The prevalence of a failed first intubation attempt was 31.4% (95% CI = 30.2–32.6) among patients managed in an out-of-hospital setting. The following variables were independently associated with a failed first intubation attempt, operator with ≤ 50 prior intubations, small inter-incisor space, limited head extension, macroglossia, ENT tumor, cardiac arrest, and vomiting.
A prior meta-analysis indicated a significant difference in the success rate of the first intubation attempt with rapid sequence induction in an out-of-hospital setting between physician and non-physician operators (88% [95% CI = 83–93%] and 78% [95% CI = 65–89%], respectively [1]. The eight studies on physicians in that analysis were heterogeneous: two retrospective [20, 21] and four prospective [13, 14, 18, 19] observational studies, one controlled trial [9], and one study concerning head trauma only [10] (Table 1). In studies of management by physician-led teams in an out-of-hospital setting and involving patients who had and had not experienced cardiac arrest, the failure rate of the first intubation attempt was 12.6–32% [8,12,15-18] (Table 1). Adnet et al. published a prospective multicenter observational study in 1998 involving 691 patients, which showed a failure rate of the first intubation attempt of 32% [8]. Two later observational studies involving only anesthesiologists yielded failure rates of 12.6% and 22.4% [12, 18]. However, 70% and 20% and 20% and 54% in the first and second studies, respectively, had trauma and cardiac arrest [12, 18]. In an analysis of 2674 patients, Combes et al. reported a failure rate of 26% and another study of 653 intubations a rate of 29.9% [15, 17]. In these two studies, the main operators were emergency physicians, but intubations were also performed by residents or anesthetist nurses under the supervision of the physician. A prospective multicenter study analyzed 1941 intubations managed by paramedics, nurses or physicians, a priori [31]. Paramedics performed 94% of the intubations and the first-attempt failure rate was 31.5%. These observational studies were largely similar to this work. A randomized multicenter trial in an out-of-hospital setting showed a failure rate of 28.8% [16]. Our results are, therefore, similar to those of other studies (Table 1).
Numerous variables are associated with DI, defined as more than two intubation attempts or bad glottic visualization. However, few studies analyzed those associated with first-attempt failure. The seven variables we identified were similar to those for DI [12, 17, 23-25], the exceptions being operator status (resident), maxillofacial trauma, history of ENT tumor, body mass index (BMI), operator position [24], patient on the floor, hyoid-mental length less than three fingerbreadths, airway obstruction [17], blood, secretion or vomit in the upper airway, anatomical factors, patient position, and bright ambient light [12]. Most of these variables—inter-incisor space, macroglossia, limited extension of the head, ENT tumor and maybe vomiting—cannot be foreseen. This hampers prediction of DI.
The risk of first-attempt failure is higher for operators with experience of fewer than 50 intubations [24, 32]. Moreover, the survival rate is higher among patients with cardiac arrest intubated by rescuers with more than 50 prior intubations, and among non-cardiac arrest patients intubated by rescuers with more than 26 prior intubations [33]. A review of the literature concluded that at least 50 intubations with no more than two intubation attempts are necessary to obtain a 90% success rate [34]. Emergency physicians perform a median of 10 intubations per year and 25% perform four or fewer intubations per year [35]. Among 5,245 out-of-hospital rescuers, > 67% had performed two or fewer TIs and > 39% had never performed TI [36]. A retrospective study involving intensive-care paramedics trained in the management of vital distress reported a failure rate of the first intubation attempt of 10.6% [37]. This highlights the importance of the initial training of rescuers, who are likely to be confronted with vital distress. Also, in this study all patients with spontaneous breathing received sedative and myorelaxant drugs. However, paramedics cannot administer curare. This could explain some of the reported risk factors, e.g., clenched jaw, trismus, inability to pass the endotracheal tube through the vocal cords, and an intact gag reflex [25, 36].
Cardiac arrest was a risk factor for failure of the first intubation attempt. Timermann et al. showed that intubation difficulty is more frequent in patients with cardiac arrest (17% vs. < 10%) but neither Combes et al. nor Freund et al. found such an association [11, 17, 24]. One possible explanation is continuation of cardiopulmonary resuscitation during intubation in > 50% of patients.
BMI was not a risk factor for failure, contrary to prior reports [24, 38, 39]. In a study similar to ours, BMI was associated with DI (OR = 1.0 [95% CI = 1.0–1.1]) [24]. In this study, the dependent variable was DI (frequency, 7%), defined as an IDS > 5 [24]. In a retrospective analysis involving paramedics, a BMI ≥ 40 kg.m-2 was significantly associated with DI (OR = 3.68 [95% CI = 1.3–10.6] [38]. However, the body weight of 39% of the patients was not available and the multivariate analysis was adjusted for only age, sex, pathology, and BMI [38]. In the emergency department (ED), obese patients (N = 342) had a failure rate of the first intubation attempt of 40.7% vs. 29.1% for leaner patients (N = 5370) (adjusted OR = 0.62 [95% CI = 0.49–0.79]) [39]. Obesity remains a controversial risk factor for DI; therefore, caution is required when intubating obese patients in emergency settings.
The overall rate of adverse events was > 13%, and significantly increased with the number of intubation attempts. This result is consistent with prior studies in out-of-hospital settings or in the emergency department [4,6, 8]. Indeed, Adnet et al. reported an incidence of mechanical complications of 12.2% and general complications of 12.3% in an out-of-hospital setting; complications were associated with a significantly higher number of intubation attempts [8]. In an emergency department, the incidence of adverse events was 14.2%, increasing to 47.2%, 63.6%, and 70.6% after two, three, and four or more intubation attempts, respectively [4]. This is consistent with a previous report of adverse event rates for two attempts or fewer and three attempts or more of 9% and 35%, respectively [6].
Overall, the failure rate of the first intubation attempt was high. Some of the associated factors can be improved (operator training and experience), but most cannot. A randomized control trial performed in an emergency department showed that systematic use of a bougie during the first intubation attempt improved the success rate [40]. However, a secondary analysis of a study conducted in an out-of-hospital setting involving paramedics found no significant difference according to bougie use [41].