Results
Twenty-two of the 24 ICUs provided data. These 22 units account for 94% of patients admitted annually to ICUs and 91.5% of funded ICU (level 3) bed capacity in Scotland. During the study period, there were 3279 admissions to ICUs; of which, 2260 (69%) had tracheal intubation performed at some time during their hospital stay. We recorded 877 tracheal intubations performed by critical care doctors in the study period; in 83 instances, tracheal intubation was performed during cardiopulmonary resuscitation—we excluded these intubations for separate analysis (not reported here). Of the remaining 794 intubations, 56 did not have a unique Wardwatcher® identifier and a further 28 had incomplete outcome data recorded on Wardwatcher®. Thus, although data from all 794 intubations were available for the intubation procedure and complications, full patient characteristics and outcome data were available for 710 tracheal intubations ( Table 2 ).
Seventy per cent of intubations occurred in ICUs with another 18% in EDs ( Table 3 ). Around 40% took place at night and in three-quarters, the degree of urgency was considered immediate or urgent ( Table 3 ). Most intubations were performed in patients newly admitted to intensive care requiring mechanical ventilation, and one-third of intubations were for problems with the existing airway, failed trial of extubation, or accidental extubation ( Table 4 ). Many patients were very unwell before intubation; 23% were receiving non-invasive ventilation, 15% were receiving vasopressor by infusion, and 75% were breathing >60% oxygen. Furthermore, 39% had
, 11% had
, 20% had systolic arterial pressure (AP) <90 mm Hg, and 6% systolic AP <80 mm Hg. In 22 patients, tracheal intubation was known to have been difficult on a previous occasion. Two hundred and four predictors of difficult tracheal intubation were recorded in 197 patients; patient obesity ( n=113), cervical spine immobilization after trauma ( n=41), and abnormal airway anatomy ( n=50) such as reduced mouth opening, reduced thyromental distance, or grade 3 or 4 Mallampati view. Seventeen of these 197 patients (8.6%) had grade 3 or 4 laryngoscopic view compared with 31 of 555 patients (5.6%) with no predictors of difficulty.
The drugs used and the main details of the intubation procedure are shown in Table 3 . Induction drugs were given in 96% of intubations and neuromuscular blocking drugs were given in 92%. Tracheal intubation at the first attempt was obtained in 91%, five patients (1%) had three attempts at intubation, and no patient underwent more than three attempts at intubation. Two failed primary intubations required insertion of a Proseal laryngeal mask, one of which required progression to surgical tracheostomy to secure the airway. Videolaryngoscopy and fibreoptic laryngoscopy were each used in six intubations. We did not find any statistically significant difference in first-time success rate or the number of intubation attempts, when we compared those patients who were given a neuromuscular blocking drug and those who were not (one attempt 90% vs 95%; two attempts 9% vs 3.5%; three attempts 1% vs 1.5%, respectively).
Immediate complications associated with tracheal intubation are shown in Table 5 . Twenty-two per cent of patients had severe hypoxaemia (
) and 20% had severe hypotension (systolic AP <80 mm Hg) (Fig. 1). Neither the severity of hypoxaemia nor hypotension after intubation was related to outcome (
, ICU mortality 23.5%; 80–89%, 31.4%; 70–79%, 31.2%, <70%, 25.6%; P=0.208: systolic AP ≥90 mm Hg, ICU mortality 24.6%; 80–89 mm Hg, 32.2%; 70–79 mm Hg, 25.7%, <70 mm Hg, 31.4%, P=0.356). However, choice of induction drug appeared to be related to the development of hypotension. Ninety-four per cent of patients had systolic AP ≥80 mm Hg before induction of anaesthesia. Sixty of these patients had systolic AP <70 mm Hg after induction; the drugs used were propofol in 38 patients, etomidate in two, thiopental in eight, ketamine in one, and midazolam in 11. The relative risk (95% confidence intervals) for developing systolic AP <70 mm Hg after each induction drug was: etomidate 0.37 (0.09–1.46); thiopental 0.69 (0.33–1.42); ketamine 0.74 (0.11–5.02); propofol 0.99 (0.60–1.64); and midazolam 4.49 (2.56–7.88). Mechanical complications were infrequent with oesophageal intubation being recorded in only 2% of cases, and vomiting or regurgitation of gastric contents also occurring in 2%.
(Enlarge Image)
Figure 1.
Percentage of patients with hypoxaemia or hypotension before and after tracheal intubation. *P=NS, **P<0.001 compared with preintubation.
Three-quarters of intubations were performed by doctors with >24 months of formal anaesthetic training, and all but one doctor with <6 months training had senior supervision ( Table 6 ). As the number of attempts at intubation increased so did the proportion of intubating doctors with >12 months anaesthetic training (attempt 1, 82%; attempt 2, 93%, attempt 3, 100%). Intubation in patients with predictors of difficulty was more likely to be performed by a senior doctor. The success rate of tracheal intubation was related to the duration of formal anaesthetic training undertaken by the intubator (P<0.001, Table 6 ). One hundred and thirty-five intubations were done by trainees with <12 months of formal anaesthetic training; 44 of these occurred overnight and only five were not supervised by a more experienced doctor.