Endocrine DOI 10.1007/s12020-015-0583-0
ORIGINAL ARTICLE
Inverse association between serum total bilirubin levels and diabetic peripheral neuropathy in patients with type 2 diabetes Eun Sook Kim1,3 • Sung Won Lee2,3 • Eun Young Mo1,3 • Sung Dae Moon1,3 Je Ho Han1,3
•
Received: 20 January 2015 / Accepted: 17 March 2015 Ó Springer Science+Business Media New York 2015
Abstract Several studies have suggested that bilirubin, a potent innate antioxidant, plays a protective role against cardiovascular and microvascular disease. This study investigated the association between serum concentrations of total bilirubin (TB) and the presence of diabetic peripheral neuropathy (DPN) in Korean diabetic patients. This crosssectional study involved 1207 patients aged more than 30 years with type 2 diabetes. DPN was assessed according to clinical symptoms and physical examinations using Michigan Neuropathy Screening Instrument examination score, 10-g monofilament sensation, and current perception threshold. The subjects were stratified into gender-specific tertiles based on TB values, and the relationship between the TB values and DPN was analyzed. Compared with patients within the lowest TB tertile, those with higher TB levels consisted of patients with shorter duration of diabetes, lower HbA1c, better renal function, and less autonomic neuropathy, retinopathy, and albuminuria. Serum TB levels were inversely associated with DPN. In multivariate analysis for the development of DPN after adjusting for potential confounding factors including retinopathy,
& Sung Dae Moon [email protected] & Je Ho Han [email protected] 1
Division of Endocrinology and Metabolism, Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, Korea
2
Division of Hepatology, Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, Korea
3
The Catholic University of Korea Incheon St. Mary’s Hospital, Incheon, Korea
albuminuria, and autonomic neuropathy, the TB levels were inversely associated with the presence of DPN, both as a continuous variable [odds ratio (OR) per log standard deviation (SD) 0.79; 95 % confidence interval (CI) 0.65–0.97; P = 0.022] and when categorized in tertiles (the highest vs. the lowest tertile; OR 0.63; 95 % CI 0.40–0.99; P = 0.046). Low serum bilirubin levels are significantly associated with DPN, independently of classic risk factors and other microvascular complications. Further investigation is necessary to determine whether serum bilirubin has a prognostic significance on DPN. Keywords Bilirubin Diabetic peripheral neuropathy Type 2 diabetes
Introduction Diabetes mellitus has become a major health issue worldwide with an increasing prevalence, resulting in grave health consequences and a heavy economic burden [1]. Diabetic peripheral neuropathy (DPN) is one of the common chronic complication of type 2 diabetes, with an estimated lifetime prevalence of more than 50 % [2]. DPN is frequently accompanied by retinopathy and nephropathy possibly due to the common pathogenesis of oxidative stress caused by long-term exposure to hyperglycemia which leads to microvascular damage [3]. Despite the high frequency of DPN, the only therapeutic option is still strict glycemic control. Moreover, even when the glycemic control has remained strict, the neuropathy often fails to improve [4] which makes finding modifiable risk factors for DPN an urgent issue. Bilirubin is the end product of heme catabolism and has been regarded merely as a waste with potential toxicity.
123
Endocrine
However, numerous recent studies have reported the beneficial effect of elevated serum bilirubin on cardiovascular health and its antioxidant properties at physiological concentrations [5]. Low total bilirubin (TB) level was significantly correlated with subclinical atherosclerotic changes and associated with the incidence of coronary artery disease, stroke, and peripheral artery disease independent of conventional risk factors in general population [6–9]. In patients with T2DM, it has been reported that low TB levels were associated with diabetic retinopathy, microalbuminuria, and an increased risk of amputation. However, no study has been carried out yet on the relationship between TB levels and DPN in T2DM patients. Considering that DPN is interrelated with other microvascular complications and is the leading cause of limb amputation, we hypothesized that there may be a negative correlation between the TB level and DPN. The aim of this study was to analyze the association between serum concentrations of TB and the presence of DPN independently of potential confounding factors including other diabetic microvascular complications in Korean patients with type 2 diabetes.
Methods Subjects We retrospectively analyzed 1327 patients with type 2 diabetes older than 30 years who visited the Incheon St. Mary’s Hospital between August 2011 and November 2013 for the purpose of glucose control. Patients with a medical history of chronic liver disease or abnormal liver function test results (defined as aspartate aminotransferase (AST) [ 100 IU/L or alanine aminotransferase (ALT) [ 100 IU/L) (n = 47), renal dysfunction (creatinine C2 mg/dL) (n = 12), severe illness such as systemic inflammatory disease, progressive malignancy, or those taking warfarin or corticosteroids (n = 58) were excluded. A total of 1207 patients were included after additionally excluding those with missing values in the final analysis. The Institutional Review Board of the Clinical Research Coordinating Center in Incheon St. Mary’s Hospital approved the study protocol. Clinical and biochemical assessment Demographic and clinical data were verified by reviewing the electronic medical records. Body mass index (BMI) was calculated by dividing the patients’ weight in kilograms by height in square meters. After overnight fasting, venous blood was taken for laboratory measurement and estimated for plasma HbA1c, serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinine,
123
total cholesterol (TC), triglyceride, and high-density lipoprotein (HDL-C) cholesterol. Low-density lipoprotein cholesterol (LDL-C) was indirectly measured using the Friedewald formula in subjects with serum triglyceride concentrations below 400 mg/mL. Hyperlipidemia was defined as a triglyceride (TG) concentration C150 mg/dL, LDL-C concentration C100 mg/dL, and/or taking cholesterol-lowering medication. Assessment of diabetic peripheral neuropathy The assessments of the DPN were performed by one expert nurse using the Michigan Neuropathy Screening Instrument (MNSI) which consisted of two parts. First, assessment of neuropathic symptoms with a 15-item questionnaire and second, physical examinations of the ankle reflexes, vibratory sensation at the first toes, appearance of feet, and foot ulceration. DPN was defined as seven or more positive responses on the MNSI questionnaire or a score [2.0 on the MNSI examinations [10]. Pressure sensation was assessed the 10-g monofilament at 10 sites on each foot as previously described [11]. At least 8 of 10 correct responses were considered to indicate normal sensibility and less than eight perceived applications indicated reduced sensibility. CPT was measured using a NeurometerÒ CPT/C (Neurotron, Inc., Baltimore, Maryland, USA) at 2000-, 250-, and 5-Hz on the distal phalange of the ring finger and the great toe. CPT values were graded by NeuvalÒ software (grade 0 = no abnormal measures, grade 12 = completely anesthetic) and any CPT between 2.5 SD below and 3–4 SD above the mean was considered abnormal. Definitions of diabetic complications Diabetic nephropathy was defined using urine albuminuria/creatinine ratio (ACR) and graded as follows: normoalbuminuria (ACR \ 30.0 mg/g), microalbuminuria (ACR of 30.0–299.9 mg/g), and macroalbuminuria (ACR C 300.0 mg/g). The estimated glomerular filtration rate (eGFR) was estimated using the Modification of Diet in Renal Disease (MDRD) study equation [12]. Diabetic retinopathy (DMR) was assessed and classified as no DMR, nonproliferative DMR (NPDR), or proliferative DMR (PDR), and positive retinopathy included NPDR and PDR. Cardiac autonomic neuropathy (CAN) was assessed by analyzing heart rate responses to deep breathing, lying to standing, and Valsalva with an automated computer-based system (DICAN, Medi-core Co. Ltd, Korea). The patients were asked to take deep breaths at a rate of six breaths per minute. During respiration, beat-to-beat variability in heart rate was assessed as the mean of the difference between maximum and minimum
Endocrine
heart rates (normal response C15 beats/min, borderline 11–14 beats/min, abnormal B10 beats/min). The heart rate response to standing was evaluated by the ratio of the shortest R–R interval at or around the 15th beat to the largest R–R interval at or around the 30th beat (normal response C1.04, borderline 1.01–1.03, abnormal response B1.00). The heart rate response to the Valsalva maneuver was calculated by the ratio of the longest R–R interval to the shortest R–R interval during forced expiration into the mouthpiece of a manometer at 40 mmHg for 15 s (normal Valsalva ratio C1.21, borderline 1.11–1.20, abnormal B1.10) [13]. The degree of CAN was categorized into absent, early, or definite according to the numbers of abnormal tests (0 = absent, 1 = early, 2 or 3 = definite), and patients with definite abnormality were considered to have CAN. Prevalent cardiovascular disease (CVD) was defined as coronary heart disease encompassing angina pectoris, myocardial infarction, or coronary revascularization (percutaneous transluminal coronary angioplasty, coronary stent placement, and coronary artery bypass graft), stroke, or peripheral vascular disease, which were confirmed by review of medical records.
Statistical analysis Statistical analyses were performed using the SAS software (ver. 9.1; SAS Institute, Cary, NC, USA). Data were expressed as mean ± SD or number (percentages). The comparison of continuous variables was done by the Student’s t test or one-way analysis of variance (ANOVA) and categorized variables by the v2 test. Univariate linear regression was used to examine the association between logtransformed TB levels with various parameters of neuropathy. Multivariate Logistic regression analyses were performed to estimate the age- and sex-adjusted odds ratios (ORs) and 95 % confidence intervals (CI) for DPN according to the categorized and continuous values of TB levels. A P value \0.05 was considered statistically significant.
was observed in the liver enzymes and autonomic neuropathy between the two groups (Table 1).
Table 1 Characteristics of the subjects with and without diabetic peripheral neuropathy Without DPN n
1056
151
1207
Age (years)
55.8 ± 10.4
56.0 ± 10.8
0.668
Male sex (%)
520 (49.2)
57 (37.8)
0.008
Diabetes duration (years)
6.6 ± 7.2
9.1 ± 8.5
\0.001 0.002
B 5 years
588 (55.7)
67 (44.4)
6–10 years
250 (23.7)
37 (24.5) 47 (31.1)
[10 years
218 (20.6)
BMI (kg/m2)
25.0 ± 3.6
25.0 ± 4.1
0.789
HbA1c (%)
8.4 ± 2.2
8.7 ± 2.4
0.194 \0.001
TB (mg/dL)
0.76 ± 0.31
0.65 ± 0.29
GGT (mg/dL)
43.8 ± 60.4
36.8 ± 43.7
0.077
AST (IU/L) ALT (IU/L)
26.0 ± 12.0 28.2 ± 16.9
25.7 ± 13.1 27.8 ± 18.9
0.792 0.814
SBP (mmHg)
129.8 ± 17.5
130.6 ± 19.3
0.581
DBP (mmHg)
78.3 ± 10.4
79.0 ± 11.2
0.435
eGFR min/1.73 m2
106.6 ± 29.2
98.7 ± 30.5
0.002
History of CVD (%)
74 (7.0)
17 (11.3)
0.064
Insulin use (%)
263 (24.9)
49 (32.5)
0.048
CCB
188 (17.8)
36 (23.8)
0.074
ARB/ACE inhibitor
326 (30.9)
60 (39.7)
0.029
Beta blockers
87 (8.2)
14 (9.3)
0.668
322 (31.4)
53 (35.1)
0.367
Antihypertensive use (%)
Hyperlipidemia (%) Autonomic neuropathy (%)
139 (13.2)
27 (17.9)
0.115
Normal
554 (52.5)
68 (45.0)
0.051
Early
363 (34.4)
56 (37.1)
139 (13.1) 256 (24.2)
27 (17.9) 72 (47.7)
\0.001
None
800 (75.8)
79 (52.3)
\0.001
Nonproliferative
229 (21.6)
61 (40.4)
Proliferative
27 (2.6)
11 (7.3)
Definite Retinopathy (%)
252 (23.9)
74 (49.0)
\0.001
Normoalbuminuria
804 (76.1)
77 (51.0)
\0.001
Microalbuminuria
197 (19.6)
46 (30.5)
Macroalbuminuria
55 (5.3)
28 (18.5)
Diabetic nephropathy (%)
Results Clinical characteristics of the subjects Of the 1207 subjects, 151 (12.5 %) were diagnosed with DPN and mean bilirubin level was 0.75 ± 0.31 mg/dL. Compared with subjects without neuropathy, the DPN group consisted of more female patients with longer duration of diabetes, lower TB, and eGFR, and more patients with diabetic retinopathy and nephropathy. No difference
With DPN
Data are expressed as mean ± SD or number (percentage) unless otherwise indicated eGFR was calculated using the Modification of Diet in Renal Disease study equation BMI indicates body mass index, TB total bilirubin, GGT c-glutamyl transferase, AST aspartate aminotransferase, ALT alanine aminotransferase, SBP systolic blood pressure, DBP diastolic blood pressure, eGFR estimated glomerular filtration rate, CVD cardiovascular disease, CCB calcium channel blocker, ARB/ACE inhibitor angiotensin receptor blockers/angiotensin-converting enzyme inhibitor
123
Endocrine
Comparison of parameters according to TB levels in men and women The characteristics of the patients according to the tertiles of serum TB levels are shown in Table 2. Compared with patients within the lowest TB tertile, those with higher TB levels consisted of patients with shorter duration of diabetes, lower HbA1c, higher hepatic enzyme levels, and less patients with retinopathy, albuminuria, and DPN, whereas BMI and blood pressure were similar among all groups and in both genders.
smoking status, history of CVD, HbA1c, SBP, ALT, hyperlipidemia, eGFR, and the use of insulin and antihypertensive agents, the TB levels showed inverse association with the presence of DPN [odds ratio (OR) per log standard deviation (SD) 0.74; 95 % confidence interval (CI) 0.61–0.89; P = 0.002] (Table 4, model 2). Also, after additional adjustments for other microvascular complications including autonomic neuropathy, retinopathy, and nephropathy, the inverse
Table 3 Association between log-transformed total serum bilirubin levels and parameters of neuropathic impairment
Association between serum TB levels and the parameters of neuropathy
ba
TB (mg/dL)
Serum TB levels were associated with higher MNSI scores, impaired ankle reflexes, reduced vibration perception, reduced pressure sensation, abnormal CPT, and showed inverse association with DPN (Table 3). Logistic regression analysis for the risk factors of DPN
t
P
Diabetic peripheral neuropathy
-0.15
-4.44
\0.001
MNSI questionnaire score
-0.02
-4.12
\0.001
MNSI score
-0.06
-3.47
\0.001
Absent ankle reflexes
-0.21
-2.51
0.012
Abnormal vibration perception
-0.11
-2.65
0.008
Abnormal pressure perception
-0.14
-3.27
0.001
Abnormal CPT
-0.06
-2.12
0.034
a
In multivariate analysis, after adjusting for risk factors including age, sex, BMI, duration of diabetes, alcohol and
Standardized coefficient MNSI Michigan Neuropathy Screening Instrument, CPT current perception threshold
Table 2 Characteristics of patients according to serum total bilirubin tertiles TB (mg/dL)
Men T1 (\0.7)
P T2 (0.7–0.9)
T3 ([0.9)
Women T1 (\0.6)
P T2 (0.6–0.7)
T3 ([0.7)
N
183
212
182
Age (years)
55.0 ± 9.8
54.2 ± 10.0
53.4 ± 9.6
Diabetes duration (years)
8.0 ± 8.1
6.0 ± 7.0
4.4 ± 6.3
BMI (kg/m2)
24.3 ± 3.7
24.8 ± 3.6
25.2 ± 3.7
0.050
25.2 ± 4.1
25.0 ± 3.6
25.5 ± 3.9
0.358
HbA1c (%)
9.0 ± 2.6
8.3 ± 2.2
8.2 ± 2.0
0.001
8.7 ± 2.1
8.2 ± 2.4
8.1 ± 2.0
0.048
AST(IU/L) ALT (IU/L)
22.9 ± 9.5 25.6 ± 15.0
25.7 ± 11.2 29.2 ± 15.5
28.1 ± 13.3 33.1 ± 20.2
25.8 ± 13.2 26.0 ± 17.6
25.2 ± 11.5 25.8 ± 16.5
28.2 ± 13.1 29.8 ± 16.8
0.032 0.030
0.283 \0.001
\0.001 \0.001
232
204
194
57.1 ± 11.1
57.4 ± 10.2
57.7 ± 11.0
9.7 ± 8.4
7.2 ± 6.7
5.7 ± 6.3
0.811 \0.001
SBP (mmHg)
130.1 ± 16.4
128.8 ± 17.1
128.4 ± 17.2
0.615
130.4 ± 18.7
131.0 ± 17.8
130.3 ± 18.8
0.929
DBP (mmHg)
78.4 ± 11.0
78.5 ± 10.1
78.5 ± 10.9
0.957
76.8 ± 11.1
76.9 ± 10.1
77.6 ± 10.9
0.689
eGFR min/1.73 m2
105.6 ± 27.6
105.4 ± 23.2
108.1 ± 29.0
0.546
99.8.5 ± 35.2
108.4 ± 30.7
107.7 ± 28.6
0.007
TG (mg/dL)
214.5 ± 165.4
175.6 ± 96.9
191.6 ± 133.7
0.028
172.4 ± 109.4
159.8 ± 135.2
160.6 ± 93.6
0.451
HDL-C (mg/dL)
42.1 ± 9.9
44.6 ± 10.2
45.5 ± 9.9
0.004
45.9 ± 11.8
49.0 ± 13.3
47.9 ± 10.2
0.033
LDL-C
97.9 ± 37.6
103.9 ± 36.8
98.2 ± 37.8
0.231
102.7 ± 40.2
107.3 ± 34.4
102.3 ± 36.3
0.343
History of CVD (%)
19 (10.4)
22 (10.4)
9 (5.0)
0.065
14 (6.0)
12 (5.9)
15 (7.7)
0.707
Current smoking (%)
79 (43.2)
74 (34.9)
62 (34.1)
0.072
21 (9.1)
8 (3.9)
11 (5.7)
0.131
Alcohol (%)
103 (56.3)
123 (58.0)
115 (63.2)
0.181
41 (17.7)
45 (22.1)
37 (19.1)
0.675
Autonomic neuropathy (%)
39 (21.3)
25 (11.8)
17 (9.3)
0.001
34 (17.7)
21 (22.1)
29 (19.1)
0.986
Retinopathy (%) Albuminuria (%)
67 (36.6) 57 (31.2)
58 (27.4) 54 (25.5)
28 (15.4) 34 (18.7)
\0.001 0.006
95 (41.0) 83 (35.8)
41 (20.1) 50 (24.5)
39 (20.1) 48 (24.7)
\0.001 0.010
DPN (%)
28 (15.3)
16 (7.6)
13 (7.1)
0.009
48 (20.7)
23 (11.3)
23 (11.9)
0.008
123
Endocrine
association between the TB levels and DPN persisted, both as a continuous variable [OR per log SD 0.79; 95 % CI 0.65–0.97; P = 0.022] and when categorized in tertiles (the highest vs. the lowest tertile; OR 0.63; 95 % CI 0.40–0.99; P = 0.046). In a receiver operating characteristics curve analysis, a TB value B0.55 mg/dL was identified as the optimal cutoff for predicting DPN (area under the curve 0.61, 95 % CI 0.56–0.66, sensitivity 74.0 %, specificities 54.3 %, P \ 0.001). As microvascular complications often co-exist, we analyzed the association between the total number of microvascular complications (0, 1, 2, 3, or greater) (Fig. 1). After multivariate adjustment, a significant inverse association between the TB levels and the number of microvascular complications was shown (P \ 0.001).
Discussion DPN is a heterogeneous disorder, and distal, symmetrical, sensory/sensorimotor polyneuropathy (DSPN), the most common type of DPN, comprised various clinical manifestations ranging from asymptomatic to numbness, dysesthesia, Table 4 Odds ratio for DPN according to serum TB levels OR
95 % CI
P
0.68
0.57–0.81
\0.001
Model 1 Per SD of log TB By TB tertiles T1
1
T2
0.47
0.31–0.72
\0.001
T3
0.48
0.31–0.73
\0.001
0.74
0.61–0.89
0.002
Model 2 Per SD of log TB By TB tertiles T1
1
T2 T3
0.51 0.55
0.35–0.86 0.35–0.86
0.002 0.009
0.79
0.65–0.97
0.022
Model 3 Per SD of log TB By TB tertiles T1
1
T2
0.57
0.37–0.89
0.014
T3
0.63
0.40–0.99
0.046
Model 1: adjusted for age, sex Model 2: adjusted for the variables in model 1 and body mass index, duration of diabetes, drinking and smoking status, history of cardiovascular disease, HbA1c, systolic blood pressure, alanine aminotransferase, hyperlipidemia, estimated glomerular filtration rate, and use of insulin and antihypertensive agents Model 3: adjusted for the variables in model 2 and autonomic neuropathy, diabetic retinopathy, and albuminuria OR odds ratio, CI confidence interval
Fig. 1 Multivariate-adjusted total bilirubin levels according to numbers of microvascular complications. P for trend \0.001 after adjustment for age, sex, body mass index, duration of diabetes, drinking and smoking status, history of cardiovascular disease, HbA1c, systolic blood pressure, alanine aminotransferase, hyperlipidemia, estimated glomerular filtration rate, and use of insulin and antihypertensive agents
pain in the extremities, and painful cramp. Diabetic patients are at a greater risk of foot ulceration and lower limb amputations which are often consequences of peripheral neuropathy that causes muscle weakness and sensory loss in the long term. DPN has negative impact on the quality of life and is an independent predictor of all-cause and diabetes-related mortality. This study showed that serum TB levels are inversely associated with the development of DPN independently of potential confounding factors and other microvascular complications. This association was further validated by consistent results of reduced pressure perception and abnormal CPT in the low TB group. Moreover, mean TB levels were negatively correlated with the numbers of microvascular complications, suggesting that there may be a graded relationship between TB concentrations and the extent of microvessel injuries. Recently, serum bilirubin levels have gained particular attention for the possible protective role against the development of CVD. Serum bilirubin was inversely correlated with carotid IMT, arterial stiffness, and CVD in a crosssectional analysis, and was an independent predictor of the development of CV outcomes such as MI, stroke, and PVD. In population with diabetes, a study that compared diabetes patients with (n = 96) and without (n = 426) Gilbert syndrome, the most common hereditary genetic disorder causing hyperbilirubinemia, reported a lower prevalence of vascular complications such as coronary artery disease, retinopathy, and macroalbuminuria in the patients [14].
123
Endocrine
Moreover, several cross-sectional studies have shown that serum bilirubin levels are inversely associated with diabetic microvascular complications including retinopathy [15], nephropathy [16], and cardiovascular autonomic neuropathy [17]. Furthermore, low serum TB levels were reported to be an independent predictor of the development of T2DM [18], diabetic nephropathy [19], and lower limb amputation [20] in prospective studies. However, to our knowledge, no study has been carried out yet on the association between the serum TB levels and DPN. The underlying biological mechanism linking the concentrations of serum bilirubin to DPN is not clear, but several explanations may be postulated. First, bilirubin might directly protect nerve tissue by lowering reactive oxygen species (ROS). DPN is a multifactorial disease, resulting from long-term hyperglycemia-mediated cellular damage cascades such as polyol pathway hyperactivity, AGE overproduction, hexosamine pathway, and integrated oxidative stress. Previous in vitro data found that increased ROS leads to axonal damage via direct activation of PARP, protein kinase C, MAPK, and NF-kB or indirectly through ischemia/ reperfusion from injured endoneurial vessels and proinflammatory process [21]. Although extremely high levels of bilirubin can cause kernicterus, a rare type of brain damage in newborn infants [22], recent experimental studies demonstrated that bilirubin is a potent antioxidant that scavenges a 10,000-fold higher concentration of hydrogen peroxide at very low concentrations (10 nmol/L) [5]. In addition, bilirubin lowers ROS and protects against nerve injury by inhibiting the protein kinase C and NAD(P)H oxidase pathway which generate oxidants, uncouples eNOS, leads to endothelial dysfunction, and suppresses the peroxidation of lipid and lipoproteins [23, 24]. This is further supported by the reports that serum uric acid, exerting antioxidant properties in situations associated with oxidative stress, is significantly associated with peripheral neuropathy [25] and sudomotor dysfunction in T2DM [26]. Second, the anti-inflammatory properties of bilirubin could decrease the nerve injury. Experimental data have shown that bilirubin acts against inflammation by interfering with the expression of cell adhesion molecules, complement activity, and T cell differentiation [27]. Moreover, clinical studies have consistently reported an inverse relationship between the levels of bilirubin and CRP, a robust marker of inflammation [28]. Third, the neuroprotective effects of bilirubin may be related to concomitant alterations in the activity of enzymes involved in the bilirubin metabolism pathway. Increased activity of heme oxygenase-1 (HMOX1), which catalyzes heme into carbon monoxide, iron, and biliverdin (subsequently converted to bilirubin), has been reported to exert anti- inflammatory and antioxidant properties [29]. In the same context, HMOX gene promoter polymorphism with longer length of (GT)n repeats which leads to reduced
123
enzyme activities, has been shown to be associated with increased susceptibility to CAD in diabetic patients through its influence on serum bilirubin and ferritin levels [30]. In addition, bilirubin levels may reflect uridine diphosphate glycosyltransferase 1 (UGT1A1) genetic variants which has been reported to be related with a heritable CVD risk. However, the association between genetic variations of UGT1A1 and CVD risk is still on controversy in contrast to consistent data showing protective effects of serum bilirubin on CVD, suggesting a dominant influence of bilirubin over UGT1A1 genetic variants [31]. Clinical implications The results shown in our study suggest that serum bilirubin may be used as a simple, inexpensive biomarker to identify patients at risk for DPN. Also, given that there is lack of effective treatments for diabetic neuropathy, further investigational studies would be necessary on whether modulating bilirubin levels could be a novel therapeutic target for DPN. Limitations Our study has certain limitations. First, its cross-sectional design limits the analysis of causal relationships. Second, we had only single values of serum TB levels available, which is not as desirable as using the means of several measures. However, the use of standardized methods set in a single center, and measurements taken from subjects in a fasting state should improve reliability because there will be less fluctuation in the levels of TB than in the postprandial state. Third, we only measured TB levels without data on unconjugated bilirubin, which has been reported to confer protection from CVD in a previous study [32]. Forth, although we adjusted the history of CVD for correlation of serum bilirubin level with DPN, we still cannot rule out some residual or undetected confounding effects; it is notable that diagnosis of CVD can be missed or delayed in a considerable portion of diabetic patients because silent ischemia is common in coronary artery disease [33] and peripheral vascular disease [34, 35]. Conversely, the strengths of our study are its relatively large sample size, diagnostic criteria that included both subjective and objective measurements, and the validation of the association using different tests in the assessment of neuropathy. In conclusion, low serum bilirubin levels are significantly associated with DPN, independently of classic risk factors and other microvascular complications. Further investigations are necessary on the underlying mechanism that links bilirubin to neuroprotection and the prognostic significance of serum bilirubin on DPN.
Endocrine Acknowledgments This study was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korea government (MSIP, 2014R1A1A1006144) and the Grant (N.H.K., 2009) from the Korean Diabetes Association, and the Grant from the Catholic Medical Center Research Foundation made in the program year of 2013. Conflict of interest The authors declare that there is no conflict of interest associated with this manuscript.
References 1. International Diabetes Federation: IDF Diabetes Atlas, 6th edn. (International Diabetes Federation, Brussels, 2013). http://www. idf.org/diabetesatlas 2. S. Tesfaye, A.J.M. Boulton, P. Dyck, R. Freeman, M. Horowitz, P. Kempler, G. Lauria, R. Malik, V. Spallone, A. Vinik, L. Bernardi, P. Valensi, Diabetic neuropathies: update on definitions, diagnostic criteria, estimation of severity, and treatments. Diabetes Care 33, 2285–2293 (2010) 3. M. Brownlee, The pathobiology of diabetic complications: a unifying mechanism. Diabetes 54, 1615–1625 (2005) 4. A.J. Boulton, F.A. Gries, J.A. Jervell, Guidelines for the diagnosis and outpatient management of diabetic peripheral neuropathy. Diabet. Med. 15, 508–514 (1998) 5. R. Stocker, Y. Yamamoto, A.F. McDonagh, A.N. Glazer, B.N. Ames, Bilirubin is an antioxidant of possible physiological importance. Science 235, 1043–1046 (1987) 6. H. Kimm, J. Yun, J. Jo, S. Jee, Low serum bilirubin level as an independent predictor of stroke incidence: a prospective study in Korean men and women. Stroke 40, 3422–3427 (2009) 7. T. Perlstein, R. Pande, J. Beckman, M. Creager, Serum total bilirubin level and prevalent lower-extremity peripheral arterial disease: National Health and Nutrition Examination Survey (NHANES) 1999 to 2004. Arterioscler. Thromb. Vasc. Biol. 28, 166–172 (2008) 8. P.N. Hopkins, L.L. Wu, S.C. Hunt, B.C. James, G.M. Vincent, R.R. Williams, Higher serum bilirubin is associated with decreased risk for early familial coronary artery disease. Arterioscler. Thromb. Vasc. Biol. 16, 250–255 (1996). doi:10.1161/ 01.atv.16.2.250 9. L. Horsfall, I. Nazareth, I. Petersen, Cardiovascular events as a function of serum bilirubin levels in a large, statin-treated cohort. Circulation 126, 2556–2564 (2012) 10. E.L. Feldman, M.J. Stevens, P.K. Thomas, M.B. Brown, N. Canal, D.A. Greene, A practical two-step quantitative clinical and electrophysiological assessment for the diagnosis and staging of diabetic neuropathy. Diabetes Care 17, 1281–1289 (1994) 11. A.N. Paisley, C.A. Abbott, C.H.M. van Schie, A.J.M. Boulton, A comparison of the neuropen against standard quantitative sensory-threshold measures for assessing peripheral nerve function. Diabet. Med. 19, 400–405 (2002) 12. L.A. Stevens, J. Coresh, T. Greene, A.S. Levey, Assessing kidney function-measured and estimated glomerular filtration rate. N. Engl. J. Med. 354, 2473–2483 (2006) 13. V. Viswanathan, D. Prasad, S. Chamukuttan, A. Ramachandran, High prevalence and early onset of cardiac autonomic neuropathy among South Indian type 2 diabetic patients with nephropathy. Diabetes Res. Clin. Pract. 48, 211–216 (2000) 14. T. Inoguchi, S. Sasaki, K. Kobayashi, R. Takayanagi, T. Yamada, Relationship between Gilbert syndrome and prevalence of vascular complications in patients with diabetes. JAMA J. Am. Med. Assoc. 298, 1398–1400 (2007)
15. M. Yasuda, Y. Kiyohara, J. Wang, S. Arakawa, K. Yonemoto, Y. Doi, T. Ninomiya, T. Ishibashi, High serum bilirubin levels and diabetic retinopathy: the Hisayama Study. Ophthalmology 118, 1423–1428 (2011) 16. M. Fukui, M. Tanaka, E. Shiraishi, I. Harusato, H. Hosoda, M. Asano, G. Hasegawa, N. Nakamura, Relationship between serum bilirubin and albuminuria in patients with type 2 diabetes. Kidney Int. 74, 1197–1201 (2008) 17. J.O. Chung, D.H. Cho, D.J. Chung, M.Y. Chung, Physiological serum bilirubin concentrations are inversely associated with the prevalence of cardiovascular autonomic neuropathy in patients with Type 2 diabetes. Diabet. Med. (2013). doi:10.1111/dme.12338 18. C. Jung, M. Lee, Y. Kang, J. Hwang, J. Jang, J. Leem, J.-Y. Park, H.-K. Kim, W. Lee, Higher serum bilirubin level as a protective factor for the development of diabetes in healthy Korean men: a 4 year retrospective longitudinal study. Metab. Clin. Exp. 63, 87–93 (2014) 19. I. Riphagen, P. Deetman, S.J.L. Bakker, G. Navis, M. Cooper, J. Lewis, D. de Zeeuw, H. Lambers Heerspink, Bilirubin and progression of nephropathy in type 2 diabetes: a post hoc analysis of RENAAL with independent replication in IDNT. Diabetes 63, 2845–2853 (2014) 20. K.H. Chan, R.L. O’Connell, D.R. Sullivan, L.S. Hoffmann, K. Rajamani, M. Whiting, M.W. Donoghoe, M. Vanhala, A. Hamer, B. Yu, R. Stocker, M.K.C. Ng, A.C. Keech, Plasma total bilirubin levels predict amputation events in type 2 diabetes mellitus: the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study. Diabetologia 56, 724–736 (2013) 21. S. Yagihashi, H. Mizukami, K. Sugimoto, Mechanism of diabetic neuropathy: where are we now and where to go? J. Diabetes Investig. 2, 18–32 (2011) 22. C. Petrich, U. Gobel, Letter: idiopathic hyperbilirubinemia and phototherapy—evidence suggesting a decreased incidence of kernicterus. Pediatrics 54, 654–655 (1974) 23. J.Y. Kwak, K. Takeshige, B.S. Cheung, S. Minakami, Bilirubin inhibits the activation of superoxide-producing NADPH oxidase in a neutrophil cell-free system. Biochim. Biophys. Acta 1076, 369–373 (1991) 24. S. Son, Reactive oxygen and nitrogen species in pathogenesis of vascular complications of diabetes. Diabetes Metab. J. 36, 190–198 (2012) 25. N. Papanas, N. Katsiki, K. Papatheodorou, M. Demetriou, D. Papazoglou, T. Gioka, E. Maltezos, Peripheral neuropathy is associated with increased serum levels of uric acid in type 2 diabetes mellitus. Angiology 62, 291–295 (2011) 26. N. Papanas, M. Demetriou, N. Katsiki, K. Papatheodorou, D. Papazoglou, T. Gioka, S. Kotsiou, E. Maltezos, D.P. Mikhailidis, Increased serum levels of uric acid are associated with sudomotor dysfunction in subjects with type 2 diabetes mellitus. Exp. Diabetes Res. 2011, 346051 (2011). doi:10.1155/2011/346051 27. E.D. Petronella, J.L.B. Stephan, P.F.D. Robin, High sensitive C-reactive protein and serum amyloid A are inversely related to serum bilirubin: effect-modification by metabolic syndrome. Cardiovasc. Diabetol. 12, 166 (2013). doi:10.1186/1475-284012-166 28. H.J. Hwang, S.W. Lee, S.H. Kim, Relationship between bilirubin and C-reactive protein. Clin. Chem. Lab. Med. 49, 1823–1828 (2011) 29. D.M. Lundvig, S. Immenschuh, F.A. Wagener, Heme oxygenase, inflammation, and fibrosis: the good, the bad, and the ugly? Front. Pharmacol. 3, 81 (2012). doi:10.3389/fphar.2012.00081 30. Y.-H. Chen, L.-Y. Chau, J.-W. Chen, S.-J. Lin, Serum bilirubin and ferritin levels link heme oxygenase-1 gene promoter polymorphism and susceptibility to coronary artery disease in diabetic patients. Diabetes Care 31, 1615–1620 (2008). doi:10.2337/dc07-2126
123
Endocrine 31. L. Vitek, The role of bilirubin in diabetes, metabolic syndrome, and cardiovascular diseases. Front. Pharmacol. 3, 55 (2012). doi:10.3389/fphar.2012.00055 32. P.J. Bosma, J.R. Chowdhury, C. Bakker, S. Gantla, A. de Boer, B.A. Oostra, D. Lindhout, G.N. Tytgat, P.L. Jansen, R.P. Oude Elferink, The genetic basis of the reduced expression of bilirubin UDP-glucuronosyltransferase 1 in Gilbert’s syndrome. New Engl. J Med. 333, 1171–1175 (1995) 33. C. Gazzaruso, A. Coppola, T. Montalcini, C. Valenti, A. Garzaniti, G. Pelissero, F. Salvucci, P. Gallotti, A. Pujia, C. Falcone, S. Solerte, A. Giustina, Erectile dysfunction can improve the
123
effectiveness of the current guidelines for the screening for asymptomatic coronary artery disease in diabetes. Endocrine 40, 273–279 (2011) 34. J. Apelqvist, Diagnostics and treatment of the diabetic foot. Endocrine 41, 384–397 (2012) 35. C. Gazzaruso, A. Coppola, C. Falcone, C. Luppi, T. Montalcini, E. Baffero, P. Gallotti, A. Pujia, S. Solerte, G. Pelissero, A. Giustina, Transcutaneous oxygen tension as a potential predictor of cardiovascular events in type 2 diabetes: comparison with ankle-brachial index. Diabetes Care 36, 1720–1725 (2013)
ORIGINAL ARTICLE
Inverse association between serum total bilirubin levels and diabetic peripheral neuropathy in patients with type 2 diabetes Eun Sook Kim1,3 • Sung Won Lee2,3 • Eun Young Mo1,3 • Sung Dae Moon1,3 Je Ho Han1,3
•
Received: 20 January 2015 / Accepted: 17 March 2015 Ó Springer Science+Business Media New York 2015
Abstract Several studies have suggested that bilirubin, a potent innate antioxidant, plays a protective role against cardiovascular and microvascular disease. This study investigated the association between serum concentrations of total bilirubin (TB) and the presence of diabetic peripheral neuropathy (DPN) in Korean diabetic patients. This crosssectional study involved 1207 patients aged more than 30 years with type 2 diabetes. DPN was assessed according to clinical symptoms and physical examinations using Michigan Neuropathy Screening Instrument examination score, 10-g monofilament sensation, and current perception threshold. The subjects were stratified into gender-specific tertiles based on TB values, and the relationship between the TB values and DPN was analyzed. Compared with patients within the lowest TB tertile, those with higher TB levels consisted of patients with shorter duration of diabetes, lower HbA1c, better renal function, and less autonomic neuropathy, retinopathy, and albuminuria. Serum TB levels were inversely associated with DPN. In multivariate analysis for the development of DPN after adjusting for potential confounding factors including retinopathy,
& Sung Dae Moon [email protected] & Je Ho Han [email protected] 1
Division of Endocrinology and Metabolism, Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, Korea
2
Division of Hepatology, Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, Korea
3
The Catholic University of Korea Incheon St. Mary’s Hospital, Incheon, Korea
albuminuria, and autonomic neuropathy, the TB levels were inversely associated with the presence of DPN, both as a continuous variable [odds ratio (OR) per log standard deviation (SD) 0.79; 95 % confidence interval (CI) 0.65–0.97; P = 0.022] and when categorized in tertiles (the highest vs. the lowest tertile; OR 0.63; 95 % CI 0.40–0.99; P = 0.046). Low serum bilirubin levels are significantly associated with DPN, independently of classic risk factors and other microvascular complications. Further investigation is necessary to determine whether serum bilirubin has a prognostic significance on DPN. Keywords Bilirubin Diabetic peripheral neuropathy Type 2 diabetes
Introduction Diabetes mellitus has become a major health issue worldwide with an increasing prevalence, resulting in grave health consequences and a heavy economic burden [1]. Diabetic peripheral neuropathy (DPN) is one of the common chronic complication of type 2 diabetes, with an estimated lifetime prevalence of more than 50 % [2]. DPN is frequently accompanied by retinopathy and nephropathy possibly due to the common pathogenesis of oxidative stress caused by long-term exposure to hyperglycemia which leads to microvascular damage [3]. Despite the high frequency of DPN, the only therapeutic option is still strict glycemic control. Moreover, even when the glycemic control has remained strict, the neuropathy often fails to improve [4] which makes finding modifiable risk factors for DPN an urgent issue. Bilirubin is the end product of heme catabolism and has been regarded merely as a waste with potential toxicity.
123
Endocrine
However, numerous recent studies have reported the beneficial effect of elevated serum bilirubin on cardiovascular health and its antioxidant properties at physiological concentrations [5]. Low total bilirubin (TB) level was significantly correlated with subclinical atherosclerotic changes and associated with the incidence of coronary artery disease, stroke, and peripheral artery disease independent of conventional risk factors in general population [6–9]. In patients with T2DM, it has been reported that low TB levels were associated with diabetic retinopathy, microalbuminuria, and an increased risk of amputation. However, no study has been carried out yet on the relationship between TB levels and DPN in T2DM patients. Considering that DPN is interrelated with other microvascular complications and is the leading cause of limb amputation, we hypothesized that there may be a negative correlation between the TB level and DPN. The aim of this study was to analyze the association between serum concentrations of TB and the presence of DPN independently of potential confounding factors including other diabetic microvascular complications in Korean patients with type 2 diabetes.
Methods Subjects We retrospectively analyzed 1327 patients with type 2 diabetes older than 30 years who visited the Incheon St. Mary’s Hospital between August 2011 and November 2013 for the purpose of glucose control. Patients with a medical history of chronic liver disease or abnormal liver function test results (defined as aspartate aminotransferase (AST) [ 100 IU/L or alanine aminotransferase (ALT) [ 100 IU/L) (n = 47), renal dysfunction (creatinine C2 mg/dL) (n = 12), severe illness such as systemic inflammatory disease, progressive malignancy, or those taking warfarin or corticosteroids (n = 58) were excluded. A total of 1207 patients were included after additionally excluding those with missing values in the final analysis. The Institutional Review Board of the Clinical Research Coordinating Center in Incheon St. Mary’s Hospital approved the study protocol. Clinical and biochemical assessment Demographic and clinical data were verified by reviewing the electronic medical records. Body mass index (BMI) was calculated by dividing the patients’ weight in kilograms by height in square meters. After overnight fasting, venous blood was taken for laboratory measurement and estimated for plasma HbA1c, serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinine,
123
total cholesterol (TC), triglyceride, and high-density lipoprotein (HDL-C) cholesterol. Low-density lipoprotein cholesterol (LDL-C) was indirectly measured using the Friedewald formula in subjects with serum triglyceride concentrations below 400 mg/mL. Hyperlipidemia was defined as a triglyceride (TG) concentration C150 mg/dL, LDL-C concentration C100 mg/dL, and/or taking cholesterol-lowering medication. Assessment of diabetic peripheral neuropathy The assessments of the DPN were performed by one expert nurse using the Michigan Neuropathy Screening Instrument (MNSI) which consisted of two parts. First, assessment of neuropathic symptoms with a 15-item questionnaire and second, physical examinations of the ankle reflexes, vibratory sensation at the first toes, appearance of feet, and foot ulceration. DPN was defined as seven or more positive responses on the MNSI questionnaire or a score [2.0 on the MNSI examinations [10]. Pressure sensation was assessed the 10-g monofilament at 10 sites on each foot as previously described [11]. At least 8 of 10 correct responses were considered to indicate normal sensibility and less than eight perceived applications indicated reduced sensibility. CPT was measured using a NeurometerÒ CPT/C (Neurotron, Inc., Baltimore, Maryland, USA) at 2000-, 250-, and 5-Hz on the distal phalange of the ring finger and the great toe. CPT values were graded by NeuvalÒ software (grade 0 = no abnormal measures, grade 12 = completely anesthetic) and any CPT between 2.5 SD below and 3–4 SD above the mean was considered abnormal. Definitions of diabetic complications Diabetic nephropathy was defined using urine albuminuria/creatinine ratio (ACR) and graded as follows: normoalbuminuria (ACR \ 30.0 mg/g), microalbuminuria (ACR of 30.0–299.9 mg/g), and macroalbuminuria (ACR C 300.0 mg/g). The estimated glomerular filtration rate (eGFR) was estimated using the Modification of Diet in Renal Disease (MDRD) study equation [12]. Diabetic retinopathy (DMR) was assessed and classified as no DMR, nonproliferative DMR (NPDR), or proliferative DMR (PDR), and positive retinopathy included NPDR and PDR. Cardiac autonomic neuropathy (CAN) was assessed by analyzing heart rate responses to deep breathing, lying to standing, and Valsalva with an automated computer-based system (DICAN, Medi-core Co. Ltd, Korea). The patients were asked to take deep breaths at a rate of six breaths per minute. During respiration, beat-to-beat variability in heart rate was assessed as the mean of the difference between maximum and minimum
Endocrine
heart rates (normal response C15 beats/min, borderline 11–14 beats/min, abnormal B10 beats/min). The heart rate response to standing was evaluated by the ratio of the shortest R–R interval at or around the 15th beat to the largest R–R interval at or around the 30th beat (normal response C1.04, borderline 1.01–1.03, abnormal response B1.00). The heart rate response to the Valsalva maneuver was calculated by the ratio of the longest R–R interval to the shortest R–R interval during forced expiration into the mouthpiece of a manometer at 40 mmHg for 15 s (normal Valsalva ratio C1.21, borderline 1.11–1.20, abnormal B1.10) [13]. The degree of CAN was categorized into absent, early, or definite according to the numbers of abnormal tests (0 = absent, 1 = early, 2 or 3 = definite), and patients with definite abnormality were considered to have CAN. Prevalent cardiovascular disease (CVD) was defined as coronary heart disease encompassing angina pectoris, myocardial infarction, or coronary revascularization (percutaneous transluminal coronary angioplasty, coronary stent placement, and coronary artery bypass graft), stroke, or peripheral vascular disease, which were confirmed by review of medical records.
Statistical analysis Statistical analyses were performed using the SAS software (ver. 9.1; SAS Institute, Cary, NC, USA). Data were expressed as mean ± SD or number (percentages). The comparison of continuous variables was done by the Student’s t test or one-way analysis of variance (ANOVA) and categorized variables by the v2 test. Univariate linear regression was used to examine the association between logtransformed TB levels with various parameters of neuropathy. Multivariate Logistic regression analyses were performed to estimate the age- and sex-adjusted odds ratios (ORs) and 95 % confidence intervals (CI) for DPN according to the categorized and continuous values of TB levels. A P value \0.05 was considered statistically significant.
was observed in the liver enzymes and autonomic neuropathy between the two groups (Table 1).
Table 1 Characteristics of the subjects with and without diabetic peripheral neuropathy Without DPN n
1056
151
1207
Age (years)
55.8 ± 10.4
56.0 ± 10.8
0.668
Male sex (%)
520 (49.2)
57 (37.8)
0.008
Diabetes duration (years)
6.6 ± 7.2
9.1 ± 8.5
\0.001 0.002
B 5 years
588 (55.7)
67 (44.4)
6–10 years
250 (23.7)
37 (24.5) 47 (31.1)
[10 years
218 (20.6)
BMI (kg/m2)
25.0 ± 3.6
25.0 ± 4.1
0.789
HbA1c (%)
8.4 ± 2.2
8.7 ± 2.4
0.194 \0.001
TB (mg/dL)
0.76 ± 0.31
0.65 ± 0.29
GGT (mg/dL)
43.8 ± 60.4
36.8 ± 43.7
0.077
AST (IU/L) ALT (IU/L)
26.0 ± 12.0 28.2 ± 16.9
25.7 ± 13.1 27.8 ± 18.9
0.792 0.814
SBP (mmHg)
129.8 ± 17.5
130.6 ± 19.3
0.581
DBP (mmHg)
78.3 ± 10.4
79.0 ± 11.2
0.435
eGFR min/1.73 m2
106.6 ± 29.2
98.7 ± 30.5
0.002
History of CVD (%)
74 (7.0)
17 (11.3)
0.064
Insulin use (%)
263 (24.9)
49 (32.5)
0.048
CCB
188 (17.8)
36 (23.8)
0.074
ARB/ACE inhibitor
326 (30.9)
60 (39.7)
0.029
Beta blockers
87 (8.2)
14 (9.3)
0.668
322 (31.4)
53 (35.1)
0.367
Antihypertensive use (%)
Hyperlipidemia (%) Autonomic neuropathy (%)
139 (13.2)
27 (17.9)
0.115
Normal
554 (52.5)
68 (45.0)
0.051
Early
363 (34.4)
56 (37.1)
139 (13.1) 256 (24.2)
27 (17.9) 72 (47.7)
\0.001
None
800 (75.8)
79 (52.3)
\0.001
Nonproliferative
229 (21.6)
61 (40.4)
Proliferative
27 (2.6)
11 (7.3)
Definite Retinopathy (%)
252 (23.9)
74 (49.0)
\0.001
Normoalbuminuria
804 (76.1)
77 (51.0)
\0.001
Microalbuminuria
197 (19.6)
46 (30.5)
Macroalbuminuria
55 (5.3)
28 (18.5)
Diabetic nephropathy (%)
Results Clinical characteristics of the subjects Of the 1207 subjects, 151 (12.5 %) were diagnosed with DPN and mean bilirubin level was 0.75 ± 0.31 mg/dL. Compared with subjects without neuropathy, the DPN group consisted of more female patients with longer duration of diabetes, lower TB, and eGFR, and more patients with diabetic retinopathy and nephropathy. No difference
With DPN
Data are expressed as mean ± SD or number (percentage) unless otherwise indicated eGFR was calculated using the Modification of Diet in Renal Disease study equation BMI indicates body mass index, TB total bilirubin, GGT c-glutamyl transferase, AST aspartate aminotransferase, ALT alanine aminotransferase, SBP systolic blood pressure, DBP diastolic blood pressure, eGFR estimated glomerular filtration rate, CVD cardiovascular disease, CCB calcium channel blocker, ARB/ACE inhibitor angiotensin receptor blockers/angiotensin-converting enzyme inhibitor
123
Endocrine
Comparison of parameters according to TB levels in men and women The characteristics of the patients according to the tertiles of serum TB levels are shown in Table 2. Compared with patients within the lowest TB tertile, those with higher TB levels consisted of patients with shorter duration of diabetes, lower HbA1c, higher hepatic enzyme levels, and less patients with retinopathy, albuminuria, and DPN, whereas BMI and blood pressure were similar among all groups and in both genders.
smoking status, history of CVD, HbA1c, SBP, ALT, hyperlipidemia, eGFR, and the use of insulin and antihypertensive agents, the TB levels showed inverse association with the presence of DPN [odds ratio (OR) per log standard deviation (SD) 0.74; 95 % confidence interval (CI) 0.61–0.89; P = 0.002] (Table 4, model 2). Also, after additional adjustments for other microvascular complications including autonomic neuropathy, retinopathy, and nephropathy, the inverse
Table 3 Association between log-transformed total serum bilirubin levels and parameters of neuropathic impairment
Association between serum TB levels and the parameters of neuropathy
ba
TB (mg/dL)
Serum TB levels were associated with higher MNSI scores, impaired ankle reflexes, reduced vibration perception, reduced pressure sensation, abnormal CPT, and showed inverse association with DPN (Table 3). Logistic regression analysis for the risk factors of DPN
t
P
Diabetic peripheral neuropathy
-0.15
-4.44
\0.001
MNSI questionnaire score
-0.02
-4.12
\0.001
MNSI score
-0.06
-3.47
\0.001
Absent ankle reflexes
-0.21
-2.51
0.012
Abnormal vibration perception
-0.11
-2.65
0.008
Abnormal pressure perception
-0.14
-3.27
0.001
Abnormal CPT
-0.06
-2.12
0.034
a
In multivariate analysis, after adjusting for risk factors including age, sex, BMI, duration of diabetes, alcohol and
Standardized coefficient MNSI Michigan Neuropathy Screening Instrument, CPT current perception threshold
Table 2 Characteristics of patients according to serum total bilirubin tertiles TB (mg/dL)
Men T1 (\0.7)
P T2 (0.7–0.9)
T3 ([0.9)
Women T1 (\0.6)
P T2 (0.6–0.7)
T3 ([0.7)
N
183
212
182
Age (years)
55.0 ± 9.8
54.2 ± 10.0
53.4 ± 9.6
Diabetes duration (years)
8.0 ± 8.1
6.0 ± 7.0
4.4 ± 6.3
BMI (kg/m2)
24.3 ± 3.7
24.8 ± 3.6
25.2 ± 3.7
0.050
25.2 ± 4.1
25.0 ± 3.6
25.5 ± 3.9
0.358
HbA1c (%)
9.0 ± 2.6
8.3 ± 2.2
8.2 ± 2.0
0.001
8.7 ± 2.1
8.2 ± 2.4
8.1 ± 2.0
0.048
AST(IU/L) ALT (IU/L)
22.9 ± 9.5 25.6 ± 15.0
25.7 ± 11.2 29.2 ± 15.5
28.1 ± 13.3 33.1 ± 20.2
25.8 ± 13.2 26.0 ± 17.6
25.2 ± 11.5 25.8 ± 16.5
28.2 ± 13.1 29.8 ± 16.8
0.032 0.030
0.283 \0.001
\0.001 \0.001
232
204
194
57.1 ± 11.1
57.4 ± 10.2
57.7 ± 11.0
9.7 ± 8.4
7.2 ± 6.7
5.7 ± 6.3
0.811 \0.001
SBP (mmHg)
130.1 ± 16.4
128.8 ± 17.1
128.4 ± 17.2
0.615
130.4 ± 18.7
131.0 ± 17.8
130.3 ± 18.8
0.929
DBP (mmHg)
78.4 ± 11.0
78.5 ± 10.1
78.5 ± 10.9
0.957
76.8 ± 11.1
76.9 ± 10.1
77.6 ± 10.9
0.689
eGFR min/1.73 m2
105.6 ± 27.6
105.4 ± 23.2
108.1 ± 29.0
0.546
99.8.5 ± 35.2
108.4 ± 30.7
107.7 ± 28.6
0.007
TG (mg/dL)
214.5 ± 165.4
175.6 ± 96.9
191.6 ± 133.7
0.028
172.4 ± 109.4
159.8 ± 135.2
160.6 ± 93.6
0.451
HDL-C (mg/dL)
42.1 ± 9.9
44.6 ± 10.2
45.5 ± 9.9
0.004
45.9 ± 11.8
49.0 ± 13.3
47.9 ± 10.2
0.033
LDL-C
97.9 ± 37.6
103.9 ± 36.8
98.2 ± 37.8
0.231
102.7 ± 40.2
107.3 ± 34.4
102.3 ± 36.3
0.343
History of CVD (%)
19 (10.4)
22 (10.4)
9 (5.0)
0.065
14 (6.0)
12 (5.9)
15 (7.7)
0.707
Current smoking (%)
79 (43.2)
74 (34.9)
62 (34.1)
0.072
21 (9.1)
8 (3.9)
11 (5.7)
0.131
Alcohol (%)
103 (56.3)
123 (58.0)
115 (63.2)
0.181
41 (17.7)
45 (22.1)
37 (19.1)
0.675
Autonomic neuropathy (%)
39 (21.3)
25 (11.8)
17 (9.3)
0.001
34 (17.7)
21 (22.1)
29 (19.1)
0.986
Retinopathy (%) Albuminuria (%)
67 (36.6) 57 (31.2)
58 (27.4) 54 (25.5)
28 (15.4) 34 (18.7)
\0.001 0.006
95 (41.0) 83 (35.8)
41 (20.1) 50 (24.5)
39 (20.1) 48 (24.7)
\0.001 0.010
DPN (%)
28 (15.3)
16 (7.6)
13 (7.1)
0.009
48 (20.7)
23 (11.3)
23 (11.9)
0.008
123
Endocrine
association between the TB levels and DPN persisted, both as a continuous variable [OR per log SD 0.79; 95 % CI 0.65–0.97; P = 0.022] and when categorized in tertiles (the highest vs. the lowest tertile; OR 0.63; 95 % CI 0.40–0.99; P = 0.046). In a receiver operating characteristics curve analysis, a TB value B0.55 mg/dL was identified as the optimal cutoff for predicting DPN (area under the curve 0.61, 95 % CI 0.56–0.66, sensitivity 74.0 %, specificities 54.3 %, P \ 0.001). As microvascular complications often co-exist, we analyzed the association between the total number of microvascular complications (0, 1, 2, 3, or greater) (Fig. 1). After multivariate adjustment, a significant inverse association between the TB levels and the number of microvascular complications was shown (P \ 0.001).
Discussion DPN is a heterogeneous disorder, and distal, symmetrical, sensory/sensorimotor polyneuropathy (DSPN), the most common type of DPN, comprised various clinical manifestations ranging from asymptomatic to numbness, dysesthesia, Table 4 Odds ratio for DPN according to serum TB levels OR
95 % CI
P
0.68
0.57–0.81
\0.001
Model 1 Per SD of log TB By TB tertiles T1
1
T2
0.47
0.31–0.72
\0.001
T3
0.48
0.31–0.73
\0.001
0.74
0.61–0.89
0.002
Model 2 Per SD of log TB By TB tertiles T1
1
T2 T3
0.51 0.55
0.35–0.86 0.35–0.86
0.002 0.009
0.79
0.65–0.97
0.022
Model 3 Per SD of log TB By TB tertiles T1
1
T2
0.57
0.37–0.89
0.014
T3
0.63
0.40–0.99
0.046
Model 1: adjusted for age, sex Model 2: adjusted for the variables in model 1 and body mass index, duration of diabetes, drinking and smoking status, history of cardiovascular disease, HbA1c, systolic blood pressure, alanine aminotransferase, hyperlipidemia, estimated glomerular filtration rate, and use of insulin and antihypertensive agents Model 3: adjusted for the variables in model 2 and autonomic neuropathy, diabetic retinopathy, and albuminuria OR odds ratio, CI confidence interval
Fig. 1 Multivariate-adjusted total bilirubin levels according to numbers of microvascular complications. P for trend \0.001 after adjustment for age, sex, body mass index, duration of diabetes, drinking and smoking status, history of cardiovascular disease, HbA1c, systolic blood pressure, alanine aminotransferase, hyperlipidemia, estimated glomerular filtration rate, and use of insulin and antihypertensive agents
pain in the extremities, and painful cramp. Diabetic patients are at a greater risk of foot ulceration and lower limb amputations which are often consequences of peripheral neuropathy that causes muscle weakness and sensory loss in the long term. DPN has negative impact on the quality of life and is an independent predictor of all-cause and diabetes-related mortality. This study showed that serum TB levels are inversely associated with the development of DPN independently of potential confounding factors and other microvascular complications. This association was further validated by consistent results of reduced pressure perception and abnormal CPT in the low TB group. Moreover, mean TB levels were negatively correlated with the numbers of microvascular complications, suggesting that there may be a graded relationship between TB concentrations and the extent of microvessel injuries. Recently, serum bilirubin levels have gained particular attention for the possible protective role against the development of CVD. Serum bilirubin was inversely correlated with carotid IMT, arterial stiffness, and CVD in a crosssectional analysis, and was an independent predictor of the development of CV outcomes such as MI, stroke, and PVD. In population with diabetes, a study that compared diabetes patients with (n = 96) and without (n = 426) Gilbert syndrome, the most common hereditary genetic disorder causing hyperbilirubinemia, reported a lower prevalence of vascular complications such as coronary artery disease, retinopathy, and macroalbuminuria in the patients [14].
123
Endocrine
Moreover, several cross-sectional studies have shown that serum bilirubin levels are inversely associated with diabetic microvascular complications including retinopathy [15], nephropathy [16], and cardiovascular autonomic neuropathy [17]. Furthermore, low serum TB levels were reported to be an independent predictor of the development of T2DM [18], diabetic nephropathy [19], and lower limb amputation [20] in prospective studies. However, to our knowledge, no study has been carried out yet on the association between the serum TB levels and DPN. The underlying biological mechanism linking the concentrations of serum bilirubin to DPN is not clear, but several explanations may be postulated. First, bilirubin might directly protect nerve tissue by lowering reactive oxygen species (ROS). DPN is a multifactorial disease, resulting from long-term hyperglycemia-mediated cellular damage cascades such as polyol pathway hyperactivity, AGE overproduction, hexosamine pathway, and integrated oxidative stress. Previous in vitro data found that increased ROS leads to axonal damage via direct activation of PARP, protein kinase C, MAPK, and NF-kB or indirectly through ischemia/ reperfusion from injured endoneurial vessels and proinflammatory process [21]. Although extremely high levels of bilirubin can cause kernicterus, a rare type of brain damage in newborn infants [22], recent experimental studies demonstrated that bilirubin is a potent antioxidant that scavenges a 10,000-fold higher concentration of hydrogen peroxide at very low concentrations (10 nmol/L) [5]. In addition, bilirubin lowers ROS and protects against nerve injury by inhibiting the protein kinase C and NAD(P)H oxidase pathway which generate oxidants, uncouples eNOS, leads to endothelial dysfunction, and suppresses the peroxidation of lipid and lipoproteins [23, 24]. This is further supported by the reports that serum uric acid, exerting antioxidant properties in situations associated with oxidative stress, is significantly associated with peripheral neuropathy [25] and sudomotor dysfunction in T2DM [26]. Second, the anti-inflammatory properties of bilirubin could decrease the nerve injury. Experimental data have shown that bilirubin acts against inflammation by interfering with the expression of cell adhesion molecules, complement activity, and T cell differentiation [27]. Moreover, clinical studies have consistently reported an inverse relationship between the levels of bilirubin and CRP, a robust marker of inflammation [28]. Third, the neuroprotective effects of bilirubin may be related to concomitant alterations in the activity of enzymes involved in the bilirubin metabolism pathway. Increased activity of heme oxygenase-1 (HMOX1), which catalyzes heme into carbon monoxide, iron, and biliverdin (subsequently converted to bilirubin), has been reported to exert anti- inflammatory and antioxidant properties [29]. In the same context, HMOX gene promoter polymorphism with longer length of (GT)n repeats which leads to reduced
123
enzyme activities, has been shown to be associated with increased susceptibility to CAD in diabetic patients through its influence on serum bilirubin and ferritin levels [30]. In addition, bilirubin levels may reflect uridine diphosphate glycosyltransferase 1 (UGT1A1) genetic variants which has been reported to be related with a heritable CVD risk. However, the association between genetic variations of UGT1A1 and CVD risk is still on controversy in contrast to consistent data showing protective effects of serum bilirubin on CVD, suggesting a dominant influence of bilirubin over UGT1A1 genetic variants [31]. Clinical implications The results shown in our study suggest that serum bilirubin may be used as a simple, inexpensive biomarker to identify patients at risk for DPN. Also, given that there is lack of effective treatments for diabetic neuropathy, further investigational studies would be necessary on whether modulating bilirubin levels could be a novel therapeutic target for DPN. Limitations Our study has certain limitations. First, its cross-sectional design limits the analysis of causal relationships. Second, we had only single values of serum TB levels available, which is not as desirable as using the means of several measures. However, the use of standardized methods set in a single center, and measurements taken from subjects in a fasting state should improve reliability because there will be less fluctuation in the levels of TB than in the postprandial state. Third, we only measured TB levels without data on unconjugated bilirubin, which has been reported to confer protection from CVD in a previous study [32]. Forth, although we adjusted the history of CVD for correlation of serum bilirubin level with DPN, we still cannot rule out some residual or undetected confounding effects; it is notable that diagnosis of CVD can be missed or delayed in a considerable portion of diabetic patients because silent ischemia is common in coronary artery disease [33] and peripheral vascular disease [34, 35]. Conversely, the strengths of our study are its relatively large sample size, diagnostic criteria that included both subjective and objective measurements, and the validation of the association using different tests in the assessment of neuropathy. In conclusion, low serum bilirubin levels are significantly associated with DPN, independently of classic risk factors and other microvascular complications. Further investigations are necessary on the underlying mechanism that links bilirubin to neuroprotection and the prognostic significance of serum bilirubin on DPN.
Endocrine Acknowledgments This study was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korea government (MSIP, 2014R1A1A1006144) and the Grant (N.H.K., 2009) from the Korean Diabetes Association, and the Grant from the Catholic Medical Center Research Foundation made in the program year of 2013. Conflict of interest The authors declare that there is no conflict of interest associated with this manuscript.
References 1. International Diabetes Federation: IDF Diabetes Atlas, 6th edn. (International Diabetes Federation, Brussels, 2013). http://www. idf.org/diabetesatlas 2. S. Tesfaye, A.J.M. Boulton, P. Dyck, R. Freeman, M. Horowitz, P. Kempler, G. Lauria, R. Malik, V. Spallone, A. Vinik, L. Bernardi, P. Valensi, Diabetic neuropathies: update on definitions, diagnostic criteria, estimation of severity, and treatments. Diabetes Care 33, 2285–2293 (2010) 3. M. Brownlee, The pathobiology of diabetic complications: a unifying mechanism. Diabetes 54, 1615–1625 (2005) 4. A.J. Boulton, F.A. Gries, J.A. Jervell, Guidelines for the diagnosis and outpatient management of diabetic peripheral neuropathy. Diabet. Med. 15, 508–514 (1998) 5. R. Stocker, Y. Yamamoto, A.F. McDonagh, A.N. Glazer, B.N. Ames, Bilirubin is an antioxidant of possible physiological importance. Science 235, 1043–1046 (1987) 6. H. Kimm, J. Yun, J. Jo, S. Jee, Low serum bilirubin level as an independent predictor of stroke incidence: a prospective study in Korean men and women. Stroke 40, 3422–3427 (2009) 7. T. Perlstein, R. Pande, J. Beckman, M. Creager, Serum total bilirubin level and prevalent lower-extremity peripheral arterial disease: National Health and Nutrition Examination Survey (NHANES) 1999 to 2004. Arterioscler. Thromb. Vasc. Biol. 28, 166–172 (2008) 8. P.N. Hopkins, L.L. Wu, S.C. Hunt, B.C. James, G.M. Vincent, R.R. Williams, Higher serum bilirubin is associated with decreased risk for early familial coronary artery disease. Arterioscler. Thromb. Vasc. Biol. 16, 250–255 (1996). doi:10.1161/ 01.atv.16.2.250 9. L. Horsfall, I. Nazareth, I. Petersen, Cardiovascular events as a function of serum bilirubin levels in a large, statin-treated cohort. Circulation 126, 2556–2564 (2012) 10. E.L. Feldman, M.J. Stevens, P.K. Thomas, M.B. Brown, N. Canal, D.A. Greene, A practical two-step quantitative clinical and electrophysiological assessment for the diagnosis and staging of diabetic neuropathy. Diabetes Care 17, 1281–1289 (1994) 11. A.N. Paisley, C.A. Abbott, C.H.M. van Schie, A.J.M. Boulton, A comparison of the neuropen against standard quantitative sensory-threshold measures for assessing peripheral nerve function. Diabet. Med. 19, 400–405 (2002) 12. L.A. Stevens, J. Coresh, T. Greene, A.S. Levey, Assessing kidney function-measured and estimated glomerular filtration rate. N. Engl. J. Med. 354, 2473–2483 (2006) 13. V. Viswanathan, D. Prasad, S. Chamukuttan, A. Ramachandran, High prevalence and early onset of cardiac autonomic neuropathy among South Indian type 2 diabetic patients with nephropathy. Diabetes Res. Clin. Pract. 48, 211–216 (2000) 14. T. Inoguchi, S. Sasaki, K. Kobayashi, R. Takayanagi, T. Yamada, Relationship between Gilbert syndrome and prevalence of vascular complications in patients with diabetes. JAMA J. Am. Med. Assoc. 298, 1398–1400 (2007)
15. M. Yasuda, Y. Kiyohara, J. Wang, S. Arakawa, K. Yonemoto, Y. Doi, T. Ninomiya, T. Ishibashi, High serum bilirubin levels and diabetic retinopathy: the Hisayama Study. Ophthalmology 118, 1423–1428 (2011) 16. M. Fukui, M. Tanaka, E. Shiraishi, I. Harusato, H. Hosoda, M. Asano, G. Hasegawa, N. Nakamura, Relationship between serum bilirubin and albuminuria in patients with type 2 diabetes. Kidney Int. 74, 1197–1201 (2008) 17. J.O. Chung, D.H. Cho, D.J. Chung, M.Y. Chung, Physiological serum bilirubin concentrations are inversely associated with the prevalence of cardiovascular autonomic neuropathy in patients with Type 2 diabetes. Diabet. Med. (2013). doi:10.1111/dme.12338 18. C. Jung, M. Lee, Y. Kang, J. Hwang, J. Jang, J. Leem, J.-Y. Park, H.-K. Kim, W. Lee, Higher serum bilirubin level as a protective factor for the development of diabetes in healthy Korean men: a 4 year retrospective longitudinal study. Metab. Clin. Exp. 63, 87–93 (2014) 19. I. Riphagen, P. Deetman, S.J.L. Bakker, G. Navis, M. Cooper, J. Lewis, D. de Zeeuw, H. Lambers Heerspink, Bilirubin and progression of nephropathy in type 2 diabetes: a post hoc analysis of RENAAL with independent replication in IDNT. Diabetes 63, 2845–2853 (2014) 20. K.H. Chan, R.L. O’Connell, D.R. Sullivan, L.S. Hoffmann, K. Rajamani, M. Whiting, M.W. Donoghoe, M. Vanhala, A. Hamer, B. Yu, R. Stocker, M.K.C. Ng, A.C. Keech, Plasma total bilirubin levels predict amputation events in type 2 diabetes mellitus: the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study. Diabetologia 56, 724–736 (2013) 21. S. Yagihashi, H. Mizukami, K. Sugimoto, Mechanism of diabetic neuropathy: where are we now and where to go? J. Diabetes Investig. 2, 18–32 (2011) 22. C. Petrich, U. Gobel, Letter: idiopathic hyperbilirubinemia and phototherapy—evidence suggesting a decreased incidence of kernicterus. Pediatrics 54, 654–655 (1974) 23. J.Y. Kwak, K. Takeshige, B.S. Cheung, S. Minakami, Bilirubin inhibits the activation of superoxide-producing NADPH oxidase in a neutrophil cell-free system. Biochim. Biophys. Acta 1076, 369–373 (1991) 24. S. Son, Reactive oxygen and nitrogen species in pathogenesis of vascular complications of diabetes. Diabetes Metab. J. 36, 190–198 (2012) 25. N. Papanas, N. Katsiki, K. Papatheodorou, M. Demetriou, D. Papazoglou, T. Gioka, E. Maltezos, Peripheral neuropathy is associated with increased serum levels of uric acid in type 2 diabetes mellitus. Angiology 62, 291–295 (2011) 26. N. Papanas, M. Demetriou, N. Katsiki, K. Papatheodorou, D. Papazoglou, T. Gioka, S. Kotsiou, E. Maltezos, D.P. Mikhailidis, Increased serum levels of uric acid are associated with sudomotor dysfunction in subjects with type 2 diabetes mellitus. Exp. Diabetes Res. 2011, 346051 (2011). doi:10.1155/2011/346051 27. E.D. Petronella, J.L.B. Stephan, P.F.D. Robin, High sensitive C-reactive protein and serum amyloid A are inversely related to serum bilirubin: effect-modification by metabolic syndrome. Cardiovasc. Diabetol. 12, 166 (2013). doi:10.1186/1475-284012-166 28. H.J. Hwang, S.W. Lee, S.H. Kim, Relationship between bilirubin and C-reactive protein. Clin. Chem. Lab. Med. 49, 1823–1828 (2011) 29. D.M. Lundvig, S. Immenschuh, F.A. Wagener, Heme oxygenase, inflammation, and fibrosis: the good, the bad, and the ugly? Front. Pharmacol. 3, 81 (2012). doi:10.3389/fphar.2012.00081 30. Y.-H. Chen, L.-Y. Chau, J.-W. Chen, S.-J. Lin, Serum bilirubin and ferritin levels link heme oxygenase-1 gene promoter polymorphism and susceptibility to coronary artery disease in diabetic patients. Diabetes Care 31, 1615–1620 (2008). doi:10.2337/dc07-2126
123
Endocrine 31. L. Vitek, The role of bilirubin in diabetes, metabolic syndrome, and cardiovascular diseases. Front. Pharmacol. 3, 55 (2012). doi:10.3389/fphar.2012.00055 32. P.J. Bosma, J.R. Chowdhury, C. Bakker, S. Gantla, A. de Boer, B.A. Oostra, D. Lindhout, G.N. Tytgat, P.L. Jansen, R.P. Oude Elferink, The genetic basis of the reduced expression of bilirubin UDP-glucuronosyltransferase 1 in Gilbert’s syndrome. New Engl. J Med. 333, 1171–1175 (1995) 33. C. Gazzaruso, A. Coppola, T. Montalcini, C. Valenti, A. Garzaniti, G. Pelissero, F. Salvucci, P. Gallotti, A. Pujia, C. Falcone, S. Solerte, A. Giustina, Erectile dysfunction can improve the
123
effectiveness of the current guidelines for the screening for asymptomatic coronary artery disease in diabetes. Endocrine 40, 273–279 (2011) 34. J. Apelqvist, Diagnostics and treatment of the diabetic foot. Endocrine 41, 384–397 (2012) 35. C. Gazzaruso, A. Coppola, C. Falcone, C. Luppi, T. Montalcini, E. Baffero, P. Gallotti, A. Pujia, S. Solerte, G. Pelissero, A. Giustina, Transcutaneous oxygen tension as a potential predictor of cardiovascular events in type 2 diabetes: comparison with ankle-brachial index. Diabetes Care 36, 1720–1725 (2013)
