Background

Chlorhexidine (CHX) is a cationic bisbiguanide [1] and an antiseptic agent active against Gram-positive and -negative bacteria widely used in clean surgeries [2,3,4,5]. CHX binds to the negatively charged bacterial cell wall to disrupt the cell barrier [1]. CHX is bacteriostatic at low concentrations and bactericidal at higher concentrations [1].

In the past decades, CHX has been used to prevent surgical-related infections [6, 7], such as for decontamination of the oropharynx to avoid respiratory tract infection [8] or for gingival health [9], preoperative skin preparation to avoid surgical site infection [10, 11], or disinfection of medical appliances to avoid nosocomial infection [1, 9, 12,13,14]. CHX bathing is an effective measure in reducing the levels of pathogens on the skin, and it can also prevent catheter colonization and central line-associated bloodstream infection [15]. CHX is associated with reduced postoperative surgical site infections (SSI) compared with povidone-iodine in clean-contaminated surgery [16].

Although several meta-analyses reported the impact of different types of CHX uses for patients undergoing various types of surgery [17,18,19,20,21,22,23,24], no meta-analysis summarized the overall effectiveness of CHX specifically for cardiac surgery. Indeed, cardiac surgery is highly invasive, usually long, and carries a high risk of infection. Infectious complications occur in 5–21% of the patients after cardiac surgery [25, 26]. After cardiac surgery, the risk of superficial wound infection is 0.5%-8%, and deep sternal wound infections occur in 0.4–2.0% of the cases [25, 26]. Infectious complications prolong the hospital stay, increase healthcare costs, and have dismal outcomes [25, 26].

Therefore, this meta-analysis aimed to examine the impact of CHX on infections after cardiac surgery compared with other cleansers or antiseptics.

Methods

This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [27]. The relevant articles were searched based on the patient, intervention, comparison, outcome (PICO) principle [28], followed by screening based on the eligibility criteria: (1) population: patients who underwent open cardiac surgery, irrespective of the indication; (2) intervention or exposure: any type of application of CHX in the treatment or exposed group; (3) outcome: number of patients with infections (SSI, pneumonia, bloodstream infection, urinary infection, or nosocomial infection); (4) comparison: placebo, other antiseptic agents, or without CHX; (5) full-text article published in English. PubMed, Embase, and the Cochrane Library were searched from inception up to October 2020 for potentially eligible studies using the MeSH terms of “general surgery” AND “chlorhexidine”, as well as relevant key words such as cardiac or cardiovascular. The exact strategies for all three databases are presented in the Additional file. The literature search and selection of the studies were performed independently by two investigators (Fengxia Weng and Lingying He). Any discrepancy was solved by discussion.

Data extraction

Study characteristics (authors, year of publication, the country where the study was performed, type of the study design, and number, age, and sex of the patients), exposure parameters (method for the application of CHX, density, and frequency of the treatments); primary outcome (SSI), and secondary outcomes (superficial infection, deep wound infection, bloodstream infection, urinary tract infection, nosocomial infection, and pneumonia) were extracted independently by two investigators (Fengxia Weng and Lingying He). Any discrepancy was solved by discussion.

Quality of the evidence

The level of evidence of all articles was assessed independently by two authors (Fengxia Weng and Lingying He) according to the Cochrane Handbook for randomized controlled trials [29, 30] and the Newcastle–Ottawa Scale (NOS) criteria for observational studies [31]. Discrepancies in the assessment were resolved through discussion until a consensus was reached.

Statistical analysis

All analyses were performed using STATA SE 14.0 software (StataCorp, College Station, Texas, USA). The results were summarized as odds ratios (ORs) and 95% confidence interval (95% CI). Statistical heterogeneity among the studies was evaluated using Cochran’s Q-test and the I2 index. A Q-test P-value < 0.10 and I2 > 50% indicated high heterogeneity. Considering the different types of agents in the control group and the various CHX regimens among the included studies, the random-effect model was applied for all analyses to avoid an overestimation of the results. Possible publication bias was not evaluated by funnel plots and Egger’s test because the numbers of studies included in each quantitative analysis were less than 10, in which case the funnel plots and Egger’s test could yield misleading results [29].

Results

Literature search

Figure 1 and the Additional file present the literature search process. The initial searched yielded 513 records, and 424 were left after removing the duplicates. These records were screened, and 212 were excluded. The 212 full-text articles or abstracts were assessed for eligibility, and 198 were excluded (52 because of study design/aim, eight for the outcomes, 30 for the populations, 57 for the intervention/exposure, 21 for non-human studies, 10 for no accessible full-text, five for being meta-analyses, and 15 for being published in a language other than English). Finally, 14 studies were included.

Fig. 1
figure 1

PRISMA 2009 flow diagram

Full size image

Characteristics of the studies

Among the 14 studies (Table 1), there were five randomized controlled trials [32,33,34,35,36], six prospective cohort studies [37,38,39,40,41,42], two retrospective cohort studies [43, 44], and one case–control study [45]. Five studies examined the effect of an oral rinse [32,33,34,35, 37], four examined skin antiseptic [38,39,40, 43], and five examined the disinfection of surgical-related appliance [36, 41, 42, 44, 45]. There were 8235 patients in the CHX group and 6901 in the control group.

Table 1 Literature search and study characteristic
Full size table

Among the five randomized controlled trials [32,33,34,35,36], only one had an unclear risk of bias for two items [36] (Additional file 1: Table S1). Among the cohort studies [37,38,39,40,41,42,43,44], six studies scored 7 stars [37, 38, 41,42,43,44], one scored 8 stars [39], and one scored 9 stars [40] (Additional file 1: Table S2). The case–control study scored 9 stars [45] (Additional file 1: Table S3).

Surgical site infection

Eight studies could be included for the impact of CHX on SSI [34, 38,39,40,41, 43, 44]. CHX did not influence the risk of SSI compared with control interventions (OR = 0.77, 95% CI: 0.57–1.04, P = 0.090; I2 = 63.5%, Pheterogeneity = 0.008) (Fig. 2a and Table 2). When considering the type of control, CHX did not influence the risk of SSI compared with placebo or isopropyl alcohol (IPA), CHX protected against SSI when without CHX was used as control (OR = 0.46, 95% CI: 0.22–0.94, P = 0.032; I2 = 59.2%, Pheterogeneity = 0.061) (Fig. 2b).

Fig. 2
figure 2

a Forest plot of surgical site infection (SSI) comparing patients treated with CHX vs. control (placebo, isopropyl alcohol (IPA), or without CHX). b Forest plot of surgical site infection comparing patients treated with CHX vs. control, according to the type of control (placebo, IPA, or without CHX)

Full size image
Table 2 Subgroup analyses: CHX vs. control
Full size table

Superficial and deep wound infections

Three studies could be included to analyze superficial/deep wound infection [38, 40, 41]. CHX was protective against superficial wound infection after cardiac surgery compared with control interventions (OR = 0.42, 95% CI: 0.26–0.70, P = 0.001; I2 = 19.4%, Pheterogeneity = 0.289) (Fig. 3a and Table 2). There was no difference between CHX and IPA, while the difference was driven by no intervention as control (Fig. 3b). No protective effect of CHX was observed for deep wound infection (OR = 0.79, 95% CI: 0.40–1.58, P = 0.509, I2 = 51.7%, Pheterogeneity = 0.126) (Fig. 4a,b and Table 2).

Fig. 3
figure 3

a Forest plot of superficial infection comparing patients treated with CHX or control (isopropyl alcohol (IPA) or without CHX). b Forest plot of surgical site infection comparing patients treated with CHX vs. control, according to the type of control (IPA or without CHX)

Full size image
Fig. 4
figure 4

a Forest plot of deep wound infection comparing patients treated with CHX or control (isopropyl alcohol (IPA) or without CHX). b Forest plot of deep wound infection comparing patients treated with CHX vs. control, according to the type of control (IPA or without CHX)

Full size image

Effect of the type of CHX intervention

Figure 5 and Table 2 show that the lack of association between CHX and SSI (OR = 0.77, 95% CI: 0.57–1.04, P = 0.090; I2 = 63.5%, Pheterogeneity = 0.008) remained insignificant when considering only the oral rinse (OR = 0.88, 95% CI: 0.58–1.33, P = 0.549), skin antiseptic (OR = 0.82, 95% CI: 0.61–1.11, P = 0.201; I2 = 51.4%, Pheterogeneity = 0.084), and disinfection of surgical-related appliance (OR = 0.20, 95% CI: 0.01–3.75, P = 0.279; I2 = 86.9%, Pheterogeneity = 0.006).

Fig. 5
figure 5

Subgroup analysis on SSI comparing patients treated with CHX or control (IPA or without CHX)

Full size image

Effect of CHX on infections other than surgical wound

The meta-analysis of four studies [32, 34, 37, 44] showed that CHX was not protective against urinary tract infection (OR = 0.80, 95% CI: 0.46–1.38, P = 0.415; I2 = 18.1%, Pheterogeneity = 0.300) (Fig. 6 and Table 2). On the other hand, the use of CHX protected against bloodstream infection [32, 36, 42, 44] (OR = 0.36, 95% CI: 0.16–0.80, P = 0.012; I2 = 16.3%, Pheterogeneity = 0.310) (Fig. 7 and Table 2), nosocomial infection [32, 34, 35, 37] (OR = 0.55, 95% CI: 0.44–0.69, P < 0.001; I2 = 0.0%, Pheterogeneity = 0.520) (Additional file 2: Fig. S1 and Table 2), and pneumonia [32,33,34,35, 37, 44] (OR = 0.26, 95% CI: 0.11–0.61, P = 0.002; I2 = 76.6%, Pheterogeneity = 0.001).

Fig. 6
figure 6

Forest plot of urinary tract infection comparing patients treated with CHX or control (IPA or without CHX)

Full size image
Fig. 7
figure 7

Forest plot of BSI comparing patients treated with CHX or control (IPA or without CHX)

Full size image

Sensitivity analyses

Additional file 3: Fig. S3 shows the analyses according to RCTs (A) and observational studies (B). The same conclusions were reached for the two study types for nosocomial infections, pneumonia, SSI, and urinary tract infections. The meta-analysis of RCTs showed no impact of CHX on bloodstream infection, while the meta-analysis of the observational studies showed an impact.

Discussion

Although several meta-analyses reported the impact of different types of CHX applications for patients undergoing various types of surgery [17,18,19,20,21,22,23,24], no meta-analysis summarized the overall effectiveness of CHX specifically for cardiac surgery, which carries a high risk of infections [25, 26]. Therefore, this study aimed to examine the impact of CHX compared with other cleansers or antiseptics used in cardiac surgeries. The results suggest that CHX was not protective for SSI, deep wound infection, and urinary tract infections but was protective against superficial SSI, bloodstream infection, nosocomial infections, and pneumonia.

The use of CHX before surgery is already well-documented and well-supported by a large amount of evidence, as revealed by many meta-analyses on the subject [17,18,19,20,21,22,23,24]. This meta-analysis adds further evidence by showing that CHX can decrease the risk of superficial SSI, bloodstream infection, nosocomial infections, and pneumonia, leading to poor outcomes in patients who are already affected by their cardiac condition, surgery itself, and often multiple comorbidities. This is supported by other meta-analyses and studies regarding superficial wound infection [12, 40, 41, 46, 47], bloodstream infection [15, 18, 48], nosocomial infections [17, 19, 24, 32,33,34,35, 37, 49], and pneumonia [8, 17, 18, 24, 33, 35, 37, 49,50,51]. Nevertheless, heterogeneity is observed in those previous studies and meta-analyses, mainly due to the different methods of using CHX, the different concentrations, and the different frequencies of use. Nevertheless, the studies agree that CHX generally contributes to the prevention of those infections. On the other hand, the present meta-analysis showed no significant impact of CHX for SSI, which is probably driven by the lack of association with deep wound infection. In addition, there was no association with urinary infections. Nevertheless, previous studies did report associations between CHX and lower odds of those infections [1, 6, 12, 13, 16, 17, 19, 20, 22, 23, 34, 38,39,40,41, 44,45,46,47, 49]. The discrepancies could be due to several factors such as the included studies, the study populations, and the CHX regimen. Nevertheless, previous meta-analyses that included cardiac surgery patients support the present meta-analysis [18, 24, 51, 52].

Mechanistically, the control of respiratory infection by CHX is probably due to the use of oral CHX [18, 52,53,54,55,56]. Nevertheless, pneumonia can also lead to bloodstream infection [57,58,59], and preventing one can prevent the other. Regarding the decreased odds of bloodstream infection, this is probably driven by using CHX for the skin and the medical devices since vascular devices are among the first causes of bloodstream infection [60,61,62]. Of course, decreased superficial SSI is linked to skin disinfection [1, 12, 13, 40, 46]. Future studies could look at the interactions between specific uses of CHX with specific infections, but the present meta-analysis could not perform such analyses.

The conclusions of the meta-analysis must be considered in the light of its limitations. First, we included observational studies in our analysis due to the small number of RCTs in this study field, but at the price of introducing heterogeneity and bias. Second, the treatment regimens in the control group were variable among the studies. Accordingly, the random-effect model was applied to all quantitative analyses regardless of the results of Cochran’s Q test and the I2 index. Third, mortality could not be analyzed because it was not reported by enough studies. Finally, despite a relatively large number of patients, the numbers of patients and studies in each subanalysis were small. More studies are required.

Conclusions

In conclusion, in patients undergoing cardiac surgery, CHX is not protective for SSI, deep wound infection, and urinary tract infections but is protective against superficial infection, bloodstream infection, nosocomial infections, and pneumonia. Therefore, CHX can be useful in cardiac surgeries to prevent infections, especially superficial infection, bloodstream infection, nosocomial infection, and pneumonia. Future studies can work on a standardized protocol to determine the recommended concentration and frequency for different application methods of CHX in patients undergoing cardiac surgery.