Article Text

Download PDFPDF

Perioperative laryngospasm management in paediatrics: a high-fidelity simulation study
  1. Daphné Michelet1,2,
  2. Jennifer Truchot1,3,
  3. Marie-Aude Piot1,4,
  4. David Drummond1,5,
  5. Pierre-François Ceccaldi1,6,
  6. Patrick Plaisance1,3,
  7. Antoine Tesnière1,7,
  8. Souhayl Dahmani2,8
  1. 1 Ilumens Simulation Department, University Paris Descartes, Paris, France
  2. 2 Department of Anesthesia, Robert Debré Hospital, Paris, France
  3. 3 Emergency Department, Lariboisière Hospital, Paris, France
  4. 4 Psychiatry Department, Monsouris Mutualiste Institute, Paris, France
  5. 5 Paediatric Pulmonology Department, Necker Hospital, Paris, France
  6. 6 Gynecology and Obstetric Department, Beaujon Hospital, Clichy, France
  7. 7 Department of Anaesthesia and Intensive Care, George Pompidou European Hospital, Paris, France
  8. 8 INSERM U 1141, Robert Debré Hospital, Paris, France
  1. Correspondence to Dr Daphné Michelet, Department of Anaesthesia, Robert Debré Hospital, Paris 75019, France; daphne.michelet{at}aphp.fr

Abstract

Aim Paediatric anaesthesia is a very specialised domain lacking training during the traditional curriculum. The laryngospasm is a stressful and life-threatening event that requires immediate action. The main objective of this study was to assess the effect of knowledge of a simple algorithm on the management of laryngospasm by trainee anaesthetists and nurse anaesthetists during a high-fidelity simulation session.

Method Residents in paediatric anaesthesia and training nurse anaesthetists with similar curriculum were randomly allocated to undergo a simulation session of laryngospasm with the help of a simple algorithm administered to them 5 min before the simulation session (group A) or as taught in their curriculum (group C). The primary endpoint was the assessment score of laryngospasm management using 10 technical items (validated in experienced paediatric anaesthetists). Secondary endpoints were: the non-technical skills using the Anaesthetists' Non-Technical Skills (ANTS) score and timing of critical management steps. The correlation between the technical and non-technical scores was also studied. Data are expressed as median (range).

Results 72 participants (18 teams in each group) were included in this study. There was a statistically significant difference between group A and C on the primary endpoint: 8.5 (2–10) vs 5 (2–8), respectively (p<0.0001). There was also a significant difference between the two groups for the ANTS score 12 (7–16) vs 8 (6–12), respectively (p<0.0001). No difference in timing of management was observed. Finally, there was a strong correlation between the technical skills and all the non-technical skills categories in the A group.

Conclusion A simple algorithm improved the technical and non-technical skills of students during the management of a simulated laryngospasm.

  • laryngospasm
  • paediatric anaesthesia
  • High-fidelity simulation training

Statistics from Altmetric.com

Introduction

Laryngospasm is defined as a closure of the larynx that blocks the passage of air to the lungs. It is one of the most frequent respiratory complications during paediatric anaesthesia. A recent European multicentre cohort study of children undergoing elective, emergency diagnostic or surgical procedures (APRICOT trial1) found a high rate of severe critical events during the perioperative period (5.2%). The incidence of respiratory critical events was 3.1%, and laryngospasm was one of the most frequent events (incidence 0.2%–6.7%). The incidence rate may increase largely if preventive measures are not taken in patients with a high risk (type of surgery and patients’ health condition such as upper respiratory tract infection).2–6

Laryngospasm may lead to severe complications including hypoxia, negative pressure pulmonary oedema, bradycardia and cardiac arrest. It remains one of the leading cause of perioperative cardiac arrest in children.7

When it occurs, the treatment of laryngospasm in children requires appropriate diagnosis, followed by prompt and intensive management. Specialists recommend stopping the stimulation, airway manipulation including removal of the irritant stimulus and administration of pharmacological agents if necessary. However, the optimal sequence of treatment is not clearly defined. Many methods and techniques of airway manipulation have been described in the literature: chin lift,8 jaw thrust,9 pressure on the laryngospasm notch,10 continuous positive airway pressure (CPAP)11 and positive pressure ventilation with a face mask and 100% oxygen.12 The pharmacological management of laryngospasm includes the administration of propofol and/or rapid onset muscle relaxant. Given that the pathophysiology of laryngospasm involves an inappropriate depth of anaesthesia, literature suggests propofol for deepening anaesthesia as the first line of treatment.13 14 In addition, the administration of this compound allows upper respiratory tract relaxation.15 Then, management of laryngospasm naturally includes several steps: airway optimisation, pharmacological treatment (to deepen anaesthesia or to act directly on the laryngeal muscles) but also a necessary time of reassessment after steps such as stopping stimulation, pressure support ventilation (efficacy) and pharmacological treatment (time to act). This short time of revaluation seems important in order to appreciate management effectiveness and avoid unnecessary supplementary treatment with potential side effects.

The use of structured algorithms has been proposed in order to improve the management of perioperative complications.16 Particularly, one study suggests that if correctly applied, an algorithm would result in earlier recognition and/or improved management of this complication in 16% of cases.14 Given the paucity of data concerning the practical management of laryngospasm, we undertook a survey in French paediatric anaesthetists in order to explore the management of laryngospasm.17 Results found that only 30% of practitioners used an algorithm in their institutions. According to results of this survey and previously published algorithms on the management of laryngospasm, we proposed a new algorithm (figure 1). We then thought to study the impact of applying this algorithm on laryngospasm management in a population of inexperienced anaesthetists. In fact, many physicians continue to perform occasionally paediatric anaesthesia after graduation, while studies have found perioperative complications to decrease when the paediatric anaesthetist were specialised and experienced. Both technical performance and non-technical skills were assessed as previous studies found an improvement of both technical and non-technical skills with the use of cognitive aids or checklists during simulated cases of anaesthesia emergencies.18 19 Given the unpredictability and complexity of this event, we used high-fidelity medical simulation. We designed a randomised, controlled trial involving teams of residents in anaesthesia and nurse anaesthetists in training. We hypothesised that the algorithm would significantly improve the management of laryngospasm in terms of technical skills as well as non-technical skills.

Figure 1

Algorithm for the management of laryngospasm. CPAP, continuous positive airway pressure.

Material and methods

This randomised controlled simulation study was performed from November 2016 to January 2017. The study was not registered given that randomisation was not involving patients.

Participants

Residents were recruited from the class of anaesthesia residents from Paris and the nurse anaesthetists in training from the different schools. Residents and nurses were from the same faculty school (Paris V University for residents and Paris VII University for nurses), had received the same online courses on paediatric anaesthesia (14 hours on paediatric anaesthesia including two on respiratory complication) and all succeeded to the online theoretical evaluation concerning this subspecialty. In addition, all recruited residents were in their paediatric anaesthesia rotation in one of the three Parisian paediatric hospitals and had an equivalent level of clinical experience in paediatric anaesthesia at the time of the simulation sessions (3 and 2 months of clinical practice for residents and nurses, respectively). Of note, all paediatric hospitals exhibit the same case mix of patient and the same teaching strategy. We selected residents and students because of their little experience in the field of paediatric anaesthesia, one can hypothesise that experienced anaesthetists might have acquired the necessary skills in managing this complication. We selected teams of resident and nurse because it is the usual organisation of anaesthesia in France and many European countries.

Design

Each team was exposed to a simulated scenario of laryngospasm of a 9-month-old girl (8 kg body weight). We used the high-fidelity manikin SimBaby (Laerdal, Stavanger, Norway) and recorded the sessions through a dedicated software (SimView, Laerdal, Stavanger, Norway) using three cameras and one microphone. The scenario is provided as supplementary material (online supplementary appendix 1). The teams (one resident and one nurse anaesthetist in training) were randomly assigned (http://www.randomization.com/) to manage the laryngospasm with the algorithm (group A, algorithm) or as usual (group C, control). Before the beginning of the simulation case, there was a briefing on the manikin, the material and the possible actions (intubation, call for help…). All the participants were told about the subject of the scenario. The algorithm was then provided to group A participants (a coloured laminated card). The algorithm is shown in figure 1, it was a ‘smart art’ graphic produced in Microsoft word, coloured in a purple gradient and designed with the help of a user experience designer (a professional who turn a presented figure to something that people like and want to use; no graphical designer was consulted). All participants had 5 min of preparation in team. A dedicated physician was available for any explanation about the algorithm as requested by participant of the group A. Importantly, the algorithm was not available as a cognitive aid during the simulation session.

Supplementary file 1

The objective of the session was the management of two episodes of laryngospasm with the following points to achieve, according to the assessment technical score:

  1. First incomplete laryngospasm before intravenous line placement (four points): stop stimulation, airway manipulation (100% oxygen, jaw thrust and gentle CPAP), deepen anaesthesia with sevoflurane and revaluation before continuing.

  2. Second complete laryngospasm at the beginning of the surgery with intravenous line (six points): stop surgeon’s stimulation, airway manipulation (100% oxygen, jaw thrust and gentle CPAP), deepen anaesthesia with intravenous propofol, revaluation (evaluation of treatment effectiveness), second doses of propofol or muscle relaxants (suxamethonium) and revaluation before continuing surgery.

Technical and non-technical skills assessments were performed twice by two raters (paediatric anaesthetists and simulation teachers) blinded to the allocated group using video tapes after the termination of the study.

The technical assessment score was the primary endpoint. Given the absence of validated score for the management of laryngospasm in the literature, the 10 items technical score was derived from evidence-based practices and recommendations (the emergency quick reference guide from the European Society of Anaesthesiology) and rated in a binary way (yes or no: table 1).20 Order of performance of items was also considered: in case of inversion in the order, only one item was rated. This score was validated on 10 experienced paediatric anaesthetists (associated with experienced nurse anaesthetists) not aware of the algorithm and using same scenarios. Those anaesthetists were recruited, after consent, from three different teams. Evaluations of their technical and non-technical skills were performed by the same raters that the trainees. The technical score was analysed with overall all items (/10 points) and according to the following three subscores: ventilation management (/2 points: items 2 and 6 table 1), pharmacological treatment (/3 points: items 3, 7 and 9 table 1) and revaluation (/3 points: items 4, 8 and 10 table 1). This was performed in order to explore possible irrelevant high scores related to the performance of subsets. Items of the score were assessed on the physician or the anaesthetist nurse depending on who perform the scored step.

Table 1

Items of the technical score

Secondary endpoints of this study were: the non-technical skills assessed using the Anaesthetists' Non-Technical Skills (ANTS) score21 and timing of performance of steps of the score. ANTS was assessed on the resident (given that the anaesthesia process is under the responsibility of the physician). The ANTS scores were recorded as the overall category scores on a scale of 1–4 as described by the authors of the ANTS framework. Inter-rater reliability was assessed for all these analysis. We also assessed the timing of performance of the ventilation step in the second scenario (time from the beginning of the second scenario to the initiation of the ventilation) and the first intravenous propofol step in the second scenario (time from the beginning of the scenario to the administration of the first propofol bolus), on the physician or nurse who performed the specific step. Reason for considering those steps and not previous ones was the fact that some steps might not be performed making the timing artificially decreasing.

Finally, we studied the correlation between the technical score (total and subscores) and the ANTS score (total score and subscores) in all participant and in each group of the study.

Statistical analysis

Data are presented as median (range) for continuous data given the non-normal distribution of data and as count (percentage) for categorical parameters. We compared the independent groups using the Mann-Whitney test. Finally, correlations were assessed using the Spearman non-parametric test. Agreement of blinded ratters on the technical analysis was assessed by inter-rater agreement test (Fleiss’ kappa). Importantly, statistical analyses were performed using the mean values displayed by the two ratters. All tests were two-tailed, and statistical significance was considered for p<0.05. Finally, given the number of comparisons, a Bonferroni correction was applied for multiple comparisons at each stage of the statistical analysis.

According to previous simulation sessions among residents trained on the same simulator, assessments in a panel of 10 simulations revealed a median (range) score (without the algorithm) of 5 (2–8). Those results allowed an estimation of the number of pair to be included in the study to 13 with an expected increase in the value of score of at least two point, an alpha risk of 5% and a beta risk of 10%.

Statistical analyses were performed using SPSS V.22.0 software (IBM).

Results

Overall 72 participants were included in the study, 36 residents and 36 nurse anaesthetists in training. Eighteen teams of two participants were randomly assigned to the group A and 18 to the group C. The median age of the residents was 28 (26–40) years, compared with 31 (27–42) years for nurse anaesthetists. Among the residents, 47% were women compared with 74% of the nurse anaesthetists. Eighty five per cent of residents had a previous experience in simulation compared with only 25% of nurse anaesthetist’s students. However, no participant had previous experience in paediatric anaesthesia simulation. Finally, the median year of resident’s training was 4 out of 5, and all the nurse anaesthetist’s students were in second (last year) of training.

Primary endpoint: analysis of the technical management

Applying the score to experienced anaesthetists found those physicians to perform the two scenarios with a score of 10 (8–10), a timing of ventilation of 10 (5–12) s and a timing to achieve propofol administration of 25 (17–28) s. Comparison between those data and the ones displayed by the control group of participants (see below and in tables for values) were all significant (p=0.0001, 0.002 and 0.03, respectively). The ANTS score in those experienced anaesthetists was 14 (13/16) and was significantly different from the study’s groups (p=0.0001 for both comparisons).

The technical score was higher in the algorithm group: 8.5 (2–10) than in the control group: 5 (2–8) (p<0.0001, significant after the Bonferroni correction: table 2). The comparison of the technical subscores (ventilation, revaluation and pharmacological treatment) found only the revaluation and pharmacological treatment subscores to be significantly different between the two randomised groups (table 2). Inter-rater agreement was 0.76 for the technical score.

Table 2

Technical performance of participants without (control) and with (algorithm) a simple algorithm. Values are median (minimum–maximum)

Secondary endpoints: analysis of the non-technical skills and timing of management

The median value of the ANTS score was higher in the algorithm group: 12 (7–16) than the median value of the score in the control group: 8 (6–12) (p<0.0001, significant after the Bonferroni correction, table 3). The detailed analysis of the ANTS categories shows that the algorithm group had higher ANTS scores in all categories compared with the control group (table 3). Inter-rater agreement was 0.73 for the ANTS score.

Table 3

Non-technical performance of participants without (control) and with (algorithm) a simple algorithm. Values are median (minimum–maximum)

Analysing the timing of realisation of the ventilation step (of the second scenario) did not found any significant difference between the two groups nor concerning the timing of ventilation initiation (12(1– 28) s in group A vs 13 (9–23) s in group C, p=0.5) neither concerning the administration of propofol (30 (3-88) s in group A vs 34 (16–91) in group C, p=0.3).

Correlations between the technical score (total and subscores) and the ANTS (total and categories) are displayed in table 4. In the algorithm group, the total ANTS correlated significantly with the total technical score and subscores and the total technical score correlated significantly with the total ANTS score and all its subcategories.

Table 4

Pearson coefficient of the correlation analysis of the clinical score (total and subscore evaluation and pharmacological management) and the Anaesthetist’s  Non- Technical Skills score (total and subcategories)

Discussion

The main finding of this study can be summarised as following: the knowledge of a simple algorithm improved the management of laryngospasm by trainees but had no impact on timing of management. It also improved all categories of the ANTS score. Finally, correlations were found between the technical and non-technical skills in the algorithm group during the management of laryngospasm.

In this trial, even though we enrolled teams of students already in residency in paediatric anaesthesia, we observed that the control group was less efficient in the management of laryngospasm (median score of 5 out of 10). The areas of improvement were those concerning the revaluation and the pharmacological treatment of laryngospasm. The pharmacological treatment either by deepening the anaesthesia (using a volatile agent or intravenous propofol) or by acting directly on the release of laryngeal muscles (by administering propofol or muscle relaxant) represents the core of the treatment of this complication which confirms the utility of the algorithm. Concerning the revaluation step, assessing the effectiveness of the treatments, especially after deepening anaesthesia or administering muscle relaxant, allows to give the treatments time to act and release the spasm. This short observational step might help physician better managing their stress, communicating with team-mates and avoid rushing to additional unnecessary and harmful treatments (readministering drugs with haemodynamic effects). Overall, our results indicate that the knowledge of a simple algorithm improved largely the performance of students, confirming the positive impact of an algorithm to manage such critical situations.

The score used for assessing the primary outcome of the study was not validated. However, it was derived from the emergency quick reference guide from an international scientific society (the European Society of Anaesthetists)20 and, then, could be considered as scoring the best practice in managing laryngospasm. This is supported by the high scores found in experienced anaesthetists (significantly higher than the control study group) and the better timing of realisation of some key steps in this group of experienced physicians. However, as both the algorithm and the assessment score were based on scientific society recommendations, literature analysis and experience of anaesthetists, there are clearly some similarities between these two tools.

The knowledge of the studied algorithm also improved the non-technical skills as displayed by the improvement of all the categories of the ANTS score. The improvement of decision making is an interesting finding given its classical association with experience.22 This result strongly supports the weight of our findings concerning the application of the algorithm. Improving non-technical skills of practitioners is recognised as a major goal in medicine.23 However, the current study cannot be compared with previous ones using cognitive aids available during the simulation session. Many studies have shown such cognitive aids to improve the management of many critical medical situations, such as cardiac arrest of malignant hyperthermia.19 24–26 Our study is different to the extent that it explores the knowledge of a simplified algorithm prior to the occurrence of the critical event. Consequently, improvement in the management of laryngospasm observed in the current study would involve both the knowledge of the algorithm (supported by results of the 10-item score) and the algorithm per se (supported by the fact that trainee participants in the algorithm group obtained 10-item scores close to those of experienced physicians).

Interestingly, we also found a strong correlation between the technical score and the ANTS score in the algorithm group. These findings have been described in other settings like surgery training18 or intraoperative cardiac arrest management.19 Technical subscores correlated significantly with teamwork, situation awareness and decision making. This result strongly supports our hypothesis concerning the important role of revaluation in improving the management of the studied complication. One can hypothesise that the revaluation time might improve teamwork, situation awareness and decision making by allowing caregivers, as previously suggested, staying calm and communicating. The resulted improvement in decision making might then explain the better pharmacological management of laryngospasm given the significant relation found between these two factors.

To our knowledge, the current study is one of the first showing an improvement in the management of laryngospasm with the knowledge of a simple algorithm in multidisciplinary student teams. Paediatric training for every resident in anaesthesia requires 3 months of continuous training in a specialist paediatric centre according to European guidelines and 6 months according to French guidelines. This is very short20 when considering the required high level of experience necessary to decrease the occurrence of perioperative complications. Consequently, one can expect that physicians with a low experience in paediatric anaesthesia might improve their management of laryngospasm using this simple algorithm.

Our study has some limitations. First, the delivered algorithm was not formatted as a cognitive aid. Although, this would impact results, this is unlikely given that it was not available during the simulation session. Similarly, one cannot exclude the effect of memory recency on the application of steps of the algorithm. Second, our evaluation was performed during a simulation session and not in real conditions. However, the rarity of the event and its random occurrence would render a clinical evaluation very laborious. In addition, simulation appears to be an elegant and validated tool to assess a clinical hypothesis, especially for emergency critical situations or rare events. A recent study27 highlighted this advantage by comparing results of simulation-based study versus real clinical situations on drug administration errors. Results found the same conclusions for primary endpoints with differences in the size effects, leading to the conclusion that simulation might be a valid method to evaluate clinical issues. Similarly, the algorithm was available 5 min before the beginning of the scenario; this does not represent real life where the tasks could not be reviewed immediately prior to the event. Consequently, further studies are needed in order to assess the efficacy of the knowledge of the algorithm in more realistic conditions. Fourth, we did not collect the baseline measures of participant’s technical and non-technical skills because of organisation difficulties. This clearly impairs the quality of evidences concerning this improvement of this outcome in the algorithm group. However, the two studied groups were at the same training level and experience at the time of the study. This supports the equivalence in technical and non-technical skills between groups. This is largely supported by (a) the ANTS score that was higher in the experienced physicians in comparison to both the C and A studied groups and (b) the observed correlation between technical and non-technical skills in the algorithm group but not in the control one. In addition, non-technical skills are partly related to technical skills and to the present situation. Finally, the choice of the teams relaying on a binomial association of a physician and a specialised nurse was based on real life organisation of paediatric anaesthesia practice in France and in many European countries.

In conclusion, our study highlights the efficacy of the knowledge of a simple algorithm in improving both technical and non-technical skills during the management of laryngospasm during paediatric anaesthesia. This appears to be important given the rarity and life threatening character of this complication. Such simple algorithms should be considered for rare and emergency situations in order to prevent their life-threatening consequences.

References

Footnotes

  • Contributors DM: participated in study design, student recruitment, high-fidelity simulation sessions, video analysis, data analysis and interpretation, drafting and revising the manuscript and approved the final version. JT: participated in study design, high-fidelity simulation session, video analysis, data interpretation, drafting and revising the manuscript and approved the final version. M-AP, DD : participated in video analysis and approved the final version. P-FC, PP, AT: participated in data collection analysis and interpretation, drafting and revising the manuscript and approved the final version. SD: participated in study design, data collection analysis and interpretation, drafting and revising the manuscript and approved the final version.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Ethics approval This study was approved by our institutional IRB (Comité d’Evaluation de l’Ethique des projets de Recherche Biomédicale de l’hôpital Robert Debré; # 2015/312(2)).

  • 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.