Scar quality in children with burns 5–7 years after injury: A cross‐sectional multicentre study

Abstract Long‐term scar formation is an important adverse consequence in children with burns, however, information regarding scar quality in the long‐term is lacking. Therefore, we evaluated scar quality and its predictors in children with burns 5–7 years after injury. Parents of children with mild/intermediate burns (≤10% total body surface area burned), and of children with severe burns (>10% burned) completed the patient scale of the Patient and Observer Scar Assessment Scale (POSAS 2.0) for their children's—in their opinion—worst scar 5–7 years post‐burn. Outcomes and predictive factors of scar quality were studied, and, for children with severe burns, POSAS parent scores were compared with observer scores. We included 103 children with mild/intermediate burns and 28 with severe burns (response rate: 51%). Most children (87%) had scars that differed from normal skin, with most differences reported for colour, and least for pain. Except for colour, children with severe burns had significantly higher scores (difference 0–2 points) on all scar characteristics (representing poorer scar quality) compared with children with mild/intermediate burns. Parent POSAS scores were on average 2.0–2.6 points higher compared to observer scores. Number of surgeries predicted both the mean POSAS and the mean overall opinion of a scar. In conclusion, 5–7 years post‐burn, the scar of the majority of children differed from normal skin, especially on the characteristic colour. The uncovered insights are useful in counselling of children and their parents on expectations of the final outcome of their (children's) scar(s), and help in further targeting scar prevention strategies for the individual child.

investigating outcomes of treatment and rehabilitation is gaining interest. 3 Scar formation is an important adverse consequence of burns. Paediatric scars can cause long-term disfigurement, as well as physical and psychological problems, and may result in a diminished health-related quality of life. 2 Assessment of patient perspectives is important as they may differ from clinician perspectives. 4,5 Patients have to live with their scars and by assessing the severity of their own scar(s), therapy is more likely to fit to their needs. 6 Studies on scar quality in children with burns have been performed; however, scar quality was only assessed up-to 28 months after injury. 4,[7][8][9][10] Research on scar quality in the longer term is lacking.
Studies showed that most change in scar quality is seen in the first months after burn injury. 7,8 Several factors have been reported to predict paediatric scar quality, including time to wound healing, a greater percentage total body surface area (%TBSA) burned, full thickness wounds, and multiple surgeries. 4,[8][9][10][11][12][13][14] However, it is also known that scar maturation can take several years, 15 and therefore, it is important to examine scar quality and associated risk factors in the longer term.
Therefore, the aim of this study was to evaluate scar quality and its predictors in children with burns 5-7 years after injury, separately for children with mild/intermediate burns and severe burns.

| METHODS
This study is part of the Burden of Burn Injuries study; a cross-sectional study on long-term consequences of burns. 16 This study was conducted according to the principles of the Declaration of Helsinki, and approved by the Ethics Committee (NL59981) and the institutional review boards of the three participating hospitals. The study was registered at the Netherlands Trial Register (NTR6407). A written informed consent form was signed by both parents, and by the child if ≥12 years old.

| Study population
Children (<18 years old at study) admitted in a Dutch burn centre between 08/2011 and 09/2012 were selected from the Dutch Burn Repository R3. 17 As this is a 5-7 year follow-up study, we included children <13 years old at burn injury. The rationale for this selection is that children up to age 18 are treated as children in Dutch health care.
As only a limited number of Dutch paediatric burn patients have severe burn injuries (about 22 each year), we extended this sample with children with severe burns (>10% TBSA burned if aged <10 years old at burn, >20% TBSA if aged ≥10 years old, or more than 5% full thickness burns 18 ) admitted between 01/2010 and 03/2013 to elucidate scar quality after severe burns. Patients were not eligible when parents were unable to answer the questionnaires, the patient was deceased, or when contact details were missing.

| Study procedure
Parents of eligible children were invited to participate through a postal letter containing an information leaflet, an informed consent form and the first survey. A second survey was sent after receiving informed consent, and children with severe burns were invited for scar assessment and physical measures at the outpatient clinic. Surveys were filled in by a parent(s). If no response was obtained, parents were called after 3 weeks to discuss participation. A postal reminder was sent if parents were not reached by telephone or did not return the informed consent form and survey after promising they would.

| Study parameters
The first survey included the patient scale of the Patient and Observer Scar Assessment Scale (POSAS) 2.0. 19 This scale assesses six scar characteristics: pain, itching, colour, pliability, thickness and relief, and includes an overall opinion item. Parents completed the POSAS for their child's-in their opinion-worst scar. We defined scars located on the hands/arms/feet/legs/face/neck as functional site scars. Scars on hands/face were indicated as visible scars. The six characteristics were scored between 1 and 10, with a score of 1 representing 'no differences with normal skin' and a score of 10 'very different to normal skin'. The POSAS score was calculated by summing up the six scores and dividing this by six. In case five of the six characteristics were completed, the POSAS score was calculated based on these five scores. The overall opinion item was measured on a 10-point scale, with 1 meaning 'as normal skin' and 10 'worst scar imaginable'. 20 The POSAS also includes an observer scale. 20 For children with severe burns visiting the outpatient clinic, two experienced and trained observers independently assessed the worst scar. The mean score of the observers was used as the observer score. Colour, thickness, relief and pliability scores of observers and parents were compared. A mean POSAS observer score was calculated by summing up the mean score of these four characteristic scores and dividing this by four. The same was done for the parent scores.
Other study parameters were extracted from the Dutch burn repository 17 and included patient characteristics (age, gender); burn characteristics (%TBSA burned, % full thickness burns, aetiology); and clinical characteristics (date of injury, number of surgeries, length of hospital stay, reconstructive surgery, artificial ventilation).

| Statistical analyses
A non-response analysis was done to compare children of participating parents with children of non-participating parents. All variables were checked for normality. If variables were not normally distributed, the median as well as the interquartile range (IQR) was reported. Continuous variables were compared with Mann-Whitney U-tests and categorical variables with chi-square tests, except for small numbers (n < 5), then the Fisher's exact test was used.
Long-term scar quality, including the six POSAS items, the POSAS score, and the overall opinion score, was assessed using descriptive statistics. Outcomes of children with mild/intermediate burns (<10% TBSA burned) were compared with children with severe burns (≥10% TBSA burned, or > 5% full thickness burns), and, parent scores were compared against observer scores.
To identify predictive factors of scar quality, univariate linear regression was used. Burn centre dependency was tested in a mixed model analysis. None of the univariate analyses were centre dependent; therefore, we used linear regression analysis. All variables with a p-value <0.10 in univariate analyses were checked for collinearity (>0.8 or <À0.8) and included in multivariate linear regression.
A p-value of <0.05 was considered statistically significant. Analyses were performed with IBM SPSS Statistics 25.

| Participants
In total, 289 children were selected, of whom 261 were eligible ( Figure 1). Half of the parents (n = 131) participated; including parents of 103 children with mild/severe burns (response rate: 49%) and 28 children with severe burns (response rate: 55%). Children of non-responders did not differ significantly from children of responders, except that children of non-responders were somewhat older (Appendix 1). Scars of 22 children with severe burns (79%) were assessed by observers; the other children were not willing to visit the outpatient clinic.
The median age of the included children was 2 years at burn (IQR = 1.0-3.0). Slightly more than half of the children (57%) were boys. Median %TBSA burned was 6% (IQR = 3-9), and median length of stay 5.0 days (IQR = 2.0-15.0). Most children (61%) did not undergo surgery. Scalds were most often (83%) the cause of burn and median time since burn was 5.5 years (IQR 5.4-5.8). The worst scar was often located on the arm (31%), trunk (23%) or leg (12%). The two subgroups (mild/intermediate vs. severe burns) differed significantly for most characteristics studied (Table 1). Of the twenty-two children who visited the outpatient clinic, one child had both legs amputated, five children had persistent contractors, and one child had eyelid and mouth deformities.

| Parent-reported scar quality
The median parent-reported POSAS score was 2.7 (IQR = 1.5-4.8) ( Table 2). For seventeen children (13%) a score of 1.0 was reported, meaning that the parents did not notice any differences with normal skin. These were sixteen children with mild/intermediate burns and   (Figure 2A,B).
Children in this group who underwent surgery had a worse POSAS score compared to children who did not have surgery (median score 4.3 vs. 1.8; Appendix 2). For the majority of children with severe burns, differences on pliability (96%), thickness (96%), and relief (89%) were reported. Itching and pain were respectively reported for 54% and 25% of these children. Especially high scores were reported for pliability and thickness; a score of ≥8 was reported for 29% of the children ( Figure 2B). Of the children with mild/intermediate burns, for 5% pain was reported, and for 16% itching. Pliability differed from normal skin in 64%, thickness in 67%, and relief in 69% of these children's worst scar ( A median score of 4.0 (IQR 2.0-6.0) was reported for the parent's overall opinion of their child's worst scar ( Table 2). The overall opinion did not significantly differ between the subgroups, though a trend was seen with higher scores reported in the severe group

| Observer scores versus parent scores
The    Length of hospital stay was highly correlated with %TBSA full thickness and %TBSA, and therefore not included in multivariate regression analysis.

| Predictive factors of long-term parentreported scar quality
scar quality on average 28 months after burns and included a subpopulation of children with a median TBSA of 5%. 7 Outcomes slightly differed from our study. The median score on colour, pliability, thickness and relief was one point higher (representing a worse outcome) in our study. There were small differences in the study populations, and also time since injury and setting were different. Time since injury might have led to a changed coping behaviour, and the setting (hospital vs. home) to an unintended influence by the clinician and/or socially accepted answers. Shorter after burns, parents might view the scar in the light of the earlier phases during the evolution of the scar, whereas 5-7 years after injury they might be more aware that the final outcome is reached.
Pain and itching were scored very low; the median of both was one, which is equal to no pain/itching. Both pain and itching are common problems experienced by burn patients, [21][22][23][24][25] which are also reported in the long-term. [26][27][28] However, an earlier study in children reported comparable low parent-reported pain scores. 8 This might be caused by the parent-reported base of the study as neuropathic-like sensations are hard to assess by parents. 8 Similar results were found for anxiety; burned children reported substantially worse anxiety levels compared to their parent-proxy outcomes. 29 In our study, it is thus unsure whether pain (and itching) are indeed experienced by only few children, or that these symptoms were underestimated by their parents. Child self-reported outcomes are important to fill this gap.
Parent-reported outcomes differed on average at least two points from observer reported outcomes. This finding is in line with earlier studies that described that patient perspectives differ from observer perspectives. 4 Multivariate analysis showed that number of surgeries was the only independent predictive factor for both POSAS and overall opinion score. This is in line with earlier studies that also found more surgeries to be a predictor of worse short-term scar quality. 9 30,31 Another proxy measure of burn severity; a greater burn size (%TBSA burned), also predicted a reduced short-term scar quality in previous studies. 4,8,9 In univariate analyses, %TBSA and other indicators of burn severity (full thickness burns and length of hospital stay) were found to be associated, but not in multivariate analyses. Results on whether full thickness burns is a predictor are inconsistent, an earlier study found that partial thickness burns were associated with a better scar quality, 8 whereas another study did not find this result. 12 Our study also showed that scarring on a functional site predicted scar quality in univariate analyses. As far as we know, this factor has not yet been studied as a predictor before.
Nevertheless, in clinical practice children with scars on a functional site seem to experience more contractures and a poorer scar quality, possibly due to a scar on a functional site being more often stretched. The findings of our study that children's age, sex and aetiology of the injury were not associated with scar quality are in line with earlier studies. 8,9,12 In the vast majority of children with mild/intermediate burns and severe burns, their worst scar differed from normal skin 5-7 years postburn. Informing children and parents about the final outcome is very important. An earlier study showed that many children expect that their scar will look as normal skin after scar maturation. 2 Also, a recent review reported that children and parental concern, and appearance are problems that are prevalent in the long-term. 3 It is thus important to counsel both children and their parents on the expectations of the final outcome of their (children's) scar(s). This is important for those with severe burns and who needed surgery as they were at a higher risk of a poorer longterm scar quality, but also other children and their parents should be counselled on realistic expectations.
This study contains strengths and limitations. Strengths include the multicentre approach, the relatively large sample size, the small amount of missing data, and the small differences between responders and nonresponders. Another strength is the use of the POSAS instrument; it is validated, includes all most relevant scar characteristics, and is the most frequently used scale. [32][33][34][35] A limitation was our inability to study other potential predictors, like skin type 12 and time to wound healing. 4,[9][10][11]13,14,36 Also, in children with multiple scars, only the worst scar was evaluated, which might have led to a slight overestimation of how scars differed from normal skin. Another limitation is the relatively small sample size to develop robust prediction models. For good prediction modelling large numbers are needed. 37,38 However, it is hard to collect large sample sizes in burns and our sample was relatively large for burns. 39 If available, combining existing datasets might overcome this problem. 40 Other limitations include the lack of information regarding scar treatment, which may have influenced scar outcome, and the fact that parentproxy outcomes were used instead of children's own evaluation.
Outcomes, in particular pain and itching, might be scored differently if children evaluated them themselves. However, to the best of our knowledge, no paediatric scar quality assessment instrument was available at time of study.

| CONCLUSION
Five to seven years after burns, the worst scar differed in most children from normal skin, with most and largest differences reported for colour. Children with severe burns had significantly higher scores on