INTRODUCTION
Tracheostomy is a surgical procedure that establishes an artificial airway by creating an opening in the anterior tracheal wall. It is performed emergently or electively to facilitate long-term ventilation, airway protection, and/or secretion management. In the United States, more than 100000 tracheostomies are performed every year to aid with prolonged ventilation or in patients with weaning difficulties[1,2]. Tracheostomy offers several clinical advantages with a 35% to 50% decrease in airway resistance, improved tidal volume and airflow dynamics, contributing to better patient-ventilator synchrony[1-4], patient comfort, reduced sedation needs, improved oral hygiene, ability to communicate and swallow, and earlier transition out of the intensive care unit (ICU)[1-5]. However, tracheostomy is associated with complications such as bleeding, pneumothorax, tracheal/Laryngeal injuries, tracheo-innominate artery fistula, mediastinitis, and vocal cord dysfunction[1,5-8], as well as longer-term effects including impaired body perception/self-esteem, decreased quality of life, and a potentially significant caregiving and financial burden for patients and families[9].
The optimal timing of tracheostomy placement remains a topic of debate, with conflicting evidence regarding its impact on patient outcomes, such as mortality, ventilator days, and ICU length of stay (LOS)[6,8]. The 2013 TracMan trial, a large multicenter randomized controlled trial, highlighted this uncertainty by demonstrating no significant difference in 30-day mortality or duration of respiratory support between early (< 4 days post-intubation) and late (> 10 days) tracheostomy[7]. A meta-analysis by Chorath et al[10] in 2021 included 3145 patients and found a significant reduction in the incidence of ventilator-associated pneumonia (VAP), shorter ICU-LOS and duration of mechanical ventilation, but no difference in mortality with early tracheostomy[10]. A recent systematic review and meta-analysis from 2024 reported decreased ICU-LOS, duration of mechanical ventilation, and moderate reduction in mortality with early tracheostomy, but did not affect the incidence of VAP, highlighting inconsistencies in the existing literature and emphasizing the need for further trials[11].
Interestingly, the period covered by the 2021 meta-analysis (articles published before 2019) and the period covered by the 2024 meta-analysis (articles published before 2023) may partially reflect differences between the pre-pandemic and pandemic era. The coronavirus disease 2019 (COVID-19) pandemic has had a significant impact on health care practices, including altered critical care delivery, changes in airway management strategies, and resource allocation. These changes may have led to shifts in the timing, indications, and procedural approach to tracheostomy in critically ill patients. Accordingly, our study aims to examine differences in tracheostomy practices before and after the COVID-19 pandemic, providing insight into how the pandemic-era and evolving clinical guidelines may have reshaped contemporary practice patterns.
MATERIALS AND METHODS
Study design and participants
This single-center retrospective cohort study included consecutive adult patients (18 years or older) undergoing tracheostomy at a community hospital setting in New York in two discrete periods between October 2012 to February 2020 (pre-pandemic/pre-COVID) and June 2022 to July 2024 (post-pandemic/post-COVID). Patients admitted during the COVID-19 pandemic, defined in our study as the period between March 2020 and May 2022, and those with active COVID-19 diagnosis were excluded. Institutional Review Board approval was obtained by Flushing Hospital Medical Center, approval No. 2234080-1. Epic System Inc. SlicerDicer was used to extract the records of eligible patients and variables of interest were automatically or manually extracted by three authors. The dataset was deidentified and patient consent was waived.
Variables
The study population was grouped according to study periods (pre- and post-COVID). Variables of interest included patient demographics (age, gender, race/ethnicity), comorbidities (diabetes mellitus, hypertension, coronary artery disease, chronic kidney disease, and chronic obstructive pulmonary disease), time-based outcomes (length of hospital stay, time to palliative care consultation, time to surgical consultation, time from admission to tracheostomy, and time from intubation to tracheostomy), and mortality at 2 months and 6 months.
Statistical analysis
Continuous variables were summarized as mean and SD or median and (Q1, Q3), based on the shape of the variable distribution. Categorical variables were reported as counts and percentages. Univariate analysis between continuous variables were conducted with t-test/Mann-Whitney U test and between categorical variables with χ2/Fisher’s exact test as appropriate.
Baseline characteristics and outcomes between the pre-COVID and post-COVID cohorts were compared using standardized mean differences, along with corresponding 95% confidence intervals (CIs). The standardized mean difference quantifies the magnitude and direction of differences between groups on a standardized scale, with a range of -1 to 1 and with values closer to -1 or 1 indicating larger effect sizes. This approach was used to assess differences in baseline characteristics between study periods (pre- vs post-COVID period) and between survivors and non-survivors.
Multivariable analyses were performed to evaluate factors associated with mortality, defined as death occurring within 6 months of the index hospitalization and modeled as a binary outcome. Logistic regression models were constructed using a stepwise approach, sequentially incorporating covariates to assess their independent effects. The initial model included COVID-19 period (pre-COVID vs post-COVID), followed by models adding interval-related variables, baseline comorbidities, and finally a fully adjusted model including all relevant demographic and clinical covariates. Model results were reported as odds ratios with corresponding 95%CI.
All statistical analyses were performed with the use of R language version 4.2.3 within the R Studio software version 2023.03.0 + 386.
DISCUSSION
In this single-center cohort study examining the differences in tracheostomy practices between the pre- and post-COVID eras, we observed a significant shift toward earlier involvement of surgical and palliative care providers, and shorter times from admission and intubation to tracheostomy following the pandemic. Of note, these changes occurred without any improvement in the 2-month and 6-month survival. These findings in the post-COVID period may reflect the changes in the ICU workflow, decision-making and practice patterns that were adopted during the pandemic[12].
Timing of tracheostomy and outcomes
Despite decades of investigation, the ideal timing of tracheostomy remains controversial. In the 2013 large multicenter randomized controlled Tracman trial, that included 909 patients and compared early [< 4 days of critical care unit (CCU) stay] to late tracheostomy (> 10 days of CCU stay) in critically ill mechanically ventilated patients, there were no significant differences in 30-day all-cause mortality (30.8% vs 31.5%, respectively), in 1- and 2-year survival, days of respiratory support, or LOS in the CCU/hospital[7]. A delayed tracheostomy after 10 days of CCU stay was recommended based on these findings and on the rationale that tracheostomy before 10 days does not offer any significant benefit and increases the number of unnecessary procedures, while a waiting approach avoids procedure-related complications. Further, the multicenter randomized trial SETPOINT2 by Bösel et al[13] found no difference in the 6-month functional outcomes, mortality, ICU-LOS and duration of mechanical ventilation between early (≤ 5 days) vs late (≥ 10 days) tracheostomy for patients with severe ischemic or hemorrhagic stroke on invasive mechanical ventilation[13]. Subsequent meta-analyses have yielded conflicting results. A meta-analysis by Chorath et al[10] in 2021 included 3145 patients and found a significant reduction in the incidence of VAP, shorter ICU-LOS and duration of mechanical ventilation, but no difference in mortality with early tracheostomy[10]. However, a recent systematic review and meta-analysis from 2024 by Merola et al[11], including 3586 patients from 19 randomized trials, reported decreased ICU-LOS, duration of mechanical ventilation, and moderate reduction in mortality with early (within 7-10 days of intubation) vs late tracheostomy (after 7-10 days of intubation), but did not affect the incidence of VAP, highlighting inconsistencies in the existing literature and emphasizing the need for further trials[11]. In our study, in alignment with most current evidence, despite a shift towards earlier placement of tracheostomy in the post-COVID era, there was no benefit in terms of short- and long-term mortality. Different findings between studies may be partially explained by substantial heterogeneity in the definitions of early vs late tracheostomy. Variable timing thresholds have been used ranging from 4 days to 14 days of endotracheal intubation, while some studies have included populations based on anticipated duration of mechanical ventilation[7] or clinical trajectory rather than fixed temporal cutoffs. This lack of standardization complicates synthesis of evidence and the generalizability of meta-analyses findings, and may contribute to the inconsistent association between tracheostomy timing and main outcomes such as mortality, ventilator duration, VAP incidence, and ICU-LOS.
Impact of COVID-19 pandemic on airway management
The COVID-19 pandemic fundamentally changed the airway management strategies in critically ill patients. In the beginning of the pandemic, concerns regarding aerosolization, staff safety, and limited resources led many institutions to delay tracheostomy placement, while subsequent experience and evolving guidelines supported earlier intervention[14,15]. Early tracheostomies were performed to facilitate weaning from mechanical ventilation and to decrease side effects associated with high requirements of sedation and analgesia in intubated severely-ill COVID-19 patients, such as paralytic ileus, delirium, nausea, drug accumulation and tachyphylaxis[12]. Early data from the United Kingdom showed a decrease in ventilator days, decreased ICU-LOS, and a higher 30-day survival despite similar APACHE-II scores between severely-ill COVID-19 patients with and without tracheostomies[14]. In a meta-analysis that summarized the findings of 14 non-randomized studies, early tracheostomy was associated with a decrease in mortality and ICU-LOS in COVID-19 patients, increasing the availability of ICU beds[12,15].
In our institution, the decision to perform a tracheostomy prior to the pandemic had historically been towards late tracheostomy. The post-COVID cohort showed earlier time to surgical consults (9.5 days vs 15.0 days, P < 0.001), shorter time to tracheostomy from admission (16 days vs 18.5 days, P = 0.003) and shorter time to tracheostomy from intubation (11 days vs 15.5 days, P = < 0.001). Our institution, located at the epicenter of the pandemic, demonstrated a durable shift toward earlier surgical consultation and tracheostomy placement even after the pandemic subsided. This finding suggests that pandemic-driven adaptations may have affected clinician risk tolerance, interdisciplinary coordination, and may have facilitated earlier procedural decision-making in high-acuity non-COVID-19 patients.
Role of palliative care in airway and tracheostomy decision-making
One of the most notable findings of this study is the significantly reduced time to palliative care consultation in the post-COVID era. Early involvement of palliative care is one of the most important factors to facilitate decision-making by the patient and/or the surrogate decision maker. Only 14% of the world’s population have access to palliative care aimed at improving quality of life and relieving pain for patients facing serious illness, regardless of disease stage or diagnosis, improve communication with and provide education to families, decrease non-necessary interventions and to provide resources for counseling and bereavement services[16,17]. Despite its growth since its early inception in the 1960s, and robust data supporting that earlier incorporation of palliative care improves patient outcomes in the ICU, awareness of its role and potential benefits varies among providers[17,18]. In addition, palliative care is often viewed by patients and family members as “giving up” and “end of life”, and this perspective hinders effective communication and timely consultation[17].
However, during the COVID-19 pandemic, palliative care played a critical role to initiate goals of care (GOC) discussions, end of life care and a bridge between patients, providers and families. Due to quarantining measures, as families were unable to visit and physicians were tasked with a high workload, palliative care provided emotional support to stress, depression and burnout[19-23]. GOC is a term commonly referred to the entire process of medical decision-making whereby interactive discussions between clinicians and their patients (or patient surrogates) provide a better understanding of the illness, elicit values and preferences to direct and formulate a personalized care plan for future medical and/or end-of-life care[18-20]. On average, less than 30% of patients with serious illness have GOC conversations or these are conducted too late to make a difference[18,22,24]. It is likely that the pandemic created stronger bonds between team members, leading to earlier inclusion for determination of GOC[24]. With the onset of the COVID-19 pandemic, many hospitals had to quickly adapt to increased hospital admissions, ICU volume, and mortality rates, which in turn created a need to conduct GOC conversations in a fast-paced environment in the setting of rapid patient deterioration[20,24,25]. These interactions, mostly conducted via telephone or telehealth due to social-distancing/quarantine, created challenges for building rapport, decision-making, and providing care aligning with patients’ values[26].
Different institutions implemented variable systems to manage the increase in patient volume. Hospitals implemented multimodal interdisciplinary GOC programs, created emergency department palliative-care teams, or applied guidelines that facilitated timely GOC conversation and documentation, resulting in decreased ICU and in-hospital mortality, shorter LOS, fewer non-necessary interventions, and lower health care related expenses at the end of life[21,24,27-29]. In the post-COVID era, providers and palliative care teams are encouraged to drive conversations surrounding harm reduction, patients’ and families’ views on future interventions, risks and benefits for the patient as well as an emphasis on making the most of life through validated GOC frameworks, positive role-modelling, and experiential learning[29-31]. Clinicians who are unfamiliar with the GOC process may utilize GOC frameworks to anticipate challenges during each conversation and address uncertainties through guided resources to promote a more effective care during challenging times[25].
Decision-making about tracheostomy and prolonged mechanical ventilation is emotionally complex, often with expectations from physicians being different than those of surrogate decision-makers faced with significant caregiving and financial burden and concerns about their loved ones’ quality of life. It should take into account patient wishes, past and active medical problems, short- and long-term outcomes, and meaningful recovery. The 2021 study by Xu et al[32] about communication in tracheostomy decision making reported that 45.3% of the documented conversations did not address patients’ preferences, baseline status, and long-term effects, while involvement of palliative care was associated with more effective documentation and communication.
The sustained earlier involvement observed in the post-COVID cohort (decrease in time to consult with palliative care (3 days vs 7 days, P = 0.001)) reflects increased recognition of the value of structured GOC discussions and interdisciplinary collaboration. This shift may partially explain earlier tracheostomy decision-making, as better prognostic understanding and communication of care preferences streamline procedural planning. Earlier interdisciplinary involvement is also partially attributed to an improved ICU workflow, improved decision-making post-pandemic, or a greater tendency among clinicians to advocate for earlier intervention.
Quality of life, short-term and long-term outcomes
While tracheostomy can facilitate ventilator weaning and reduce sedation requirements, it also carries important short- and long-term consequences, such as bleeding, infection, tracheal injury, and pneumonia, but patient-centered impacts - including impaired communication, altered self-identity, reduced mobility, lower quality of life, and prolonged rehabilitation - are often underemphasized[31,33,34]. More overarching patient centered aspects surrounding tracheostomy are issues involving patient autonomy and self-esteem[31,35]. Limited patient motility results in deconditioning, impaired daily activities and potentially hinders the rehabilitation process. The significant mental and physical toll and affected quality of life should be taken into consideration[9].
The effects of early tracheostomy on mortality are still unclear. A meta-analysis by Chorath et al[10] in 2021 included 3145 patients and showed no difference in mortality with early tracheostomy, while a 2024 meta-analysis showed a moderate reduction in mortality[11]. Inconsistent findings between these two studies may be explained by differences in the definition of early tracheostomy in the included data, with timing thresholds ranging from 4 days to 14 days of endotracheal intubation.
In this study, the 2-month and 6-month mortality rates were not statistically different between the pre- and post-COVID cohorts, despite earlier tracheostomy placement in the post-COVID era. These findings align with prior studies showing that early tracheostomy may reduce certain resource utilization metrics without translating into consistent survival benefits or functional outcomes[11,13].
Limitations
Despite its strengths, including real-world data, consistent data extraction, and pragmatic design, this study has several limitations. First, its single-center retrospective design limits the generalizability of the findings. Second, several clinically relevant and patient-centered outcomes were not captured, such as standardized illness severity scores (e.g., APACHE II, SOFA, Charlson Comorbidity Index), which were not consistently available/extractable throughout the study period, time to discontinuation of sedation, discharge destination, level of care at 6 months, time to initiation of ventilator weaning, success of weaning in the ICU and after hospital discharge, and detailed post-discharge outcomes. Mortality is an important, but incomplete, endpoint for assessing the clinical value of tracheostomy. Earlier tracheostomy may facilitate ventilator weaning, patient comfort, ICU throughput, and goal-concordant care, which could not be evaluated due to the retrospective design of the study and data availability. The absence of a mortality benefit should not be interpreted as a lack of overall clinical value. Residual confounding by disease severity may affect our findings. In addition, measures of patient and caregiver burden, quality of life, and long-term functional status were not assessed. Further, secular trends over a prolonged pre-COVID era may contribute to observed differences, which cannot be attributed solely to the COVID-19 pandemic. Temporal confounding may have affected our findings, including potential influences from evolving ICU protocols and guidelines, institutional culture, access to palliative care, and population changes over time. Lastly, the post-COVID cohort was small (n = 68 patients), thus limiting the statistical power and increasing the risk of type II error, which may explain the lack of statistically significant associations in multivariable models. Future prospective, multicenter studies investigating these outcomes are needed to better characterize the short- and long-term impact of tracheostomy timing and evolving practice patterns.
