| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
Interact CardioVasc Thorac Surg 2007;6:759-767. doi:10.1510/icvts.2007.162511 © 2007 European Association of Cardio-Thoracic Surgery
Nosocomial infections after off-pump coronary artery bypass surgery: frequency, characteristics, and risk factors
a Alfa Institute of Biomedical Sciences (AIBS), Athens, Greece Received 2 July 2007; received in revised form 6 September 2007; accepted 7 September 2007
*Corresponding author. 9 Neapoleos Street, Marousi 151 23, Greece. Tel.: +30-694-6110000; fax: +30-210-6839605.
We evaluated the frequency, risk factors, and characteristics of infections in 360 patients after off-pump coronary artery bypass grafting (OPCABG). A prospective study was performed during the period June 2004–October 2005 at Henry Dunant Hospital, Athens, Greece. C-reactive protein (CRP) and procalcitonin were assayed from 222 patients preoperatively, and 1–3 days following OPCABG. Variables independently associated with infection were identified by a multivariable logistic regression model. Eighteen of 360 (5%) patients developed postoperative infections; 1.7% developed superficial wound infection, 1.4% pneumonia, 1.1% bacteremia, 0.3% mediastinitis, and 0.3% intra-aortic balloon pump related infection. The mean increase of CRP and procalcitonin levels in the first two or three days, respectively, after surgery was significantly higher (P<0.05) in patients with infection. Independent risk factors of infection (P<0.05) were history of major nervous system disorder, left ventricular heart failure preoperatively, emergent operation, transfusions of red blood cells during ICU stay, and duration of central venous catheter placement. The identification of risk factors for infection in combination with the appropriate evaluation of the increased CRP and procalcitonin values may help clinicians for the early diagnosis of infection after OPCABG.
Key Words: Postoperative complications; Mortality; Procalcitonin; C-reactive protein; Sternal wound cultures; Predictive value
In a previous prospective cohort study on microbiologically documented nosocomial infections, we found that postoperative infections occurred in 2.7% of patients that underwent off-pump coronary artery bypass grafting surgery (OPCABG) [1]. Risk factors independently associated with development of infection were arterial hypertension, previous vascular surgery, urgent operation, postoperative atrial fibrillation, number of inotropes used during operation and after operation, transfusions of fresh frozen plasma during ICU stay, and duration of ICU stay until development of infection. In an attempt to examine all postoperative infections (including non-microbiologically documented nosocomial infections) we performed this prospective study where cases were all patients with postoperative infection. Furthermore, we investigated the role of laboratory tests such as the role of sternal wound cultures in the diagnosis of sternal wound infection, and the role of C-reactive protein (CRP) and procalcitonin levels in the diagnosis of postoperative infections.
This prospective cohort study was performed at Henry Dunant Hospital, a tertiary care hospital in Athens, Greece. Approval was received from the Institutional Review Board of the hospital. We studied all adult patients who underwent OPCABG during the period June 2004–October 2005. Patients who underwent valve surgery combined with coronary artery bypass surgery were excluded from the study. Cases were patients who developed postoperative infection.
All patients underwent OPCABG using the An investigator prospectively collected data from all patients who underwent OPCABG. Variables that were considered risk factors for infections were entered into a database. We recorded all infectious complications during patients' in-hospital stay. Patients were followed-up for a 6-month period. Definitions of infections were based on the guidelines published from the Centers for Disease Control and Prevention [3]. Bacteremia was diagnosed in the presence of clinical signs or symptoms: fever >38 °C, chills or hypotension and at least one positive blood culture drawn through a peripheral line. The absence of a known focus of infection in another site was necessary for the diagnosis. Sepsis was defined as a systemic inflammatory response syndrome (SIRS) in the presence of infection [4]. The primary outcome of the study was the incidence of postoperative infections. Cure of infection was defined as improvement and finally resolution of presenting symptoms and signs by the end of treatment and discharge from the hospital; unresponsiveness was defined as persistence or worsening of presenting symptoms and/or signs of the infection despite treatment. Secondary outcome measured was in-hospital mortality.
Continuous variables were compared by the Student's t-test or the Mann–Whitney test, and categorical variables by
Out of 360 cases undergoing OPCABG surgery, 18 patients (5%) developed postoperative infections. Fifty-six patients were excluded from the analysis, because they underwent combined valve surgery. The characteristics of nosocomial infections are shown in Table 1. Seven of 18 patients (1.9%) developed sternotomy wound infection [one patient (0.3%) developed mediastinitis, and 6 (1.7%) superficial wound infection]. Five of 18 patients (1.4%) developed pneumonia, 4/18 (1.1%) bacteremia, 1/18 (0.3%) intra-aortic balloon pump related infection, and afterwards candidemia, and 1/18 (0.3%) pressure sore infection. Three patients had infections from two different sites. Thus, there were 21 episodes of infections in 18 patients. Four of 21 (19.1%) episodes of infections were non-microbiologically documented.
Three hundred and seventy-seven sternal wound cultures were received from 359 patients. Thirteen patients had more than one culture. Ninety-five microorganisms were isolated from 80 out of 377 (21.2%) sternal wound cultures, while no microbes were isolated from the rest 297 (78.8%). The isolated organisms are presented in Table 2. Gram-negative microbes were less common isolated than gram-positive cocci but had greater positive predictive value in sternal wound infections (5/18, 27.8%) compared with gram-positive cocci (7/74, 9.5%), although this difference was not significant (P=0.054, Fisher exact test).
C-reactive protein and procalcitonin levels measured the day before surgery, the first, the second, and the third day after surgery are presented in Table 3. There was not a significant difference in CRP and procalcitonin levels between patients with infection and patients without infection the day before surgery. CRP levels increased significantly from the day before surgery to day 1, 2, and 3 after surgery in both groups of patients. However, CRP levels measured one or two days after OPCABG were significantly higher in patients with postoperative infection compared to patients without infection (P<0.05). CRP levels measured on the third day after surgery were not significantly different between the two groups. Procalcitonin levels increased significantly from the day before surgery to day 1 and day 2 after surgery in both groups of patients (comparison of means of procalcitonin levels in each group of patients, P<0.05). Furthermore, procalcitonin levels measured the first, the second, and the third day after OPCABG were significantly higher in patients with post-operative infection compared to patients without infection (P<0.05).
Variables that were considered risk factors for infections were compared between the two groups of patients (Table 4). It is noteworthy that bivariable analysis showed that kind of antibiotic prophylaxis was not associated with the development of postoperative infection. Furthermore, we did not find significant difference when we compared the rate of patients who developed infection and received cefuroxime after operation and the rate of patients who developed infection and received a combination of teicoplanin and cefuroxime (13/224 vs. 3/112, P=0.28, Fisher exact test).
Transfusion rates of all studied patients were 40% in ICU vs. 28% in operation room. In addition, about 65% of studied patients did not require inotropes during their hospitalisation, while one inotropic agent was used in 21% of patients, two agents in 9% of patients, three agents in 4% and four agents in 1%. The mean duration of the use of inotrops was 23 h. The backward stepwise multivariable logistic regression model revealed that independent risk factors of nosocomial infections were a history of major nervous system disorder (nervous system condition or disease that severe affecting ambulation or day-to-day functioning), history of left ventricular heart failure, an emergent operation, transfusions of red blood cells (RBC) during ICU stay, and duration of central venous catheter placement (days) during ICU and ward hospitalisation (Table 5).
Clinical response of the infection was observed in 18/21 episodes (85.7%). Unresponsiveness was observed in 3/21 (14.3%) episodes in two patients. Thus, mortality was 2/18 (11.1%) for patients with infection, and 2/242 (0.6%) for the patients without infection (P<0.05). Although infection was statistically associated with mortality in the bivariable analysis (Table 6), the multivariable analysis revealed that independent risk factors for death were: history of major nervous system disorder (OR=13.3, 95% CI 1.40–127.07), and perioperative use of inotrops (OR=2.11, 95% CI 1.01–4.40). Two patients were admitted to hospital during a 6-month follow-up in order to perform plastic reconstruction of the infected sites. Mortality for patients who developed postoperative infection did not change in a 6-month follow-up. Mortality in the group of patients without infection was 6/242 (2.5%).
The frequency of nosocomial infections (5%) in this cohort study was higher than the frequency of microbiologically documented infections that was identified in a previous study of our group (2.7%) [1]. Among risk factors of infection of the two studies only emergent operation was an independent risk factor in both studies. From the rest independent risk factors of infection, history of left ventricular heart failure, transfusions of RBC during ICU stay, duration of central venous catheter placement, history of peripheral arteriopathy, and transfusion of fresh frozen plasma during ICU stay were found to be associated with the development of infection in the bivariable analyses of both studies but were not independently associated with infection in both multivariable analyses. History of major nervous system disorder, arterial hypertension, postoperative atrial fibrillation, and ICU stay until development of infection were risk factors associated with postoperative infection only in one of the two studies. The differences in risk factors for infection could be explained with the different methodology between the two studies, i.e. differences in data collection, definition of cases, and differences of patients' comorbidity between the two cohorts. Specifically, history of major nervous system disorder was remarkably frequent among patients with postoperative infection in this study, while history of arterial hypertension was more frequent among patients with postoperative infection in the previous study. Heart failure, prolonged use of central venous catheters and inotropes were associated with the development of nosocomial infection in previous studies [5, 6]. In addition, emergent surgery is often associated with unstable hemodynamic status and also requires prolonged placement of central venous catheters and prolonged use of inotropes. On the contrary, patients undergoing elective CABG surgery associated with early tracheal extubation resulted in shorter ICU and hospital length of stay with less postoperative infectious complications [7]. Furthermore, the association of RBC transfusion with infection was dose dependent [8]. In our study, patients who developed infection received a higher number of RBC transfusions (per patient) in ICU than in the operation room (Table 4). In addition, transfusion rates of all studied patients were greater during ICU stay than during cardiac surgery. Thus, the immunomodulation from transfusions during ICU stay was greater than in the operation room. Although the combination of teicoplanin and ceftazidime decreased postoperative infections compared to cefuroxime, that decrease was not significant. Thus, the recommended antimicrobials for cardiothoracic operations include cefazoline or cefuroxime that should be administered up to 24–72 h [9, 10]. However, vancomycin could be used in hospitals in which multi-resistant gram-positive pathogens are frequent causes of postoperative wound infection, and for patients previously colonized from multi-resistant isolates. Previous studies had shown that procalcitonin is not a specific marker for infection. High concentrations of procalcitonin were found in patients with three or more criteria for the diagnosis of SIRS, postoperative pulmonary dysfunction, myocardial infarction and cardiogenic shock [11–15]. The dynamics of procalcitonin levels, rather than absolute values, were more important for identifying patients with infections after cardiac surgery. Procalcitonin was found to be superior to CRP in discriminating infection after cardiac surgery. Our study is not without limitations. First, we did not have a sufficient number of patients that underwent on-pump CABG surgery in our institution in order to perform a statistical comparison between the two techniques. Second, although we studied a relatively large group of patients, the small number of patients with the endpoint of interest increases the probability of not finding a true association between a variable and the outcome of interest. In addition, postoperative CRP and procalcitonin values were missing for a considerable number of patients, especially those that were obtained 72 h after OPCABG. In conclusion, patients with left ventricular heart failure and movement disabilities that undergo an emergent operation, receive a high number of blood transfusions, and need prolonged central venous catheter placement, are in high risk of developing infection after OPCABG. Procalcitonin levels (and less CRP levels) increased markedly after postoperative infections. Sternal wound cultures should be taken only in cases of suspected infection because of the high proportion of negative results and the frequent isolations of microbes which are common skin contaminants. Cefuroxime remains an effective antibiotic prophylaxis regimen and thus broad-spectrum antibiotics should be avoided.
We would like to thank Mr Alexis Spanos and Mr Evangelos Issaris for their assistance with the statistical analysis.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
Top
Abstract
1. Introduction
-graft technique when indicated 
8 g/dl or in case of severe hemorrhage. Central venous catheters were changed every seven days. Samples of serum for assay of CRP and procalcitonin were obtained from a subgroup of 222 patients preoperatively, 24, 48, and 72 h following OPCABG. 
2 or Fisher exact test. The variables that were found to be associated with the studied outcome in the bivariable analysis (P