Population and pathology data at baseline. Of the 102 hospitalized patients or outpatients experiencing a leg ulcer considered at risk of secondary infection recruited into this clinical trial, 52 comprised the silver group and 50 the control group. The included population was predominantly composed of outpatients (72, 70.6%), female (71, 69.6%), markedly overweight (mean body mass index [BMI] 28.7), and had diabetes mellitus (19, 18.6%) and a history of deep venous thrombosis (33, 32.4%).
This clinical trial was the accumulation of 4796 days of treatment and documented 800 medical evaluations and 2461 local care operations. The median treatment time for each patient was 56 days and 51 days in the silver and control groups, respectively. Time between 2 dressing changes was similar (2.12 days vs. 1.84 days in the silver and control groups, respectively).
At baseline, wounds in both groups had a large surface area (average of 20 cm²); mean surface area was 22.3 cm2 (range 3.11–119.90 cm²) and 17.5 (range 4.89–73.49 cm²) for the silver and nonsilver groups, respectively (P = .84). Wounds were of long duration (average 10.8 months); mean duration was 11.0 (range 3.11–119.90) months and 10.5 months (range 4.89– 73.49) months for the silver and nonsilver groups, respectively (P = .11). Sixty-six (66) wounds (64.7%) were recurrent and had a wound bed covered by ~50% of sloughy tissue, with healthy perilesional skin present in only 3 patients. In total, investigators considered 81 (79.4%) of the wounds stagnant or worsening despite appropriate wound bed preparation and provision of specialized services by vascular or dermatology physicians, no necrosis after wound debridement, and provision of compression therapy (provided to more than 86% of the treated wounds at baseline before the start of the trial).
No significant differences were noted between the 2 groups for all the evaluated parameters (ie, demographic and wound data such as age, gender, weight, BMI, medical and surgical history, wound area, wound duration, wound recurrence, and condition of the perilesional skin), except for the Clinical Score, which was significantly higher in the silver than in the control group (3.84 vs. 3.40; P = .0032), due to a higher number of existing clinical signs of infection in the silver group (see Table 1). Of note: at least 3 of the 5 selected clinical signs were mandatory to include the patient in this RCT.
Efficacy. In terms of the primary endpoint, at week 4 with more than 95% of the patients complying with their venous compression treatment, the median surface area decreased 4.2 cm2 (29.1%) vs. 1.1 cm2 (9.5%) in the silver and control groups, respectively (P = .023). After week 4, when all patients in the silver group switched to the nonsilver-containing contact layer, ulcer area continued to decrease while no clinically relevant change was observed in the control group. At week 8, the median absolute wound surface reduction was 5.9 cm2 and 0.8 cm2, respectively, in the silver and nonsilver groups (P = .002). The same trends were observed when surface area changes were expressed as percentage reduction from baseline. After 8 weeks of treatment, relative median wound area decreased 47.9% in the silver group versus 5.6% in the control group (P = .036) (see Figure).
Hence, wound closure was significantly faster by week 4 in the silver group (median 0.145 cm2/day vs. 0.044 cm2/day; P = .009); even after switching to the nonsilver dressing, this parameter remained unchanged between week 4 and week 8, remaining significantly higher in the sequential group (0.135 cm2/day vs. 0.023 cm2/day; P = .001).
Regression analyses in a model that included ulcer outcome prediction factors (age, BMI, ulcer duration, area at inclusion, and recurrence) showed the likelihood of reaching an ulcer area regression of 40% in 8 weeks was 2.7 times greater in the silver group than in the control group (OR: 2.7; 95% confidence interval: 1.1-6.7; P = .038). None of the other factors included in the model was significant.
In terms of the secondary endpoints, at the end of the 8-week treatment period, the Clinical Score (based on the presence of the 5 predetermined clinical signs of infection) was significantly lower in the silver group than in the control group (1.43 vs. 2.31, respectively; P = .0001). At week 4, the percentage of ulcers with no predetermined clinical signs (clinical score equal to zero) was significantly higher in the silver than in the control group (39.2% vs. 16.7%; P = .0097) and remained significant at week 8 (35.3% vs. 20.8%; P = .044).
Safety. A total of 22 local adverse events (11 in each treatment group and not different between the 2 groups in their nature and frequency), possibly related to the tested dressings, were reported in 20 patients. In the first period (day 0 to week 4), no secondary infection occurred in the silver group; 2 infections were noted in the control group.
At the end of the treatment period, the perilesional skin was considered healthy in more cases in the silver group than in the control group (20 vs. 9 patients, 39.2% vs. 18.8%, in the silver and control groups, respectively), compared to the 3 patients with healthy perilesional skin at baseline.
Assays of blood silver (at day 0, week 4, and week 8) performed in the silver group to document biological safety showed the silver dressing did not increase blood silver levels. In addition, no local perilesional signs (which could evoke an argyria) or general signs (which could suggest a toxicity of silver salts) occurred during the trial.
Acceptability. Because the study was sequential for the silver arm, the acceptability results were divided into 2 periods. In the first period (day 0 to week 4), the silver dressing was significantly superior in terms of easier dressing removal (P = .03), less pain (P <.001), and less bleeding at removal (P = .001); no significant differences were observed between the 2 treatment groups in terms of maceration, dressing adherence to the wound, ease of application, and dressing conformability. In the second period (week 4 to week 8), no significant difference between the 2 groups was noted in the parameters studied.