Articles

Airway management complications in children with difficult

tracheal intubation from the Pediatric Difficult Intubation
(PeDI) registry: a prospective cohort analysis
John Edem Fiadjoe, Akira Nishisaki, Narasimhan Jagannathan, Agnes I Hunyady, Robert S Greenberg, Paul I Reynolds, Maria E Matuszczak,
Mohamed A Rehman, David M Polaner, Peter Szmuk, Vinay M Nadkarni, Francis X McGowan Jr, Ronald S Litman, Pete G Kovatsis

Summary
Background Despite the established vulnerability of children during airway management, remarkably little is Lancet Respir Med 2015
known about complications in children with difficult tracheal intubation. To address this concern, we developed a Published Online
multicentre registry (Pediatric Difficult Intubation [PeDI]) to characterise risk factors for difficult tracheal December 16, 2015
http://dx.doi.org/10.1016/
intubation, establish the success rates of various tracheal intubation techniques, catalogue the complications of
S2213-2600(15)00508-1
children with difficult tracheal intubation, and establish the effect of more than two tracheal intubation attempts
See Online/Comment
on complications. http://dx.doi.org/10.1016/
S2213-2600(15)00519-6
Methods The PeDI registry consists of prospectively collected tracheal intubation data from 13 children’s hospitals in Children’s Hospital of
the USA. We established standard data collection methods before implementing the secure web-based registry. After Philadelphia, Philadelphia, PA,
establishing standard definitions, we collected and analysed patient, clinician, and practice data and tracheal USA (A Nishisaki MD,
J E Fiadjoe MD,
intubation outcomes. We categorised complications as severe or non-severe. Prof F X McGowan Jr MD,
Prof V M Nadkarni MD,
Findings Between August, 2012, and January, 2015, 1018 difficult paediatric tracheal intubation encounters were done. Prof R S Litman DO); Perelman
School of Medicine,
The most frequently attempted first tracheal intubation techniques were direct laryngoscopy (n=461, 46%), fibre-optic
University of Pennsylvania,
bronchoscopy (n=284 [28%]), and indirect video laryngoscopy (n=183 [18%]) with first attempt success rates of 16 (3%) Pennsylvania, PA, USA
of 461 with direct laryngoscopy, 153 (54%) of 284 with fibre-optic bronchoscopy, and 101 (55%) of 183 with indirect (J E Fiadjoe, Prof F X McGowan Jr,
video laryngoscopy. Tracheal intubation failed in 19 (2%) of cases. 204 (20%) children had at least one complication; Prof V M Nadkarni,
Prof M A Rehman MD,
30 (3%) of these were severe and 192 (19%) were non-severe. The most common severe complication was cardiac
Prof R S Litman); Children’s
arrest, which occurred in 15 (2%) patients. The occurrence of complications was associated with more than two Hospital Boston, Boston, MA,
tracheal intubation attempts, a weight of less than 10 kg, short thyromental distance, and three direct laryngoscopy USA (P G Kovatsis MD);
attempts before an indirect technique. Temporary hypoxaemia was the most frequent non-severe complication. Harvard Medical School,
Boston, MA, USA (P G Kovatsis);
Ann and Robert H Lurie
Interpretation More than two direct laryngoscopy attempts in children with difficult tracheal intubation are associated Children’s Hospital of Chicago,
with a high failure rate and an increased incidence of severe complications. These results suggest that limiting the Chicago, IL, USA
number of direct laryngoscopy attempts and quickly transitioning to an indirect technique when direct laryngoscopy (N Jagannathan MD); Feinberg
School of Medicine,
fails would enhance patient safety.
Northwestern University,
Illinois, IL, USA (N Jagannathan);
Funding None. The Children’s Hospital Denver,
Denver, CO, USA
(Prof D M Polaner MD);
Introduction Project (NAP4) of the UK Royal College of Anaesthetists University of Colorado School of
Tracheal intubation is a potentially life-saving procedure and the Difficult Airway Society did a study7,8 of major Medicine, Denver, CO, USA
done by many clinicians and is usually easily accomplished complications of airway management in their National (Prof D M Polaner); Seattle
with conventional direct laryngoscopy. Difficult tracheal Health Service hospitals during 1 year. They reported only Children’s Hospital, Seattle, WA,
USA (A I Hunyady MD);
intubation requires unique expertise and methods such ten events in children younger than 10 years, four of University of Washington
as extraglottic airway devices, fibre-optic bronchoscopy, which were related to difficult intubation. Complications Medical School, Seattle, WA,
and video laryngoscopy.1 Despite the widespread use of included subglottic narrowing, aspiration, and death. A USA (A I Hunyady); University of
these indirect techniques by various clinicians (eg, knowledge gap exists about the efficacy of various indirect Michigan Medical School,
Michigan, MI, USA
emergency room physicians, neonatologists, intensivists, tracheal intubation methods in children, related (P I Reynolds MD); University of
surgeons, and anaesthesiologists), little is known about complications, and their risk factors. We successfully Texas Southwestern Medical
related adverse events.2–4 Children under care of an designed and implemented a collaborative, multicentre Center, Dallas, TX, USA
anaesthetist have more airway-related adverse events web-based registry (the Pediatric Difficult Intubation (Prof P Szmuk MD); Dallas and
Children’s Medical Center at
than adults.5 Analysis of the American Society of (PeDI) registry) under the auspices of the Society for Dallas and Outcome Research
Anesthesiologists (ASA) closed claims database showed Pediatric Anesthesia to address these concerns and Consortium, Cleveland, OH,
that respiratory events were more common in children improve airway management in children with difficult USA (Prof P Szmuk); University
than in adults (43% vs 30% respectively; p≤ 0·01) with tracheal intubation. of Texas Health Science Center
Houston, Texas, TX, USA
greater mortality in paediatric claims than in adult claims The goals of the present study were to define the type and (M E Matuszczak MD);
(50% vs 35%; p ≤0·01).5,6 The Fourth National Audit incidence of complications that arise from airway

www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1 1

 Articles

Children’s Memorial Herman

Hospital, Houston, TX, USA
(M E Matuszczak); and
The Johns Hopkins Medical
Insitutions, Baltimore, MD, USA
(R S Greenberg MD)
Correspondence to:
Dr John E Fiadjoe, Children’s
Hospital of Philadelphia,
Department of Anesthesiology
and Critical Care, Philadelphia, PA
19104, USA
FIADJOEJ@email.chop.edu
Research in context
Evidence before the study
Findings of single centre studies in adults have shown that
repeated conventional tracheal intubation attempts in critically
ill patients contribute to patient morbidity. Multicentre studies
from paediatric intensive care units suggest that critically ill
children are at high risk of complications during airway
management in the intensive care unit. Little is known about
the complications and associated risk factors of airway
management in children with difficult tracheal intubation cared
for by anaesthesiologists. Before undertaking this study we
searched the scientific literature published between October,
1975, and October, 2011, with the terms “pediatric difficult
intubation”, “difficult airway”, “difficult direct laryngoscopy”,
“difficult tracheal intubation”, “complications and pediatric
intubation”, “anesthesiology and pediatric difficult intubation”
in various combinations. We filtered our results to exclude adult
studies, case reports, and case series. We found no relevant
large multicentre trials directly relating to this topic. The aim of
our study was to establish the complications and their
associated risk factors in children included in a prospective
multicentre difficult tracheal intubation registry.
Added value of this study
This study is the first to assess the complications of children
with difficult tracheal intubation as established by
anaesthesiologists. This study establishes the complication
rates in these patients, identifies the risk factors for
complications, and estimates the incidence of difficult tracheal
intubation in children under anaesthesia care. This study has
identified a high complication rate. This finding should
encourage further investigations and a change in clinical
practice patterns to enhance patient safety.
Implications of all the available evidence
Children with difficult tracheal intubation are a high-risk group
and multiple tracheal intubation attempts are a key risk factor for
complications. Clinicians should treat every tracheal intubation
attempt as a critical intervention and should limit the number of
direct laryngoscopy and tracheal intubation attempts in this
population. Future research should investigate interventions to
reduce these complications such as checklists and care protocols.
Additionally, the role and efficacy of passive oxygenation during
tracheal intubation is unclear and should be investigated.

management in children with difficult tracheal intubation;
establish the success of various tracheal intubation
techniques; identify associations between patient, clinician,
and practice characteristics, and the occurrence of
complications; and establish the effect of multiple tracheal
intubation attempts (>2 attempts) on complications.
Methods
Study design
The PeDI Collaborative Group was created as a special
interest group within the Society for Pediatric Anesthesia
with the goal to make possible multicentre collaboration
and quality improvement in patients with difficult
tracheal intubation. In December, 2010, members of the
Society for Pediatric Anesthesia were solicited via
electronic mail to take part in the special interest group;
known experts in the paediatric anaesthesia community
were also invited to participate. 48 members of the society
responded and were included in the group; however, ten
active members were instrumental in formulating the
definitions for the registry. The group held four consensus
development meetings between Oct 14, 2011, and Oct 17,
2012, which were attended by all the core members in
person or by telephone conference. In addition to
meetings held at the Society for Pediatric Anesthesia and
ASA annual meetings, task force discussions were
continued using a dedicated group listserv and telephone
conferences to define paediatric airway management
related terminology and adopt standard definitions in
order to create a multicentre registry. The group defined
relevant outcomes, developed a standard data collection
sheet, standardised data definitions, and established
standard operating procedures by expert consensus. This
consensus was reached iteratively by open discussions
until all ten core members reached agreement.
Study population
We collected tracheal intubations supervised or
performed by anaesthesiologists in elective and non-
elective tracheal intubation situations. These situations
included tracheal intubations in various anaesthetic
locations and non-elective tracheal intubations from
other hospital locations such as the emergency
department and intensive care unit whenever the
anaesthesiology team assisted with airway management.
We used the following inclusion criteria in children
younger than 18 years in whom an anaesthesiologist
supervised or performed tracheal intubation: children
with difficult laryngeal exposure with conventional direct
laryngoscopy as assessed by the attending
anaesthesiologist (Cormack and Lehane classification
≥3);9 children in whom conventional direct laryngoscopy
was physically impossible because of anatomical reasons
(eg, severely limited mouth opening or other craniofacial
anomalies); children who had failed conventional direct
laryngoscopy within the preceding 6 months; children
in whom the attending anaesthesiologist deferred
conventional direct laryngoscopy because of an
unfavourable (predictive of a difficult laryngoscopy)
airway physical examination (eg, neonatal Robin
sequence) or the clinical situation in which a non-
attending clinician obtains an unfavourable view that is
unconfirmed with a subsequent conventional direct
laryngoscopy by the attending anaesthesiologist.

2 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1

 Articles

The airway physical examination was not standardised We defined an intubation attempt as the act of inserting
but was based on the assessment of the attending an airway device into the pharynx or naris with the
anaesthesiologists. Several cases met more than one intent to perform tracheal intubation. Many attempts to
inclusion criteria and all tracheal intubation attempts were pass the tracheal tube through the vocal cords were
collected if one of the inclusion criteria was met. Separately regarded as one attempt provided the intubating device
we also queried each anaesthesiologist attending about
whether he or she anticipated the difficulty with mask
Cases entered Complication
ventilation or tracheal intubation or both.
Centre A 395 92 (23%)

Site enrolment and data capture Centre B 231 33 (14%)

After Institutional Review Board approval, we developed Centre C 117 23 (20%)

and piloted a secure, password-protected web-based data Centre D 76 22 (29%)
entry portal at the data coordinating centre (at the Centre E 37 9 (24%)
Children’s Hospital of Philadelphia, PA, USA). We Centre F 33 7 (21%)
expanded this portal to 13 academic children’s hospitals Centre G 31 4 (13%)
in the USA. The contributing sites are listed in the Centre H 26 4 (15%)
acknowledgments section. Each site identified a project Centre I 24 2 (8%)
coordinator (paediatric anaesthesiologist attending). The Centre J 24 3 (13%)
project coordinator educated their faculty about the Centre K 14 2 (14%)
registry, developed a site-specific compliance plan to Centre L 10 3 (30%)
ensure 100% data capture and did monthly audits for Total 1018 204 (20%)
data capture and accuracy. Our compliance plan included Median (IQR) 33 (24–97) 18% (14–24)
monthly data audits at each site and a standardised
Data are n or n (%) unless stated otherwise. Centres sorted by the number of
electronic medical record or manual search for missed difficult cases of tracheal intubation.
cases. The data coordinating centre did monthly audits
of the aggregate data to ensure data entry completeness Table 1: Case distribution across centres
and sent monthly missing data reports to the project
coordinators at each site. A steering committee reviewed
and approved each centre’s compliance plan before their Anticipated Unanticipated Total (n=1018)
full participation. One site failed to meet compliance difficult airway difficult airway
with audits and 100% data capture and was excluded (n=821) (n=197)

from the analysis. The web-based data entry portal used Success* 810 (99%) 188 (95%) 998 (98%)
the Research Electronic Data Capture (REDCap) method Surgical or failed airway* 10 (1%) 9 (5%) 19 (2%)
hosted at the data coordinating centre.10 Any complications 157 (19%) 47 (24%) 204 (20%)
We prospectively created a standard data collection sheet Severe complications† 19 (2%) 11 (6%) 30 (3%)
that was completed either by a member of the anaesthesia Cardiac arrest 10 (1%) 5 (3%) 15 (1%)
care team, the site principal investigator, or one of the Severe airway trauma 8 (1%) 6 (3%) 14 (1%)
principal investigator’s designees whenever a difficult Death 3 (<1%) 2 (1%) 5 (<1%)
intubation was encountered. We had three methods of Aspiration 1 (<1%) 0 1 (<1%)
data collection and entry: an electronic medical record hot Pneumothorax 1 (<1%) 0 1 (<1%)
key alerted a research assistant by pager whenever a Non-severe complications† 148 (18%) 44 (22%) 192 (19%)
difficult intubation was encountered and the data was Hypoxaemia 65 (8%) 29 (15%) 94 (9%)
attempted to be collected shortly after the intubation; the Minor airway trauma 36 (4%) 8 (4%) 44 (4%)
attending physician completed the data entry sheet Oesophageal intubation with immediate 21 (3%) 11 (6%) 32 (3%)
immediately after encountering a difficult intubation; and recognition
the care team was interviewed and data entry sheet was Laryngospasm 24 (3%) 8 (4%) 32 (3%)
completed after a discovered case via an electronic medical Epistaxis 12 (1%) 2 (1%) 14 (1%)
record search or quality assurance process. In most cases Bronchospasm 7 (1%) 5 (3%) 12 (1%)
the data sheet was collected immediately after the difficult Pharyngeal bleeding 10 (1%) 2 (1%) 12 (1%)
intubation occurred. The site principal investigator or a Arrhythmia 3 (<1%) 1 (1%) 4 (<1%)
member of the research team was responsible for verifying Emesis 4 (<1%) 0 4 (<1%)
the data on the collection forms and entering the data into
Data are n (%). *Outcome is missing in one case in anticipated difficult airway group. †Note that each case can have
the online registry.
more than one complication. The complication rates were not significantly different between the anticipated and the
unanticipated groups; however, severe complications were more commonly observed in the unanticipated difficult
Definitions airway group (p=0·015, χ² test).
We deemed any patient who met our inclusion criteria
Table 2: Airway management outcomes and complications
(category 1–4) as having difficult tracheal intubation.

www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1 3

 Articles

Anticipated difficult airway (n=821) Unanticipated difficult airway (n=197) Total (n=1018)
No complications Complications p value No Complications p value No Complications p value
(n=664) (n=157) complications (n=47) complications (n=204)
(n=150) (n=814)
General
Age (months) 79 (10–156) 34 (8–147) 0·08 25 (4–149) 7 (2–34) 0·009 68 (8–154) 23 (5–131) 0·003
Weight (kg) 20 (8–35) 12 (6–30) 0·007 12 (5–41) 7 (4–20) 0·009 18 (7–35) 11 (5–28) 0·0002
Weight <10 kg 200 (30%) 71 (45%) 0·0003 66 (44%) 28 (60%) 0·06 266 (33%) 99 (49%) <0·0001
Male 369 (56%) 96 (61%) 0·21 108 (72%) 32 (68%) 0·61 477 (59%) 128 (63%) 0·28
ASA status (mean, median, IQR)* 2·9 (3:3–3) 3·1 (3:3–3) 0·02 2·6 (3:2–3) 1·8 (3:2–3) 0·08 2·9 (3:3–3) 3 (3:3–3) 0·01
ASA-Emergency* 30 (5%) 14 (9%) 0·03 8 (6%) 2 (4%) 0·73 38 (5%) 16 (8%) 0·07
Location
Operating room 606 (91%) 144 (92%) 0·86† 128 (86%) 36 (77%) 0·16† 734 (90%) 180 (88%) 0·41†
Off-site location 33 (5%) 3 (2%) .. 14 (9%) 2 (4%) .. 47 (6%) 5 (3%) ..
Intensive care unit 14 (2%) 6 (4%) .. 2 (1%) 1 (2%) .. 16 (2%) 7 (3%) ..
Other 11 (2%) 4 (2%) .. 6 (4%) 8 (17%) .. 17 (2%) 12 (6%) ..
Syndrome
Syndrome diagnosis‡ 498 (75%) 119 (76%) 0·97‡ 62 (42%) 16 (34%) 0·80‡ 560 (69%) 135 (66%) 0·54‡
No syndrome diagnosis 143 (22%) 34 (22%) .. 83 (55%) 27 (57%) 226 (28%) 61 (30%)
Undefined§ 23 (3%) 4 (2%) 5 (3%) 4 (9%) 28 (3%) 8 (4%)
Specific syndrome
Pierre-Robin sequence 94 (14%) 22 (14%) 0·96 4 (3%) 2 (4%) 0·58 98 (12%) 24 (12%) 0·91
Goldenhar syndrome 74 (11%) 11 (7%) 0·13 2 (1%) 0 (0%) 0·43 76 (9%) 11 (5%) 0·07
Difficult airway exam¶
Normal 32 (5%) 11 (7%) 0·27 70 (47%) 17 (36%) 0·21 102 (13%) 28 (14%) 0·65
Abnormal 631 (95%) 146 (93%) .. 80 (53%) 30 (64%) .. 711 (87%) 176 (86%) ..
Specific exam finding
Micrognathia 290 (44%) 71 (45%) 0·73 21 (14%) 12 (26%) 0·07 311 (38%) 83 (41%) 0·52
Limited mouth opening 269 (41%) 58 (37%) 0·41 11 (7%) 7 (15%) 0·12 280 (34%) 65 (32%) 0·49
Cervical spine immobility 137 (21%) 30 (19%) 0·67 3 (2%) 1 (2%) 0·96 140 (17%) 31 (15%) 0·49
Short thyromental distance 77 (12%) 33 (21%) 0·002 9 (6%) 5 (11%) 0·28 86 (11%) 38 (19%) 0·002
Anticipated difficulty with:
Mask ventilation and 94 (14%) 30 (19%) NA NA NA NA 94 (12%) 30 (15%) ··
laryngoscopy
Laryngoscopy only 565 (85%) 126 (80%) NA NA NA NA 565 (69%) 126 (62%) ··
Mask ventilation only 5 (1%) 1 (1%) NA NA NA NA 5 (1%) 1 (0%) ··
Unanticipated 0 (0%) 0 (0%) NA NA NA NA 150 (18%) 47 (23%) 0·13||
Difficult direct laryngoscopy
(inclusion criteria)**
Cormack scale 3–4 215 (32%) 85 (54%) <0·001 143 (95%) 43 (91%) 0·32 358 (44%) 128 (63%) <0·0001
Direct laryngoscopy impossible 169 (25%) 37 (24%) 0·62 0 (0%) 1 (2%) 0·24 169 (21%) 38 (19%) 0·50
Direct laryngoscopy failure within 109 (16%) 25 (16%) 0·88 2 (1%) 0 (0%) 1·00 111 (14%) 25 (12%) 0·60
6 months
Direct laryngoscopy is possible 231 (35%) 29 (18%) <0·001 4 (3%) 2 (4%) 0·58 235 (29%) 31 (15%) <0·0001
but harmful

Data are n (%) or median (IQR), unless otherwise stated. ASA=American Society of Anesthesiology. NA=not applicable. *Datapoints were missing in 17 cases. †p value based on operating room vs outside of
operating room. ‡Syndrome diagnosis was marked as yes if any known syndrome was diagnosed or suspected at the time of airway management; p value based on no syndrome vs syndrome undefined or
confirmed. §Undefined was marked when a syndrome was suspected but not confirmed. ¶Difficult airway exam was performed by attending anaesthesiologists, data missing in one case. ||p value based on
anticipated vs unanticipated difficulty before the case. **Difficult direct laryngoscopy (inclusion criteria): each case may meet more than one criterion. Syndrome diagnosis: marked as yes if any known syndrome
is diagnosed or suspected at the time of airway management.

Table 3: Patient demographics and outcomes

(eg, direct laryngoscope or video laryngoscope) Children (NEAR4KIDS) operational definitions.11,12 The
remained in place. following were categorised as severe complications: severe
We categorised complications as severe and non-severe, airway trauma (glottic or subglottic injury), clinical evidence
modified from the National Emergency Airway Registry for of aspiration (chest radiograph or bronchoscopy evidence),

4 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1

 Articles

Anticipated difficult airway (n=821) Unanticipated difficult airway (n=197) Total (n=1018)
No Complications p value No complications Complications p value No Complications p value
complications (n=157) (n=150) (n=47) complications (n=204)
(n=664) (n=814)
First attempt provider*
Trainee 389 (59%) 99 (63%) 0·053 80 (54%) 29 (62%) 0·39 469 (58%) 128 (63%) 0·18
CRNA 116 (17%) 15 (10%) ·· 37 (25%) 7 (15%) ·· 153 (19%) 22 (11%) ··
Attending anaesthesiologist 139 (21%) 33 (21%) ·· 29 (19%) 9 (19%) ·· 168 (21%) 42 (21%) ··
Otolaryngologist 18 (3%) 8 (5%) ·· 1 (1%) 0 (0%) ·· 19 (2%) 8 (4%) ··
Other 2 (0%) 1 (1%) ·· 2 (1%) 2 (4%) ·· 4 (0%) 3 (1%) ··
Successful provider†
Trainee 312 (47%) 55 (37%) ·· 31 (22%) 3 (6%) ·· 343 (43%) 58 (30%) ··
CRNA 72 (11%) 7 (5%) ·· 6 (4%) 0 (0%) ·· 78 (10%) 7 (4%) ··
Attending anaesthesiologist 236 (36%) 71 (48%) 0·007 96 (68%) 38 (83%) 0·051 332 (41%) 109 (56%) 0·001
Otolaryngologist 39 (6%) 14 (9%) ·· 6 (4%) 4 (9%) ·· 45 (5%) 18 (9%) ··
Other 2 (0%) 2 (1%) ·· 3 (2%) 1 (2%) ·· 5 (1%) 3 (1%) ··
Attending only 107 (16%) 22 (14%) 0·52 29 (19%) 8 (17%) 0·72 136 (17%) 30 (15%) 0·49
Attending experience
Median (year, IQR) 8 (3–20) 7 (3–14) 0·23 8 (4–18) 9 (3–18) 0·77 8 (3–20) 8 (3–17) 0·37
Fellowship training (%)‡ 631 (97%) 151 (97%) 0·68 137 (94%) 44 (94%) 0·73 768 (96%) 195 (97%) 0·96

Data are n (%) unless otherwise stated. CRNA=certified registered nurse anaesthetist. *Data missing in two cases; p value was based on trainee vs non-trainee. †17 intubations were never successful; data were
missing in three cases; p value was based on attending anaesthesiologist vs other. ‡Attending fellowship training status is missing in 17 cases.

Table 4: Provider characteristics

cardiac arrest, emergent surgical airway, oesophageal

1200 Trainee
intubation with delayed recognition, pneumothorax, and CRNA
death. Non-severe complications included minor airway Otolaryngologist
Other anaesthesiologist
trauma (dental or lip), pharyngeal bleeding, arrhythmia 1000
Primary anaesthesiologist
without haemodynamic consequences, bronchospasm,
epistaxis, oesophageal intubation with immediate
800
recognition, hypoxaemia, laryngospasm, and emesis
Number of cases

without aspiration. We defined hypoxaemia as a 10%

decrease from the pre-intubation oxygen saturation for 600
more than 45 s.
We collected other patient, clinician, and practice data 400
related to airway management. These included patient
demographics, airway physical examination findings,
mask ventilation efficacy (assessed with the Han scale13), 200

anaesthesia induction method, airway interventions, and

outcome (intubation success or failure). Patients were
a-priori categorised by weight (<10 kg vs ≥10 kg).14–18
Clinician data included the clinician type (attending,
trainee, or registered nurse anaesthetist), the attending’s
fellowship training status and years of experience (time
since completion of fellowship training). We also
documented the patients’ ASA physical status
classification (1–6), a system to describe the patients’
medical condition before their anaesthetic.19 The addition
of E to the classification denotes an emergent procedure.
Statistical analysis
We did our statistical analysis with Stata (version 11.2).
Our sample included all available data during the study
period. We report summary statistics using means and
standard deviation for parametric variables and medians
0
1st 2nd 3rd 4th 5th
Tracheal intubation attempt
Figure 1: Proportion of airway providers for each attempt
CRNA=certified registered nurse anaesthetist.
with interquartile ranges for non-parametric variables.
We used histograms of the variable distributions to
determine normality. We used a contingency table
method with χ² test for categorical variables with a
dichotomous outcome. For non-parametric variables, we
used the Wilcoxon rank sum test. For multivariable
analysis, occurrence of any complications was a
dichotomous dependent variable, multiple attempts
(≤2 attempts vs >2 attempts) was an independent variable,
and factors associated with multiple attempts in

www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1 5

 Articles

Anticipated difficult airway (n=821) Unanticipated difficult airway (n=197) Total (n=1018)
No Complications p value No Complications p value No Complications p value
complications (n=157) complications (n=47) complications (n=204)
(n=664) (n=150) (n=814)
Induction technique 0·36 0·15 0·083
Mask induction 430 (65%) 99 (63%) ·· 93 (62%) 24 (51%) ·· 523 (64%) 123 (60%) ··
Intravenous induction 180 (27%) 43 (27%) ·· 53 (36%) 20 (43%) ·· 233 (29%) 63 (31%) ··
Intravenous sedation 36 (6%) 8 (5%) ·· 2 (1%) 0 (0%) ·· 38 (5%) 8 (4%) ··
Tracheal induction 9 (1%) 1 (1%) ·· 0 (0%) 0 (0%) ·· 9 (1%) 1 (0%) ··
NA 9 (1%) 6 (4%) ·· 2 (1%) 3 (6%) ·· 11 (1%) 9 (4%) ··
Anaesthesia approach 0·11 0·04 0·005
General 610 (92%) 145 (92%) ·· 145 (97%) 44 (94%) ·· 755 (93%) 189 (93%) ··
Sedation 47 (7%) 7 (4%) ·· 3 (2%) 0 (0%) ·· 50 (6%) 7 (3%) ··
Awake 6 (1%) 4 (3%) ·· 2 (1%) 0 (0%) ·· 8 (1%) 4 (2%) ··
None 1 (0%) 1 (1%) ·· 0 (0%) 3 (6%) ·· 1 (0%) 4 (2%) ··
Intubation route* 0·12 0·22 0·088
Oral 488 (74%) 109 (73%) ·· 128 (90%) 40 (87%) ·· 616 (77%) 149 (76%) ··
Nasal 166 (25%) 37 (25%) ·· 10 (7%) 5 (11%) ·· 176 (22%) 42 (21%) ··
Surgical 0 (0%) 1 (0%) ·· 0 (0%) 1 (2%) ·· 0 (0%) 2 (1%) ··
Other 5 (1%) 3 (2%) ·· 4 (3%) 0 (0%) ·· 9 (1%) 3 (2%) ··
First attempt device† 0·006 0·30 0·001
Direct laryngoscope 220 (33%) 69 (45%) ·· 128 (87%) 44 (94%) ·· 348 (43%) 113 (56%) ··
Flexible fibreoptic bronchoscope 226 (34%) 51 (33%) ·· 5 (3%) 2 (4%) ·· 231 (29%) 53 (26%) ··
Glidescope 158 (24%) 19 (12%) ·· 5 (3%) 1 (2%) ·· 163 (20%) 20 (10%) ··
Other or combined 57 (9%) 15 (10%) ·· 10 (7%) 0 (0%) ·· 67 (8%) 15 (7%) ··
Successful device‡ 0·06 0·49 0·037
Direct laryngoscope 41 (6%) 17 (12%) ·· 46 (34%) 17 (37%) ·· 87 (11%) 34 (18%) ··
Flexible fibreoptic bronchoscope 281 (43%) 69 (47%) ·· 22 (16%) 6 (13%) ·· 303 (39%) 75 (39%) ··
Glidescope 241 (37%) 42 (29%) ·· 48 (36%) 13 (28%) ·· 289 (37%) 55 (29%) ··
Other or combined 89 (14%) 18 (12%) ·· 18 (13%) 10 (22%) ·· 107 (14%) 28 (15%) ··
Neuromuscular blockade use 268 (40%) 63 (40%) 0·96 72 (48%) 30 (64%) 0·06 340 (42%) 93 (46%) 0·32

Data are n (%), unless otherwise stated. NA=not applicable. *Data are missing in 19 cases. †Data are missing in eight cases. ‡Intubation was not successful in 17 intubations.

Table 5: Association between practice characteristics and complications

univariate analysis as covariates. We deemed a p value of
less than 0·05 as statistically significant.
Role of the funding source
There was no funding source for this study. The
corresponding author had full access to all the data in the
study and had final responsibility for the decision to
submit for publication.
Results
The study period included data collected from our centres
between Aug 6, 2012, and Jan 31, 2015. The 13 centres
reported 1061 cases of tracheal intubation during this
period. The four largest centres contributed most of the
cases in the registry; table 1 reports the distribution of
cases across centres. One centre was excluded because of
lack of full compliance with data capture rates and audits,
leaving 1018 cases for analysis (table 1). Three centres
reported 112, 164, and 20 cases of difficult tracheal
intubation in 2014 with anaesthetic case volumes that year
of 38 813, 35 000, and 8000, respectively. This finding
translates to 0·28%, 0·47%, and 0·25% of anaesthetics,
respectively suggesting a range of 2–5 difficult tracheal
intubations in 1000 anaesthetised children. 80% (821/1018)
of the difficult tracheal intubation cases were anticipated.
20% (204/1018) of cases had at least one complication.
Cardiac arrest (n=15; 2%) was the most common severe
complication, and hypoxaemia (n=94; 9%) was the most
common non-severe complication. All cardiac arrests
were preceded by hypoxaemia. Table 2 shows tracheal
intubation-related outcomes. Four of the cardiac arrests
were non-elective urgent intubations; two cases in the
intensive care unit and two in the emergency room. 11 of
the cardiac arrest cases were elective non-urgent tracheal
intubations. Five patients died shortly after tracheal
intubation (within 7 days), four were unrelated to tracheal
intubation (cardiac arrest and withdrawal of care), and one
was related to difficult tracheal intubation (hypoxaemic
arrest). The complication rates were not significantly
different between the anticipated and unanticipated

6 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1

 Articles

Anticipated difficult airway (n=821) Unanticipated difficult airway (n=197) Total (n=1018)
No Complications p value No Complications p value No complications Complications p value
complications (n=157) complications (n=47) (n=814) (n=204)
(n=664) (n=150)
Mask ventilation <0·0001 0·03 <0·0001
Easy mask ventilation 430 (65%) 66 (42%) ·· 118 (79%) 30 (64%) ·· 548 (67%) 96 (47%) ··
Airway adjunct needed 119 (18%) 39 (25%) ·· 24 (16%) 11 (23%) ·· 143 (18%) 50 (25%) ··
Difficult mask ventilation 39 (6%) 28 (18%) ·· 4 (3%) 6 (13%) ·· 43 (5%) 34 (17%) ··
Impossible for mask 0 (0%) 5 (3%) ·· 1 (<1%) 0 (0%) ·· 1 (<1%) 5 (2%) ··
ventilation
Not attempted 76 (11%) 19 (12%) 3 (2%) 0 (0%) 79 (10%) 19 (9%)
Extraglottic airway* <0·0001 0·45 <0·0001
Not attempted 553 (84%) 100 (64%) ·· 120 (80%) 37 (79%) ·· 673 (83%) 137 (67%) ··
Easy placement and 96 (14%) 34 (22%) ·· 23 (15%) 6 (13%) ·· 119 (15%) 40 (20%) ··
ventilation
Easy placement and poor 10 (2%) 11 (7%) ·· 5 (3%) 3 (6%) ·· 15 (2%) 14 (7%) ··
ventilation (tidal volume
<5 mL/kg)
Difficult placement and 2 (<1%) 4 (2%) ·· 1 (1%) 0 (0%) ·· 3 (<1%) 4 (2%) ··
easy ventilation
Difficult placement and 0 (0%) 6 (4%) ·· 0 (0%) 1 (2%) ·· 0 (0%) 7 (3%) ··
difficult ventilation
Impossible to place 1 (0%) 2 (1%) ·· 1 (1%) 0 (0%) ·· 2 (<1%) 2 (1%) ··
Number of attempts 2 (1–3) 3 (2–4) <0·0001 3 (3–5) 4 (3–6) 0·009 2 (1–3) 3 (2–5) <0·0001
(median, IQR)†
1 271 (41%) 23 (15%) ·· 10 (7%) 2 (4%) ·· 281 (34%) 25 (13%) ··
2 181 (27%) 33 (21%) ·· 20 (13%) 4 (9%) ·· 201 (25%) 37 (18%) ··
3 107 (16%) 36 (23%) ·· 46 (31%) 9 (19%) ·· 153 (19%) 45 (22%) ··
4 60 (9%) 26 (16%) ·· 32 (21%) 11 (23%) ·· 92 (11%) 37 (18%) ··
5 22 (4%) 11 (7%) ·· 23 (15%) 7 (15%) ·· 45 (6%) 18 (9%) ··
6 8 (1%) 8 (5%) ·· 7 (5%) 3 (6%) ·· 15 (2%) 11 (5%) ··
7 or more 15 (2%) 20 (13%) ·· 12 (8%) 11 (24%) ·· 27 (3%) 31 (15%) ··

Data are n (%), unless otherwise stated. *Extraglottic airway: data are missing in two cases. †p value calculated by Wilcoxon rank-sum test.

Table 6: Process characteristics

groups; however, severe complications were more
common in the unanticipated group (χ² test p=0·015;
table 2).
Patients who weighed less than 10 kg had more tracheal
intubations with complications (49% [99/204]) than
tracheal intubations without complications (33%
[266/814]; table 3). This was true in both the anticipated
and unanticipated groups. Patients entered in the registry
because of a failed direct laryngoscopy by the attending
anaesthesiologist. Patients in category 1 (cases with
Cormack classification 3 and 4) had a greater proportion
of complications (n=128; 63%; p<0·0001), and patients in
category 4 (suspected difficulty without attending direct
laryngoscopy confirmation) had fewer complications
(n=31 [15%]; p<0·0001; table 3).
The median number of years of experience of the
supervising anaesthesiologist was 8 years (IQR 3–20;
table 4). Although attending physicians made 21% of
first intubation attempts, they were the eventual
successful clinician in 44% of cases, which meant they
often took over the intubation from the other clinicians
(trainee and certified registered nurse anaesthetist).
Figure 1 shows the distribution of clinician types
attempting the first five tracheal intubations in patients
who had five or more attempts. Of note trainees made
20% of fifth tracheal intubation attempts. The first
attempt provider type was similar between anticipated
and unanticipated difficult airway cases. The attending
anaesthesiologist was the more successful provider in
unanticipated difficult airway cases than in anticipated
difficult airway cases.
Inhalational induction of anaesthesia via facemask was
the most common induction method used (64%
[646/1018]; table 5). 77 (8%) of 1018 cases were difficult to
ventilate by facemask. Six patients were impossible to
ventilate by facemask. This difficulty was anticipated in
five of these patients. Two patients received neuro-
muscular blockade without improvement in the ability to
ventilate. Two patients were ventilated successfully using
an extraglottic airway device after facemask ventilation
failed. Tracheal intubation was successful in all six patients
(table 6). The median number of tracheal intubation

www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1 7

 Articles

(95% CI 1·4–1·6; p<0·0001). The median number of

0·6
attempts was 3 (IQR 2–5) for cases with complications
while the median was two attempts (IQR 1–3) for cases
Proportion of cases with complications

0·5
without complications (p<0·0001; table 6). Oxygen was
0·4 rarely given during tracheal intubation attempts. The
approaches reported included nasal cannula in 17 tracheal
0·3 intubations, modified nasopharyngeal airway in
76 tracheal intubations, and an endoscopy mask in four
0·2 tracheal intubations.
Tracheal intubation failed in 2% (19/1017) of cases
0·1
(table 2). The success rate of the first tracheal intubation
0 attempt in the registry cohort was only 30% (n=288);
1 2 3 4 5 6 ≥7 however, 98% of the patients were eventually successfully
Number of tracheal intubation attempts intubated. Figure 3 shows the distributions of devices
Figure 2: Proportion of cases with complications by the number of attempts used for each attempt. 46% (461/1010) of cases used
for airway management direct laryngoscopy for the first tracheal intubation
The average proportion of cases with complications by the number of required attempt, while it was the successful device in 12%
intubation attempts. Each error bar represents mean (SD). (121/978) of cases. Flexible fibre-optic laryngoscope was
the first attempt device in 28% (284/1010) of patients,
Other
whereas it was the successful device in 37% (378/978) of
1200
Glidescope cases. The most frequently attempted first tracheal
Flexible fibre-optic bronchoscope intubation techniques were direct laryngoscopy (direct
Direct laryngoscope
1000 laryngoscopy; n=461 [46%]), indirect video laryngoscopy
(n=183 [18%]), and fibre-optic bronchoscopy (n=284
800
[28%]) with success rates of 16 (3%) of 461, 101 (55%) of
183, and 153 (54%) of 284, respectively. An extraglottic
Number of cases

airway was attempted in 208 (20%) of 1016 cases.

600
400
200
0
1st
Figure 3: Type of device for each attempt
Other includes Airtraq, Karl Storz C-MAC, optical stylet, Anterior Commissure
scope, and others.
See Online for appendix
attempts was 2·5 (IQR 2–6) in these six patients. Three
patients were intubated by otolaryngologists—one with a
rigid bronchoscope and two with standard direct
laryngoscopy with a grade 4 view. Flexible fibre-optic
intubation was done successfully through an extraglottic
airway in two patients and the GlideScope Cobalt was
used in one patient. Two of these patients had cardiac
arrest during tracheal intubation attempts and were
successfully resuscitated.
43% (433/1018) of patients in the registry were managed
with neuromuscular blockade. Cases where direct
laryngoscopy was attempted for the first three attempts
(direct laryngoscopy persistence) had more complications
(48/158 [30%] vs 62/297 [21%]; p=0·02). The occurrence of
a complication was associated with the number of
attempts (figure 2): odds ratio (OR) 1·5 per attempt
Ventilation with the extraglottic airway was poor in 17%
(36/208) cases in which it was attempted, and extraglottic
airway placement was impossible in 2% (4/208) of
insertion attempts (table 6). Of the 36 failed extraglottic
airways 17% (14/84) received neuromuscular blockade
and 19% (22/118) received no neuromuscular blockade.
There were 455 cases with direct laryngoscopy as the
2nd 3rd 4th 5th
initial approach that required three or more tracheal
Tracheal intubation attempt
intubation attempts. Of these cases, an early transition
(after the first or second direct laryngoscopy attempt) to a
non-direct laryngoscopy technique was associated with
fewer complications (62/297 [21%]) than cases with late
transition (>2 direct laryngoscopy attempts; 48/158 [30%];
p=0·024). The appendix shows associations between the
first attempted device and patient and provider
characteristics (appendix p 1).
After adjusting for clustering by site, patient factors,
weight less than 10 kg, thyromental distance, ASA
physical status, inclusion criteria, and first attempt
device, many attempts (>2) were independently
associated with the occurrence of any complication
(OR 3·1, 95% CI 2·1–4·6; p<0·0001; table 7)
Discussion
This prospective multicentre study estimated that
difficult tracheal intubation occurred in 2–5 per
1000 paediatric anaesthesia cases in large academic
centres in the USA. Furthermore, this registry showed
that 20% of children with difficult tracheal intubations
had a complication. We identified the following

8 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1

 Articles

Univariate analysis Multivariate analysis

OR (95% CI) p value OR (95% CI) p value
Multiple attempt (>2 attempts) 3·48 (2·48–4·89) <0·0001 3·08 (2·08–4·57) <0·0001
Weight <10 kg 2·09 (1·51–2·88) <0·0001 1·61 (1·13–2·31) 0·009
Physical exam: short thyromental distance 1·96 (1·28–3·00) 0·002 2·72 (1·71–4·35) <0·0001
ASA status
ASA-1 Reference Reference
ASA-2 1·60 (0·44–5·74) 0·47 1·97 (0·53–7·27) 0·31
ASA-3 1·36 (0·39–4·72) 0·63 1·92 (0·54–6·85) 0·32
ASA-4 3·22 (0·89–11·57) 0·07 4·04 (1·07–15·20) 0·038
Difficult direct laryngoscopy as inclusion criteria 2·31 (1·66–3·21) <0·0001 2·01 (1·20–3·37) 0·008
(Cormack-Lehane 3–4 or failed)
First attempt device
Direct laryngoscope Reference Reference
Flexible fiberoptic bronchoscope 0·66 (0·45–0·97) 0·03 2·13 (1·19–3·81) 0·011
Glidescope 0·37 (0·22–0·63) <0·0002 0·92 (0·48–1·76) 0·79
Other device 0·80 (0·43–1·49) 0·48 1·13 (0·57–2·25) 0·72

Univariate and multivariate analysis with multiple attempts (≤2 attempts vs >2 attempts), weight (<10 kg vs ≥10 kg), American Society of Anesthesia (ASA) physical status,
Direct laryngoscopy failure or Cormack-Lehane 3–4 as inclusion criteria, and first attempt device, adjusted for site-level clustering with random-effect model. Univariate analysis
is adjusted for clustering by site. 993 cases in the multivariate analysis (ASA status was missing in 17 cases, device data was missing in eight cases and one case with ASA-5 status
completely predicted outcome, therefore dropped from analysis). Overall model was significant (p<0·0001). OR=odds ratio. ASA=American Society of Anesthesia.

Table 7: Multivariate analysis for patient, provider, and process characteristics

associations with any complication (severe or non-
severe): multiple tracheal intubation attempts (>2),
weight less than 10 kg, short thyromental distance,
abnormal airway physical examination, and persistent
direct laryngoscopy attempts (for the first three tracheal
intubation attempts). Multivariable analysis confirmed
the independent association of multiple attempts with
complications (severe or non-severe). Furthermore, we
noted an incremental increase in the occurrence of
complications with each additional tracheal intubation
attempt.
The high percentage of difficult tracheal intubations
that were anticipated (80%) probably meant that the team
was prepared with the appropriate equipment and
personnel and had information about the previously
successful tracheal intubation approach. In the
unanticipated difficult tracheal intubation group, teams
were probably unprepared and not ready with the
optimum equipment. Despite these population
differences, increasing attempts and lower weight were
associated with more complications in both groups.
Understandably unanticipated difficult tracheal
intubations had more severe complications and more
tracheal intubation attempts and the attending
anaesthesiologist was more likely to be the successful
provider. These findings suggest that preparation matters
and might lower the likelihood of severe complications.
Hypoxaemia was the most common non-severe
complication related to tracheal intubation and occurred
in 9% of tracheal intubation attempts. Strategies to
prevent hypoxaemia during tracheal intubation include
pre-oxygenation via facemask and passive oxygen
administration during the attempt. Facemask pre-
oxygenation is routinely done before tracheal intubation
and is consistent with the ASA guideline to “administer
facemask pre-oxygenation before initiating management
of a difficult airway”.20,21 The ASA further states that
practitioners should “actively pursue opportunities to
deliver supplemental oxygen throughout the process of
difficult airway management”. Although pre-oxygenation
is presumably practised, few practitioners in participating
centres provided supplemental oxygen during the actual
intubation (10%). Children have higher oxygen
consumption rates than adults and their rate of arterial
oxygen desaturation when apnoeic is consequently much
faster. This rapid desaturation rate creates a time
pressure to intubate these children. This pressure is
heightened by the common practice of intubation by
trainees with senior supervision. When hypoxaemia
occurs during tracheal intubation, the team has to
interrupt the intubation attempt to ventilate the patient.
This results in more tracheal intubation attempts to
secure the airway. We speculate that passive oxygen
administration during tracheal intubation might reduce
the number of tracheal intubation attempts and the
incidence of hypoxaemia. Anaesthesiologists often
leverage the benefits of passive oxygenation during
routine anaesthetic cases such as rigid bronchoscopy for
airway examinations. Passive oxygenation during rigid
bronchoscopy increases the time available for the
otolaryngologist to complete the airway examination
without oxygen desaturation. Likewise, passive
oxygenation (eg, via nasal cannula, modified nasal airway
or extraglottic airway) during the intubation will delay

www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1 9

 Articles
the onset of hypoxaemia and provide more time for the
team to secure the tracheal tube with fewer overall
tracheal intubation attempts. Passive oxygenation is
effective in adults and was anecdotally reported to be
effective in children with pyloric stenosis requiring rapid
sequence intubation; we plan to explore the effectiveness
more systematically in the future.22,23
The 15 children with difficult tracheal intubation who
required cardiopulmonary resuscitation for cardiac arrest
translates to an incidence of one in 68 cases experiencing
cardiac arrest. This cardiac arrest incidence is substantially
higher than the 1·4 in 10 000 reported in the general
paediatric anaesthesiology population.24 11 of these
procedures were non-emergent tracheal intubations in
the operating room. This unexpectedly high incidence is
a potential target for quality improvement. Reduction of
hypoxaemia and persistent direct laryngoscopy, and the
introduction of a protocol in which the most experienced
laryngoscopist intervenes quickly when a trainee fails
might reduce the incidence of cardiac arrest.
We studied cases with five tracheal intubation attempts
to understand the role the device and personnel had in
the airway management. Standard direct laryngoscopy
had a poor overall success rate (12%), but it still accounted
for 21% of fifth attempts (figure 3). Attending physicians
made only 21% of first tracheal intubation attempts and
yet were the successful laryngoscopist more than 44% of
the time indicating that they rescued the other clinicians
(trainee, certified registered nurse anaesthetist)
frequently. If these additional attempts were for the
educational benefit of the trainees, the experience gained
by trainees has to be carefully weighed against the
consequences of multiple tracheal intubation attempts.
The transition to the most experienced clinician should
happen quickly in these patients. These data could be
relevant to paediatric practitioners in other acute care
areas including emergency departments, intensive care
units, and intubations in the field.
The original ASA practice guidelines for the management
of patients with a difficult airway defined a difficult tracheal
intubation as three failed conventional attempts and
recommended that subsequent attempts use alternate
devices.20 Our data suggest that this guideline was often
not followed because repeated attempts with conventional
direct laryngoscopy happened in many patients (figure 3).
Could the use of an alternative technique after the failed
attempt expedite intubation with fewer complications?
Our data showed the tracheal intubations with late
transition from repeated direct laryngoscopy attempts
were associated with higher occurrence of complications.
Patients in category 1 had a greater proportion of overall
complications than those in category 4. This finding might
suggest that patients in category 4 might have been easier
to intubate (their difficulty was suspected and not
diagnosed by direct laryngoscopy) or the fact that they did
not have direct laryngoscopy translated to fewer attempts
and less complications.
10 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1
Other investigators have shown that repeated
intubation attempts are associated with adverse events.
Mort and colleagues25 noted that adult emergency room
patients with more than two attempts had more
hypoxaemia, oesophageal intubation, and cardiac arrest.
Our study examined a different population (children),
under controlled conditions (mostly elective operating
room intubations) done by more experienced clinicians
(mainly paediatric anaesthesiology fellows and
attendings), but the effect of more than two tracheal
intubation attempts is similar. Graciano and colleagues11
also reported similar findings in the paediatric intensive
care unit. Neonatal and paediatric intensive care data
suggest that inexperienced clinicians, frequent attempts,
and younger patients were associated with greater
complication rates. Based on our results, using a specific
airway checklist may reduce adverse events in these
patients. Deriving and implementing a standardised
airway bundle checklist has been described in previous
studies and seems like a natural progression of our
work.26 The airway checklist could incorporate the plan to
use an extraglottic airway early, consider and prepare for
passive oxygenation during tracheal intubation attempts,
and restrict the number of direct laryngoscopy attempts.
Our study has several limitations. First, although
13 paediatric centres contributed data into the registry,
four large centres contributed most of the cases. The
specialised nature of the participating institutions and the
proportion of cases reported by a small number of centres
might limit the ability to generalise the results and their
implications. Second, although we established a standard
operational plan including rigorous data collection and
audit processes, cases could still have been missed if
difficult tracheal intubations were not documented in
patient’s charts or reported by the care team. Missed cases
could lead to under-reporting of multiple attempts or
occurrence of complications. Furthermore tracheal
intubation attempts and complications data are probably
an under-representation of the true numbers because they
were not independently collected and were self reported
by the team. A video-based assessment of tracheal
intubations in a paediatric emergency department showed
that clinicians reported fewer adverse events than were
recorded on independent video review. This result is
probably a major limitation of our study and implies that
we might have under-reported our complications.27 Third,
we classified training level as trainees versus attendings,
and we did not capture the actual trainee’s postgraduate
year or whether the trainee was a resident or fellow.
Fourth, patients in category 4 of our inclusion criteria
were deemed difficult tracheal intubations at the discretion
of the attending based on physical examination. The
anaesthesiologist did not perform direct laryngoscopy in
these patients, and it is possible that they might have been
intubated easily with conventional direct laryngoscopy.
Next, our statistical analysis assumes that each of the
1018 tracheal intubations are independent; however, if

there is any dependency between tracheal intubations (eg,
repeat patients), there might be less information in our
dataset than we think. Although there were probably
repeat patients, we believe that treating each tracheal
intubation as independent is reasonable in view of the fact
that the airway in children is in transition and because of
this constant change, conditions during one tracheal
intubation can be different from a subsequent tracheal
intubation. This idea is well known in paediatric
anaesthesia practice—eg, children with Pierre Robin
sequence can be extremely difficult to intubate at birth but
often become easier to intubate as they grow. Finally, we
estimated the rate of difficult tracheal intubation to be
roughly 0·2–0·5% of anaesthetic cases. Our centres
tracked the number of anaesthetics done every year not
the number of tracheal intubations done in anaesthetised
patients. Because many patients are not intubated when
anaesthetised, our calculations are an approximation of
the true rate.
In conclusion, we have defined the incidence and
types of complications in children with difficult tracheal
intubation across 12 paediatric centres and reported the
tracheal intubation success rates of commonly used
tracheal intubation devices and techniques. 20% of the
patients meeting difficult tracheal intubation criteria
had at least one complication. Taken together, we have
learned that attempting more than two tracheal
intubations in children with difficult tracheal intubation
is associated with a high failure rate and increased
incidence of complications. This high complication rate
is multifactorial in origin. The patient-level factors
include abnormal airway physical examination and
lower patient age and weight. Process factors included
multiple tracheal intubation attempts and multiple
direct laryngoscopy attempts. Additionally, delays in
transition from trainee to attending tracheal intubation
attempts could have contributed to complications. We
are now poised to create and disseminate a quality
improvement bundle to address these complications
across our centres. Our findings should affect the clinical
decisions of all clinicians who perform this potentially
life-saving procedure in children with difficult tracheal
intubation.
Contributors
JEF, NJ, DMP, MEM, PS, VMN, FXM, RSL, MAR, PGK, PIR, and AIH
contributed to manuscript authorship, study design, data interpretation,
and data contribution; AN contributed to statistical analysis, study
design, data interpretation, and data contribution; and RSG contributed
to manuscript authorship, study design, and data interpretation.
PeDI collaborative group
Paul Stricker, Brad Taicher, Vidya Raman, Ralph Beltran, Ian Lewis,
Bishr Haydar, Prabhat Koppera, Shobha Malviya, Katherine Gentry,
Adrian Bosenberg, Patrick Olomu, Edgar Kiss, Judit Szolnoki,
Jennifer Zieg, Madhankumar Sathyamoorthy, Vikram Patel,
Codruta Soneru, Ricardo Falcon, Jimmy Windsor, Tim Peterson,
Raymond Park, James Peyton, Denise Chan, Chris Glover, Paul Hopkins,
Somaletha T Bhattacharya, Ricardo Riveros, Nicholas Dalesio,
Martina Richtsfeld, Kumar Belani, Bin Zhang, John McCloskey,
Olutoyin A Olutoye, Christopher Estrada, and Tarun Bhalla.
www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1
Articles
Participating centres
Department of Anesthesia and Critical Care Medicine, Children’s Hospital
of Philadelphia, PA, USA; Department of Anesthesiology Perioperative and
Pain Medicine, Boston Children’s Hospital, Boston, MA, USA; Department
of Anesthesiology and Pain Management, UT Southwestern Medical
Center, Dallas, TX, USA; Department of Anesthesiology, Seattle Children’s
Hospital, WA, USA; Department of Anesthesiology, Texas Children’s
Hospital, Houston, TX, USA; Department of Pediatric Anesthesiology,
University of Michigan Health Center, MI, USA; Department of
Anesthesiology, University of Texas Medical School at Houston, Houston,
TX, USA; Department of Anesthesiology, Children’s Hospital of Colorado,
CO, USA; Department of Pediatric Anesthesiology, Lurie Children’s
Hospital Chicago, Chicago, IL, USA; Department of Anesthesiology and
Pain Medicine, Nationwide Children’s Hospital, Columbus, OH, USA;
Department of Anesthesiology, the University of Mississippi Medical
Center, Jackson, MI, USA; Department of Anesthesiology, Duke University,
Durham, NC, USA; and Department of Anesthesiology and Pain
Management, Children’s Hospital of Cleveland Clinic, Cleveland, OH, USA
Declaration of interests
RG discloses patents in the public domain relevant to airway
management, FXM discloses grant funding from the Agency for
Healthcare Research Quality and the Laerdal Foundation Center for
Excellence, and AN discloses grant funding from the Agency for
Healthcare Research Quality and the Laerdal Foundation Center for
Excellence. All other authors declare no competing interests.
Acknowledgments
We thank the following individuals for their contributions to this work:
Bill Greeley, Mark Schreiner, Jeff Feldman, Bin Zhang, Bishr Haydar,
Kenneth Peeples, Ariel Vincent, Thea Goebel, Allison Wright,
Shana Emery, Terri Voepel-Lewis, Sinead Rivard, Roxana Ploski,
Jennifer Spears, John Hajduk, Ranu Jain, Rachel Bernier, Mark Breibart,
Stephanie Coulombre, Ian Lewis, and Jessica Burkhardt.
References
1 Fiadjoe J, Stricker P. Pediatric difficult airway management:
current devices and techniques. Anesthesiol Clin 2009; 27: 185–95.
2 Fiadjoe JE, Gurnaney H, Dalesio N, et al. A prospective randomized
equivalence trial of the GlideScope Cobalt® video laryngoscope to
traditional direct laryngoscopy in neonates and infants.
Anesthesiology 2012; 116: 622–28.
3 Vanderhal AL, Berci G, Simmons CF Jr, Hagiike M.
A videolaryngoscopy technique for the intubation of the newborn:
preliminary report. Pediatrics 2009; 124: e339–46.
4 Long E, Sabato S, Babl FE. Endotracheal intubation in the pediatric
emergency department. Paediatr Anaesth 2014; 24: 1204–11.
5 Morray JP, Geiduschek JM, Caplan RA, Posner KL, Gild WM,
Cheney FW. A comparison of pediatric and adult anesthesia closed
malpractice claims. Anesthesiology 1993; 78: 461–67.
6 Jimenez N, Posner KL, Cheney FW, Caplan RA, Lee LA,
Domino KB. An update on pediatric anesthesia liability: a closed
claims analysis. Anesth Analg 2007; 104: 147–53.
7 Cook TM, Woodall N, Harper J, Benger J, and the Fourth National
Audit Project. Major complications of airway management in the
UK: results of the Fourth National Audit Project of the Royal College
of Anaesthetists and the Difficult Airway Society. Part 2: intensive
care and emergency departments. Br J Anaesth 2011; 106: 632–42.
8 Cook TM, Woodall N, Frerk C, and the Fourth National Audit
Project. Major complications of airway management in the UK:
results of the Fourth National Audit Project of the Royal College of
Anaesthetists and the Difficult Airway Society. Part 1: anaesthesia.
Br J Anaesth 2011; 106: 617–31.
9 Cormack RS, Lehane J. Difficult tracheal intubation in obstetrics.
Anaesthesia 1984; 39: 1105–11.
10 Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG.
Research electronic data capture (REDCap)--a metadata-driven
methodology and workflow process for providing translational
research informatics support. J Biomed Inform 2009; 42: 377–81.
11 Graciano AL, Tamburro R, Thompson AE, Fiadjoe J, Nadkarni VM,
Nishisaki A. Incidence and associated factors of difficult tracheal
intubations in pediatric ICUs: a report from National Emergency
Airway Registry for Children: NEAR4KIDS. Intensive Care Med 2014;
40: 1659–69.
11
 Articles
12 Nishisaki A, Turner DA, Brown CA 3rd, Walls RM, Nadkarni VM,
and the National Emergency Airway Registry for Children
(NEAR4KIDS), and the Pediatric Acute Lung Injury and Sepsis
Investigators (PALISI) Network. A National Emergency Airway
Registry for children: landscape of tracheal intubation in 15 PICUs.
Crit Care Med 2013; 41: 874–85.
13 Han R, Tremper KK, Kheterpal S, O’Reilly M. Grading scale for
mask ventilation. Anesthesiology 2004; 101: 267.
14 Benumof JL, Dagg R, Benumof R. Critical hemoglobin desaturation
will occur before return to an unparalyzed state following 1 mg/kg
intravenous succinylcholine. Anesthesiology 1997; 87: 979–82.
15 Denman WT, Goudsouzian NG. Position of the laryngeal mask
airway. Anesthesiology 1992; 77: 401–02.
16 Dubreuil M, Laffon M, Plaud B, Penon C, Ecoffey C. Complications
and fiberoptic assessment of size 1 laryngeal mask airway.
Anesth Analg 1993; 76: 527–29.
17 Park C, Bahk JH, Ahn WS, Do SH, Lee KH. The laryngeal mask
airway in infants and children. Can J Anaesth 2001; 48: 413–17.
18 Duggan LJN. Unique airway issues in the pediatric population:
McGraw Hill, 2012.
19 Sankar A, Johnson SR, Beattie WS, Tait G, Wijeysundera DN.
Reliability of the American Society of Anesthesiologists physical
status scale in clinical practice. Br J Anaesth 2014; 113: 424–32.
20 Practice guidelines for management of the difficult airway. A report
by the American Society of Anesthesiologists Task Force on
Management of the Difficult Airway. Anesthesiology 1993; 78: 597–602.
21 Apfelbaum JL, Hagberg CA, Caplan RA, et al, and the American
Society of Anesthesiologists Task Force on Management of the
Difficult Airway. Practice guidelines for management of the difficult
airway: an updated report by the American Society of
Anesthesiologists Task Force on Management of the Difficult
Airway. Anesthesiology 2013; 118: 251–70.
12 www.thelancet.com/respiratory Published online December 16, 2015 http://dx.doi.org/10.1016/S2213-2600(15)00508-1
22 Bhagwan SD. Levitan’s no desat with nasal cannula for infants with
pyloric stenosis requiring intubation. Paediatr Anaesth 2013;
23: 297–98.
23 Wimalasena Y, Burns B, Reid C, Ware S, Habig K. Apneic
oxygenation was associated with decreased desaturation rates
during rapid sequence intubation by an Australian helicopter
emergency medicine service. Ann Emerg Med 2015; 65: 371–76.
24 Morray JP, Geiduschek JM, Ramamoorthy C, et al. Anesthesia-
related cardiac arrest in children: initial findings of the Pediatric
Perioperative Cardiac Arrest (POCA) Registry. Anesthesiology 2000;
93: 6–14.
25 Mort TC. Emergency tracheal intubation: complications associated
with repeated laryngoscopic attempts. Anesth Analg 2004;
99: 607–13.
26 Li S, Rehder KJ, Giuliano JS Jr, et al, and the National Emergency
Airway Registry for Children (NEAR4KIDS) Investigators and
Pediatric Acute Lung Injury and Sepsis Investigator (PALISI)
Network Investigators. Development of a quality improvement
bundle to reduce tracheal intubation-associated events in pediatric
ICUs. Am J Med Qual 2014; published online Aug 20.
DOI:10.1177/1062860614547259.
27 Kerrey BT, Rinderknecht AS, Geis GL, Nigrovic LE, Mittiga MR.
Rapid sequence intubation for pediatric emergency patients: higher
frequency of failed attempts and adverse effects found by video
review. Ann Emerg Med 2012; 60: 251–59.