Study design

This is a cross-sectional study conducted by the Endocrine Care Center of the Fourth People’s Hospital in Guiyang, China. In 2021, a health management program was launched for elderly patients with diabetes. The aim of this program is to evaluate the health condition of elderly diabetic patients through a series of physical examinations and health monitoring, thereby preventing the development of chronic complications such as diabetic nephropathy, diabetic foot, and diabetic retinopathy. Additionally, the program includes monitoring for sarcopenia.

Study population

This study prospectively enrolled patients who participated in a health management program for older patients with diabetes at our endocrine care center from May 2021 to February 2023. The patient inclusion process was non-selective and consecutive. The diagnosis of type-2 diabetes mellitus was based mainly on the Chinese Diabetes Control Guidelines 2020 Edition published by the Chinese Diabetes Association. The inclusion criteria were: (1) age greater than or equal to 65 years, (2) clear history or diagnosis of type 2 diabetes. Exclusion criteria were: (1) age younger than 65 years, (2) liver or renal insufficiency, (3) various types of acute infections in combination, (4) tumors, (5) thyroid disease, (6) autoimmune diseases, (7) combined connective tissue diseases, (8) acute myocardial infarction, (9) acute cerebral infarction and cerebral hemorrhage, (10) type 1 diabetes, diabetic nephropathy or hyperosmolar coma, (11) long-term bed rest, local braking or hemiplegia. This study was approved by the Ethics Committee of the Fourth People’s Hospital (approval number: [2021]No 005). All patients and their blood relatives were informed in detail about the health management program for elderly diabetic patients, signed informed consent forms, and agreed to use their clinical information and measurement data for data analysis under the premise of anonymisation of personal privacy. Initially, 318 elderly diabetic patients were enrolled, and based on a series of nadir criteria, 288 remained for final data analysis. The specific screening process is detailed in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The process of enrolling patients involved in this study.

https://doi.org/10.1371/journal.pone.0292654.g001

Variable

Missing data addressing

Because the percentage of missing data in this study was less than 5%, no multiple imputation was employed (S2 Table in S1 File).

Statistical analysis

We summarized continuous variables using means and standard deviations, and categorical variables using numbers and proportions. Additionally, we categorized participants into three groups based on their fasting c-peptide tertiles and used chi-square tests for categorical variables and one-way ANOVA for continuous variables to analyze differences between the three groups. Multivariable logistic regression was used to investigate the relationship between fasting c-peptide and sarcopenia. We displayed model 1 with no adjusted covariates, model 2 which was only adjusted for age, sex, education level and marital status, and model 3 which was adjusted for all covariates presented in Table 1 simultaneously. We divided the fasting C-peptide by 100 to facilitate observation of changes in OR. Therefore, the results should be interpreted per 100 unit change.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of elder patients with diabetes categorized by fasting C-peptide (tertile).

https://doi.org/10.1371/journal.pone.0292654.t001

As logistic regression is not well-suited for analyzing non-linear associations, we utilized a generalized additive model (GAM) to investigate the relationship between fasting C-peptide and sarcopenia. Initially, we smoothed the curve to observe the true association between fasting C-peptide and sarcopenia, and then determined the inflection point values using a recursive algorithm. Finally, effect values (Odds ratio) and 95% confidence intervals for fasting C-peptide and sarcopenia were calculated using a two-piecewise linear model on both sides of the inflection point. To ensure the robustness of our findings, we conducted a series of sensitivity analyses: (1) Fasting C-peptide was divided into three categories and P for trend was calculated to observe the robustness of the association between fasting C-peptide as a continuous variable and as a categorical variable with sarcopenia. (2) We used a log-likelihood ratio test to determine whether a two-piecewise linear model or a logistic regression model more accurately fit the association between fasting C-peptide and sarcopenia. All analyses were performed using the statistical packages R Foundation (http://www.R-project.org) and EmpowerStats (http://www.empowerstats.com,X&Y Solutions, Inc., Boston, MA). Statistical significance was defined as a two-sided p-value less than 0.05.

The baseline characteristics of Chinese HF patients

We analyzed and compared the distribution of patient characteristics between the sarcopenic and non-sarcopenic groups. The results are presented in Table 1. Our study included elderly diabetic patients, with a mean age of 69.74 ± 6.28 years, and a male-to-female ratio of 1:1.6, with males accounting for 37.5% (108/288) of the total patients. Of the patients, 27.43% (79/288) were found to have sarcopenia. Compared to non-sarcopenic patients, those in the sarcopenic group had a lower body mass index, fasting plasma C-peptide levels, hemoglobin levels, and Bartel Index (indicating poorer physical function). In contrast, elderly diabetic patients in the sarcopenic group had higher fat mass, duration of diabetic load, combined cardiovascular disease, hypertension, diabetes mellitus, osteoporosis, smoking and alcohol consumption compared to the non-sarcopenic group. higher. There were no statistically significant differences in HbA1C levels, globulin, education, gender, marital status, hyperlipidemia, and COPD between the two groups.

The association between fasting C-peptide and sarcopenia obtained from logistic regression model, both univariate and multivariable

We investigated the association between fasting C-peptide and sarcopenia using different covariate adjustment strategies (Table 2). In model 1 (the unadjusted model), the results indicated that there was no significant correlation between fasting C-peptide and the prevalence of sarcopenia (odds ratio [OR], 0.92; 95% confidence interval [CI], 0.85 to 1.00; P = 0.065). The same results were obtained from model 2. Even after adjusting for sex, age, education level, and marital status, we still did not observe a substantial association between fasting C-peptide (per 100 pmol/L) and sarcopenia (odds ratio [OR], 0.92; 95% confidence interval [CI], 0.84 to 1.01; P = 0.085). When we adjusted for all the covariates listed in Table 1 (model 3, fully-adjusted model), we found a 4% decrease in the probability of sarcopenia for each 100 pmol/L increase in fasting C-peptide. Nevertheless, due to the wider distribution of confidence intervals, we could not reject the null hypothesis (odds ratio [OR], 0.96; 95% confidence interval [CI], 0.85 to 1.07; P = 0.442).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associations between fasting C-peptide and sarcopenia using logistic regression models.

https://doi.org/10.1371/journal.pone.0292654.t002

To conduct a sensitivity analysis, we transformed fasting C-peptide into a categorical variable based on tertile and assessed whether the outcomes remained consistent when fasting C-peptide was used as either a categorical or continuous variable (Table 2). The p-values of the trend test for the three models were 0.056 (model 1), 0.093 (model 2), and 0.418 (model 3), respectively. These outcomes were consistent with the findings when fasting C-peptide was applied as a continuous variable.

The nonlinearity addressing

To investigate the potential nonlinear relationship between fasting C-peptide and sarcopenia, we utilized a GAM model and smoothed curve fitting to reveal a U-shaped correlation between these two variables (Fig 2). To better explicate this U-shaped curve, we applied a recursive algorithm to calculate the inflection points of fasting C-peptide, which were determined to be 744 and 939 pmol/L. Thereafter, we deployed a two-piece linear model to estimate the effect values and confidence intervals on both sides of the inflection points (Table 3). Our results showed that for every 100 pmol/L increase in fasting C-peptide within the range of 50–744 pmol/L, there was a tendency for the probability of sarcopenia to decrease by 37% (odds ratio [OR] 0.63; 95% confidence interval [CI] 0.49 to 0.83; P < 0.001). On the contrary, we did not observe a substantial association between fasting c-peptide and the probability of sarcopenia when fasting c-peptide was within the 744–939 pmol/L interval (odds ratio [OR] 1.27; 95% confidence interval [CI] 0.34 to 4.59; P = 0.720). However, when fasting C-peptide was in the 939–1694 range, each 100 pmol/L increase in fasting C-peptide was associated with a 76% increase in the probability of sarcopenia (odds ratio [OR] 1.76; 95% confidence interval [CI] 1.11 to 2.81; P = 0.017)). Additionally, upon comparing the logistic regression model fit with the two-piece linear model fit, the log likelihood ratio yielded a P value of less than 0.001, indicating that the two-piece linear model provided a better fit for the association between fasting C-peptide and sarcopenia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. A non-linear relationship between fasting C-peptide and sarcopenia.

The level of fasting C-peptide is plotted on the horizontal axis, while the probability of sarcopenia is plotted on the vertical axis. The middle line reflects the trend in the association between C-peptide and sarcopenia, while the upper and lower lines indicate the corresponding 95% confidence intervals.

https://doi.org/10.1371/journal.pone.0292654.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Exploration of nonlinear association between fasting C-peptide and sarcopenia using two-piecewise linear model.

https://doi.org/10.1371/journal.pone.0292654.t003

We conducted a single-centre cross-sectional study in an elderly Chinese diabetic population to investigate the association between fasting plasma C-peptide and sarcopenia. The results showed that both low and high fasting C-peptide levels were associated with a higher probability of sarcopenia, indicating a U-shaped association between C-peptide and sarcopenia. We observed that fasting C-peptide was associated with a lower probability of sarcopenia within the range of 774 to 939 pmol/L.

Previous evidence on the association of fasting C-peptide with sarcopenia is insufficient. Lu YX et al [20]. found a negative association between fasting C-peptide and the risk of sarcopenia in 41 diabetic patients and 205 non-diabetic patients, but their study only included 41 diabetic patients and diabetes status was only adjusted as a covariate. In addition, they only adjusted for sex, age, diabetes status, and total body fat mass. Our study explored possible non-linear associations and found negative associations between fasting C-peptide and sarcopenia in the 50–774 pmol/L range, which is partially consistent with the results of Hitomi Miyake et al [21]. and Tanaka K et al [22]. Hitomi Miyake et al. found that fasting C-peptide values were significantly higher in patients in the high skeletal muscle content group than in the low skeletal muscle content group in a study on the association between low skeletal muscle content and all-cause mortality in a diabetic population. Tanaka K et al. similarly found a negative association between C-peptide and sarcopenia in 191 men with type 2 diabetes but did not explore non-linear associations.

In contrast to earlier findings, our study discovered a link between high fasting C-peptide levels (939 pmol/L–1700 pmol/L) and sarcopenia in elderly diabetes patients, which we believe is the first of its kind. We propose many explanations for this unusual U-shaped association trend: First, C-peptide is an insulin precursor that, unlike insulin, is filtered by the kidneys and eliminated in the urine rather than being metabolized by the liver. As a result, monitoring C-peptide levels can reflect islet cell function without being influenced by external influences such as food [23]. Low fasting C-peptide levels may indicate insulin production and release inhibition, resulting in metabolic inefficiency, decreased muscle mass, and an increased risk of sarcopenia [24]. Previous research has shown that insulin is a key regulator of glucose absorption and use in peripheral tissues, including muscle [25]. Inadequate insulin secretion and release caused by low C-peptide levels can result in diminished insulin signaling and impaired glucose metabolism, resulting in decreased amino acid absorption and protein synthesis in muscle, resulting in decreased muscle mass and an increased risk of sarcopenia [25, 26]. Furthermore, low insulin levels stimulate glucagon secretion, which causes muscle proteolysis and amino acid release into the circulation, increasing the risk of sarcopenia [27]. High fasting C-peptide levels, on the other hand, may indicate insulin resistance [28, 29]. As a result, increased insulin signaling, particularly in peripheral tissues, might result in decreased muscle mass. Sarcopenia is caused by three mechanisms: (1) increased proteolytic metabolism and decreased protein synthesis in skeletal muscle; (2) increased expression of the FoxO family, which either directly or indirectly attenuates skeletal muscle; and (3) autophagy in skeletal muscle cells [29]. Furthermore, Sara Manrique-Arija et al [30]. discovered that insulin resistance was associated with higher levels of CRP and interleukin-1 in a cross-sectional study of arthritis patients, whereas previous literature confirmed that muscle atrophy is a major determinant of vulnerability in old age and is associated with high levels of inflammation [31]. Thus, despite the fact that this study was merely cross-sectional in nature, the association between high C-peptide, insulin resistance, inflammation, and muscle atrophy may be the cause of the positive correlation between these variables. To establish definitive evidence, additional high-level investigations, including cohort studies and randomized controlled clinical trials, are required. Second, high levels of C-peptide point to poorly controlled diabetes. On the other hand, prior clinical research indicates that vulnerability in older people is frequently linked to inadequate chronic disease management [32]. As a result, this may also be one of the possible causes of the positive association between high C-peptide levels and muscle atrophy. Lastly, 25-hydroxyvitamin D3 in elderly people is frequently linked to muscular atrophy [33]. We used the cut points of 7.74 and 9.39 to divide the population into three categories. The levels of 25-hydroxyvitamin D3 were substantially greater in the two low C-peptide groups (7.74 and 7.74–9.39) than in the high C-peptide group (>9.39), according to our research (S1 Fig in S1 File). Therefore, the considerable drop in 25-hydroxyvitamin D3 levels may potentially be linked to the positive connection between a high C-peptide group and muscle atrophy.

Our study reports a U-shaped association between fasting C-peptide levels and the risk of sarcopenia, which, to our knowledge, has not been previously reported in the literature. Our findings suggest that in elderly diabetic patients, maintaining appropriate C-peptide levels is crucial for normal metabolism and muscle mass and may prevent the development of sarcopenia. While our study provides relevant clinical data for future research on this topic, we acknowledge that cross-sectional evidence is relatively weak and further studies, such as prospective cohort studies, are required for validation. We continue to follow up with elderly people with diabetes mellitus but without concomitant sarcopenia in our center, and as such, our findings have important implications for future research on this topic. By utilizing easily accessible, simple, and inexpensive assessments such as fasting C-peptide levels, healthcare professionals can identify patients at risk of sarcopenia early on and take preventive measures accordingly. In summary, our study supports the notion that monitoring C-peptide levels in elderly diabetic patients is important for maintaining normal metabolism and muscle mass, thus avoiding the development of sarcopenia. We hope that our findings will promote further research in this area and eventually lead to the adoption of more effective prevention and treatment strategies for sarcopenia.

Our study has several strengths that contribute to its validity. Firstly, we utilized advanced algorithms (GAM model and two-piece linear model) which allowed us to effectively detect a true association between fasting C-peptide levels and the risk of sarcopenia in elderly Chinese diabetics. Secondly, we conducted a series of sensitivity analyses and utilized different variable tuning strategies, reducing the possibility that our findings were merely incidental and increasing the robustness of our results. Thirdly, our study’s prospective, continuous, and non-selective inclusion of patients, combined with the application of blinding to reduce possible distortion of results due to selection bias and observation bias.

However, our study does have some limitations that must be taken into account when interpreting our findings. Firstly, the study population was limited to Chinese diabetics and caution should be taken when extrapolating our findings to other races. Secondly, our results are only applicable to diabetic patients over 65 years of age and cannot be generalized to other populations (e.g., type 1 diabetes, hepatic and renal insufficiency, tumor patients, autoimmune system diseases, etc.). Thirdly, as an observational cross-sectional study, our findings are subject to confounding, although we have rigorously adjusted for confounders and assessed the robustness of our results through sensitivity analysis. Fourthly, the nature of the cross-sectional study design limits our ability to assess causality; we can only observe associations. Finally, we can only adjust for measurable confounders, making it impossible to adjust for unmeasurable confounding factors. Therefore, future clinical studies with higher levels of evidence in larger populations are warranted to validate our findings.