In vivo effect of sustained-release silver sulphadiazine foam on bioburden and wound closure in infected venous leg ulcers

Transcription

1 In vivo effect of sustained-release silver sulphadiazine foam on bioburden and wound closure in infected venous leg ulcers l Objective: To determine the in vivo effect of a sustained-release silver sulphadiazine powder foam dressing on the bacterial burden of venous leg ulcers (VLUs), with a view to correlating the wound closure rate with the degree of bioburden and to assess other markers of its progression towards healing. l Method: Patients attending a tertiary care wound practice were screened for a VLU bacterial count of >10 5 colony forming units (cfu/g) per gram of tissue. Patients were treated with the above topical dressing plus multilayer compression bandaging for 12 weeks. Quantitative cultures were taken at weeks 0, 2, 4 and 8. Wounds were assessed at each weekly visit using photography and planimetry. l Results: Twenty-four of the 33 screened patients met the inclusion criteria. The average baseline wound size was 12.3cm 2. At week 8, the bioburden had reduced to <10 5 cfu/g in 54.2% of patients, with 41.7% achieving this reduction by week 2. At week 8, the median reduction was 0.7log 10 (p<0.001). The median percentage reductions in bioburden were 50.5%, 56.8% and 80.4% at weeks 2, 4 and 8 respectively. Over the study period, 79.2% of the patients had >75% wound area reduction and 45.8% achieved ulcer closure in a median of 80.5 days. l Conclusion: Although the wound size and bacterial counts reduced significantly, there was no statistical correlation between the two. Nevertheless, the in vivo data show that this active antimicrobial dressing was associated with a very high healing rate in these hard-to-heal wounds. l Conflict of interest: This research was supported by a grant from Smith & Nephew, Hull, UK. In addition, the principal investigator receives other grants from Smith & Nephew, Hull, UK. hard-to-heal venous leg ulcers; bioburden; silver sulphadiazine; wound closure J.C. Lantis II, MD, Chief of Division, Vascular/ Endovascular Surgery, St Luke s Roosevelt Hospital, Associate Clinical Professor Surgery, Columbia University, New York, NY, USA; C. Gendics, RN, Senior Research Coordinator, Division of Vascular/ Endovascular Surgery, St Luke s Roosevelt Hospital, New York, NY, USA. JLantis@chpnet. org Most randomised controlled trials of venous leg ulcers (VLUs) report 12-week closure rates of approximately 30 35% for standard care, 1-3 (rising to 50% in most active therapy arms. 2,3 However, these trials usually have strict inclusion criteria, requiring that ulcers be clean (often defined as having a bioburden of less than 10 5 cfu/g), 4 have only been present for relatively short periods, often less than 1 2 years, 4 and be of a certain size. Very few VLUs fit such limited inclusion criteria: approximately one in eight patients at our centre would be eligible for inclusion in such trials. Some studies suggest that % of VLUs are clinically colonised with bacteria, which further decreases their ability to close. 5-7 A World Health Organization 6 trial found that 26% of VLUs visually assessed as clean actually have a bioburden greater than 10 5 cfu/g. 4 Because of this perceived correlation between bioburden and poor closure rates, many clinicians use topical antibacterial therapies in an attempt to control critically colonisation in VLUs Despite their widespread use, very little data is available on the in vivo efficacy of topical antibacterials on this wound type, with the focus instead being placed on wound closure rates. To date, the best studied of these topical antibacterial agents is cadhexomer iodine, which was found to significantly decrease the bioburden of Staphylococcus aureus in VLUs over 6 weeks. 11 However, of the 10 well-designed trials conducted on the use of cadhexomer iodine under compression in VLUs, only three have reported on bioburden Harcup et al. noted a 44% reduction in ulcer size for those treated with standard care versus a 66% reduction for those treated with cadhexomer iodine for 6 weeks. 12 Skog et al. reported a closure rate of 32% at 4 weeks for cadhexomer iodine versus 3% for the control, and a mean area reduction of 34% at 6 weeks (versus -5% for the control). 11 The more widely used topical silver dressings have even less evidence on their in vivo efficacy. Just three studies have been reported, with a total of 847 participants, examining three different, absorbent, sustained-release silver dressings. They did not find any significant reduction in bioburden with treatment (although this was not adequately assessed), or show improved wound healing rates. 14

2 We designed the current study to address the poverty of data on the in vivo efficacy of topical antibacterials. As we wanted our sample to better reflect the reality of VLU patients seen at our centre, we avoided the use of strict inclusion criteria that assessed low bioburden wounds of short duration and small size. We set out to assess the actual efficacy of silver sulphadiazine powder foam, a topical antibacterial, in chronically infected VLUs. Very little data exist on its efficacy in this wound type. Our goal was to follow the in vivo effect of the topical antibacterial on the bacterial burden (primary endpoint) and to track the closure rate (secondary endpoint). Materials and method We initiated a hospital-sponsored and Institutional Review Board-approved clinical protocol. The patients were recruited from the Division of Vascular/Endovascular Practice in an inner city tertiary care hospital. Inclusion criteria were: l VLUs that were at least 4cm 2 and at least 4 weeks in duration l VLUs with a biopsy-proven bioburden of 10 5 colony forming units per gram (cfu/g) of wound tissue l VLUs that were full-thickness but not extending to muscle or bone l Confirmation of venous insufficiency by ultrasound l An ankle brachial pressure index (ABPI) of 0.8 and <1.3 and, hence, suitability for compression l Concordance with compression therapy l One or more clinical sign(s) of infection (oedema, malodor, local/periwound erythema, spontaneous pain between dressing changes, increased exudate, discoloration of granulation tissue, increased temperature at wound, non-progression of wound closure, purulent exudate or friable granulation tissue). Patients were excluded if they had a sensitivity to silver sulphadiazine, polyurethane foams or other components of the test dressing, were undergoing chemotherapy, being treated with immunosuppressants or corticosteroids, had an autoimmune disease, or had participated in an experimental drug or device study within the last 30 days. The initial visit consisted of assessment of eligibility, the (written) informed consent process, two adjacent 3mm punch biopsies, confirmatory ultrasound studies and non-invasive flow studies. Initial laboratory values included analysis of glucose levels, renal insufficiency, anaemia, white blood cell count, albumin/total protein level and pregnancy. One week later, a non-adhesive silver sulphadiazine, polyurethane foam dressing (Allevyn Ag Non-Adhesive, Smith & Nephew, Hull, UK) was applied to the debrided wound. The leg was wrapped in a multilayer compression wrap (Profore, Smith & Nephew, Hull, UK). Patients whose biopsies showed <10 5 cfu/g were continued on the same therapy and included in the safety analysis, but were not included in the efficacy data analysis. Manual wound tracings (planimetry), digital photography and visual wound assessment were performed on a weekly basis for 12 weeks or until wound closure, together with the assessment of any adverse events. Wound closure was defined as complete epithelialisation and a lack of drainage for one week. The patients marked their pain level themselves on a validated visual analogue scale (0 100) and were graded using a 0 5 faces scale by a blinded observer; both were undertaken on a weekly basis. Any patient who interrupted the evaluation dressing regimen for longer than 7 days was withdrawn from the study. Patients who did not respond to therapy were withdrawn only when it became necessary to change their dressings on clinical grounds. Biopsies were also obtained at weeks 2, 4 and 8, and semi-quantitative swab cultures at weeks 2, 4, 8 and 12, to provide additional data on the wound surface microflora. If the ulcer was too small for biopsy, then the data from previous assessment date was used. All wounds were debrided to some degree at each weekly dressing change. Any additional dressing changes were recorded. Quantitative cultures were performed using tissue from the above biopsies a 0.1g sample was placed in a 30ml polypropylene container and stored at room temperature. Using sterile forceps or a sterile swab, it was removed from the container and ground in a sterile pestle and placed in a clean Petri dis7h; care was taken not to expose the tissue to long periods of air contamination. One ml of thioglycollate broth was then placed in the tube and the tissue was reground, then streaked onto the medium. In addition, both the thioglycollate and CMG broth were inoculated using the same loop. The plates were then incubated at 35 37ºC for 48 hours. Following incubation, the number of colonies that grew for each organism was counted and the cfu/g for that colony was calculated. If no growth was noted at 5 days, the specimen was considered sterile. Statistical analysis Statistical analysis was carried using the SAS software package 9.1 (Cary, NC, USA). Spearman rank correlation coefficients were calculated for risk factor analysis, as it related to healing, and to determine the significance of a correlation of bioburden to wound closure. Kaplan Mier life table analysis was used to assess mean and median times to achieve wound closure and bioburden reduction. Wilcoxian signed rank testing was used to assess the significance of bioburden reduction over time. McNemars test for difference was used to assess changes in primary and secondary endpoints over time. References 1 Vin, F., Teot, L., Meaume, S. The healing properties of Promogran in venous leg ulcers. J Wound Care 2002; 11: 9, Mostow, E.N., Haraway, G.D., Dalsing, M. et al. Effectiveness of an extracellular matrix graft (OASIS Wound Matrix) in the treatment of chronic leg ulcers: a randomized clinical trial. J Vasc Surg 2005; 41: 5, Falanga, V., Margolis, D., Alvarez, O. et al. Rapid healing of venous ulcers and lack of clinical rejection with an allogeneic cultured human skin equivalent. Arch Dermatol 1998; 134: 3, Robson, M.C., Phillips, T.J., Falanga, V. et al. Randomized trial of topically applied repifermin (recombinant human keratinocyte growth factor-2) to accelerate wound healing in venous ulcers. Wound Repair Regen 2001; 9: 5, Brook, I., Frazier, E.H. Aerobic and anaerobic microbiology of chronic venous leg ulcers. Int J Dermatol 1998; 37: 6, Halbert, A.R., Stacey, M.C., Rohr, J.B., Jopp-McKay, A. The effect of bacterial colonization on venous ulcer healing. Australas J Dermatol 1992; 33: 2, Harker, J. The effect of bacteria on leg ulcer healing. Br J Community Nurs 2001; 6: 3, Doughty, D.B., Sparks-DeFriese, B. Wound healing physiology. In: Bryant, R., Nix, D. (eds). Acute and Chronic Wounds: Current Management Concepts. (3rd edn). Mosby, Davies, C.E., Hill, K.E., Newcombe, R.G. et al. A prospective study of the microbiology of chronic venous leg ulcers to reevaluate the clinical predictive value of tissue biopsies and swabs. Wound Repair Regen. 2007; 15: 1, continued on page 94 s

3 Table 1. Baseline medical conditions (n=24) Medical current Prior Never Related to the condition reference ulcer Table 2. Distribution of wound microflora (n=24 patients) Initial biopsy no. of Final biopsy no. of patients patients Enterococcus faecalis 9 Staphylococcus aureus 8 Staphylococcus aureus 8 Serratia marcescens 7 Pseudomonas aeruginosa 4 Enterococcus faecalis 3 Klebsiella oxytoca 3 Citrobacter koseri 1 Enterobacter cloacae 2 Enterococcus avium 1 Morganella morganii 2 Klebsiella oxytoca 1 Proteus mirabilis 2 Pseudomonas aeruginosa 1 Streptococcus agalactiae 2 Staphylococcus schleiferi 1 Acinetobacter baumannii 1 Streptococcus dysgalactiae 1 Enterococcus avium 1 Escherichia coli 1 Klebsiella pneumoniae 1 Proteus penneri 1 Staphylococcus haemolyticus 1 Serratia marcescens 1 Streptococcus dysgalactiae 1 no. (%) No. (%) No. (%) No. (%) Anaemia 6 (25%) 2 (8.3%) 16 (66.7%) N/A Diabetes 7 (29.2%) 0 17 (70.8%) N/A Stroke 0 5 (20.8%) 19 (79.2%) N/A Hypertension 19 (79.2%) 0 5 (20.8%) N/A Congestive heart failure 4 (16.7%) 3 (12.5%) 17 (70.8%) N/A Osteoarthritis 7 (29.2%) 1 (4.2%) 16 (66.7%) 0 Deep vein thromobosis 3 (12.5%) 8 (33.3%) 13 (54.2%) 5 (20.8%) Varicose veins 14 (58.3%) 2 (8.3%) 8 (33.3%) 6 (25%) Results Population and treatment time: Thirty-three patients were recruited between June 2008 and September Eight were excluded during screening for not meeting the minimum level of bioburden and one patient was withdrawn during the study for not wearing the compression bandages. Therefore, a total of 24 patients were eligible for complete assessment. The gender distribution was 15 (62.5%) males and nine (37.5%) females. The mean age was 59.8 years (range: years). The mean ABPI of the reference limb was 1.1 (median: 1.2, range: ) and all patients had venous disease confirmed by ultrasound. Current and previous medical conditions are shown in Table 1. In terms of mobility, 11 (45.8%) patients could walk unaided, 12 (50%) could walk with an aid and the remaining one (4.2%) patient was chairbound. Eleven patients (45.8%) completed the 12-week study period and 11/24 (45.8%) discontinued early due to ulcer closure. Two patients (8.3%) were withdrawn during the course of the study, but were included in the intent-to-treat analysis: one died of unrelated causes (respiratory failure/pneumonia) on day 33 and the other requested to be withdrawn after missing their 10-week follow up. Primary endpoint analysis The analysis set consisted of 24 patients that had a bioburden of 10 5 cfu/g, confirmed by biopsy at baseline assessment. The safety population consisted of all 33 patients recruited into the study who attended baseline assessment and had at least one dressing application. At initial screening, the median log 10 level of bioburden from ulcer biopsy was 5.5log 10 cfu/g (mean 5.5, range ). The median level of bioburden, also from biopsy, was 330,000cfu/g (mean 347,917, range 100, ,000). The distribution of wound microflora is shown in Table 2. Seventy-nine per cent (26/33) of patients had Staphylcoccus aureus present at some point during treatment. Of these 26 patients, 69% (18/26) had a meticillin-resistant strain. At initial biopsy, the most prevalent species was Enterococcus faecalis. At week 8, the level of bioburden had reduced to <10 5 cfu/g tissue in 13 patients (54.2%) (95% confidence interval, range: %), which compares with 10 patients (41.7%) at week 2 (Table 3). At week 8 the median log 10 level of bioburden was 4.8log 10 cfu/g tissue (mean 3.3, range: 0 5.6) (Fig 1). The median level of bioburden in the ulcer, from biopsy sample, was 60,000cfu/g tissue (mean: 122,435, range: 0 400,000). Eight patients had no recorded ulcer bioburden,as their ulcers had closed or were too small for biopsy (<1.5cm 2 ). There was significant evidence of a reduction in the log 10 level of bioburden from the tissue biopsy after 8 weeks from the baseline assessment (p<0.001) (Table 3). Compared with baseline, at weeks 2, 4 and 8 the median percentage reductions in bioburden were 50.5% (mean 37.2%), 56.8% (mean 35.1%) and 80.4% (mean 54.0%), respectively. The median time from baseline to achieve a level of bioburden

4 <10 5 cfu/g tissue was 56 days. This represents the associated number of days to achieve <10 5 cfu/g tissue for 50% of the patients. However, there was no evidence of an association between the reduction in the log 10 level of bioburden from the tissue biopsy per week and the percentage reduction in ulcer area per week (Spearman s rank correlation: 0.103, p=0.512). The ulcer swab was collected to assess its practicability as a primary test to check the density of bacteria in the wound. Based on the ulcer swabs, the median percentage reductions in bioburden compared with baseline were 80.9% at 8 weeks and 100% at 12 weeks. Secondary endpoints l Ulcer closure and area The median ulcer area at baseline was 12.3cm 2 (mean 20.1, range ). The general appearance of the reference ulcer is indexed in Table 4. The most common tissue type was dull red (unhealthy granulation) tissue. All ulcers were exuding and all had at least three clinical signs of infection at baseline, with a mean of 5.5 per patient. In all patients, failure to progress towards healing was a marker of chronic infection. Most had increased exudate (19/24, 79.2%), spontaneous pain between dressing changes (20/24, 83.3%), local periulcer erythema (18/24, 75%) and oedema (22/24, 91.7%). There was significant evidence of a reduction in ulcer area from baseline to treatment discontinuation (p<0.001) (Fig 2). The median percentage reduction at treatment discontinuation was 96.9% (mean 81.5%, range %). During the study, 19/24 (79.2%) patients achieved at least a 75% reduction in ulcer area, all within 8 weeks of the baseline assessment. For the patients who remained in the study, the median percentage reductions after 2, 4 and 8 weeks were 28.1% (n=24), 56.4% (n=23) and 93.4% (n=23), respectively. At discontinuation of treatment, ulcers with a baseline area less than or equal to the median (12.3cm²) had a slightly higher percentage area reduction (median 100%, mean 83.2%), compared with those with an area greater than the median (median 91.3%, mean 79.7%). Eleven of 24 (45.8%) patients achieved ulcer closure by treatment discontinuation (median duration of 80.5 days). The percentage of patients who achieved ulcer closure at 2, 4 and 8 weeks following baseline assessment were 0%, 4.2% (1/24) and 29.2% (7/24), respectively. Of note, 5 of the 11 patients whose ulcer had closed at treatment discontinuation, the ulcer had re-opened by the week 12 follow-up assessments. The re-opened ulcers were not included in the statistical model but will be addressed in the discussion. Considering ulcer closure rates, the median percentage reduction per week was slightly higher (9.6% per week) for smaller ulcers, those with areas Table 3. Level of bioburden from the tissue biopsy Log 10 cfu/g Baseline Week 2 Week 4 Week 8 (n=24) (n=24) (n=24) (n=23) (4.2%) 2 (8.3%) 8 (34.8%) >0 to < to <4 0 4 (16.7%) 3 (12.5%) 1 (4.3%) 4 to <5 0 5 (20.8%) 3 (12.5%) 4 (17.4%) 5 24 (100%) 14 (58.3%) 16 (66.7%) 10 (43.5%) Mean (median) 5.5 (5.5) 4.7 (5.1) 4.6 (5.3) 3.3 (4.8) (range) (5 5.8) (0 5.9) 0 0 ( ) cfu/g Mean 347, , , ,435 (median) (330,000) (135,000) (180,000) (60,000) (range) (100, ,000) (0 750,000) 0 700,000) (0 400,000) Fig 1. Bioburden at week 1 and week 8 Level of bioburden (log 10 cfu/g tissue) Mean Multiple results Individual results Median Baseline Week 8 Assessment less than or equal to the baseline median, than for larger ones (7.8%/week) with areas greater than the median. Using this data, the median time to achieve closure (the time it took for 50% of patients to close) was 91 days. Other surrogate markers of wound health, including the appearance of the VLU, the condition of the surrounding skin, exudate level, pain level and clinical signs of infection were all reviewed. There was a reduction in the percentage of dull red (unhealthy granulation) tissue from baseline (39.0%) to treatment discontinuation (15.4%). There was also a reduction in the percentage of devitalised (yellow slough or black necrotic) tissue from baseline (21.9%) to treatment discontinuation (5.4%). This s

5 Fig 2. Median ulcer area and bioburden reduction over time reduction over time Reduction in log 10 level of bioburden Week 10 British Medical Association (BMA). British National Formulary 58. BMA, Skog, E., Arnesjö, B., Troëng, T. et al. A randomized trial comparing cadexomer iodine and standard treatment in the out-patient management of chronic venous ulcers. Br J Deramatol 1983; 109: 1, Harcup, J.W., Saul, P.A. A study of the effect of cadexomer iodine in the treatment of venous leg ulcers. Br J Clin Pract 1986; 40: 9, Steele, K., Irwin, G., Dowds, N. Cadexomer iodine in the management of venous leg ulcers in general practice. Practitioner 1986; 230: 1411, % reduction in ulcer area Reduction in log 10 level of bioburden Table 4. General appearance of the reference ulcer (n=24) Tissue types % reduction in ulcer area % of the reference wound Pink (epithelial) 18.8 (0) 0 90 Beefy red (healthy granulation) 15.6 (0) 0 80) Dull red (unhealthy granulation) 39.0 (25) (0 100) Friable granulation 1.9 (0) (0 40) Yellow (sloughy) 18.8 (10) (0 80) Black (necrotic) 3.1 (0) (0 70) Other 2.9 (0) (0 30) corresponded with an increase in the percentage of pink (epithelial) tissue from 18.8% at baseline to 51.5% at treatment discontinuation. There was very little observed difference in periwound skin condition between baseline and discontinuation. There was, however, a significant reduction in the level of exudate (p<0.001) and a reduction in the level of pain during the last week of the study. At baseline, 14 patients (58.3%) noted moderate pain from the ulcer during the preceding week. At treatment discontinuation, eight (33.3%) noted no pain during the previous week, four (16.7%) noted minimal pain, and only two (8.3%) noted moderate pain. All patients had at least one clinical sign of infection at a post-baseline assessment (100%). There was a significant reduction (p<0.001) in the percentage of patients with clinical signs of infection (from 100% to 50%) from baseline to treatment discontinuation. The median time to resolving all clinical signs of infection for the remainder of the treatment period was 91 days (Table 5). In addition, there was a significant reduction in the number of clinical signs of infection between baseline assessment and treatment discontinuation (p=0.001); the median number of clinical signs of infection at treatment discontinuation was 0.5 (mean: 1.6, range: 0 7) (Table 5). l Treatment frequency and duration The mean treatment duration for the full analysis set (24 patients) was 69.5 days (median: 80.5 days). Patients who completed the full 12-week study period all had a treatment duration of between 84 and 91 days (mean: 84.6 days), obviously longer than for those who achieved ulcer closure (mean: 57.3 days) or were withdrawn (mean: 53 days). The mean time between assessments was 7.2 days (range: 4 14 days). The median duration until the final tissue biopsy was 28 days (mean: 45.4 days, range: days) and until the final swab was 56 days (mean: 61 days, range: days). Treatment with topical dressing plus compression was not interrupted for longer than 7 days in any patient, although in 2 cases, the dressing regimen was interrupted due to emergency room visits. In 22 assessments, the dressings were left on for longer than 7 days (14 days in 2 cases). l Safety There were 11 adverse events across eight patients, of which one was an unanticipated, possibly device-related adverse event: a report of severe pain on dressing removal at week 6. No action was taken and the patient did not report significant pain from any further dressing removals. In three patients, there were five serious adverse events, which were all non-device related. Ten of the 11 adverse events were resolved with no sequelae at study discontinuation. The remaining one was a serious non-device related adverse event resulting in the patient s death. Discussion We set out to assess the antibacterial efficacy of sustained release silver sulphadiazine foam dressing plus compression in infected, non-healing VLUs. VLUs that are almost one year old, have a high bioburden and an area of >10cm 2 have been defined as hard to heal. 15,16 While our inclusion criteria were designed to specifically collect VLUs with a high bioburden, we also collected patients with relatively large ulcers (mean 20.1cm 2 /median 12.3cm 2 ) but did not look at the actual duration. It is this specific cohort we showed a significant reduction in bioburden, as measured by quantitative tissue biopsy. Compared with baseline values, the median percentage reduction in bioburden was 50.5% at week 2, 56.8% at week 4 and 80.4% at week 8. Surprisingly, there is little relevant data available for the comparison of bioburden in chronic wounds, with most of the literature being based on in vitro studies. 17

6 Table 5. Clinical signs of infection over time Our current data is the first obtained through the assessment of bioburden within a wound being managed with SSPF. While the use of this dressing has been shown to result in a 4log 10 reduction in the numbers of some bacteria in vitro, the in vivo wound environment is perhaps a more challenging setting, as multiple species are present and colonisation is not uniform. For most of the patients in whom a significant reduction in bioburden was achieved, the effect was rapid: it reduced to <10 5 cfu/g at 2 weeks in 10 patients (41.7%). This suggests that short-term biopsy data may be of use in judging the efficacy of future antibacterial therapies. Previous studies have reported that % of VLUs are colonised with bacteria, with most studies supporting the concept that bioburden delays healing. 5 7,15,21 23 It is thought that bacterial diversity and bacterial density both contribute to delayed healing The most common organisms previously noted in chronically colonised VLUs are Pseudomonas aeruginosa, Staphylococcus aureus and β-haemolytic Streptococcus. 15,22,27 While our data are similar, we have found a higher proportion of Ente- 0rococcus faecalis than previously reported. Over the course of the study, the diversity and Baseline Week 2 Week 4 Week 8 Treatment (n=24) (n=24) (n=23) (n=23) discontinuation (n=24) no. (%) No. (%) No. (%) No. (%) No. (%) Any clinical signs of infection 24 (100%) 22 (91.7%) 18 (78.3%) 14 (60.9%) 12 (50%) p=0.219 p=0.033 p=0.002 p<0.001 No progression towards healing 24 (100%) 3 (12.5%) 5 (21.7%) 4 (17.4%) 7 (29.2%) Increased exudate 19 (79.2%) 9 (37.5%) 3 (13%) 0 3 (12.5%) Spontaneous pain between 20 (83.3%) 17 (70.8%) 11 (47.8%) 7 (30.4%) 9 (37.5%) dressing changes Increased temperature around ulcer 4 (16.7%) 1 (4.2%) Discoloration of 6 (25%) 2 (8.3%) 1 (4.3%) 3 (13.0%) 1 (4.2%) granulation tissue Friable granulation tissue 5 (20.8%) 4 (16.7%) 5 (21.7%) 4 (17.4%) 3 (12.5%) Local peri-ulcer erythema 18 (75%) 16 (66.7%) 10 (43.5%) 8 (34.8%) 5 (20.8%) Oedema 22 (91.7%) 16 (66.7%) 13 (56.5%) 5 (21.7%) 5 (20.8%) Purulent exudate 4 (16.7%) 6 (25%) 4 (17.4%) 2 (8.7%) 1 (4.2% Malodour 10 (41.7%) 9 (37.5%) 8 (34.8%) 4 (17.4%) 5 (20.8%) No. of clinical signs of infection Mean (median) (range) 5.5 (5) (3 9) 3.5 (4) (0 6) 2.6 (3) ( (1) (0 6) 1.6 (0.5) (0 7) - p<0.001 p<0.001 p<0.001 p<0.001 quantity of bacteria decreased. The incidence of MRSA was very high but this mimics the increasing incidence seen in the community. One of the main drawbacks of this study is that it did not take biofilms into account. In the future, such studies should include polymerase chain reaction assays to identify difficult-to-assess bacterial types that may be in a dormant phase. More important to the clinician and patient is the closure rate. Here, it was 45.8% (11/24) at 80.5 days and the average wound area reduction was 93.4% after 8 weeks. In this very-hard-to-close wound population, these results are very positive, when compared with published wound closure rates for active agents such as matrix metalloproteinase-absorbing dressings (41% closure; 54.4% average wound area reduction at 12 weeks), bilayered skin substitutes (63% closure at mean 61 days) and extracellular matrices (50% closure at mean 3 months). 1 3 This may relate to the under-recognition of critical colonisation in previous trials, or a policy of severely restricted debridement, seen in trials involving skin replacements and extracellular matrices. 4 However, a high re-opening rate was noted after closure. This study, like many others, regarded a wound being 14 Vermeulen, H., van Hattem, J.M., Storm- Versloot, M.N., Ubbink, D.T. Topical silver for treating infected wounds. Cochrane Database Syst Rev 2007; 1: CD Madsen, S.M., Westh, H., Danielsen, L., Rosdahl, V.T. Bacterial colonization and healing of venous leg ulcers. APMIS 1996; 104: 12, Vowden, P., Romanelli, M., Peter, P.R. et al. The effect of amelogenins (Xelma) on hard-to-heal venous leg ulcers. Wound Repair Regen 2006; 14: 3, Lansdown, A.B., Williams, A., Chandler, S., Benfield, S. Silver absorption and antibacterial efficacy of silver dressings. J Wound Care 2005; 14: 4, Woodmansey, E. Antimicrobial activity of ALLEVYN Ag Non Adhesive dressing against a broad spectrum of microorganisms. Smith & Nephew Data 2007; Woodmansey E. Antimicrobial activity of ALLEVYN Ag over 7 days. Smith & Nephew Data. 2007; Woodmansey E. Antimicrobial activity of ALLEVYN Ag dressings in comparison to other silver-based antimicrobial dressings against Pseudomonas aeruginosa using a dynamic shake flask. Smith & Nephew Data. 2007; Robson, M.C., Stenberg, B.D., Heggers, J.P. Wound healing alterations caused by infection. Clin Plast Surg. 1990; 17: 3, Kotz, P., Fisher, J., McCluskey, P. et al. Use of a new silver barrier dressing, ALLEVYN Ag in exuding chronic wounds. Int Wound J 2009; 6: 3, continued on page 96 s

7 The statistical analysis was carried out by Trevor Mole of Smith & Nephew, Hull, using SAS software package 9.1 (Cary, NC, USA). Study protocol support, data collection and analysis was supported by Samantha Hartwell, PhD, of Smith & Nephew, Hull, UK closed on two successive visits as a successful wound closure. While some studies have linked the use of a specific dressing to prolonged post-treatment closure times, 2 many post-treatment variables including concordance with compression, nutritional and ambulatory status affect long-term closure rates. As this was only a 12-week study with limited followup, we did not assess all of these post-treatment variables. The present study also had a high level of debridement at almost 100% of visits, which may have had a significant impact on wound healing. 27 As one would expect with such high wound closure and improvement rates, the non-classical signs of infection, as described by Fierheller et al., appeared to be significantly reduced by the application of this sustained-release antibacterial dressing. 28 The median number of clinical signs of infection decreased from five at initiation to three at week 4 (p=0.033). As one would expect with improved healing rates, the general appearance of wounds improved, with increased granulation tissue and a decrease in devitalised tissue, and a decrease in exudate was noted. While this finding supports a previously published paper on this dressing combination, 22 we did not have a control arm using compression bandaging alone, so it is impossible to establish whether or not the topical dressing is responsible for this decrease. Importantly, a significant decrease in wound pain was observed, which allowed for good patient concordance within the cohort. Like others before it, this study does not offer a link between bioburden and delayed closure. When considering such a link, the assumption has been that bacteriology of the wound makes it difficult for the wound to close, due to metabolic abnormalities. However, large analysis has previously shown it is very difficult to prove a correlation between bioburden and wound healing. 29 In this study, rather than assuming a negligible bioburden in closed wounds, we analysed data based on the final biopsy results prior to wound closure. It is proposed that by removing barriers to healing in these hard-to-heal wounds in this case, reducing bioburden to below the threshold level known to reduce epithelialisation and affect normal wound healing (>10 5 cfu/g) the clinician creates an environment in which the wound can progress to closure. 21,30 In our study, closure may have been facilitated by a relatively aggressive debridement protocol instead of a change in the bioburden. We assessed swab cultures to see if they correlated with quantitative tissue culture. The number of patients with a bioburden >10 5 cuf/g reduced from 20/24 (83.3%) at baseline to 5/22 (22.7%) at week 12. While there was a correlation with quantitative tissue culture results, this not significant enough (p>0.05) to allow us to abandon quantitative biopsy data as our gold standard. There are study limitations. While the dressing effect appears important, especially given such positive clinical outcomes of these hard-to-close wounds, there is no control group to control for the bias of simple, good wound care. While quantitative culture remains the gold standard, it is only representative of the region biopsied (in this study, the area that visually appeared the worst), not the whole wound. We made no effort to assess for biofilm, which is by definition not culturable, and which may respond better to serial debridement than to topical antibacterials. Future studies would be well served to incorporate polymerase chain reaction assays, although baselines for critical colonisation are yet to be established. Finally, we did not include a control arm because we believe that there is sufficient baseline data for wound closure rates in this population. Conclusion In this study of difficult to heal venous leg ulcers, managed with SSPF and MLCW, we demonstrated evidence of a statistically significant (p<0.001) reduction in bacterial burden. This occurred at a lower level than is sometimes thought to be clinically significant, but with a much higher wound closure rate than expected. In addition, clinical signs of infection were reduced while being well tolerated by the patient. The clinical implication is that this particular algorithm appears to work well, but one may not expect the same results without frequent debridement. If one already employs frequent debridement as part of their VLU treatment protocol, but does not have a standardised post debridement dressing, this combination of SSPF/MLCW now has clinical evidence to support it use. n 23 Hansson, C., Hoborn, J., Möller, A., Swanbeck, G. The microbial flora in venous leg ulcers without clinical signs of infection. Repeated culture using a validated standardised microbiological technique. Acta Derm Venereol 1995; 75: 1, Trengove, N.J., Stacey, M.C., McGechie, D.F., Mata, S. Qualitative bacteriology and leg ulcer healing. J Wound Care 1996; 5: 6, Bowler, P.G. The 10(5) bacterial growth guideline: reassessing its clinical relevence in wound healing. Ostomy Wound Manage 2003; 49: 1, Davies, C.E., Hill, K.E., Newcombe, R.G. et al. A prospective study of the microbiology of chronic venous leg ulcers to reevaluate the clinical predictive value of tissue biopsies and swabs. Wound Repair Regen 2007; 15: 1, Cardinal, M., Eisenbud, D.E., Armstrong, D.G. et al. Serial surgical debridement: a retrospective study on clinical outcomes in chronic lower extremity wounds. Wound Repair Regen 2009; 17: 3, Fierheller, M., Sibbald, R.G. A clinical investigation into the relationship between increased periwound skin temperature and local wound infection in patients with chronic leg ulcers. Adv Skin Wound Care 2010; 23: 8, O Meara, S.M., Cullum, N.A., Majid, M., Sheldon, T.A. Systematic review of antimicrobial agents used for chronic wounds. Br J Surg 2001; 88: 1, Robson, M.C., Lea, C.E., Dalton, J.B., Heggers, J.P. Quantitative bacteriology and delayed wound closure. Surg Forum 1968; 19: