CT Scans & Early Tracheostomy

Today's FLARE will address two clinical questions:

Subscribe to the latest updates from FLARE Advances in Motion

Thank you for subscribing!

Error: Please enter a valid email address.

Email Address Submit

1. What is the utility of chest CT in patients with suspected or confirmed COVID-19?

2. What is the role for early tracheostomy in patients with respiratory failure from COVID-19?

In this section, we review the early literature on the role of CT scanning in the screening, diagnosis, and management of COVID-19 patients.

Courtesy of Dr. Vlad Vinarsky

What are the Typical Chest CT Findings in COVID-19?

Several case series have described the incidence of various CT abnormalities in patients with COVID-19. Typical CT findings include bilateral, multifocal rounded, patchy, and peripheral ground glass opacities (GGO), sometimes with superimposed septal thickening, or rounded consolidations with surrounding GGOs (Simpson et al. 2020). These findings are nonspecific, and similar findings are common in organizing pneumonia, aspiration, viral and some bacterial pneumonias.

In fact, Bai et. al. compared CT scans of COVID-19 patients to scans from patients with other viral pneumonias (Bai et al. 2020). COVID-19 pneumonia was more likely to have a peripheral distribution, ground-glass opacity, fine reticular opacity (56% vs. 22%, p<0.001), and vascular thickening (59% vs. 22%, p<0.001) compared to other viral pneumonias. COVID-19 pneumonia was less likely to have a central and peripheral distribution (14.% vs. 35%, p<0.001), pleural effusion (4.1 vs. 39%, p<0.001) and lymphadenopathy (2.7% vs. 10.2%, p<0.001). Thus, while some findings are more common in COVID-19, none of the abnormalities are present or absent exclusively in this disease. There is, however, general consensus that consolidation with a cavitation or tree-in-bud opacities would be atypical for COVID-19.

What is the Sensitivity of Chest CT in Diagnosis of COVID-19?

The fundamental challenge to evaluating the sensitivity and specificity of a CT scan for the diagnosis of COVID-19 is the absence of a “gold standard.” The most sensitive assay is likely viral testing on BAL fluid (W. Wang et al. 2020), but it is rarely performed due to exposure risk to providers.

In a study of 51 patients from Wuhan, Hubei, with RT-PCR-proven SARS-CoV-2 infection, the disease was correctly diagnosed based on the CT scan and clinical information in 49 patients (Li and Xia 2020). Although the results are intriguing, it is impossible to extrapolate to the sensitivity and specificity of CT scanning in the diagnosis of COVID-19 from this retrospective analysis of a small data set of positive controls. In another intriguing study, Bai et. al. compared the ability of radiologists in the US and China to differentiate between CT scans of patients with RT-PCR-confirmed COVID-19 (total 219) and patients with other viral pneumonias (total 205, from archives) (Bai et al. 2020). The sensitivity of correctly identifying COVID-19 varied from 67% to 97% depending on the radiologist, but the specificity only varied 93% to 100% in this group of CT scans.

In a large cohort of 1014 patients in Wuhan recruited from 1/6/2020 to 2/6/2020, Ai et. al. compared CT scanning with viral testing in diagnosing COVID-19 (Ai et al. 2020). Two radiologists made the diagnosis of COVID-19 by consensus after reviewing imaging and clinical information (but blinded to the RT-PCR results). This was a provocative study because initial RT-PCR assay only diagnosed 601/1014 (59%) of the patients, whereas CT scanning picked up 888/1014 (88%) of the patients, giving CT scanning a remarkable 97% sensitivity if subsequent positive RT-PCR is taken as the gold standard (580/601). Furthermore, many of the patients with a “positive” CT scan and an initial negative RT-PCR had a later positive RT-PCR test, suggesting that CT scanning could diagnose COVID-19 earlier than viral testing.

Concerns have been raised with this study. Throat swabs were used for viral testing, which appear to be much less sensitive than nasal swabs in several independent studies (Zou et al. 2020; W. Wang et al. 2020). Indeed, the authors admit that using “RT-PCR assays with relatively low positive rate as reference” can overestimate the utility of CT scans. And, in fact, in a retrospective analysis of CT findings of COVID-19 relative to the onset of symptoms, with early (0-2 days) and late (6-12 day) symptom-onset groups, more than half (20/56) of the COVID-19 patients in the early group had normal CT scans (Bernheim et al. 2020) . Furthermore, in a study of 55 asymptomatic patients with positive pharyngeal swab for SARS-CoV-2, CT scan was normal in 29% (16/55) of the cases (Y. Wang et al. 2020). In aggregate, these studies suggest that CT scanning is a less sensitive test for COVID-19 compared to viral testing, especially early in the disease.

To Scan or Not To Scan?

The most likely explanation for the discrepancy between the sensitivity of CT scanning in different studies is the difference in the quality of viral RT-PCR test, possibly related to sampling variability. The sensitivity of viral testing likely increases with additional sampling. After a consensus gold-standard viral test for COVID-19 becomes widely adopted and available, future larger studies that compare adequately-sampled viral tests to CT scans, preferable at the same stage of the disease, will determine the actual sensitivity of CT scanning in COVID-19 diagnosis.

For now, the American College of Radiology recommends:

• CT chest should not be used to screen for or as a first-line test to diagnose COVID-19
• CT should be used sparingly and reserved for hospitalized, symptomatic patients with specific clinical indications for CT. Appropriate infection control procedures should be followed before scanning subsequent patients, and portable CT should be considered
• A normal chest CT does not exclude COVID-19 infection; An abnormal CT is not specific for COVID-19 diagnosis

Similarly, in a statement published on 3/10/2020, Society of Thoracic Radiology (STR) declares:

• At this time, the STR does not recommend routine screening for the diagnosis of patients under investigation for COVID-19
• CT scanning should be restricted to COVID+ patients who are suspected of having complicating features such as abscess or empyema, where imaging may alter management

In the cases where clinical suspicion for COVID-19 remains high despite negative viral testing, with acknowledgement of the limitations of radiographic studies, a multidisciplinary discussion between clinical care teams should guide decision-making regarding the risks and benefits of additional imaging. Given that PCR testing is also low sensitivity, many organizations are evaluating strategies for improving clinical decision making by using both CT and PCR in appropriate patients. The challenge is always integrating all of these individual reviews together into sensible care.

What is the Role of Early Tracheostomy in Patients with Respiratory Failure from COVID-19?

Courtesy of Dr. Jason Maley

Summary:

• With anticipated limitations to ventilator and ICU bed capacity, some clinicians have wondered if early tracheostomy could facilitate earlier discontinuation of sedation and thus earlier vent liberation, freeing resources for other patients in need
• Overall, existing high-quality evidence does not support routine early tracheostomy. Specifically, studies do not demonstrate improvement in survival, VAP rate, or duration of ventilation for patients with respiratory failure undergoing early tracheostomy (generally occurring within 7 days in studies) as opposed to tracheostomy after approximately 14 days of ventilation
• An early tracheostomy approach will increase the overall rate of tracheostomy, resulting in tracheostomy for some patients who would have been successfully extubated by 14 days
• As an aerosol-generating procedure, careful precautions should be taken if tracheostomy placement is indicated

In our usual clinical practice, we often discuss tracheostomy around 14 days if continued need for ventilation is anticipated. However, investigators have long sought to understand if select patients could benefit from earlier tracheostomy. In the current SARS-CoV-2 pandemic, anecdotal reports of prolonged ventilation (>2-3 weeks) have circulated within the critical care community, once again stimulating discussion around the timing of tracheostomy. However, published reports to-date do not describe the duration of ventilation. For reference, the median duration of ventilation was 8 days (IQR: 4-16d) in LUNG SAFE and the median length of ICU stay was 10 days (IQR: 5-19d). The median duration of ventilation among survivors in the ARMA trial was 8 days in both arms. Here, we explore the evidence regarding the timing of tracheostomy and discuss the potential risks and/or benefits of an early versus late approach.

As with many aspects of ICU management, significant variation in practice patterns exists nationwide around the typical timing of tracheostomy for respiratory failure. This is demonstrated in a study examining the 2012 national inpatient sample, a survey of acute care hospitalizations representative of ~95% of inpatient care in the US (Mehta 2016). This study found that hospitals varied in their rates of early tracheostomy (in the first week of mechanical ventilation) for patients with sepsis or pneumonia, ranging from average hospital-level early trach rate of 14.9-38.3% across the US. For patients with respiratory failure secondary to trauma, the rate of early trach was 21.9-81.9% across US hospitals. Early tracheostomy was not associated with improved survival.

Variation also exists in the definition of “early” in prospective studies. This exact definition is crucial to understanding the potential impact on outcomes – let’s dig into this more before discussing the studies. Like many other studies of the timing of interventions in the ICU (see early CRRT literature), performing an “early” intervention has important downstream effects. An “early” intervention will allocate therapy to some patients who would not have needed the therapy if it was withheld for longer. The earlier we shift the “early” intervention, the more likely it is to be overused – as we wait longer, prognostic confidence may increase and we can better predict the need for trach. In everyday practice, if we choose a specific therapy at a specific time, we will never have a counterfactual outcome for that patient – a world in which the patient did not get the therapy – to see if we made the right choice such as timing of intervention. Randomization allows us to understand these average causal effects across different treatment strategies. So now let’s look at the prospective, randomized studies in this space.

Randomized, controlled trials have not demonstrated improved survival when patients undergo early tracheostomy versus late tracheostomy. Further, they have established that it appears equally safe (i.e. no differences in ventilator-associated pneumonia (VAP) or tracheal injury) to wait until at least 10-14 days. A large U.K study serves as a representative example of numerous randomized studies on this topic (Young 2013). Termed the TracMan study, this multicenter randomized trial enrolled 909 adults (majority medical ICU admission with pneumonia) who were within the first 4 days of mechanical ventilation and were predicted by their physician to require at least 7 more days of mechanical ventilation. Patients were randomized to “early” tracheostomy (within 4 days) or “late” tracheostomy (after 10 days if still indicated). Approximately 92% of patients in the early arm received trach, whereas 44.9% of patients in the late arm underwent tracheostomy. There was no difference in 30-day or 2-year mortality. ICU length of stay was the same in both groups.

It is important to note that mortality benefit may not be the only concern in the setting of our current pandemic. Given limited resources, outcomes measuring resource use (ICU length of stay, duration of mechanical ventilation) are important additional considerations. Unfortunately, like the TracMan study, nearly all studies to date do not demonstrate reduced resource use with early tracheostomy. However, one randomized trial is worth noting for completeness. A 2010 randomized trial of early tracheostomy (~ 7 days) versus late (~14 days) enrolled 419 patients total, with the primary intention of determining the effect of treatment strategies on ventilator-associated pneumonia (Terragni 2010). Similar to prior studies, more patients in the early tracheostomy group received a tracheostomy (69% vs 56%). VAP rate was 14% in the early group and 21% in the late group (p=0.7 between groups). Ventilator-free days at day 28 were greater in the early tracheostomy group (11 vs 6d). However, there was no difference in hospital length of stay, 1 year mortality, or need for care at a long-term health facility. Importantly, the reduced rates of tracheostomy in both groups complicates interpretation of outcomes.

In conclusion, current evidence does not support a routine early approach to tracheostomy, even if the intention is to reduce healthcare resource use. As data emerge regarding the ventilator trajectories in patients surviving ARDS from SARS-CoV-2, the timing of tracheostomy can be carefully reassessed. However, any practice changes must also thoughtfully weigh the downstream effects on post-acute care needs and the capacity of the healthcare system to support those needs.

View all COVID-19 updates

Learn about research in the Division of Pulmonary and Critical Care Medicine