Association Between Serum Total Bilirubin Levels and the Morphology of Corneal Nerve Fibers in Japanese Patients With Uncontrolled Type 2 Diabetes

Oxidative stress (OS) plays a key role in the development of diabetes complications. Bilirubin is a potent antioxidant, and serum total bilirubin (TB) levels in the low-normal range are associated with diabetic retinopathy (1) and nephropathy (2) and are predictive of lower-limb amputation (3). However, the relationship between serum TB and diabetic neuropathy has not been reported. Due to constant exposure to OS (sunlight and ambient air) and poor antioxidant activity due to its avascularity, the cornea is vulnerable to OS (4). Diabetes increases the OS burden on the cornea. Corneal nerve fibers (CNFs) are composed of unmyelinated C-fibers, which are the first nerve fibers to undergo damage and subsequent repair. We examined the association between serum TB and CNF morphology in patients with uncontrolled type 2 diabetes (T2D).

Cross-sectional data were collected from 384 Japanese T2D patients with severe (HbA_1c ≥8.5% [69.4 mmol/mol], n = 143), moderate (7.2% [55.2 mmol/mol] to <8.5% [69.4 mmol/mol], n = 104), or mild (6.5% [47.5 mmol/mol] to <7.2% [55.2 mmol/mol], n = 137) hyperglycemia and compared with age- and sex-matched controls (n = 57). CNF morphological parameters (CNF density [CNFD] and length [CNFL], CNF branch density [CNBD] and length [CNBL], tortuosity, and beading frequency) were examined by corneal confocal microscopy (5). HbA_1c, serum TB, γ-glutamyl transpeptidase (GGT), high-sensitivity C-reactive protein (hsCRP), creatinine, and the urinary albumin-to-creatinine ratio (ACR) were determined. The independent associations of serum TB with CNF morphological parameters and clinical factors were assessed. Written informed consent was obtained from all subjects prior to participation. The ethics committee of Ishibashi Clinic approved the study protocol.

Serum TB (12.3 ± 0.45–14.3 ± 0.76 μmol/L) was similar among the four groups. GGT and hsCRP in T2D patients with severe hyperglycemia were higher than those in the other two groups (P < 0.0001), despite similar BMIs and alcohol consumption. All CNF parameters except CNBL in T2D patients differed from those of control subjects (P < 0.0001 to <0.001). In patients with severe hyperglycemia, serum TB was positively associated with CNFD, CNFL, CNBD, and CNBL, and was inversely associated with tortuosity. By contrast, log-hsCRP, GGT, triglycerides, log-ACR, and smoking habit were negative predictors of CNFD, and systolic blood pressure (SBP), GGT, triglycerides, log-ACR, and smoking habit were negative predictors of CNFL. SBP, GGT, log-ACR, and smoking habit were positive predictors of tortuosity (Table 1). In patients with moderate, but not mild, hyperglycemia, serum TB levels were associated only with CNFD (P = 0.039) and CNFL (P = 0.023). Serum TB levels were negatively correlated with hsCRP (P = 0.017–0.031), triglycerides (P = 0.005 to <0.0001), and smoking habit (P = 0.022 to <0.0001) in all diabetes subgroups.

In uncontrolled T2D patients with severe hyperglycemia who were exposed to marked OS, normal serum TB was associated with CNF morphology. Triglycerides, hsCRP, and smoking might influence CNF pathology in part by lowering serum TB levels. The present cross-sectional study identified an association between the TB level and CNF changes. The prospective study clarifies a pivotal role of lower serum TB levels in CNF degeneration in uncontrolled T2D that is similar to its role in lower-limb amputation (3).

• © 2014 by the American Diabetes Association.