Abstract
Abstract
Background:
The use of 2% chlorhexidine in 70% alcohol (CHG) has been associated with reduction in catheter-related bloodstream infections and surgical site infection (SSI) in general surgery. Also, improved awareness of best practice from the perspective of the operative team is likely to result in reductions in SSI rates.
Methods:
This is an ambispective cohort study of patients undergoing elective cardiac surgery. Between January 2010 and December 2010, patients underwent surgical preparation using Alcohol Povidone Iodine (API). Between January 2011 and December 2011, the surgical team received education and switched to CHG for surgical preparation. Univariate analysis was performed to identify the impact of known risk factors for SSI. A logistic regression model was then fit to estimate the effect of education and CHG in the reduction of SSI in 2011, controlling for known SSI risk factors.
Results:
There was a substantial reduction in overall SSI rate in 2011 following staff education and the introduction of CHG. The overall unadjusted SSI rate was 4.67% versus 2.08% (p<0.05) for 2010 and 2011 respectively. Using a logistic regression model, the combined effect of education and CHG in 2011 was a 63% reduction in SSI in cardiothoracic surgery (OR 0.37, 95% CI: 0.17–0.83, p=0.016), controlling for age, major co-morbidities, and SSI risk factors.
Conclusions:
Using CHG as pre-operative antiseptic in cardiothoracic surgery in a risk-adjusted cohort with education of the surgical team is associated with significantly lower SSI infection rates when compared with API. Emphasis must be placed on the multifactorial approach required to prevent postoperative wound infections.
S
It is well recognized that certain interventions can and do reduce the risk of SSI. These include avoiding hair removal at the site, correctly timed antibiotic prophylaxis, pre-operative skin disinfection and an interactive wound dressing applied in theatre and left undisturbed for 48 h minimum [1–7,8]. Recent studies have demonstrated the benefits of chlorhexidine-based skin preparations in reducing bloodstream infections, and the debate regarding the most effective skin disinfectant is ongoing [9]. To date, neither the U.S. Centers for Disease Control and Prevention (CDC) or the World Health Organization (WHO) have suggested the optimum antiseptic with which to prevent post-operative SSIs. The CDC do, however, recommend a minimum of 0.5% chlorhexidine skin preparation with 70% alcohol for cleaning and disinfecting the site of insertion of vascular catheters [2].
Several studies comparing alcoholic povidone iodine (API) with 2% chlorhexidine in 70% alcohol (CHG) have been published. The most notable of these, by Darouiche et al., demonstrated a substantial decrease in SSI with CHG over PI using a single use sponge application system (Chloraprep) in general surgery [6]. A systematic review of research comparing CHG with API carried out in 2012 by Kamel et al. found that although quite a number of studies had been conducted, the concentration of chlorhexidine varied from study to study and the API (includes both alcoholic and non-alcoholic API in these studies and at different concentration) was sometimes combined with alcohol and other times with water [11]. Authors concluded that there is a lack of large well conducted randomized controlled trials to prove unequivocally the effectiveness of one antiseptic preparation over another [11].
The purpose of this study was to assess the combined efficacy of staff education and changing from API to CHG in reducing the incidence of SSI in cardiothoracic surgery over a 2-y period with prospective data collection.
Patients and Methods
The study was performed in a single tertiary referral centre for cardiothoracic surgery (CTS). The 2-y study period was from January 2010 through to December 2011. All patients having CABG, valve or combination CABG and valve surgery were prospectively enrolled in the study including those undergoing re-do cardiac surgery. Patients undergoing CTS other than those mentioned above were excluded. All patients' baseline characteristics were prospectively recorded on a proforma admission document and entered into an electronic database for data analysis. CTS morbidity and mortality risk factors were assessed using parameters assessed to compute the EuroSCORE a European system for cardiac operative risk evaluation [12]. Risk factors for infection were documented at the time of admission and using the National Nosocomial Infection Surveillance (NNIS) the US system for assessing risk factors for SSI.
The study was divided into two separate time periods. The first (January 1, 2010 to December 31, 2010) enrolled patients undergoing surgical preparation using API. No intervention in terms of team education was performed during this period. The second period (January 1, 2011, to December 31, 2011) involved the introduction of two interventions; a 2-mo education and training period, during which the surgical team attended sessions on the optimum surgical preparation technique and a switch to CHG for surgical pre-operative anaesthetic. Detail with regard to the pre-operative preparation is outlined, there was no alteration to this or to post-operative management during the study period.
Pre-operative preparation
All patients underwent the same pre-operative investigation and theatre preparation for cardiac surgery. Surgical antibiotic prophylaxis was vancomycin 1 gram BD and gentamicin 5mg/kg OD for 24 h. This was administered at induction of anaesthesia prior to surgical checklist completion [13]. Pre-operative skin preparation was three chlorhexidine 4% aqueous solution showers pre-operatively. These were timed as follows: The day before surgery, the evening before surgery and the morning of surgery. Prior to showering the patient was instructed to use chlorhexidine 4% aqueous solution to scrub chest (from chin to umbilicus), arms, and legs using a clean sterile dressing pad for each shower. The patient then dressed in his/her theatre gown or fresh nightwear. As near as possible to the time of surgery, male patients had leg, arm, and chest hair clipped where necessary by a porter. Female patients had leg hair clipped where necessary [10]. Where patient was unable to shower because of clinical limitations such as unstable angina, a healthcare worker applied chlorhexidine 4% aqueous solution to the chest (chin to umbilicus), arms, and legs using sterile pads. This was then allowed to dry on the skin. This process remained constant throughout the study period. Patients with a known history of diabetes mellitus or noted to by hyperglycaemic based on fasting glucose concentrations or HbA1c concentrations preoperatively were reviewed by a consultant endocrinologist and placed on an appropriate regimen to ensure pre-operative normoglycaemia.
Surgical technique
There was no change in cardiac surgeons or cardiac surgical fellows during this 2-y period. An educational session for the surgical team regarding the implementation of the change in surgical preparation was undertaken at the time of the change to CHG in 2011. This involved weekly teaching sessions conducted by the infection control nurse (A.H.) and involved an outline of the optimum painting and draping techniques using CHG over a 2-mo period. Meticulous attention to detail was emphasized to the surgical fellows.
Following the induction of anesthesia, the patient was positioned and API (2010) or CHG (2011) applied to the anterior neck, chest, abdomen, groin, and circumferentially around the lower limbs including feet. In the case of API, this was applied using sterile sponges, which were discarded frequently and when moving to an unpainted area of skin. All painted areas were allowed to dry fully. Two percent chlorhexidine gluconate and 70% isopropyl alcohol (Chlorprep) was applied using pre-impregnated sponges supplied by manufacturer and allowed to dry completely for a minimum of 3 min. Sterile surgical drapes were placed under the patients' legs while they were elevated by an assistant grasping the ankles with a sterile pad. Adhesive strips were placed along the lateral abdomen and thorax. In the case of a radial graft harvest for coronary artery bypass grafting, the chosen arm was painted also and laid on top of a sterile drape covering an armboard. For cases where a radial harvest was not required, both arms were fully covered from the shoulder and placed adjacent to the patient's thorax. An integrated genital cover was applied on top of a sterile towel and sealed with an adhesive strip horizontally just above the level of the umbilicus. When the posterior aspect of the lower limbs was fully painted, foot covers were applied in a sterile fashion and the legs lowered to the barrier drape overlying the operating table. A separate thoracic drape was applied with a clear window extending from above the supraclavicular notch to below the xiphisternum. The clear adhesive film was securely applied to the skin and the remaining part of the drape was unfolded both laterally and toward the head and feet. The upper segment was secured vertically to create a barrier separating the upper part of the table from the operative field.
In CABG, the long saphenous vein was exposed using a continuous longitudinal incision. No endoscopic vein harvests were performed. The adventitial layer was stripped and side branches ligated using ligaclips. Tissue was closed in layers using 2-0 vicryl and skin was closed using 3-0 monocryl monofilament. Skin was cleaned with a 0.5% Chlorhexidine with 70% denatured ethanol soaked swab solution and a sterile mepore dressing was applied to the length of the surgical incision. Median sternotomy was carried out in the standard fashion from just below the sternal notch to the tip of the xiphoid process using a scalpel and electrocautery for pinpoint haemostasis. Electrocautery was used to divide the pectoral fascia and a sternal saw used to divide the sternum in the midline. Once the sternum was split, the two edges were retracted and bleeding from the sternal edge controlled with electrocautery and bone wax.
Where internal mammary artery harvest was undertaken, the left internal mammary artery (LIMA) was used and harvested in a pedicled fashion. No patient underwent bilateral internal mammary artery (BIMA) harvest. Cardiopulmonary bypass was instituted allowing CABG, valve replacement or CABG with concomitant valve replacement to proceed. All cases were performed using cardiopulmonary bypass. A combination of single and double Myo wires was used to reunite the sternum; three single wires in the manubrium, three Myo wires around the body of the sternum, and one to two single wires at the lower portion of the sternum depending on sternal length. The wound was closed in three layers; two deep layers with 1-0 Prolene and 3-0 Monocryl for skin closure. The wound was cleaned using a 0.5% Chlorhexidine with 70% denatured ethanol soaked swab and a sterile mepore dressing applied to the length of the incision. All dressings remained unchanged for 5 d except in emergency, and all patients were nursed on the same cardiothoracic ward following discharge from the intensive care unit and until discharge home. A tight glycemic control protocol was adhered to in the post-operative period using an intravenous insulin infusion in the intensive care unit (ICU). Any patient with a history of diabetes mellitus or evidence of newly diagnosed diabetes mellitus was transitioned from that to the appropriate oral hypoglycemic agents or insulin by a consultant endocrinologist.
Surgical Site Infection Surveillance
Surgical site surveillance had been established in the hospital in 1995 and was carried out by the infection prevention and control (IPC) team using a combination of laboratory based and ward liaison surveillance, reported quarterly and annually. There was no change to this process during the study period. Regarding pre-study SSI rates, the overall incidence in 2009 was 3.8%, with a sternal wound rate and radial/saphenous graft infection rate of 2.45% (n=10) and 2% (n=8) respectively. All CTS patients were reviewed in outpatient clinics by their operating surgeon approximately 6 wks post-surgery and any late infections were recorded at that time. Any patient re-admitted within 12 mo of their surgery (in the case of valve surgery) with infection were also included. The CDC definitions of SSI were used to define infections and all infections confirmed by the IPC Nurse and IPC doctor in consultation with the surgical team.
Statistical Analysis
A univariate analysis was conducted on patient baseline characteristics as well as associated SSI rate between January 2010 and December 2011. Pearson chi squared test was utilized for dichotomous variables and Wilcoxon signed-rank test for continuous variables. A logistic regression model was then fit to estimate the effect of CHG in the reduction of SSI in 2011, controlling for age, major co-morbidities (e.g., left ventricular ejection fraction (LVEF), history of acute myocardial infarction (MI), elevated creatinine, and diabetes mellitus), body mass index (BMI), and other known SSI risk factors, for example, Anesthetic Surgical Assessment (ASA) score, presence of active infection at the surgical site, and duration of surgery.
Results
Patient characteristics
A total of 844 patients were enrolled in the study; 364 were enrolled in the 2010 pre-intervention cohort and 480 in the 2011 post-intervention cohort. There was no substantial difference in baseline patient characteristics identified between the two cohorts (Table 1). The median (inter-quartile range) age was 68 (61 to 75 y) y versus 68 (61 to 74 y) y for patients in the pre-intervention versus post intervention cohort, respectively (p=0.29).
IQR=interquartile range; CABG=coronary artery bypass surgery; SSI=surgical site infection; ASA=American Society of Anesthesiology; LVEF=left ventricle ejection fraction; MI=myocardial infarction; NS=not significant.
Operative risk factors
Patients in the pre-intervention 2010 cohort had a higher frequency of urgent versus elective surgery (28 vs. 18.8% p=0.0016), a higher family history of heart disease (46.2 vs. 36.9% p=0.0072), incidence of cerebrovascular (7.4 vs. 3.1% p=0.0061), carotid (6 vs. 1% p<0.0001) and hepatic disease (7.7 vs. 1.9% p=0.0001). Patients in the post-intervention 2011 cohort had a higher incidence of major co-morbidities such as previous myocardial infarction (29.8 vs. 17.6% p=0.0001), pulmonary disease (7.5 vs. 0.8% p<0.0001), and asthma (8.3 vs. 0.8% p<0.0001) and a higher incidence of renal failure at the time of surgery (5.8 vs. 0.8% p<0.0001). There was no difference in in-hospital mortality rates between the two groups (1.3 vs. 1.4% NS).
Surgical site infection risk factors
Univariate analysis of risk factors for surgical site infection failed to reveal any statistically significant correlation with the incidence of SSI. Specifically, there was no association between age (p=0.26), baseline creatinine (p=0.81), previous MI (p=0.36), diabetes mellitus (p=0.12), reduced LVEF (p=0.37), ASA grade (p=0.65), incision to closure time (p=0.37), or BMI (p=0.68) and the occurrence of SSI identified (Table 2).
MI=myocardial infarction; LVEF=left ventricle ejection fraction; ASA=American Society of Anesthesiology; BMI=body mass index.
Unadjusted SSI rates
There was a substantial reduction in overall SSI rate in 2011 following staff education and the introduction of CHG. The overall unadjusted SSI rate was 4.67% versus 2.08% (p<0.05) for 2010 and 2011 respectively (Table 3). The most noticeable reduction in SSI was in the graft site 2.75% versus 0.21%, (p<0.03) for the year 2010 and 2011 respectively. Likewise, the most noticeable reduction in terms of the classification of infection was from in superficial infection rate from 3.57% versus 1.46% (p<0.1) for the year of 2010 and 2011 respectively.
Multivariable adjusted effect of 2% Chlorhexidine with alcohol and education on SSI rates
Using multivariable adjusted rates the estimated effect of CHG and education in reduction of SSI in 2011 was 63% (OR: 0.37, 95% CI: 0.17–0.83, p=0.016) compared with API in 2010, controlling for age, major co-morbidities, and known SSI risk factors (Table 4). An outline of the specific microorganisms isolated from respective wound sites is detailed (Table 5).
Figures calculated using multivariate logistic regression.
ASA=American Society of Anesthesiology; LVEF=Left ventricle ejection fraction; MI=Myocardial infarction; BMI=body mass index; SSI=Surgical site incision.
Creatinine >1.49, LVEF <25%, MI history, or diabetes mellitus, SSI; surgical site infection.
ASA >2, incision to closure time >250 min, or BMI >30.
Discussion
Surgical site infection in CTS contributes to significant morbidity and mortality post-operatively and the development of mediastinitis is a devastating complication of sternal wound infection [14]. Recently, implementation of multi-disciplinary prevention measures as a care bundle have been shown to reduce the incidence of deep sternal wound infections [15]. Such bundles encompass a multitude of measures at each point throughout the entire pre- to post-operative time course of a patient's stay and include screening for MRSA carriage, hair clipping, skin preparation, prophylactic antibiotics, post-operative glycemic control, and appropriate wound dressings [15]. Regular audit and ongoing assessment of clinical practice is crucial and must be performed in a multi-disciplinary fashion as part of any successful Infection Prevention and Control (IPC) program [16].
Our study demonstrated a statistically significant reduction in SSI rates for graft sites from 2.75% in our 2010 cohort, to 0.21% in our 2011 cohort. These were associated with a substantial decrease overall in the numbers of patients classified as superficial SSI infection from 3.57% in 2010 to 1.46% in 2011. The major difference between the two cohorts was the reduction in graft infection, many of which were caused by gram-negative microorganisms in 2010. There was no substantial change in the number of deep SSI in the two cohorts with two deep sternal wound infection identified in each cohort. Of note there were no cases of organ space infection or mediastinitis in either cohort.
Other infection control strategies investigated in cardiac surgery in recent years include the use of microbial sealants to immobilize antimicrobials missed by skin prep. To date, one study failed to demonstrate successful decline in SSI incidence resultant from their use [17]. However, a number of authors report to the contrary; Iyer et al. performed a randomized trial of Integuseal (Halyard, Alpharetta, Georgia), demonstrating a 2.1 versus 25.5% incidence of donor site infections when compared with conventionally prepared skin [18]. Similar results were obtained by Dohmen et al. although their study was conducted in a non-randomized fashion [19]. A total of 280 patients underwent standard skin preparation with 300 allocated to the microbial skin sealant group. The incidence of surgical site infection was lower in the sealant group (2.3 vs. 6.8% p=0.011). Interestingly, Falk-Brynhildsen et al. investigated further and noted that at the time of surgery patients in whom a skin sealant was used had almost no intraoperative bacterial presence on the skin, whereas a relatively high incidence of late wound infection developed, suggesting that the wound contamination occurred post-operatively [20]. One additional strategy has been the use of topical vancomycin and Lazar et al. demonstrate a reduction in superficial sternal wounds in patients receiving topical vancomycin in addition to tight glycemic control [21].
When the risk factors for the two cohort groups were compared, most of the baseline characteristics such as BMI, age, incision to closure time, gender, and previous cardiac operations were similar. However, on univariate analysis there were substantial differences in some clinical characteristics between the 2010 and 2011 cohorts. Interestingly, in the post-intervention group, the patients had higher ASA scores and higher incidence of diabetes mellitus, both of which are well recognized as increasing infection risk. This finding of substantially lower infection rates despite substantially higher risk factor for in the 2011 cohort adds further strength to the conclusions of this paper that the use of CHG and a period of education is a crucial intervention. It has been shown that education has a positive effect on retention of knowledge, attitudes and practices of staff with respect to hospital infection control measures [22]. Undoubtedly, focused education of registrars and fellows in skin preparation technique and drying time resulted in more attention to detail with regard to the process in 2011 versus 2010 [22]. However, many studies have indicated that education alone will not create sustained improvement [23]. Organisms cultured from sternal and donor site infections are in keeping with those reported by other series [24]. It should be noted that methicillin-resistant Staphylococcus aureus was cultured from three superficial donor site infections.
This study does have a number of limitations. This is a cohort study rather than a blinded, randomized study, which is a recognized limitation. Additionally, our data is gathered from a single center. Recent evidence suggests that there is substantial center-level variation in the incidence of infection rates following cardiac surgery with as much as 18.2% variability identified between centers, not explained by individual patient characteristics [25]. Thus, ideally our intervention would be tested in multiple centers and the effect assessed. It is not possible based on this study's design to segregate the benefit obtained from each element of the study; in other words, what impact education compared with the change in agent made the outcome results. Since completion of this study, we have further revised our institutional policy and now weight-dose prophylactic vancomycin administered pre-operatively, which was not our practice at the time of this study.
Conclusion
This study demonstrates combined effect of CHG surgical skin preparation and education on the reduction in SSI rates following cardiac surgery. The next logical step of investigation is to repeat this study in a multi-center study to assess reproducibility of the interventions and to do so in a fully randomized fashion. We have demonstrated the importance of a multi-factorial approach required to prevent post-operative wound infections.
Footnotes
Author Disclosure Statement
No competing financial interests exist.