The goal of this study was to determine the association of first pass success with the incidence of adverse events (AEs) during emergency department (ED) intubations.
The goal of this study was to determine the association of first pass success with the incidence of adverse events (AEs) during emergency department (ED) intubations.
This was a retrospective analysis of prospectively collected continuous quality improvement data based on orotracheal intubations performed in an academic ED over a 4-year period. Following each intubation, the operator completed a data form regarding multiple aspects of the intubation, including patient and operator characteristics, method of intubation, device used, the number of attempts required, and AEs. Numerous AEs were tracked and included events such as witnessed aspiration, oxygen desaturation, esophageal intubation, hypotension, dysrhythmia, and cardiac arrest. Multivariable logistic regression was used to assess the relationship between the primary predictor variable of interest, first pass success, and the outcome variable, the presence of one or more AEs, after controlling for various other potential risk factors and confounders.
Over the 4-year study period, there were 1,828 orotracheal intubations. If the intubation was successful on the first attempt, the incidence of one or more AEs was 14.2% (95% confidence interval [CI] = 12.4% to 16.2%). In cases requiring two attempts, the incidence of one or more AEs was 47.2% (95% CI = 41.8% to 52.7%); in cases requiring three attempts, the incidence of one or more AEs was 63.6% (95% CI = 53.7% to 72.6%); and in cases requiring four or more attempts, the incidence of one or more AEs was 70.6% (95% CI = 56.2.3% to 82.5%). Multivariable logistic regression showed that more than one attempt at tracheal intubation was a significant predictor of one or more AEs (adjusted odds ratio [aOR] = 7.52, 95% CI = 5.86 to 9.63).
When performing orotracheal intubation in the ED, first pass success is associated with a relatively small incidence of AEs. As the number of attempts increases, the incidence of AEs increases substantially.
El objetivo de este estudio fue determinar la asociación entre el éxito del primer intento con la incidencia de eventos adversos (EA) durante las intubaciones en el servicio de urgencias (SU).
Análisis retrospectivo con una recogida de forma prospectiva continua de datos de mejora de la calidad, basados en intubaciones orotraqueales realizadas en un SU universitario en un periodo de cuatro años. Tras cada intubación, el operador completó un informe de datos sobre aspectos de la intubación que incluyeron: las características del paciente y el operador, el método de intubación, el dispositivo empleado, el número de intentos requeridos y los EA. Se monitorizaron numerosos EA, e incluyeron eventos como la aspiración presenciada, la desaturación de oxígeno, la intubación esofágica, la hipotensión, las arritmias y la parada cardiorrespiratoria. Se realizó un análisis multivariable mediante regresión logística para evaluar la relación entre la principal variable predictora de interés, éxito del primer paso, y la variable resultado, la presencia de uno o más EA, tras controlar por otros posibles factores de riesgo potenciales y de confusión.
Durante el periodo de estudio de cuatro años, se realizaron 1.828 intubaciones orotraqueales. Si la intubación fue exitosa en el primer intento, la incidencia de uno o más EA fue del 14,2% (IC 95% = 12,4% a 16,2%). En caso de requerir dos intentos, la incidencia de uno o más EA fue del 47,2% (IC 95% = 41,8% a 52,7%), en caso de requerir tres, la incidencia fue del 63,6% (IC 95% = 53,7% a 72,6%), y en caso de requerir cuatro o más, la incidencia fue del 70.6% (IC 95% = 56,2% a 82,5%). El análisis de regresión multivariable mediante regresión logística mostró que más de un intento de intubación orotraqueal fue un factor predictor significativo de uno o más EA (OR ajustada 7,52, IC 95% = 5,86 a 9,63).
Cuando la intubación orotraqueal se realiza en el SU, el éxito del primer intento se asocia con una incidencia relativamente baja de eventos adversos. Conforme aumenta el número de intentos, la incidencia de eventos adversos se incrementa de forma substancial.
The concept of first pass success is frequently promoted as the goal of emergency intubation. However, there is surprisingly little published evidence to support this. Data from the anesthesia literature have revealed that multiple attempts at tracheal intubation are associated with an increase in the incidence of adverse events (AEs).[1-6] There are limited data in the emergency medicine (EM) literature evaluating the effect of first pass success on the incidence of AEs in the emergency department (ED).
Critically ill and injured patients in the ED frequently require intubation. These patients are often at risk of AEs during intubation due to the urgency of the situation, lack of time for preparation, the presence of a full stomach, hemodynamic and respiratory decompensation, and decreased physiologic reserves. Therefore, it is important to understand the effect of first pass success and the risk of AEs associated with multiple attempts at orotracheal intubation in the ED. The goal of this study was to determine the association between multiple intubation attempts and the incidence of AEs during ED intubations, with an emphasis on the effect of first pass success on AEs.
This was a retrospective analysis of all orotracheal intubations prospectively recorded in a continuous quality improvement database from July 1, 2007, to June 30, 2011. This project was granted exemption by the University of Arizona Institutional Review Board prior to conducting the study.
This study was conducted in a 61-bed tertiary care academic ED with approximately 70,000 annual visits. This ED, also a Level I trauma center, has an Accreditation Council for Graduate Medical Education–accredited 3-year EM residency program and a 5-year combined pediatric/EM residency program. All residents in these programs participate in a 1-month rotation on the anesthesiology service where they learn and practice airway management techniques. On this rotation, residents average approximately 30 intubations. During the course of residency, EM residents perform an average of 30 intubations in this ED. The vast majority are performed with a direct laryngoscope or a video laryngoscope (VL). The two VLs most commonly used in our ED are the GlideScope (Verathon, Bothell, WA), and the C-MAC VL (Karl Storz, Tuttlingen, Germany). Typically, intubations are performed by EM residents under direct supervision of an EM attending physician. During academic conference one morning a week, the ED is staffed by EM attending physicians who perform the intubations primarily. On rare occasions, at the discretion of the EM attending physician, off-service residents and medical students are permitted to intubate in the ED. All intubations performed in the ED are supervised by the EM attending. All patients undergoing orotracheal intubation in the ED during the study period were included.
Following each intubation, the operator completed a data collection form, which included the following information: patient demographics, occurrence of a failed prehospital intubation attempt, operator specialty, operator postgraduate year (PGY), indication for intubation, method of intubation, paralytic agent, sedative agent, reason for device selection, device(s) used, presence of difficult airway characteristics (DACs), number of attempts at intubation, outcome of each attempt, and occurrence of AEs.
Only patients undergoing orotracheal intubation in the ED were included in this study. This included patients who underwent unsuccessful attempts at intubation in the field. Methods of intubation included rapid sequence intubation (RSI), in which a paralytic agent was used; oral intubation, in which a sedative agent was used (SED); and oral intubation, in which no medications were used (OTI).
The operator had three options to choose from for the reason for device selection. If the intubation was a routine airway with no anticipated difficulty, then the device selection was marked “standard.” If the device was selected with the expectation of a difficult airway, the reason for device selection was “difficult.” If the operator was using the device to gain educational experience with the device, then it would be classified as “education.”
Standard preoperative difficult airway predictors have been shown to be challenging to apply in the emergency setting.[8, 9] Thus, we developed a list of DACs that were feasible for the operator to determine prior to intubation in an emergent setting by brief examination of the patient. These include the presence of cervical immobility, obesity, large tongue, short neck, small mandible, facial or neck trauma, airway edema, blood in the airway, and vomit in the airway.
An attempt at orotracheal intubation was defined as insertion of the laryngoscope blade into the oropharynx, regardless of whether an attempt was made to pass the endotracheal tube. Each attempt was documented with one of three possible outcomes: 1) successful tracheal intubation with no additional attempts required, 2) inability to intubate with additional attempt(s) required, or 3) inadvertent esophageal intubation with additional attempt(s) required. Successful intubation was defined as correct placement of the endotracheal tube in the trachea as confirmed by end-tidal CO2 capnometry, pulse oximetry, chest auscultation, observation of chest excursion, absence of epigastric sounds, and misting of the endotracheal tube.
Adverse events tracked in this study include the following: esophageal intubation, oxygen desaturation, witnessed aspiration, mainstem intubation, accidental extubation, cuff leak, dental trauma, laryngospasm, pneumothorax, hypotension, dysrhythmia, and cardiac arrest. Cricothyrotomy was not considered an AE as we considered it an alternative way to secure the airway (see Table 1 for definitions of these AEs).
|Accidental extubation||Accidental removal of the ETT, requiring reintubation.|
|Aspiration||Presence of vomit at the glottic inlet visualized during intubation in a previously clear airway.|
|Cardiac arrest||Pulseless dysrhythmia occurring during intubation.|
|Cuff leak||Air leak around a cuffed ETT, requiring replacement of the ETT.|
|Dental trauma||Fracture or avulsion of tooth during intubation.|
|Dysrhythmiaa||Bradycardia or any ventricular dysrhythmia during intubation.|
|Esophageal intubation||Inadvertent placement of the ETT in the esophagus, requiring removal and reintubation.|
|Hypotension||Decrease in systolic blood pressure to <90 mmHg, unexplained by underlying pathophysiology.|
|Laryngospasm||Adduction of vocal cords, preventing passage of the ETT through the glottic inlet.|
|Mainstem intubation||Radiographic identification of the tip of the ETT in a mainstem bronchus.|
|Oxygen desaturation||A decrease in oxygen saturation greater than or equal to 10%.|
|Pneumothorax||Radiographic identification of air in the pleural space, without another obvious cause.|
The data forms were reviewed by the senior author (JCS). If the form had any missing data, it was returned to the operator for completion. If information on the form contained inconsistencies, the operator was interviewed by the senior author for clarification. The data forms were cross-referenced to professional billing and pharmacy records to identify any intubations performed without a corresponding data form. If an intubation was identified without a data form, the operator was given a data form to complete as soon as possible to ensure a maximal capture rate. During the study period, 93.8% of the airway forms were turned in at the time of intubation, and the remaining 6.2% were captured by cross-referencing, for an overall 100% capture rate.
The data were then entered into the electronic database program HanDBase 4.0 (DDH Software, Wellington, FL, www.ddhsoftware.com) for the Palm Pilot and iPad and were subsequently transferred to Excel for Windows 2010 (Microsoft, Redmond, WA). The primary outcome measures were the incidence of one or more AEs and the incidence of specific AEs.
Means with standard deviation (SD) for continuous variables, and proportions with interquartile range (IQR) for categorical variables, are reported along with 95% confidence intervals (CIs) calculated using the standard method for continuous data and the Clopper-Pearson method for categorical data. The primary risk factor for AEs was failure of a first pass intubation attempt, with the reference group consisting of patients successfully intubated on the first pass and the comparison group consisting of patients who required multiple intubation attempts. Patient demographics and intubation characteristics for these two groups were compared using Fisher's exact test and Student's t-test as appropriate. A list of patient and case characteristics was compiled a priori based on clinical experience and previous literature.[5, 10, 11] These characteristics were then used as covariates for multivariable logistic regression analysis with the incidence of at least one AE as the outcome variable. Covariates included for regression analysis were age, sex, trauma designation, presence of at least one DAC, method of intubation, paralytic used, sedative used, indication for intubation, reason for device selection, operator specialty, operator PGY, device chosen, and number of attempts required. Number of attempts (dichotomized as one vs. more than one attempt) was the main predictor variable of interest. The backward-elimination stepwise method was used to build logistic regression models. All covariates with univariate p-values of ≤0.2 were included into an initial starting model, and then covariates with the largest p-values (using the likelihood ratio chi-square test) > 0.05 were eliminated iteratively until only covariates with p-values ≤ 0.05 were left. All covariates that were eliminated during the first round of stepwise elimination were then introduced back into the preliminary final model and reassessed as independent risk factors (p ≤ 0.05) and/or confounders, with significant confounding considered if the addition of a covariate changed the regression coefficient for the primary risk factor (more than one intubation attempt) by ≥10%. We used univariate and multivariable fractional polynomial regression to assess the scale for the continuous variable age against the outcome variable (one or more AEs) to ensure a linear relationship in the logit scale. We assessed model fit with the Hosmer-Lemeshow goodness-of-fit test and model discrimination with the area under the receiver operating characteristics curve (AUC). In addition, we examined model diagnostics by examining the leverage, residual, and influence statistics for covariate patterns and tested for collinearity between covariates in our final model. Finally we tested for the presence of significant interaction effects for several clinical plausible scenarios (number of DACs × number of attempts, reason for intubation × number of attempts, and number of DACs × reason for intubation). We conducted a secondary analysis where only oxygen desaturation during intubation was considered an AE. We used the same methods as the primary logistic regression analysis. All statistical analyses were performed with Stata v.12.1 (StataCorp, College Station, TX).
During the study period, 1,850 intubations were performed. Of these, 1,828 (98.8%, 95% CI = 98.2% to 99.3%) were orotracheal intubations. Only these cases were included in this analysis. Patient, operator, and intubation characteristics for cases of successful intubation on the first attempt (“first attempt group”), and cases requiring more than one attempt (“multiple attempts group”) are described in Table 2. Of the 1,828 cases, 1,333 were intubated successfully on the first attempt (72.9%, 95% CI = 70.8% to 74.9%), while two or more attempts were required in 495 cases (37.1%, 95% CI = 34.6% to 39.8%).
|First Attempt (n = 1,333)||Multiple Attempts (n = 495)||p-valuea|
|% (n)||95% CI||% (n)||95% CI|
|Age (yr), mean (±SD)||43.9 (±23.4)||42.7–45.2||42.9 (±22.9)||40.9–45.0||0.4|
|Male||63.5 (846)||60.8–66.1||70.1 (347)||65.9–74.1||0.009|
|Trauma||47.9 (638)||45.2–50.6||45.7 (226)||41.2–50.2||0.4|
|One or more DACs||61.2 (816)||58.5–63.8||76.2 (377)||72.2–79.9||<0.001|
|EM||97.4 (1299)||96.5–98.2||96.0 (475)||93.8–97.5||0.20|
|Non-EM||2.6 (34)||1.8–3.6||4.0 (20)||2.5–6.2|
|0||1.8 (24)||1.2–2.7||2.6 (13)||1.4–4.5||0.07|
|1||19.0 (253)||16.9–21.2||24.0 (119)||20.3–28.1|
|2||36.3 (484)||33.8–38.9||36.0 (178)||31.7–40.4|
|3||41.2 (549)||38.5–43.9||35.8 (177)||31.5–40.2|
|Attending||1.7 (23)||1.1–2.6||1.6 (8)||0.7–3.2|
|Indication for intubation|
|Airway protection||63.7 (849)||61.0–66.2||59.4 (294)||54.9–63.8||<0.001|
|Respiratory failure||17.2 (229)||15.2–19.3||16.8 (83)||13.6–20.4|
|Cardiac arrest||8.9 (119)||7.5–10.6||16.4 (81)||13.2–19.9|
|Patient control||8.6 (115)||7.2–10.3||6.5 (32)||4.5–9.0|
|Hypoxia||1.6 (21)||0.1–2.4||1.0 (5)||0.3–2.3|
|Method of intubation|
|RSI||87.0 (1159)||85.0–88.7||80.0 (396)||76.2–83.4||0.001|
|OTI||10.7 (143)||9.1–12.5||16.4 (81)||13.2–19.9|
|SED||2.3 (31)||1.6–3.3||3.6 (18)||2.2–5.7|
|None||13.1 (174)||11.3–15.0||20.0 (99)||16.6–23.8||0.003|
|Rocuronium||45.2 (603)||42.6–47.9||40.6 (201)||36.3–45.1|
|Succinylcholine||41.4 (552)||38.8–44.1||39.0 (193)||34.7–43.4|
|Vecuronium||0.3 (4)||0.1–0.8||0.4 (2)||0.1–1.5|
|None||12.8 (171)||11.1–14.7||18.0 (89)||14.7–21.7||0.08|
|Etomidate||76.7 (1022)||74.3–78.9||71.9 (356)||67.7–75.8|
|Ketamine||5.6 (74)||4.4–6.9||5.9 (29)||4.0–8.3|
|Propofol||2.6 (34)||1.8–3.6||1.8 (9)||0.8–3.4|
|Other||2.4 (32)||1.7–3.4||2.4 (12)||1.3–4.2|
|Initial Device Used|
|Other||3.8 (51)||2.9–5.0||6.1 (30)||4.1–8.5||<0.001|
|Direct laryngoscope||52.0 (693)||49.3–54.7||61.4 (304)||57.0–65.7|
|Glidescope||28.7 (383)||26.3–31.2||21.4 (106)||17.9–25.3|
|C-MAC||15.5 (206)||13.6–17.5||11.1 (55)||8.5–14.2|
|Reason for device selection|
|Standard airway||62.2 (829)||59.5–85.9||65.9 (326)||61.5–70.0||0.01|
|Difficult airway||22.9 (305)||20.7–25.2||24.4 (121)||20.7–28.5|
|Education purposes||14.9 (199)||13.1–17.0||9.7 (48)||7.2–12.7|
The incidence of one or more AEs and the incidence of specific AEs versus number of attempts are shown in Figure 1. One or more AEs occurred in 14.2% (95% CI = 12.4% to 16.2%) of cases in the first attempt group versus in 53.1% (95% CI = 48.6% to 57.6%) of cases in the multiple attempts group (see Table 3 for the incidence of specific AEs in the first attempt and multiple attempts groups). Oxygen desaturation occurred in 9.2% (95% CI = 7.7% to 10.8%) of cases in the first pass group, while it occurred in 37.8% (95% CI = 33.5% to 42.2%) of cases in the multiple attempts group, specifically, 32.9% (95% CI = 27.9% to 38.2%) of intubations requiring two attempts, 43.9% (95% CI = 34.3% to 53.9%) of intubations requiring three attempts, and 56.9% (95% CI = 42.3% to 70.7%) of intubations requiring four or more attempts. Of patients experiencing one or more AEs in the first attempt group, oxygen desaturation occurred in 64.6% (95% CI = 57.3% to 71.4%) and aspiration occurred in 11.0% (95% CI = 7.5% to 15.5%) of these patients. Of patients experiencing one or more AEs in the multiple attempts group, oxygen desaturation occurred in 71.1% (95% CI = 65.2% to 76.5%), aspiration occurred in 11.0% (95% CI = 7.5% to 15.5%), and esophageal intubation occurred in 29.7% (95% CI = 24.2% to 35.6%) of these patients.
|Variable||First Attempt Group (n = 1,333)||Multiple Attempts Group (n = 495)|
|% (n)||95% CI||% (n)||95% CI|
|One or more AEs||14.2 (189)||12.4–16.2||53.1 (263)||48.6–57.6|
|Oxygen desaturation||9.2 (122)||7.7–10.8||37.8 (187)||33.5–42.2|
|Aspiration||0.8 (11)||0.4–1.5||5.9 (29)||4.0–8.3|
|Esophageal intubation||0.0 (0)||0.0–0.3||15.8 (78)||12.7–19.3|
|Mainstem intubation||3.2 (43)||2.3–4.3||4.0 (20)||2.5–6.2|
|Accidental extubation||0.5 (7)||0.2–1.1||0.2 (1)||<0.01–1.1|
|Cuff leak||0.5 (6)||0.2–1.0||0.8 (4)||0.2–2.0|
|Dental trauma||0.2 (2)||0.02–0.5||0.4 (2)||0.05–1.5|
|Laryngospasm||0.1 (1)||<0.01–0.4||0.4 (2)||0.05–1.5|
|Pneumothorax||0.1 (1)||<0.01–0.4||0.0 (0)||0.0–0.7|
|Dysrhythmia||0.2 (3)||0.05–0.7||1.4 (7)||0.6–2.9|
|Hypotension||0.2 (3)||0.05–0.7||0.4 (2)||0.05–1.5|
|Cardiac arrest||0.1 (1)||<0.01–0.4||0.4 (2)||0.05–1.5|
The adjusted odds ratios (aOR) for the occurrence of one or more AEs for our main risk factor and other covariates included in the final model, as well as crude ORs (cORs) for all covariates, are shown in Table 4. The final logistic regression model contained the following covariates: number of attempts (one vs. multiple attempts), presence of at least one DAC (0 vs. one or more), age (years), reason for intubation (cardiac arrest, protection of the airway, respiratory failure, patient control, hypoxia), and trauma-related versus medical condition. Univariate fractional polynomial regression showed that a cubed transformation of age improved model fit; however, multivariable fractional polynomial regression showed that untransformed age as a continuous variable was linear in the logit scale, so we used untransformed age (per 5 years) in our final model. Multiple attempts at orotracheal intubation was a significant predictor of the occurrence of one or more AEs (cOR = 6.86, 95% CI = 5.43 to 8.67; aOR = 7.52, 95% CI = 5.86 to 9.63). The presence of one or more DACs was also a predictor of the occurrence of one or more AEs (cOR = 1.68; 95% CI = 1.33 to 2.13; aOR = 1.67, 95% CI = 1.26 to 2.22). Model diagnostics did not reveal any outliers or overly influential covariate patterns.
|One or More AEs, n/N (%)||cOR||95% CI||aORa||95% CI||p-valueb|
|First attempt success||189/1333 (14.2)||[Reference]||—||[Reference]||—||<0.001|
|Multiple attempts (≥2)||263/495 (53.1)||6.86||5.43–8.67||7.52||5.86–9.63|
|Presence of DACs|
|One or more||333/1193 (27.9)||1.68||1.33–2.13||1.67|
|Age, yr (per 5 years)||0.86||0.80–0.93||0.93||0.90–0.95||<0.001|
|Indication for intubation|
|Airway protection||288/1143 (25.2)||[Reference]||—||[Reference]||—||<0.001|
|Respiratory failure||85/312 (27.2)||1.11||0.84–1.47||1.16||0.83–1.61|
|Cardiac arrest||32/200 (16.0)||0.57||0.38–0.84||0.32||0.20–0.50|
|Patient control||39/147 (26.5)||1.07||0.73–1.58||1.11||0.82–1.96|
|Type of emergency|
|Specialty of operator|
|Medical student||10/37 (27.0)||[Reference]||—||NS||0.29|
|Year 1||116/372 (31.2)||1.22||0.57–2.61|
|Year 2||157/662 (23.7)||0.84||0.40–1.77|
|Year 3||162/726 (22.3)||0.78||0.36–1.64|
|Reason device chosen|
|Difficult airway||115/426 (27.0)||1.15||0.90–1.49|
|Direct laryngoscope||127/489 (26.0)||1.05||0.82–1.34|
|Failed prehospital intubation|
For our secondary analysis using only oxygen desaturation as an AE, multiple attempts at orotracheal intubation was a significant predictor of desaturation (cOR = 6.03, 95% CI = 4.65 to 7.82; aOR = 6.85, 95% CI = 5.21 to 9.00). The presence of one or more DACs was also a predictor of desaturation (cOR = 1.62; 95% CI = 1.23 to 2.13; aOR = 1.65, 95% CI = 1.20 to 2.27). The Hosmer-Lemeshow goodness-of-fit p-value was 0.16 and the AUC was 0.767 for the final model. Model diagnostics did not reveal any outliers or overly influential covariate patterns.
In this study, the majority of patients were successfully intubated on the first attempt. When the first pass was successful, the incidence of AEs was 14.2%. However, if more than one attempt was required, patients suffered significantly more AEs. Patients requiring two attempts had 33% more AEs (47.2%) than those intubated on the first attempt. Patients requiring three attempts had 16% more AEs (63.6%) than those intubated in two attempts. Patients intubated in four or more attempts had 7% more AEs (70.6%) than those intubated in three attempts. This suggests that as the number of attempts increases, so does the risk of AEs, with the largest increase in AEs occurring between an unsuccessful first attempt and the second intubation attempt. This supports the notion that when performing intubations in the ED, every effort should be made to secure the airway on the first pass. When entered into a multiple logistic regression model, multiple attempts at tracheal intubation and the presence of one or more DACs were predictors of the occurrence of one or more AEs during intubation. Multiple attempts at intubation was the main predictor of the occurrence of one or more AEs, with an adjusted OR of 7.6. This nearly eightfold increase in risk of an AE underscores the importance of minimizing the number of attempts during ED intubations.
Oxygen desaturation was the most frequently reported AE, occurring in 9.2% in the first pass group and 37.8% in the multiple attempts group. The incidence of oxygen desaturation increased substantially as the number of attempts increased. In the multiple regression analysis, multiple attempts at intubation proved to be a predictor of oxygen desaturation (aOR = 6.85). These findings are easily understandable, as time taken during an intubation attempt is typically limited by oxygen saturation levels. Abandonment of an attempt is usually predicated on decreasing oxygen saturation.
The incidence of aspiration was also found to increase as the number of attempts increased. If the intubation was successful on the first pass, the incidence of aspiration was only 0.8%. However, if multiple attempts were required, the incidence of aspiration climbed to 5.9%. This is likely due to the fact that if an attempt fails, it is usually necessary to perform bag-valve-mask ventilation, which runs the risk of insufflating the stomach with air, thereby potentially causing vomiting.
The incidence of esophageal intubation also increased dramatically as the number of attempts increased. As the intubation becomes more difficult and the number of attempts increases, the operator may become more inclined to take risks and intubate a poorly visualized glottic inlet. This may explain the increase in incidence of esophageal intubation as the number of attempts increases. Regardless of the cause, esophageal intubation has already been shown in previous literature to be associated with an increase in other complications, including hypoxemia, bradycardia, and cardiac arrest.[2, 3]
Multiple attempts at orotracheal intubation were not found to be a predictor of some AEs, such as mainstem intubation or accidental extubation. This is logical, as the number of attempts should not have any bearing on how deep the endotracheal tube is placed or whether it comes out accidentally. Also, the cardiovascular AEs (hypotension, dysrhythmia, and cardiac arrest) all were not found to be increased in patients requiring multiple attempts. Previous literature is conflicting regarding the effect of the incidence of hemodynamic AEs. Some studies have shown an increased risk of cardiac arrest associated with multiple attempts.[2, 3] Others have shown it to be difficult to distinguish airway-related hemodynamic complications from the patient's underlying pathophysiology.
We found that the operator PGY was not a predictor of multiple attempts or AEs. One might suspect that more junior operators would require more attempts. However, significant selection bias probably explains this, as junior operators are probably intubating routine patients in the ED, while the senior operators are likely managing the more difficult airways.
Patients who underwent intubation due to cardiac arrest were found to be more likely to have required multiple attempts, but these intubations were associated with fewer AEs. The increase in multiple attempts is likely due to the fact that operators who respond to cardiac arrests have less time to assess the patient adequately and make preparations to optimize first pass success. The decreased incidence of AEs noted is likely due to the fact that if a patient is already in cardiac arrest, many of the other AEs, such as oxygen desaturation, aspiration, and dysrhythmia, are not possible.
This was a single-center study in an academic tertiary care center, where residents performed the vast majority of intubations. This limits the ability to generalize the findings to other clinical settings. Because quality improvement data collection forms were completed by the operators following each intubation, the data are subject to self-report and recall bias. In addition, because all data were self-reported, there is the possibility of underreporting of AEs by operators. However, if AEs were underreported, it is likely that they would be underreported in both the first pass group and the multiple group. To account for possible confounders, we used a logistic regression model that incorporated a priori selected clinically relevant observed confounders. However, it is possible that there are unknown confounders that we were unable to incorporate into the model that could account for the differences observed here. Therefore, these results must be interpreted with caution. We recognize that there are limitations to the definitions we chose for AEs. For example, the definition of aspiration used in this study is limited. It is possible that some patients had microaspiration and that this was not clinically apparent to the operator. Following up on patients after intubation would be one way to better determine if patients developed aspiration pneumonia. However, this would be problematic as well, since it would be impossible to know whether the patient developed aspiration pneumonia from aspiration in the field prior to intubation, or aspiration while on the ventilator, after the intubation.
We present quality improvement data from an academic ED that demonstrate a higher incidence of adverse events when multiple intubation attempts are required as compared to a single attempt. This supports the concept that every effort should be made to maximize first pass success during ED intubations. Future research should attempt to prospectively validate these findings and focus on strategies to improve first pass success.