p b 0.0001, Table 1). Multivariable logistic regression models of prediabetes risk across sex-specific quartiles of total serum bilirubin levels confirmed an inverse relationship between bilirubin and prediabetes, as shown in Table 2. Sensitivity analysis using waist to hip ratio instead of BMI in the fully adjusted model provided very similar results with ORs (95%CI) of 1.0 (reference), 0.83 (0.59; 1.17); 0.57 (0.41; 0.78) and 0.52 (0.37; 0.72) across increasing quartiles of bilirubin. Similar findings were also obtained when hemoglobin was added as a potential confounder to the fully adjusted
M. Bossard et al. / International Journal of Cardiology 171 (2014) e24–e25 Table 1 Baseline characteristics stratified by normoglycemics and prediabetics. n = 1758
Normoglycemia (n = 1322)
Prediabetes (n = 436)
pValue⁎
Male sex, % Age, years Body mass index, kg/m2 Smoking, % Current Past Never Highest education level achieved, % High school College University degree Regular consumption of fruits/ vegetables, % Alcohol consumption, g/day Vigorous physical activity, % Fat mass, % Muscle mass, % Body water, % Hypertension, %a Bilirubin, µmol/l LDL cholesterol, mmol/l HDL cholesterol, mmol/l High sensitive-CRP, mg/l Alanine transaminase, U/l Aspartate transaminase, U/l γ-glutamyl transferase, U/l Uric acid, µmol/l
599 (45.3) 37 (32; 40) 24.4 ± 3.7
237 (54.4) 39 (34; 41) 25.4 ± 3.8
0.001 b0.0001 b0.0001 0.0002
252 (19.1) 328 (24.9) 738 (56.0)
124 (28.4) 88 (20.2) 224 (51.4) 0.74
475 (37.0) 320 (25.0) 388 (30.2) 247 (18.7)
162 (37.8) 107 (25.0) 120 (28.0) 92 (21.1)
0.64 (0.00; 1.71) 631 (47.7) 25 ± 7 35 ± 4 55 ± 5 173 (13.1) 10.3 (6.8; 13.7) 2.92 ± 0.80 1.55 ± 0.41 0.9 (0.5; 1.8) 20.0 (15.0; 29.0) 22.0 (18.0; 26.0) 14.0 (11.0; 22.0) 286.1 ± 76.0
0.64 (0.00; 2.05) 239 (54.8) 25 ± 7 36 ± 4 54 ± 5 74 (17.0) 8.6 (6.8; 12.0) 3.26 ± 0.95 1.47 ± 0.43 1.0 (0.5; 2.0) 23.0 (16.0; 32.0) 22.0 (19.0; 27.0) 17.0 (12.0; 27.0) 299.4 ± 79.0
0.27 0.21 0.01 0.35 0.16 0.87 0.04 b0.0001 b0.0001 0.0004 0.01 0.001 0.34 b0.0001 0.02
Data are presented as mean ± SD, median (interquartile range) or number (percentage). ⁎ p values were based on t-tests, Wilcoxon rank sum tests or Chi-square tests, as appropriate. a Hypertension was defined as a systolic blood pressure ≥140 mm Hg or a diastolic blood pressure ≥90 mm Hg or hypertensive treatment.
model (OR (95%CI) 1.0 (reference), 0.86 (0.60; 1.21); 0.60 (0.44; 0.83) and 0.58 (0.41; 0.81) across increasing quartiles of bilirubin). Women had significantly lower bilirubin levels compared with men (median (interquartile range) 8.55 (6.84; 11.97) versus 10.26 (7.35; 13.68) umol/L, p b 0.0001). In sex-stratified models, we found an OR (95% CI) for prediabetes of 0.34 (0.23; 0.50), p b 0.0001 per one unit increase in log-transformed bilirubin levels among men and 0.72 (0.51; 1.02), p = 0.07 among women (p for interaction = 0.0006). Quartilespecific ORs confirmed these differential relationships (data not shown). In addition, the inverse relationship between bilirubin levels and risk of prediabetes tended to be stronger in overweight/obese individuals (p for interaction 0.07) and in hypertensive participants (p for interaction 0.08). None of the other subgroups by bilirubin interaction tests were statistically significant. In this study, we found a strong inverse relationship between bilirubin levels and risk of prediabetes, even after adjustment for multiple confounders. These data are in agreement with prior studies and extend prior findings to a well characterized sample of young and healthy European individuals [7,8]. Our analysis therefore suggests that oxidative stress and inflammation are implicated early in the development
e25
of diabetes mellitus [3]. In addition, we found that the strong inverse relationship between bilirubin and prediabetes was mainly confined to men. Gender dependent differences in diabetogenesis and associated risk for consecutive complications are a topic of interest and possible influencing factors remain controversial [10]. It seems that lifestyle, diet, fat and muscle distribution and hormonal difference may play a pivotal role [1,8,10]. Higher oxidative stress burden among men may make the protective effect of elevated bilirubin levels particularly important. It is also important to consider that females have lower hemoglobin levels, which is a precursor of bilirubin. However, adjusting for hemoglobin levels in this study did not influence our results. A trend to a stronger inverse association was also observed among overweight/ obese individuals and among those with hypertension. Inflammation and oxidative stress have been associated with obesity and hypertension. Thus, these data again suggest that the protective effect of bilirubin may be particularly important among those with an elevated oxidative stress burden. These data are in agreement with experimental studies showing that reactive oxygen species (ROS) and reactive nitrogen species (RNS) may particularly damage pancreatic β-cells [3]. A strength of this study is the population based design and the large sample of well-characterized participants. Potential limitations that should be taken into account are the following: First, we performed a cross-sectional analysis which does not allow for causal inferences. Second, the study population mainly consists of white adults and the generalizability to other populations is uncertain. Third, we specifically assessed the relationship between bilirubin and prediabetes. It is unclear whether the same associations would be found for overt type 2 diabetes. In conclusion, we found a highly significant inverse association between total bilirubin levels and prediabetes. These data support the concept that oxidative stress may be an important factor in the early course of diabetes development. This effect may be stronger in men. References [1] Nolan CJ, Damm P, Prentki M. Type 2 diabetes across generations: from pathophysiology to prevention and management. Lancet Jul 9 2011;378(9786):169–81. [2] Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. J Clin Invest Jul 2006;116(7):1793–801. [3] Houstis N, Rosen ED, Lander ES. Reactive oxygen species have a causal role in multiple forms of insulin resistance. Nature Apr 13 2006;440(7086):944–8. [4] Meigs JB, Larson MG, Fox CS, Keaney Jr JF, Vasan RS, Benjamin EJ. Association of oxidative stress, insulin resistance, and diabetes risk phenotypes: the Framingham Offspring Study. Diabetes Care Oct 2007;30(10):2529–35. [5] Stocker R, Yamamoto Y, Mcdonagh AF, Glazer AN, Ames BN. Bilirubin is an antioxidant of possible physiological importance. Science Feb 27 1987;235(4792):1043–6. [6] Tomaro ML, Batlle AM. Bilirubin: its role in cytoprotection against oxidative stress. Int J Biochem Cell Biol Mar 2002;34(3):216–20. [7] Oda E. Bilirubin is negatively associated with A1C independently of fasting plasma glucose, age, obesity, inflammation, hemoglobin, and iron in apparently healthy Japanese men and women. Diabetes Care Oct 2010;33(10):e131. [8] Choi SH, Yun KE, Choi HJ. Relationships between serum total bilirubin levels and metabolic syndrome in Korean adults. Nutr Metab Cardiovasc Dis Jan 2013;23(1):31–7. [9] Conen D, Schoen T, Aeschbacher S, et al. Genetic and phenotypic determinants of blood pressure and other cardiovascular risk factors (GAPP). Swiss Med Wkly 2013;143:w13728. [10] Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA Mar 15 2006;295(11):1288–99.
Table 2 Quartiles of serum bilirubin levels and risk of prediabetes.
Unadjusted model Age- and sex adjusted model Fully adjusted model
n
Quartile1
Quartile 2
Quartile 3
Quartile 4
p for linear trend
1758 1758 1652
Ref. Ref. Ref.
0.83 (0.61; 1.14) 0.82 (0.59; 1.13) 0.83 (0.59; 1.17)
0.59 (0.44; 0.79) 0.55 (0.41; 0.74) 0.56 (0.41; 0.78)
0.53 (0.39; 0.71) 0.50 (0.37; 0.68) 0.51 (0.36; 0.71)
b0.0001 b0.0001 b0.0001
Fully adjusted model is adjusted for age, sex, body mass index, hypertension, smoking status, alcohol consumption, low density lipoprotein cholesterol, high density lipoprotein cholesterol, education, physical activity, fruit/vegetable consumption, body composition, C-reactive protein, uric acid, alanine transaminase, aspartate transaminase and gammaglutamyl transferase. The lower sample size in the fully adjusted model is due to missing covariate data.
