Study design and participants

In this retrospective cohort study, we reviewed the medical records of patients with diabetic foot infections at Taiyuan Central Hospital from January 2020 to September 2023.During the study period, 406 patients were diagnosed with diabetic foot. After excluding patients with Wagner stage 2 and below, cancer and patients with incomplete data. And because the study population was patients with diabetes-related foot ulcers, other inflammatory causes of foot breakdown such as Buerger’s disease, inflammatory bowel disease, and rheumatic diseases in diabetic patients were also excluded. A total of 301 patients with Wagner stage 3 or higher diabetes-related foot ulcers combined with moderate to severe infections were included in the research (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of the patient selection process.

https://doi.org/10.1371/journal.pone.0323704.g001

This was a cross-sectional study and based on the access to the incidence of diabetic foot infection was 15% [11] with a tolerance error of 0.03, a two-sided test was set up with a = 0.05, and the sample size was calculated to be196.Considering the 20% loss to follow-up rate, a minimum of 236 patients with diabetic foot infection would need to be investigated.

Covariates

The following potential covariates were analysed in this study, including sex, age, body mass index (BMI) and smoking status, mean arterial pressure, hypertension, coronary heart disease, cerebrovascular disease, fasting plasma glucose, glycosylated haemoglobin, erythrocyte count, haemoglobin, white blood cell count, neutrophil percentage, platelet count, C-reactive protein (CRP), albumin, duration of diabetes, DFI microbial community, severity of diabetic foot infection, amputation status. Smoking status was categorized as current, former, or never smoker [12]. The BMI is the ratio of height to weight squared. Comorbidities(Hypertension, coronary heart disease, and cerebrovascular disease)was determined by documenting in the electronic medical record whether or not a physician has made a previous diagnosis. The DFI microbial community is the result of general bacterial cultures of localised tissue from affected foot ulcers. Amputations were classified as minor or major. Minor amputations were defined as those where the ankle was left intact, and major amputations were above the ankle [13].

Identification of DFI severity

The International Working Group on the Diabetic Foot has expanded the criteria for diagnosing and evaluating the severity of DFI [14]. 1) Mild infection: Diabetic foot ulcer with no systemic manifestations involving only the skin or subcutaneous tissue (not any deeper tissues) or any erythema that does not extend > 2 cm. 2) Moderate infection:The ulcers deep to bone and fascia with erythema extending > 2 cm and without systemic manifestations.3) Severe infection: The presence of two or more systemic symptoms, such as body temperature exceeding 38°C or falling below 36°C, a heart rate surpassing 90 beats/min, a respiratory rate higher than 20 breaths/min, a partial pressure CO2 level below 4.3 kPa (32 mmHg), a WBC count exceeding 12,000 or falling below 4,000 cells/mm³, or the presence of more than 10% immature (banded) leukocytes.

Ethics statement

Approval for this research was granted by the Ethics Review Committee of Taiyuan central hospital (2023046), ensuring full compliance with the ethical guidelines set forth by the Declaration of Helsinki and its successive updates. The requirement for informed consent was waived because the patient data were anonymized.

Statistical analysis

This study aimed to investigate the association between CRP and DFI severity. There is currently no established diagnostic threshold for C-reactive protein in clinical practice. Patients in this study were divided into three groups based on CRP quantiles to ensure a relatively balanced population between groups and to explore some new findings. Data are presented as mean ± standard deviation (SD) or median (interquartile range, IQR) for continuous variables, and as frequencies (percentages) for categorical variables. Logistic regression analyses was used to estimate the odds ratio (OR) and corresponding 95% CI for DFI severity according to serum CRP levels.

Both non-adjusted and multivariateadjusted models were used. Variables were adjusted if they changed the odds ratio by at least 10% after being added to the model, and clinically significant variables were also adjusted. Model 1 was adjusted covariates including age and sex, and model 2 adjusted for variables from model 1 plus smoking, diabetes duration, hypertension, coronary heart disease, cerebral infarction, BMI, Mean arterial pressure. A multivariate-adjusted restricted cubic spline model was constructed to establish the OR curves at 3 knots to examine the possible nonlinear dose response association between CRP and DFI severity.We used a two-piece-wise logistic regression model with smoothing to analyze the association threshold between CRP and DFI severity after adjusting the variables in Model 3. The likelihood-ratio test and the bootstrap resampling method were used to determine inflection points.

In addition, it was determined whether the relationship between CRP and DFI was stable in different population groups. Subgroup analyses were performed according to age, sex, BMI, smoking status, hypertension, coronary heart disease and cerebrovascular disease.

The R statistical software (version 4.2.2, http://www.R-project.org, The R Foundation) and the Free Statistics Analysis Platform Version 1.9 [15] were employed to carry out all the analyses, and p value less than 0.05 represented statistically significance.

Baseline characteristics

Baseline characteristics of study participants by CRP category are summarised in Table 1. Of the 406 patients with DFI, 301 eligible individuals were included. The participants comprised 208 male and 93 female, with an average age of 65.2 ± 12.0 years, and 21.9% of the patients had severe DFIs. Differences were noted among the three groups in terms of the severity of DFI, FBG, RBC, hemoglobin, WBC, neutrophil percentage, platelet count, and albumin (all P < 0.05). Otherwise, the distribution of patient characteristics was similar (P > 0.05).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline Characteristics.

https://doi.org/10.1371/journal.pone.0323704.t001

Association between CRP and DFI severity

In the multivariable logistic regression analysis, the odds ratio (OR) of CRP (per 1mg/dl increase) were consistently significant in all three models (OR:1.01) (Table 2). After adjustment for all covariates, patients with CRP T3 showed a 43.12-fold increase in the DFI severity (OR = 43.12, 95%CI: 9.87 ~ 188.25, p < 0.001, model 2), compared to CRP T1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Multivariable logistic regression analyses of CRP and Severity of diabetic foot infection.

https://doi.org/10.1371/journal.pone.0323704.t002

Restricted cubic spline analysis was used to investigate the dose-response relationship between CRP and DFI severity. Fig 2 shows a non-linear relationship between CRP and DFI severity (Fig 2). In the two-piecewise regression models, the adjusted OR for severe diabetic foot infection was 1.039 (95% CI, 1.019–1.06; P ＜ 0.001) in participants with CRP ＜ 105mg/L, whereas there was no association between CRP and DFI severity in participants with CRP ≥ 105mg/L (S1 Table).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Nonlinear dose-response relationship between CRP and severity of diabetic foot infection.

https://doi.org/10.1371/journal.pone.0323704.g002

Subgroup analyses

Fig 3 shows the results of the subgroup analyses. In the subgroup analyses, for patients who were at advanced age,BMI and smoking or with coexisting conditions (hypertension, coronary heart disease, and cerebrovascular disease), the associations between CRP and DFI severity were statistically significant, and no significant interaction was detected.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Stratified analyses of the association between CRP and severity of diabetic foot infection according to baseline characteristics.

https://doi.org/10.1371/journal.pone.0323704.g003

Comparison with Previous Studies

A multitude of epidemiological research has substantiated the role of CRP associated with the intensity of diabetic foot infections. A study that included a retrospective cohort of 245 patients with DFI found an association between necrosis, serum albumin, erythrocyte sedimentation rate, CRP, and neutrophil-to-lymphocyte ratio values, and severe DFI [10]. Furthermore, In a clinical investigation conducted in South Korean hospitals, which included 123 patients diagnosed with DFI, it was found that higher concentrations of procalcitonin and CRP correlated with increased severity of DFI [16]. Several studies have confirmed the strong relationship between DFI and serum CRP levels [2,10,17,18], but no reports of a non-linear relationship between the two have been found. In our study population, we confirmed a positive correspondence between CRP levels and the severity of DFI. And there is an inverted L-shaped relationship between the two. In other words, serum CRP levels was associated with increased DFI severity when CRP was low than 105 mg/L, and there was no association between the two when CRP was above 105 mg/L.

Mechanistic insights

It’s pertinent to note that CRP is an acute-phase reactant synthesized within hepatocytes [19]. When infection occurs in patients with diabetes-related foot ulcers, plasma levels of CRP rise rapidly and dramatically, up to 1,000 times higher than normal [2,20]. Specific mechanisms may be that CRP is promptly produced by hepatocytes under the regulation of cytokines such as interleukin-6 and interleukin-1, in which upregulation occurs through the stimulation of gene expression by the STAT3, C/EBP, and NF-κB pathways [20].CRP binds to phospholipoproteins of localised bacterial cells in diabetes-related foot ulcers or exerts an immune effect by activating the complement converting enzyme C3. In addition, CRP binds to phagocytes and platelet-activating factors to reduce the inflammatory response [2].The results of this study recommend that CRP is not only a sign of inflammation but is also associated with inflammatory and systemic immune responses [20].

Clinical implications

Our study found a nonlinear relationship between CRP levels and DFI severity, which was not found in previous studies. In the early stages of severe infections, the immune system can be over-activated, producing large amounts of inflammation-associated factors such as CRP, calcitoninogen and IL-6. However, as the infection progresses, the immune cells may enter a state of exhaustion and the ability to produce inflammatory factors decreases or even reaches saturation, preventing further immune response. Hence, high CRP levels may indicate a more severe infection or complications. These findings highlight the importance of considering low and high CRP levels when assessing DFI severity and tailoring treatment strategies accordingly. Additional research is required to elucidate the fundamental mechanisms behind this non-linear relationship.

The study showed that CRP was not useful to distinguish between noninfected and mild IDFU, and is the most informative marker for differentiating more serious foot infections [17,18]. Our study population of patients with diabetes-related foot ulcers grade 3 and above with moderate to severe infections further confirms that CRP is associated with severe infections, which is consistent with the above findings.

This study had some limitations that need to be noted. First, although we excluded some inflammatory conditions that may affect CRP levels, we cannot exclude the possibility that CRP may be affected by other inflammatory conditions. Second, as with all observational studies, selection bias may have occurred. Thirt,t he study population is from a single institution, which may limit generalisability. Finally,t he wide 95% confidence intervals for the correlation between CRP and DFI severity may be related to sample size, population heterogeneity, and the non-specific nature of CRP as a marker of inflammation, necessitating larger standardised studies with multi-biomarker validation to improve clinical interpretability.