INTRODUCTION

Coronavirus disease 2019 (COVID-19) has caused more than 5.4 million deaths worldwide and is the most important health crisis since the influenza pandemic of 1918 ¹. As of July 24, 2023, nearly 4.5 million cases of COVID-19 have been diagnosed in Peru, with 221,203 deaths and a case fatality rate of 4.9% ². Peru is the seventh country with the highest number of deaths in the world due to COVID and the third in Latin America ³.

Patients with acute respiratory distress syndrome (ARDS) and COVID-19 usually remain intubated for prolonged periods ⁴. Tracheostomy plays a fundamental role in the management of these patients when the intubation and mechanical ventilation time exceeds 14 days (prolonged mechanical ventilation) by allowing better airway management and facilitating the withdrawal of invasive ventilatory support. Tracheostomy has been associated with less need for deep sedation, shorter weaning time from the mechanical ventilator, shorter stay in the intensive care unit (ICU) and shorter hospital stay ⁵.

However, most ICU patients who are tracheostomy candidates have a higher risk of death due to the following characteristics: (a) ICU stay increases the risk of nosocomial infections ⁶; (b) prolonged mechanical ventilation, which requires longer sedation, analgesia and muscle relaxation, causes polyneuropathy and myopathy associated with immobility ⁷, (c) delirium upon withdrawal of sedation and analgesia; (d) the severe inflammatory process (cytokine storm) caused by COVID-19 directly affects gas exchange ⁸ and (e) the advanced age and presence of comorbidities that patients admitted to the ICU usually present ⁹.

Likewise, tracheostomy is a procedure that is not free of risks and potential complications such as hemorrhage, surgical site infection, pneumothorax and tracheal stenosis. So, due to all these conditions, tracheostomy is not always safe ⁴. Therefore, in order to improve clinical decision making and optimize results, the factors associated with higher mortality should be identified even if a patient with severe COVID-19 has the indication for tracheostomy, taking into account the high complexity and the accumulated mortality due to the baseline conditions of these patients.

So far, the factors related to mortality in critically ill patients tracheostomized due to COVID-19 are the following: high values of C-reactive protein (CRP) at the time of tracheostomy ¹⁰, higher inspired oxygen fraction (FiO₂) requirement ¹¹ and early timing of tracheostomy ¹². Unfortunately, most of this information comes from developed countries and does not reflect the reality of Peru, a country with a fragile health system. In this context, critical illness due to COVID-19 was more frequent, aggressive and fatal ¹³, since after the three waves of this disease, the number of respiratory critical patients and tracheostomized patients increased, with a mortality rate of 32.9% in our ICU ¹⁴.

This study aimed to describe the demographic, clinical, laboratory and therapeutic characteristics as well as to identify predictors that allow timely identification of patients with COVID-19 who are at higher risk of death after percutaneous tracheostomy, since this group of patients have prolonged length of stay and high mortality rates ¹⁵.

MATERIALS AND METHODS

RESULTS

Of the 329 critically ill patients hospitalized in the COVID-ICU, 83 (25.2%) underwent tracheostomy as part of their evolution, but 10 did not meet the inclusion criteria (one patient was admitted with a diagnosis other than COVID-19 pneumonia, 2 patients did not have the outcome of death or decannulation and 7 patients had incomplete clinical and laboratory data). We included 73 medical records of patients admitted to COVID-ICU with a diagnosis of SARS-CoV-2 infection, COVID-19 pneumonia and percutaneous tracheostomy carriers.

The median age was 59 (IQR: 52-66) years, 72.6% of the patients were male and only 6.8% were vaccinated against SARS-CoV-2. The most frequent comorbidities were obesity (68.4%), type 2 diabetes mellitus (35.6%) and arterial hypertension (34.2%).

The median ICU and mechanical ventilation (MV) length of stay was 32 (IQR: 24-41) and 31 (IQR: 24-43) days, respectively, the median tracheostomy time was 18 (IQR: 10-25) days (decannulation or death) and the time at which tracheostomy was performed was 17 (IQR: 15-20) days counted from the time of intubation.

Regarding laboratory tests, the median PaO₂/FiO₂ at the time of tracheostomy was 213 (IQR: 175-270), procalcitonin was 0.22 ng/mL (IQR: 0.13-0.40), CRP of 6.35 mg/dL (IQR: 3.2-12.1), and leukocytes of 11,950 cells/mL (8.810-15.510). Concerning mechanical ventilation variables, FiO₂ at the time of tracheostomy was 30% (IQR: 28-35), PEEP (positive end-expiratory pressure) was 7 cm H₂O (IQR: 6-8), inspiratory pressure was 15 cm H₂O (IQR: 14-18) and the tidal volume was 466 mL (IQR: 406-509).

Regarding the treatment, 71 (97.3%) patients received antibiotics, 20 (27.4%) received antifungals and 9 (12.3%) received vasopressors at the time of tracheostomy. MV-associated pneumonia was found in 66 patients (90.4%), bacteremia in 14 (19.2%) and urinary tract infection in 9 (12.3%) patients. We found that 27 (37%) tracheostomized patients died during ICU stay and 46 (63%) were successfully decannulated and discharged (Table 1).

No significant differences were found regarding the mortality in patients according to age group, sex, vaccination status, comorbidities, length of stay in the ICU, time on mechanical ventilation, time of tracheostomy, or according to the treatment. However, the time of tracheostomy was longer in those who achieved decannulation than in those who died (21 vs. 15 days; p=0.043). Likewise, PaO₂/FiO₂ was lower in the group of patients who died (p<0.001), while a procalcitonin value >0.5mg/dL, higher FiO₂, inspiratory pressure and peak pressure at the time of tracheostomy were associated with a higher risk of death.

Factors associated with mortality were evaluated by Cox proportional hazards regression. In the adjusted analysis, procalcitonin >0.50 at the time of tracheostomy was shown to be a factor associated with mortality with a aHR of 2.40 (95%CI: 1.03-2.59). Likewise, a PaO₂/FiO₂ level ≤150 mmHg was independently associated with higher mortality, with an aHR of 4.40 (95%CI: 1.56-12.60) (Table 2). Figure 1 presents the Kaplan-Meier survival curves for the variables PaO₂/FiO₂ (p<0.001) and procalcitonin (p<0.001).

Figure 1 Kaplan-Meier survival curves according to PaO₂/FiO₂ and procalcitonin level, censored at 30 days. 

DISCUSSION

We found that 36.9% of all patients who were tracheostomized by COVID-19 died. Furthermore, we were able to establish two risk factors for mortality, which are the procalcitonin value >0.5 ng/dL, and a low PaO₂/FiO₂ ratio (less than or equal to 150) at the time of tracheostomy.

Reported mortality rates in patients with tracheostomy and COVID-19 ranges from 15 to 66% ^18-21. A meta-analysis of 5268 tracheostomized patients reported a pooled mortality of 22%; however, there was great heterogeneity among the included studies ¹⁹. It has also been reported that critically ill patients who underwent tracheostomy had lower mortality than those who were not tracheostomized, even in severely ill patients with APACHE II >17 ^18,21. In Peru, the overall mortality due to ARDS associated with COVID-19 was 25% ²¹, which is lower than that in our cohort of tracheostomized patients, perhaps due to the low proportion of patients immunized against COVID-19 and longer mechanical ventilation and intensive care stay.

Although the fact that vaccines arrived in Peru in February 2021, vaccination in our region started some months later ² and that the number of COVID-ICU beds was not constant throughout the pandemic ¹⁴, we could not find statistical differences in relation to deaths and decannulated patients according to the infection period (2020, 2021, 2022), probably because most of our patients were not vaccinated at the time of the study, and the management of these critical patients was always protocolized and homogeneous, since it was always in charge of highly trained human resources in intensive care ²³.

The benefits of performing an early tracheostomy (≤7 days) in critically ill patients with pneumonia are the reduction of ventilator-associated pneumonia (VAP), more days free of mechanical ventilation, and shorter stay in the ICU ²². However, these benefits have not been demonstrated in critically ill patients diagnosed with COVID-19 ²³, where even higher mortality has been found in those patients with COVID-19 undergoing early tracheostomy (≤ 14 days from orotracheal intubation) compared to late tracheostomy ^16,24,25. In our study the time from intubation to tracheostomy was not associated with mortality, and although most of our patients underwent late tracheostomy, we did not find differences in the incidence of VAP, days of mechanical ventilation, or ICU length of stay, between decannulated and deceased patients. We presume that, due to the complexity of these patients, achieving ventilatory stability was preferred to be able to perform the procedure in most cases.

Usually, tracheostomy is a safe procedure in which the most frequently reported complications are bleeding (7-20%), emphysema (4.4%), stenosis or obstruction (2%) and false airway, with a slightly lower number of complications in percutaneous tracheostomies ^20,26,27. Another alternative is bronchoscopy-guided percutaneous tracheostomy, which allows direct visualization of the airway, increasing the safety of the procedure by allowing the location of the cannulation site, with a low incidence of late complications (<1%) ²⁸. And although tracheostomy is a procedure that can cause aerosols and therefore a higher risk of SARS-CoV-2 infection, a study showed that no health worker involved in the tracheostomy procedure developed COVID-19 infection ⁹.

Peak pressure is the pressure reached during inspiration, when air is pushed into the lungs and is a measure of airway resistance ²⁹. In the management of the patient with respiratory distress syndrome it is recommended that this value be <25 cm H₂O to avoid pressure damage (barotrauma) ³⁰. We found that this value was higher in patients who died, both in the bivariate analysis and in the crude regression model, although the adjusted regression model did not show a significant value.

Procalcitonin is an indicator of COVID-19 severity, and serial procalcitonin measurements may be useful in predicting disease prognosis ^31,32. We found that 85% of patients with a high procalcitonin value died. These patients probably had a more intense hyperinflammatory state due to COVID-19, or bacterial superinfections ³³⁾ at the time of tracheostomy, which limited their recovery. Likewise, several studies have reported an inversely proportional association between PaO₂/FiO₂ and mortality in patients with COVID-19 ^13,22,33,34. In our study, the median PaO₂/FiO₂ at the time of tracheostomy was higher than that reported by other authors, who coincidentally observed higher mortality ¹⁸, a value less than or equal to 150 mmHg, was the most important factor associated with mortality in our cohort. Due to this finding, it is suggested to avoid tracheostomy when PaO₂/FiO₂ lower or equal to 150 mmHg.

Our study has some limitations. The main one was its retrospective nature, which did not allow us to fully evaluate other confounding variables. Another important limitation was the small number of patients, which reduced the possibility of controlling for some confounding factors, and this could also explain why the confidence intervals were so wide and why many variables did not present statistical association, such as the time of tracheostomy (early vs. late), period of infection or exposure to vaccines. Likewise, we cannot affirm whether the elevated procalcitonin level at the time of tracheostomy is due to persistence of COVID-19 or to an added infectious complication and therefore explain the lower survival of this group of patients as in other studies ³⁵. Likewise, complications related to the percutaneous tracheostomy procedure in the short term (bleeding, pneumothorax, false airway, etc.) and long term (stenosis, tracheomalacia, fistulas, etc.) were not studied. Finally, it was impossible to measure the impact of changes in treatment strategies (corticosteroids, tocilizumab, etc.) and control methods implemented throughout the pandemic ²⁰, which directly affect the practice of tracheostomy.

In conclusion, approximately one third of the patients in this cohort died during the study period. Among the factors associated with higher mortality in tracheostomy patients were a procalcitonin value >0.50ng/dL, and a PaO₂/FiO₂ value ≤150 at the time of performing the tracheostomy. Timely recognition of these factors should call for reflection of the personnel performing this procedure, to defer it until oxygen and infectious conditions improve, and thus decrease mortality in this group of critically ill patients.