This retrospective cohort study included 292 individuals diagnosed with type 2 diabetes who received examinations at our hospital over the period from January 2023 to December 2023. They were divided into the DF group (n = 82) and nondiabetic foot group (n = 210, NDF group). This study was approved by the institutional review board and ethics committee of Shaanxi Provincial People’s Hospital. Given its retrospective design, this study only used data from unidentified patients and informed consent was waived.

The inclusion criteria were as follows: (1) Diagnosis of type 2 diabetes as per World Health Organization guidelines[14]; (2) Aged between 18 and 80 years; and (3) Ineligibility for surgical revascularization.

The exclusion criteria were as follows: (1) Prior occurrence of acute coronary syndrome, myocardial infarction, or transient ischemic stroke within the last 6 months; (2) Presence of uncontrolled immune disorders or active severe systemic infections; (3) Severe hematologic disorders or coagulation abnormalities; (4) History of malignant tumors; (5) Participation in other clinical trials within the preceding 3 months; or (6) Other concerns identified by the investigators that may impede compliance or safety.

The patients were grouped based on the presence or absence of DF: DF group (n = 82) and NDF group (n = 210). The diagnostic criteria were based on the 2012 Infectious Diseases Society of America clinical practice guidelines for DF infections[14].

Patient demographic data were acquired from the medical records system. Upon admission, 5 mL of fasting blood sample was obtained from the antecubital vein in the morning for blood testing. Hematological parameters such as hemoglobin (g/dL), hematocrit (%), white blood cell count (× 10⁹/L), and platelet count (× 10⁹/L) were measured using a fully automated coagulation analyzer (HC00608166, STA Compact, China).

Biochemical parameters and genomic and proteomic data were analyzed using venous blood samples of 5-7 mL (including whole blood, plasma, and serum). The measured biochemical markers included HbA1c, plasma glucose, total cholesterol (measured by the CHOD-PAP method), low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, urea, and creatinine. All quantitative measurements were performed using a BS 400 auto-analyzer in accordance with the procedures provided by Dia Sys Diagnostic Systems GmbH, Germany. The glomerular filtration rate (GFR) was computed from serum creatinine levels using the formula: GFR (mL/minute/1.73 m²) = (140 - age) × weight (kg)/72 × serum creatinine (mg/dL).

Delay in seeking medical care (patient delay) refers to the behavior of individuals who fail to seek timely medical attention after noticing abnormal symptoms due to objective or subjective reasons. The time of medical delay is defined as the interval between the time the patient first noticed the symptoms and their first visit to a healthcare facility, with a duration of over 3 months considered as medical delay.

Attitudes Toward Medical Help Seeking Scale (ATMHSS): The ATMHSS is a self-assessment scale consisting of 35 items divided into four dimensions: Behavioral intention (12 items), nonfatalism (11 items), medical trust (7 items), and nonavoidant attitudes (5 items). Each item is rated on a Likert 4-point scale, with “disagree” to “agree” corresponding to scores of 0-3. The total score ranges from 0 to 105, where a high score indicates a positive attitude toward seeking medical care. The scale has a Cronbach's α coefficient of 0.82.

The sample size was mainly depended on the number of patients included in the inclusion time frame complied with the events per variable > 10 principle. Using G*Power 3.1.9.7, we conducted a post hoc analysis based on the "Means: Difference between two independent means (two groups)" option for t-tests. We configured the analysis with a two-tailed mode, an effect size of d = 0.6, and a significance level (α err prob) of 0.05. After inputting the sample sizes for the two groups, we calculated the power (1 - β err prob), which yielded a result of 0.969.

Patient characteristics were compared between the two groups using independent t-tests for continuous variables and χ² tests for categorical variables. The normality of continuous variables was checked using the Shapiro-Wilk test. Normally distributed data were expressed as mean ± SD, and categorical variables were reported as counts and percentages. A P value of less than 0.05 was deemed statistically significant. All analyses were conducted with SPSS software, version 29.0 (SPSS Inc., Chicago, IL, United States).

Spearman correlation analysis was conducted. Indicators showing significant differences in the differential and correlation analyses were included as covariates in logistic regression analysis. The diagnostic efficiency of delayed medical care for DF was assessed using the area under the receiver operating characteristic (ROC) curve, and a combined predictive model was established by incorporating blood glucose levels, glycated protein levels, and delayed medical care.

Between-group comparison showed no statistically significant differences in general demographic data such as age and gender in the presence of DF (P > 0.05; Table 1). However, the body mass index (BMI) of the DF group was significantly higher than that of the NDF group (23.51 ± 3.57 vs 25.17 ± 3.57, P < 0.001). This finding suggests that an increase in BMI may elevate the incidence of DF.

Differential analysis was conducted on the occurrence probabilities of diabetes duration, classification, staging, and complications between the two groups (Table 2). No significant differences were observed in classification, staging, diabetic retinopathy, and diabetic vascular disease (P > 0.05). The duration of diabetes in the DF group was significantly longer than that in the NDF group (10.12 ± 4.95 vs 11.74 ± 4.84, P = 0.012). In addition, the occurrence rates of diabetic neuropathy (22.86% vs 40.2%, P = 0.005) and diabetic nephropathy (0.48% vs 4.88%, P = 0.035) were markedly lower in the NDF group compared with those in the DF group. These findings suggest that an increase in diabetes duration and the occurrence of some complications may contribute to the increased incidence of DF.

Routine blood examination indicated no statistical differences in hemoglobin concentration, white blood cell count, red blood cell count, neutrophil count, and platelet count between the two groups (P > 0.05; Figure 1). This finding suggests that preoperative blood routine indicators have no impact on the research results.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Blood Routine Examination. A: Hemoglobin concentration; B: Red blood cell count; C: White blood cell count; D: Neutrophil count; E: Platelet. NS: No significant difference.

Examination of the patients' blood glucose and lipid levels indicated no statistical differences in total cholesterol, triglycerides, LDL cholesterol, and HDL cholesterol between the two groups (P > 0.05; Figure 2). This finding suggests that preoperative blood routine indicators have no impact on the research results. However, the DF group exhibited significantly higher levels of plasma glucose (5.71 ± 0.61 vs 10.24 ± 4.72, P < 0.001) and HbA1c (7.41 ± 1.87 vs 9.12 ± 1.65, P < 0.001) than the NDF group. This result indicates that elevated blood glucose levels have a negative impact on DF occurrence.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Blood glucose and lipid levels. A: Plasma glucose; B: HbA1c; C: Total cholesterol; D: Triglycerides; E: Low-density lipoprotein cholesterol; F: High-density lipoprotein cholesterol. ^cP < 0.001; NS: No significant difference.

Differential analysis was conducted on the renal function indicators of the two groups, including plasma creatinine, plasma urea, and estimated GFR (Table 3). No significant differences in renal function indicators were found between the two groups (P > 0.05). This finding suggests that intraoperative indicators have no impact on the research results.

The patients’ delayed medical care of more than 3 months was recorded, and their willingness to seek medical care was assessed using the ATMHSS (Table 4). The NDF group had significantly higher ATMHSS scores (68.71 ± 10.41 vs 59.84 ± 9.78, P < 0.001) and lower rate of delayed medical care (13.41% vs 72.38%) compared with the DF group. This finding indicates that low willingness to seek medical care and delayed medical care may lead to DF.

A significant correlation was observed between various indicators and DF occurrence (Table 5). On the one hand, BMI, duration of diabetes, plasma glucose, HbA1c, and presence of diabetic peripheral neuropathy and diabetic nephropathy were positively correlated with DF occurrence. On the other hand, the ATMHSS score and duration of medical delay were negatively correlated with DF occurrence. These results underscore the predictive potential of these indicators for DF occurrence.

Logistic regression analysis revealed a significant finding (Table 6). The multivariate regression model demonstrated that BMI, duration of diabetes, plasma glucose, HbA1c, presence of diabetic peripheral neuropathy and diabetic nephropathy, ATMHSS score, and medical delay time all had a strong diagnostic value for DF.

A validation cohort of patients was used to establish a predictive model for DF occurrence based on factors such as BMI, duration of diabetes, blood glucose levels, complications, and medical delay (Table 7). The results demonstrated that plasma glucose [area under the curve (AUC) = 0.819], HbA1c (AUC = 0.804), and medical delay (AUC = 0.795) exhibited good predictive value. The combined predictive model yielded an AUC of 0.933 (Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Analysis of the combined predictive model for plasma glucose, HbA1c, and medical delay. ROC: Receiver operating characteristic; AUC: Area under the curve.