Timing of tracheostomy in critically ill patients refers to the time at which tracheostomy is performed during the clinical course. Even though indications and advantages of tracheostomy have been largely described in the critical care setting, timing remains controversial. More than 25 years ago, a consensus conference on artificial airways in patients receiving mechanical ventilation produced the following recommendation: “The appropriate duration of translaryngeal intubation cannot be defined at present. Clinical consideration or complications may dictate changing the artificial airway to another route. However, no data exist that give adequate direction as to when it is routinely advisable to change from a translaryngeal intubation to a tracheostomy” [4]. Nowadays, some uncertainty remains about this topic. A strong indication for tracheostomy arises when there is the need for prolonged endotracheal intubation for mechanical ventilation. The concept of “prolonged” intubation has changed over time, starting from an older approach with indication for tracheostomy given after several days of the course of the critical illness (usually after 15 days or more of endotracheal intubation) to more recent approaches considering tracheostomy earlier (e.g. after 3–10 days). Moreover, other aspects have influenced this trend. In 1960s, ETs were made of rigid and inflexible material with low-volume, high-pressure cuff. [27] Consequently, it was common to perform tracheostomy “early” during the clinical course to minimise injuries to upper airways, larynx and trachea resulting from translaryngeal intubation. During the following decades, progress in materials and equipment leads to less injuries and complications. In 1981 Stauffer et al. [28] reported data about risks associated with tracheostomy describing a significant increase in morbidity (stomal haemorrhage and infection rates >30 %, rate of tracheal stenosis >50 %) and a significant increase in mortality (4 %). Thus, the trend started to change leading to a progressive delay of the procedure. During the last 25 years, advances in techniques and equipment caused great improvement in the safety of the procedure that became easily performable at bedside [3, 29, 30]. Generally, the different approaches in timing have been defined as early and late. However, no consensus exists about what exactly constitutes early vs late tracheostomy. The question about tracheostomy timing is complex because several factors should be considered: (1) an estimate of the probability of prolonged mechanical ventilation, especially for certain categories of critically ill patients (e.g. neurological or trauma patients), and (2) the best time for tracheostomy during the course of the critical illness. In the case of patients considered at risk of prolonged mechanical ventilation, early tracheostomy strategy would expose them to an unnecessary procedure if the prediction failed. On the other hand, patients considered to have unduly relatively short mechanical ventilation during their stay in the ICU may undergo an unnecessary prolonged exposure to translaryngeal endotracheal intubation. This fact may increase potential complications, without the advantages of tracheostomy in terms of weaning, and other aspects of care.

A recent analysis [31] of seven national surveys performed in France [32], Germany [33], Italy [9], the Netherlands [34], Spain [35], Switzerland [36] and the UK [37] demonstrated that the presence of a shared clinical practice across Europe about tracheostomy timing (from 7 to 15 days from ICU admission). A review of the Project IMPACT database (109 ICUs) reported that tracheostomy was performed at a median of 9 days after ICU admission with an interquartile range of 5–14 days [38]. From many years, evidence to guide the decision about tracheostomy timing came from observational, retrospective and relatively small and/or single-centre randomised trials [10, 22, 39–42]. Several factors may have reduced the quality of evidence from studies: variable protocols quality, sub-optimal sample size, heterogeneity in populations enrolled and patients’ characteristics, lack of standardised protocol for co-interventions and inconsistency in outcomes selection across studies. Many of these studies supported the concept that early tracheostomy was beneficial, generating the hypothesis to be confirmed by larger studies. Recently, two large prospective multi-centre randomised trials enrolling general ICU population [20, 21] and two meta-analyses [43, 44] have investigated early versus late strategy for tracheostomy, adding robust data and new insight about this topic.

In 2010, Terragni et al. [20] reported data from a multi-centre randomised trial performed in 12 Italian ICUs from June 2004 to June 2008 enrolling 600 adult patients without pneumonia (clinical pulmonary infection score, CPIS -<6), ventilated for 24 h, who had a Simplified Organ Failure Assessment score between 35 and 65 and a Sequential Organ Failure Assessment (SOFA) score of 5 or more. Patients were randomised after 72 h (48 h after the enrolment) if having PaO₂ <60 mmHg (FiO₂ = 0.5, PEEP = 8 cmH₂O), SOFA score = >5 and no pneumonia. The two arms of the study were: (1) early tracheostomy (6–8 days of laryngeal intubation, 209 patients assigned) and (2) late tracheostomy (13–15 days of laryngeal intubation, 210 patients assigned). The primary outcome of the study was the incidence of VAP; secondary outcomes during the 28 days immediately following randomisation were number of ventilator-free days, number of ICU-free days and survival. VAP were diagnosed using the simplified CPIS (CPIS score >6) at study entry, at randomisation and every 72 h until day 28. The study included medical patients (40 % of the early group and 36 % of the late group), scheduled surgical patients (8 % vs 10 %), unscheduled surgical patients (41 % vs 45 %) and trauma patients (11 % vs 9 %) without significant difference between groups. Many randomised patients (31 % in the early group vs 43 % in the late group) did not undergo tracheostomy due to proximity to extubation or death. All tracheostomies were performed bedside with percutaneous techniques (Griggs technique in 72 % in early group and 73 % in late group, PercuTwist technique in 25 % in early group and 22 % in late group). Adverse events occurred in 39 % of both groups. The incidence of VAP was not significantly different between groups (14 % early vs 21 % late; P = 0.07). The number of ventilator-free and ICU-free days and the incidences of successful weaning and ICU discharge were significantly greater in patients randomised to the early tracheostomy group compared with patients randomised to the late tracheostomy group; there were no difference between the groups in survival at 28 days. The authors concluded that early tracheostomy did not result in a significant reduction in incidence of VAP compared to late tracheostomy. They also underlined that long-term outcomes did not differ and that more than one-third of patients experienced adverse events related to the tracheostomy. Basing on these findings, they suggest that tracheostomy should not be performed earlier than 13–15 days. One limitation of this study was the use of CPIS score for the diagnosis of VAP since its diagnostic performance in this setting was not high, especially in surgical and trauma patients [45]. Another limitation was related to some exclusion criteria (such as chronic obstructive pulmonary disease, active pneumonia, anatomic deformity of the neck, lung cancer) partially limiting the clinical applicability of the results to general ICU population.