3 Clinical evidence

3 Clinical evidence

Summary of clinical evidence

3.1 Full details of all clinical outcomes considered by the Committee are available in the assessment report overview.

3.2 The key clinical outcomes for the ReCell Spray‑On Skin system ('ReCell') presented in the decision problem were:

  • speed of healing

  • length of stay in hospital

  • degree of scarring

  • degree of pigmentation

  • number and size of donor sites

  • device‑related adverse events.

3.3 The sponsor presented 11 studies: 3 peer‑reviewed articles (Gravante et al. 2007, Park et al. 2013 and Wood et al. 2012) and 8 conference abstracts. An abstract by Sen et al. (2012) reported findings from a case series that overlapped with a study by Philp et al. (2013), submitted as academic in confidence. The External Assessment Centre considered 2 of the conference abstracts (Echlin 2012b and Palombo 2012) to be outside the decision problem because these studies evaluated the treatment of donor site wounds and hypertrophic scars rather than acute burns. The External Assessment Centre identified a further 9 conference abstracts, of which 2 were unavailable (Rawlins 2010 and 2011b), 4 contained data that overlapped with references presented by the sponsor and 1 other may have contained overlapping data.

3.4 The sponsor also presented 5 additional published studies in support of the degree of pigmentation, which was an outcome included in the decision problem. The External Assessment Centre considered these studies to be outside the scope of this guidance because they evaluated treatments for people with vitiligo or hypertrophic scars and not with burns.

Studies included in the sponsor's submission

3.5 Gravante et al. (2007) compared ReCell with skin grafting for treating deep partial thickness burns in a randomised controlled trial. Time to complete epithelialisation (completely healed), pain, aesthetic and functional scar quality and procedure time were assessed in 82 adults, 42 treated with ReCell alone and 40 with split thickness skin grafting. Time to complete epithelialisation was 13±2 days for the ReCell group and 12±2 days for the skin grafting group (difference not statistically significant). Postoperative pain (measured using a visual analogue score) in the ReCell group was statistically significantly lower than pain in the skin grafting group (3.3±1.6 in the ReCell group compared with 6.8±1.2 in the skin grafting group, p=0.03). Postoperative analgesia levels were the same in both groups, although patients in the skin grafting group 'complained of an additional painful site (the area of harvesting)'. The total donor site in the ReCell group was statistically significantly smaller than that in the skin grafting group (p<0.001). Aesthetic scar quality was measured using the Vancouver scar scale and the development of contractures after 1 month was also measured as an indication of function in the burned area. Quantitative Vancouver scar scale values were not reported but were described as not different between the groups according to the judgment of 2 plastic surgeons. In the ReCell Spray‑On Skin system group, 12 patients (29%) developed at least 1 contracture, as did 15 patients (38%) in the skin grafting group (difference not statistically significant). Procedure time was significantly longer for the ReCell group than for the skin grafting group (59±4 minutes [mean ± standard deviation] for ReCell and 20±6 minutes for grafting, p<0.001). A second procedure was needed for 7 patients (17%) in the ReCell group and 6 (15%) in the comparator group.

3.6 Park et al. (2013) compared outcomes for 722 patients who were admitted to an Australian burns centre between January 2004 and December 2011 and who needed skin grafting or a skin replacement procedure. A total of 770 patients were enrolled in the study but 48 were later excluded from this analysis. The authors reported that 724 enrolled patients were divided into 3 groups: ReCell alone (n=73), ReCell plus standard skin graft (n=264), or standard skin graft alone (n=387). It is assumed that 2 of these patients were later excluded but no information was given about their treatment group. The study reported that the type of surgical intervention did not influence the likelihood of the patient having a burn wound infection. ReCell alone was associated with a shorter hospital stay than standard skin grafting alone (actual values not reported, odds ratio 0.7, 95% confidence interval [CI] 0.57 to 0.82, p<0.01), but ReCell plus standard skin grafting was not associated with a shorter hospital stay (odds ratio 0.98, 95% CI 0.88 to 1.10, p=0.85). The authors concluded that patients treated with ReCell alone had a 30% shorter length of hospital stay (p<0.01) compared with standard skin grafting. The External Assessment Centre questioned the validity of the statistical analysis, but noted that the authors stated that the shorter hospital stay for those treated with ReCell alone should be interpreted carefully. This is because wound depth and surgery timing differed between the ReCell and standard skin grafting groups. The study authors also speculated that the reduction in donor skin harvesting associated with the ReCell technique may have reduced the length of hospital stay.

3.7 Wood et al. (2012) evaluated the use of ReCell and a biosynthetic dressing (Biobrane) compared to standard treatment (dressings every 2–3 days with definitive surgery at 10–14 days after the injury). A total of 13 children were included and followed up over 6 months; 5 received ReCell with Biobrane, 4 received Biobrane alone and 4 received standard treatment. At 10 days after the initial burn, none of the patients in the ReCell plus Biobrane group were assessed as needing surgery; 1 patient in the Biobrane group needed surgery, and 3 out of 4 patients in the standard treatment group needed surgery. The median time to complete healing for ReCell group was similar to the time for the Biobrane group and the median time was longer in the standard treatment group (median [interquartile range] 16.0 [11.5–18.0], 16.0 [14.25–23.0] and 36.5 [18.5–47.7] days respectively; no statistical analysis provided). The patients treated with ReCell plus Biobrane showed a higher proportion of wound area healing at both 10 and 21 days after the burn (95% and 100% respectively) compared with Biobrane alone (83.2% and 97.7%) or with standard treatment (71.2% and 90.1%) although no statistical analysis was reported.

3.8 Dunne and Rawlins (2012a) observed 40 children in the UK who were treated either by ReCell plus Biobrane (mid and deep dermal burns, n=13), or Biobrane (superficial dermal burns, n=20), or standard skin grafting (full thickness burns, n=7). The authors reported that hospital stay was shorter and scar assessment was better in the ReCell plus Biobrane and the Biobrane groups, although neither numbers nor statistical significance were reported. The External Assessment Centre noted that this was a non‑comparative review of treatments in burns of different depths and could be considered as 3 separate case series.

3.9 Echlin et al. (2012a) observed 5 patients with mid to deep dermal facial burns (3 scalds and 2 flame burns) in the UK. Four patients were treated with ReCell plus non‑adherent dressings because they were assessed 9–11 days after the burn and their burns were deemed unlikely to heal within 3 weeks. One further patient was assessed 23 days after injury and was treated with ReCell plus an allograft. The authors concluded that ReCell increased the speed of wound healing (mean healing time 5 days for those treated 9–11 days after the burn, maximum healing time 7 days) and decreased the rate of standard skin grafting and subsequent scar formation, but no comparative data or statistical analyses were reported. The authors reported that as a result of this study the burns service involved changed its practice to treat these wounds with ReCell rather than a skin graft.

3.10 Rawlins et al. (2011a) compared the outcomes for 15 adults with deep dermal flame burns treated 48–72 hours after injury with either ReCell plus Biobrane (n=5), or standard skin grafting (n=10). The mean time to wound healing (no healing measurement defined) was 18 days in the ReCell plus Biobrane group and 48 days for the standard skin grafting group. No statistical analyses were reported. The authors reported that less analgesia was needed in the ReCell plus Biobrane group than in the standard skin graft group and that scar quality was better in the ReCell plus Biobrane group, but values were not reported.

3.11 Rawlins (2013) described the outcomes for 26 children with deep dermal burns who were treated with ReCell (n=11) or a standard skin graft (n=15). The mean visual analogue scale score for scar quality, assessed by independent clinicians, was very similar in the 2 groups (3.9, 95% CI 2.8 to 4.9 for ReCell and 3.9, 95% CI 3.3 to 4.5 for standard skin graft; p=0.97). Operative time was longer for ReCell than for standard skin grafting (mean 87 minutes compared with 58 minutes; p=0.05), although the total burn surface area was greater for the ReCell group than the standard skin graft group (mean total burn surface area of 6.5% compared with 2.9%; p=0.04).

3.12 Sen et al. (2012) reported data from a case series. The data from this case series overlap with a study by Philp et al. (2013), a report on which was available in full to the Committee as academic in confidence. Sen et al. reported narrative findings from a case series of 5 patients with deep partial or full thickness burns over more than 50% of their total body surface area. Burns were covered with split thickness skin grafts and split thickness dermal grafts. The split thickness dermal grafts and donor sites were treated with ReCell. The graft and donor sites were assessed by 2 independent observers. Graft take was reported as being complete for all patients and healing was reported to be similar to that for patients treated with split thickness skin grafting alone.

Additional studies identified by the External Assessment Centre

3.13 Dunne and Rawlins (2012b) reported early results from the same study as Rawlins (2013).

3.14 Dunne and Rawlins (2013) reported outcomes for 11 children treated with ReCell plus Biobrane for scalds and 10 adults, 8 of whom were treated with ReCell plus Biobrane for flame burns. The nature of the burns in the other 2 adults was unclear. This was a retrospective review with possible overlap with Rawlins (2011a, 2013), Rawlins et al (2011a and 2011b) and Dunne and Rawlins (2012b). Wound coverage, pigmentation, hypertrophic scarring and donor site morbidity were assessed. One child needed a standard skin graft after treatment with ReCell but early wound coverage and good pigmentation were reported with minimal hypertrophic scarring or donor site morbidity.

3.15 Hiller et al. (2013) (possible overlap with Rennekampff et al. 2011, see section 3.17) described the outcomes for 5 patients who had partial thickness facial burns and who were treated with ReCell. Only narrative outcomes were reported, with the authors reporting accelerated healing time and an improvement in scar quality.

3.16 Rawlins (2011a) and Rawlins et al. (2011b) (overlap with Rawlins et al. 2011a, see section 3.10) described outcomes for 4 patients treated with ReCell plus Biobrane compared with 10 matched controls who received standard skin grafts. Time to healing (healing measurement not defined) was 18 days in the ReCell plus Biobrane group and 48 days in the standard skin grafting group. Analgesia needs and length of hospital stay were reported as being reduced in the ReCell group compared with standard skin grafting, but no statistical analyses were reported. After 6 months, an assessment using the Vancouver scar scale indicated better scar outcomes in the ReCell group (5.3) than in the standard skin graft group (6.5).

3.17 Rennekampff et al. (2011) (possible overlap with Hiller et al. 2013, see section 3.15) reported outcomes in a case series of 5 patients with facial burns who were treated with ReCell. The depth of the burns was not clearly reported, although the authors' discussion implied that partial and full thickness burns were included. The authors reported that the full thickness burns needed skin grafting, however they did not state clearly that these burns were also treated with ReCell. Time to epithelialisation was reported as being 7–9 days after surgery and skin pigmentation was described as being slightly reduced compared with skin surrounding the area. No hypertrophic scars or severe contractions occurred.

3.18 Sood et al. (2009) reported findings from an intra‑patient comparative study in 10 patients with partial thickness burns. Each patient was treated with ReCell in 1 area and meshed skin grafting in another. Skin graft take was the main outcome reported, but the method of measurement was not reported. Results showed the overall graft take was 93.6% at the ReCell sites and 98.2% at the standard skin graft sites. Eight patients had 100% graft take in the ReCell group. In the standard skin grafting group, 9 patients had 100% graft take.

3.19 The External Assessment Centre identified 1 ongoing multicentre randomised, within‑patient controlled feasibility study that fitted the decision problem, but no findings were available.

Additional work requested by the Committee

3.20 The Committee requested additional information on the potential effect of the use of ReCell on improving skin colour match in burn scars. The External Assessment Centre gathered this information from an additional literature search and a survey of clinical experts, in conjunction with further work requested by the Committee to enable revision of the sponsor's cost modelling (see sections 5.7–5.9).

3.21 The External Assessment Centre identified 3 comparative studies examining the effect of using ReCell on the repigmentation of stable vitiligo lesions.

3.22 Daniel et al. (2011) reported interim results in a conference abstract from an intra‑patient randomised comparison of ReCell compared with mini‑grafting in 14 patients with stable vitiligo. Repigmentation at 3 months was 27% for the ReCell‑treated areas compared with 11% for mini‑grafting, but at 12 months the proportions were 15% and 12% respectively (not statistically tested).

3.23 Venugopal et al. (2009) reported results in a conference abstract from an intrapatient randomised comparison of ReCell compared with mini‑grafting in 12 patients with stable vitiligo who completed a 6‑month follow‑up. Pigmentation results were 'highly variable' with no difference between treatments, although ReCell produced a more uniform repigmentation in some cases.

3.24 Mulekar et al. (2007) conducted an intrapatient comparison of ReCell against melanocyte‑keratinocyte transplantation in 5 patients with stable vitiligo. At 4 months postoperatively results were comparable, with 100% repigmentation in both sites in 2 patients, no repigmentation in either site in 1 patient, 65% for ReCell compared with 100% for melanocyte‑keratinocyte transplantation in 1 patient and 40% for ReCell compared with 30% for melanocyte‑keratinocyte transplantation in 1 patient.

3.25 The External Assessment Centre stated that results from the comparative studies were inconclusive. The clinical experts surveyed by the External Assessment Centre pointed out that there were physiological differences between burns and surgically created wounds, such as debrided vitiligo lesions prepared for treatment. The opinion of the experts about whether outcomes from the treatment of vitiligo could be transferable to burns was divided. Most experts stated that they might consider using ReCell for acute burns if it could be shown to have benefit in hypopigmentation conditions.

Additional evidence

3.26 During consultation, 4 additional unpublished clinical studies were identified. These included interim findings from an ongoing clinical trial comparing ReCell with split thickness skin grafts for healing burn wounds and with standard care for healing donor sites; 2 comparative studies (Komen et al., Aust et al.) using ReCell in vitiligo and healed burn scars; and 1 case study from the United States (Holmes et al.) describing the use of ReCell to treat a large area burn, comparing outcomes with an institutional average. These studies were presented to the Committee in full (1 as commercial in confidence and 3 as academic in confidence) and were reviewed by the External Assessment Centre. The External Assessment Centre was of the opinion that the interim findings from the ongoing trial agreed with data presented in the sponsor's submission and the assessment report. The External Assessment Centre advised the Committee that the information submitted on the 2 comparative studies was not sufficiently detailed to draw robust conclusions, and that the findings did not suggest that outcomes for these indications would be transferable to acute burn wounds. The External Assessment Centre described the single case study as not being representative of the NHS care pathway and providing limited comparative data.

Adverse events

3.27 No adverse event reports relating to ReCell were found in a search of the Manufacturer and User Facility Device Experience (MAUDE) database operated by the US Food and Drug Administration.

Committee considerations

3.28 The Committee heard clinical expert advice that a key motive for using ReCell would be to reduce burn healing time, which could reduce the risk of infection and the degree of scarring, as well as shortening length of hospital stay. It also noted the potential benefits associated with using ReCell for treating burns by generating viable skin cells from a small donor site. However, the Committee judged that there was insufficient evidence to quantify any improvements in burn healing time, shorter hospital stays and improvements in pigmentation in healed burns to make a recommendation for routine adoption.

3.29 The Committee heard expert advice that small (up to 10% total body surface area) partial thickness burns usually heal with conventional dressings or a biosynthetic dressing such as Biobrane, without the use of skin grafting. The Committee considered the clinical evidence taken together with the clinical expert advice gathered by the External Assessment Centre (see sections 5.7–5.9). It accepted the views of the majority of clinical experts obtained by the External Assessment Centre, who agreed that these burns would heal without the use of skin substitutes. The Committee concluded that ReCell was unlikely to be beneficial in treating patients with small partial thickness burns.

3.30 The Committee noted that the available clinical evidence suggested that ReCell may be an effective alternative to skin grafting for mid‑dermal to deep dermal partial thickness burns and the need for a smaller donor site may offer advantages. The Committee was aware that most of these studies were conducted outside the UK, and it accepted that current clinical opinion would not favour the use of ReCell as an alternative to skin grafting in these burns in NHS practice.

3.31 The Committee heard clinical expert advice that ReCell might be used alone on mid to deep dermal or indeterminate depth burns where the need for grafting is initially unclear. The External Assessment Centre confirmed that there was some support for this scenario in the clinical expert views obtained (see section 5.15). The Committee concluded that ReCell may offer some benefit in this group from avoiding skin grafting and better scar outcomes, but that further evidence about these outcomes was needed.

3.32 The Committee was advised by experts that the patients most likely to benefit from the use of ReCell were those with large full thickness or deep partial thickness burns that need meshed skin grafting. It was advised that potential applications also include the use of ReCell on skin graft donor sites to reduce healing time and allow further grafts to be taken sooner from the same site to treat very large burns. The Committee was aware that there was little published evidence on these outcomes. The Committee considered additional clinical expert advice subsequently gathered by the External Assessment Centre and concluded that use of ReCell in conjunction with meshed skin grafts in these burns and donor sites showed promise, but the evidence was inconclusive.

3.33 The Committee considered the possibility that ReCell might reduce the occurrence of hypopigmentation in healed burns, as a result of the retention of melanocytes in the cell suspension. It noted that scar skin colour match was an important consideration for burns patients, but that primary evidence for pigmentation outcomes when using ReCell in burns was scarce. At its first meeting, the Committee noted a difference in clinical expert opinion on the relevance of data on using ReCell for treating hypertrophic scars or for non‑burn indications such as vitiligo. It therefore asked the External Assessment Centre to include questions about skin colour match in burns scars in its additional work, as well as reviewing any relevant evidence. The Committee accepted the External Assessment Centre's conclusion from its additional work that both the study evidence and expert opinion on pigmentation outcomes when using ReCell were inconclusive.

3.34 The Committee judged that further evidence should be gathered from additional research to investigate the clinical utility of using ReCell in:

  • full thickness or deep partial thickness burns that need skin grafting (see section 3.32) and

  • partial thickness or indeterminate depth burns where the need for grafting is unclear (see section 3.31).

    The Committee was advised that observational data, such as those collected in the International Burn Injury Database or from well‑designed clinical audits, could be useful in helping to resolve some of the clinical uncertainties in these patient groups.

3.35 For larger burns needing wide‑meshed skin grafting, the Committee heard advice from clinical experts that cultured cells are sometimes used in the NHS to promote healing. Their advice concurred with the findings in the External Assessment Centre's additional report, that the choice of whether to use cultured cells or ReCell was based on the availability of cultured cells and the personal preferences of clinicians because of a lack of good comparative evidence. The Committee was advised by clinical experts that using cultured cells has the advantage of the availability of large volumes of viable cells, but may be associated with a fragile epithelium and problems with skin loss in the healed burn. The External Assessment Centre confirmed that there was currently no evidence comparing ReCell with cultured cells. The Committee concluded that data from further research comparing these treatments would be useful.

3.36 The Committee was advised by clinical experts that multicentre research into the effectiveness of ReCell in treating burns would be possible, but difficult, particularly for large full‑thickness burns that occur in only a small number of people each year. The Committee was also advised that it might be hard to determine the effect of the use of ReCell or other treatment strategies on length of hospital stay because many other factors influence this in patients with large, deep burns. Careful selection of patients would be necessary to minimise the confounding effect of comorbidities.

3.37 The Committee heard expert clinical advice that time to 95% healing is a standard measure and would be an appropriate endpoint to include in any research, in addition to length of hospital stay, cosmetic appearance of scars (assessed using a validated scale) and a measure of function in the burned area.

3.38 At its meeting to discuss the comments received during consultation, the Committee considered the additional evidence obtained by the External Assessment Centre, described in section 3.26. The Committee concluded that these studies did not provide sufficient clinical evidence to alter its recommendations. However, the Committee considered that findings from these studies could be useful to inform further research.

  • National Institute for Health and Care Excellence (NICE)