Prospective study if wound infections in dermatologic surgery in the absence of prophylactic antibiotics.
Anthony J. Dixon1, MBBS FACRRM
Mary P. Dixon2, B Appl Sci (Nursing)
Deborah A. Askew3, PhD
David Wilkinson4 MBChB, FRACGP, MRCP, DSc, PhD
1 Dermasurgeon, Senior Lecturer, School Of Medicine, University of Queensland
2 Nurse Manager, Skincanceronly, Geelong, Australia
3 Post Doctoral Research Fellow, School of Medicine, University of Queensland, Australia
4 Professor of Primary Care and Deputy Head, School of Medicine, University of Queensland, Australia
A prospective study of wound infections in dermatologic surgery in the absence of prophylactic antibiotics.
We undertook a 3 year prospective study of 5091 lesions, (predominantly non melanoma skin cancer) treated on 2424 patients from July 2002 to June 2005. No patient was given prophylactic antibiotics, and no patient ceased warfarin or aspirin. The study aimed to identify wound sites / procedures where infection incidence was over 5%, such that these circumstances may warrant wound infection antibiotic prophylaxis.
Overall infection incidence was 1.47%. Individual procedures had the following infection incidence: curettage 0.73% (3/412); skin flap repairs 2.94% (47/1601); simple excision and closure 0.54% (16/2974); skin grafts 8.70% (6/69); and wedge excision 8.57% (3/35).
Analysis of regions of the body demonstrated that surgery below the knee (n=448) had an infection incidence of 6.92% (31/448) (p<.0001). Sub-analysis demonstrated that all regions below the knee were at high infection risk. Elsewhere, groin excisional surgery had an infection incidence of 10% (1/10) (p=.027). No other body site demonstrated an infection incidence beyond 5% of statistical significance.
Procedures on the face demonstrated infection incidence of 0.81% (18/2209). Diabetic patients, those on warfarin and / or aspirin, and smokers showed no difference in infection incidence.
Based on a prediction of infection incidence over 5%, the following cutaneous oncologic procedures warrant consideration of oral antibiotic wound infection prophylaxis; all procedures below the knee, wedge excisions of lip and ear, all skin grafts, and lesions in the groin.
Other than under these circumstances, surgery to the nose, ear, fingers, lips, skin flap surgery, and surgery on diabetics, smokers, and those on anticoagulants have previously been considered for wound infection prophylaxis but do not warrant such intervention based on our data.
Low infection rates have been reported with the surgical excision of skin tumours, including Mohs’ surgery 50, 71 with incidence ranging from below 1 to over 4% following skin surgery to clean wounds 120, 125, 146.
In a comprehensive review of the subject, Maragh 121 and colleagues points out there are two indications for antibiotic prophylaxis in skin surgery: (i) wound infection prevention (therapeutic) and (ii) distant site infection prophylaxis, predominantly endocarditis and recent joint prosthetic surgery considerations. Maragh and colleagues identifies oral dicloxacillin and cephalexin as suitable choices for wound infection prophylaxis in selected cases of dermatologic surgery. Antibiotic wound infection prophylaxis for dermatologic surgery can be topical (including within the wound itself 147 ) and systemic. However, in a prospective randomized controlled trial, we have demonstrated a lack of efficacy of topical mupirocin in preventing wound infection in skin cancer surgery 2.
Usage of wound infection prophylaxis in dermatologic surgery varies greatly from surgeon to surgeon with some studies reporting widespread usage of antibiotics. 148 In general, antibiotics are not indicated for non-inflamed skin involving uncontaminated wounds. 120, 143, 149 Antibiotic prophylaxis has a greater role when the skin is inflamed or infected prior to surgery 121, the skin is contaminated 141, or if the patient is immunocompromised. 121
The location on the body is a factor in considering antibiotics. Extremities are believed to be at higher risk of wound infection.85 In particular the lower limb has been identified as at high risk of surgical wound infection.84 Other body locations reported to be at higher risk of wound infection following dermatologic surgery include the following: nose150, ear71, lip121, groin135, and axilla.151
Aspects of the patients’ health may also affect their propensity to wound infection. In particular, people with diabetes are considered to be at increased risk of wound infection. This has been demonstrated in diabetics who have lacerations 152, have cesarean sections153, cardiothoracic surgery139, 154, laryngectomy155, and knee replacement surgery.156 In contrast, diabetes was not shown to be a risk factor for wound infection in studies regarding bariatric surgery 157, leg vein harvesting 84, and lower extremity revascularization surgery 85. The dermatologic surgery wound infection incidence in diabetics compared with non diabetics is not well understood, yet diabetes is recognized as a possible consideration when considering wound infection prophylaxis 121. Smokers may also be at increased risk of wound infection such as has been identified in hip surgery 158, 159.
More data are needed to determine sites of the body and techniques of dermatologic surgery that place the patient at increased risk of wound infection. With this added information the clinician can make a more informed choice regarding situations whereby wound infection prophylaxis is appropriate.
We aimed to identify circumstances where skin cancer surgery (SCS) wound infection incidence of over 5% is predictable, such that in those situations surgery might have oral preoperative antibiotic administration considered. Specifically, through a three year prospective analysis of wound infections in a referral based skin cancer surgery centre, we aimed to identify body locations, dermatologic procedures and patient circumstances that, in the absence of antibiotic prophylaxis, place the patient at greater than 5% risk of wound infection.
This prospective study of wound infections in SCS without antibiotic prophylaxis was approved as part of a larger trial by the Barwon Health District Ethics Committee.
This study involved patients, and their lesions, managed in the 3 years from 1 July 2002 to 30 June 2005 at the Skincanceronly Centre, Geelong, Australia. Patients were not administered antibiotics either before or following SCS. In the event that wound infection developed, antibiotics were prescribed as appropriate. If a lesion was clinically infected at presentation, antibiotics were administered. Surgery was not undertaken until the lesion no longer had clinical features of infection and antibiotic therapy had ceased.
All patients presenting during the study period were considered for enrolment.
Inclusion criteria required that the patient:
Be referred to the Skincanceronly Centre for management of skin lesion(s).
Have a skin lesion for which incisional or excisional surgery or curettage is deemed appropriate in the patient’s management. Surgical procedures included: Mohs surgery; small and large excision and closure of lesions; curettage (with or without electrodessication); skin flaps; full thickness and partial- thickness skin grafts; and wedge excision surgery.
Exclusion criteria included:
Lesions managed entirely by cryotherapy.
Lesions treated with topical imiquimod or other non surgical means.
Lesions beyond 15 on any given patient were excluded to avoid over representation of some individuals.
Partial thickness and full thickness skin graft donor sites were excluded.
Patients requiring prophylactic antibiotics due to heart valve or recent prosthetic joint surgery.
One dermasurgeon (AD) performed all procedures in one of two operating rooms at either Skincanceronly Centre or Geelong Private Hospital. All surgery was undertaken using sterile surgical gloves, drapes and equipment. The dermasurgeon wore a surgically clean gown and a face mask. At Geelong Private Hospital the surgical gown was sterile and disposable hair cover was worn.
The site of all removed lesions was recorded and all specimens underwent histopathologic examination. All wounds were closed with nylon or polyamide interrupted skin sutures. Absorbable deep sutures were used in closure only if layers deep to the subcutis required direct closure. The absorbable sutures were either braided polyglycolic acid suture or monofilament poliglecaprone. Following surgery, all wounds had an application of an occlusive dressing unless site specific or patient’s allergy considerations made this impractical.
Because some patients had multiple tumors excised, the tumor deemed by the surgeon to be most urgent was removed first and a sub analysis of these lesions was carried out.
Patient details noted included warfarin and aspirin therapy, diabetes, and smoking status. In keeping with current practice 160, patients on medication affecting coagulation such as Warfarin or Aspirin did not have this medication ceased prior to surgery unless INR levels were excessive to therapeutic range.
Patients were followed up at least until removal of sutures. Patients were telephoned by nursing staff the day following surgery. Patients were strongly encouraged to return for removal of sutures. In the event that a patient attended elsewhere for removal of sutures, they were followed up by phone.
Any wound infection was recorded and classified in the following predetermined groupings: low grade infection with or without purulent suture site; suture abscess; large subcuticular abscess; infective necrosis; cellulitis; regional lymphadenitis; and septicemia. Wound infection was assessed clinically unless there was evidence of involvement beyond the local site. Under these circumstances, a wound swab for culture was taken. In the absence of suppuration, a wound was considered infected if three of the following features were present; discharge; pain; induration; and erythema. All wound infections were treated with oral dicloxacillin 500mg orally four times daily unless sensitivity, patient size or allergy deemed this to be inappropriate.
Infection incidence for each location / procedure undertaken was compared with all other locations / techniques using the chi square test. A site or method of SCS would be considered at high risk of wound infection if an existing incidence of 5% or more was reported.
A total of 5091 skin lesions from 2424 patients were treated by excision or curettage in the 3 year study period. The average age of patients was 57.5 years (median = 50 years) with 52.5% of patients being female. The 5091 lesions managed included 1263 squamous cell carcinoma(SCCs) (24.8%), 1194 basal cell carcinoma (BCCs) (23.5%), 132 melanoma (2.6%), 130 dysplastic melanocytic naevi (2.6%) and 2203 other benign lesions (43.3%). The benign lesions included 789 actinic keratoses, 110 cysts, 88 solar lentigenes, 63 angiomata, 26 lipomata, 31 warts and 280 benign non dysplastic pigmented naevi.
All patients were followed up at least until removal of sutures. Ten patients were not seen at or following removal of sutures and were followed up through telephone contact. All others were seen by surgeon and nursing staff in the follow-up period.
Two patients (three lesions) were excluded because antibiotic therapy was commenced prior to surgery. One patient had a heart valve prosthesis and another had been administered intravenous antibiotics by the anesthetist with the management of conscious sedation. One patient had more than 15 lesions and had all lesions after the 15th excluded from the study. No patient was excluded due to recent joint prosthetic surgery.
Out of the 5091 lesions, 75 wound infections were recorded (1.47%). There was no difference in the infection rate between men and women. Ten wounds were swabbed. One did not demonstrate a pathogen. The other nine demonstrated heavy growth of staphylococcus aureus. In eight of these cases, the bacteria were resistant to Penicillin but sensitive to dicloxacillin and cephalexin. At no stage were antibiotics adjusted based on culture results. Dicloxacillin was the antibiotic of choice other than for three Penicillin sensitive patients who were prescribed roxithromycin.
Two patients had serious infections. One patient had extensive cellulitis that did not respond to oral dicloxacillin. The infection later responded to oral ciprofloxacin. Cultures did not identify a pathogen. Another patient did not respond to oral dicloxacillin and was subsequently managed by another doctor. She required hospitalization, debridement and intravenous antibiotics. A partial thickness skin graft was applied. This failed to take and the defect on the shin healed by second intent over many months. These were the only two patients who did not respond to the initial empirical antibiotic selection.
Infection in relation to surgical technique and site on body:
Infection incidence from each surgical technique is summarized in Table 1, and at different sites on the body in Table 2. All sites below the knee demonstrated and increased infection incidence. In total, there were 31 infections below the knee, or 6.92% (p<.0001). The groin was the only other body location that demonstrated infection incidence over 5%.
We hypothesized that perhaps only the larger defects below the knee were experiencing a high infection incidence. A subanalysis revealed that 225 defects below the knee were over 11 mm in diameter. The incidence of infection in this subgroup was 11.6% with 26 wound infections identified. In contrast, 223 defects smaller than 11 mm in diameter below the knee demonstrated an infection incidence of 2.24% with five wound infections identified. (p = .0001).
The types of infection are detailed in Table 3. It was noteworthy that abscess formation was not a feature of all 31 infections below the knee in contrast to 22 of 44 infections above the knee (p < .0001). Low grade infections below the knee either developed flap necrosis or cellulitis rather than proceeded to abscess formation.
While the infection incidence on fingers was 4.88%, this was not significantly different to elsewhere. Only 41 finger lesions were managed in the 3 year period. Similarly, tumors treated with Mohs surgery demonstrated no significant difference in infection incidence. All three infected curettage wounds were on the shin from 35 lesions curetted below the knee (8.57%). A sub analysis of procedures undertaken in the difficult region; nose, ear, lip, hand and below knee is detailed in Table 4. Excluding wedge resection and graft operations, there were five infections in 346 operations on the ear and lip (1.48%).
The types of skin flaps (n=1601) undertaken in this 3 year period were as follows; 445 transposition (27.8%), 295 O - S (18.4%), 282 bilobed (17.6%), 175 reducing opposed multilobed (ROM)6 (10.9%), 93 twin advancement (5.8%), 92 V – Y (5.7%), 85 A – T (5.3%),and 134 others (8.4%).
“At risk” patients
Diabetes was reported by 135 patients (5.6%) who accounted for 285 lesions. There were three infections among these wounds (1.1%). This was not different from the non-diabetic population (p = 0.54). A total of 369 (15.2%) patients were on either warfarin (n = 64) or aspirin (n = 299) or both (n = 6). These patients had 799 operations and experienced 10 wound infections (1.25%), which did not differ from the general population of patients (p = .57). Patients taking warfarin had 162 operations with three wound infections (1.85%), those taking aspirin had 647 operations with seven wound infections (1.1%) and patients taking both aspirin and warfarin had 10 operations and no infections.
There was no difference in infection rates between smokers and non smokers (p = .24). A total of
296 (12.2%) patients declared themselves as smokers. These patients had 634 lesions managed, having six wound infections, (0.95%).
Other potential influencing factors
We recognized that as some patients had many procedures, they may be weighted excessively in the overall analysis. However, a sub analysis of the infection incidence of the first lesion on each patient treated was not different from the overall infection incidence demonstrated (not shown).
We also recognized that infection incidence may have been different between the two surgical venues studied. One hundred seventy-four tumors excised at Geelong Private Hospital resulted in five subsequent wound infections (2.9%) This was not dissimilar to the overall infection incidence identified, (p = 0.12).
This study aimed to identify circumstances where SCS wound infection rates of greater than 5% can be predicted and therefore, circumstances where prophylactic antibiotics may be indicated.
We identified all SCS procedures, including curettage, at any site below the knee had a high infection incidence. This was expected given the compromise in circulation that is recognized in these locations. Of note was the finding that when infection progressed below the knee, it developed into cellulitis or infective necrosis but not abscess. We postulate that this lack of abscesses is due to the tight skin below the knee and hence reduced dead space. The higher wound infection incidence below the knee is relevant, given the increased risk of wound dehiscence, delayed healing, flap necrosis and poor cosmesis that can occur in association with wound infection. These adverse outcomes can also result in a reduced confidence in the clinician and even legal action. Consequently, wound infection prophylaxis for below knee procedures should be considered.
Skin grafts and wedge excisions were the only types of procedures performed above the knee that were associated with a high infection incidence. This contributed to the higher rates of infection incidence on the lips and ears, compared with the rest of the body (with the exception of below knee procedures). Lip and ear closures that did not require a graft or wedge excision had the same infection incidence as elsewhere. It was noteworthy that diabetics did not suffer a higher incidence of wound infection than non diabetics, either above or below the knee.
Flap surgery has a significantly higher infection incidence than elliptical excision and closure. Invariably these are larger / more complicated procedures with skin tension and perfusion being considerations in explaining this finding. While skin flaps are designed to reduce skin wound tension (among other features), wound tension is still frequently higher than noted in small simple elliptical closures. The skin flap often “controls” wound tension across a large part of the repair in a large defect where tension would have been excessive centrally if closed by ellipse alone. An example of this controlling of wound tension is seen with the ROM flap below the knee previously reported by this author.6
This study has significant limitations. It involves a single dermasurgeon working in two surgical venues in one southern Australian regional city. This region has a temperate climate. The wound infection incidence reported may not reflect practice elsewhere, such as more tropical / humid climates. Wound infections were assessed clinically rather than by culture of wound swabs. A relatively short follow-up period on some cases may have led to an underreporting of late wound infections. The trial setting was that of a community based dermatologic surgery centre where wound swabs would not be routinely taken on wounds where infection is a concern. This mimics routine community dermasurgical practice in Australia. Other than warfarin and aspirin, other medications affecting coagulation such as non steroidal anti inflammatory medications were not recorded in this database.
Whenever the question of antibiotic prophylaxis is considered, disadvantages include the risk of antimicrobial resistance arising from such liberal administration. Further, antibiotics subject the patient to adverse reactions including allergic skin manifestations. A significant improvement in wound outcomes would be needed in future studies to justify the extra intervention of prophylactic antibiotics, given these significant reservations. This study has identified sites and procedures of high risk; it does not determine whether antibiotic wound infection prophylaxis will afford the patient benefit in these circumstances. While there is existing evidence that antibiotic prophylaxis will reduce the subsequent wound infection incidence, a randomized controlled trial specific to these high infection scenarios is needed to determine whether the sites / procedures identified improve sufficiently with such prophylaxis.
Our results suggest that a 5% or higher incidence of wound infection in the absence of antibiotic prophylaxis is more likely following;
all SCS procedures at any site below the knee, although small defects do not demonstrate the same infection rate as larger ones; wedge resections of lip or ear, all skin grafts; and lesions in the groin.
Therefore, oral antibiotics need to be considered prior to surgery in these circumstances.
While flap surgery has over five times the infection rate of elliptical excisions, the only locations where flap surgery suffers an infection rate above 5% are those below the knee. As such, wound infection prophylaxis for flap surgery above the knee is not indicated. Mohs’ surgery requires special consideration. The repeated surgery and intermittent wound packing at a single site results in wounds being clean-contaminated and are as such, at an increased infection risk. Furthermore, wound infection prophylaxis may need to be repeated in these cases due to the duration from the commencement to the conclusion of surgery in some cases.121
Patients who are smokers, diabetics and those on aspirin or warfarin medication do not have an infection incidence different from the general population, and as such need not be considered for wound infection prophylaxis, other than under the circumstances for the population detailed above.
When deciding about antibiotic wound infection prophylaxis, the clinician needs to involve the patient in the decision making process. Advice to the patient should include the likelihood of infection and other factors such as whether or not the wound is contaminated, whether the patient is immune-compromised, whether the patient is considered at increased risk of endocarditis, and whether the patient has had recent joint prosthesis surgery. To assist clinicians to give the best possible advice to patients about the efficacy of antibiotic wound infection prophylaxis, more research that is specific to the sites and circumstances indicated is needed to provide a comprehensive evidence base to support decision making in this area.