Baseline characteristics of the study participants
In 2011, a total of 7,190 participants were involved in this study, including 3,384 (47.1%) men and 3,806 (52.9%) women. The median age of the participants was 58.0 (13.00) years. The baseline characteristics of the participants in 2011 were analysed based on whether they had sarcopenia, as shown in Table 1. A total of 1,139 (15.8%) participants had sarcopenia, and there were significant differences (p<0.001) in age, BMI, muscle mass, grip strength, education level, and triglyceride levels between those with and without sarcopenia. The TyG index of participants with sarcopenia was significantly lower than that of participants without sarcopenia (p<0.001). The median TyG index was 8.51 (0.73). The baseline characteristics of participants in 2011 were analysed based on quartiles of the TyG index, as shown in Table S1. Regardless of sex, the group with the highest TyG index had greater muscle mass and BMI than did the group with the lowest TyG index (p<0.001). The incidence of sarcopenia in the group with the lowest TyG index (21.0%) was significantly greater than that in the group with the highest TyG index (8.9%) (p<0.001).
Table 1 Study participants characteristics according to sarcopenia in 2011.
|
No sarcopenia
n=6051
|
Sarcopenia
n=1139
|
P-value
|
Age (years)
|
57.00(13.00)
|
69.00(12.00)
|
<0.001
|
Sex
|
|
|
0.362
|
Male
|
2862(47.3)
|
522(45.8)
|
|
Female
|
3189(52.7)
|
617(54.2)
|
|
BMI(kg/m2)
|
23.53(4.30)
|
19.05(2.10)
|
<0.001
|
ASM/Ht2(kg/m2)
|
7.00(1.58)
|
5.18(1.87)
|
<0.001
|
HS(kg)
|
31.50(14.00)
|
24.50(12.00)
|
<0.001
|
Married or not
|
|
|
<0.001
|
Yes
|
5446(90.0)
|
847(74.4)
|
|
No
|
605(10.0)
|
292(25.6)
|
|
Smoke or not
|
|
|
0.010
|
Yes
|
2371(39.2)
|
493(43.3)
|
|
No
|
3679(60.8)
|
646(56.7)
|
|
Drinking or not
|
|
|
0.003
|
Yes
|
2082(34.4)
|
340(29.9)
|
|
No
|
3969(65.6)
|
799(70.1)
|
|
Education
|
|
|
<0.001
|
Primary school and below
|
4051(67.0)
|
1015(89.2)
|
|
Middle school
|
1354(22.4)
|
90(7.9)
|
|
High school and above
|
645(10.7)
|
33(2.9)
|
|
TG(mg/dl)
|
103.545(74.34)
|
86.730(52.22)
|
<0.001
|
FBG(mg/dl)
|
100.26(14.22)
|
99.54(14.40)
|
0.136
|
TyG index
|
8.54(0.74)
|
8.36(0.63)
|
<0.001
|
HbA1c(%)
|
5.10(0.50)
|
5.10(0.50)
|
0.478
|
HDL-C(mg/dl)
|
49.485(18.56)
|
56.830(20.88)
|
<0.001
|
TC(mg/dl)
|
189.434(47.17)
|
188.661(49.87)
|
0.073
|
LDL-C(mg/dl)
|
114.434(43.01)
|
110.954(43.69)
|
<0.001
|
Hypertension
|
1307(21.7)
|
194(17.1)
|
0.001
|
Dyslipidemia
|
258(4.3)
|
23(2.1)
|
<0.001
|
Cancer or Malignant Tumour
|
59(1.0)
|
7(0.6)
|
0.242
|
Digestive disease
|
1384(22.9)
|
325(28.6)
|
<0.001
|
Chronic Lung diseases
|
532(8.8)
|
198(17.5)
|
<0.001
|
Heart Problems
|
578(9.6)
|
128(11.3)
|
0.081
|
Stroke
|
99(1.6)
|
22(1.9)
|
0.478
|
Emotional, Nervous,
or Psychiatric Problems
|
84(1.4)
|
12(1.1)
|
0.370
|
Memory-Related disease
|
56(0.9)
|
18(1.6)
|
0.045
|
Arthritis or Rheumatism
|
2081(34.5)
|
439(38.5)
|
0.008
|
Asthma
|
191(3.2)
|
63(5.5)
|
<0.001
|
BMI, body mass index; ASMI, ASM/height (m)2 ; HS, hand strength; TG, triglyceride; FBG, fasting blood glucose; TyG, triglyceride glucose; HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol.
During the 4-year follow-up, after excluding those who were lost to follow-up or who had incomplete information, 520 of the 4,034 participants newly developed sarcopenia. Baseline analyses were performed accordingly, as shown in Table S2. There were significant differences (p<0.001) in age, BMI, muscle mass, grip strength, smoking status, education level, and triglyceride levels between patients with newly developed sarcopenia and those without sarcopenia, consistent with the cross-sectional results from 2011. The TyG index of participants with newly developed sarcopenia was significantly lower than that of participants without sarcopenia (p<0.001).
Cross-sectional associations of the TyG index with sarcopenia
Table 2 illustrates the relationships between the TyG index and sarcopenia as well as its components according to the 2011 cross-sectional data. According to the unadjusted model, the TyG index showed a significant negative correlation with sarcopenia (OR=0.537, 95% CI: 0.476, 0.607). The components of sarcopenia, including low muscle mass (0.500 [0.447, 0.560]), low muscle strength (0.847 [0.741, 0.967]), and low physical function (0.914 [0.840, 0.994]), were also negatively correlated with the TyG index (p<0.05). After adjusting for sex, age, education level, marital status, smoking status, and alcohol consumption, a significant negative correlation remained between the TyG index and sarcopenia [0.454 (0.391, 0.527)] as well as low muscle mass [0.442 (0.388, 0.503) (p<0.001)], although the effect was attenuated after adjustment. However, after further adjustment for glycated haemoglobin and high-density lipoprotein, a significant negative correlation was observed between the TyG index and sarcopenia, low muscle mass, and low physical function (correlation coefficient > 0.6). According to the fully adjusted model (Model 5), the TyG index showed a significant negative correlation with sarcopenia, low muscle mass, and low physical function (p<0.01). As the model was adjusted, the negative correlation between the TyG index and sarcopenia as well as low muscle mass became increasingly stronger. The negative correlation between low physical function (OR=0.840, 95% CI: 0.739, 0.955) and the TyG index was stronger than that between sarcopenia (OR=0.699, 95% CI: 0.588, 0.830) and low muscle mass (OR=0.677, 95% CI: 0.583, 0.786) with the TyG index. The results in Table S3 also show that compared with those of the lowest quartile of the TyG index, the TyG indices of Q3 and Q4 were significantly negatively correlated with sarcopenia (p<0.05). The restricted cubic spline regression revealed a nonlinear association, showing a significant negative correlation with sarcopenia when TyG > 8.51 (p<0.001), as shown in FIGURE 2. The combined TyG indices, including TyG-WC, TyG-BMI, and TyG-WtHR, exhibited a significant negative correlation with sarcopenia (p<0.001), but the negative correlation was weaker than that of the TyG index, as shown in Table S4.
Table 2 Association of the TyG index with sarcopenia in 2011(n=7190).
Models
|
Sarcopenia
(n=1139)
OR(95%CI)
|
P-value
|
Low muscle mass
(n=1422)
OR(95%CI)
|
P-value
|
Low muscle strength
(n=817)
OR(95%CI)
|
P-value
|
Low physical function
(n=4062)
OR(95%CI)
|
P-value
|
Model 1
|
0.537(0.476, 0.607)
|
<0.001
|
0.500(0.447, 0.560)
|
<0.001
|
0.847(0.741, 0.967)
|
0.014
|
0.914(0.840, 0.994)
|
0.036
|
Model 2
|
0.457(0.394, 0.531)
|
<0.001
|
0.446(0.392, 0.507)
|
<0.001
|
0.907(0.782, 1.051)
|
0.195
|
0.916(0.823, 1.020)
|
0.109
|
Model 3
|
0.454(0.391, 0.527)
|
<0.001
|
0.442(0.388, 0.503)
|
<0.001
|
0.898(0.774, 1.042)
|
0.157
|
0.915(0.822, 1.019)
|
0.106
|
Model 4
|
0.672(0.569, 0.795)
|
<0.001
|
0.650(0.562, 0.752)
|
<0.001
|
0.856(0.723, 1.014)
|
0.071
|
0.852(0.752, 0.966)
|
0.012
|
Model 5
|
0.699(0.588, 0.830)
|
<0.001
|
0.677(0.583, 0.786)
|
<0.001
|
0.861(0.724, 1.024)
|
0.091
|
0.840(0.739, 0.955)
|
0.008
|
Model 1: No adjustment; Model 2: Adjusted for age, sex, education, and marital status; Model 3: model 2 + ever/current smoke, ever/current alcohol; Model 4: model 3 + glycated haemoglobin, HDL-C; Model 5: model 4 + comorbidities.
comorbidities: hypertension, dyslipidaemia, cancer or malignant tumour, stomach or other digestive disease, chronic Lung Diseases, heart problems, stroke, emotional, nervous, or psychiatric problems, memory-related disease, arthritis or rheumatism, asthma.
Longitudinal association between the baseline TyG index and incident sarcopenia at follow-up, 2011−2015
Table 3 indicates that during the four-year follow-up, 520 (12.9%) participants were found to have sarcopenia. After full adjustment of the model, the TyG index remained significantly negatively correlated with the incidence of sarcopenia (p=0.005). Each SD increase in the TyG index was associated with a 23.8% reduction in the risk of developing sarcopenia (HR=0.762, 95% CI: 0.629, 0.923). Specifically, the adjusted hazard ratio for sarcopenia in the highest quartile of the TyG index compared to the lowest quartile was 0.719 (p=0.025), indicating a 28.1% reduced risk of incident sarcopenia, which is consistent with the cross-sectional results. The TyG-WC, TyG-BMI, and TyG-WtHR also showed significant negative correlations with incident sarcopenia (p<0.001), as shown in Table S5. Subgroup analysis by age and sex revealed that the highest quartile of the TyG index was negatively correlated with incident sarcopenia among subjects aged 60 and above (HR=0.618, 95% CI: 0.433, 0.882, p=0.008) and men (HR=0.634, 95% CI: 0.414, 0.971, p=0.036) but was not associated with incident sarcopenia among women, as shown in Table S6.
Table 3 Cox-proportional hazard models for the association between triglyceride glucose index and incident sarcopenia.
|
Case/total
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
TyG index
|
|
|
|
|
|
|
Per SD increase
|
|
0.631(0.538, 0.741)
|
0.576(0.486, 0.681)
|
0.578(0.488, 0.684)
|
0.719(0.595, 0.869)
|
0.762(0.629, 0.923)
|
P-value
|
|
<0.001
|
<0.001
|
<0.001
|
0.001
|
0.005
|
Quartiles
|
|
|
|
|
|
|
Q1
|
161/984
|
Ref
|
Ref
|
Ref
|
Ref
|
Ref
|
Q2
|
149/1033
|
0.882(0.705, 1.102)
0.267
|
0.872(0.697, 1.090)
0.229
|
0.869(0.695, 1.088)
0.221
|
0.951(0.759, 1.193)
0.666
|
0.976(0.775, 1.230)
0.838
|
Q3
|
122/988
|
0.755(0.596, 0.955)
0.019
|
0.702(0.554, 0.890)
0.004
|
0.704(0.556, 0.893)
0.004
|
0.848(0.662, 1.085)
0.190
|
0.881(0.683, 1.136)
0.328
|
Q4
|
88/1029
|
0.523(0.403, 0.678)
<0.001
|
0.484(0.372, 0.629)
<0.001
|
0.485(0.373, 0.631)
<0.001
|
0.664(0.499, 0.885)
0.005
|
0.719(0.539, 0.960)
0.025
|
Model 1: No adjustment; Model 2: Adjusted for age, sex, education, and marital status; Model 3: model 2 + ever/current smoke, ever/current alcohol; Model 4: model 3 + glycated haemoglobin, HDL-C; Model 5: model 4 + co-morbidities. Q1<8.19, 8.19≤Q2<8.55, 8.55≤Q3<8.93, Q4≥8.93.