Article Text

Download PDFPDF

Recovery position significantly associated with a reduced admission rate of children with loss of consciousness
  1. Sebastien Julliand1,
  2. Marie Desmarest2,
  3. Leticia Gonzalez3,
  4. Yolanda Ballestero4,
  5. Abel Martinez5,
  6. Raffaella Moretti6,
  7. Aristides Rivas7,
  8. Laurence Lacroix8,
  9. Armand Biver9,
  10. Emilie Lejay2,
  11. Lakshmi Kanagarajah10,
  12. Nancy Portillo4,
  13. Giovanni Crichiutti6,
  14. Chiara Stefani11,
  15. Liviana Da Dalt11,
  16. Daniele Spiri12,
  17. Patrick Van De Voorde13,
  18. Luigi Titomanlio2
  1. 1Pediatric Mobile Service for Emergencies and Resuscitation, Robert Debré Hospital, Paris Diderot University, Paris, France
  2. 2Pediatric Emergency Department, Robert Debré Hospital, Paris Diderot University, Paris, France
  3. 3Pediatric Emergency Department, Rio Hortega Hospital, Valladolid, Spain
  4. 4Pediatric Emergency Department, Cruces University Hospital, Barakaldo, Spain
  5. 5Pediatric Emergency Department, Barcelona University Hospital, Spain
  6. 6Pediatric Emergency Department, University of Udine, Udine, Italy
  7. 7Pediatric Emergency Department, Gregorio Marañon General Univerisity Hospital, Madrid, Spain
  8. 8Pediatric Emergency Department, Geneva University Hospital, Switzerland
  9. 9Pediatric Emergency Department, Centre Hospitalier Luxembourg, Luxembourg, Luxembourg
  10. 10Basildon University Hospital, Essex, UK
  11. 11Pediatric Emergency Department, University of Padova and Treviso Hospital, Italy
  12. 12Pediatric Emergency Department, Milan University Hospital, Italy
  13. 13Pediatric Emergency Department, Ghent University Hospital, Ghent, Belgium
  1. Correspondence to Dr Luigi Titomanlio, Pediatric Emergency Department, Robert Debré Hospital—Assistance Publique Hôpitaux de Paris (APHP), Paris Diderot University, Sorbonne Paris Cité, 48, Bld Sérurier— Paris 75019, France; luigi.titomanlio{at}


Background Loss of consciousness (LOC) is often seen in children. The response of caregivers to a child with LOC has been poorly investigated. Potential caregivers (parents, teachers) seem to have a poor knowledge of the recovery position (RP)—that is, the position into which an unconscious child should be placed in order to protect the airway.

Objectives To report the management and diagnoses of LOC in childhood, and to evaluate variables associated with an increased hospital admission rate.

Methods We conducted a prospective cohort study of consecutive children aged between 0 and 18 years diagnosed with LOC at 11 paediatric emergency departments (PEDs) of 6 European countries. The enrolment period was 3 months. Data were obtained from parental interviews, PED reports and clinical examination.

Results 553 children were enrolled. The most frequent final diagnoses causing LOC were seizures (n=278, 50.3%), and vasovagal syncope (n=124, 22.4%). Caregivers put the child in the RP in 145 cases (26.2%). The RP was independently associated with a significant decrease in the admission rate (aOR=0.28; 95% CI 0.17 to 0.48; p<0.0001).

Conclusions Our study demonstrates for the first time that the RP may reduce the admission rate of infants with LOC. Caregivers often perform inadequate manoeuvres when a child becomes unconscious. Campaigns aiming at increasing knowledge of the RP should be promoted.

  • Accident & Emergency
  • Comm Child Health
View Full Text

Statistics from

What is already known on this topic

  • Loss of consciousness (LOC) is often seen in children, but the response of caregivers has been poorly investigated.

  • The recovery position (RP), into which an unconscious child can be placed to protect the airway, is a well-known first-aid manoeuvre.

What this study adds

  • In this European study, the RP was performed in only a quarter of children with LOC.

  • The RP is significantly associated with reduced hospital admission of infants with LOC.

  • Campaigns aiming at increasing knowledge of the RP should be promoted.


Loss of consciousness (LOC) is a common symptom in the paediatric population, with as many as 15% of children presenting with at least one syncopal event before the end of adolescence.1–4 LOC has a wide variety of causes. Although often benign, it may be the manifestation of a potentially severe underlying cardiac, neurological or metabolic disorder.5 ,6 When parents see their child losing consciousness they are understandably shocked, many of them believing that their child may die.7 Most parents have inadequate knowledge of first aid,8 and, more generally, the level of first-aid knowledge among caregivers is low.9 ,10

The recovery position (RP) is a lateral recumbent position of the body, into which an unconscious child must be placed as part of first-aid treatment.11–13 The European Resuscitation Council Guidelines for Resuscitation recommend that “an unconscious child whose airway is clear, and who is breathing normally, should be turned on his side into the recovery position”.14 The basic principle of the RP is to protect the airway; the mouth is downward so that fluid can drain from the patient's airway, while the chin is up to keep the epiglottis open. Arms and legs are locked to stabilise the position of the patient.

The response and behaviour of parents and caregivers when a child becomes unconscious have not been reported. Although the RP is part of the guidelines of emergency care organisations, its effect on children with LOC is not known. Our primary objective was to study the potential effect of the RP on children with LOC by analysing its association with the hospital admission rate.



We performed a prospective, observational multicentre study of consecutive children consulting for LOC at one of the 11 recruiting paediatric emergency departments (PEDs) between 1 July and 30 September, 2014. Centres were European tertiary university hospitals from Spain, France, Italy, Luxembourg, Belgium and Switzerland. Written informed consent from parents and assent from the children was obtained when possible. The institutional review board of each country approved the study protocol.


We identified all new patients at each centre, aged 0–18 years, with either a history of LOC within the preceding 24 h or presenting with ongoing LOC, regardless of the cause. LOC was defined as an interruption of consciousness without response to stimulation, regardless of the length of that interruption.

Data collection procedures

Parents were asked to answer questions from a structured questionnaire translated into the language of each country. The questionnaire was administered face to face by the PED physician caring for the child. Physicians and parents were informed that the aim of the study was to explore causes of LOC in children and to describe the manoeuvres performed by caregivers during the LOC. Caregivers present at the PED were all interviewed in order to retrieve the most accurate information about the LOC. Demographic and medical information was collected by clinical examination and from PED data sheets.

The following clinical data were obtained by parental interviews and clinical examination: demographic data, coexisting chronic disease, duration of the LOC, clinical signs associated with the LOC, response of caregivers during the LOC (manoeuvres performed, including the RP, administration of drugs, call to the national emergency number), clinical parameters at PED arrival, final diagnosis of the cause of the LOC at the PED and outcome (immediate discharge from the PED, admission to a short-stay observation unit followed by discharge, standard ward admission, admission to a paediatric intensive care unit (PICU), death during the hospital stay). We also asked caregivers how they learnt the manoeuvres that they performed to reanimate the child, and asked the physician for his/her judgement about the potential usefulness or dangerousness of the performed manoeuvres for that child, as they were reported by the caregivers.

Statistical analyses

The primary objective was an assessment of the association between the RP and hospital admission rate as primary outcome. This primary outcome took into account all admissions, including short-stay observational unit admissions. Secondary objectives were the highlighting of other risk factors associated with this primary outcome, the analysis of factors related to setting the RP, and analysis of the association between the RP and prolonged hospitalisation (standard ward or PICU hospitalisation vs direct discharge from the PED or after an admission to a short-stay observational unit). To assess the possible bias due to different causes of LOC in younger children, we performed a subgroup analysis for patients aged either 0–2 years or ≥2–18 years.

Variables were compared by mixed logistic regression with a random effect for the enrolling centre for univariate analyses and multivariable models. The chosen level of significance was 0.05 (two-sided). All collected variables were analysed in a univariate model with both hospital admission and RP as reference outcomes. The final multivariable model included variables that were significant in the univariate analysis at the 20% nominal level. The excluded variables were all investigated as additions to the final model to verify that no association had been missed (factors that were only significant after accounting for others). Estimators used for multivariable models were adjusted ORs (aOR) and the 95% CIs.

Statistical analyses were performed with R (R Development Core Team) software.


Of the 101 926 patients consulting at the participating PEDs during the study, 557 (0.55%) fitted the inclusion criteria. Four families refused consent, and thus 553 patients were enrolled from the participating centres.

The clinical characteristics of these patients are shown in table 1.

Table 1

Characteristics of the 553 included children

In our series, LOC represented 0.55% of all consultations at the PED. The median age at PED visit was 3.2 years (centiles 0–25–75–100: 0.01–1.5–9.6–17.7 years). A coexisting chronic disease was identified in about one in five cases, of which epilepsy was the most common (n=41, 7.4%). Nearly half of all patients had had a previous episode of LOC. The median duration of the acute episode of LOC was 2 min (centiles 0–25–75–100: 0–1–5–480 min). In 33 cases (6%) the LOC lasted for more than 20 min. Among accompanying signs of LOC, cyanosis was only peripheral in 111 cases (20.1%) and generalised in 51 cases (9.2%). At PED arrival, 69 children (12.5%) still had impaired consciousness, with a Glasgow Coma Scale ≤8 in 19 children (3.4%), and rated from 9 to 14 in 50 patients (9%).

Caregivers put the child in the RP in 145 cases (26.2%). Drugs given by caregivers during the LOC were mainly antiepileptic drugs (intrarectal or intrabuccal benzodiazepines in 25 cases). In 293 cases (53.0%), manoeuvres other than the RP were made (details shown in table 1). The PED physician considered that the caregivers had performed potentially dangerous manoeuvres for the child in 95 cases (17.2%). Notably, children were shaken in 91 cases (16.5%). Infants younger than 1 year, for whom shaking can lead to severe cerebral injuries, were shaken in 25/82 (30.5%) cases, and in 18 cases (22.0%) the shaking was considered potentially dangerous. We also investigated how caregivers learnt the manoeuvres: RP training was primarily given by doctors (n=74, 51%) or obtained from first-aid training (n=29, 20%), whereas dangerous manoeuvres were primarily learnt from family (n=28, 29.5%) or media (n=17, 17.9%).

The most frequently reported final diagnoses causing the LOC episode were febrile seizures (n=147, 26.6%), non-febrile seizures (n=131, 23.7%) and vasovagal syncope (n=124, 22.4%). In children younger than 2 years, febrile seizures were the most common diagnosis (48.4%), whereas vasovagal syncope was the most common diagnosis in children aged >2 years (31.9%) (figure 1).

Figure 1

Most frequently reported diagnoses of loss of consciousness according to age groups.

After PED consultation, 43.8% of children were immediately discharged and 56.2% were admitted (33.5% to a short-stay observation unit, 17.5% to a standard ward and 5.2% to a PICU).

Six patients (1.1%) died during the hospital stay; these included one acute myocarditis, one septic shock, one foreign body aspiration, one traumatic brain injury and two cardiac arrests of unknown aetiology. Overall, missing data represented 3.8%, which did not concern performed manoeuvres or the outcome.

The results of univariate analyses of variables associated with the primary outcome (admission rate) are reported in table 2.

Table 2

Univariate analyses: variables associated with the admission rate

The variables associated with a RP setting by caregivers are reported in table 3. The RP was more often set in older children (≥2 years old), when they had a chronic epileptic disease and a history of LOC, when they presented a prolonged LOC (≥2 min) or abnormal movements and when the final diagnosis was a seizure.

Table 3

Univariate analyses: variables associated to a setting in the recovery position

The primary analysis (table 4) showed an independent association between the RP and a decreased admission rate with an adjusted OR of 0.28 (95% CI 0.17 to 0.48, p<0.0001). We found no statistical interaction between the variables—notably, between the RP and the age. Variables significantly associated with the admission rate were also the same when those variables not significant at the p<0.2 level in the univariate analysis were added to the final model.

Table 4

Multivariable primary analysis: association with the admission rate

Manoeuvres considered as dangerous by physicians were independently associated with an increased rate of admission, with an aOR of 2.12 (95% CI 1.19 to 3.78, p=0.01). Other variables significantly associated with a higher admission rate were LOC duration ≥2 min, impaired consciousness at PED arrival and vomiting. Vasovagal syncope was associated with a significantly lower admission rate.

The RP was still associated with a decreased admission rate when a longer hospitalisation was considered as outcome (conventional or PICU hospitalisation vs direct discharge from the PED or admission in a short-stay observational unit): aOR=0.43 (95% CI 0.21 to 0.88, p=0.02).


The aim of our study was to report caregivers’ behaviour when faced with a child with LOC, and to evaluate if their actions, use of the RP in particular, might influence the clinical outcome. Although LOC occurs frequently in childhood,2 its management by parents or other caregivers can be confused. We report results in children aged 0–18 years, whereas the only previous study on the subject focused solely on children aged >2 years.15 We chose to use the admission rate as the primary outcome measure because of the very low frequency of deaths after LOC and owing to the relatively high number of admissions of children with LOC across Europe. We show for the first time that the RP may reduce the admission rate in children with LOC.

Causes of LOC in our population differed according to age. In accordance with the literature, the most frequent diagnoses were febrile seizures (common in children aged <2 years) and vasovagal syncope (common in older children).16–18 In our series, 43.8% of children were immediately discharged after the PED consultation, reflecting the benign cause of LOC in about half of the cases. Only six patients (1.1%) died during hospital admission following an episode of LOC owing to the severity of underlying disease.

We found an association between hospitalisation rate and use of the RP for children aged <2 years (by 10-fold). Interestingly, this association was independent of other significant variables for admission, such as a diagnosis of seizures, Glasgow Coma Scale <15, and a longer duration of the LOC.

It is therefore possible that by putting a child into the RP leads to a better recovery by improving the child's breathing. This study, by its design, did not retrieve laboratory or radiological investigation results after PED admission, so that some data, including complications such as the occurrence of aspiration pneumonia, are not known. Future prospective studies recording clinical complications, laboratory and radiological investigation results are needed to explore the potential mechanisms of an association between the use of RP and fewer hospital admissions after LOC.

The RP is a simple manoeuvre which is commonly recommended in first aid for all unconscious people,14 in order to protect the airway against aspiration, which is a recognised cause of death in patients with epilepsy.19 Ideally, everyone should be able to position a child on his side after LOC. A study from Norway found that 4–5-year-old children can learn and apply basic first aid. Tested 2 months after course completion, 70% of the children assessed consciousness correctly and knew the correct emergency telephone number; 60% demonstrated correct assessment of breathing and 40% accomplished other tasks, including the RP.20 However, in our population we found that only 26.2% of the caregivers put an unconscious child in the RP. Many studies have already indicated the lack of knowledge of parents about first aid in children, but only two studies, which focused on febrile seizures in young children, have evaluated the parents’ ability to put the child into the RP (RP in 15% and 29% of cases).8 ,21 In children with a first seizure, 60% of parents thought that ‘the child was already dead’ and 72% ‘did not know what to do’,21 or believed that the child was dying.22 This stress can lead to inappropriate behaviour, as reported in a study which found that 20% of the mothers who witnessed their child having a febrile seizure put the child in warm water, held the child tightly, opened the child's mouth, hugged the child and blew air into the child's mouth.23

School teachers are also likely to encounter LOC in a child, but previous studies have shown that their knowledge of emergency care is often deficient24—notably, only 3.7% of them in one study had an adequate knowledge of first aid.9 In our own study, manoeuvres other than the RP were made in 53% of cases, and more frequently included shaking, putting water on the face, slapping and blowing on the face. Of these manoeuvres, 17.2% were considered by the PED physician to be potentially dangerous. Interviews of caregivers showed that whereas the RP was principally learnt from doctors or first-aid training courses, dangerous procedures were learnt mainly from family or the media. Also, to our surprise, considering the expected stress of the situation, only in about half of all cases was the national emergency number called, although this did not significantly affect the admission rate.

Our study has several limitations. First, the duration of the LOC may be overestimated, as this is primarily reported by parents or caregivers in a stressful situation in which an accurate assessment of time is often difficult. However, the aim of our study was to report frequent real-life situations. To limit a potential recall bias, we decided to retrieve the information during the PED consultation that followed the initial episode. Second, the outcome might have been biased by caregivers’ behaviour during the LOC. It is possible that parents who put their child into the RP made doctors more confident, thus lowering the admission rate, and vice versa for parents who carried out dangerous manoeuvres. To limit these biases, physicians were only informed that the aim of the study was to explore causes of LOC. All information about the management of LOC was retrieved without any particular focus being placed on the RP, which could potentially have caused bias. Additionally, physicians were unaware that the admission rate was an outcome measure. Third, we used a classic logistic regression model, which might have caused some problems of colinearity, instead of a recursive partitioning analysis to develop a multivariate prediction model.25 Another potential limitation is that we enrolled only children consulting at a PED, which might not be representative of all children presenting with LOC. However, we designed the study to serve as a proof of concept. We chose to perform a multicentre, multinational study so as to include a sufficient number of patients, thereby increasing the reliability of our findings.

In conclusion, our study showed that the RP may reduce the hospital admission rate of children with LOC, but often is not performed. Campaigns aimed at increasing the knowledge and skills of adults to deliver safe, effective first aid to children, and specifically those highlighting the importance of using the RP, should be promoted. These campaigns could lead to significant benefits for a child's health, including reduced morbidity and mortality, as well as reducing the costs of hospital admissions. Large, multicentre studies of children who have a LOC are needed to evaluate the reasons for the significant association between the RP and reduced hospital admission.


We thank all the families and physicians who agreed to participate in our study. We thank the Societe Francaise de Medecine d'Urgence (SFMU) for their support and the Department Hospitalo-Universitaire (DHU) Protect of Paris, France, and the Research in European Paediatric Emergency Medicine (REPEM) Network for their support.


View Abstract


  • SJ and MD contributed equally to this work.

  • Contributors SJ: statistical analysis, drafting and revision of the article. MD: acquisition of data, drafting and revision of the article. LG, YB, AM, RM, EL, NP, LL, AB, CS, DS, PVDV: acquisition of data, interpretation of data, revision of the article. LK, GC, LDD: interpretation of data, revision of the article. LT: conception and design, interpretation of data, drafting and revision of the article.

  • Competing interests None declared.

  • Ethics approval Institutional review board, Comité d’évaluation de l’éthique des projets de recherche biomédicale, Paris Nord, France.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles