The external assessment centre (EAC) considered 5 publications:
1 randomised controlled trial (RCT; Stanirowski et al. 2016a)
2 pilot RCTs (Totty et al. 2019; Stanirowski et al. 2016b)
1 non-RCT (Bua et al. 2017) and
1 unpublished audit (Taylor et al. 2020).
The EAC excluded 5 studies identified by the company because 4 did not include Leukomed Sorbact and there were significant uncertainties about the design of 1 study.
Stanirowski et al. 2016a and 2016b were both done in Poland in women having elective or emergency caesarean section. Totty et al. 2019 and Bua et al. 2017 were UK studies in people having vascular surgery. Taylor et al. 2020 contained audit data provided by the company on women having caesarean section in 1 UK health board.
Up to 30 days after surgery, surgical site infection (SSI) rates were lower for people having Leukomed Sorbact compared with those having standard dressings. The difference in infection rates was not always statistically significant depending on the trial size. The largest RCT was considered to have the least risk of bias (Stanirowski et al. 2016a). In this study, the SSI rate was 1.8% for Leukomed Sorbact compared with 5.2% for standard dressings at 14 days after caesarean section (statistically significant, p=0.04). In Stanirowski et al. 2016b, the SSI rate was 2.8% for Leukomed Sorbact compared with 9.8% for standard dressings at 14 days after caesarean section (not statistically significant; p=0.08). In Bua et al. 2017, the SSI rate was 1% for Leukomed Sorbact and 10% for standard dressings at 5 to 7 days after vascular surgery (statistically significant, p<0.05). In Totty et al. 2019 and Bua et al. 2017, SSI rates were 16% and 9% at 30 days respectively for Leukomed Sorbact after vascular surgery, compared with 26% and 10% for standard dressings. The differences were not statistically significant (p=0.161 and p=0.83, respectively).
In 3 studies there was less need for antibiotic treatment with Leukomed Sorbact compared with standard dressings (Bua et al. 2017, Stanirowski et al. 2016a and 2016b). In all studies the number of people reported as having antibiotics was low in both arms, and the reported differences were not statistically significant in Stanirowski et al. 2016a (0 in Leukomed Sorbact group, 4 in control group, p=0.13).
In Stanirowski et al. 2016a, women with SSI in the standard dressings group each had 2.9 outpatient hospital visits. Women with SSI in the Leukomed Sorbact group had 4.6 visits, a difference that was statistically significant, p=0.02. However, this was a secondary analysis in a small subgroup of women. The same study found that women with SSI who had Leukomed Sorbact had fewer additional days in hospital (0 days compared with 8.2 days for standard dressings, p=0.22).
The economic analysis in the Stanirowski et al. 2016a and Stanirowski et al. 2019 studies showed that Leukomed Sorbact is cost saving when compared with standard surgical dressings. In Stanirowski et al. 2016a, total costs for preventing and treating SSI were 5,775 euros in the standard dressings group compared with 1,065 euros in the Leukomed Sorbact group. In Stanirowski et al. 2019, the same data were used and a decision-analytic model was applied from a UK NHS perspective. This showed a cost saving of £119.07 per person in favour of Leukomed Sorbact.
The company submitted a simple decision tree model with 2 interventions, Leukomed Sorbact or standard surgical dressings. There were 2 outcomes, SSI or no SSI. The time horizon was 30 days. The company reported base-case cost savings per person with Leukomed Sorbact of £107.43 for caesarean section, £23.55 for vascular surgery, and £20.56 for all surgery. The company's sensitivity analyses found these results to be robust to parameter changes.
The EAC agreed with the company's model and its assumptions and made 1 change, to the cost of an SSI episode for vascular surgery. Leukomed Sorbact remained cost saving but the cost savings were lower than those estimated in the company's model for vascular surgery, at £17.82 per patient. The cost savings remained robust to parameter changes. The EAC chose not to model the use of Leukomed Sorbact for all types of surgery because it considered that there was insufficient clinical evidence to do so.