Introduction

Poor outcomes in critical patients infected with SARS-CoV-2 admitted to ICUs have been reported [1,2,3,4,5,6,7,8,9]. In Italy, [3] 58% of 1591 COVID-19 patients were still in ICU 5 weeks after admission and mortality rates have been over 50% in large case series [1, 2]. In this pandemic scenario, with limited ICUs resources [10], tracheostomy seems to help COVID-19 patients to get off the mechanical ventilation [11], reducing the respiratory effort in patients with limited pulmonary reserves, shortening the dead space and enabling the suctioning of accumulated mucous.

Current protocols have recommended to delay tracheostomy for at least 14 days or longer or advocated to wait until a negative PCR [12,13,14]. However, all recommendations are neither based on the experiences of patients infected by the SARS-CoV-2 nor in a pandemic situation with overcrowded ICUs lacking proper equipment for mechanical ventilation.

The questions of an optimal strategy and outcomes in COVID-19 tracheostomies have not been answered yet. Spanish otolaryngologists began tracheostomies very early in the pandemic, on March 11, 2020. The rationale was to achieve fast weaning to enable incoming patients to take advantage of the released mechanical ventilation equipment. The critical focus of our case study is the unprecedented amount of tracheostomies in Spain that may serve as a lesson for this and potential future pandemics.

Materials and methods

Patients

On March 25, the Spanish COVID ORL Group was created to share experiences in the management of upper airway during this pandemic. The group has performed a national multicentric prospective observational study on 1890 COVID-19 critical patients undergoing tracheostomy in a total of 120 hospitals in Spain. The study prospectively collected data during seven consecutive weeks from March 28 to May 15, 2020 in each participating hospital.

Data collection

Clinical data were collected daily with an innovative approach: instant messaging via WhatsApp by participants. A daily report analysed data to keep all surgeons updated. Each participating hospital included: new COVID-19 patients undergoing tracheostomy, achieved weaning and decannulation of patients (as non-identified data). Also, the accumulated COVID-19-related mortality rate was registered. Data were available for all collaborators at any given time. Members of the COVID ORL group shared experiences in tracheostomy indications, performance and outcomes, but also in self-protection in COVID-19 scenario.

Follow-up

To assess the impact of tracheostomy in intubated COVID-19 patients, the following outcome measurements were assessed on a daily basis: (a) indication of tracheostomy: conventional or emergency tracheostomy; (b) timing for tracheostomy (early < 10 days or late > 10 days); (c) technique: open or percutaneous tracheostomy; (d) perioperative complications; and (e) final outcomes: the number of patients in whom weaning and/or decannulation could be achieved, complications and survival. Follow-up of tracheotomized patients was performed until 1 month after the procedure.

Results

Over 254 otolaryngologists had registered on March 25, 2020 to participate in the COVID ORL messenger Group. This social media platform allows the large group of ENT colleagues from 120 hospitals to exchange critical information and development of a common strategy in the pandemic.

As of May 14, 2020, with 11,493 ICUs patients confirmed by the Health Ministry of Spain [15], data for 1890 COVID-19 patients who underwent tracheostomy across 120 hospitals in Spain were available. Most of tracheostomies were performed in a very short time: 1890 tracheostomies in less than 2 months (Figs. 1, 2). From March 28 to April 20 alone, during the 3 weeks of the pandemic peak in Spain, data were entered for 1400 cases. After the peak, from April 20 to May 15, further 490 cases were added. The unequal spread of COVID-19 in Spain had an impact on the number of tracheostomies in each hospital (Fig. 3).

Fig. 1
figure 1

Cumulative COVID-19 tracheostomies performed by COVID ORL ESP Collaborative Group

Fig. 2
figure 2

Tracheostomies on COVID-19 patients per day

Fig. 3
figure 3

Distribution of tracheostomies in COVID-19 patients per hospital

Most tracheostomies (n = 1461; 81.3%) were open and the rest percutaneous (n = 429; 22.7%). Indication and timing of elective tracheostomy were usually established based on patient respiratory status by the ICU staff. Only 4.6% of COVID-19 patients (88 of 1890 patients) underwent tracheostomy at a very early stage within 7 days of intubation. The median timing of tracheostomy was 12 days after intubation (range 4–42 days). Most procedures were performed at bed-side in the ICU. While almost all tracheotomies were regular procedures, four (0.2%) were vital emergency tracheostomies.

Surgeons shared some technical modifications, such as a sub-isthmus tracheotomy approach or the use of electrocautery and harmonic scalpel, both used exclusively before opening the trachea. Most of the endotracheal tubes were found partially or subtotally blocked with sticky secretions. Only cuffed non-fenestrated tubes were used in all COVID-19 patients. Isolated cases of tracheomalacia with cuff leaking were reported.

Complication rate was low; hemorrhage was the notable major adverse outcome in 49 patients (2.6%). One patient died from bleeding. Among other adverse events were desaturation with cardiac arrest (n = 8; 0.42%), right after opening the trachea, with 5 (0.2%) subsequent intraoperative deaths. Pneumothorax after tracheostomy was reported in 3 cases.

All surgeons were able to use personal protective equipment (PPE) in all procedures. Standard PPE consisted of glasses, N95/FFP2 with an additional surgical mask placed above, two pairs of gloves and a face shield. Initially, PPE with powered air-purifying respirator (PARP) was available only in two hospitals. Just one vital emergency tracheotomy was performed without PPE under less optimal conditions, with only double surgical mask and glasses. The surgeons who performed this procedure were asymptomatic at 2 weeks after procedure and tested antibodies proved negative.

An overview of all outcomes is listed in Fig. 4. A close follow-up of 1616/1890 (85.5%) patients at the time of reporting showed that in 842 (52.1%) patients, weaning was achieved, while 391 (24.2%) were still under mechanical ventilation, while 383 (23.7%) patients died from COVID-19. Among those in whom weaning was successful (n = 842), decannulation could be achieved in 683/842 (81%) at the cut-off time.

Fig. 4
figure 4

Outcome of 1890 tracheostomies for critical COVID-19 patients in Spain

Discussion

On early March, 2020, at a very critical moment of the pandemic in Spain and without published data yet from initial experiences in China or Italy, Spanish otolaryngologists began their first tracheostomies in Madrid (March, 11) and Barcelona (March, 16), with the onrush of more than a thousand patients in their respective overcrowded ICUs, which doubled its current capacity. A large group of otolaryngologists started an innovative exchange of critical information focused on tracheostomy on COVID-19 patients under mechanical ventilation in an attempt to coordinate among themselves and to develop a common strategy.

On May 15, the Spanish Health Ministry confirmed 11,493 ICU accumulative admissions [15]. Between March 28 and May 15, the Spanish COVID Group, encompassing 120 hospitals, performed 1890 tracheostomies, the equivalent of 16.4% of all registered ICU patients with mechanical ventilation in Spain. Herein, ENT Departments not participating in our social media group have not been considered; therefore, the number of total tracheostomies in Spain can be presumed to be even higher. To the best of our knowledge, there are no published data about such a high number of tracheostomies in COVID-19 critical care units.

The rate of tracheostomies (16.4%) is similar to that reported in large cohort studies on Acute Respiratory Distress Syndrome (ARDS) in the LUNG-SAFE trial [16]; 13% of ARDS patients underwent tracheostomy in ICUs of 50 countries in five continents, reaching 15.7% in European countries alone. The most recent literature in COVID-19 patients (Table 1) shows only small series that add up to 150 tracheostomies and a more large case report about 564 from 78 UK NHS hospitals, the COVID TRACH Collaborative Group [11].

Table 1 Tracheostomy in COVID-19 patients (Refs. [1117,18,19,20,21,22,23,24])

Timing of the tracheostomy remains controversial [25]. Our data about the median timing for tracheostomy are 12 days (range 4–42 days) after tracheostomy. Only 4.6% of our COVID-19 patients (88/1890) underwent tracheostomy at a very early stage within 7 days of intubation. In the COVID Trach Collaborative Group [11], only 18% of patients underwent tracheostomy before 10 days of intubation. In the recent literature (Table 1), tracheostomy related to COVID-19 was indicated after 10–21 days of mechanical ventilation. A systematic review of 222,641 patients from 43 studies published by Adly [26] could show that early tracheostomy within 7 days reduces the complications and mortality. However, most of the patients had healthy lungs, whereas critical COVID-19 patients presented severe Acute Respiratory Distress Syndrome (ARDS). Several ENT academies have advocated to delay tracheostomies in COVID-19 patients for 14 days or longer after intubation, some including PCR negativity, to reduce the risk of contagion [12,13,14]. An international consensus suggests that COVID-19 tracheostomy should be performed after day 10 of invasive ventilation and considered only when patients display signs of clinical improvement [27]. In any case, all recommendations are currently based on patients not infected by the SARS-CoV-2. Also, the pressure of overcrowded ICUs with a high demand for equipment for mechanical ventilation had not been considered in any recommendation. This calls into question whether delaying a useful technique that potentially supports weaning and eventual decannulation has to be delayed because of a potential risk of contagion that has not been demonstrated so far. Due to the type of data collection, it could not be analyzed whether early or late tracheostomy made a difference in time of rate of weaning or decannulation. No comparison was made with long-term intubated, non-tracheostomized patients.

Technical issues in the execution for COVID-19 tracheostomy were highlighted after lessons learned in the 2003 SARS outbreak [28]: expert surgical teams, avoiding repeated connections and disconnections of ventilator circuits, minimal use of diathermy, stopping mechanical ventilation just before opening the trachea, closed suction circuits and negative-pressure room whenever feasible, as recommended in the consensus document elaborated later based on this experience [29]. But in COVID-19 pandemic, new scenarios have come up. In some centers, the onrush of cases in need of mechanical ventilation led to the creation of COVID-19 critical care units, sometimes located far from the operation theaters and most of our tracheostomies were performed at the bed-side. In the CovidTrach collaborative cohort [11], 55% of tracheostomies were performed in the operation theatre and 45% in intensive care. In our cohort, the decision whether to perform a surgical or a percutaneous tracheostomy depended on the experience with the latter. More open tracheostomies (81.3%) were performed by the group, as were in the CovidTrach collaborative cohort [11].

Complication rates in tracheostomy were low. With 2.6% hemorrhages requiring surgical revision was the most frequent one, similarly as in other reports [11, 22] or the ARDS-related TRAC-MAN trial study [30]. Eventually, one patient died from bleeding. Among other adverse events, 5 (0.2%) intraoperative deaths right after opening the trachea occurred.

Poor survival rates have been published for intubated critical patients without tracheostomy infected with SARS-CoV-2 admitted to ICUs (Table 2), with mortality rates in large case reports of 26–54.4% [1, 2]. Case reports about tracheostomies in COVID-19 are scarce and with too few cases (Table 2) to allow further conclusion. The overall COVID-19-associated mortality in our cohort at the time of reporting was 23.7% (n = 383/1616). In the CovidTrach collaborative cohort [11], only 12% mortality rate was found; however, 48% (n = 221/465) of patients were still under mechanical ventilation at the time of completing the study. In our study, 391/1616 (24.2%) were still under mechanical under ventilation at the cut-off time. A review of clinical studies about tracheostomies in Acute Respiratory Distress Syndrome (ARDS) patients undergoing tracheotomy, with very damaged lungs, such as COVID-19 pneumoniae, shows a mortality rate of 23–30% [16, 30].

Table 2 Mortality rates of critical COVID-19 patients in ICUs without tracheostomy (Refs. [1,2,3,4,5,6,7,8,9])

In this pandemic scenario, with limited ICUs resources [10], our data show that about 52.1% (n = 842/1.616) of patients undergoing tracheostomy get off the ventilator. The Covid Trach collaborative group from UK NHS reports similar weaning success rates (52%; n = 219/465). The results of both reports suggest that the tracheostomy could be effective in the weaning and decannulation of COVID-19 patients, although data are preliminary yet and further studies are required.

Conclusion

COVID-19 recommendations on tracheostomy technique, perioperative management and self-protection have been very similar to SARS 2003 outbreak protocols. SARS-CoV-2 is a new virus that has led to a pandemic, with an onrush of COVID-19 cases that needed to be dealt with. In Spain, a large group of ENT colleagues from 120 hospitals created an innovative exchange of critical information through social media platform that proved usual for exchange of information and development of a common strategy. From 1,6161 patients, weaning could be achieved in 52% with an overall mortality of 23.7%. Nevertheless, 24.2% still remain connected to mechanical ventilation at the cut-off time despite the long-term follow-up due to chronicity and severity of COVID-19 disease. Tracheostomy may have helped to the recovery of more than 800 critically ill COVID-19 patients and subsequently released more than 800 mechanical ventilators for incoming patients, but it remains unclear whether tracheostomy influences the survival of these patients and which might be the ideal timing to perform it. Further studies are required to try to clarify these issues.

Limitations

This study has several limitations. Data were collected prospectively by an instant messaging group on-site. Subsequently, data are incomplete or missing in 14.5% of the cases. The follow-up time of 1 month seems to be too short as at the cut-off time. 24.2% patients were still tracheotomized and under mechanical ventilation.