Children in Sweden admitted to intensive care after trauma

Introduction 

Sweden has been regarded as one of the safest countries in the world for children. In a report from UNICEF in 2001, the death rate from injury of children aged 1–14 years in Sweden was the lowest across the member countries of the OECD, with an incidence of 5.2/100,000.14 Nevertheless, approximately 60 children under 16 years of age are killed each year by trauma,10 which is therefore the single leading cause of childhood mortality in Sweden as well as in other developed countries.5, 9, 10, 13, 14, 22, 23, 24, 34 Children admitted to paediatric intensive care units (PICUs) are considered to have the highest mortality and morbidity of paediatric trauma patients surviving to hospital admission.19 Studies have previously been mainly from the US; the present study is, to our knowledge, the first survey of this patient population in Sweden.

The aims of this study were to describe the demographics, injuries, mechanisms and severity of injury, prehospital and hospital care during the first 24h, and outcome among children treated at our PICU following trauma.

Methods 

We studied the medical records of 131 traumatised children (0–16 years of age), treated at the Queen Silvia Children's Hospital (TQSCH) in Gothenburg from January 1990 to October 2000; 45 of these children had multiple injuries. The selection was made from the patient database at the PICU at TQSCH. All medical records with trauma-related diagnoses were included. Drowning accidents and intoxications were excluded. The majority of children with head injury had been admitted to the neurosurgery department at Sahlgrenska hospital.

The medical records were examined regarding demographics, injuries, mechanisms and severity of injury, prehospital and hospital care during the first 24h, and outcome. Injuries and mechanisms of injury were registered according to the International Statistical Classification of Diseases and Related Health Problems (ICD-10). The severity of injury was estimated according to the Glasgow Paediatric Coma Score (GCS),31 Injury Severity Score (ISS),31 Triage Revised Trauma Score (T-RTS),31, 12 Revised Trauma Score (RTS),31, 12 Trauma Score Injury Severity Score (TRISS)31, 12 and Pediatric Risk of Mortality Score (PRISM).24, 34 All scores except PRISM were calculated according to data obtained at the time of arrival at hospital. PRISM scores were calculated from data at 24h after admission. Missing data were scored as normal. Collected data for prehospital care included type of transport to hospital, need for airway control, immobilisation (spine board, rigid cervical collar, vacuum mattress), intravenous infusion, administration of analgesia or other drugs and cardiac resuscitation. The data for hospital care during the first 24h after admission included hospitals where initial evaluation or treatment had taken place, time of arrival at the emergency department, airway control, radiological examinations, surgical procedures, fixation of fractures and need for and duration of mechanical ventilation. Time in the PICU and the total length of stay at the hospital were registered, as well as results of 30-day follow-up.

GCS (range 15–3), a neurotrauma score, was estimated using information from the medical records. Calculating GCS poses a special problem with intubated and sedated patients.16, 17 One patient in this study, intubated at arrival at the hospital, was registered as GCS unknown.

ISS (range 0–75) >10 is used to define high-risk patients; ISS>15 has been reported to predict a mortality rate of at least 10%.5, 6, 12, 23

T-RTS (range 12–0) is a physiological scoring system designed for prehospital use. T-RTS<12 has been proposed as the cut-off point for patients requiring trauma centre care.7, 12, 23, 25 The RTS (range 7.84–0) is designed for retrospective outcome analysis. RTS of 6 is considered predictive of >90% probability of survival.31

In our study, information about respiratory rate was missing in 26 cases (65%) at hospital arrival. This was dealt with by assigning a respiratory rate value of 4 (respiratory rate 10–29) to 13 patients whose breathing was considered normal or unaffected. A respiratory rate value of 3 (respiratory rate>29) was used for 9 patients with hyperventilation, and a respiratory rate value of 2 (respiratory rate 6–9) was used for 3 patients who were intubated and in deep coma. In one case, information about breathing was missing.

TRISS and ISS analyses (ranges 100–0%) were used to determine the probability of survival. Although TRISS involves coefficients derived from the predominantly adult Major Trauma Outcome Score (MTOS) database, TRISS has been validated for paediatric trauma outcome analysis.17, 23, 33

PRISM (range 0–100%) has been included to predict mortality, being the most widely used scoring system for estimation of mortality among patients at a PICU.5, 19, 24, 34 It is based on data from the first 24h after admission to the ICU. Not all variables need be measured for the sole purpose of mortality risk prediction; only the maximum physiological derangement is used for scoring.5, 24, 34 One or more PRISM score variables were not registered for each patient in this study (905 of 1965 or 46%), being assigned the PRISM score zero or normal. The variables GCS and respiratory rate were dealt with as above.

Registration, statistical analysis of collected data and calculation of RTS, ISS and TRISS were processed mainly with the support of KVITTRA, software developed for the national quality register of trauma patients in Sweden.20

Results 

During the period 1 January 1990 to 31 October 2000, 131 children were admitted to the PICU at TQSCH in Gothenburg following trauma: 89 were boys (68%) and 42 were girls (32%). The predominant age was 10 years and mean age was 7.9 years (S.D. 4.7 years), with girls in general younger than boys; 45 children had multiple injuries. Most of the accidents occurred at weekends and in June. In 72 cases (55%) the accident had occurred in Gothenburg.

Transport-related accidents were the main cause of injury, with cyclists predominating. The second most common cause of injury was falls, followed by sport and burns. Other causes of injuries in this study included suffocation, self-destructive action, assault and attack by mammals (Table 1).

Table 1. Mechanism of injury in 131 children with major injuries in Sweden

	Mechanism of injury	Number of children	(%)
	Traffic and transport related injuries	53	40
	Cyclist	21
	Pedestrians	14
	Passenger in vehicle	11
	Other	7
	Falls	44	34
	Sport	19	14
	Burns	10	8
	Other	5	4

In total, 270 injuries were diagnosed in the 131 cases. The most commonly injured body region was the head. All but five injuries were caused by blunt trauma. Penetrating trauma included injury with glass and animal bites; one infant was attacked by a polecat (Table 2).

Table 2. Injuries in 131 children with major injuries in Sweden

	Injured body region	Number of children	(%) of in total 270 injuries
	Head	66	17
	External	46	16
	Abdominal	43	14
	Face	36	13
	Thoracic	27	10
	Lower extremity	25	9
	Upper extremity	14	5
	Pelvis	9	3
	Spinal	4	1

GCS indicated that the majority of children had a normal or slightly depressed level of consciousness at the time of arrival at hospital (74 children or 56% with GCS 15–14); 70 children (53%) had T-RTS <12, which is suggested as the triage point for patients requiring trauma centre care. According to ISS, 83 children (63%) were considered high-risk patients because they scored >9, and 63 children (48%) were estimated to have a mortality risk of at least 10% as they scored >15. One child with spinal injuries scored ISS 75, and died shortly after admission to the PICU. Probability of survival was estimated as 90% or more at the time of arrival at hospital for 111 children (85%) according to TRISS, and for 100 children (76%) scoring RTS >5. Probability of death 24h after trauma was estimated as <10% for 129 of the children (98%) according to PRISM. Table 3 shows the range and median of each scoring system in this study.

Table 3. Scoring in 131 children with major injuries in Sweden

	Scoring system	Median	(Range)
	Neurotrauma scores
	Reaction Level Scale 85 (RLS)	2	1–8
	Glasgow Peadiatric Coma Score (GCS)	14	3–15
	Anatomical injury scoring systems
	Injury Severity Score (ISS)	14	3–75
	Physiological scoring systems
	Revised Trauma Score (RTS)	7.55	1–7.84
	Triage-Revised Trauma Score (T-RTS)	11	2–12
	Outcome analysis systems
	Trauma Score Injury Severity Score (TRISS)	98%	0.4–100%
	Pediatric Risk of Mortality Score (PRISM)	0.7%	0.2–21%

Most of the children arrived at the hospital by ambulance; prehospital information was missing in 19 of these cases. The most frequent prehospital proceedings in the remaining 59 children were intravenous infusion and immobilisation with a rigid cervical collar. Two children underwent intubation (Table 4).

Table 4. Prehospital management of 131 children with major injuries in Sweden

	Prehospital management	Children	(%) of 59 children transported with ambulance
	Intravenous fluid	24	41
	Cervical collar	19	32
	Analgesics	11	19
	Immobilisation of extremities	10	17
	Other drugs	9	15
	Oropharyngeal airway	5	8
	Spineboard	5	8
	Intubation	2	3
	CPR	1	2

In all, 49 children were initially assessed at a hospital other than TQSCH. Hospital management during the first 24h of admission included some form of radiological examination in all cases; CT predominated, followed by chest and skeletal radiography. Surgical procedures were performed in 64 cases, 11 of these at a hospital other than TQSCH.

The average length of stay at the PICU was 4.2 days, and corresponded to 19% of the overall hospital stay. The overall mean hospital stay was 22.6 days (Table 5).

Table 5. Hospital management the first 24 hours after admission of 131 children with major injuries in Sweden

	Hospital management	Children	(%)
	Radiology exam, CT	104	79
	Radiology exam, chest X-ray	67	51
	Intubation at ER	66	50
	Radiology exam, X-ray of skeleton	60	45
	Surgery	58	44
	Orthopaedic	22
	Neuro	14
	Plastic/Face	13
	Abdominal	10
	Vascular	5
	Minor	9
	Blood transfusion	45	34
	Radiology exam, ultrasonography	24	18
	Chest tube	6	4
	Fixation of fracture at ER	6	4

The overall survival rate was 97% (n=128), so the mortality rate was 3%. There were three deaths, all within 48h of admission to hospital. Causes of death were cerebral complications and multi-organ failure, and all involved multiple injuries in different transport-related incidents. One boy, 14 years old, was ejected from a car when it rolled over. He sustained multiple abdominal organ injuries, scored an ISS of 34, and died from multi-organ failure. However, his TRISS score was 52%, so death was unexpected (Table 6).

Table 6. Injuries and scores of the three deaths in 131 children with major injuries in Sweden

		Death 1	Death 2	Death 3
	Age (years)	0.8	14	11
	Gender	Female	Male	Male
	Mechanism of injury	Transport, passenger on bicycle	Traffic, passenger in car	Traffic hit by a tram
	Injured body region	Head and spinal	Multiple abdominal	Head, lower extr
	Scoring
	Injury Severity Score (ISS)	75	34	29
	Trauma Score Injury Severity Score (TRISS)	0.4%	52%	27%
	Pediatric Risk of Mortality Score (PRISM)	dead	2%	22%
	Cause of death	Transection of spinal cord	Multi organ failure	Intracranial edema

At follow-up 30 days after trauma, 79 children (60%) were at home and 26 children (20%) still in hospital. One child was transferred to another hospital, and two to a rehabilitation centre. In 20 cases follow-up information was missing.

Discussion 

In all industrialised countries, injury is the leading cause of death in children of 1–14 years of age, accounting for almost 40% of deaths in that age group.14 For every injured child who dies, many more live on with varying degrees of disability due to trauma.

Children admitted to PICUs have the highest mortality and morbidity of paediatric trauma patients surviving to hospital admission.19 In our study, the aim was to give a picture of this important patient group in an urban area of Sweden. The number of patients in this study was small, and accordingly extensive statistical processing was not possible.

In our study, boys were more frequently injured than girls, and transport-related accidents were the most common mechanism of injury, as reported in other studies.2, 8, 9, 17, 18, 23 Traffic is the most dangerous environment for children in Sweden as well as internationally. Although a great deal of trauma prevention has been carried out in Sweden since the 1950s, traffic is still the single leading cause of death among children in Sweden.10 All deaths in our study were caused by traffic and transport related accidents.

Different scoring systems have been used in this study to characterise the severity of trauma and estimate probability of survival. Except for the neurotrauma score Reaction Level Scale 85,27, 28 commonly used in Sweden, no scoring system was used routinely at TOSCO. Consequently, each score has been retrospectively calculated from medical records. In the present study, 48% of the children had an ISS>15. The three deaths scored high values according to the scoring systems based on the severity of injuries, but one of these deaths scored TRISS 0.53 (53%) and hence was defined as unexpected. When the scores are determined as recommended in the literature, diagnostic accuracy is about 80%.22, 26 Whether the systems including physiological variables are influenced by therapeutic interventions or not has been debated.22 In this study, we used the variables registered before therapy where possible. Most scoring systems have limitations, and the limitations of each scoring system have been considered in our estimate. However, this study did not aim to evaluate the different scoring systems.

An important indication of a trauma centre's performance is the outcome of the most severely injured trauma patients surviving to hospital admission, i.e. those admitted to the PICU.3, 19, 29 On completion of this study we found few other studies of this patient population with multiple injuries.

A US paper by Yian et al. reports a study of a population of 91 multi-trauma patients with orthopaedic injuries.33 The mortality rate was 4%, but causes of death were not presented.

In a German study, Ott et al. compared the prognostic value of different trauma scores in 261 paediatric patients with multiple injuries.23 The mortality rate in the German series was 27%. Causes of death were cerebral complications and multi-organ failure.

Treatment of head injury is considered key to survival in a paediatric population.1, 4, 5, 13 In Gothenburg, children with severe head trauma (particularly isolated head trauma) are either directly admitted or transferred to the department of neurosurgery at Sahlgrenska University Hospital. This is a plausible explanation for the lower frequency of head trauma in our study population.

An analysis of the deaths in our population showed that they all occurred within 48h of admission to the hospital, and that they could all be explained by the severity of the trauma. The outcome of trauma care in this study, measured in terms of mortality rate, showed that the Queen Silvia Children's Hospital in Gothenburg has a standard similar to that of other centres.23, 33

The major variable of outcome was patient survival. These scoring systems do not specifically address long-term risk of impairment, and therefore overlook one of the most crucial elements of paediatric trauma.15, 23, 30 The measurement of outcome of trauma care should incorporate not only survival, but also take into account quality of life, morbidity and disability.

A follow-up of the survivors in our study would present a more complete analysis of major trauma and outcome in a Swedish paediatric population.

This study is limited by its retrospective design. Specific concerns include determination of level of consciousness, and scoring in general. Missing data were assigned a normal value or left out, which may have resulted in scoring trauma as less severe than in reality it was.

Conclusion 

Major trauma with admission to a PICU is rare in a Swedish paediatric population. When these children are cared for at a centre with the necessary facilities and trained personnel, they have a good chance of survival.