Article Text

Tracheostomy in children admitted to paediatric intensive care
  1. Dora Wood1,
  2. Philip McShane2,
  3. Peter Davis1
  1. 1Paediatric Intensive Care Unit, Bristol Royal Hospital for Children, Bristol, UK
  2. 2Paediatric Epidemiology Group, University of Leeds, Leeds, UK
  1. Correspondence to Dora Wood, Bristol Royal Hospital for Children, Paediatric Intensive Care Unit, Upper Maudlin Street, Bristol, Avon BS2 8BJ, UK; dora{at}


Purpose Tracheostomy is a common intervention for adults admitted to intensive care; many are performed early and most are percutaneous. Our study aimed to elucidate current practice and indications for children in the UK admitted to paediatric intensive care and undergoing tracheostomy.

Design A questionnaire covering unit guidelines, practice, and the advantages and disadvantages of tracheostomy was sent to all UK paediatric intensive care units (PICUs) participating in the Paediatric Intensive Care Audit Network (PICANet). These results were combined with data from PICANet on all children in the UK reported to have had a tracheostomy performed during a PICU admission between 2005 and 2009 inclusive.

Results Over 5 years, 1613 children had tracheostomies performed during their PICU admission (2.05% of all admissions). The death rate was 5.58% with tracheostomy versus 4.72% overall, but differences were not significant when risk-adjusted using the Paediatric Index of Mortality 2 (PIM2). All 29 units participating in PICANet responded to the survey. Prolonged invasive ventilation was an indication for tracheostomy in 25/29 units, but the definition varied between 14 and 90 days, and most respondents considered timing on an individual basis. Children undergoing tracheostomy during PICU admission account for 9% of PICU bed days in the UK.

Conclusions In contrast with current adult UK practice, tracheostomy for children admitted to intensive care is infrequent, performed late following admission and usually surgical. Practice varies significantly. The death rate for children having a tracheostomy performed was not significantly higher than for children admitted to PICU who did not undergo tracheostomy.

Statistics from

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.


Tracheostomy was not widely performed until the nineteenth century when its use in children became common during epidemics of infectious disease including poliomyelitis and diphtheria. Before routine trans-laryngeal intubation, it was often the only option for bypassing upper airway obstruction or providing positive pressure ventilation.

The indications for tracheostomy evolved following the introduction of improved equipment and techniques for endotracheal intubation. There has been a steep decline in immunisation-preventable infectious diseases causing airway obstruction over the past 50 years. Conversely, there has more recently been an increase in the number of children surviving with complex medical needs for whom tracheostomy and/or home ventilation is part of their chronic disease management.1

What is already known on this topic

  • In hospitalised children, tracheostomy is increasingly used in children with complex and/or chronic conditions, which are associated with higher background mortality.

  • Reported death rates for children undergoing tracheostomy are high, particularly for infants.

What this study adds

  • Tracheostomy for children admitted to paediatric intensive care units (PICU) is infrequent, performed late following admission and usually surgical.

  • PICU mortality in children following tracheostomy was not significantly higher than for children not undergoing tracheostomy in the PICU.

  • Children undergoing tracheostomy during PICU admission account for a disproportionate number of PICU bed days nationally.

Many advantages of tracheostomy over translaryngeal intubation have been postulated. In adult practice, an increasing body of evidence has supported the role of tracheostomy in reducing length of intensive care unit (ICU) and hospital stay and improving patient comfort. The evidence for other benefits, such as patient safety and better long-term laryngeal function, is less convincing.2

In critically ill adults, tracheostomy is one of the most commonly performed procedures on the ICU, and the trend is for it to be carried out ever earlier in the patient's ICU stay. A survey of UK adult practice showed that the overwhelming majority of respondents would consider tracheostomy indicated at less than 10 days of mechanical ventilation.3 Indeed, recent research and debate increasingly looks at a cut-off of 72 h or even less.4 This trend has been mirrored by the increase in tracheostomy being performed by intensivists as a bedside percutaneous procedure, rather than as a surgical procedure that necessitates theatre and surgical time plus potentially risky patient transfers.3

Much less is known about current practice in paediatric intensive care units (PICUs). With few exceptions, studies of tracheostomy in children are retrospective reviews of practice and outcomes in single units involving only small numbers of children. Larger studies have looked at tracheostomy in all hospitalised children, rather than the specialised case of those on PICU.1 ,5 There are few trial data on the benefits or risks of tracheostomy in this setting and therefore there are currently no national or international recommendations to guide practice.

By combining data from the UK Paediatric Intensive Care Audit Network (PICANet) with information from a survey of lead clinicians from PICUs nationally, our aim was to clarify current practice with regard to tracheostomy on PICU.


PICANet is a national audit of paediatric intensive care that has collected a standardised dataset prospectively on all admissions to PICUs in England and Wales since 2002 and subsequently from PICUs in Scotland, Northern Ireland and Ireland. The dataset consists of demographics, admission characteristics, diagnostic information, interventions and outcomes. Data quality is ensured by regular training of PICU staff on data definitions and by means of local and central validation checks. PICANet has research MREC ethics committee approval (05/MRE04/17) and National Information Governance Board (4-07(c)/2002-PICANet) approvals to collect patient identifiable data without informed consent.

Data were extracted from the PICANet dataset on all children reported to have had a tracheostomy performed during an admission to PICU between 2005 and 2009 inclusive. This included date of admission, age at admission, gender, Paediatric Index of Mortality 2 (PIM2) score at admission,6 admission diagnosis and diagnostic group, and survival at PICU discharge. Where relevant, comparison was made with the data for all admissions over the same period.

A questionnaire about practice related to tracheostomy in paediatric intensive care was developed in accordance with local institutional policy. Agreement was obtained from the UK Paediatric Intensive Care Society Study Group for the questionnaire to be distributed nationally. The questionnaire, available in electronic or paper format, was sent to the lead clinicians of all UK and Irish PICUs participating in PICANet (see online supplementary appendix). If no response was received after two reminders, clinicians were contacted by telephone. Information requested included the availability of a unit guideline, details about type of tracheostomy, where and by whom these were performed, and any minimum size and age limits. There were also questions on attitudes to tracheostomy for children on PICU, advantages, indications and other factors influencing the decision-making process. Free text annotations and additions were encouraged.

Percentages were calculated using Microsoft Excel and further analysis carried out using Stata (V.11).7 p Values for PICANet results were derived using the χ2 test for categorical variables, Mann–Whitney test for comparing length of stay, and t test to compare predicted mortality. The relationship between unit size and tracheostomy rate was expressed using the Pearson product-moment correlation coefficient. The OR for mortality with and without tracheostomy was corrected for admission PIM2 score using logistic regression with unit treated as a random effect.


Characteristics of children undergoing tracheostomy

Over the 5 years studied, 1613 of 78 504 patients admitted to PICU had tracheostomies performed during their PICU stay (2.05% all admissions). Rates varied between units from 0.13% to 5.66% of admissions. There was no relationship between unit size and tracheostomy rate (correlation coefficient −0.05). Compared with all admissions, children who had tracheostomies stayed in PICU longer (mean 24.7 vs 5.7 days, median 9 vs 3 days) with 21.2% (vs 2%) staying for more than 28 days (table 1). Children undergoing tracheostomy used 9.0% of all the PICU bed days over this period.

Table 1

Comparison of characteristics of children undergoing tracheostomy with all admissions

The age and gender distribution of children having tracheostomy was similar to the profile for all admissions. The average PIM2-derived predicted mortality risk for children having tracheostomy performed was slightly raised at 6.1% compared with 5.5% for all admissions. These children were more likely to have been admitted to intensive care (neonatal intensive care unit/PICU/ICU) during that hospital stay and less likely to be post surgery.

Children who underwent tracheostomy were more likely to have their admission diagnosis grouped as respiratory (48.0 vs 25.2% of all admissions) and less likely as cardiovascular (13.7 vs 29.0% of all admissions). In children who had a tracheostomy performed as part of their PICU admission, the most common diagnoses were respiratory failure, lower respiratory tract infection, subglottic stenosis and head injury. Comparatively for all admissions, the most common diagnoses were ventricular septal defect, respiratory failure, tetralogy of Fallot and status epilepticus.

The death rate before PICU discharge was 5.58% for children undergoing tracheostomy; this was higher than the rate of 4.72% overall, but differences were not significant when adjusted for PIM2 score at admission (OR 1.09, 95% CI 0.84 to 1.42, p=0.5). It should be noted that death rates were higher in infants compared with older children.

Current practice within paediatric intensive care

A consultant from all 29 units responded to the survey; 27 respondents knew of at least one child having a tracheostomy in their unit over the preceding 5 years. One respondent reported that their unit never performed tracheostomy. Only four (13.8%) had a unit guideline or policy regarding timing and indications for tracheostomy.

In 21 units, all tracheostomies were surgical procedures. In six others, percutaneous tracheostomy was additionally performed on selected adolescents, but never on younger children.

Ear, nose and throat surgeons usually performed any tracheostomy required in all responding units. In addition, general (adult) intensivists carried out some procedures in four units, and paediatric intensivists, thoracic surgeons, paediatric surgeons, neonatal surgeons or cardiac surgeons carried out procedures in one unit each. Tracheostomies were usually performed in the operating theatre, but for the units undertaking percutaneous procedures these were completed on PICU. Five other units responded that they had facilities for emergency tracheostomy on PICU, but all noted that this happened very rarely if ever.

No unit that performed tracheostomy had a lower size or age limit for this procedure, although three respondents reported that they would have a higher threshold or more practical difficulty in infants.

Indications and advantages of tracheostomy

Of the 28 units which reported undertaking tracheostomies, all reported ‘airway obstruction’ as an indication, while ‘providing prolonged mechanical ventilation’ (25/28) and ‘to aid weaning from ventilation’ (20/28) were also common indications. There was little agreement over the length of invasive ventilation that might indicate need for a tracheostomy. Of the 25 respondents who listed it as an indication, seven respondents gave no indication of the length they would consider appropriate, four suggested >14 days of ventilation, seven >28 days, five between 14 and 28 days and two indicated >90 days. Two respondents also reported that they would consider tracheostomy after a number of failed extubations. Other indications included unsafe airway reflexes, tracheomalacia, difficult intubation, cervical spine/neurological injuries, neuromuscular problems and complex syndromes.

In terms of other possible advantages of tracheostomy, six units responded that there were none. Twelve thought it was helpful if long-term or repeated support was anticipated. Eleven noted the advantage of a safe airway, five of those specifically if intubation or the airway was ‘difficult’. Nine respondents commented that tracheostomy allowed clinicians to reduce or stop sedation with improved comfort, normalisation, development or quality of life. Similarly, eight reported that tracheostomy allowed easier weaning from the ventilator, with two specifically mentioning ability to discharge the patient from PICU while they still required respiratory support. Four felt tracheostomy allowed improved tracheal toilet or clearance of secretions. Only one noted protection of the vocal cords specifically.

Other factors influencing tracheostomy decision making

Of other factors which would influence the decision to proceed to tracheostomy, ‘underlying condition’ was selected by 18 respondents, ‘co-morbidities’ by 17, ‘risk of complications’ by 11, ‘parent factors’ by 11, ‘size of child’ by 9, ‘age of child’ by 8 and ‘cosmetic result’ by 1. Five respondents added that they would be influenced by factors such as prognosis, futility or the overall best interests of the child.


To our knowledge, this is the largest study to date of tracheostomy on the PICU. The comprehensive nature of the data from PICANet, as well as the high response rate from the survey, means that our results are likely to be representative of current practice in the UK.

Only 2% of children admitted to PICU undergo a tracheostomy, compared with much higher rates in adult intensive care units. Tracheostomies in children are predominantly surgical, whereas in adult practice in the UK the vast majority are now percutaneous.3 In adults, tracheostomy is increasingly performed for relatively short-term ventilator dependence, often of less than 7 days,4 whereas no paediatric intensivist in this study reported undertaking tracheostomy for durations of ventilation of less than 14 days and for some the suggested timing was over 90 days of mechanical ventilation.

In some ways, current paediatric practice is similar to that in adult critical care a couple of decades ago. So should paediatric practice move closer to that of our adult colleagues? The changes in adult practice have been driven by research data which are largely absent in the paediatric population, making it difficult to make evidence-based recommendations. In the absence of evidence, paediatric practice will continue to be determined by personal experience and expertise, as reflected in the variation in tracheostomy rates seen between PICUs in this study. Consistent with a recent survey of Canadian paediatric intensivists,8 attitudes towards tracheostomy among paediatric intensivists in the UK are generally positive but also widely varied. For example, despite all but one unit having children undergoing tracheostomy, over 20% of our respondents said they could see no advantages to tracheostomy at all. As few units reported having a guideline, even within an individual PICU practice between intensivists may vary significantly. Similarly, although percutaneous tracheostomy has been reported in children since the mid-1990s,9 only six PICUs in the UK reported undertaking this procedure, most likely dependent on the operator experience of individual clinicians.

The main limitation of our study is that although we had data on a large number of children undergoing tracheostomy, we had limited information available on individual patients, so we are unable to report on, for example, indications for tracheostomy, complication rates, or whether mortality was tracheostomy-related. However, we were able to analyse the outcomes of children undergoing tracheostomy in terms of mortality before PICU discharge. A recent large study examined mortality in children, including neonates, following tracheostomy and quoted a death rate of 7.7–8.5% before discharge from hospital.1 An older systematic review of reported complication rates following tracheostomy in children reported a 7% death rate overall and a 1.3% tracheostomy-related death rate.10 By comparison, our observed death rate was 5.58% in children with a tracheostomy compared to 4.72% overall, but differences between these groups were not significant when adjusted for admission PIM2 score of predicted mortality. One reason for the reduced death rate in this study may be that it relates to PICU discharge rather than hospital discharge. Also death rates in neonates tend to be higher and these infants would not normally be included in PICU data, unlike hospital-wide data.

More evidence is needed to demonstrate that tracheostomy in selected patients does not increase the risk of death or serious adverse events, but in a limited way our data do contradict older studies that describe paediatric tracheostomy as a hazardous undertaking. The systematic review found an increased rate of both complications and mortality in younger children, consistent with the perceptions of many respondents.10 The death rate in children under 1 year of age admitted to PICU is about 50% higher than in those aged over 1, and this increase is similar in children with or without tracheostomy. It is clear that the population of children requiring tracheostomy represents a significant workload, using 9% of PICU bed days nationally over this 5-year period. Tracheostomy per se does not necessarily increase admission length; this result is more likely to reflect a group of children with complex medical needs in a setting with increasing acceptance of aggressive medical interventions, leading to frequent and/or prolonged admissions.


In contrast with current adult practice, tracheostomy for children admitted to PICU is infrequent, is performed relatively late, and is usually a surgical procedure. However, children undergoing tracheostomy use a significant proportion of PICU bed days. Practice varies significantly between PICUs, most likely partly determined by local expertise, as there is a paucity of good evidence regarding advantages or risks in the paediatric population. The death rate for children undergoing tracheostomy in this study was lower than previously reported, and was not significantly higher than the rate in children admitted to PICU who did not undergo tracheostomy. With increasing numbers of children with complex medical needs and ever higher parental expectations of medical care, good evidence to guide practice is urgently needed, especially given the likely effect on PICU occupancy of increased numbers of children undergoing tracheostomy.


The authors would like to thank all the staff in participating hospitals who have collected data for PICANet. The authors are grateful to the Paediatric Intensive Care Society for continued support and to the members of the PICANet Steering Group and Clinical Advisory Group. A list of contributing centres and Steering Group and Clinical Advisory Group members may be found at


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Funding PICANet is funded by the National Clinical Audit and Patient Outcomes Programme via the Healthcare Quality Improvement Partnership (HQIP), Health Commission Wales Specialised Services, NHS Lothian/National Service Division NHS Scotland, the Royal Belfast Hospital for Sick Children, Our Lady's Children's Hospital, Dublin, Children's University Hospital, Dublin and The Harley Street Clinic, London.

  • Competing interests None.

  • Ethics approval PICANet has research MREC ethics committee approval (05/MRE04/17) and National Information Governance Board (4-07(c)/2002-PICANet) approvals to collect patient identifiable data without informed consent.

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