Skip to main content
Premium Trial:

Request an Annual Quote

Breath-Based Diagnostics Gain Traction as Potential Method for Detecting SARS-CoV-2


NEW YORK – As the COVID-19 pandemic has ravaged the globe and strained diagnostic resources, many players have entered or expanded their presence on the scene to interrogate new methods of testing for SARS-CoV-2. 

One of those methods is the use of breath samples for diagnosing the virus, and a variety of companies and research teams from around the world are developing tests that seek to detect SARS-CoV-2 in the breath, although no such test has received Emergency Use Authorization from the US Food and Drug Administration.   

When it comes to breath testing, there are two main ways to measure the volatile organic compounds in a patient's breath, explained Raed Dweik, the chair of Cleveland Clinic's Respiratory Institute. There are thousands of VOCs, which are impacted by a variety of factors, such as the environment and what people consume, Dweik said.

The first way to measure VOCs is a sensor array, sometimes called an electronic nose, that analyzes breath. The electronic nose has at least two but sometimes up to 64 sensors and is simple, portable, and trainable, Dweik said. The downside is that it "doesn't recognize anything it's not trained to recognize," he said. An electronic nose can "guide you down the wrong path" because you don't know specifically what you're looking at, Dweik added.

The other method is using mass spectrometry to "analyze everything in the breath," Dweik said. Mass spec instruments are bulkier and can't be trained, so they can't be used on their own for diagnostics, but they can detect everything in the breath. "You have to do some work behind the scenes, to do some statistical analysis … if you want to make a diagnosis," he said.

Dweik's team at the Cleveland Clinic is using the "best-of-both-worlds approach," combining the methods and using mass spectrometry in conjunction with pattern recognition to determine what Dweik called a "breathprint," allowing the team to diagnose and monitor a variety of ailments, including lung disease, kidney disease, and heart failure. There's also no limit on how much breath can be taken as a sample, unlike a blood test, he added.

The team is ready to start clinical trials to evaluate the technology, which can provide "almost instantaneous" results, takes less than five minutes, and is non-invasive, Dweik said. The protocol for the trial has been finalized, and he said that he and his colleagues expect answers about how the test works before the summer of 2021.

However, there are limitations when developing breath tests for detecting SARS-CoV-2. A major one Dweik noted is the potential for spreading infection during the collection process. The team had to "make sure we do it safely," and avoid infecting more people than they're diagnosing, he said. To solve this issue, the clinic is using disposable collection devices to ensure there's no contamination.

Dweik's team at the clinic is in charge of the sample collection, using a process developed internally to ensure there's no spread of disease when collecting the breath. The interpretation of the results is being done in collaboration with the companies. Although the group has their own mass spectrometry device, it is also working with other companies to use a variety of devices that may be more scalable. The team is testing some of the breath samples across multiple devices to determine accuracy and utility, he added, and continuing to test with mass spectrometry for quality control.

Due to the size of the mass spectrometers being used, Dweik said scaling up using those instruments isn't possible, and they can't be moved to point-of-care settings. However, his team has partnered with a subsidiary of Astrotech, BreathTech, that has smaller devices, about "the size of a desktop printer," that it can use to accurately determine a patient's breathprint, he said. He noted that his team is "technology agnostic" and "not committed or wedded to any single technology," and is in contact with other companies to discuss using their tech. The promise of BreathTech's machine is in its scalability, he added, because having an accurate test doesn't do much if it can't be run on a lot of people.

One potential limiting factor is that the instrument does require some training to be used, although Dweik said anyone "with a reasonable ability" could be trained to run the test and it doesn't have to be a lab technician. The goal is for the device and test to be used at the point of care in physicians' offices and hospitals, he said.

Dweik emphasized that many in the breath testing space are not developing new technology to test for COVID-19, but rather adapting whatever technology they've been working on, whether its mass spectrometry or sensor arrays. "We know that these tests work, we know that they detect other diseases, so they are more likely to work for COVID," he said.

There are three potential outcomes of the clinical trial, he noted: one is that the test can determine whether someone is sick;  that the test can determine someone is sick and that the illness is in their lungs; and the target outcome, which is to determine that someone is sick in their lungs with COVID-19. Even if only the first outcome is achieved, Dweik said it's still helpful to know someone is sick, and knowing that could help increase capacity of swab tests. Breath testing could also be complementary to swab testing and used for screening, he continued. 

Screening options

Meanwhile, at the University of Miami, BioSafety Technologies, a subsidiary of the Israel-based company TeraGroup, has run a validation study testing their COVID-19 breath analyzer on students, staff, and faculty for screening. According to Khemraj Hirani, associate vice chair of research and regulatory compliance for the department of medicine and one of the members of the study team, the university partnered with the company to test its breath analyzer on 1,000 students who were also receiving PCR swab tests.

To conduct the test, a user breathes slowly and deeply into an approximately eight-centimeter tube, called the TeraTube, that they hold three times for 40 seconds. The tube is then closed and put into an alcohol dip for five to 10 minutes, and then into a freestanding scanner that can detect COVID-19 using proprietary technology. The data is provided in less than a minute and can be sent to a user's phone based on a barcode scan, Hirani said. The result is based on the moisture content of the sample, and if there is any suspected infection, the user would go for a confirmatory PCR test, he added. The study aimed to further calibrate the outcome data for increased accuracy, he continued.

The study at the University of Miami was for BioSafety to calibrate the results of its test, refine sensitivity and specificity measurements, and prepare for submission to regulatory agencies, including the US FDA for EUA, Hirani said. The company has tested its analyzer at other sites globally and "data are promising," Hirani said. According to the company's website, it could run up to 2,000 tests per day with the scanner and analyzer.

Hirani noted that students were intrigued by the non-invasiveness of the test, as well as the convenience and rapid result turnaround, especially in comparison to the "discomfort" of PCR tests. The "slow and deep breathing" is easier for patients and that ease led to more engagement from the faculty, staff, and students and increased their willingness to participate in the study.

Although the test hasn't received clearance from any major regulatory agencies, Hirani said that many organizations, including the university use the test if it's approved, likely in conjunction with PCR testing and for surveillance.

A variety of other companies are also working on breath tests for COVID-19, including Imec, which uses chips to collect aerosols and droplets and then uses PCR, Rapid Biosensor, which has primarily focused its development on rapid testing for tuberculosis, and Canary Health Technologies, which is working to detect the metabolic response to the virus.

Rapid Biosensor is working to adapt the technology it uses for its TB test for COVID-19, said CEO Dennis Camilleri. The TB test is an antigen test that detects the presence of TB using an evanescent wave immunosensor in a disposable tube that is coated with a bioassay, he said. The tube is then inserted into a non-disposable optical bio reader that has laser optics, which produces a change in fluorescent signal if tuberculosis is present. The process takes approximately two minutes to determine negativity or positivity, Camilleri said. 

He said he is confident the company can adapt this process for COVID-19, although he noted the viral loads for SARS-CoV-2 are much smaller than TB and added the company would have to develop a new bioassay for the virus. The biosensor, tube, and reader are patented in the US for breath and other fluids, but not cleared by the FDA, and a COVID-19 test would use the same battery-operated reader. Conceivably, he said, a user could have one reader with a different number of tubes testing for tuberculosis, COVID-19, and pneumonia, for example.

Camilleri said the company is actively pursuing funding to develop a COVID-19 test, but said the main hurdle is getting the public to accept that breath testing and analysis "can actually work," because it is still a relatively young technology. But the benefits of Rapid Biosensor's instrument, namely the rapid results, low cost, ease of use, and lack of electricity required, could allow widespread testing, especially in countries with limited access to centralized laboratories. 

In contrast to tests directly detecting the presence of SARS-CoV-2, Canary Health's test is looking for the metabolic response to the virus – namely the "unique VOC signature that correlates to COVID," said Anna Wang, senior vice president of corporate affairs and strategic partnerships. The company claims its test can detect the infection earlier than other tests, possibly as soon as one to two days after infection, since there is no need to detect the viral load. "We believe our sensors are sensitive enough to detect that before the onset of symptoms," Wang said. 

Canary's nanosensor detects breath biomarkers using gas sensors that convert the VOC gas molecules to digital signal patterns and translates the electronic signatures of the biomarkers to an AI analytical "lab in the cloud," which then uses pattern recognition to try and match the signatures to the COVID-19-positive breath samples the software has been trained with.

The disposable sensor with a blow tube plugs into the reusable device hub, which Wang also called a "handheld station." Users breathe in for one minute, sending signals to the cloud, and the operator of the station presses a button for the result, which says the sample is positive or negative for the virus. The analysis depends on the speed of the location's Wi-Fi, Wang said, but the entire testing process takes about three minutes. The sensor has a unique ID for each person and the results can be linked to someone's phone or computer, Wang said.

The device can be used up to 100,000 times and the disposable sensor can be customized, she added. The company originally began as a cancer detection firm for lung cancer and was going into pivotal trials for the device but pivoted to COVID-19 and changed its prototype once the pandemic began, she said.

The expectation for the cost of each disposable tube is about $30, and the instrument will likely cost $1,500, which Wang said was "nominal" considering how many times it can be used. The AI is being trained and the device is currently in validation trials for COVID-19 in India, using 750 subjects. Wang said the company expects the trial to end this month but needs 250 positive cases to properly train the system and added that the sensitivity is believed to be above 90 percent.

The Cleveland-based firm is also looking to start a trial in the US, likely in Las Vegas, where it will test around 3,000 subjects. The US trial will be used to submit to the FDA for EUA, which the company hopes to do in March. The company has begun manufacturing the device and allotted itself three months to manufacture and validate the test, with a plan to put the product on the market in April, Wang said.

There is also potential for the device and tubes to be used for screening at places like airports and resorts, she said, or if an airline requires either a vaccine or a negative test to board a plane. "We designed the whole platform as point of care, there's no lab, you don't need to send a sample," Wang said. "It's non-invasive, easy, ultra-fast, and provides early detection." The test could also be good for repeat testing and could be performed every day if necessary, she added.

Canary's test is unique because it doesn’t require mass spectrometry like many other breath tests in development and the disposable units do away with cross-contamination risks, Wang said, although many other companies are adapting to disposable collection devices for COVID-19 testing.

Hurdles to overcome

All of these tests could change the way COVID-19 testing functions across the world and could help alleviate the strain on labs and test manufacturers. But Chris Claxton, head of investor relations for Owlstone Medical, and Marc van der Schee, the company's head of clinical, aren't sure the tests can live up to the promises being made. Owlstone Medical, based in Cambridge, United Kingdom, is a breath-based diagnostic firm developing tests for a variety of diseases, including cancer and liver disease.

For purely sensor-based technology that relies on pattern recognition, van der Schee said, there are a lot of attractive qualities – it's small, portable, and cheap. However, the downside is that you don't actually know what you're measuring, and false positives may show up from everyday tasks like brushing your teeth or eating certain foods. The approach of many companies to these potential false positives has been to ignore them, he added, which is risky because the population being studied for validation trials isn't the same population who will be using the tests in practice. In use in the general population, the majority of people would have no symptoms. 

For tests using analytical chemistry and mass spectrometry to build a library, van der Schee said the dataset would be richer because the exact chemical compound in breath would be measured, so users would know exactly what they're looking at. However, the library of biomarkers would have to be built carefully, and mass spectrometry is neither cheap nor fast. He said if biomarkers related to COVID-19 were identified, a company could make a sensor detecting those compounds in particular. Claxton noted that identifying the biomarkers is a major challenge, since COVID-19 symptoms overlap with many other respiratory infectious diseases.

Companies would also have to determine if they're looking to detect a general infection or COVID-19 specifically – it would be easier to detect any kind of infection but detecting COVID-19 would require detecting the structural part of the virus, Claxton said. Someone could detect the consequences of the virus, but only the response of the immune system could be detected, he added. If a patient is positive but asymptomatic, the test could be negative.

Owlstone Medical currently isn't working on a test for COVID-19, despite being a leader in the breath-testing space, but van der Schee said one option for breath-testing could be using it to determine the severity of an infection and measure how the body responds. A rule-in test for the presence of COVID-19 "is probably best done by looking for its genomic signature through techniques like PCR," van der Schee said. "A breath test would look at more aspecific responses of the body to an infection, meaning it would detect a lot of general respiratory infections" that aren't COVID-19, which could clog up the system, he added.

He said many organizations and companies have overstated their claims and that he's skeptical of the promises, since the virus doesn't generate its own specific metabolites. Despite the virus likely being aspecificic, "it can still be valuable to look at the metabolites our bodies generate in response to the virus to look at severity," van der Schee said. "That's an area where genomic tests don't have strong performance." 

Instead, Owlstone has focused on optimizing sample collection to understand the risk of spreading the virus via exhaled aerosols and droplets, and if sensitive enough this could lead to a "valuable at-home test," van der Schee said.

Van der Schee noted that breath analysis technology needs to mature further, and he'd be surprised if these companies are able to create a working, reliable test in a reasonable time frame. In addition, if the tests didn't work, overall confidence in breath testing would drop, he said. 

Many companies are also focusing on accuracy, but van der Schee emphasized the question of utility. "Will the outcome be better? Will it help stop the spread and improve treatment?" he asked.