| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Departments of Anesthesiology and Intensive Care Medicine, and Cardiovascular Surgery, University Hospital Charité, Campus Charité Mitte/Campus Virchow Klinikum, Charité Universitätsmedizin, Berlin, Germany; Department of Anesthesiology, St. Mary’s Medical Center, San Francisco, California
Address correspondence and reprint requests to Michael Sander, MD, Department of Anesthesiology and Intensive Care Medicine, Charité University Medicine Berlin, Charité Campus Mitte, Schumannstr. 20/21, 10117 Berlin, Germany. Address e-mail to michael.sander{at}charite.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
, interleukin (IL)-6, and IL-10. We investigated the predictive ability of these variables for infection after cardiac surgery. Forty-six patients with impaired left ventricular ejection fraction (<60%), scheduled for cardiac surgery, were included. Plasma samples were drawn 1 day before and immediately before surgery, on admission to the intensive care unit, and on days 1, 3, and 7 after surgery. Infection was identified according to the criteria of the Centers for Disease Control and Prevention. After surgery 13 patients developed an infection. In patients with infection, confirmed a median of 4 days after surgery, all measurements of IL-6, and IL-10 on postoperative day 3 were significantly increased. Tumor necrosis factor-, leukocytes, and C-reactive protein were not increased in these patients. Immediately after surgery blood glucose was significantly increased in patients with infection. Increased IL-6 after CPB is predictive of infection after cardiac surgery in patients with impaired left ventricular function. |
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
An inflammatory response after CPB is associated with systemic release of tumor necrosis factor (TNF)- and interleukin-6 (IL-6) (5). In septic patients IL-6 is induced by TNF- and is a surrogate for localized TNF- activity, which is more difficult to detect (6). In perioperative settings, IL-6 is a marker of tissue injury and destruction (7). A postoperative increase in C-reactive protein (CRP) is associated with an increase in IL-6. However, CRP levels may be increased even when the inflammatory stimulus has stopped (8). The inflammatory response to cardiac surgery is balanced by a phased antiinflammatory cytokine response, including IL-10 (9). IL-10 is a potent inhibitor of the production of TNF- and IL-6 (9).
The aim of this study was to investigate the predictive ability of perioperative plasma levels of TNF-, IL-6, and IL-10 for postoperative infection in patients as seen daily in specialized university cardiac surgical and cardiac anesthesiological departments.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Microbial screening was started on ICU admission. In the presence of clinical signs of infection, we obtained nose, throat, and wound swabs and cultures from tracheal aspirate or bronchoalveolar lavage. Infections were diagnosed according to the criteria of the Centers for Disease Control and Prevention (10). Diagnosis of pneumonia was made in the presence of systemic signs of infection, new or worsening infiltrates on the chest radiograph, and new onset of purulent sputum or a change in sputum with bacterial evidence (11). A superficial or deep surgical wound infection was diagnosed if it occurred within 30 days after surgery and if at least one of the following criteria was met: purulent drainage from the incision, microorganisms isolated from the incision, and at least one of the following signs or symptoms: pain, tenderness, swelling, redness, or heat. Central venous infection was diagnosed if the patient had clinical signs of infection and/or a positive blood culture was obtained before or within 48 h after removal of the central venous catheter and the catheter-tip culture had 
1000 colony-forming units (12).
Oral premedication was with midazolam 0.1 mg/kg. General anesthesia was induced with etomidate (0.2 mg/kg), fentanyl 5 µg/kg, and pancuronium 0.1 mg/kg. Maintenance of anesthesia was with infusion of fentanyl 5–10 µg · kg–1 · h–1, boluses of midazolam 0.1 mg/kg, pancuronium 0.03 mg/kg, and 0.6%–1% end-tidal isoflurane. All patients were ventilated with an oxygen-air mixture (Fio2 0.5) to maintain an end-tidal Pco2 of 35–45 mm Hg. After orotracheal intubation, a 4-lumen central venous catheter (Arrow, Reading, PA) was inserted into the right internal jugular vein.
After sternotomy, aprotinin was administered in a dose of 1.5 x 106 IU. Before CPB, heparin 350 U/kg and additional boluses of 50 U/kg were given to maintain an activated clotting time of at least 480 s. Normothermic CPB was performed using a membrane oxygenator and centrifugal pump adjusted to a cardiac index of 3 L/min/m2. Warm blood cardioplegia was used. The initial prednisolone loading dose was 1000 mg for each patient. Hemoconcentration and cell saving techniques were not applied in this study.
Transfusion management was performed according to our standard operating procedure, aiming at a hematocrit level >0.27 in patients with reduced oxygen transporting capacity.
We aimed at blood glucose levels of 80–180 mg/dL. Adjustments of insulin dose were based on measurements of whole-blood glucose in arterial blood, performed at 1-h to 4-h intervals with the use of a glucose analyzer (ABL700; Radiometer Medical, Copenhagen). In the case of increased measured blood glucose a continuous infusion of insulin was started and the infusion was adjusted to maintain a level within the selected range.
For perioperative antibiotic prophylaxis patients received 1.5 g cefotaxime 30 min before skin incision, 1.5g cefotaxime immediately after CPB and 1.5 g cefotaxime 6 h after ICU admission.
Durations of anesthesia, surgery, and aortic occlusion, and number of coronary artery bypass grafts were recorded. On ICU admission, the Acute Physiology and Chronic Health Evaluation (APACHE II) scores, and the use of inotropes and vasopressors were recorded. The durations of ICU and hospital stays were recorded. Transfusion requirements and the occurrence of SIRS, acute renal failure (ARF) (diagnosed by requirement of renal replacement therapy [RRT] or continuous IV loop diuretics) and death were recorded.
Laboratory tests including cytokines were drawn on the day before surgery, preoperatively on the day of surgery, on admission to the ICU, and on days 1, 3, and 7 after surgery. IL-6, IL-10, and TNF- were determined by enzyme-linked immunosorbent assay (ELISA; IEMA, Immunotech, Beckman Coulter Company, Marseille, France).
Data are expressed as median, 25th and 75th percentiles. Mann-Whitney U-test and Fisher’s exact test were used for intergroup differences. Dichotomous variables were examined with the 
2 test. All cytokine parameters with respect to time were analyzed using nonparametric multivariate analysis of variance (MANOVA) for repeated measurements, longitudinal data and small sample sizes in a two-factorial design (1st factor [group]: patients with infections versus patients without infections, 2nd factor [time]) (13). Therefore, we compared all 6 measurements simultaneously on the corresponding response curves. Predictive ability of laboratory tests was evaluated by receiver operating characteristics analysis. Statistical significance was accepted when P did not exceed 0.05. We used SAS 8.02 (SAS Institute Inc., Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Demographic characteristics such as age, gender, weight, and body mass index did not differ between the groups with and without infection (Table 1). Patients with infection had significantly lower preoperative LVEF (Table 1). There were no differences between the groups in durations of anesthesia, surgery, or aortic occlusion (Table 2). There were no differences in transfusion requirements (0 [0–0] U packed red blood cells versus 1 [0–2] units packed red blood cells; P = 0.14), and number of coronary artery bypass grafts (3 [2–3] versus 2 [1–3]; P = 0.26). Acute Physiology and Chronic Health Evaluation (APACHE) II scores did not differ between groups. The duration of mechanical ventilation, length of ICU stay, and the length of hospital stay were prolonged in patients with infections (Table 2). Patients with infections had an increased incidence of ARF (Table 2). Median onset of renal failure was on day 1 (interquartile range, 0–2).
|
|
IL-6 was significantly increased in patients with infection at all times of measurement (Fig. 1). IL-10 was significantly increased in patients with infection only on day 3 after surgery (Fig. 2). TNF- did not differ when comparing all times of measurement (Fig. 2). CRP and white blood cell (WBC) count (Fig. 3) did not differ between patients with and without infections.
|
|
|
Even though tight glucose control was a goal throughout the study period it could not be achieved for all patients. The aim for blood glucose control was not achieved in 79 (34%) measurements throughout the study period. Patients with postoperative infections displayed significantly higher levels of blood glucose on admission to the ICU. This increased level was predictive for the later onset of infection (Fig. 4). Thereafter, we did not observe differences between both groups (Fig. 4). Spearman correlation analysis demonstrated that increased IL-6 on admission to the ICU was independent from increased blood glucose (r = 0.24, P = 0.14).
|
Receiver operating characteristics demonstrated that conventional laboratory tests failed to predict an infection after surgery, whereas all measurements of IL-6 after surgery were predictive. The only value of IL-10 that was predictive of infection was on day 3 after surgery (Fig. 2).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Patients with postoperative infection had a significantly lower LVEF. This reduced cardiac function might have contributed to the risk of postoperative infection by immunological dysregulation in these patients. It has been proven that impaired cardiac function is associated with increased inflammatory cytokines such as IL-6 (15). IL-6 mRNA was increased in patients with cardiomyopathy undergoing cardiac transplantation (16). Human myocardium may produce IL-6 during myocardial infarction, ischemia, and reperfusion (17,18). Therefore compromised cardiac function might be a trigger for immunologic dysregulation increasing susceptibility to infection, and thus it is associated with an increase in IL-6. Moreover the surgical trauma itself may lead to increased IL-6 and to immunosuppression, resulting in a higher risk of infection. This has been reported in patients with myocardial ischemia undergoing noncardiac surgery (7).
The more frequent incidence of ARF is consistent with our finding of more severely compromised cardiac function in patients with infection. Patients with RRT show unchanged (19) or increased (20) clearance of cytokines, depending on the membrane used. As we observed, significantly more patients with ARF in the group with postoperative infections, the levels of IL-6 might have been slightly decreased after starting RRT in these patients. However, as infected patients had significantly increased level of IL-6 throughout the study, we believe that ARF and RRT only had a minor impact on the cytokines measured, if they had any at all.
Blood glucose levels were also significantly increased in patients with postoperative infections, even though tight glucose control was a study goal. However, this tight glucose control could not be achieved in 79 measurements throughout the study period (34%), which might have put some patients at a risk of contracting infections (21). Glucose levels are inherently difficult to control in this setting. At least in part, this might reflect the plethora of variables having an impact on levels of blood glucose, including feeding regimen, catecholamine and steroid administration, and stress response (22). We choose 80 to 180 mg/dL blood glucose as an aim for postoperative glucose control in cardiac surgical patients, as tighter glucose control might lead to severe problems with electrolyte disturbances and hypoglycemia, with the risk of even increased morbidity and mortality (22). Increased levels of blood glucose immediately after surgery were predictive for the later onset of infections. All other measurements did not differ between patients with and without infection. Even though blood glucose differed between patients with and without infection on admission to the ICU, this difference could not be observed afterwards. In contrast to this (and in contrast to conventional laboratory variables such as like CRP and WBC), IL-6 was predictive for postoperative infections throughout the study period. As the increase of IL-6 at admission to the ICU was not correlated with increased blood glucose levels at that time, different mechanisms might have been relevant for the immunological imbalance observed in the patients with postoperative infections. Nevertheless we do not discount that the increased blood glucose level on admission to the ICU might also have put these patients at a higher risk of contracting postoperative infections.
The site of infection was in most cases pulmonary. Thus patients with an infection were mechanically ventilated significantly longer than patients without infection. This was also associated with a significant increase in the length of ICU treatment, generating a significant amount of additional costs per patient with postoperative infection.
An inflammatory cytokine response after surgery is physiologic and is associated with an adequate host response to surgical trauma (23). In selected patients, a threefold increased rate of postoperative complications has been described if this physiologic inflammatory response is absent (23). However, an exaggerated systemic inflammatory response to surgical trauma can be harmful (24). Inflammatory response can lead to a compensatory antiinflammatory reaction. This may explain the significant increase in IL-10 on postoperative day 3 in our patients. This may have led to immune depression and made the patients susceptible to postoperative infection (25). Our finding of increased IL-6 after surgery is in accordance with previous studies (3,7). In our patients, IL-6 peaked immediately after surgery. Previous studies have also reported peak values of IL-6 within 4 hours after start of CPB (26). Increased IL-6 in patients with postoperative infection suggests more severe tissue trauma in these patients (7). This may cause derangement of the physiologic cytokine pattern, leaving the patient prone to postoperative infection.
TNF- did not vary in the immediate postoperative period and did not differ between patients with and without postoperative infection. Similarly, Mokart et al. (27) found no difference in TNF on days 1–3 after major surgery between septic patients and patients with uneventful postoperative course. However, the peak of TNF- might have been missed and therefore, we might not have been able to detect any differences regarding this parameter.
We have confirmed previous (28) findings that cardiac surgery with CPB leads to a postoperative increase in IL-10. Our new finding is that levels of IL-10 were significantly increased in patients developing postoperative infection. Increased antiinflammatory cytokine IL-10 has been described to be a predictor of an adverse outcome after surgery (29). Infection and sepsis especially correlate with increased IL-10 after surgery and trauma (30). A more pronounced antiinflammatory state seems to be hazardous in patients with inflammatory diseases (31). In septic patients increased levels of IL-10 were associated with increased morbidity and mortality (32).
As previously reported, CRP and WBC increased in all patients after surgery and were not different between the groups (2). Gabay and Kushner (8) found CRP to be increased even after the inflammatory stimulus stopped.
Even though tight glucose control was a goal throughout the study period, blood glucose levels were significantly increased in patients with postoperative infections immediately after surgery. However, this difference could not be observed afterwards. In contrast, IL-6 was predictive for postoperative infections throughout the study period. Nevertheless we cannot exclude that some patients were at a higher risk of subsequent onset of infection as a result of their increased blood glucose levels. The routine use of aprotinin and steroids interfered with the levels of IL, CRP, WBC, and other variables (26). However, as both groups received aprotinin and steroids it should not interfere with our findings. We measured only a few of the more than 140 known chemokines in only 46 patients. Therefore, we are not able to conclude whether any other marker might be more valuable to be predictive for the later onset of infections. However, this was not the aim of the study. We sought to determine an association between these chosen markers and postoperative infections, and therefore shed a light on alternative strategies to conventional laboratory markers to evaluate infection in this postoperative setting, such as CRP and WBC, which have been proven ineffective. Nevertheless we understand that our results need to be confirmed by studies with more patients.
Increased IL-6 after CPB is predictive of infection after cardiac surgery in patients with impaired left ventricular function. Increased preoperative IL-6 as a result of cardiac dysfunction identifies patients at higher risk of postoperative infection. Conventional laboratory variables failed to predict infection. However, we cannot conclude a cause-and-effect relationship, nor do we discount that the reduced ventricular function and the increased blood glucose level on admission to the ICU might also have put these patients at a higher risk of contracting postoperative infections.
The authors thank Mrs. Gerda Siebert, Dipl.-Math., Department of Medical Biometry, Charité University Medicine Berlin, Germany, for the detailed statistical advice for analyzing the data and Mrs. Sirka Sander for the diligent linguistic revision of this manuscript.
|
Footnotes |
|---|
Supported, in part, by departmental funding and institutional research grants of the Charité Medical School.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. L. Ford, I. M. Mains, E. J. Hilton, S. T. Reeves, R. E. Stroud, F. A. Crawford Jr, J. S. Ikonomidis, and F. G. Spinale Endothelin-A Receptor Inhibition After Cardiopulmonary Bypass: Cytokines and Receptor Activation Ann. Thorac. Surg., November 1, 2008; 86(5): 1576 - 1583. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Radaelli, C. Loardi, M. Cazzaniga, G. Balestri, C. DeCarlini, M. G. Cerrito, E. N. Cusa, L. Guerra, S. Garducci, D. Santo, et al. Inflammatory Activation During Coronary Artery Surgery and Its Dose-Dependent Modulation by Statin/ACE-Inhibitor Combination Arterioscler. Thromb. Vasc. Biol., December 1, 2007; 27(12): 2750 - 2755. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
