The saphenous vein harvest wound is well recognized

Transcription

1 Occlusive Wra Dressing Reduces Infection Rate in Sahenous Vein Harvest Site Franklin L. Rosenfeldt, FRACS, Justin Negri, FRACS, Damien Holdaway, FRACS, Bruce B. Davis, FRACS, Julie Mack, BS, Michael J. Grigg, FRACS, Cambell Miles, FRACS, and Donald S. Esmore, FRACS Cardiac Surgical Research Unit, Deartments of Cardiothoracic Surgery and Vascular Surgery, Alfred Hosital, Melbourne, Victoria, Australia Background. Infection in the sahenous vein harvest site is a common roblem. We develoed an occlusive circumferential wra dressing technique that reduces skin edge tension, eliminates dead sace, and revents external contamination. We comared the surgical site infection rate using the wra dressing technique with that of standard longitudinal dressings. Methods. One hundred fifty-two consecutive atients were randomly assigned to receive either standard dressings or the wra dressing. Data were collected in the hosital and then 4 to 6 weeks ostoeratively. Suerficial and dee wound infections were defined by the standard criteria from the Centers for Disease Control and Prevention. Results. The infection rate in the wra grou was 14% comared with 35%, for the standard grou ( 0.006). Multivariate analysis showed that wra technique was the only significant redictor (negative) of infection (odds ratio, 0.19; 0.001). Conclusions. In sahenous vein harvest wounds, the occlusive wra dressing technique has the otential to reduce the rate of infection by 50%. This simle and inexensive technique is also readily alicable to the radial artery harvest site in the arm and may rovide similar benefit. (Ann Thorac Surg 2003;75:101 5) 2003 by The Society of Thoracic Surgeons The sahenous vein harvest wound is well recognized as a frequent site of infectious comlications [1]. Many studies have addressed methods of reducing the rate of infection, but most have focused on the method of closure [2, 3], the use of rohylactic antibiotics, or intraoerative wound lavage [4]. Few have focused on the management of the wound after closure as a method of reducing infection. Sahenous vein harvest wounds are generally longer than most surgical incisions. Patients undergoing cardiac revascularization include many with multile risk factors for wound infection, including obesity, diabetes, and occlusive arterial disease. When a long segment of vein is required, the incision is carried into the groin region where infection is more likely [5]. Often the harvest roduces an undermined fla in which the skin incision fails to arallel the variable ath of the sahenous vein. The weight of this fla laces tension across the wound, which may lead to dehiscence. In addition a dead sace may remain after wound closure, esecially in the thigh of an obese atient. This allows a collection of fluid to form, which redisoses to infection. We hyothesized that a broad, occlusive, olyurethane wra would suort, comress, and seal the wound and that if it were left intact for 10 days it would reduce the Acceted for ublication July 30, Address rerint requests to Dr Rosenfeldt, Deartment of Cardiothoracic Surgery, Alfred Hosital, Commercial Rd, Prahran 3181, Melbourne, Victoria, Australia; f.rosenfeldt@alfred.org.au. rate of infection. We comared the infection rate using the wra dressing with that observed with standard dressings. Material and Methods Between January and July 1998, 152 consecutive atients were enrolled in the study. As the study was art of a quality assurance rogram for wound infections and only dressing techniques already in use were comared, informed consent was not obtained from the atients. Patients were randomly assigned to either the occlusive wra dressing or a standard dressing by use of coded enveloes. Demograhics including body mass index, the resence of diabetes, tobacco use (resent or ast), renal imairment (serum creatinine above the normal range), and concurrent corticosteroid use were recorded. The following oerative data were recorded: byass time, the seniority of the surgeon who harvested the vein, whether a dee stitch was used, the distal extent and maximal deth of the wound, and whether a drain was used. Surgical Technique Antibiotic rohylaxis continued for 24 hours was cehazolin, 1 g on induction of anesthesia, then every 8 hours for 24 hours. If the atient had been in the hosital for 3 or more days, then vancomycin, 1 g intravenously, was administered on induction of anesthesia, then every 12 hours for 48 hours. The harvest sites were closed with 2003 by The Society of Thoracic Surgeons /03/$30.00 Published by Elsevier Science Inc PII S (02)

2 102 ROSENFELDT ET AL Ann Thorac Surg OCCLUSIVE DRESSING REDUCES INFECTION IN VEIN HARVEST SITE 2003;75:101 5 Table 1. Patient Demograhics Characteristic Wra Fig 1. Alication of adhesive dressing stris under tension osteriorly to anteriorly to suort the wound. Male Age (y) Hosital stay (median days) Body mass index a 26 (0.5) 26.9 (0.5) 0.23 Diabetes (%) Periheral vascular disease (%) Tobacco use (%) Renal imairment (%) Steroid use a Results are mean and standard error of the mean. subcuticular sutures. The choice of a deeer subcutaneous stitch was left to the surgeon s discretion. In the wra grou, dressing commenced with a vigorous wash with aqueous chlorhexidine-cetrimide solution, after which the leg was wraed in a series of 20-cm by 8-cm olyurethane adhesive stris (OSite dressing, Smith and Nehew, Clayton, Melbourne, Australia). The stris were alied firmly from osterior to anterior to suort the osterior wound edge beginning distally (Fig 1). The dressing covered only 60% to 70% of the circumference of the leg to revent distal swelling. Each stri had a 20% overla with the adjacent stri. The dressings were left intact for a minimum of 10 days (unless they were soiled or lost contact with the skin) and then usually removed by the atient after leaving the hosital. dressings consisted of a normal saline wash then either a dry nonstick cotton acrylic ad mounted on adhesive sterile aer (Primaore, Smith and Nehew, Clayton, Victoria, Australia), transarent, adhesive film (Tegaderm, 3M Health Care, St Paul, MN), or flexible, absorbent gel (Duo Derm, Bristol-Meyers Squibb Canada Inc, Montreal, Quebec, Canada). All these standard dressings were laced longitudinally and were suorted for 24 hours by cotton wool and a cree bandage. These dressings were removed shortly before the atient left the hosital. Patients were allowed to shower as soon as they were fit to do so regardless of the tye of dressing. The Centers for Disease Control and Prevention (CDC) definition of a suerficial wound infection was that it involved only the skin or subcutaneous tissue with at least one of the following: urulent discharge; an organism isolated on microbiological swab; or some sign of inflammation, ie,: tenderness, swelling, erythema, heat, or surgical intervention (the incision was reoened). Dee wound infection involved the dee tissues and at least one of the following: urulent discharge; abscess formation; sontaneous dehiscence; or surgical intervention. Finally, according to the CDC definition, the diagnosis of infection could be made at the surgeon s discretion. At follow-u the surgeon had no knowledge of the grou to which the atient had been assigned. Wounds were assessed until discharge and then in the outatient deartment after 4 to 6 weeks. Those who did not attend the outatient deartment were followed u with a combination of telehone interview and a questionnaire to their local medical ractitioner. All atients identified as having wound infection had their microbiological and medical records reviewed, and data were checked by the infection control deartment. Statistical Methods s given are mean standard error of the mean unless otherwise stated. The Student s t test was used for arametric data, the Wilcoxon test for nonarametric data, and the 2 and Fisher s exact test for categorical data. Infected and noninfected grous were comared to identify univariate redictors of infection. Significant univariate redictors were then entered into a multivariate, stewise, logistic regression analysis. This was initially done as a forward rocedure and then validated by a backward rocedure. Statistical significance was defined as a value less than Results Seventy-two atients had wra dressings and 78 had standard dressings. There were no significant demograhic differences between the two grous (Table 1). Oerative data were also similar (Table 2). Table 2. Oerative Data Wra Byass time (min) a Closure % Senior resident Junior resident No. of grafts Location (%) Groin Below groin Maximal wound deth (mm) a a Results are mean standard error of the mean.

3 Ann Thorac Surg ROSENFELDT ET AL 2003;75:101 5 OCCLUSIVE DRESSING REDUCES INFECTION IN VEIN HARVEST SITE Table 3. Infection Rate Follow-u was 99% comlete with 65% assessed in the outatient deartment and 35% by telehone survey. The infection rate in the wra grou (14%) was less than half that in the standard grou (35%; 0.006; Table 3). Of the total of 27 infections, only one (in the control grou) was detected in the hosital; the remainder were diagnosed after discharge from hosital. All infected atients received either oral or intravenous antibiotics. Three atients (2%) had infections serious enough to warrant further surgical rocedures, one being debridement and two requiring skin grafts. Two of these atients had received standard dressings and one the wra. In the wra grou 80% of infections were culture ositive. Of the causative organisms 50% were methicillin-resistant Stahylococcus aureus, 50% methicillinsensitive S aureus, and none were gram-negative. In the standard grou 85% of wounds were culture ositive: 39% grew methicillin-resistant S aureus, 22% methicillinsensitive S aureus, and 39% other organisms, redominantly gram-negative. Univariate analysis of infection both on demograhic and oerative data indicated wra, wound deth, and diabetes were significant redictors of infection (Table 4). Multivariate analysis showed the use of the wra was the only significant, indeendent, negative determinant of infection with an odds ratio of 0.19 ( 0.001; Table 5). Comment Wra Total infections 27 (35%) 10 (14%) Suerficial 22 (81%) 8 (80%) Dee 5 (19%) 2 (20%) 1.00 Purulent discharge 21 (27%) 8 (11%) (ACHCS definitions) Inflammation only 6 (8%) 2 (3%) 0.33 ACHCS Australian Council of Health Care s. This study showed a high incidence of leg wound infections in sahenous vein harvest sites according to the CDC definition and that the infection rate could be reduced substantially and gram-negative sesis eliminated by an occlusive wra dressing technique. This inexensive and novel dressing technique is now also Table 4. Univariate Predictors of Infection Noninfected (n 113) Infected (n 37) Wra used (n) Diabetes (%) Location (%) groin Wound deth (mm) a a Results are mean standard error of the mean. Table 5. Predictors of Infection by Multivariate Analysis Odds Ratio 95% CI Wra Wound deth Diabetes CI confidence interval. 103 routinely alied in our unit to the radial artery harvest site in the uer limb as well as the sahenous vein site. It is generally believed that different dressings have little effect on the outcome of a clean surgical wound. This is true when the wound is likely to heal rimarily because if the infection rate is low it is difficult to demonstrate any gains with one dressing tye over another. Consequently most reorts focus on other issues such as the ease of a dressing change or the number of dressings and the ain involved [6]. However in the case of sahenous vein harvest wounds, comarison of wound dressings may reveal differences in outcomes owing to the high rate of infections and other roblems secific to these wounds. One study similar to ours is that by Angelini and associates [7], comaring four tyes of skin closure: clis, nylon mattress sutures, subcuticular sutures, or sutureless adhesive dressing. In keeing with the roblematic nature of this wound, only 44% were well healed by the 10th ostoerative day. The adhesive dressing closure caused no difference in the rate of infection but achieved an imrovement in cosmetic result. Mechanism of Action of Wra Technique We ostulate that the wra works by a combination of three mechanisms. First, it suorts the wound by oosing distracting forces on the wound edges. These forces occur as a result of the gain in tissue volume after caillary leak as a result of the systemic inflammatory resonse triggered during cardioulmonary byass. The wra reduces tension by roviding broad-based suort for the skin edge to facilitate aosition and revent dehiscence (Fig 2). Second, the wra rovides comression. This reduces the accumulation of fluid in the subcutaneous lane. Third, the membrane seals the wound area and rotects the wound from secondary inoculation. The wound edges are often not sealed within 48 hours and are exosed to direct contamination, esecially when the roximal harvest sites were the inguinal area, a fertile site for colonization with gram-negative organisms. That no gram-negative infection occurred in the wra grou is in accordance with a sealing action of the wra. An imortant comonent of the wra technique was the vigorous washing of the wound edges with aqueous chlorhexidine and cetrimide before alying the dressing. This rocedure was designed to ensure that the entire surface underlying the occlusive dressing was sterile at the moment of alication of the dressing. Chlorhexidine sterilizes the wound surface, and cetrimide has a detergent action, which removes fat residues from the skin that can revent adhesion of the wra. An

4 104 ROSENFELDT ET AL Ann Thorac Surg OCCLUSIVE DRESSING REDUCES INFECTION IN VEIN HARVEST SITE 2003;75:101 5 infection or necrosis; and class 3, erigraft infection [1]. Other end oints used are any discharge (infective or not) or simly failure of the wound to heal comletely [8]. Our rate of infection aears unaccetably high, but we believe that this can be exlained in art by our use of the broad CDC definition of infection. To confirm this we reanalyzed our data using the narrower definition currently used by the Australian Council of Health Care s. This simler definition includes only wounds from which urulent material drains on or after the fifth ostoerative day and excludes a tissue reaction around the suture material, which could be included in the CDC definition. This reanalysis reduced the rate of infection from 14% to 11% in the wra grou and from 35% to 27% in the control grou. The difference in infection rate between the wra and the control grous remained significant ( 0.025). Fig 2. Mechanisms of action of the dressing: (left) aosition of wound edges suorting the suture line; (middle) comression of the tissue to eliminate dead sace and revent accumulation of fluid; (right) sealing of the wound area to revent secondary inoculation. additional benefit of the olyurethane stris over most other dressings is that they are transarent. This allows the wound to be checked for signs of exudation, bleeding, or dehiscence without removing the dressing. Critical Evaluation In the study by Mullen and colleagues [2], wound deth was a significant redictor of infection. Diabetes is the major risk factor in most studies of wound infection. In the resent study diabetes and wound deth were univariate redictors of infection but fell short of statistical significance on multivariate analysis. The infection rate for surgical rocedures is highly deendent on the definition of infection adoted. Published infection rates from sahenous vein harvest vary from 1% to 44% [1]. Infection rates are also affected by the length of follow-u. In the study by Johnson and coworkers [8], the eak incidence of wound infection occurred 4 weeks after oeration. Not all studies follow the atient until the wound has healed or even beyond the inhosital stay, which would roduce a suriously low infection rate. In a revious study from our institution, 58% of infections were identified after discharge [9]. In the resent study we found that 97% of infections were detected after the atient had been discharged from the hosital. Only 1 atient (a control grou atient), had an infection documented while in the hosital, ie, 3% of the combined grous. Different definitions of infection cause difficulties in comaring comlication rates among different institutions. In many cases local wound roblems such as skin edge necrosis, noninfective serous discharge, or simle local dehiscence may not be counted as infections. To overcome this some authors use descritive terms such as class 1, lymh leak, or skin edge necrosis; class 2, Economics of Infection The economic imact of infection and inconvenience for atients are robably greater than the clinical morbidity suggests. Only a minority of wound infections requires reoeration, and infection of the sahenous harvest site rarely causes death. However, the regular dressings and ongoing antibiotic use often means a long hosital stay and rotracted outatient medical care. Nelson and Dries [10] noted that wound infections including the sternum and the leg increased the average length of stay by 16.7 days and increased the cost er atient by $8,118. Conclusions The broad-based CDC definitions for wound infection and rolonged ostoerative surveillance can roduce a higher infection rate in sahenous vein harvest wounds than with more narrow definitions and inatient surveillance alone. The simle intervention of using an occlusive olyurethane wra can significantly reduce infection in the sahenous vein harvest site. This technique, which is readily alicable to the radial artery wound in the arm, should roduce similar benefits to those seen in the leg. We thank Carol M. and Ian G. Clarke for financial assistance and Smith and Nehew for their donation of OSite dressings for this study. We thank the oerating room nurses for their assistance. We acknowledge the statistical assistance of Michael Bailey in the Deartment of Eidemiology and Preventative Medicine, Monash University, Melbourne; and the advice of Denis Selman, MD, Deartment of Infectious Diseases, Alfred Hosital, Melbourne, Victoria, Australia. References 1. Reifsnyder T, Bandyk D, Seabrook G, Kinney E, Towne JB. Wound comlications of the in situ sahenous vein byass technique. J Vasc Surg 1992;15: Mullen JC, Bentley MJ, Mong K, et al. Reduction of leg wound infections following coronary artery byass surgery. Can J Cardiol 1999;15: Corder AP, Schache DJ, Farquharson SM, Tristram S. Wound infection following high sahenous ligation. A trial comaring two skin closure techniques: subcuticular oly-

5 Ann Thorac Surg ROSENFELDT ET AL 2003;75:101 5 OCCLUSIVE DRESSING REDUCES INFECTION IN VEIN HARVEST SITE glycolic acid and interruted monofilament nylon mattress sutures. J R Coll Surg Edinb 1991;36: Wong SW, Fernando D, Grant P. Leg wound infections associated with coronary revascularization. Aust N Z J Surg 1997;76: Lorentzen JE, Nielsen OM, Arendru H, et al. Vascular graft infection: an analysis of sixty two graft infections in 2411 consecutively lanted synthetic vascular grafts. Surgery 1985;98: Rasmussen H, Larsen MJ, Skeie E. Surgical wound dressing in outatient aediatric surgery: a randomised study. Danish Med Bull 1993;40: Angelini GD, Butchart EG, Armistead SH, Breckenridge IM. Comarative study of leg wound skin closure in coronary artery byass graft oerations. Thorax 1984;39: Johnson RG, Cohn WE, Thurer RL, McCarthy JR, Sirois CA, Weintraub RM. Cutaneous closure after cardiac oerations. Ann Surg 1997;226: Kent P, McDonald M, Harris O, Mason T, Selman D. Post-discharge surgical wound infection surveillance in a rovincial hosital: follow-u rates, validity of data and review of the literature. Aust N Z J Surg 2001;71: Nelson RM, Dries DJ. The economic imlications of infection in cardiac surgery. Ann Thorac Surg 1986;42: INVITED COMMENTARY The surgical rinciles of surgical wound management to minimize infection are well recognized and include strict adherence to sterile technique to avoid bacterial contamination; gentle surgical technique in handling tissue to avoid devitalization; wound closure without tension to avoid local ischemia; avoiding and treating imaired regional circulation; avoiding dead sace and hematoma; and minimizing the resence of foreign bodies. Rosenfeldt and associates describe a relative inexensive occlusive wra wound dressing to reduce the infection rate of the sahenous vein harvest site. Even though it was a limited study, there was a 60% reduction in infection rate that reached statistical significance. While it may be difficult to aly in the groin region, this dressing technique does adhere to some of the rinciles outlined above, namely reduce contamination, dead sace, and tension. Recent advances in the technique of endoscoic vein harvest have been shown to significantly reduce the rate of sahenous vein harvest wound comlications, including infection. The endoscoic technique aears to have the same advantage that Rosenfeldt and associates describe, but without the long incision. Stanley K. C. Tam, MD Harvard Medical School 300 Mount Auburn Street Suite 516 Cambridge, MA by The Society of Thoracic Surgeons /03/$30.00 Published by Elsevier Science Inc PII S (02)