ABSTRACT
To study the risk factors associated with multidrug‐resistant bacteria (MDRB) infection in diabetic foot (DF) wounds and to develop a nomogram model to predict the risk of MDRB infection in patients with DF. A total of 157 patients with DF between January 2013 and December 2023 were included in this study. A retrospective analysis was conducted to examine the characteristics of MDRB infections. Univariate and multivariate logistic regression analyses were performed to identify risk factors associated with MDRB infection. Based on these risk factors, a predictive model was built using R software and a nomogram was constructed. Multivariate logistic regression analysis revealed that the wound area, previous hospitalization, prior use of antibacterial agents, lower extremity ischaemia grade, and hypoproteinaemia were independent risk factors for MDRB infection in DF wounds (p < 0.05). Construction of the nomogram model for MDRB infection in DF wounds: A nomogram model was developed using five identified risk factors—wound area, previous hospitalization, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia—as predictors. The area under the receiver operating characteristic (ROC) curve (AUC) was 0.863. The model has a good ability to predict MDRB infections in DF wounds. Wound area, previous hospitalization, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia were identified as independent risk factors for MDRB infections in DF wounds. The nomogram model constructed on the basis of these five factors demonstrated good predictive performance for MDRB infection.
Keywords: diabetic foot, multidrug‐resistant bacteria, nomogram model, risk factors
1. Introduction
Diabetes mellitus is a chronic metabolic disorder primarily characterized by impaired glucose metabolism. In recent years, the global prevalence of diabetes has been steadily increasing, driven by changes in lifestyle, dietary patterns and population ageing. A study revealed that approximately 463 million people were living with diabetes worldwide in 2019, and this number is projected to rise to 578 million by 2030 and 700 million by 2045 [1]. Poor long‐term glycaemic control and prolonged disease duration can lead to a range of acute and chronic complications in patients with diabetes. DF is one of the most common complications and is defined as a chronic condition resulting from hyperglycaemia‐induced neuropathy, vasculopathy and other contributing factors, which lead to poor wound healing or infection of the foot. According to an epidemiological study, the global prevalence of DF is approximately 6.3% [2]. DF significantly reduces patients' quality of life and may result in severe consequences such as foot ulcers, gangrene, amputation and even cardiovascular and cerebrovascular events [3]. It is a major cause of disability and mortality among patients with diabetes.
DF is frequently associated with bacterial infections. According to previous reports, 58% of patients with DF already present with infections at the time of medical consultation [4]. In recent years, increasing attention has been given to the spectrum and antimicrobial resistance of pathogens in DF wound infections, particularly infections caused by MDRB. MDRB are defined as pathogens resistant to three or more classes of clinically relevant antimicrobial agents [5]. The incidence of MDRB infections in DF wounds is increasing, with detection rates reportedly reaching as high as 90% [6, 7]. Owing to the complexity of MDRB, infection not only increases the difficulty of treating foot infections, but it also may prolong hospitalization, increase medical costs, increase the risk of amputation and even lead to death [8, 9]. Thus, MDRB play a critical role in delayed wound healing in DF. Early identification of MDRB infection in DF wounds is essential for timely intervention. However, the following questions remain unanswered: What are the risk factors? How can clinicians identify MDRB infection in DF wounds before the results of wound culture and sensitivity tests are available? To date, few studies have addressed these issues. The development of risk prediction models can help quantify the probability of infection, facilitate early diagnosis and support clinical decision‐making. Furthermore, internal validation of such models helps determine their clinical applicability and reliability.
Therefore, the aim of this study was to retrospectively analyse the clinical data of patients with DF in our hospital, identify important risk factors by univariate and multivariate logistic regression analyses, and construct a nomogram model for predicting the risk of MDRB infection based on these factors. Ultimately, this study seeks to provide a reliable clinical basis for the early identification of MDRB infections in DF wounds.
2. Materials and Methods
2.1. Patient Characteristics
A total of 157 patients with DF admitted to the Department of Endocrinology at the Bethune International Peace Hospital from January 2013 to December 2023 were included. The inclusion criteria were as follows: (1) patients who met the diagnostic criteria for type 2 diabetes mellitus issued by the World Health Organization (WHO) in 1999 and the diagnostic criteria for diabetic foot ulcers established by the International Working Group on the Diabetic Foot/Infectious Diseases Society of America (IWGDF/IDSA 2023); and (2) patients whose complete clinical records and examination data were available. The exclusion criteria were as follows: (1) foot ulcers due to diseases such as varicose veins; and (2) the presence of malignancy, rheumatic diseases, severe hepatic or renal dysfunction, cardiac dysfunction, cognitive impairment, or other severe infectious diseases. This retrospective study was approved by the Ethics Committee of the Bethune International Peace Hospital, which waived informed consent. All procedures were conducted in compliance with relevant guidelines, regulations and the Declaration of Helsinki.
2.2. Data Collection
Clinical data including sex, age, diabetes duration, wound duration and wound area, Wagner grade, previous use of antibacterial drugs, previous hospitalization, ankle‐brachial index (ABI), C‐reactive protein (CRP), complete blood count, liver and kidney functions, electrolytes, glycated haemoglobin (HbA1c) levels, urinary albumin‐to‐creatinine ratio and bacterial cultures with antimicrobial susceptibility from wound secretions or tissue samples were extracted from the hospital information system. Data were collected and independently verified by two researchers. Patients were divided into MDRB and non‐MDRB groups based on the presence of MDRB infection.
2.3. Specimen Collection
Specimen collection was performed by trained personnel in strict accordance with standard clinical procedures for diabetic foot infections. Prior to sampling, the wound surface was washed with sterile normal saline to reduce surface contamination. Swabs or syringe needles were inserted deeply into the wound and placed in close contact with the ulcer base, and tissue samples were excised with a scalpel from the interface between viable tissue and the deep wound bed for microbiological analysis.
2.4. Bacterial Culture, Identification and Antimicrobial Susceptibility Testing
All samples were promptly transported to the microbiology laboratory and inoculated onto Columbia blood agar and MacConkey agar plates (Autobio Biochemistry Co. Ltd., China). The plates were incubated at 35°C for 24–48 h and examined for bacterial growth. Bacterial identification and antimicrobial susceptibility testing were performed using the MA120 microbial identification and antimicrobial susceptibility analyser (Zhuhai Meihua Medical Technology Co. Ltd., China) following the manufacturer's instructions. Antimicrobial susceptibility was determined using the minimum inhibitory concentration (MIC) method. Antimicrobial susceptibility testing requires selecting antibiotics according to the bacterial genus. The choice of antibiotic types and interpretation of susceptibility results were based on the relevant standards established by the Clinical and Laboratory Standards Institute (CLSI) [10]. Susceptibility results were reported as susceptible (S), intermediate (I), or resistant (R). MDRB were defined as isolates exhibiting resistance to ≥ 3 different classes of antimicrobial agents.
Quality control was ensured using standard reference strains, including Pseudomonas aeruginosa ( P. aeruginosa ) ATCC 27853, Staphylococcus aureus ( S. aureus ) ATCC 29213, Escherichia coli ( E. coli ) ATCC 25922, and Enterococcus faecalis ( E. faecalis ) ATCC 29212, which were used to verify antimicrobial agents, culture media and the overall performance of laboratory procedures.
2.5. Diagnostic Criteria and Definitions
(1) Patients meeting the diagnostic criteria for type 2 diabetes mellitus issued by the World Health Organization (WHO) in 1999 [11]. (2) DF diagnosis following the 2023 IWGDF guidelines [12]. (3) Wagner wound classification was performed by clinicians according to the criteria proposed by Wagner FW in 1981 [13]. (4) Chronic kidney disease (CKD) was diagnosed on the basis of the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease [14]. (5) The diagnosis of MDRB was based on the interim international expert consensus criteria published in 2012 [5]. (6) Anaemia was diagnosed according to the WHO haemoglobin threshold guidelines [15]. (7) The ABI was used to assess the lower extremity ischaemia grade [16], with severity categorized as follows: Grade 0: ABI > 0.90; Grade 1: 0.70 < ABI ≤ 0.90; Grade 2: 0.40 < ABI ≤ 0.70; Grade 3: ABI ≤ 0.40. (8) Hypoproteinaemia was defined as a serum albumin concentration < 30 g/L. (9) Previous antimicrobial use was defined as a history of intravenous antimicrobial use in the present foot wound. (10) Previous hospitalization was defined as a history of hospitalization for the present foot wound.
2.6. Statistical Analyses
SPSS 26.0 software and R software (version 4.4.1) were used for the statistical analysis. Categorical variables are expressed as frequencies (n, %) and were compared using chi‐square tests. Continuous variables with a normal distribution are presented as the means ± SDs and were compared using t‐tests; nonnormally distributed variables are reported as medians (P25, P75) and were compared using Mann–Whitney U tests. Variables significant in the univariate analyses were included in the multivariate binary logistic regression to identify independent predictors of MDRB infection, with statistical significance set at p < 0.05. A nomogram model was constructed on the basis of the regression results using the “rms” package in R. The ROC curve and area under the curve (AUC) were calculated using the “pROC” package. Calibration curves were generated with the “rms” package, and clinical decision curves were plotted using the “rmda” package. Internal validation was performed using the bootstrap method.
3. Results
Baseline characteristics of the patients.
Among the 157 patients with DF included in this study, 103 were males and 54 were females. There were no significant differences between the MDRB and non‐MDRB groups in terms of age, diabetes duration, white blood cell (WBC) count, CRP level, HbA1c level, sex, or CKD status (p > 0.05). However, significant differences were observed between the two groups regarding the wound area, wound duration, previous hospitalization, Wagner grade, previous use of antibacterial drugs, lower extremity ischaemia grade, anaemia and hypoproteinaemia (p < 0.05) (Table 1).
TABLE 1.
Comparison of the general data between the two groups is presented.
| Variable | Non‐MDRB group (n = 40) | MDRB group (n = 117) | z/χ2/t | p |
|---|---|---|---|---|
| Age (years) | 64.00 (57.25, 72.75) | 65.00 (59.50, 71.50) | z = −0.381 | 0.703 |
| Diabetes duration (years) | 14.00 (5.25, 20.00) | 13.00 (7.00, 20.00) | z = −0.414 | 0.679 |
| Wound area (cm2) | 4.00 (1.00, 8.75) | 11.00 (4.00, 21.00) | z = −4.099 | < 0.001* |
| Wound duration (days) | 7.50 (5.00, 18.75) | 30.00 (14.50, 40.00) | z = −4.338 | < 0.001* |
| WBC count (109/L) | 8.04 (6.82, 12.41) | 8.25 (6.52, 10.88) | z = −0.242 | 0.809 |
| CRP level (mg/L) | 82.07 ± 52.58 | 72.42 ± 52.66 | t = 1.002 | 0.318 |
| HbA1c level (%) | 8.60 (7.53, 9.50) | 8.20 (6.90, 9.70) | z = −0.532 | 0.595 |
| Sex | χ2 = 2.100 | 0.147 | ||
| Female | 10 (25.00) | 44 (37.61) | ||
| Male | 30 (75.00) | 73 (62.39) | ||
| Wagner grade | χ2 = 17.173 | < 0.001* | ||
| Grade 2 | 24 (60.00) | 29 (24.79) | ||
| Grade 3 | 13 (32.50) | 61 (52.14) | ||
| Grade 4 | 3 (7.50) | 27 (23.08) | ||
| Previous hospitalization | χ2 = 15.050 | < 0.001* | ||
| No | 25 (62.50) | 33 (28.21) | ||
| Yes | 15 (37.50) | 84 (71.79) | ||
| Previous use of antibacterial drugs | χ2 = 11.439 | < 0.001* | ||
| No | 33 (82.50) | 61 (52.14) | ||
| Yes | 7 (17.50) | 56 (47.86) | ||
| Lower extremity ischaemia grade | χ2 = 9.613 | 0.022* | ||
| Grade 0 | 19 (47.50) | 30 (25.64) | ||
| Grade 1 | 6 (15.00) | 16 (13.68) | ||
| Grade 2 | 13 (32.50) | 46 (39.32) | ||
| Grade 3 | 2 (5.00) | 25 (21.37) | ||
| Anaemia | χ2 = 5.872 | 0.015* | ||
| No | 28 (70.00) | 56 (47.86) | ||
| Yes | 12 (30.00) | 61 (52.14) | ||
| Hypoproteinaemia | χ2 = 9.051 | 0.003* | ||
| No | 32 (80.00) | 62 (52.99) | ||
| Yes | 8 (20.00) | 55 (47.01) | ||
| CKD | χ2 = 2.986 | 0.084 | ||
| No | 22 (55.00) | 46 (39.32) | ||
| Yes | 18 (45.00) | 71 (60.68) |
Abbreviations: CKD, chronic kidney disease; CRP, C‐reactive protein; HbA1c, glycated haemoglobin; MDRB, multidrug‐resistant bacteria; WBC, white blood cell.
p < 0.05.
4. Characteristics of MDRB Infections in DF Wounds
Among the 157 patients with DF who were enrolled, 117 had MDRB infections, representing an infection rate of 74.52%. A total of 211 pathogenic bacterial strains were isolated, among which 144 (68.25%) were MDRB and 67 (31.75%) were non‐MDRB. The predominant Gram‐positive MDRB strains included S. aureus , Staphylococcus epidermidis and E. faecalis , accounting for 32.63%, 5.56% and 3.47%, respectively. The major Gram‐negative MDRB strains were E. coli , P. aeruginosa and Proteus mirabilis , accounting for 8.33%, 5.56% and 5.56%, respectively (Table 2).
TABLE 2.
Characteristics of MDRB infection in diabetic foot wounds.
| Pathogen category | Strains | Proportion of isolates (%) | MDRB strains | MDRB detection rate (%) |
|---|---|---|---|---|
| Gram‐positive bacteria | 112 | 53.08 | 77 | 68.75 |
| S. aureus | 67 | 31.75 | 47 | 70.15 |
| Staphylococcus epidermidis | 9 | 4.27 | 8 | 88.89 |
| E. faecalis | 6 | 2.84 | 5 | 83.33 |
| Staphylococcus simulans | 5 | 2.37 | 4 | 80.00 |
| Others | 25 | 11.84 | 13 | 52.00 |
| Gram‐negative bacteria | 99 | 46.92 | 67 | 67.68 |
| E. coli | 17 | 8.06 | 12 | 70.59 |
| P. aeruginosa | 14 | 6.64 | 8 | 57.14 |
| Proteus vulgaris | 12 | 5.69 | 8 | 66.67 |
| Klebsiella pneumoniae | 9 | 4.27 | 4 | 44.44 |
| Others | 47 | 22.27 | 35 | 74.47 |
| Total strains | 211 | 100.00 | 144 | 68.25 |
Note: Data are presented as n (%). Proportion of isolates (%) = (Number of bacterial strains in a given pathogen category/Total number of bacterial strains isolated) × 100. MDRB detection rate (%) = (Number of multidrug‐resistant bacterial strains within a given pathogen category ÷ Total number of bacterial strains in the same category) × 100.
Abbreviations: E. coli, Escherichia coli; MDRB, multidrug‐resistant bacteria; P. aeruginosa, Pseudomonas aeruginosa; S. aureus, Staphylococcus aureus .
5. Univariate Logistic Regression Analysis
Univariate logistic regression analysis was conducted with MDRB infection status as the dependent variable, and sex, age, wound area, wound duration, diabetes duration, WBC count, CRP levels, HbA1c levels, previous hospitalization, previous use of antibacterial drugs, anaemia, hypoproteinaemia, CKD, Wagner grade and lower extremity ischaemia grade as independent variables. The analysis revealed that the wound area, wound duration, previous hospitalization, previous use of antibacterial drugs, anaemia, hypoproteinaemia, Wagner grade and lower extremity ischaemia grade were significantly associated with MDRB infection in DF wounds. (p < 0.05) (Table 3).
TABLE 3.
Univariate logistic regression analysis of MDRB infection in diabetic foot wounds.
| Variable | β | S.E | Wald | p | OR (95% CI) |
|---|---|---|---|---|---|
| Sex | −0.592 | 0.412 | 2.067 | 0.151 | 0.553 (0.247–1.240) |
| Age | 0.006 | 0.016 | 0.426 | 0.690 | 1.006 (0.976–1.038) |
| Wound area | 0.075 | 0.026 | 8.502 | 0.004* | 1.078 (1.025–1.133) |
| Wound duration | 0.023 | 0.009 | 6.268 | 0.012* | 1.024 (1.005–1.043) |
| Diabetes duration | −0.013 | 0.024 | 0.304 | 0.581 | 0.987 (0.943–1.034) |
| WBC count | −0.008 | 0.049 | 0.024 | 0.878 | 0.992 (0.901–1.093) |
| CRP level | −0.003 | 0.003 | 1.003 | 0.317 | 0.997 (0.990–1.003) |
| HbA1c level | −0.011 | 0.094 | 0.014 | 0.907 | 0.989 (0.822–1.190) |
| Previous hospitalization | 1.445 | 0.386 | 14.028 | < 0.001* | 4.242 (1.992–9.037) |
| Previous use of antibacterial drugs | 1.465 | 0.455 | 10.349 | 0.001* | 4.328 (1.773–10.566) |
| Anaemia | 0.933 | 0.392 | 5.676 | 0.017* | 2.542 (1.180–5.475) |
| Hypoproteinaemia | 1.266 | 0.437 | 8.417 | 0.004* | 3.548 (1.508–8.348) |
| CKD | 0.635 | 0.370 | 2.944 | 0.086 | 1.886 (0.914–3.895) |
| Wagner grade | |||||
| Grade 2 | 15.604 | < 0.001* | |||
| Grade 3 | 1.357 | 0.412 | 10.861 | 0.001* | 3.883 (1.733–8.702) |
| Grade 4 | 2.008 | 0.668 | 9.030 | 0.003* | 7.448 (2.010–27.594) |
| Lower extremity ischaemia grade | |||||
| Grade 0 | 8.421 | 0.038* | |||
| Grade 1 | 0.524 | 0.561 | 0.872 | 0.351 | 1.689 (0.562–5.075) |
| Grade 2 | 0.807 | 0.430 | 3.527 | 0.060 | 2.241 (0.965–5.202) |
| Grade 3 | 2.069 | 0.791 | 6.838 | 0.009* | 7.917 |
Abbreviations: CKD, chronic kidney disease; CRP, C‐reactive protein; HbA1c, glycated haemoglobin; MDRB, multidrug‐resistant bacteria; WBC, white blood cell.
p < 0.05.
6. Multivariate Logistic Regression Analysis
A multivariate logistic regression analysis was conducted to identify independent risk factors for MDRB infection in DF wounds. The presence of MDRB infection was set as the dependent variable, whereas the wound area, wound duration, previous hospitalization, previous use of antibacterial drugs, anaemia, hypoproteinaemia, Wagner grade and lower extremity ischaemia grade were included as independent variables. The results demonstrated that the wound area, previous hospitalization, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia were independent risk factors for MDRB infection in patients with DF wounds (p < 0.05) (Table 4).
TABLE 4.
Multifactorial Logistic Regression Analysis of MDRB Infection in Diabetic Foot Wounds.
| Variable | β | S.E | Wald | p | OR (95% CI) |
|---|---|---|---|---|---|
| Wound area | 0.066 | 0.029 | 5.025 | 0.025* | 1.068 (1.008–1.131) |
| Wound duration | 0.000 | 0.010 | 0.002 | 0.965 | 1.000 (0.981–1.020) |
| Previous hospitalization | 1.019 | 0.494 | 4.249 | 0.039* | 2.770 (1.051–7.298) |
| Previous use of antibacterial drugs | 1.197 | 0.604 | 3.926 | 0.048* | 3.312 (1.013–10.826) |
| Anaemia | 0.288 | 0.480 | 0.360 | 0.549 | 1.333 (0.521–3.431) |
| Hypoproteinaemia | 1.233 | 0.535 | 5.318 | 0.021* | 3.432 (1.203–9.789) |
| Lower extremity ischaemia grade | |||||
| Grade 0 | 6.872 | 0.076 | |||
| Grade 1 | 0.331 | 0.707 | 0.220 | 0.639 | 1.393 (0.349–5.565) |
| Grade 2 | 0.881 | 0.526 | 2.801 | 0.940 | 2.414 (0.860–6.773) |
| Grade 3 | 2.161 | 0.903 | 5.724 | 0.017* | 8.684 (1.478–51.014) |
| Wagner grade | |||||
| Grade 2 | 3.716 | 0.156 | |||
| Grade 3 | 0.590 | 0.520 | 1.288 | 0.256 | 1.803 (0.651–4.993) |
| Grade 4 | 1.389 | 0.769 | 3.257 | 0.710 | 4.009 (0.887–18.109) |
Abbreviation: MDRB, multidrug‐resistant bacteria.
7. Establishment of the Nomogram Model
A nomogram model was developed using R software version 4.4.1. The model incorporated five predictors identified as independent risk factors in the multivariate logistic regression analysis, including wound area, previous hospitalization, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia. These variables were used to construct the nomogram, which calculates the predicted probability of MDRB infection on basis of the total score derived from the five risk factors (Figure 1).
FIGURE 1.
Nomogram model of DF wound infection with MDRB.
8. Evaluation of the Nomogram Model
8.1. Discrimination Ability of the Nomogram Model
An ROC curve analysis was performed to evaluate the predictive performance of the five risk factors—wound area, previous hospitalisation, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia—as well as the overall nomogram model. The results revealed that the nomogram achieved an AUC of 0.863 (95% CI: 0.802–0.923), with a sensitivity of 70.9% and a specificity of 90.0% (p < 0.001), indicating good discriminative ability of the model (Table 5, Figure 2).
TABLE 5.
Analysis of model effectiveness.
| Variable | Cut‐off value | AUC | Sensitivity (%) | Specificity (%) | 95% CI | p |
|---|---|---|---|---|---|---|
| Wound area | 9.500 | 0.717 | 0.547 | 0.850 | 0.630–0.804 | < 0.001 |
| Previous hospitalization | 0.500 | 0.671 | 0.718 | 0.625 | 0.572–0.771 | 0.001 |
| Previous use of antibacterial drugs | 0.500 | 0.652 | 0.479 | 0.825 | 0.559–0.745 | 0.004 |
| Lower extremity ischaemia grade | 1.500 | 0.654 | 0.607 | 0.625 | 0.559–0.749 | 0.004 |
| Hypoproteinaemia | 0.500 | 0.635 | 0.470 | 0.800 | 0.540–0.730 | 0.011 |
| Nomogram model | 0.801 | 0.863 | 0.709 | 0.900 | 0.802–0.923 | < 0.001 |
Abbreviation: AUC, area under the curve.
FIGURE 2.
ROC curves. ROC, receiver operating characteristic.
8.2. Calibration of the Nomogram Model
The Hosmer–Lemeshow goodness‐of‐fit test yielded a χ2 value of 13.197 with a p value of 0.105 (> 0.05), emphasizing an appreciable consistency between the predicted and actual probabilities. Internal validation was performed using the bootstrap method with 1000 resamples, and the C‐index was 0.861 (95% CI: 0.795–0.916). Calibration curves based on both the original dataset and internal bootstrap validation were used to assess the agreement between the predicted and observed probabilities. This observation suggests that the nomogram model has high predictive accuracy. (Figures 3 and 4).
FIGURE 3.
Calibration curves for the nomogram model in the original.
FIGURE 4.
Calibration curves for internal validation of the nomogram model.
8.3. Clinical Application of the Nomogram Model
The clinical applicability of the nomogram model was evaluated using DCA. The model demonstrated high clinical applicability. Within a certain threshold range, the decision curve showed greater net benefit than did the strategies of treating all patients or treating none. (Figure 5).
FIGURE 5.
Clinical decision‐making curve.
9. Discussion
The results of this study demonstrated that Gram‐positive bacteria were the predominant pathogens associated with MDRB infections in DF wounds, with MDR S. aureus , MDR E. coli and MDR P. aeruginosa being the most frequently isolated organisms. Five independent risk factors for MDRB infection were identified: wound area, previous hospitalization, prior antimicrobial use, lower limb ischaemia grade and hypoproteinaemia. A nomogram model constructed on the basis of these variables exhibited good predictive performance.
In this study, the prevalence of MDRB infection in diabetic foot wounds was 68.25%, which was higher than the 50.86% reported in a previous meta‐analysis [17]. Among the MDRB isolates, Gram‐positive bacteria accounted for 53.47%, while Gram‐negative bacteria accounted for 46.53%, indicating a predominance of Gram‐positive organisms. Similar bacterial spectrum patterns have been reported by Shen [18] and Sannathimmappa [19]. Over the past decade, an increasing number of studies have reported a progressive rise in the prevalence of Gram‐negative MDRB, which has become more prevalent than Gram‐positive organisms in some regions among patients with diabetic foot infections [20, 21]. Another meta‐analysis confirmed that the prevalence of multidrug‐resistant Gram‐positive bacteria in diabetic foot ulcers was 19.81%, whereas that of multidrug‐resistant Gram‐negative bacteria was 32.84%. The pathogen spectrum was characterized primarily by MDR S. aureus (12.13%), MDR E. coli (6.93%) and MDR P. aeruginosa (6.01%) [17]. The results of this study align with the meta‐analysis. In the present study, the bacterial spectrum was dominated by MRSA, MDR E. coli and MDR P. aeruginosa . Due to differences in study periods, study centres, and other factors, several studies have reported distinct pathogen spectrum profiles [20, 22, 23]. These findings underscore the importance of regional surveillance and standardized testing to guide empirical antimicrobial therapy and infection control strategies.
Previous use of antibacterial drugs was significantly associated with MDRB infection in DF wounds, which is consistent with the findings of Zhang [24]. This association may be attributed to the repeated and prolonged use of antibiotics, which promotes the emergence of resistance‐related gene mutations and the development of multidrug‐resistant gene clusters in pathogens. Therefore, rational antibiotic prescribing and patient education on standardized diagnosis and treatment are essential to mitigate the risk of MDRB infections due to inappropriate antimicrobial exposure.
A history of previous hospitalization was also identified as an important risk factor for MDRB infection. Liu [25] reported a significant association between prior hospitalization and MDRB infection, which is consistent with the findings of this study. Another study further demonstrated that MDRB infection was closely related to nosocomial infection, with longer and more frequent hospitalisations increasing the risk of infection [26]. This may be attributed to the widespread presence of MDRB in hospital environments and the greater likelihood of cross‐infection due to invasive procedures such as repeated wound debridement and frequent medical interventions in patients with DF [27, 28]. Therefore, improving health care workers' aseptic awareness is essential to reduce the risk of transmission.
The wound area was also identified as an independent risk factor, which is consistent with the findings of Kang [29]. Larger wound areas increase the likelihood of MDRB colonization from the environment [21]. Several studies [30, 31] have further concluded that a wound area greater than 4 cm2 is a significant risk factor for MDRB infection. Collectively, these findings suggest that a larger wound size may substantially increase the risk of MDRB infection by providing greater opportunities for bacterial exposure and colonization.
Hypoproteinaemia has been identified as an independent risk factor for MDRB infection. Our findings are consistent with those reported by Feng [32]. A decrease in serum albumin is associated with impaired immune function, which compromises the host's ability to defend against pathogenic bacteria and thereby increases susceptibility to MDRB infection. Additionally, reduced albumin levels may lower the protein‐binding rate of antimicrobial agents, leading to decreased effective drug concentrations at the site of infection, delayed therapeutic response and an increased risk of antimicrobial resistance.
Peripheral arterial disease is very common in patients with DF [33], and lesions can cause varying degrees of lower limb ischaemia. Lower limb ischaemia has now been shown to be an independent risk factor for MDRB infection [34], which is consistent with the results of this study. The mechanism may include insufficient blood supply to the trauma in the case of severe vascular lesions of the lower limbs, leading to local tissue hypoxia and a reduced nutrient supply, making it difficult for the trauma to heal and increasing the susceptibility to MDRB. Second, severe lower limb ischaemia may lead to a decrease in the local antimicrobial drug concentration during trauma, making it difficult to achieve an effective bactericidal concentration; a lower blood drug concentration and longer antibiotic treatment time may further lead to infection with MDRB [34].
Our study used a nomogram model to visualize the probability of MDRB infection in patients with DF. Predictors identified by multivariate logistic regression—including prior antimicrobial use, previous hospitalization, hypoproteinaemia, wound area and lower extremity ischaemia grade—were used to construct the nomogram. Differences exist between this study and the nomogram model developed by Ma [35] in terms of predictor selection. The model by Ma included age, diabetes duration, previous use of antibacterial drugs, types of antibacterial agents and osteoporosis as predictors. These predictors may reflect differences in variable selection criteria, study populations and study designs.
This study has several limitations. As a single‐centre, retrospective study, the findings may be subject to selection and information bias and may have limited generalisability to other clinical settings. In addition, external validation was not performed; therefore, the applicability of the nomogram beyond the study population remains uncertain. Future prospective multicentre studies with external validation are needed to further confirm and refine the predictive performance and clinical utility of the proposed model.
10. Conclusion
A nomogram model was developed using the five identified risk factors—wound area, previous hospitalization, previous use of antibacterial drugs, lower extremity ischaemia grade and hypoproteinaemia—as predictors. The model has a good ability to predict MDRB infections in DF wounds.
The findings of this study have significant clinical implications. Early identification of MDRB infection risk allows clinicians to initiate targeted empirical therapy and implement more rigorous infection control measures. The nomogram model can aid in personalized clinical decision‐making, reduce the likelihood of inappropriate antibiotic use, and potentially improve wound healing outcomes while minimizing complications such as amputation.
Funding
This work was supported by the Medical Science Research Project of Hebei (Grant No. 20240696).
Ethics Statement
This retrospective study was approved by the Ethics Committee of the Bethune International Peace Hospital, which waived informed consent.
Conflicts of Interest
The authors declare no conflicts of interest.
Acknowledgements
This study was supported by the Medical Science Research Project of Hebei (Grant No. 20240696).
Data Availability Statement
Research data are not shared.
References
- 1. Saeedi P., Petersohn I., Salpea P., et al., “Global and Regional Diabetes Prevalence Estimates for 2019 and Projections for 2030 and 2045: Results From the International Diabetes Federation Diabetes Atlas, 9(Th) Edition,” Diabetes Research and Clinical Practice 157 (2019): 107843. [DOI] [PubMed] [Google Scholar]
- 2. Zhang P., Lu J., Jing Y., Tang S., Zhu D., and Bi Y., “Global Epidemiology of Diabetic Foot Ulceration: A Systematic Review and Meta‐Analysis,” Annals of Medicine 49, no. 2 (2017): 106–116. [DOI] [PubMed] [Google Scholar]
- 3. Edmonds M., Manu C., and Vas P., “The Current Burden of Diabetic Foot Disease,” Journal of Clinical Orthopaedics and Trauma 17 (2021): 88–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Prompers L., Huijberts M., Apelqvist J., et al., “High Prevalence of Ischaemia, Infection and Serious Comorbidity in Patients With Diabetic Foot Disease in Europe. Baseline Results From the Eurodiale Study,” Diabetologia 50, no. 1 (2007): 18–25. [DOI] [PubMed] [Google Scholar]
- 5. Magiorakos A. P., Srinivasan A., Carey R. B., et al., “Multidrug‐Resistant, Extensively Drug‐Resistant and Pandrug‐Resistant Bacteria: An International Expert Proposal for Interim Standard Definitions for Acquired Resistance,” Clinical Microbiology and Infection: The Official Publication of the European Society of Clinical Microbiology and Infectious Diseases 18, no. 3 (2012): 268–281. [DOI] [PubMed] [Google Scholar]
- 6. Atlaw A., Kebede H. B., Abdela A. A., and Woldeamanuel Y., “Bacterial Isolates From Diabetic Foot Ulcers and Their Antimicrobial Resistance Profile From Selected Hospitals in Addis Ababa,” Ethiopia Frontiers in Endocrinology 13 (2022): 987487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mashaly M., Kheir M. A. E., Ibrahim M., et al., “Aerobic Bacteria Isolated From Diabetic Foot Ulcers of Egyptian Patients: Types, Antibiotic Susceptibility Pattern and Risk Factors Associated With Multidrug‐Resistant Organisms,” Germs 11, no. 4 (2021): 570–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Adeyemo A. T., Kolawole B., Rotimi V. O., and Aboderin A. O., “Multicentre Study of the Burden of Multidrug‐Resistant Bacteria in the Aetiology of Infected Diabetic Foot Ulcers,” African Journal of Laboratory Medicine 10, no. 1 (2021): 1261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Zubair M., Malik A., and Ahmad J., “Clinico‐Microbiological Study and Antimicrobial Drug Resistance Profile of Diabetic Foot Infections in North India,” Foot (Edinburgh, Scotland) 21, no. 1 (2011): 6–14. [DOI] [PubMed] [Google Scholar]
- 10. Clinical and Laboratory Standards Institute , “Performance Standards for Antimicrobial Susceptibility Testing; Nineteenth Informational Supplement. In: CLSI Document M100‐S23. Wayne, Pa: Clinical and Laboratory Standards Institute,” (2013).
- 11. Grimaldi A. and Heurtier A., “Diagnostic Criteria for Type 2 Diabetes,” La Revue du Praticien 49, no. 1 (1999): 16–21. [PubMed] [Google Scholar]
- 12. Senneville É., Albalawi Z., Van Asten S. A., et al., “IWGDF/IDSA Guidelines on the Diagnosis and Treatment of Diabetes‐Related Foot Infections (IWGDF/IDSA 2023),” Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America 40 (2023): e3687. [DOI] [PubMed] [Google Scholar]
- 13. F. W. Wagner, Jr. , “The Dysvascular Foot: A System for Diagnosis and Treatment,” Foot & Ankle 2, no. 2 (1981): 64–122. [DOI] [PubMed] [Google Scholar]
- 14. KDIGO , “2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease,” Kidney International 105, no. 4s (2024): S117–s314. [DOI] [PubMed] [Google Scholar]
- 15. World Health Organization , “WHO Guidelines Approved by the Guidelines Review Committee [M]. Guideline on Haemoglobin Cutoffs to Define Anaemia in Individuals and Populations. Geneva,” (2024). [PubMed]
- 16. Crawford F., Welch K., Andras A., and Chappell F. M., “Ankle Brachial Index for the Diagnosis of Lower Limb Peripheral Arterial Disease,” Cochrane Database of Systematic Reviews 9, no. 9 (2016): Cd010680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Yang S., Hu L., Zhao Y., Meng G., Xu S., and Han R., “Prevalence of Multidrug‐Resistant Bacterial Infections in Diabetic Foot Ulcers: A Meta‐Analysis,” International Wound Journal 21, no. 4 (2024): e14864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Shen Q., Lin D., Zhu H., et al., “Clinical Distribution and Antimicrobial Resistance Analysis of 754 Pathogenic Bacteria in Diabetic Foot Infection,” Zhonghua Yi Xue Za Zhi 94, no. 12 (2014): 889–894. [PubMed] [Google Scholar]
- 19. Sannathimmappa M. B., Nambiar V., Aravindakshan R., et al., “Diabetic Foot Infections: Profile and Antibiotic Susceptibility Patterns of Bacterial Isolates in a Tertiary Care Hospital of Oman,” Journal of Education Health Promotion 10 (2021): 254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Chen T., Yu J., Ye J., et al., “Infection Characteristics and Drug Susceptibility of Multidrug‐Resistant Bacteria in Patients With Diabetic Foot Ulcers,” Clinical Laboratory 69, no. 9 (2023): 230309. [DOI] [PubMed] [Google Scholar]
- 21. Yan X., Song J. F., Zhang L., and Li X., “Analysis of Risk Factors for Multidrug‐Resistant Organisms in Diabetic Foot Infection,” BMC Endocrine Disorders 22, no. 1 (2022): 46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Miyan Z., Fawwad A., Sabir R., and Basit A., “Microbiological Pattern of Diabetic Foot Infections at a Tertiary Care Center in a Developing Country,” Journal of the Pakistan Medical Association 67, no. 5 (2017): 665–669. [PubMed] [Google Scholar]
- 23. Pessoa E. C. T., Duarte B., João A. L., et al., “Multidrug‐Resistant Bacteria in Diabetic Foot Infections: Experience From a Portuguese Tertiary Centre,” International Wound Journal 17, no. 6 (2020): 1835–1839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Zhang J., Chu Y., Wang P., et al., “Clinical Outcomes of Multidrug Resistant pseudomonas aeruginosa Infection and the Relationship With Type III Secretion System in Patients With Diabetic Foot,” International Journal of Lower Extremity Wounds 13, no. 3 (2014): 205–210. [DOI] [PubMed] [Google Scholar]
- 25. Liu X., Ren Q., Zhai Y., Kong Y., Chen D., and Chang B., “Risk Factors for Multidrug‐Resistant Organisms Infection in Diabetic Foot Ulcer,” Infection and Drug Resistance 15 (2022): 1627–1635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Bakthavatchalam Y. D., Rao S. V., Isaac B., et al., “A Comparative Assessment of Clinical, Pharmacological and Antimicrobial Profile of Novel Anti‐Methicillin‐Resistant Staphylococcus aureus Agent Levonadifloxacin: Therapeutic Role in Nosocomial and Community Infections,” Indian Journal of Medical Microbiology 37, no. 4 (2019): 478–487. [DOI] [PubMed] [Google Scholar]
- 27. Guo H., Song Q., Mei S., Xue Z., Li J., and Ning T., “Distribution of Multidrug‐Resistant Bacterial Infections in Diabetic Foot Ulcers and Risk Factors for Drug Resistance: A Retrospective Analysis,” PeerJ 11 (2023): e16162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Chaoui L., Mhand R., Mellouki F., and Rhallabi N., “Contamination of the Surfaces of a Health Care Environment by Multidrug‐Resistant (MDR) Bacteria,” International Journal of Microbiology 2019 (2019): 3236526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Kang W. J., Shi L., Shi Y., et al., “Analysis on Distribution, Drug Resistance and Risk Factors of Multi Drug Resistant Bacteria in Diabetic Foot Infection,” Biomedical Research (India) 28 (2017): 10186–10196. [Google Scholar]
- 30. Zubair M. and Ahmad J., “Potential Risk Factors and Outcomes of Infection With Multidrug Resistance Among Diabetic Patients Having Ulcers: 7 Years Study,” Diabetes & Metabolic Syndrome 13, no. 1 (2019): 414–418. [DOI] [PubMed] [Google Scholar]
- 31. Gadepalli R., Dhawan B., Sreenivas V., Kapil A., Ammini A. C., and Chaudhry R., “A Clinico‐Microbiological Study of Diabetic Foot Ulcers in an Indian Tertiary Care Hospital,” Diabetes Care 29, no. 8 (2006): 1727–1732. [DOI] [PubMed] [Google Scholar]
- 32. Feng S. H., Chu Y. J., Wang P. H., Jun X., Min D., and Li X. M., “Risk Factors and Gene Type for Infections of MRSA in Diabetic Foot Patients in Tianjin, China,” International Journal of Lower Extremity Wounds 12, no. 2 (2013): 106–112. [DOI] [PubMed] [Google Scholar]
- 33. Rhee S. Y., Guan H., Liu Z. M., et al., “Multi‐Country Study on the Prevalence and Clinical Features of Peripheral Arterial Disease in Asian Type 2 Diabetes Patients at High Risk of Atherosclerosis,” Diabetes Research and Clinical Practice 76, no. 1 (2007): 82–92. [DOI] [PubMed] [Google Scholar]
- 34. Xia W., He W., Luo T., and Tang N., “Risk Factors for Multidrug‐Resistant Bacterial Infections in Patients With Diabetic Foot Ulcers: A Systematic Review and Meta‐Analysis,” Annals of Palliative Medicine 10, no. 12 (2021): 12618–12630. [DOI] [PubMed] [Google Scholar]
- 35. Ma Y. N., Zhang L. X., Hu Y. Y., and Shi T. L., “Nomogram Model for Predicting the Risk of Multidrug‐Resistant Bacteria Infection in Diabetic Foot Patients,” Infection and Drug Resistance 14 (2021): 627–637. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Research data are not shared.