Epidemiology And Outcomes Of Radiator Burns At A High Throughput Burns Centre

Summary

Indoor radiator burns can cause significant morbidity and mortality, especially in vulnerable patients. However, the epidemiology and clinical outcomes are poorly characterized. A retrospective study of all radiator burns referred to a tertiary regional burns centre between 2013-2019 was conducted. Four hundred and forty-seven patients (median age 25.6 years, male:female ratio 1.4:1) were referred; 109 (24%) admitted, 201 (45%) managed in outpatients and 137 (31%) managed locally. The incidence of burns leading to referral was 0.65/100,000/year, but increased annually. Contact burns consisted of 99.6%. Age distribution was bimodal: <5 years (43%), >65 years (27%), although all ages were affected. Median TBSA was 0.75% (0.1-11.5%), but 79% had injuries <2%. Childhood burns were predominantly superficial dermal grab injuries, usually manageable as outpatients. Elderly patients had larger burns usually secondary to falls or impaired sensorium, and were more likely to die (p<0.05). Burns due to impaired sensorium were deeper and more likely to require surgery (p<0.05). Sixty-three (14%) patients required surgery. Thirty-day mortality was 1.1%. Age was the only significant predictor of mortality. This study quantifies the epidemiology and outcomes of a growing problem. It identifies at risk populations (extremes of age), and important features of the clinical assessment.

Introduction

Central heating is used by over 95% of households in the United Kingdom (UK) and United States (US).^1,2 The majority of UK households still rely on gas-fueled, wall-mounted radiators which are positioned on the reachable parts of wall surfaces and underneath windows.³ Radiator surfaces can reach 60-80○C.⁴ This is important to note since contact with a surface with a temperature >70○C for less than one second can cause partial thickness burns.⁵ Radiators are therefore a particular risk for contact burns but can also cause scalds when being bled,⁶ or even radiant heat burns.

There is more literature about burns resulting from vehicle radiators^7-13 than those from indoor radiators. Studies discussing radiator burns include three small British clinical studies of five, thirteen and fifty patients.^4,14,15 Only one of these studies included radiator-related burns alone.⁴ An epidemiological study in Avon, England found that of 166 children sustaining burns in a longitudinal follow up, 12% were caused from radiator contact injuries.¹⁶

One US study reported outcomes in ten children sustaining burns from standard hot water radiators, but also the now largely historical steam radiators.¹⁷ One Greek group studied all traumatic injuries from domestic radiators, of which burns formed only 2.4%.¹⁸

One published study included all ages but its sample size of only 50 patients limited the strength of conclusions that could be drawn.4 Moreover, the short study periods precluded formal analysis of annual trends in radiator burn incidence. The publication of several newspaper reports suggests there is a public interest in these injuries.^19-23

There is a gap in the literature for a highly powered study of the epidemiology and outcomes of radiator burns.¹

The authors’ institution is a regional burns centre covering a population of 9.8 million in England. This allows the capture of large cohorts from which definitive answers to clinical questions can be offered. The aims of this study were to characterize the epidemiological features and clinical outcomes of patients treated for indoor radiator burns in a high throughput setting. Determinants of surgical outcome were also analyzed.

Methods

A single-centre retrospective analysis of medical records was performed. The unit’s burn injury databases were searched in January 2020 for referrals containing the term “radiator*” between 2013-19. All patients sustaining burns from indoor radiators were included. Data were collected from prospectively held databases and electronic medical records. Fig.1 summarizes the demographic data, burn characteristics, treatments and clinical outcomes that were recorded.

Fig. 1. Demographic data, burn characteristics, treatments and clinical outcomes recorded.

Statistical analysis

Descriptive statistics were used to analyze patient demographics: medians for continuous data with skewed distribution and absolute value (proportion) for categorical data. The relationships between burn mechanisms and age and burn depth and level of sensation/consciousness were investigated using Chi-squared tests. Stepwise multiple logistic regressions were used to analyze risk factors for hospital admission and 60-day mortality. Probability values of <0.05 were considered significant.

The study did not require ethical approval as this was a retrospective service evaluation with no change to patient pathways (audit number CA19-196). All data were handled according to Caldicott principles.²⁴

Results

There were 9952 referrals to the unit during this sevenyear study period. Of these, 447 patients (4.5%) sustained indoor radiator burns. Considering the centre’s catchment population of 9.8 million, the incidence of radiator burns in the East of England leading to secondary care consults was an estimated 0.65/100,000 population/year. There were 264 males and 183 females (1.4:1). The median age was 25.6 years (range one month to 100.6 years). There was a bimodal age distribution with the most affected age groups being <5 years (43%) and >65 years (27%) (Fig.2). There was a steady annual increase in the number of radiator burn referrals over the study period (Fig.3). There was a median of 40 referrals/month (range 3-76). Most injuries occurred in the winter months between December and February (45%). However, 4.7% occurred during summer (Fig.4).

Fig. 2. Number of referrals per age category.

Fig. 3. Number of referrals per year.

Fig. 4. Number of referrals per month.

Vastly more patients sustained contact burns than were scalded (99.6% vs. 0.4%). There were no radiant heat burns. The most common mechanisms of contact burn were falls and accidental grab injuries (36% and 34% respectively). Children under five were significantly more likely to sustain burns due to accidentally grabbing radiators (p<0.001), whilst over 65s were the most likely to sustain a burn following a fall (p<0.001) (Table I). There was no significant variation in mechanism by gender. All mechanisms of contact burns are illustrated in Fig.5. There were three recorded burns secondary to assault, one in a 33-yearold woman and two girls under five. All burns secondary to assaults were at least deep dermal depth, and were located on extremities.

The median total body surface area (TBSA) was 0.75% (range 0.1-11.5%), with the majority of patients sustaining injuries <1% (50%) and 1-2% (28%). Seizures and falls accounted for 86% of burns >4% TBSA (Table II). There was an average of 1.2 anatomical regions injured per patient: the most common being hands (35%), forearm/arm (20%) and trunk (14%) (Fig. 6). With the exception of grab injuries, scalds (both affecting the hands) and peripheral neuropathy (primarily affecting the feet), all other mechanisms mainly caused injury to the arms/forearms (Table III).

Burn depth was associated with injury mechanism (Table IV). Grab injuries and scalds predominantly caused superficial dermal burns (p<0.001), whilst injuries associated with impaired sensorium (intoxication, peripheral neuropathy, seizures and deep sleep) were significantly more likely to cause full thickness injuries (p<0.001) (Table IV).

Overall, 109 (24%) patients had at least one admission (range 0-3), including five who had admissions to the intensive care unit (ICU) for postoperative monitoring or comorbidity management. There were 201 (45%) patients managed as outpatients and 137 (31%) that were minor enough for local management to be advised. Total burn surface area, intoxication and peripheral neuropathy were all positively associated with admission (p<0.05), whilst grab injuries usually did not require admission (p<0.05) (Table V). No significant associations were observed between age, interval to referral or other mechanism of injury and hospital admission. Common comorbidities for admitted patients included epilepsy (14%), diabetes mellitus (11%), psychiatric disorders (6.7%) and dementia (3.8%).

Sixty-three patients (14%) underwent 84 surgical procedures (average 1.33 operations per patient, range 0-7). The greatest surgical burden was seen in patients who sustained burns secondary to seizures, intoxication or peripheral neuropathy (35%, 33% and 29% respectively) (Table VI). The types of surgical procedure are detailed in Table VII. The median length of hospital stay (LoS) was 2 days (range 0-70). The median number of dressings clinic appointments either following discharge or as a primary treatment location was 2 (range 0-32).

Fig. 5. Relative contribution of each mechanism to contact burn injury.

Fig. 6. Body map showing proportion of burn injuries in each anatomical area.

Table I. Mechanism per age group.

Table II. Mechanism per TBSA.

Table III. Anatomical burn injury location per mechanism.

Table IV. Burn depth per mechanism.

Table V. Multiple logistic regression for variables affecting hospital admission.

Table VI. Number of patients requiring surgery per mechanism.

The 30- and 60-day mortality rates were 1.1% and 1.8% respectively. All of the eight patients who died in the 60-day post burn period were aged 55+, with a relatively equal gender split (1.7:1 male:female). Four of the patients had severe comorbidities, including ischaemic heart disease, cerebrovascular disease and dementia. The burns were all at least deep dermal in thickness and all occurred following a fall. All patients who died were conservatively managed, with no in-hospital deaths. The cause of community deaths could not be ascertained. Increased age was the only variable found to be significantly associated with 60-day mortality upon stepwise multiple logistic regression (p<0.05).

Discussion

This is the largest study of indoor radiator burns to date. It quantifies epidemiological, aetiological and clinical outcome data from one of the largest burns centres in Europe.

The median TBSA of 0.75% (range 0.1-11.5%), which is similar to previous estimates of 1.6%,4 belies the importance of these injuries to vulnerable populations. The results of this detailed analysis are important to both clinicians and policy-makers.

The incidence rate of radiator burns found in this series is relatively low (0.65/100,000/year). As an illustration, 0.65/100,000/year is equal to the incidence of systemic lupus erythematosus in the UK.²⁵ Although a 1993 Centre for Disease Control (CDC) survey found 1881 children burned by domestic radiators in the US,²⁶ there is no previous literature on radiator burns incorporating all age groups on which to base temporal comparisons. Despite this low overall incidence, there was a steady annual increase in the number of referrals over the seven-year study period (Fig. 3). Radiator burns currently comprise 4.5% of all referrals, but should the increasing trend continue, in parallel with an ageing population, these injuries may become a more significant part of the burns take.

Further analysis of the cases reveals that although all ages are affected, those at the vulnerable extremes of age such as children <5 years (43%), the elderly (27%), and the comorbid are at increased risk (Table I). This bimodal age distribution, and the involvement of all ages, was also observed by Harper and Dickson in their 1996, 3-year retrospective review of 50 radiator burns in Wales.⁴

This study found a mean age of 43.4 years, which is older than the median age of 25.6 years in the present study. This is likely because the present study included all patients, whereas the Welsh study included inpatients only. Previous studies have also highlighted the particular burn risks of radiators to children.^18,26 Similarly, Hurren and Dunn recognized the risk of radiators to the elderly in a 1993 report.¹⁵

The high incidence in toddlers is largely explained by children undergoing the “pull to stand” phase of development where they use furniture and appliances for support.^27,28 This is in keeping with the finding that the majority of childhood radiator burns are grab injuries (p<0.001) (Table I), which largely cause superficial dermal injuries (Table IV) of <1% TBSA to the hands since this population has intact reflexes (Tables II and III). However, the publication of national news reports on these seemingly minor injuries shows public concern at any avoidable injuries to this age group.²² Deeper, more extensive injuries in children raise the possibility of diseases of altered sensorium and non-accidental injury. Only two children in this series had been assaulted and received the appropriate medicosocial management.

The elderly, although affected less than children, are also disproportionately burned by radiators for reasons including higher comorbidity levels (e.g. dementia and diabetes) and reduced reaction times. Similar elder susceptibility has been discussed in previous small case series.^4,14,15

Falls are the leading mechanism of injury (Table I), in keeping with previous reports.¹⁵ Falls cause burns which affect the widest range of anatomical locations, only sparing perinea (Table III). Almost 75% of burns secondary to falls are deep dermal and full-thickness (Table IV). Falls also cause most radiator burns >1% TBSA (Table II). Falling onto radiators in the elderly may be due to loss of consciousness, and can cause entrapment, explaining the greater extent and depth in the population. Concordantly, 21% of patients presenting after falls require surgery due to the deeper, larger burns.

This is also a high perioperative risk group owing to comorbidity and reduced physiological reserve. These factors, and the finding that increased age was the only variable predictive of mortality (p<0.05), demonstrate the seriousness of radiator burns in this cohort. Assessment must be meticulous to identify cases of neglect, and documentation accurate since these cases have led to litigations.²¹

Another important cause of radiator burns is impaired sensorium (21%), as also identified in the literature.^29-31 All age groups are affected by intoxication, peripheral neuropathy, seizures and deep sleep burns to varying extents.

The ensuing burns were often deeper (full thickness [FT]) (p<0.05) (Table IV) and more extensive (Table II). Increased injury severity was also reflected in hospital admissions, with burns secondary intoxication and peripheral neuropathy significantly more likely to require inpatient hospital stay and surgery (p<0.05) (Tables V and VI). The surgical burden for radiator burns in this study was relatively high, with 14% of referred patients undergoing surgery, of which 80% required skin autografting.

Falls and impaired sensorium had the greatest surgical burden, accounting for 87% of all procedures performed. These figures are higher than for hot water bottle burns, which the authors’ unit previously reviewed,³² reflecting a higher injury severity in these vulnerable patient cohorts. In contrast, the median LoS was low at only 2 days, since many patients were young with injuries <2%.

This study revealed 30- and 60-day mortality rates of 1.1% and 1.8% respectively, with no inhospital deaths. All of these patients were aged over 55, with over half having significant comorbidities.

We believe this is a low mortality rate, ascribable to the relatively low TBSA observed in radiator burn injuries. However there are very few studies which give acute mortality estimates,^33,34 therefore precluding direct quantitative comparisons.

In order to tackle the rising incidence of radiator burns, interventions need to reduce radiator surface temperatures whilst maintaining mechanical efficiency. British standards for radiator heat output are based on boiler outflow temperatures of 75°C.³⁵ Surface temperatures can therefore exceed 70°C – sufficient to cause a partial thickness burn in less than 1 second.⁵ For this reason, engineering regulations dictate a maximum surface temperature of 43°C for new radiators in NHS institutions, schools and care homes.^36-38 However, this is not enough, since patients with impaired sensorium are at risk of prolonged contact, and can still burn at this temperature.

Other measures to complement temperature regulation include installation of protective casings, repositioning beds away from radiators, regular supervision of vulnerable persons and educational public health campaigns. Surgeons are ideally placed to lead these efforts.

The main limitation of this study was that patients not referred to secondary care were not captured in our dataset. This may have led to an underestimation of radiator burn incidence.

The location of radiator burn injury could not be reliably ascertained due to incomplete recording. Notwithstanding these limitations, this is the first study to attempt to ascertain the incidence of radiator burns, and therefore is important.

Conclusion

This study quantifies the epidemiology and outcomes of radiator burns. It identifies at risk populations, important features of the clinical assessment and treatment considerations. The burden of radiator burns could be reduced by preventative measures such as protective covers, low surface temperature systems and public health campaigns.