| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2003;126:992-999
© 2003 The American Association for Thoracic Surgery
Surgery for acquired cardiovascular disease |
a Division of Infectious Diseases, Department of Internal Medicine, Washington University School of Medicine, St Louis, Mo, USA
b Department of Surgery, Mayo Clinic, Rochester, Minn, USA
c Department of Surgery, Washington University School of Medicine, St Louis, Mo, USA
d Department of Infection Control, Barnes-Jewish Hospital, St Louis, Mo, USA
Received for publication September 19, 2002; revisions received October 25, 2002; accepted for publication December 27, 2002.
* Address for reprints: Margaret A. Olsen, MPH, PhD, Washington University School of Medicine, Division of Infectious Diseases, 660 South Euclid Ave, Campus Box 8051, St Louis, MO 63110-1093, USA
molsen{at}im.wustl.edu
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionSummaryReferences |
|---|
METHODS: We retrospectively analyzed data collected from 1980 coronary artery bypass patients undergoing surgery at our institution from January 1, 1996, through June 30, 1999, using The Society of Thoracic Surgeons database. Independent risk factors for leg harvest site infection were identified by multivariate logistic regression.
RESULTS: Seventy-six patients (4.5%) were coded as having had a leg harvest site infection, of which 67 were confirmed by infection control. The length of hospital stay after surgery was significantly longer in patients with leg harvest site infection (mean 10.1 days) compared with that of patients without infection (mean 7.1 days, P < .001), and infected patients were more likely to be readmitted to the hospital within 30 days of surgery. Independent risk factors for leg harvest site infection included previous cerebrovascular accident (odds ratio, 2.9), postoperative transfusion of 5 units or more of red blood cells (odds ratio, 2.8), obesity (odds ratio, 2.5), age 75 years or older (odds ratio, 1.9), and female gender (odds ratio, 1.8).
CONCLUSIONS: Consistent with previous studies, female gender and obesity were identified as independent risk factors for leg harvest site infection, while previous cerebrovascular accident, postoperative transfusion, and older age are newly described risk factors. The Society of Thoracic Surgeons database is a useful tool for identification of predictors of leg harvest site infections.
Few studies have been published that have included an analysis of risk factors specifically for harvest site infection in CABG patients. The majority of the literature concerning SSI in CABG patients has focused either on sternal wound infections or has involved an analysis of all wound infections occurring after surgery (combining chest and harvest sites together). To identify patients at increased risk of these infections and to devise strategies to reduce their risk, it is important to determine specific risk factors for leg harvest site infections. This may be of particular relevance as new, less invasive but more expensive technologies are developed for conduit harvest.
Risk factors for leg harvest site infection that have been identified in previous studies include obesity,6-8 female gender,7,9 diabetes,7 wound depth,10 continuous open incision method of harvesting the saphenous vein(s),7,11 and wound length.12 Numerous studies have reported risk factors for leg wound complications or impaired wound healing that presumably correlate with risk of leg harvest site infection. The variety of risk factors for leg wound complications or impaired healing reported include female gender,5,13,14 diabetes,13,15 obesity,13,15 peripheral vascular disease,13,14 left ventricular end-diastolic pressure greater than 15 mm Hg,13 longer crossclamp time,16 postoperative placement of an intra-aortic balloon pump,14 postoperative treatment with nicardipine,17 lack of postoperative treatment with dobutamine or nitroglycerin,17 leg incision open more than 150 minutes,17 and the traditional open method of saphenous vein harvest (compared with endoscopic harvest).15,18,19 It is difficult to interpret the relevance of some of these reported risk factors for leg wound complications due to the heterogeneous nature of the wound complications included in the published studies.
We used our institutional Society of Thoracic Surgeons (STS) Adult Cardiac Surgery database to analyze the influence of various potential risk factors on the incidence of leg harvest site infection. The national STS database was established for the purpose of risk stratification for mortality after cardiac surgery, and thus the variables available for analysis are those related to the primary outcome of mortality, including a variety of underlying medical, operative, and postoperative conditions and processes of care. We used the variables in this database to determine independent risk factors for leg harvest site infections in patients undergoing CABG surgery, and to determine the effect of these infections on duration of hospital stay and mortality.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionSummaryReferences |
|---|
Risk factor analyses were performed using the total number of leg harvest site infections in the STS database and separately for confirmed leg infections meeting the National Nosocomial Infections Surveillance (NNIS) criteria with symptoms occurring within 30 days of surgery. Deep leg harvest site infections involved the fascial or muscle layers of the incision and must have been accompanied by purulent drainage, spontaneous dehiscence, or intentional opening by a surgeon in a patient with fever, local pain, or tenderness, abscess or other evidence of deep infection, or diagnosis of a deep leg harvest site infection by a physician. Superficial leg harvest site infections involved only the skin or subcutaneous layers of the incision, and must have included purulent drainage, isolation of microorganisms from an aseptically obtained culture, diagnosis of a superficial leg harvest site infection by a physician, or at least one of the following: pain, tenderness, edema in the area of the incision, erythema, or intentional opening of the incision by the surgeon. Leg infections were defined by STS during the time period of data collection (January 1996 through June 1999) as infections involving a leg vein harvest site, including at least one of the following: (1) wound opened with excision of tissue, (2) positive wound culture, and/or (3) treatment with antibiotics. Postdischarge surveillance on all patients was performed 30 days after surgery to identify leg infections that occurred in the outpatient setting by telephone contact with both the cardiothoracic surgeon and the referring physician's offices.
Potential risk factors for leg harvest site infection analyzed included the following: (1) history of diabetes (including method of treatment), (2) congestive heart failure (including severity and previous myocardial infarction), (3) renal failure, (4) hypertension, (5) hypercholesterolemia, (6) peripheral vascular disease, and (7) cerebrovascular accident (or transient ischemic attack).
Baseline characteristics tested included body mass index (BMI, weight in kilograms/(height in meters)2), smoking history, previous cardiac surgical and nonsurgical interventions, length of hospital stay before surgery, use of inotropic drugs before surgery, diseased number of vessels, and use of various preoperative medications. Operative or postoperative variables included the number and type of blood transfusions received before, during, and after surgery; use of internal thoracic artery grafts; use of intra-aortic balloon pumping; operative status; aortic crossclamp time; time on bypass; duration of antibiotic prophylaxis before incision; number of conduits; and mortality after surgery.
Since risk factors for leg harvest site infection may also be risk factors for other types of infections, control patients for all the univariate and multivariate analyses were restricted to those patients with no identified infection in the STS database within 30 days of surgery. Thus, patients with sepsis, pneumonia, urinary tract infection, or any other SSI (chest, radial artery, and intra-aortic balloon pump site) were excluded from the control population.
Statistical analyses were performed in SPSS 11.0 software (SPSS, Inc, Chicago, Ill). Univariate analyses were performed for all relevant categorical variables using contingency tables (
2 or Fisher exact test for variables with small expected cell numbers), and t tests or the Mann-Whitney U test for continuous variables. Multivariate analyses were performed by logistic regression in which all variables with P < .10 in the univariate analyses, or previously shown to be related to surgical site infection from the literature, were included in the initial models. All 2-way interaction factors involving categorical variables that were deemed relevant were tested in the models after selection of the independent risk factors (main effects) and were included in the final models only if they were significant (P < .05).20 The final models were judged to be a good fit for the data by the Hosmer and Lemeshow goodness-of-fit test.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionSummaryReferences |
|---|
Seventy-five patients had leg harvest site infections according to our STS database, with one additional patient identified by the hospital infection control specialist (total number of leg harvest site infections = 76). Sixty-seven leg harvest site infections were confirmed by the infection control specialists, resulting in an incidence of 4.5% confirmed infections and 5.1% total leg harvest site infections during the study period. No outbreaks of leg harvest site infection were recognized during the study period, and the yearly incidence did not vary significantly with time (range 4.8-7.1%, P = .755).
Eight of the 67 confirmed leg harvest site infections were characterized as deep incisional infections by the infection control specialist, 56 were superficial infections, and 3 were not characterized as to depth. The 9 harvest site infections that were not confirmed by the infection control specialists all occurred in the outpatient setting and were presumably superficial infections. Thirty-six of the total 75 leg harvest site infections (48%) were diagnosed and treated solely in an outpatient setting, while the remaining 39 patients were hospitalized for therapy of their infection. Positive bacterial cultures were obtained from only 25 patients, 12 of which contained mixed flora (3 or more species). Staphylococcus aureus was isolated in pure culture from 5 patients and was present in an additional 4 specimens with other bacteria. Gram-negative bacilli were identified in 11 of the 25 positive cultures.
The 8 patients with confirmed deep leg harvest site infections and the 3 patients with infections of unknown depth were all treated with intravenous antibiotics, whereas 26 of the patients with superficial leg harvest site infections were also treated with intravenous antibiotics (alone or in combination with oral antibiotics). Thirty of the patients with confirmed superficial leg harvest site infections were treated exclusively with oral antibiotics; 27 of the 30 patients were treated as outpatients, and 3 were treated with oral antibiotics in the hospital.
Risk factors identified by univariate analyses
For all statistical analyses, the potential risk factors for the 76 patients with leg harvest site infections were compared with the risk factors in patients with saphenous vein donor grafts who were free of recognized infection (as defined by the STS database) during the month after surgery (n = 1200). We excluded 227 patients with other postoperative infections from the control group, since risk factors for one type of infection may also be risk factors for leg harvest site infection.
Results of the univariate analyses for categorical risk factors found to be significantly associated with total leg harvest site infection are shown in Table 1, and the results for continuous variables are shown in Table 2. Significantly increased risk of leg harvest site infection was associated with age 75 years or older; female gender; obesity (BMI between 30 and 35), and morbid obesity (BMI > 35); a history of congestive heart failure, cerebrovascular accident, or insulin-dependent diabetes; circulatory arrest during surgery; and receipt of 5 to 6 units of packed red blood cells postoperatively (Table 1).
|
|
2 test). Although insulin-dependent diabetes was associated with only a marginally increased risk of leg harvest site infection (Table 1), a history of cerebrovascular accident in patients with insulin-dependent diabetes was associated with a greatly increased risk of leg harvest site infection by univariate analysis (odds ratio 8.1, 95% confidence interval: 3.7, 17.9, P < .001). There was also a significant trend toward increasing risk of peripheral vascular disease in patients with both a history of previous cerebrovascular accident and insulin-dependent diabetes compared with patients with only one illness or neither previous cerebrovascular accident or insulin-dependent diabetes (P for linear trend < .001).
Independent risk factors of leg harvest site infection
Multivariate logistic regression was used to identify independent risk factors for leg harvest site infection. Circulatory arrest during surgery was associated with the greatest risk of total leg harvest site infection (adjusted odds ratio 3.9), although the confidence interval for this estimate is large due to the small number of infected patients with circulatory arrest (Table 3). Postoperative transfusion with 5 or more units of packed red blood cells, previous cerebrovascular accident, BMI greater than 30, age over 75 years, and female gender were additional independent risk factors for total leg harvest site infection.
|
Interaction effects were also tested in the multivariate models to determine whether combinations of risk factors conferred greater risk of leg harvest site infection that would be predicted from the additive effects of the individual risk factors alone. As described above in the univariate analysis, the interaction of previous cerebrovascular accident and insulin-dependent diabetes was associated with significantly increased risk of total leg harvest site infection. This interaction remained significant after controlling for the other independent risk factors of total leg harvest site infection (Table 4, P = .035 for interaction). The interaction did not remain significant in the model using confirmed leg harvest site infections as the outcome, however (P = .132 for interaction).
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionSummaryReferences |
|---|
Independent risk factors as predictors of leg harvest site infection
Postoperative transfusion with 5 or more units of packed red blood cells, previous cerebrovascular accident, obesity, older age, and female gender were identified as independent risk factors for leg harvest site infections in our study. A history of cerebrovascular accident has not been identified previously as a risk factor for leg harvest site infection in CABG patients, although peripheral vascular disease14 and arterial occlusive disease of the legs13 have been reported to increase the risk of leg wound complications. The increased risk of leg harvest site infections in women in this study supports previous findings7,9 and may be due to gender differences in fat distribution or to more impaired peripheral circulation in women than in men.
Postoperative transfusion with 5 or more units of packed red blood cells was also associated with increased risk of leg harvest site infection in this study. Transfused blood products have been reported as risk factors of deep chest surgical site infections in CABG patients,21-23 but have not been reported previously as risk factors for leg harvest site infection, to our knowledge.
Obesity was also independently associated with both confirmed and total leg harvest site infections, confirming previous reports.6-8 Obesity may increase the risk of leg infections by a number of different mechanisms, including poor penetration of prophylactic antibiotics into adipose tissue, increased likelihood of poorly controlled serum glucose (with or without documentation of a diagnosis of diabetes), increased colonization of the skin with bacteria with associated difficulty in skin antisepsis at the time of surgery and in the postoperative period, and impaired wound healing after surgery. Obesity may be a risk factor for superficial infections in general at the site of a surgical incision, as it also has been shown to be associated with increased risk of superficial chest infections after CABG8,21 and all surgical site infections in cancer patients.24
Preoperative aspirin therapy had a protective effect against confirmed leg harvest site infections in this study but not total infections. This may be a reflection of the protective effect of aspirin in more severe infections. The infections in the 9 patients that could not be verified by the infection control specialist were all presumably diagnosed in the outpatient setting and were, most likely, minor superficial incisional infections. This apparent protective effect of aspirin therapy for more severe leg harvest site infections requires independent confirmation.
The combination of insulin-dependent diabetes and a history of cerebrovascular accident was associated with greater risk of leg harvest site infections than would be predicted based on the additive effects of the individual risk factors. The greatly increased risk of leg harvest site infection in individuals with both insulin-dependent diabetes and previous cerebrovascular accident may be due to the increased incidence of peripheral vascular disease in this subpopulation of patients compared with patients with neither pre-existing illness or to some unmeasured covariate.
Effect of leg harvest site infection on morbidity and mortality
Patients with leg harvest site infections (total infections) stayed an average 3.0 days longer in the hospital after surgery than patients with no identified infections in the month after CABG. Although this is a crude estimate of the attributable increase in length of hospital stay during the surgical admission, it suggests increased morbidity in patients with harvest site infections compared to uninfected patients. This estimate does not take into account the additional hospital days incurred during readmission(s) for treatment of infections in the patients with leg harvest site infections, and thus the attributable increase in total hospital length of stay is greater than determined in this study. There was no increase in mortality attributed to leg harvest site infection in this study, consistent with the observation that the majority of infections were superficial.
Clinical implications
The ultimate goal of this type of epidemiologic study is to identify potential risk factor(s) and use this knowledge to design strategies to prevent postoperative infections. Although several of the risk factors identified in this study are not modifiable, such as female gender, older age, and history of cerebrovascular accident, strategies to reduce the risk of leg harvest site infections in these individuals can still be devised. Obese patients, particularly women with excessive fat distribution in the thighs, may benefit from higher doses of prophylactic antibiotics (ie, 2 g of cefazolin instead of 1 g), since antibiotic penetration in fat is poor.25 Older patients and those with a history of cerebrovascular accident discharged to home may need follow-up by home health care nurses to ensure adequate local care at the site of the leg incision(s).
Several studies have been published that report a lower rate of leg harvest site infections or wound complications in patients whose saphenous vein(s) were harvested by endoscopic rather than conventional methods.7,11,15,18,19 Although only one of these studies randomized patients to either harvest method,11 the apparent association of a traditional method of harvest using a long incision with increased risk of infection is consistent with the observation that risk correlates with wound length.12 We could not test the association of harvest method with leg harvest site infection since the time period of the study predated the introduction of endoscopic harvest in our institution. It will be important to confirm the benefit of endoscopic harvest in reducing the risk of leg harvest site infection, particularly in patients at high risk for infection as identified in this study.
Study limitations and strengths
The principle limitation of this study is the relatively small number of outcome events, although the number of leg harvest site infections in our study is still larger than in almost all other published reports. As with all studies of surgical site infections, the number of superficial harvest site infections is probably underestimated due to the difficulty of identifying particularly those infections diagnosed in the outpatient setting. An additional limitation is the lack of information in the STS database about some potentially important predictor variables, such as the location and length of the leg incision and the time the leg wound was open, that have been associated with leg wound infections or complications in single reports.12,17
Our study benefits from the confirmation of 90% of the leg harvest site infections using NNIS criteria to verify the identification of deep and superficial leg harvest site infections. It is particularly important to confirm leg infections with standardized accepted criteria since the majority of infections were superficial and could not be confirmed by growth of bacteria in cultures obtained from the leg wound. The primary strengths of this study are its use of the STS database as the source of data and its use of the multivariate methods for analysis. Our institutional database contains a large quantity of information on potential risk factors in addition to potentially relevant variables not included in the national database, such as the number of units of transfused blood products and the duration of antibiotic prophylaxis, making it very useful for identification of risk factors for a variety of postoperative infections.
|
Summary |
|---|
|
TopAbstractMethodsResultsDiscussionSummaryReferences |
|---|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionSummaryReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. K L Ho, M. P. Sun, T. W K Au, and C. S W Chiu Pneumatic Pump Reduces Leg Wound Complications in Cardiac Patients Asian Cardiovasc Thorac Ann, December 1, 2006; 14(6): 452 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. B. Allen Saphenectomy wound complications: the real story. J. Thorac. Cardiovasc. Surg., May 1, 2004; 127(5): 1535 - 1535. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
