3 Evidence

Clinical evidence

Relevant evidence comes from 5 studies, 3 of which are randomised controlled trials

3.1

Of the 5 studies that met the inclusion criteria defined in the scope, 2 were randomised controlled trials in venous and mixed leg ulcers and 1 was a randomised controlled trial in diabetic foot ulcers. There is also a non-comparative study in diabetic foot ulcers and a pooled analysis of non-comparative observational studies, which included both patient groups. For full details of the clinical evidence, see section 3 of the assessment report.

Results from EXPLORER show an increase in wound closure for diabetic foot ulcers

3.2

The multicentre, double-blind, international randomised controlled trial EXPLORER (n=240; 20‑week follow-up) compared UrgoStart with UrgoTul, a non-interactive dressing (Edmonds et al. 2018). The results reported a statistically significant increase in complete wound closure in favour of UrgoStart (p=0.002), as well as a statistically significant increase in absolute wound area reduction (p=0.022). Adverse effects and quality of life were similar in the 2 groups. The external assessment centre (EAC) noted that this was a European international study with some patients recruited from UK centres, but the number of patients recruited per centre was low (median=3) and the study only included patients with neuro-ischaemic ulcers.

Results from CHALLENGE show an increase in wound area reduction in the first 8 weeks for venous leg ulcers

3.3

The multicentre, double-blind, international randomised controlled trial CHALLENGE (n=187; 8‑week study period) compared UrgoStart with UrgoTul Absorb, a non-interactive dressing (Meaume et al. 2012 and Meaume et al. 2017). Compression therapy was used in both the intervention and control groups (more than 96% at week 6). The results reported a statistically significant increase in relative wound area reduction (p=0.002) and in absolute wound area reduction (p=0.003) in favour of UrgoStart. Use of UrgoStart also resulted in a statistically significant improvement in the pain and discomfort dimensions of the EQ‑5D (p=0.022). Adverse effects and patient acceptance were similar in the 2 groups. The EAC noted that the follow-up period of 8 weeks was potentially too short to assess healing in complex wounds, and only 13 wounds in total were completely healed by the end of the study (equally spread across the 2 treatment arms). No UK sites were included in this study, and there was a small number of patients per centre (mean=4.2).

A pooled analysis of non-observational studies broadly supports the evidence from the randomised controlled trials

3.4

Evidence from a pooled analysis of non-comparative data from 8 observational studies (Munter et al. 2017) supported the healing rates of diabetic foot and venous leg ulcers seen with UrgoStart in the randomised controlled trials. The analysis included more than 10,000 patients with chronic wounds, of whom 7,903 had venous leg ulcers and 1,306 had diabetic foot ulcers. However, the EAC noted that there were a range of follow-up periods (4 to 20 weeks), outcome measures and distributions of ulcer type in the included studies.

2023 guidance review

3.5

As part of the guidance surveillance process, new clinical evidence for UrgoStart was reviewed. A total of 22 eligible new clinical studies were identified. The EAG did not identify any evidence that contradicts the current NICE guidance for the UrgoStart range. Three studies noted that wound duration at baseline was an important predictor of outcome following UrgoStart use, with a shorter wound duration leading to better wound healing outcomes. There was still not enough evidence to recommend UrgoStart for non-venous leg ulcers because of a lack of new evidence explicitly in this population. For more on the new evidence, see the review report. [2023]

Cost evidence

The company's models for both leg ulcers and diabetic foot ulcers show cost savings with UrgoStart

3.6

The company presented separate de novo cost-effectiveness models for leg ulcers and diabetic foot ulcers. The leg ulcer model was a Markov model with a 1‑week cycle length, which incorporated 3 health states. The diabetic foot ulcer model was more complicated and included 6 health states. The company presented base-case results with a time horizon of 1 year. The results showed that compared with non-interactive dressings, UrgoStart was associated with savings of £274.25 per patient per year for leg ulcers and £666.51 per patient per year for diabetic foot ulcers.

The EAC's changes to the model parameters and its calibrations more accurately reflect NHS costs and consequences

3.7

The EAC considered that both model structures presented by the company adequately captured all the relevant health states, and that the assumptions were valid and reasonable. However, it changed some parameter values with which it did not agree. The EAC also calibrated the models to align with the healing outcomes and resource use from published UK studies (Guest et al. 2018a and 2018b). In its changes to the models, the EAC assumed that:

  • diabetic foot ulcers would not heal in 20% of patients and treatment would continue for 1.4 months (6.09 weeks) on average before the dressing was changed to a different product

  • leg ulcers would not heal in 37.6% of patients and treatment would continue for 1.9 months (8.26 weeks) on average before the dressing was changed to a different product.

The EAC's updated models show that UrgoStart is likely to be cost saving

3.8

Results from the EAC's base-case analysis showed that UrgoStart compared with standard care was associated with cost savings of £541 per patient per year for leg ulcers and £342 per patient per year for diabetic foot ulcers. The main drivers of the savings were the cost of dressings, the transition parameters for healing and infection or complications, and the cost of community nursing and hospital visits. Sensitivity analyses showed that UrgoStart was always cost saving for leg ulcers, but that it became cost incurring for diabetic foot ulcers if the healing rate was assumed to be half of that reported in the EXPLORER trial. For full details of the cost evidence, see section 4 of the assessment report.