3.1 The baseline characteristics of diabetic patients with or without DSPN
The baseline characteristics of DSPN and non-DSPN group were shown in Table 1. When compared with patients of non-DSPN group, there was an increase in disease duration, HbA1c, CRP, BUN and a decrease in FINS, FC-P in those of DSPN group (all P < 0 .05). Additionally, diabetic patients with DSPN had lower levels of CTRP3 (P < 0.05). There was no significant difference in age, gender, smoking status, alcohol consumption, SBP, DBP, BMI, WC, WHR, TC, TG, LDL-C, HDL-C, FBG, HOMA-IR, Scr and UA identified between the two groups (P > 0.05).
Table 1 Baseline characteristics of the subjects
Characteristics
|
non-DSPN group(n =89)
|
DSPN group(n = 89)
|
t、Z or χ2
|
P value
|
Male/Female (n/n)
|
55/34
|
50/39
|
0.581
|
0.446
|
Age (years)
|
54 (48, 59)
|
56 (48, 64)
|
﹣1.566
|
0.120
|
Disease duration (months)
|
6.0 (1.0, 21.0)
|
12.0 (6.0, 48.0)
|
﹣4.690
|
< 0.001
|
Smoking [n (%)]
|
27 (30.3)
|
29 (32.6)
|
0.104
|
0.747
|
Alcohol intake [n (%)]
|
11 (12.4)
|
16 (18.0)
|
1.091
|
0.296
|
Systolic blood pressure (mmHg)
|
121 ± 16
|
123 ± 15
|
﹣0.845
|
0.399
|
Diastolic blood pressure (mmHg)
|
79 ± 10
|
78 ± 10
|
0.560
|
0.576
|
Body mass index (kg/m2)
|
24.60 ± 3.22
|
25.01 ± 3.60
|
﹣0.810
|
0.419
|
Waist circumference (cm)
|
85.8 ± 9.4
|
86.8 ± 10.7
|
﹣0.658
|
0.512
|
Waist–hip ratio
|
0.90 (0.85, 0.94)
|
0.92 (0.85, 0.96)
|
﹣1.352
|
0.176
|
Total cholesterol (mmol/L)
|
4.66 ± 1.15
|
4.70 ± 1.04
|
﹣0.222
|
0.825
|
Triglycerides (mmol/L)
|
1.70 (1.11, 2.64)
|
1.67 (1.04, 2.51)
|
﹣0.314
|
0.753
|
LDL-cholesterol (mmol/L)
|
2.67 ± 0.72
|
2.62 ± 0.66
|
﹣0.420
|
0.675
|
HDL-cholesterol (mmol/L)
|
1.21 (1.03, 1.42)
|
1.19 (1.02, 1.33)
|
﹣0.500
|
0.617
|
Fasting glucose (mmol/L)
|
7.32 (6.09, 9.73)
|
8.04 (6.00, 10,60)
|
- 1.191
|
0.234
|
Fasting insulin (pmo/L)
|
69.05 (39.00, 117.15)
|
51.79 (37.31, 70.13)
|
﹣2.451
|
0.014
|
Fasting C-peptide (nmo/L)
|
0.73 (0.51, 0.93)
|
0.60 (0.44, 0.79)
|
﹣2.074
|
0.038
|
HOMA-IR
|
3.20 (1.60, 5.57)
|
2.34 (1.57, 3.69)
|
﹣1.537
|
0.124
|
HbA1c (%)
|
8.1 (7.1, 10.8)
|
10.0 (7.8, 11.9)
|
﹣2.511
|
0.012
|
CRP (mg/L)
|
1.07 (0.67, 1.91)
|
1.63 (0.85, 3.63)
|
﹣2.411
|
0.016
|
Serum creatinine (umol/L)
|
58 (48, 70)
|
55 (47, 64)
|
﹣1.199
|
0.231
|
Blood urea nitrogen (mmol/L)
|
5.41 ± 1.30
|
5.88 ± 1.40
|
﹣2.328
|
0.021
|
Uric acid (umol/L)
|
311 ± 103
|
291 ± 99
|
﹣1.188
|
0.235
|
NSS score
|
0.63 ± 1.16
|
4.97 ± 2.10
|
-10.677
|
< 0.001
|
NDS score
|
0.43 ± 0.80
|
4.07 ± 1.56
|
-11.429
|
< 0.001
|
CTRP3 (ng/mL)
|
8.39 (7.60, 9.71)
|
6.38 (5.42, 7.44)
|
﹣6.069
|
< 0.001
|
Data are mean ± standard deviation, median (inter-quartile range) or percentage. HOMA-IR, homeostasis model assessment index-insulin resistance; HbA1c, glycosylated hemoglobin A1c; CRP, c-reactive protein; NSS, neuropathy symptom score; NDS, neuropathy disability score; CTRP3, C1q/TNF-related protein-3.
3.2 Relationship between serum CTPR3 levels and the severity of DSPN
Firstly, to assess the association between serum CTPR3 levels and the severity of DSPN, the participants were grouped according to numbers of neurological signs. As shown in Fig. 1A, serum CTRP3 levels were significantly decreased with increasing numbers of abnormal signs in DSPN screening (all P < 0.05). As well as we know, disorders of temperature sensation or pin-prick sensation indicate small-fiber neuropathy, and abnormal vibration sensation, pressure sensation and ankle reflexes commonly refer to large-fiber neuropathy. In Fig. 1B-C, it was found that serum CTRP3 levels were significantly lower in patients with at least one sign of damage in small-fiber nerve or large-fiber nerve as compared to those with negative signs (both P < 0.05).
Next, serum CTRP3 concentrations were stratified based on NSS score or NDS score, respectively. These two scoring systems are used to assess the severity of DSPN in diabetic patients. Similarly, we observed a progressive reduction in serum CTRP3 levels coupled with higher NSS score or NDS score, as shown in Fig. 1D-E (all P < 0.05).
At last, the subjects were divided into two groups according to the CTRP3 median. The positive rates of signs in DSPN screening, including sensation of pin-prick, pressure, vibration, temperature and ankle reflexes, were compared between the low CTRP3 group (serum CTRP3 ≤ 7.58ng/ml) and high CTRP3 group (serum CTRP3>7.58ng/ml). Interestingly, data in Fig. 2 displayed that the patients with lower CTRP3 levels had a higher proportion of positive signs in peripheral nervous system (all P < 0.05).
3.3 The correlation analysis between CTRP3 and other parameters
Spearmen correlation analysis was used to observe the correlation between CTRP3 and other variables, as shown in Table 2. Serum CTRP3 was negatively correlated with SBP (r = -0.214, P = 0.004), BMI (r =-0.388, P<0.001), WC (r = -0.249, P<0.001), WHR (r = -0.268, P<0.001), disease duration (r =-0.218, P = 0.003), NSS score (r =-0.358, P<0.001), and NDS score (r = -0.449, P<0.001). However, CTRP3 was found not to be associated with gender, age, smoking status, alcohol consumption, DBP, TC, TG, LDL-C, HDL-C, FBG, FINS, FC-P, HbA1c, CRP, Scr, BUN as well as UA (all P > 0.05).
Table 2
Correlation analysis between serum CTRP3 levels and other variables
Variables | Correlation coefficient | P value |
Gender | − 0.041 | 0.585 |
Age (years) | − 0.030 | 0.688 |
Disease duration(months) | − 0.218 | 0.003 |
Smoking [n (%)] | 0.032 | 0.670 |
Alcohol intake [n (%)] | -0.030 | 0.689 |
Systolic blood pressure (mmHg) | -0.214 | 0.004 |
Diastolic blood pressure (mmHg) | -0.144 | 0.055 |
Body mass index (Kg/m2) | -0.388 | < 0.001 |
Waist circumference (cm) | -0.249 | < 0.001 |
Waist–hip ratio | -0.268 | < 0.001 |
Total cholesterol (mmol/L) | 0.014 | 0.852 |
Triglycerides (mmol/L) | 0.002 | 0.984 |
LDL-cholesterol (mmol/L) | 0.047 | 0.534 |
HDL-cholesterol (mmol/L) | -0.097 | 0.196 |
Fasting glucose (mmol/L) | 0.102 | 0.176 |
Fasting insulin (pmo/L) | -0.048 | 0.660 |
Fasting C-peptide (nmo/L) | -0.085 | 0.280 |
HOMA-IR | -0.019 | 0.858 |
HbA1c (%) | 0.040 | 0.598 |
CRP (mg/L) | -0.132 | 0.089 |
Serum creatinine (umol/L) | 0.016 | 0.831 |
Blood urea nitrogen (mmol/L) | -0.088 | 0.243 |
Uric acid (umol/L) | -0.016 | 0.836 |
NSS score | -0.358 | < 0.001 |
NDS score | -0.449 | < 0.001 |
HOMA-IR homeostasis model assessment index-insulin resistance, HBA1c Glycosylated hemoglobin A1c, CRP C-reactive protein, NSS neuropathy symptom score, NDS neuropathy disability score, CTRP3 C1q/TNF-related protein-3. |
3.4 The effect of serum CTRP3 levels on the incidence of DSPN
The independent influence of serum CTRP3 levels on the incidence of DSPN was determined using stepwise binary logistic regression analysis. We introduced DSPN (yes or no) as an independent variable. After excluding collinearity factors, gender, age, disease duration, smoking status, alcohol consumption, SBP, DBP, BMI, WC, WHR, TG, HDL-C, FBG, FC-P, HbA1c, CRP, Scr, BUN, UA, CTRP3 served as dependent variables. Then, we entered variables using the forward method. As shown in Table 3, several variables, including CTRP3, disease duration, HbA1c, CRP and Fasting C-peptide finally entered into the regression equation. The results indicated the independent role of CTRP3 in the occurrence of DSPN for diabetic patients (β=-0.316, P < 0.001). In another word, there was a 72.9% increase in the odds of DSPN for each 1 ng/ml decrease in serum CTRP3 levels.
Table 3
Analysis of independent risk factors for the incidence of DSPN
Variables | B | SE | Wald | P value | ORs (95%CI) |
Constant | − 0.045 | 1.136 | 0.002 | 0.968 | 0.956 |
CTRP3 | − 0.316 | 0.086 | 13.665 | < 0.001 | 0.729 (0.616–0.862) |
Disease duration | 0.039 | 0.010 | 13.980 | < 0.001 | 1.039 (1.019–1.061) |
HbA1c | 0.194 | 0.085 | 5.128 | 0.024 | 1.214 (1.026–1.435) |
CRP | 0.332 | 0.116 | 8.142 | 0.004 | 1.394 (1.110–1.751) |
Fasting C-peptide | − 1.001 | 0.463 | 4.672 | 0.031 | 0.368 (0.148–0.911) |
FINS, HOMA-IR, TC, LDL-C did not enter the equation because of collinearity. CTRP3, c1q/TNF-related protein-3; HBA1c, glycosylated hemoglobin A1c; CRP, C-reactive protein. |
Additionally, to explore the predictive value of serum CTRP3 on DSPN, we analyzed the receiver operating characteristic (ROC) curves. As shown in Fig. 3, the best cutoff value for serum CTRP3 to predict DSPN was 7.55ng/ml with sensitivity, 78.7%; specificity, 79.8%; the area under the curve, 0.763 (95% CI 0.689–0.838, P < 0.001).