Escolar Documentos
Profissional Documentos
Cultura Documentos
Summary
Background: Laryngospasm is a common and often serious adverse
respiratory event encountered during anesthetic care of children. We
examined, in a case control design, the risk factors for laryngospasm
in children.
Material and Methods: The records of 130 children identified as having
experienced laryngospasm under general anesthesia were examined.
Cases were identified from those prospectively entered into the Mayo
Clinic performance improvement database between January 1, 1996
and December 31, 2005. Potential demographic, patient, surgical and
anesthetic related risk factors were determined in a 1 : 2 case–control
study.
Results: No individual demographic factors were found to be signi-
ficantly associated with risk for laryngospasm. However, multivariate
analysis demonstrated significant associations between laryngospasm
and intercurrent upper respiratory infection (OR 2.03 P = 0.022) and
the presence of an airway anomaly (OR = 3.35, P = 0.030). Among
those experiencing laryngospasm during maintenance or emergence,
the use of a laryngeal mask airway was strongly associated even when
adjusted for the presence of upper respiratory infection and airway
anomaly (P = 0.019). Ten patients experienced postoperatively one or
more complications whereas only three complications were observed
among controls (P = 0.008). No child required cardiopulmonary
resuscitation and there were no deaths in either study cohort.
Conclusions: In our pediatric population, the risk of laryngospasm was
increased in children with upper respiratory tract infection or an
airway anomaly. The use of laryngeal mask airway was found to be
associated with laryngospasm even when adjusted for the presence of
upper respiratory tract infection and airway anomaly.
Correspondence to: J. Sprung, Department of Anesthesiology, College of Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905,
USA (email: sprung.juraj@mayo.edu).
examine whether laryngospasm may be provider nance ⁄ emergence. In all cases, two-tailed P-values
dependent. Surgery or anesthesia related factors £0.05 was considered as statistically significant.
analyzed included type of anesthetic induction
(intravenous vs inhalational), maintenance volatile
Results
agent (halothane, sevoflurane, isoflurane), use of
nitrous oxide and muscle relaxant (any). Definitive Between 1996 and 2005, there were 130 patients with
airway management included either endotracheal research authorization younger than 18 who had
tube (ETT), laryngeal mask airway (LMA) or face ‘laryngospasm’ coded in the Performance Improve-
mask. The patients who experienced laryngospasm ment Database and also had documentation of a
during LMA or face mask anesthesia and required laryngospasm in their anesthesia records. Three
endotracheal intubation were counted as their initial patients did not have written research authorization
airway device (LMA or mask, respectively). Timing and were, therefore, excluded from review. Two
of laryngospasm was noted as at induction, mainte- patients had laryngospasm noted during two sur-
nance or emergence. Postoperative outcomes geries (recurrent laryngospasm). For these patients
monitored were: pulmonary aspiration, unplanned we abstracted information from only the first occur-
mechanical ventilation, unplanned intensive care unit rence. Of the 130 laryngospasm events identified as
admission for observation, unplanned chest radio- cases, 60 (46%) occurred during induction, 13 (10%)
graph and death. during maintenance and 57 (44%) during emer-
gence. The majority occurred in children £5 years of
age (57% £3 years of age, 68% £5 years of age).
Data analysis
Mean patient weights were 22.9 ± 18.9 and
All analyses were performed using conditional 24.9 ± 23.5 (OR 1.02, 95% CI 1.00–1.04, P = 0.062)
logistic regression, taking into account the 1 : 2 in laryngospasm and control groups, respectively.
matched study design. Findings are summarized Two matched controls were identified for each
using odds ratios (OR) and corresponding 95% laryngospasm case with the exception of 10 cases for
confidence intervals. Initial univariate analyses were whom only one matched control could be identified.
performed to assess each risk factor individually. Thus the 130 laryngospasm cases were matched to
Patient demographics, comorbidities and preopera- 250 controls. From univariate analyses, no indivi-
tive medications were assessed in an overall analysis dual demographic variable, comorbidity, or preoper-
that utilized all matched sets of cases and controls. In ative medication was found to be significantly
addition, two subset analyses were performed. For associated with laryngospasm (Table 1). However,
one subset analysis, premedication and induction when variables with some evidence (P £ 0.15) of a
agents were analyzed as risk factors for laryngo- univariate association were considered simul-
spasm in the subset of matched sets where the case taneously in a multivariable analysis, URI (OR =
experienced laryngospasm during induction. For 2.03, P = 0.022) and airway anomaly (OR = 3.35,
the second subset analysis, maintenance agents P = 0.030) were found to be significant risk factors
and airway devices were assessed as risk factors for laryngospasm (Table 2).
for laryngospasm that occurred during mainte- In an analysis restricted to case-control sets where
nance ⁄ emergence. In addition to assessing univariate the case patient experienced laryngospasm on
associations, a series of multivariate analyses were induction, preoperative use of midazolam was
also performed. Patient demographics and comor- found to be associated with an increased risk of
bidities found to have some evidence (P £ 0.15) of a laryngospasm on induction (OR = 4.44, P = 0.030)
univariate association with laryngospasm were (Table 3). In a multivariable model adjusting for
included in a multiple conditional logistic regression upper respiratory infection and airway anomaly this
model using all matched sets. Variables found to be finding was no longer statistically significant
statistically significant in this multivariable model (OR = 3.68, P = 0.076).
were included as adjustor variables in additional In an analysis restricted to case-control sets
multivariable models to assess risk factors for laryn- where the case experienced laryngospasm during
gospasm during induction and also during mainte- maintenance ⁄ emergence the risk for laryngospasm
Table 1
Odds ratioa Association of demographics,
Controls Laryngospasm
comorbidities and preoperative
(n = 250) No. (%) (n = 130) No. (%) Estimate 95% C.I. P-value
medications with laryngospasm
Gender 0.077
Male 154 (61.6) 69 (53.1) 1.00
Female 96 (38.4) 61 (46.9) 1.50 0.96–2.36
ASA PS 0.862
1 117 (46.8) 60 (46.2) 1.00
2 96 (38.4) 52 (40.0) 0.98 0.59–1.64
3 37 (14.8) 18 (13.8) 0.81 0.36–1.82
Urgency 0.619
Elective 240 (96.0) 126 (96.9) 1.00
Emergency 10 (4.0) 4 (3.1) 0.68 0.15–3.09
Admission status 0.256
Inpatients 88 (35.2) 40 (30.8) 1.00
Outpatients 162 (64.8) 90 (69.2) 1.46 0.76–2.82
Comorbidities
Asthma 31 (12.4) 14 (10.8) 0.85 0.44–1.63 0.625
URI 28 (11.2) 25 (19.2) 1.77 0.99–3.16 0.055
Airway anomaly 9 (3.6) 12 (9.2) 2.56 0.91–7.19 0.075
Lung disease 5 (2.0) 5 (3.8) 2.20 0.58–8.36 0.249
CHD 28 (11.2) 9 (6.9) 0.49 0.21–1.18 0.111
Neurologic disease 54 (21.7) 34 (26.2) 1.37 0.75–2.49 0.305
Syndrome 21 (8.5) 7 (5.4) 0.57 0.24–1.40 0.221
GERD 14 (5.6) 11 (8.5) 1.61 0.62–4.14 0.325
Preoperative Medications
Antibiotics 41 (16.5) 15 (11.5) 0.64 0.32–1.26 0.195
Bronchodilators 32 (12.9) 11 (8.5) 0.61 0.29–1.25 0.175
Steroids 8 (3.2) 8 (6.1) 1.75 0.62–4.94 0.287
Anticonvulsants 13 (5.2) 4 (3.1) 0.52 0.16–1.71 0.284
H1 ⁄ H2 blockers 13 (5.2) 12 (9.2) 1.57 0.62–3.98 0.340
Provider
Pediatric anesthesiologist 111 (44.4) 63 (48.5) 1.33 0.80–2.21 0.268
a
Univariate analyses were performed to assess each variable as a potential risk factor for
laryngospasm using conditional logistic regression taking into account the matched set study
design.
URI, upper respiratory tract infection; CHD, congenital heart disease; ASA PS, American Society of
Anesthesiologist Physical Status; GERD, gastroesophageal reflux disease; CI, confidence interval.
Table 3
Association of type of anesthetics Controls (n = 115) Laryngospasm Odds Ratioa
with laryngospasm during Variable No. (%) (n = 60) No.(%) Estimate 95% C.I. P-value
induction
Preoperative midazolam 5 (4.4) 9 (15.0) 4.44 1.15–17.09 0.030b
Type of induction 0.509
Intravenous 23 (20.0) 14 (23.3) 1.00
Inhalation 92 (80.0) 46 (76.7) 0.75 0.32–1.76
Anesthetic Agent
Volatilec 0.370
Sevoflurane 82 (89.1) 37 (80.4) 1.00
Halothane 6 (6.5) 6 (13.0) 2.22 0.67–7.33
Other 4 (4.3) 3 (6.5) 1.66 0.35–7.80
Intravenousc 0.260
Propofol 11 (47.8) 7 (50.0) 1.00
Thiopental 11 (47.8) 4 (28.6) 0.57 0.13–2.53
Other 1 (4.3) 3 (21.4) 4.71 0.40–54.83
a
Univariate analyses were performed to assess each variable as a potential risk factor for
laryngospasm using conditional logistic regression taking into account the matched set study
design. The analyses presented in this table include only the matched sets in which the case
experienced laryngospasm during induction.
b
In a multivariable model adjusting for URI and airway anomaly, premedication with midazolam
was not found to be statistically significant P = 0.076.
c
Analyses of types of volatile and intravenous agents were restricted to those receiving inhalation
and intravenous induction respectively. When assessing these risk factors unmatched analyses
were performed.
CI, confidence interval.
Table 4
Association of type of anesthetics Controls (n = 135) Laryngospasm Odds Ratioa
and airway device used with Variable No. (%) (n = 70) No. (%) Estimate 95% C.I. P-value
laryngospasm during
N2O 104 (77.0) 52 (74.3) 0.85 0.39–1.85 0.691
maintenance ⁄ emergence
Muscle relaxants used (any) 61 (45.2) 26 (37.1) 0.63 0.31–1.26 0.192
Agents for maintenance 0.586
Isoflurane 59 (43.7) 27 (38.6) 1.00
Halothane 5 (3.7) 3 (4.3) 1.67 0.29–9.51
Sevoflurane 61 (45.2) 37 (52.9) 1.51 0.70–3.26
Other 10 (7.4) 3 (4.3) 0.74 0.18–2.93
Type airway management 0.019b
Face mask 31 (23) 15 (21.4) 1.00
LMA 8 (5.93) 13 (18.6) 6.71 1.06–42.43
ETT 96 (71.1) 42 (60) 1.31 0.31–5.57
a
Univariate analyses were performed to assess each variable as a potential risk factor for
laryngospasm using conditional logistic regression taking into account the matched set study
design. The analyses presented in this table include only the matched sets in which the case
experienced laryngospasm during maintenance ⁄ emergence.
b
In a multivariable model adjusting for URI and airway anomaly, type of airway management was
still statistically significant P = 0.013.
N, number of patients; LMA, laryngeal mask airway, ETT, endotracheal tube; C.I., confidence
interval.
unplanned admission and unplanned postoperative pulmonary resuscitation and there were no deaths in
mechanical ventilation (n = 2), pulmonary aspira- either group.
tion (n = 3) (no such complications were identified
in the control group, P = 0.98 for all) and unplanned
Discussion
intensive care unit admission for observation (n = 6)
(three such complications were identified in our In a large retrospective study of pediatric laryngo-
control group, P = 0.054). No child required cardio- spasms, we examined demographics, comorbidities
Table 5
Demographics, type of surgery, timing of laryngospasm and complications that resulted from laryngospasm
Age Initial
gender airway Time of
ASA PS Type of surgery device laryngospasm Complication
7 years ⁄ F ⁄ 1 ENT, tonsillectomy ETT Induction Possible aspiration, unplanned chest radiogram
11 months ⁄ M ⁄ 3 ENT, microlaryngoscopy ETT Emergence ICU admission for observationa
(excision of subglottic cyst)
14 years ⁄ F ⁄ 2 Posterior spinal fusion ETT Emergence Tracheal reintubation and ICU admission for
unplanned postoperative mechanical ventilation
3 months ⁄ M ⁄ 3E Strangulated inguinal hernia ETT Induction ICU for postoperative mechanical ventilationa
23 months ⁄ F ⁄ 2 ENT, microlaryngoscopy Face mask Induction ICU admission for observation
17 months ⁄ F ⁄ 2 Magnetic resonance scan (arm) ETT Emergence Aspiration (admission for observation on ward)
2 years ⁄ M ⁄ 1 ENT, tonsillectomy ETT Emergence ICU admission for observation (aspiration)
20 months ⁄ M ⁄ 2 ENT, microlaryngoscopy ETT Emergence ICU admission for observationa
5 years ⁄ F ⁄ 3 ENT tonsillectomy Face mask Emergence ICU admission for observationa
5 years ⁄ F ⁄ 2 ENT, suspension laryngoscopy ETT Induction ICU admission for observationa
(excision subglottic papilloma)
a
Laryngospasm was a significant contributory factor for admission to ICU ⁄ mechanical ventilation.
F, female; M, male; E, emergency, ASA PS, American Society of Anesthesiologist Physical Status; ICU, Intensive Care Unit; ETT,
endotracheal tube; ENT, ear nose and throat surgery; MV, mechanical ventilation; ENT, ear, nose and throat surgery.
and anesthesia as risk factors for the occurrence of ratio 2.1, 95% confidence interval 1.21–3.45) (7).
laryngospasm. Our main findings were that the Finally, a just published prospective study con-
presence of URI, upper airway anomaly and the use firmed that the history of recent (<2 weeks) URI
of LMA were significant risk factors for laryngosp- doubles the rate of laryngospasm (odds ratio 2.6) in
asm. Furthermore, children who experienced laryn- patients who had LMA-based anesthesia (8).
gospasm had higher overall number of The type of device used to maintain airway may
postoperative complications, although none had be associated with increased risk of RAE. For
cardiac arrest or died. example, Cohen and Cameron (12) demonstrated
Known risk factors for laryngospasm, such as that the use of an ETT increased by 11-fold the
younger age and the type of surgery (1,2,7,8,10) were chance of RAE in children with URI. LMA has been
not examined in the present study as they were the considered as a superior airway device compared
matching variables in our case–control study. How- with ETT to reduce the incidence of RAE in children
ever, 68% and 83% of laryngospasm were recorded with URI (14,15). Only one study (9) reported
in children younger than 5 and 10 years old, respec- association between LMA and increased incidence
tively. of various RAE. However, within the examined
It is widely accepted among anesthesiologists that patient cohort, the rate of laryngospasm was com-
the URI poses a risk for RAE including laryngo- parable between different airway devices (1.12%
spasm. In a prospectively collected pediatric anes- ETT, 1.7% LMA and 1.5% face mask).
thesia database Cohen and Cameron (12) found that We found that the LMA poses higher risk for
children with URI had two to seven times higher laryngospasm compared with ETT and face mask,
chance to experience perioperative RAE. Schreiner even after adjusting for the presence of URI and
et al. (7) showed that the children who experienced airway anomaly. The cause of increased risk is
laryngospasm were twice as likely to have an active unclear although one may speculate that the use of
URI as defined by their parents. However, in a study an LMA and ⁄ or accumulated secretions may repre-
by Tait and Knight, (13) the presence of URI was not sent a potent stimulus to the airway during emer-
associated with an increased incidence of perioper- gence. Our database does not allow us to determine
ative respiratory complications. whether the LMA was removed prior to or after
We identified a 2.3-fold higher risk of laryngo- emergence. This may be an important factor in the
spasm in children with URI, which is in close genesis of laryngospasm and deserves future study.
agreement with the findings of Schreiner et al. (odds Furthermore, in the present study we do not have
information regarding the stage of patients’ URI and ASA PS scores, emergency ⁄ elective status,
it is possible that the severity of URI may be inpatient ⁄ outpatient status and preoperative medi-
associated with the reported difference in respira- cations scheduled until the day of surgery. Nor did
tory complications related to LMA (7,8). we find an association between comorbidities (con-
Finally, we found that the ETT only modestly genital heart disease, syndromes, pulmonary disease
increased the risk for laryngospasm compared with or asthma, presence of gastric reflux, neurologic
face mask. While this may be surprising, Mamie disease) and the risk for laryngospasm.
et al. (10) reported that the use of muscle relaxants Our children with laryngospasm had increased
concomitant with tracheal intubation reduces the number of postoperative complications, therefore,
risk of laryngospasm and most of our children who their overall morbidity was increased compared
received tracheal intubation received muscle relax- with controls. At the same time some of the patients
ants. in this group had preexisting airway pathology
In our univariate analysis the use of midazolam (Table 5) which could have contributed to the ICU
premedication was found to be associated with an admission.
increased risk for laryngospasm, however, in an
analysis that adjusted for the presence of either
Limitations of the study
upper airway anomaly or URI, the association of
midazolam with laryngospasm was no longer Although event data was entered prospectively (as
statistically significant. In one recent study (5) the a part of our Performance Improvement database)
use of midazolam premedication increased RAE this is a retrospective study with the inherent
(which included ‘any degree of laryngospasm’) limitation related to retrospective design. First, we
upon emergence from anesthesia; however, authors cannot exclude that relevant information was
did not specify the proportion of laryngospasm (if omitted from the description of events. Second,
any) in these adverse events. In contrast Schreiner the presence of mandatory reporting for perioper-
et al. (7) reported a nonsignificant trend toward a ative laryngospasm does not assure that all events
decreased likelihood of laryngospasm (P = 0.06) if were recorded and that case ascertainment was
the midazolam was used before induction. It is complete, therefore, we did not attempt to calcu-
difficult to retrospectively interpret these divergent late the incidence of laryngospasm. Third, Mayo
findings. In our study absence of significance of Clinic is a tertiary referral institution and this
midazolam in our final multivariable model sug- practice is likely not representative of the general
gests the univariate association of midazolam with population of children undergoing surgery. Our
laryngospasm may be at least partially explained by surgical population as a whole is weighted
other covariates. towards fewer routine procedures. This factor can
The role of anesthesiologists’ experience in pedi- potentially bias toward different risk factors com-
atric anesthesia, as a potential risk factor for laryn- pared with a less complex pediatric population.
gospasm, remains controversial. Mamie et al. (10) Fourth, there are no criteria for the diagnosis and
reported that children not anesthetized by pediatric recording of laryngospasm in our database, there-
anesthesiologist have a 1.7 times greater risk of fore we had to rely on assessment accuracy of
perioperative RAE (these events included laryn- anesthesia consultant involved with the respective
gospasms) (10). Similarly, children who developed case.
laryngospasm were more likely to have their anes- In conclusion, in our surgical population the
thesia supervised by a less experienced anesthesia administration of general anesthesia to children with
supervisor (7). In the present case–control study, we an URI or airway anomalies increases the risk for
could not confirm that procedures supervised by laryngospasm. The risk is also determined by the
pediatric anesthesiologist reduced the risk for laryn- type of airway used rather than by the child’s past
gospasm. A most recent study reported similar medical history. The use of LMA was associated
findings (8). with the highest risk for laryngospasm even
As was observed by Cohen (12) we did not when adjusted for the presence of URI or airway
confirm an association between laryngospasm and anomalies.