Chapter 109: Controversies: Early Tracheotomy

Tracheotomy is a common procedure with almost 1 in 4 ICU patients receiving a tracheotomy.¹ While a tracheotomy is a relatively straightforward surgical procedure, the medical complexity of patients requiring tracheotomy, and the debate about its timing and indications, make its utilization complex. Despite its common and widespread use there remains considerable debate in the literature regarding tracheotomies risks, benefits, and indications.

The indications for tracheotomy have evolved throughout the history of the procedure. With improved understanding of the disease processes affecting the airway, the indications have become more standardized and rigorous in recent years. Today there are numerous indications recognized in the adult population. These indications were defined by the American Academy of Otolaryngology and Head and Neck surgery in the 2010 Clinical Indicators Compendium,² and fall into 4 broad categories: to relieve a mechanical obstruction (or potential obstruction), to manage aspiration (and promote improved pulmonary toilet), to provide long-term ventilation (and avoid the complications of long-term translaryngeal intubation), and to promote weaning from the ventilator. The last 2 categories constitute the vast majority of patients included in studies on ET.

Tracheotomy offers a number of potential benefits over endotracheal intubation. Tracheotomy has been shown to decrease the mechanical workload of ventilator-dependent patients,^3,4 and patients with tracheotomies can be placed on and off the ventilator without any significant risk to the patient, potentially facilitating earlier extubation. Tracheotomy improves comfort level when compared with translaryngeal intubation. Patients with tracheotomies have improved tracheal suctioning, oral care, mobility, ability to take oral nutrition and articulate speech,^5,6 and decreased oropharyngeal and laryngeal trauma compared to intubated patients.⁷

Different clinicians have different opinions on the risks and timings of tracheotomy. The risk of tracheotomy has decreased over time and varies depending on the skill and experience of the person performing the procedure. In 1833 Trousseau reported a 75% mortality rate for tracheotomy patients, compared with current mortality rates of approximately 1%.^8,9,10 Today, while tracheotomy is a relatively safe procedure, complications still occur. A recent study looking at 113,653 tracheotomies performed in the United States in 2006 showed a 3.2% complication rate and 0.6% mortality rate.¹¹ However, the experience and expertize of the physician performing the procedure can greatly impact the risk of tracheotomy. A recent study showed that surgeons performing the fewest tracheotomies demonstrated the highest complication rates, and that intraoperative complication rates varied greatly among practitioners with otolaryngologists having an intraoperative complication rate is 0.39% compared to 3.5% with tracheotomies done by nonotolaryngologists.¹²

The optimal timing for a tracheotomy is dependent on the patient and the situation. There is significant controversy regarding the timing of surgery for those patients undergoing an elective tracheotomy after intubation, and it is the timing of tracheotomy in patients undergoing prolonged intubation that is the most debated, and is the focus of this chapter.

Physicians of the 1960s promoted the placement of ET tubes, as high pressure cuffs and rigid endotracheal tubes lead to significant complications rates.¹³ With the arrival of the more flexible endotracheal tubes with low-pressure cuffs in the 1970s, combined with the reported high complication rate of tracheotomy, delayed tracheotomy, 3 to 4 weeks after intubation, became the rule. Since the mid-1980s, with the introduction of percutaneous tracheotomy and the decreased complication rates associated with open tracheotomy, there has been a shift toward earlier tracheotomy placement.

The ideal time for tracheotomy in ventilated patients has not been established, despite many studies addressing this topic. The risks of prolonged intubation must be weighed against the benefits of tracheotomy as well as the risks of the procedure.¹⁴ Earlier recommendations suggested that tracheotomy be performed in patients requiring mechanical ventilation for more than 21 days.⁶ More recent guidelines advise that translaryngeal intubation be used only for patients requiring fewer than 10 days of artificial ventilation, and that tracheotomy be performed for those requiring artificial ventilation for more than 21 days. The decision is left to the physician for those patients falling between 10 and 21 days.¹⁵ A commonly used method of determining the need and timing of a tracheotomy is the “anticipatory approach.” This involves initially stabilizing the patient and treating the primary disease process. If extubation is felt to be possible in the first several days of mechanical ventilation then tracheotomy is not considered necessary. If after a week of translaryngeal intubation the patient is felt (1) likely to benefit from a tracheotomy and (2) likely to require extended intubation (more than 7 additional days), then a tracheotomy should be performed.^13,16 As the clinician’s ability to predict the duration of intubation improves, even earlier tracheotomy can be considered. However, despite its widespread use, this approach has yet to be determined to improve the morbidity and mortality of patients requiring prolonged mechanical ventilation.

There are many studies examining the appropriate timing for tracheotomy that typically fall into 3 general categories: (1) nonrandomized retrospective reviews, (2) prospective studies, (3) meta-analysis (of either retrospective reviews or prospective studies). Observational studies retrospectively comparing patients who received a tracheotomy early after intubation with those who received a tracheotomy sometime after intubation are common. However, it is very difficult to extrapolate meaningful conclusions from these observational papers as the patient populations in the 2 groups may be very different, and the difference in management in regards to the timing of tracheotomy may be reflective of differences in perceived outcomes of the patients.

A study by Arabi looked at 531 consecutive nonrandomized patients.¹⁷ They found that the time to tracheotomy was associated with increased duration of ventilation, length of ICU and hospital stay, but not survival. A similar retrospective study by Tong et al, looked at 592 patients, 128 ET patients who received a tracheotomy by day 7, and 464 LT patients who received a tracheotomy after day 7, demonstrating decreased days of mechanical ventilation, length of ICU stay, and length of hospital stays in patients receiving an ET, but no improvement in survival and ventilator-associated pneumonia (VAP).¹⁸

Shan et al¹⁹ performed a meta-analysis of observational studies comparing ET (< 7 days) with LT (7 or more days) in ICU patients. They combined the data from 6 studies (Armstrong,²⁰ Arabi,¹⁷ Moller,²¹ Flaatten,²² Zalgi,²³ and Tong¹) looking at a total of 2037 subjects. They found that the mortality in the early tracheotomy group was significantly lower than the late group at 1 year (26.1% vs 29.8% P = 0.02). In addition the duration of mechanical ventilation was shorter in the early group than the late group (mean difference 10.04 days, P = 0.001), the ICU stay was shorter in the early tracheotomy group (mean difference 8.8 days, P < 0.001), the hospital stay was shorter in the early tracheotomy group (mean 12.18 days, P < 0.001), and there was no difference in the incidence of VAP between the 2 groups.

A more recent observational study, looking at tracheotomies in the American College of Surgeon’s Trauma Quality Improvement program, used a well-balanced propensity matched cohort of 1154 patients in an attempt to decrease bias and compared early (≤ 8 days) versus late (> 8 days) tracheotomy. They found that ET was associated with fewer mechanical ventilation days (10 vs 16), shorter ICU stay (13 vs 19 days), shorter hospital length of stay (20 vs 27 days), and lower incidence of pneumonia (41.7% vs 52.7%), deep venous thrombosis (DVT) (8.2% vs 14.4%), and decubitus ulcer (4.0% vs 8.9%). Hospital mortality was similar between the 2 groups.²⁴

A number of larger observational studies exist, which, like these smaller single-center studies, usually show a benefit for early tracheotomy and contributed to the interest in performing early tracheotomy. A study by Freeman et al²⁵ looked at the data from 43,916 patients of whom 2473 had tracheotomy. It showed that tracheotomy timing correlated significantly with duration of mechanical ventilation, ICU stay, and hospital stay. A similar study looked at tracheotomy in acute care hospitals in Ontario hospitals.²⁶ Between 1992 and 2004 10,927 patients received tracheotomy: 3758 received early tracheotomy (≤ 10 days) and 7169 patients received late tracheotomy. Patients receiving early tracheotomy had lower 90 day, 1 year, and study mortality. On multivariate analysis each delay of 1 day was associated with increased mortality. While these observational studies are an excellent starting point, there are a number of reasons why their findings cannot be extrapolated to prospective tracheotomy management. Most important among these is the potential selection biases in determining who received an early versus a late tracheotomy. However, at a minimum, these observational studies promoted the concept that ET might be beneficial, and encouraged larger prospective studies.

A number of prospective studies have been performed comparing early and late tracheotomy. Significant differences in definitions, methodologies, and conclusions exist between studies. A study by Koch et al²⁷ examined 100 predominately surgical patients in the university hospital in Geiben Germany, prospectively randomized to early (≤ 4 days, average 2.8 days) or late (≥ 6 days, average 8.1 days) tracheotomy. They found no change in mortality between the ET and LT groups, but did find a statistically significant decrease in VAP incidence, duration of mechanical ventilation, ICU and hospital stay in the ET group. A study by Terragni et al²⁸ in JAMA compared ET patients (day 6-8) and LT patients (day 13-15) as part of a randomized control trial across 12 Italian ICUs. Six hundred patients were randomized to early or late tracheotomy after 48 hours: 145/209 ET group patients received tracheotomy and 119/210 LT group patients received tracheotomy. They found a 14% incidence of VAP (measured at 28 days from randomization) in the ET group compared to a 21% incidence in the LT group that approached, but did not reach, statistical significance (P = 0.07). The number of ventilator-free days and ICU-free days were statistically significantly decreased in the ET group, but there was no change in hospital stay or mortality between the 2 groups.

Blot et al²⁹ performed a prospective and randomized study in 2008 evaluating 25 ICUs in France comparing ET (day 4) with prolonged intubation. The study was halted after 2 years because of difficulty accruing patients, and only 123 patients were enlisted in the study. There were no difference in mortality, duration of ventilation, or VAP between the ET and LT group, but there were improvements in patient comfort and late laryngeal symptoms in the ET group. This study was underpowered and fell well short of its recruitment goal partially due to what the authors thought was recruitment bias and therefore it is difficult to draw meaningful conclusions from the study. Trouillet et al³⁰ performed a prospective randomized controlled study looking at patients after heart surgery at a single French academic hospital expected to need more than 7 days of ventilation. They randomized patients to immediate ET (< day 5) or prolonged intubation with tracheotomy (> day 14). Two hundred sixteen patients were enrolled, 109 in the ET group and 107 in the LT group. There were some differences in the characteristics of the 2 groups with higher rates of heart transplantation, repeat surgery, and renal replacement in the ET group. Only 27% of LT group received tracheotomies. There was no significant difference in ventilation-free days, survival ICU or hospital stay, or rates of VAP between the groups, but more sedation-free days in the ET group with patients in the ET group having more time calm awake or lightly sedated and more days comfortable with easy care, earlier nutrition, and mobilization.

Rumbak et al³¹ performed a prospective randomized trial looking at 120 patients in a Florida hospital MICU comparing early percutaneous tracheotomy (48 hours) or delayed tracheotomy (day 14-16). They found that patients in the early group had statistically significant decreased mortality (31.7% vs 61.7%) rates of pneumonia (5% vs 25%), less time in ICU (4.8 vs 16.2 days), and days on mechanical ventilation (7.6 vs. 17.4). This paper has fewer patients than other papers; however, it also stratified patients to the widest difference between ET and LT (< 48 hours vs > 2 weeks).

The TracMan trial³² is a recent prospective study that was highly anticipated. It was an open multicentered randomized clinical trial at both university and nonuniversity hospitals in the United Kingdom that identified people within the first 4 days of admission that were likely to require 7 days of ventilator support. ETs were done by day 4 and LTs were done by day 10 or later. Of the people in the LT group over half did not get a tracheotomy (37% were extubated). There was an average of 13.6 days of mechanical ventilation in the ET group versus 15.2 days of mechanical ventilation in the LT group, this approached but did not reach statistical significance (P = 0.06). There was no difference in the length of ICU stay and mortality, but less sedation was used in the ET group. This study highlights the limited ability of clinicians to predict which patients would require ventilatory support as over half of the patients in the LT group did not need tracheotomy.

While most of these prospective studies are well designed it is difficult to power a prospective study that would pick up small changes between ET and LT groups. Because of this a number of meta-analysis exists attempting to increase the patient population and the power of the study. A study by Griffiths in 2005¹⁵ looked at 5 studies with 406 participants: the combined studies by Rumbak,³¹ Bouderka,³³ Dunham,³⁴, Rodriguez,³⁵ and Saffle.³⁶ They did not find any changes in morality, although the tracheotomy had a slightly lower relative risk (RR), and no change in hospital-acquired pneumonia, (although the tracheotomy had a slighter lower RR). They found that tracheotomy patients had 8.5 fewer days of mechanical ventilation and 15.3 fewer days in the ICU. A recent (2014) meta-analysis by Wang et al³⁷ looks at all randomized controlled trials regarding early versus late tracheotomy. The analyzed 7 trials (Saffle,³⁶ Bouderka,³³ Rumbak,³¹ Barquist,³⁸ Blot,²⁹ Terragni,²⁸ and Trouillet³⁰) encompass 1044 patients. They found no change in mortality (although there again was a lower RR in the ET group) and no change in VAP (although again there was an early with lower RR in the ET group). There was no appreciable change in length of mechanical ventilation or ICU stay.

Despite the numerous studies looking at the timing of tracheotomy no definitive conclusions can be reached. There are a number of potential explanations for the lack of consensus regarding timing of tracheotomy despite numerous papers. First tracheotomy patients are a varied patient population who might require tracheotomy secondary to trauma, general medical issue, respiratory issues both acute (such as pneumonia) and chronic (such as chronic obstructive pulmonary disease [COPD]), and complications from surgery, neurologic injury, etc. It is possible that ET may benefit one type of patient while not benefiting others. Additionally, ET has many different definitions. Some studies compared people who received a tracheotomy within 48 hours versus those who did not receive a tracheotomy until 2 weeks passed, while others compared those earlier than 7 days versus those 7 days or later. It is possible that there are some cutoffs when early tracheotomy may be beneficial, but these were not measured in these studies, or if they were measured then they were underpowered. A comprehensive study would have to randomize people to every day to find ideal timing (and this could vary by disease).

Additionally the decision to perform an ET is based on the clinician’s impression that prolonged intubation will be needed while also believing that the patient has a chance of recovery. The studies examining ET are therefore evaluating clinician’s ability to determine which patients will need prolonged intubation but have a chance of recovery as much as whether ET is useful. Perhaps in the right patient population early tracheotomy is beneficial, but we are not yet able to accurately determine which patients fit these criteria. A number of studies exist examining clinician’s ability to predict ET,^39,40 but as of yet no applicable guidelines exist to determine which patients will need prolonged intubation.

When analyzing the data some studies suggest benefits for ET and some do not. Many studies find a benefit for some aspects of ET (such as increased comfort, or incidence of VAP) but not for others. However, care must be taken when drawing conclusions on some characteristics such as length of mechanical ventilation, ICU stay, and hospital stay as these data may be more indicative of system issues, or of how these patients were managed and triaged, rather than actual medical differences. For example, patients with a tracheotomy may have easier access to step down units or hospice facilities thus decreasing their length of stay. While the data do not find conclusively for or against the use of ET, more of the data suggest benefits for ET and none of the data suggest harm for ET. The meta-analysis did not find a statistically significant benefit in terms of mortality and days of mechanical ventilation, but in both cases the RR of ET was less than that of LT patients.

Many studies have been written examining the benefits of ET, but as of yet there is no clear consensus on ET. The heterogeneity of the many studies and the patient population receiving tracheotomy makes drawing definitive conclusions difficult. While it cannot be conclusively stated that ET is beneficial there are a number of studies that suggest that it may be (and none that suggest it is harmful). At a minimum more larger and better defined studies are needed before we can say for certain when, and in which patients, ET is conclusively beneficial.