Inflammatory Markers in Diabetic Foot Infection: A Meta-Analysis

Diabetic foot is one of the most serious complications of diabetes, with a global prevalence of 6.3%.¹ Currently, diabetic foot infection affects about 15% of patients with diabetes.² Diabetic foot infection is clinically considered to be soft tissue or bone infection below the ankle, including paronychia, cellulitis, myositis, abscess, necrotizing fasciitis, septic arthritis, tendonitis, and osteomyelitis.³ The considerable morbidity associated with infected diabetic foot (IDF) warrants attention, and the development of preventive measures is of particular importance. These infections account for the majority of hospitalizations associated with diabetes other than non-traumatic amputation.⁴ The treatment process increases the economic burden on patients. Additionally, according to the clinical data, up to 28% of diabetic foot ulcers with IDF may require some type of lower extremity amputation.⁵ 

Some studies have used inflammatory markers to detect diabetic foot infection, while others have not. Typical markers include white blood cell (WBC) count, C-reactive protein (CRP) level, and procalcitonin (PCT) level.⁶ Many studies of the diabetic foot include the detection of more than 1 of these inflammatory markers.^7-9 In general, elevated inflammatory markers are considered a sign of infection, rising again after initial values fall, or rising over an extended period.¹⁰ However, 1 study found no significant difference in WBC count  between patients with IDF and patients with non-infected diabetic foot (NIDF).¹¹ Another study indicated that CRP level is not a valuable marker of infection in patients with diabetic foot.¹² Additionally, a pilot study with a small number of cases showed that PCT level was not significantly higher in patients with IDF, even in those who had undergone amputation.¹³ However, 2 studies proposed that WBC count, CRP level, and PCT level were significantly different before and after infection, which suggests that these could be used as effective inflammatory markers in the detection of diabetic foot infection.^14,15

As noted previously, as of this writing no meta-analysis existed confirming the significant differences in WBC count, CRP level, and PCT level between patients with IDF and patients with NIDF. Therefore, no reliable evidence existed for the clinical use of inflammatory markers to assist in the detection of diabetic foot infection. The purpose of this study was to investigate whether WBC count, CRP level, and PCT level are helpful in detecting diabetic foot infection.

Database and literature search

PubMed, SpringerLink, and Web of Science databases were searched electronically using 3 search terms: “diabetic foot,” “infection,” and “inflammatory markers.” These search terms and their combinations with “AND” or “OR” were searched separately; searches were done separately when searching article title and abstract. Manual retrieval of relevant studies and the related data were extracted according to the standard research practice.

Study selection

The following criteria were used to select relevant studies. First, observational studies were chosen; these included prospective cohorts, retrospective cohorts, and cross-sectional surveys.¹⁶ Second, other study types were excluded; these included letters, meeting abstracts, animal experiments, biological research, non-English language studies, studies with no data, and studies that did not analyze infection and non-infection separately. Two of the authors (LXZ and JZ) reviewed the titles and abstracts and excluded literature that did not satisfy the study criteria. They evaluated the full-text versions of the remaining studies and selected studies for this meta-analysis. During the selection process, differences among reviewers and quality assessments were resolved through discussion and consensus.

Data extraction and quality assessment

Author, publication year, country, study design, study duration, patient age, male to female ratio, the proportion of patients with and without diabetic foot infection, WBC counts, CRP levels, and PCT levels were extracted. The predictive variables included the number of patients with IDF and with NIDF as well as WBC, CRP, and PCT concentrations.

The Newcastle-Ottawa Scale was used to evaluate the quality of studies and data.¹⁷ The scores were presented with concrete numerical values, and the full score was 9 points. Studies were judged based on 3 broad factors—selection, comparability, and outcome. Selection includes 4 items with a maximum value of 1 point each: representation of the exposed cohort, selection of the non-exposed cohort, certainty of exposure, and demonstration that an outcome of interest was not present at the start of the study; comparability consists of control-matched significant factors (1 point) and research controlled (1 point); and outcome consists of assessment of outcomes (1 point), follow-up long enough for outcomes to occur (1 point), and advocacy of follow-up cohorts (1 point).

Data synthesis and analysis

For continuous data, the standardized mean differences (SMD) and 95% CIs were analyzed. The present authors conducted a random-effects model to analyze statistical heterogeneity. The random-effects model also resulted in a wider range of linear relationships and more conservative accuracy estimates.¹⁸ Statistical heterogeneity was evaluated using the inconsistency statistic (I²). The I² values ranged from 0% to 100%, with higher values indicating increasing heterogeneity.¹⁹ The funnel plot was mainly used to evaluate for publication bias. Subgroup analysis was conducted by prospective and retrospective analysis. All statistical analyses were carried out using Stata15.0 (StataCorp LLC).

Identification of eligible studies

A total of 1937 potential articles met the search criteria. The study authors (LXZ and JZ) removed 1305 duplicate articles. After an initial screening of all titles and abstracts, 454 articles were excluded, and the present authors assessed 178 articles for eligibility. The final analysis included 10 studies.^7-9,12,20-25 A flow diagram of the selection process is shown in Figure 1.

Characteristics of the included studies

A total of 785 participants were identified using the search strategy noted previously. The number of participants per study ranged from 12 to 89. Participant age ranged from 28 years to 80 years. Among them, 490 participants were patients with IDF, and 295 participants were patients with NIDF. The main study characteristics included author, publication year, country, study design, number of patients, IDF/NIDF status, infection rate, age, sex (male or female), detection methods, and quality score. These characteristics are summarized in Table 1. Data including WBC count, CRP level, and PCT level were presented as mean ± SD and are summarized in Table 2. The mean quality score of the Newcastle-Ottawa Scale was 7.0 (out of a possible 9 points).

Meta-analysis of WBC count

The forest plot shows the mean WBC count of retrospective studies (SMD: 0.44, 95% CI, −0.28, 1.17) and prospective studies (SMD: 0.54, 95% CI, 0.21, 0.88) (Figure 2A). Overall, mean WBC count (SMD: 0.51, 95% CI, 0.23, 0.79) was higher in patients with IDF. The forest plot, funnel plot, and sensitivity diagram of WBC count are shown in Figure 2.  Heterogeneity was confirmed in the WBC-related data (overall I² = 61.0%). The statistical test of the combined effect of WBC count was 3.57 (P <.0001), which was significant. The funnel plot shows that the WBC values before and after infection were not symmetrical to the meta-analysis, indicating the possibility of publication bias. Judging from the sensitivity analysis, the point estimates of all the results fell within 95% CI of the combined effect. All these results show that the overall stability of this study is good. 

Meta-analysis of CRP level

The forest plot shows the mean CRP level of the retrospective studies (SMD: 0.49, 95% CI, 0.18, 0.80) and the prospective studies (SMD: 1.34, 95% CI, 0.77, 1.91) (Figure 3A). Overall, mean CRP level (SMD: 1.05, 95% CI, 0.60, 1.50) was higher in patients with IDF. The forest plot, funnel plot, and sensitivity diagram of CRP level are shown in Figure 3. Heterogeneity was high (overall I² = 86.1%). The statistical test of the combined effect of CRP level was 4.56 (P <.0001), which was significant. The funnel plot shows that CRP levels before and after infection were not symmetrical to the meta-analysis, indicating the possibility of publication bias. The remaining studies were reanalyzed after excluding the independent studies, and the point estimates of all the results fell within 95% CI of the combined effect. All these results showed that the overall stability of this study was good. 

Meta-analysis of PCT level

The forest plot shows the mean PCT level of retrospective studies (SMD: 0.95, 95% CI, 0.35, 1.54) and prospective studies (SMD: 0.78, 95% CI, 0.27, 1.28) (Figure 4A). Overall, the mean PCT level (SMD: 0.80, 95% CI, 0.36, 1.24) was higher in patients with IDF. The forest plot, funnel plot, and sensitivity diagram of PCT level are shown in Figure 4. Heterogeneity was high (overall I² = 81.1%). The statistical test of the combined effect of PCT was 3.60 (P <.0001), which was significant. The funnel plot shows that PCT level before and after infection was not symmetrical to the meta-analysis, indicating the possibility of publication bias. After the sensitivity analysis, the authors found that the inclusion of the fourth study had a significant effect on the stability of the whole PCT value; thus, the meta-analysis of PCT level is unstable.

The study reported herein found significant differences in WBC count, CRP level, and PCT level between patients with IDF and patients with NIDF, with mean WBC count (SMD: 0.51, 95% CI, 0.23, 0.79), CRP level (SMD: 1.05, 95% CI, 0.60, 1.50), and PCT level (SMD: 0.80, 95% CI, 0.36, 1.24) higher in patients with IDF (P <.0001 for all). A meta-analysis of risk factors for amputation showed that increased WBC counts (SMD: 0.8, 95% CI, 0.449, 1.146) and elevated CRP levels (SMD: 0.8, 95% CI, 0.561, 1.035) were identified as risk factors for amputation.²⁶ However, the function of measuring WBC, CRP, and PCT in detecting diabetic foot infection was not emphasized in the study above in terms of their effects on amputation owing to the different topics and patients included in the meta-analysis. But in this present study, the function of WBC, CRP, and PCT in detection of diabetic foot infection was more prominent. 

Several possible biological mechanisms can explain these findings. First, WBC count is the most basic marker of infection.²⁵ When an infection occurs in a diabetic foot, the cause of the increased WBC counts may be increased production and release of WBCs in the bone marrow. Second, as a peptide produced in the liver, CRP is primarily stimulated by interleukin-6 and is used for complement binding and macrophage phagocytosis.²⁷ Serious infection stimulates endothelial cells to produce interleukin-6, causing the liver to synthesize the acute phase proteins, including CRP. Third, PCT is a protein secreted by non-neuroendocrine parenchymal cells and is a PCT peptide.²⁸ When infection occurs, target cells secrete PCT in response to lipopolysaccharide and other related factors, which results in the absence of essential hydrolases during post-transcription, leading to the observed increases in PCT level. One study notes that the most common pathogens in diabetic foot infection are gram-positive cocci, mainly Staphylococcus aureus and S epidermidis.²⁹ Gram-negative bacteria are also common pathogens.²⁹ Their effects will directly cause inflammatory reaction, leading to an increase in CRP level, WBC count, and PCT level. 

The function of WBC count in the detection of IDF was studied both because it is a common laboratory test indicator that can assist in detecting infection and other nonspecific inflammation³⁰ and also because of the special innate immunity and acquired immunity of physiological and pathological functions.³¹ Additionally, a growing body of evidence shows that CRP, in addition to being an effective inflammatory marker, is also a direct mediator of the inflammatory response and innate immunity.³² Further studies on WBC count and CRP level could possibly help provide information for the prevention of diabetic foot infection by increasing the understanding of their immune function. Although the meta-analysis of PCT was unstable, the results show that PCT could play a role in detecting IDF. Two previous studies indicate that PCT plays a role in guiding the use of antibiotics.^33,34 Thus, it is helpful to explore whether PCT can guide the application of antibiotics in the management of diabetic foot infection. In summary, WBC count, CRP level, and PCT level are useful laboratory assistant indexes for effective clinical detection of diabetic foot infection. Measurement of these markers is convenient and rapid, which is conducive to detecting early infection and supporting timely prevention measures.

The present study does have limitations. First, the severity of diabetic foot infection varies.⁵ The study did not analyze the different severities of such infection because of the different approaches and patients of the included studies and the limited data. Second, the small sample size may lead to inadequate efficacy and publication bias. Third, although subgroup analysis was performed, considerable heterogeneity remained among the analysis groups, indicating other heterogeneity sources that could not be systematically evaluated. Further high-quality studies are needed to verify the findings of this study. 

In summary, there were significant differences in WBC count, CRP level, and PCT level between patients with IDF and patients with NIDF. These findings strengthen the evidence that diabetic foot infection is accompanied by abnormal levels of WBC, CRP, and PCT. These inflammatory markers can be effective laboratory indexes that play important roles in the detection of diabetic foot infection. Additionally, a thorough understanding of these mechanisms will help in the development of new strategies to prevent diabetic foot infection through monitoring WBC count, CRP level, and PCT level. 

Authors: Yu-Ting Wang, BS; Le-Xuan Zhang, BS; Yang Li, BS; Jun Zhao, BS; and Hong-Lin Chen, MD

Affiliation: School of Public Health, Nantong University, Nantong City, Jiangsu Province, People’s Republic of China

Corresponding Author: Hong-Lin Chen, School of Public Health, Nantong University, Nantong City, SeYuan Road 20#, Jiangsu Province 226001, People’s Republic of China; 206148587@qq.com 

Disclosure: The authors disclose no financial or other conflicts of interest.