Risk factors for infection with multidrug-resistant organisms in diabetic foot ulcer patients: A systematic review and meta-analysis

Dong Chen, Xin-Bang Liu, Bai Chang?

1. Tianjin University of Traditional Chinese Medicine, Tianjin 301617, China

2. Tianjin Medical University Chu Hsien-ⅠMemorial Hospital, Tianjin 300134, China

Keywords:Diabetic foot ulcer Multidrug-resistant organisms Risk factors Meta-analysis

ABSTRACT Objective: To systematically evaluate the risk factors for multidrug-resistant organisms(MDROs) infection in patients with diabetic foot ulcer(DFU). Methods: The quality assessment of outcome measures was performed by searching the Web of Science, Embase library, PubMed, Cochrane Library databases and screening the literature on the risk factors of MDROs infection in DFU patients according to the inclusion and exclusion criteria, and meta-analysis was performed using revman5.3 analysis software. Results: 13 literature was retrieved, involving in 1715 patients. A total of 15 risk factors were included in the analysis and the meta-analysis showed that Previous hospitalization(OR＝2.61,95%CI[1.51,4.52],P=0.0006), Previous antibiotic use(OR＝2.17,95%CI[1.24-3.78],P＜0.01), Type of diabetes(OR＝2.44,95%CI[1.29-4.63],P＜0.01), Nature of ulcer(OR＝2.16,95%CI[1.06-4.40],P=0.03),Size of ulcer(OR＝2.56,95%CI[1.53-4.28],P＜0.01), Osteomyelitis(OR＝3.50,95%CI[2.37-5.17],P＜0.01), Peripheral vascular disease(OR＝2.37,95%CI[1.41-3.99],P＜0.01),and Surgical treatment(OR＝4.81,95%CI[2.95-7.84],P＜0.01)were closely associated with MDROs infection in DFU patients. Conclusions: The risk factors of MDROs infection in patients with DFU were previous hospitalization, previous antibiotic use, type of diabetes,nature of ulcer, size of ulcer, osteomyelitis, peripheral vascular disease, and surgical treatment.This study is conducive to early detection of MDROs infection in high-risk groups and timely comprehensive treatment to delay the development of the disease.

1. Introduction

Diabetic foot ulcer (DFU) refers to the lower extremity vascular ischemia and neuropathy caused by diabetes, which leads to a certain degree of infection, ulcer and gangrene in the deep tissue of the foot. In severe cases, it can eventually lead to amputation[1,2]. Most patients with severe DFU currently require treatment with antibiotic drugs[3]. In recent years, due to the gradual increase in the frequency of antibiotic drug use, antibiotic resistant pathogens have emerged around the world, which is making the selection of appropriate empirical antibiotics for the treatment of DFU patients fewer and fewer[4, 5]. Multi-drug resistant bacteria are mainly a class of microorganisms that are resistant to a wide range of antibiotics,which cause slower wound recovery and adversely affect the healing process as well as worsen people's quality of life[6]. With further study of MDROs infection in DFU patients, numerous risk factors have been gradually recognized. Understanding these risk factors could be of immense help in the recovery of DFU patients.In order to solve these problems, we aim to quantify the risk factors of MDROs infection in DFU patients through a meta-analysis of previous research, and provide evidence-based basis for the prevention of disease infection in the future.

2.Materials and methods

2.1 Study registration

This study strictly follows the principles of the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) [7]and the protocol is registered with PROSPER under the accession number（CRD42021227683).

2.2 Literature search strategy

The literature was searched through the following databases:Web of Science, Embase library, PubMed, Cochrane Library, and the medical subject headings (Mesh) terms or keywords were mainly: “Diabetic Foot”,“Diabetic Foot Ulcer”,“Drug Resistance”,“Multiple”,“MDROS”,“risk factors”,“cohort studies”,“prospective studies”. The search was last updated in May 2021. We screened the titles and abstracts of the retrieved literature, then reviewed the full text and included the eligible literatures. If the full text was not retrieved, the completeness of the literature was assured by communicating with the original authors through relevant contacts in the literature.

2.3 Inclusion and exclusion criteria

All prospective studies and case-control studies met the inclusion criteria if the following criteria were met: (1) The patients in case group were DFU patients with MDROs infection (including patients with a type of MDROs infection), while the patients in control group were DFU patients without MDROs infection; (2) The types of studies include cohort study, case-control study and cross-sectional study; (3) There are clear diagnostic criteria of MDROs infection and detailed description of risk factors of DFU with MDROs infection.Exclusion Criteria: (1) reviews, meta-analysis, animal studies and other studies with inconsistent experimental content and incomplete literature data; (2) The full text and detailed literature on risk factors were not available.

2.4 Data extraction and quality evaluation

Literature data were extracted and screened by two researchers independently. Any disagreement could be sought from a senior researcher for discussion to reach a consensus. Firstly, repetitive literature and the literature that did not meet the research content were eliminated by browsing the abstract and title, and then the literature that met the inclusion criteria was recruited by reading the full text. The content of literature extraction includes: (1) The basic information includes author, title, study design and publication time; (2) Baseline data included the number of studies, incidence of MDROs infection, odds ratio (OR) and 95% confidence interval(CI) of risk factors in each group;(3) Outcome measures and detailed data of each group. Two researchers individually assessed the risk of the included articles according to the Newcastle-Ottawa Scale(NOS) [8]. In case of disagreement, a third researcher can participate in the evaluation and reach an agreement. The NOS consists of eight items, which are divided into three aspects: selection, Comparability of case and control groups, and exposure to an overall evaluation of the literature for risk of bias. The NOS scores ranged from 0 to 9.The higher the score, the higher the quality of literature. Among the 13 articles included in this study, the quality score of 3 articles was 6 points, the quality score of 7 articles was 7 points, and the quality score of 3 articles was 8 points, The quality of the included literature is relatively good.

2.5 Statistical analysis

The Revman 5.3 software was used to analyze OR and 95% CI of all risk factors. The heterogeneity was analyzed by Q test and I2test:If the test show statistical homogeneity (P > 0.1, I2< 50%), the fixed effect model can be used to integrate the data for analysis; If the test show statistical heterogeneity (P ≤ 0.1, I2≥ 50%), then subgroup analysis or sensitivity analysis can be selected according to the data type to check the heterogeneity. If the heterogeneity still exists,random effect model can be selected to data analysis because of the similar study baseline and certain clinical homogeneity. Funnel plot analysis was used to evaluate publication bias with the way of Egger's linear regression test. All tests were two-sided, and statistical significance was identified as P < 0.05, except for the heterogeneity test, which was considered statistically significant at P < 0.10.

3. Results

3.1 Search results and study characteristics

A total of 785 articles were initially retrieved. 148 repetitive literature were excluded, 34 literature of review, systematic review and animal experiment were excluded by title and abstract, and 590 literature which is inconsistent with experimental methods and contents were screened out by further Scanning the full text.13 literature[9-21] were ultimately incorporated in the qualification criteria of qualitative analysis. 15 risk factors for MDROs infection were included in this review. The baseline data of the included articles is summarized in Table 1

3.2 Meta‐analysis

15 risk factors for MDROs infection were analyzed by metaanalysis, 8 of them were statistically significant, including: Previous antibiotic use, Previous hospitalization, Osteomyelitis, Size of ulcer,Nature of ulcer, Type of diabetes, Surgical treatment and Peripheral vascular disease. The OR, 95% CI and heterogeneity test outcome of each risk factors were shown in Table 2

3.3 Publication bias

Only one risk factor (Size of ulcer) may have publication bias(Egger’s test P=0.004), indicating that there is a low-risk publication bias in the overall study. The outcome of each test were shown in(Table 3)

Table 1 Baseline characteristics of included 13 studies.

Table 2 The outcome of “heterogeneity test” and “OR values” of 15 risk factors

Table 3 Publication bias of risk factors for MDROs infection

3.4 Age

6 studies were included, with 431 cases in the MDROs infection group and 311 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.62, I2=0%). The results of 6 studies showed that age was not a risk factor for MDROs infection(OR＝0.98, 95%CI[0.95, 1.01], Z＝1.11, P=0.27＞0.05)

3.5 Sex

4 studies were included, with 311 cases in the MDROs infection group and 163 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.48, I2=0%). The results of 4 studies showed that sex was not a risk factor for MDROs infection(OR＝0.79, 95%CI[0.45, 1.40] , Z＝0.81, P=0.42＞0.05)

3.6 Previous hospitalization

5 studies were included, with 307 cases in the MDROs infection group and 362 cases in the non-MDROs infection group. The heterogeneity test outcome (P<0.1, I2=83%) showed that there was statistical heterogeneity between the two groups. Further sensitivity analysis found that Richard's studies[17] was the main cause of heterogeneity, so we chose a fixed effects model for this data analysis (P=0.14, I2=46%) after excluding heterogeneity. The results of 5 studies showed that previous hospitalization was a risk factor for MDROs infection(OR＝2.61, 95%CI[1.51, 4.52] , Z＝3.44,P=0.0006＜0.05)(Figure 1)

Figure1 Previous hospitalization (sensitivity analysis)

3.7 Previous antibiotic use

7 studies were included, with 429 cases in the MDROs infection group and 518 cases in the non-MDROs infection group. The heterogeneity test outcome (P<0.1, I2=87%) showed that there was statistical heterogeneity between the two groups. Further sensitivity analysis showed no studies causing heterogeneity. Considering the similar baseline and clinical homogeneity between the two groups,we chose the random effect model for analysis. The results of 7 studies showed that previous antibiotic use was a risk factor for MDROs infection(OR＝2.17, 95%CI[1.24, 3.78] , Z＝2.72,P=0.006＜0.05)(Figure 2)

Figure2 Previous antibiotic use

3.8 Type of diabetes

4 studies were included, with 311 cases in the MDROs infection group and 173 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.22, I2=31%). The results of 4 studies showed that the type of diabetes was a risk factor for MDROs infection(OR＝2.44, 95%CI[1.29, 4.63] , Z＝2.73,P=0.006＜0.05)(Figure 3)

Figure3 Type of diabetes

3.9 Duration of diabetes

5 studies were included, with 332 cases in the MDROs infection group and 260 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.12, I2=45%). The results of 5 studies showed that duration of diabetes was not a risk factor for MDROs infection(OR＝0.97, 95%CI[0.91, 1.03] , Z＝0.98, P=0.33＞0.05)

3.10 Nature of ulcer

5 studies were included, with 320 cases in the MDROs infection group and 241 cases in the non-MDROs infection group. The heterogeneity test results (P=0.03, I2=63%) showed that there was statistical heterogeneity between the two groups. Further sensitivity analysis found that Kang's studies [15] was the main cause of heterogeneity, The results of 4 studies showed that nature of ulcer was a risk factor for MDROs infection after excluding heterogeneity(OR＝3.01, 95%CI[1.58, 5.72] , Z＝3.36, P＜0.05).According to the type of ulcer, neuro ischemic ulcer was used as a risk factor for subgroup analysis, the outcome showed (OR＝4.56, 95%CI[1.95, 10.65] , Z＝3.50, P＜0.05), and there was no significant heterogeneity between the studies (P=0.31 , I2=3%). The combined results showed that necrotic ulcer was used as a risk factor for subgroup analysis (OR＝1.73, 95%CI[0.65, 4.61] , Z＝1.10,P=0.27＞0.05]) ,and there was no significant heterogeneity between the studies (P=0.48 , I2=0%).(Figure 4A,4B)

Figure4A Nature of ulcer(sensitivity analysis)

Figure4B Nature of ulcer(subgroup analysis)

3.11 Duration of ulcer

5 studies were included, with 368 cases in the MDROs infection group and 303 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.45, I2=0%). The results of 5 studies showed that the duration of ulcer was not a risk factor for MDROs infection(OR＝0.91, 95%CI[0.66, 1.25] , Z＝0.60, P=0.55＞0.05)

3.12 Size of ulcer

8 studies were included, with 559 cases in the MDROs infection group and 468 cases in the non-MDROs infection group. The heterogeneity test outcome (P＜0.1, I2= 91%) showed that there was statistical heterogeneity between the two groups. Further sensitivity analysis found that Zubair 2019's studies [21] was the main cause of heterogeneity. After exclusion of heterogeneity, subgroup analysis was performed with each 4cm2 increase in size of ulcer as a risk factor. The outcome showed (OR＝6.78, 95%CI[4.28, 10.75] , Z＝8.15, P＜0.05). There was mild heterogeneity between studies(P=0.07, I2=57%). Subgroup analysis was performed with each 1cm2 increase in size of ulcer as a risk factor. The results showed(OR＝1.49, 95%CI[1.23, 1.80] , Z＝4.11, P＜0.05), there was no significant heterogeneity between the studies (P=0.74, I2 = 0%).(Figure 5)

Figure5 Size of ulcer(subgroup analysis)

3.13 Osteomyelitis

9 studies were included, with 508 cases in the MDROs infection group and 637 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.32, I2=14%). The results of 9 studies showed that osteomyelitis was a risk factor for MDROs infection（OR＝3.50, 95%CI[2.37, 5.17] , Z＝6.30, P＜0.05)(Figure 6)

Figure6 Osteomyelitis

3.14 Retinopathy

4 studies were included, with 248 cases in the MDROs infection group and 272 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.13, I2=47%). The results of 4 studies showed that retinopathy was not a risk factor for MDROs infection(OR＝1.43, 95%CI[0.80, 2.53] , Z＝1.21, P=0.23＞0.05)

3.15 Nephropathy

3 studies were included, with 160 cases in the MDROs infection group and 122 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.32, I2=13%). The results of 3 studies showed that nephropathy was not a risk factor for MDROs infection(OR＝1.92,95%CI[0.95, 3.88] , Z＝1.81, P=0.07＞0.05)

3.16 Neuropathy

4 studies were included, with 259 cases in the MDROs infection group and 173 cases in the non-MDROs infection group. The heterogeneity test results (P=0.08, I2=55%) showed that there was statistical heterogeneity between the two groups. Further sensitivity analysis showed no studies causing heterogeneity. Considering the similar baseline and clinical homogeneity between the two groups,we chose the random effect model for analysis. The results of 4 studies showed that neuropathy was not a risk factor for MDROs infection(OR＝1.12, 95%CI[0.47, 2.70] , Z＝0.26, P=0.8＞0.05)

3.17 Peripheral vascular disease

4 studies were included, with 277 cases in the MDROs infection group and 220 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.89, I2=0%). The results of 4 studies showed that peripheral vascular disease was a risk factor for MDROs infection(OR＝2.37, 95%CI[1.41, 3.99] , Z＝3.26, P=0.001＜0.05)(Figure 7)

Figure7 Peripheral vascular disease

3.18 Surgical treatment

6 studies were included, with 446 cases in the MDROs infection group and 266 cases in the non-MDROs infection group. We chose a fixed effects model for this data analysis (P=0.16, I2=37%). The results of 6 studies showed that surgical treatment was a risk factor for MDROs infection(OR＝4.81, 95%CI[2.95, 7.84] , Z＝6.31, P＜0.05)(Figure 8)

Figure8 Surgical treatment

4. Discussion

The results of this meta-analysis provide a strong evidence-based medical basis, indicating that patients with DFU are more prone to MDROs infection when they have the following risk factors: "Previous hospitalization, Previous antibiotic use, Type of diabetes,Nature of ulcer, Size of ulcer, Osteomyelitis, Peripheral vascular disease and Surgical treatment".

Previous hospitalization, Previous antibiotic use and Surgical treatment are the risk factors of MDROs infection. MDROs infection is intensively interrelated to nosocomial infection[22], and nosocomial infection is also related to the use of antibiotics. Some studies [23, 24] show that irrational use and abuse of antibiotics can lead to nosocomial infection and increase the rate of antibiotic resistance. Over time, doctors will use higher levels of antibiotics,gradually forming a vicious circle. Kang et al [17]found that long hospital stay is easy to lead to the risk of cross infection, and taking antibiotics is easy to lead to bacterial resistance gene mutation,which leads to multiple drug resistance. Surgical treatment is also a common treatment at present. yet it may damage the wound healing for microcirculation changes. Antibiotics is needed to be used in Infected wounds,but excessive or improper use of antibiotics also lead to MDORs infection[25]. Therefore, medical personnel should strengthen aseptic operation to prevent and control cross infection in the process of dressing change and nursing, and in using antibiotics,it is recommended to use appropriate antibiotics after drug sensitivity test rather than empirical therapy[26, 27].

Type of diabetes, Nature of ulcer and Size of ulcer can also lead to MDROs infection. Some studies[28, 29]show that about 50% of DFU are neuro ischemic ulcer. Neuro ischemic ulcer can lead to microcirculation damage and local tissue can not get sufficient nutrition, thus delaying wound healing and eventually leading to MDROs infection. Once the ulcer area exceeds 4cm2, the risk of MDROs infection increase rapidly[9, 14]. Therefore, medical personnel should pay more attention to the wound recovery and timely use antibiotics according to the drug sensitivity test to prevent the occurrence of MDROs infection in some patients with large ulcer and especially the neuro ischemic ulcer.

Both of Osteomyelitis and Peripheral vascular disease can lead to the development of MDROs infection, and the outcome of this study show that osteomyelitis increases the risk of MDROs infection by 3.5-fold. Some studies [30, 31] have found that bacteria usually invade the osteocytic-canalicular networks, where they can escape the effects of mechanical debridement and antibiotics, and these bacterial colonies will form abscesses in the skin, which become a physical barrier to prevent immune cells from entering and killing bacteria, so as to make bacteria survive for a long time, thus inducing MDROs infection. Therefore, it should be careful for antibiotic treatment of patients with DFU accompanying osteomyelitis, and properly reducing the duration of antibiotic use may avoid the occurrence of MDROs.

Nevertheless, there are some limitations to this study. First, the quality of the 13 studies included in this study varies to a certain extent, and parts of studies are case-control study, which leads to a problem of insufficient demonstration strength of causal relationship between variables. Second, the treatment periods of each study were widely distributed, and it is not clear whether the results were affected. Third, in some studies, it can not be judged if the MDROs infection is pre-hospital or post-hospital, which may affect the comprehensive outcome. Finally, in some risk factor’s analyses,the limited number of studies may cause some deviations. We also found it is necessary to develop a risk prediction model, but it is limited by some problems such as the classifications of researches are simplistic and inconsistent. Therefore, a well-designed plan to provide evidence-based medical support for deeper protection against MDROs infection is highly necessary.

In summary, Previous hospitalization, Previous antibiotic use,Type of diabetes, Nature of ulcer, Size of ulcer, Osteomyelitis and Amputation are risk factors for MDROs infection in patients with DFU. So it is very important for medical personnel to pay attention to these risk factors and take not only active treatment but also prevention measures to reduce the probability of MDROs infection in patients with DFU.

Author’s contribution

The implementation of this article is mainly for Dong Chen and Bai Chang, Dong Chen and Xinbang Liu to search, screen and evaluate the quality of literature. Dong Chen wrote, planned and revised.Bai Chang guides and reviews the articles. All authors evaluated the articles and searched the literatures.