Airway management and perioperative adverse events in children with mucopolysaccharidoses and mucolipidoses: A retrospective cohort study

Children suffering from mucopolysaccharidoses (subtypes I, II, III, IV, VI, and VII) or mucolipidoses often require anesthesia, but are at high risk for perioperative adverse events. However, the impact of the disease subtype and the standard of care for airway management are still unclear.


| INTRODUC TI ON
The reported incidence of serious adverse events in pediatric anesthesia varies between 0.14% (North American registry data 7 ) and 5.2% (prospective European multicenter study 8 ). A 30-day inhospital mortality rate of 0.1% has been found in children undergoing anesthetic procedures. 8 Respiratory complications are the most frequently reported adverse events. [7][8][9] However, perioperative morbidity and mortality are dramatically higher among children with MPS/ ML. [4][5][6][9][10][11] In MPS type I, a perioperative morbidity of up to 30% 5,6 and a 30-day risk of death per procedure of approximately 0.7% 12,13 have been reported.
Airway management is challenging in children with MPS/ML. [4][5][6]10,11,13,14 Fiberoptic intubation through a supraglottic airway and videolaryngoscopy are frequently used as indirect intubation techniques with high success rates in children with difficult airways. 15 Although it is recognized that MPS/ML poses unique anesthesia-related challenges, the standard of care for airway management in these children still remains unclear, and the impact of the disease subtype and primary airway approach on perioperative adverse events is unknown.
This study aimed to assess independent risk factors for perioperative adverse events-such as airway management problems and respiratory and cardiocirculatory events-in individuals with MPS/ ML and to analyze the interaction with the primary airway technique implemented.

| PATIENTS AND ME THODS
This retrospective, two-center study was approved by the Ethical Committee of the Medical Board of Hamburg, Germany (reference number PV4832, October 28, 2014).

| Data acquisition
The electronic patient management system (

What this article adds
• The disease subtype and the choice of the primary airway technique were the most important independent risk factors for perioperative adverse events in children with mucopolysaccharidoses or mucolipidoses.
The highest success rate and lowest risk for perioperative adverse events was found if indirect intubation techniques, such as fiberoptic intubation through a supraglottic airway, were chosen for the first intubation attempt.
Center Hamburg-Eppendorf. The medical records of each identified patient were systematically reviewed for anesthesia procedures.
Individuals with MPS or ML who underwent general anesthesia, regional anesthesia, or monitored anesthesia care at either of the study centers (University Medical Center Hamburg-Eppendorf or AKK Altona Children's Hospital, Hamburg) were included, thus comprising the study cohort. Patients' baseline characteristics, medical history, comorbidities, procedural and anesthesia-related data, and outcome data, were collected from electronic medical records (Soarian™ Health Archive, Release 3.04 SP12, Siemens Healthcare) and anesthesia charts.

| Anesthesia procedures (study cases)
A standardized anesthesia protocol was not implemented during the study period, and management was left to the discretion of the supervising anesthesiologist.
The analysis was based on the following assumptions and definitions.
For general anesthesia, the primary airway technique implemented was either bag-mask ventilation, supraglottic airway, or tracheal intubation. Tracheal intubation was facilitated by direct laryngoscopy, indirect intubation techniques, or fiberoptic intubation through a supraglottic airway (FOI SGA ). Indirect intubation techniques, such as conventional fiberoptic intubation techniques (FOI CON ), video laryngoscopy, or rigid fiberscopes, were implemented in order to visualize laryngeal structures and to guide tracheal tube positioning. The category "FOI CON " subsumes any fiberoptic nasal or oral intubation prior to or after the anesthesia induction without routine use of a supraglottic conduit. A laryngeal mask was used as a supraglottic conduit to guide fiberoptic intubation during FOI SGA . Videolaryngoscopy was typically enabled by C-MAC TM (Karl Storz) in both centers. Cases of local anesthesia performed by a surgeon with solely "anesthesia standby" were excluded from the outcome analysis. Cases that did not necessitate a secured airway (eg, regional anesthesia, procedural sedation, and analgesia or solely bag-mask ventilation) were not included in the multivariate model.

| Outcome measures
The primary outcome measure was perioperative adverse events, defined as follows. • Respiratory adverse events: clearly documented, severe respiratory complication, hypoxia (drop of peripheral oxygen saturation <90% for >1 minute or any below 80%), hypercapnia (endtidal or arterial partial carbon dioxide pressure >50 mm Hg for >1 minute or any above 60 mm Hg), aspiration, reintubation, airway obstruction, or reduced lung compliance accompanied by low tidal volumes and/or differences between peak inspiratory and end-expiratory airway pressure ≥25 mbar, severe laryngospasm, or postextubation stridor requiring any therapeutic intervention.
• Cardiocirculatory adverse events: clearly documented, unstable hemodynamic conditions, resuscitation, requirement of multiple or high doses of drugs (eg, atropine, epinephrine, or norepinephrine) for severe bradycardia or unexpected severe hypotension not directly linked to a surgical complication (eg, bleeding episode).
Secondary outcomes were as follows.
• Conversion rate: frequency of the transition to a different (rescue) airway technique or abandoned anesthesia in relation to the total number of airway techniques.
Outcome measures were recorded during anesthesia and in the early postoperative period until discharge from the postanesthesia recovery unit or admission to an intensive care unit. Anesthesia charts were independently checked for perioperative adverse events by three assessors (TD, MAP, MP). Adverse events were designated based on a consensus decision.

| Statistical analysis
The primary aim of the study was to identify independent risk factors for perioperative adverse events in individuals with MPS/ML. It was an exploratory study. Continuous data are presented as medians with interquartile ranges. Categorical data are presented as frequencies (percentage values in relation to the number of valid data).

| Multivariate multilevel regression model
As a first step, potentially eligible factors and covariates for the multivariate analysis were identified through clinical considerations and literature research. The primary airway techniques were subcategorized as either supraglottic airway, direct laryngoscopy, indirect intubation techniques, or FOI SGA . MPS subtypes IV and VI and ML were clustered in order to create the category "very rare diseases." A multivariate three-level mixed-effects logistic regression model was fitted in order to identify risk factors for perioperative adverse events. In order to account for repeated anesthesia procedures within the same patient and for multiple events within a single anesthesia episode, two nested random effects were added. Four factors (disease subtype, subcategory of primary airway technique, predicted difficult airway, and gender) and five covariates (age, American Society of Anesthesiologists physical status classification, emergency procedure, duration of anesthesia, and indication for anesthesia) were included in the model as fixed effects. Additionally, a variable identifying the type of adverse event and the interaction term between this variable and the predictor "primary airway technique" was included in the model. The final model results from a complete case analysis.
Results are presented as odds ratios (OR) with 95% confidence interval (CI). The reference category for calculation of the ORs for categorical variables was based on clinical considerations. The subgroup with the lowest rate of perioperative adverse events was favored. The interaction and subgroup contrast between event subcategories (airway management problems, cardiocirculatory, or respiratory events) and primary airway approaches were visualized by calculating the marginal predicted probability with the corresponding 95% CI. Differences were considered significant with an alpha error of less than 5% (P-value <.05). Nominal P-values were reported. Statistical analysis was performed using STATA version 15.0 (StataCorp).

| Study cohort
A total of 141 inpatients suffering from either MPS or ML were identified during the study period. Within this group, we identified 67 patients (the "study cohort") who underwent 274 anesthesia procedures at the two medical centers, 63 with MPS, and four with ML.
Five procedures had to be excluded, as anesthesia charts were unavailable. This left a total of 269 anesthesia procedures (the "study cases") for analysis. Overall, patients underwent 353 surgical or diagnostic interventions. Frequently, two or more procedures were combined within a single anesthesia episode (Table 1). Ten cases were excluded from the analysis of perioperative adverse events, as only a local anesthesia was implemented by a surgeon (eg, ophthalmological procedures).
The study cohort predominantly included individuals with MPS types I (35.8%), II (16.4%), III (28.3%), IV (4.5%), VI (6.0%), and some individuals with ML (6.0%). Table 1 shows the age range of our cohort. The types of procedure recorded were either surgical (63.7%), dental (6.5%), interventional (4.2%), or diagnostic (25.5%). Twenty   ); MRI, magnetic resonance imaging; * multiple planned procedures within a single anesthesia episode; the dataset of this analysis is complete without any missing values. # Of note, the study cohort includes 10 cases which involved only local anesthesia performed by a surgeon (eg, ophthalmological procedures), these cases were included in the descriptive analysis (Table 1 and 2) but not in the outcome analysis (Table 4, Figures 1 and 2).

| Anesthesia and airway management
In most cases, general anesthesia was used (92.6%). A total of 280 airway approaches were documented; 244 primary airway techniques were used (Tables 2 and 3). Conversion to a rescue airway approach was required 36 times, as the primary or subsequent technique was regarded as insufficient. In two of these cases, anesthesia and surgery were abandoned due to failed airway attempts. Of note, FOI SGA was not attempted in these two cases. The conversion rates and pathways, if the attempted airway technique was regarded as insufficient, are illustrated in Table 3. A conversion was not noted in any case of FOI SGA . As FOI SGA was not accompanied by airway management problems ( Figure 2, middle panel); in any case, the rule of three was used to calculate the upper limit of the 95% confidence interval in that subcategory. 16

| Perioperative adverse events
In 25.6% of the cases studied, one or more perioperative adverse events were recorded (   ; § eight missing values in this category; $ the category "airway not secured" includes 5 cases in which solely bag-mask ventilation was used; # Of note, the study cohort includes ten cases which involved only local anesthesia performed by a surgeon (eg, ophthalmological procedures); these cases were included in the descriptive analysis (Table 1 and 2) but not in the outcome analysis (Table 4, Figures 1 and  2). * conventional fiberoptic intubation includes cases of nasal or oral fiberoptic tracheal intubations prior to or after induction of anesthesia without routine use of a supraglottic conduit; ** laryngeal masks were used as a supraglottic conduit to guide fiberoptic intubation.

| Multivariate analysis of risk factors for perioperative adverse events
The analysis of the interaction between the primary airway approach and event subcategory (airway management problems, cardiocirculatory and respiratory events) revealed the following.
• The adjusted event rates and 95% confidence intervals (CI) for cardiovascular events were 1.6 (0-4.8) % for direct laryngoscopy,  In nonsyndromic children, the risk for perioperative critical events decreases with age. 8 However, it has been proposed that due to disease progression during the natural history of MPS, the risk for intubation problems might paradoxically increase. 4,10,11 Nevertheless, after adjustment for multiple confounders, we found that perioperative adverse events did not increase with age in our multivariate analysis. This finding could be attributed to the therapeutic effects of bone marrow transplantation or enzyme replacement therapy, which may slow down disease progression. 4,5,17

| The disease subtype is an independent risk factor
MPS III was associated with the lowest rate of perioperative adverse events in our study. Airway management was usually Note: Overall airway approaches (n = 280) include the primary airway techniques (n = 244) and subsequent rescue airway approaches (n = 36) which include 2 cases of abandoned anesthesia # ; conversion rate: frequency of the transition to a different (rescue) airway technique or abandoned anesthesia in relation to the total number of airway techniques; *the category "others" includes various techniques such as bag-mask ventilation, rigid optics, tracheostoma, or abandoned anesthesia.
uneventful, which is consistent with previous findings. 18

| The primary airway technique is an independent risk factor
Previous case series addressed the issue of airway techniques in MPS, but did not provide direct comparative assessments taking essential confounding factors into account. 4-6,10,11

| Supraglottic airway
We found the supraglottic airway to be associated with a lower risk for airway management problems than direct laryngoscopy. However, it was associated with a higher risk for respiratory adverse events as opposed to indirect intubation techniques and FOI SGA . Thus, caution should be exercised should a supraglottic airway be implemented in individuals suffering from MPS/ML who present with progressive respiratory dysfunction or a narrowed upper airway. There is growing evidence that the laryngeal mask airway is associated with lower postoperative complication rates, as opposed to tracheal intubation in pediatric cohorts. 27,28 Our data indicate that these findings should not be extrapolated without critical consideration to children with MPS/ML.

| Indirect tracheal intubation techniques
A previous study has suggested that limiting the number of direct laryngoscopy attempts and quickly transitioning to indirect techniques might improve safety in children with a difficult tracheal intubation. 9 The initial choice of an indirect tracheal intubation technique was associated with a lower risk for airway management problems and respiratory adverse events than with direct laryngoscopy in our study cohort. Of note, in our study, fiberoptic intubation was the predominant indirect intubation technique, while videolaryngoscopy was rarely implemented.

| Fiberoptic intubation through a supraglottic airway (FOI SGA )
As awake FOI CON is often poorly tolerated by children, FOI SGA is a very popular approach in children with difficult airways. In a large multicenter trial, FOI SGA was associated with a higher overall success rate than videolaryngoscopy in children with difficult airways. 15 FOI SGA is gaining widespread popularity in children at risk for spinal cord compression, as an extension of the head can be avoided and the cervical spine remains aligned. 29 This is an extremely important issue for children with MPS/ML, who frequently exhibit a cervical spinal canal stenosis or dens hypoplasia with atlantoaxial instability. Although FOI SGA has been established for more than two decades, data concerning its implementation in children with MPS/ML are still very limited. 14,30 Our analysis is based on data stemming from two high-volume centers. Our study cohort included the largest ever case series focused on FOI SGA in MPS. During the study period, FOI SGA was typically enabled by a stepwise approach, which was finalized by the insertion of a tracheal tube railroaded straight via a bronchoscope through a disposable laryngeal mask. However, impossible laryngeal mask placement, which is a limitation for FOI SGA , has not been observed with FOI SGA in our study.
The retrospective nature of this study is a well-recognized limitation and a potential source for selection bias. Moreover, our data only reflect experiences from two centers, and caution should be exercised should our results be generalized. Appropriate caution should also be exercised if conclusions are drawn from unadjusted data.
In conclusion, our study demonstrated that the disease subtype and the primary airway technique were the most important independent risk factors for perioperative adverse events in children with MPS