Epidemiology and mortality in patients hospitalized for burns in Catalonia, Spain

Burn injuries are one of the leading causes of morbidity worldwide. Although the overall incidence of burns and burn-related mortality is declining, these factors have not been analysed in our population for 25 years. The aim of this study has been to determine whether the epidemiological profile of patients hospitalized for burns has changed over the past 25 years. We performed a retrospective cohort study of patients hospitalised between 1 January 2011 and 31 December 2018 with a primary diagnosis of burns. The incidence of burns in our setting was 3.68/105 population. Most patients admitted for burns were men (61%), aged between 35 and 45 years (16.8%), followed by children aged between 0 and 4 years (12.4%). Scalding was the most prevalent mechanism of injury, and the region most frequently affected was the hands. The mean burned total body surface (TBSA) area was 8.3%, and the proportion of severely burned patients was 9.7%. Obesity was the most prevalent comorbidity (39.5%). The median length of stay was 1.8 days. The most frequent in-hospital complications were sepsis (16.6%), acute kidney injury (7.9%), and cardiovascular complications (5.9%). Risk factors for mortality were advanced age, high abbreviated burn severity index score, smoke inhalation, existing cardiovascular disease full-thickness burn, and high percentage of burned TBSA. Overall mortality was 4.3%. Multi-organ failure was the most frequent cause of death, with an incidence of 49.5%. The population has aged over the 25 years since the previous study, and the number of comorbidities has increased. The incidence and severity of burns, and the percentage of burned TBSA have all decreased, with scalding being the most prevalent mechanism of injury. The clinical presentation and evolution of burns differs between children and adults. Risk factors for mortality were advanced age, smoke inhalation, existing cardiovascular disease, full-thickness burn, and high percentage of burned TBSA.


Burn characteristics.
The mean burned TBSA was 8.3% (Table 1) full-thickness burns were diagnosed in 23.4% of adults and 7.3% of children.In 75.1% of patients, less than 10% of TBSA was affected.In 9.7% of patients, > 20% of TBSA was affected; in adults, this incidence (10.6%) was double that of children (p < 0.001).In all age group, burns in the arms (64.8%) were most frequent, followed by the legs (51.8%), and the head and neck (35.3%).
Scalding (37.7%) was the most prevalent mechanism of injury overall; however, after adjusting for age and sex, the most prevalent mechanism of injury in adult males was flame (39%), and scalding in adult females and in children (45.9% and76.6%,respectively); 14.1% of burns occurred in the workplace.Summer was the season with the highest number of admissions (28%).
Treatment.Table 3 summarizes the treatments administered.Over half (63.8%) of the sample underwent surgery; the difference between children (41.4%) and adults (68.9%) was statistically significant (p < 0.001).Of the surgical patients, 19.5% required a median of 1 (IQR 1.2) reintervention.Median time from the burn event to surgery was 16 [IQR 12-21.8]days in children and 14 [IQR 9-20] days in adults (p < 0.001).In contrast, Table 4 shows that the median waiting time for surgery is 10 days [ IQR 7-15] days, compared to patients with minor burns, with a median of 15 days (p < 0.001).
Complications/Outcomes. Some (13%) patients required admission to the ICU; this percentage was significantly higher in children (25.9%) than in adults (10.7%) (p < 0.001), although children remained in the unit for a median of 1 (IQR 1.4) day and adults for a median of 13 [IQR 3.30] days (Table 5).
Mechanical ventilation was required in 3.9% of children and 9.4% of adults (p < 0.001).Median time on mechanical ventilation was significantly lower in children (6 days [IQR 3.15.5])compared to adults (10 days  [IQR 2.25]).
In adults, the most prevalent complications during hospital stay were sepsis (16.6%),AKI (7.9%), and cardiovascular complications (5.9%).Among children, the most frequent complication was sepsis (14.9%).The  Overall mortality in the study period was 3.5%; the age-adjusted the mortality rate in children was 0% vs. 4.3% in adults.Table 6 describes the clinical differences between survivors and deceased.
The most frequent cause of death was multiple organ failure (MOF) (49.5%); 9.5% of deaths were due to sepsis.Mortality at 30 and 90 days was 2% in the group of adults, while no children died during post-discharge follow-up.At 30 days, 148 patients had been lost to follow-up, and 235 at 90 days.The mean age of patients who died was 68.6 years.The mean LOS in our population was 14.6 (± 20.4) days; this was significantly lower in children (11.8 [± 14.8] days).The median LOS/TBSA ratio was 1.9 for adults and 1.5 for children.In the univariate analysis, the Kaplan-Meier survival curve was calculated (Figs. 2, 3, 4, 5, 6, 7, 8, 9 and 10) for each qualitative variable (Table 7).

Economic factors.
In our study, mean annual expenditure was €5,030,043.85, with a mean cost per patient of €15,179.31.

Discussion
This is the first study to describe the epidemiological profile of children and adults presenting with burns in Catalonia.It is also the first to describe the overall in-hospital complications presented by patients admitted for burns, since these complications are usually analysed in a subgroup, such as major burn patients.
The mean age in our study was 41.4 years (49.7 years in adults).This is similar to Palacios 31 and higher than the earlier study published by Barrett 21 , and reflects a global trend due to worldwide population ageing.
The proportion of burns in children (18.1%) is lower than in other developed countries 7,28,32,33 and far lower than that reported in developing countries 11,34 .One of the reasons for this low incidence is the consistently low birth rate in Spain over the past few decades and compulsory school education.Interestingly, the age range with the highest incidence of burns is usually 0-4 years, in some cases accounting for up to 50% of all admissions 32,[35][36][37][38] .In our study, the incidence in this age group was 68%.This percentage, though similar to that reported in other studies, is still very high, and suggests that the authorities should promote health education and implement preventive measures in this age group (including their parents).A relatively simply measure would be to introduce legislation to lower the maximum permitted temperature of household water, an approach that has proven effective in other countries 39,40 .

Characteristic of burns.
Scalding was the most frequent mechanism of injury (37.7%) in our review, whereas flame was the most prevalent cause of burns in the 1990s.After adjusting for age and sex, flame was the most prevalent mechanism in adult men, and scalding was the most prevalent mechanism in adult women and in children.These results are in line with studies published in other countries, such as Portugal 41 , Israel 29 , China [42][43][44] , Switzerland 9 , and Holland 7 ..This reduction in burned TBSA may be due to a lower incidence of flame burns, which are usually severe, extensive, and accompanied by smoke inhalation 45 .
The proportion of patients with burns covering less than 10% of total body surface has increased from 60% in 1995 to 75% today.This is a global trend that has also been reported in other studies 7,16,43,46,47 , and could be due to the decline of flame as the main cause of burns in our setting.For the same reason, the number of patients with 20% of TBSA affected has decreased to 9.7%.According to figures from the ABA (American Burn Association), 22.5% of burn patients in 2013 presented major burns 48 .In children, 77% of patients presented burns affecting less than 10% TBSA, in line with other studies 28,35,36,41,49 .

Comorbidities.
Our study provides an overview of comorbidities in burn patients and their effect on mortality.Smoking is the most prevalent comorbidity (25.8%), while the rate of cardiovascular disease (24%) is similar to Barret et al. 21, and significantly higher than that reported by other authors, such as 8% in Knowlin 50 .The percentage of in-hospital cardiovascular complications was also higher in our series compared with Knowlin (5.9% vs. 3%), while cardiovascular comorbidity increased mortality at a rate similar to that observe by Knowlin 50 (HR 2.12 [95% CI 1.34-3.34],p = 0.001).Alcohol abuse was reported in 6.9% of patients in our study, a high rate compared to the average of 4.2% in the Spanish population in 2019 51 .These figures are consistent with those reported by Eiroa-Orosa, who showed that a higher proportion of burn patients had a history of substance abuse compared with the general population 52 .

Variables
In Spain, according to the ENPE study, 22% of the population is obese, and obesity is more prevalent in men aged over 65 years from low-income groups 53 .In Catalonia, the percentage of obesity (16%) is lower than the Spanish average.In our series, 34.5% of adult patients were obese, a proportion that contrasts with most reports in the literature; however, according to the Center for Disease Control, 42.4% of the US population was obese in 2017 54 , while Jeschke reported a prevalence of 29% 55 .In our population, obesity was not associated with increased mortality.
Treatment.Surgery was required in 64% of our patients, a percentage that is higher than other studies, such as Dokter 7 .The significant difference in the rate of surgery observed between adults and children is consistent with other studies, in which burns in children are usually less severe and respond well to conservative treatment, particularly high protein intake.

Complications.
None of the studies identified in our database search summarise the most common complications during hospital admission.www.nature.com/scientificreports/Several authors have described AKI as an independent risk factor for mortality [56][57][58][59][60] .This complication arose in 8% of patients in our study, whereas an incidence of 30% has been reported in other studies 60 with an 80% mortality rate 56,61 .However, it is important to bear in mind that nearly all these studies were performed in critically ill patients 17,62 , so the findings cannot be extrapolated to the general population.Burn patients are at increased risk of suffering potentially fatal infection of any cause in the first 5 years after the burn 63 , and sepsis increases LOS and ICULOS 64 .In our study, 16.3% of patients were diagnosed with sepsis, somewhat higher than the figures published in Belgium 65 .
Outcomes.Mortality remained at 3.5% (0% in children under 16 years of age and 4.3% in adults).This rate is somewhat higher than that reported in most developed European countries 6,10,15,65,66 .
We compared observed vs. predicted mortality based on the ABSI to determine the accuracy of this score in our population 67 .The ABSI correlated with observed mortality, except in mild injury patients (ABSI < 6) and very severe injury patients (ABSI ≥ 12), in which mortality was lower than predicted.
A factor that has changed substantially over the years is cause of death.In Barrett et al., the main causes were acute respiratory distress syndrome (ARDS) (34%), MOF (26.8%) and sepsis (13.2%).The decrease in ARDS (13.7%) is striking, and may be due to improvements in critical care, a decrease in flame burns, and advances in mechanical ventilation.The incidence of sepsis-related deaths has also declined, and it is now the cause of death in 9.5% of patients, 50% lower than the figure reported by Barret et al. 21and in line with the findings of a Belgian study 65 .The decrease in mortality due to sepsis is a promising development, and may be due to the implementation of strict care protocols in critically ill patients 68 , together with improvements in the use of antibiotics and in the management of septic shock.www.nature.com/scientificreports/ In our study, the main cause of death was MOF (49.5%).According to the American Burn Association, MOF is the cause in 27.5% of fatalities 48,69 .In other studies, this percentage increases to 40%. 17,70.
The main predictors of in-hospital mortality identified in this study are age over 80 years, ABSI ≥ 7, smoke inhalation, cardiovascular comorbidity, full-thickness burn, and burned TBSA.These factors have also been described as predictors elsewhere 8,27,33,43,46,50,71,72 .Interestingly, there is scant evidence elsewhere of the association between cardiovascular comorbidity and increased mortality; however, in our series, this comorbidity increased the probability of dying by 112%.We, like some other authors, did not find gender to be associated with increased mortality 8,73,74 , although this contrasts with the findings of other studies 67 .
The number of admissions for burns has decreased dramatically from 6.6/10 5 population/year to 3.68/10 5 population/year This is a global trend 1,2,15,33,75 , despite significantly higher rates of admission reported in some of our European neighbours, such as 18.9/10 541 in Portugal, and 36.9/10 5 in Romania. 11Both LOS and LOS/TBSA in Catalonia are above the average in developed countries 2,4,6,11,13,71 , possibly due to 2 factors: first, the average waiting time from burn event to surgery is 18 days, far higher than the average 14.7 days in the Netherlands 7 ; second, home care in our region is underdeveloped, a situation that places an additional burden on in-hospital care.
In total, 14% of adults were admitted for burns occurring in the workplace, similar to the percentage reported by Barrett, and midway between the 5.9% reported by Palacios 31 and 20.9% by Sánchez 25 .Other countries that have published statistics relating to burns in the workplace include Germany, with an incidence that ranges from 18%-33.7% 12,76 , Switzerland with 31% 9 , the USA with 18% 77 , China with 78% 72 , Australia with 17% 74,78 , and Austria with 14.9% 8 .

Economic factors.
Care costs are currently one of the main problems in the healthcare sector.The economic downturn followed by the SARS-CoV-2 pandemic has brought public health systems in Spain to the brink of collapse, and it is now more important than ever to optimise resource management.As mentioned above, mean annual expenditure was €5,030,043.85, with a mean cost per patient of €15,179.31;these cost estimates were not TBSA-weighted.In Portugal, meanwhile, the cost per patient is €8,030 41 , in Holland it is €26,540 30 and in Finland, €25,000 3 .As our study was performed in a single hospital, our findings cannot be assumed to reflect the situation in other regions; however, this single-centre design ensures that both the study population and the treatment of burns is homogeneous.
Selection bias cannot be ruled out, since patients with non-severe burns are not initially treated at or transferred to the HUVH, so we were unable to include these data in our analysis.Another potential source of selection bias stems from our retrospective design, insofar as some patients were lost because doctors did not always strictly adhere to the 90-day follow-up schedule.
In the study of comorbidities, patients were not assessed using the gold-standard Charlson index, and the presence of psychiatric disorders was not recorded.

Conclusions
This study shows that changes have occurred in the pattern of burn injuries, their extent, and their severity, and burns are now predominantly less extensive and deep in our setting.The pattern of clinical presentation differed been children and adults.In our series, children ages between 0 and 4 years of age account for 68% of all children admitted for burns.This means that they are still most important risk group and should be the target of preventive measures and health education campaigns.The factors associated with a higher risk of mortality were age, %TBSA, full-thickness burns, smoke inhalation, and cardiovascular comorbidity.Unlike other studies, in our series female sex was not a risk factor for mortality.Prospective, multicentre studies are needed to obtain a more accurate picture of the situation of burn patients in Spain.

Figure 10 .
Figure 10.Kaplan-Meier survival curve for delay time for admission.

Table 7 .
Cox Regression analysis of factors associated with mortality.CI Confidence Interval.

Table 8 .
Multivariate analysis of factors related to mortality (adjusted for clinical factors).R 2 Nagelkerke: 0.267 CI: Confidence Interval.Significant values are in bold.

Table 9 .
Multivariate analysis of factors related to mortality (unadjusted).