text new page (beta)
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Similars in
SciELO 
Permalink
Introduction
In patients with severe traumatic brain injury (TBI), a tracheostomy is frequently performed to protect the airway and to allow withdrawal of assisted mechanical ventilation (AMV)1. The actual incidence of tracheostomy in patients with severe TBI is around 50-70%2.
The decision to perform a tracheostomy should be individualized, considering the risk of mortality, days of AMV, and neurological prognosis in the setting of severe TBI3.
Some authors have shown advantages by performing an early tracheostomy such as to reduce the rate of pneumonia4, days of AMV5,6, length of stay, and mortality5,7,8. It has been considered that patients that will require tracheostomy will profit by performing it earlier8-11.
There is still lack of consensus about if and when the tracheostomy should be performed and it is also unclear in which clinical and biological factors are associated with the need of tracheostomy in patients with severe TBI.
Patients and methods
This was a prospective, observational, and analytical study. To estimate the sample size, we used as reference a prevalence study of pneumonia associated with severe TBI because it is one of the most associated factors to perform tracheostomy, this study reported a prevalence of 45%6. The proportion of patients was calculated considering a population of 140 patients in average per year admitted to our hospital, yielding a result of 103 patients for the statistical calculation.
One hundred and twenty patients were included over a period of 18 months, between 2016 and 2018, older than 18 years old, and with diagnose in the first 24 h of admission of severe TBI. Twelve patients died before the decision of tracheostomy/not tracheostomy was made, so the final sample number was 108 patients.
The decision to perform a tracheostomy was made by the attending physician according to the patient clinical characteristics after the withdrawal of sedation.
For the diagnosis of pneumonia, we used the ventilator-associated pneumonia criteria that include new or progressive lung infiltrates, consolidation, cavitation, or pleural effusion on the chest radiograph and at least one of the following: new onset of purulent sputum or change in the macroscopical characteristics of sputum, fever, increase or decrease in the leukocyte count, positive blood cultures, or identification of a microorganism in bronchoalveolar washout or biopsy.
Results
Of our final sample, 94 men (87.04%) and 14 women (12.96%) comprise our universe of study. The most frequent mechanism of trauma was road traffic accidents (57%), followed by falls (26.6%), aggression (8.3%), gunshot wounds (4.2%), and others (3.7%). The average of Glasgow Coma Scale (GCS) at evaluation was 6.33 ± 1.4. The computed tomography (CT) scan at admission was classified according to Marshall scale and results were the following: Grade 5, 40 patients (48%); Grade 4, 13 patients (11%); Grade 3, 19 patients (16%); Grade 2, 39 patients (33%); and Grade 1, 1 patient (1%). In 74% of the cases, there was a skull fracture, 55% in the convexity and 19% in the cranial base. About 25% of the patients did not required surgical treatment. Table 1 summarizes the characteristics and significant variables.
Table 1 Comparison of the groups with tracheostomy and non-tracheostomy with significant statistical findings
| Non-tracheostomy, n = 38 | Tracheostomy, n = 70 | p | |
|---|---|---|---|
| Age | 29 ± 10 | 36 ± 15 | 0.003* |
| Glasgow | 7 ± 1 | 6 ± 1 | > 0.001* |
| Glasgow motor | 5 ± 1 | 4 ± 1 | > 0.001* |
| Revised trauma score | 5.8 ± 0.3 | 5.5 ±.7 | 0.003* |
| CRASH score | 23 ± 12 | 48 ± 22 | > 0.001* |
| Impact score | 7.8 ± 3.2 | 12.7 ± 4.5 | > 0.001* |
| APACHE II score | 14.3 ± 2.5 | 17.5 ± 3.8 | > 0.001* |
| SAPS II score | 36.6 ± 6.4 | 43.7 ± 10.2 | > 0.001* |
| Days of mechanical ventilation | 4.6 ± 2.6 | 7.2 ± 2.8 | > 0.001* |
| Subdural hematoma | 3 ± 6 | 21 ± 30 | > 0.001* |
| Midline shift | 2.5 ± 3.9 | 4.6 ± 5.5 | 0.023* |
| Pupil reactivity | |||
| Both | 31 (81.6%) | 33 (47.1%) | 0.0026¬ |
| One | 6 (15.8%) | 27 (38.6%) | |
| None | 1 (2.6%) | 10 (14%) | |
| Pneumonia | > 0.001¬ | ||
| No | 23 (60.5%) | 11 (15.7%) | |
| Yes | 15 (39.5%) | 59 (84.3%) | |
| Cisterns | |||
| Abnormal | 9 (23.7%) | 40 (57.1%) | 0.001¬ |
| Normal | 29 (76.43%) | 30 (42.9%) | |
| Surgical indication | |||
| No surgery | 11 (28.9%) | 17 (24.3%) | 0.016¬ |
| Epidural hematoma | 11 (28.9%) | 10 (14.3%) | |
| Subdural hematoma | 4 (10.5%) | 29 (41.4%) | |
| Contusion of edema | 6 (15.8%) | 8 (11.4%) | |
| Fracture | 6 (15.8%) | 6 (8.6%) | |
After the analysis, the global mortality was 23.3%, none associated with the performance of tracheostomy. Twelve patients died before we could decide to do a tracheostomy. Comparing the frequency of pneumonia in the tracheostomy group was 84.2% (Table 2).
Table 2 Comparison of groups tracheostomy versus non-tracheostomy
| Tracheostomy, n = 70 | Non-tracheostomy, n = 38 | p | |
|---|---|---|---|
| Days of mechanical ventilation | 7.2 ± 2.8 | 4.6 ± 2.6 | > 0.001 |
| Days of stay at hospital | 14.3 ± 7.4 | 9.3 ± 4.1 | > 0.001 |
| Days of stay after mechanical ventilation removal | 8.87 ± 6.2 | 4.7 ± 2.8 | > 0.001 |
| Pneumonia | 59 (84.2%) | 15 (39.4%) | > 0.001 |
| Deaths | 16 (22.9%) | 0 | 0.001 |
According to our bivariate analysis (Table 3), the main associated factors were CRASH score, GCS, and alteration of the pupil response. The pupil response was grouped for the analysis in two groups: those with both pupils reactive and those with the inadequate response of one or both pupils.
Table 3 Bivariate analysis
| Odds ratio | Confidence interval 95% | p | |
|---|---|---|---|
| CRASH score | 1.06 | 1.02-1.1 | 0.002 |
| Glasgow | 0.448 | 0.287-0.698 | > 0.001 |
| Dilated pupil, one or both | 4.965 | 1.93-12.77 | 0.001 |
Out of this bivariate analysis, we took into account those variables with p-values that showed statistical significance (Table 4). Considering the 75 percentile as the cut point for the statistical analysis, we found that the cumulative risk after 4 items was 100% (Table 5). Each factor provides the same value, considering that the cut point was adjusted by percentile.
Table 4 Prevalence risk ratio
| Odds ratio | Confidence interval 95% | p | |
|---|---|---|---|
| CRASH ≥ 60 | 1.06 | 1.02-1.1 | 0.002 |
| Impact ≥ 15 | 9.391 | 2.08-42.38 | 0.001 |
| SAPS II ≥ 49 | 18.106 | 2.33-140.35 | > 0.001 |
| APACHE II ≥ 19 | 7.326 | 2.05-26.177 | 0.001 |
| Subdural volume ≥ 20 | 21.86 | 2.83-168.91 | > 0.001 |
| Revised trauma score ≤ 5.89 | 2.781 | 1.07-7.19 | 0.031 |
| Age ≥ 44 | 3.896 | 1.23-12.33 | 0.015 |
| Glasgow ≤ 6 | 6.233 | 2.17-17.82 | > 0.001 |
| Dilated pupil, one or both | 4.965 | 1.93-12.77 | 0.001 |
| Collapse or absent of cisterns | 4.296 | 1.77-10.41 | 0.001 |
Discussion
Sociodemographic variables
In our study, patients younger than 44 years had less risk of tracheostomy; Shamim et al. and Goettler et al. found the same results as ours, as younger the patient, less frequent the tracheostomy was12,13.
Clinical variables
The GCS at admission is directly related to the need of tracheostomy. In our study, we found that a CGS of 6 points or less at admission is strongly associated with the need of tracheostomy, and we also found that the motor response of this scale is the one that correlates strongly with the need of tracheostomy. Major et al. found similar results as ours, GCS <7 at the 4th day correlated directly with the requirement of tracheostomy4. Gurkin et al. defined that on day 7, a GCS <9 is a tracheostomy predictor along with an injury severity score (ISS) >2412. Goettler et al. also found a significant p-value in terms of the GCS13.
Other clinical scales such as revised trauma score <5.89, IMPACT >15, SAPS >49, and APACHE II >19 are also associated with the need of tracheostomy4. In our study, unlike other authors, we did not found an association between the ISS and the need for tracheostomy, different from Gurkin, Shamim, and Goetler14.
Radiological variables
In our study, we founded two main radiological variables associated with the need of tracheostomy: partial or total collapse of basal cisterns and acute subdural hematoma >20 cc, in the initial CT scan.
Pneumonia, days of AMV, and total length of stay in hospitalization
About 42% of our patients were treated at neurosurgery hospitalization floor because we do not have 100% of availability at intensive care unit (ICU) all the time, we found significant differences between these two groups. The total length of stay and days of AMV were less in the neurosurgery floor group, as also the frequency of pneumonia (73.9% ICU) versus 45.1% neurosurgery floor. These results show the direct relationship between the number of days with AMV and frequency of pneumonia and also this with the total length of stay. Some other authors have found that the use of early tracheostomy decreases the number of days with AMV5,7,11,12.
Alali et al. reported in their series 30% reduction in AMV in patients with an early tracheostomy and 20% fewer days of hospitalization15. In our study, we found a reduction in the days of AMV in patients with early tracheostomy and making a new subgroup in patients who underwent tracheostomy before the day 4 of stay a radical reduction in the rate of pneumonia (63.6% ultra-early tracheostomy vs. 100% late tracheostomy). These results are similar as those reported by other authors7,4,11,15. Other authors differ in their results arguing that they found a decrease in the frequency of pneumonia5,6,16,17.
In our study, the patients that needed a tracheostomy had 5 more days of total length stay and days of AMV were greater, this can be explained because patients that required tracheostomy were critically ill.
Mortality
There was a global mortality of 23.3%, none associated with the performance nor the time of the tracheostomy, as reported in the "TracMan Randomized Trial" 18.
A total of 12 patients died before we were able to decide if they were going to need tracheostomy. Analyzing this situation and along with the findings reported by other authors that suggest that early tracheostomy has to be performed after day 3 and before the 1st week of stay4,11-13,19,20, led us to suggest that this procedure should be done between day 3 and 4 of stay because before day 3 critically ill patients should not be exposed to unnecessary procedures as they will die during hospitalization as reported by Major et al. and Bouderka et al.4,21
We did not found a relationship between the time of performing the tracheostomy and mortality rate. The deaths in the tracheostomy group were 22.9% versus 0% in the non-tracheostomy group, as also days of AMV and total length of stay were higher in the tracheostomy group, this can be explained because this group of patients was critically ill. Some authors even associate the early tracheostomy with a decrease in mortality5-7,19. This was not the case in our results.
Complications of tracheostomy
No immediate complications of the tracheostomy were identified, so at our hospital, there seems to be no contraindication to perform an early tracheostomy as other authors refer5.
Functional prognosis
The functional result was not modified by the time of performance of the tracheostomy, as in other published reports9. However, like the previous variables, the global clinical scenario of the patient reflects that those critically ill and with worse functional prognosis are the ones that will require and benefit from the tracheostomy at hospitalization.
Conclusions
According to our results, we can conclude that there are multiple factors associated with the need for tracheostomy in adult patients with severe TBI, most of them according to the clinical conditions and that it is possible to predict at admission which patients will need this procedure during their hospitalization, using scales such as CRASH, IMPACT, SAPS II, and APACHE II that have already been validated as prognostic in TBI. This scale (Table 6) can be used to calculate the need for tracheostomy; however, further studies are required to validate these results in a statistical way. At time, we are running a validation test of this scale in a new protocol.
Table 6 Predictive tracheostomy scale in a patient with severe head trauma
| Points | |
|---|---|
| If CRASH ≥ 60 | 1 |
| If impact ≥ 15 | 1 |
| If SAPS II ≥ 49 | 1 |
| If APACHE II ≥ 19 | 1 |
| If subdural volume ≥ 20 cc | 1 |
| If revised trauma score ≤ 5.89 | 1 |
| If age ≥ 44 years | 1 |
| If Glasgow Coma Score ≤ 6 | 1 |
| If dilated pupil one or both | 1 |
| If collapse or absent cisterns | 1 |
| Maximum 10 |
We propose a new term for the performance of tracheostomy as ultra-early on the 4th day or earlier, early on day 7 or before, and late if it is performed after the 7th day of hospitalization. This proposal should be analyzed in subsequent studies to define its usefulness.
Until today has not been possible to standardize the ideal time to perform the tracheostomy in patients with severe TBI or to determine what patients are candidate for this procedure, being able to perform a prognostic scale could reflect benefit for institutions, as it will decrease the total length of stay and associated morbidities such as days of AMV and pneumonia frequency.
