A total of 17 studies (n=15,335 patients) were included. These included one controlled trial (n=61 patients), one case-control study (n=670 patients), three cohort studies (n=375 patients) and 12 retrospective analyses of databases (n=14,229 patients). For study quality, three were class 2 studies and 14 were class 3 studies. Methodological flaws included the following: lack of randomisation; statistical adjustment for important confounding factors; description of control intervention and harms outcomes and information on drop-outs; historical control groups; and short-term outcomes. The maximum duration of follow-up was up to six months or hospital discharge.
In-hospital mortality (15 studies): Unadjusted odds ratios for in-hospital mortality ranged from 0.17 (95% CI 0.10 to 0.31) favouring control interventions to 2.43 (95% CI 1.78 to 3.33) favouring pre-hospital intubation); the absolute difference ranged from -21.8% to 38.2%; adjusted odds ratios (seven studies) ranged from 0.24 (95% CI 0.11 to 0.49) to 1.42 (95% CI 0.13 to 1.78).
Functional outcomes (five studies): Results were mixed. Three of five studies, assessing outcome by post-hospital destination, reported improved outcomes in control groups; one favoured pre-hospital intubation; and one small study findings were inconclusive. Two of six studies, assessing functional outcomes using various scoring systems, reported improved outcomes in pre-hospital intubation groups; two studies favoured control groups; and two studies findings were inconclusive.
Harms (seven studies): Pre-hospital intubation failure rates or complications ranged from 2.1% to 41.1% (five studies).
Pneumonia (three studies): All three studies reporting this outcome reported a higher risk of pneumonia in pre-hospital intubation groups.