Clearing the Way for Early Detection of Ventilator-Assisted Pneumonia

Hospital patient on ventilator
Article Summary
  • Miller School researchers investigated volatile organic compounds found on ventilator filters to predict the early onset of ventilator-associated pneumonia.
  • The team found that an elevated presence of carbon disulfide could serve as an early biomarker for pneumonia.
  • The work grew from a previous project with BMW, which asked Dr. Carl Schulman and team to measure exhaled breath to assess driver alertness.

The last thing patients on respirators need is to catch pneumonia from the very device that’s helping keep them alive.

Despite everyone’s best intentions, it happens all too frequently. In the United States, ventilator-associated pneumonia (VAP) is a major worry in health care settings, especially in intensive care units. As many as 40 percent of patients on mechanical ventilation develop the condition. There are approximately 150,000 to 300,000 cases annually. The death rate ranges from 30 to 70 percent.

But in a first-of-its-kind study, University of Miami Miller School of Medicine researchers investigated the use of volatile organic compounds (VOCs) found on heat and moisture exchange ventilator filters to predict the early onset of VAP. VOCs are chemicals that vaporize into air and dissolve in water. They are found in thousands of everyday products, including paint. The body produces them when exhaling.

Dr. Carl Schulman in white clinic coat
Dr. Carl Schulman

The research team, led by senior author Carl Schulman, M.D., Ph.D., MSPH, professor of surgery at the Miller School’s DeWitt Daughtry Family Department of Surgery and the school’s senior associate dean for research operations, found that an elevated presence of a VOC called carbon disulfide could serve as an early biomarker for VAP and allow for faster interventions and use of antibiotics to prevent a lethal outcome.

The discovery has been recently published in the Annals of Surgery.

Ventilator-associated Pneumonia

Ventilator-associated pneumonia is pneumonia occurring more than 48 hours after mechanical ventilation first occurs. It poses a significant threat to patients unable to breathe on their own, increases the chances of a patient dying and is financially burdensome, as well. The cost can soar to $50,000 to $60,000 per patient, accounting for extra days in the ICU, antibiotic treatments and additional tests. Current diagnostic methods are often delayed, invasive or overly subjective, frequently leading to unnecessary antibiotic use and worsening outcomes.

Patients on ventilators are prone to VAP for several reasons:

• Natural defense mechanisms, such as coughing and mucociliary clearance, are often impaired.

• The ventilator’s endotracheal tube provides a direct pathway for bacteria to reach the lungs.

• Compromised patient immune systems are more susceptible to pneumonia.

• Patients on ventilators may aspirate (inhale) their own secretions due to decreased consciousness and impaired swallowing reflexes.

• Ventilator patients are surrounded by patients who may contribute pathogens, many of which resist antibiotics.

The study used patented technology to analyze exhaled VOCs collected on heat and moisture exchange filters in ventilators.

Early Detection of Respiratory Infections

“Since exhaled breath can be collected continuously and without invasive procedures, it presents a promising avenue for detecting respiratory infections such as VAP early on,” said Dr. Schulman. “Our hypothesis was that VOCs from exhaled breath collected on the filters could serve as predictive biomarkers for VAP, providing a faster and more objective diagnostic tool. The earlier the condition is identified, the better the potential outcome.”

Air that is exhaled by patients on mechanical ventilation passes through a heat and moisture exchange filter. As part of the process, the filter captures a variety of exhaled aerosols. Over time, the filters become saturated with exhaled breath condensate and aerosols. They are typically exchanged on the ventilator circuit at least once per day.

“The filter serves as a makeshift trap for compounds of interest that can be studied to potentially diagnose respiratory disease or infection,” said Dr. Schulman.

Dr. Sylvia Daunert in white clinic coat
Dr. Sylvia Daunert
Dr. Sapna Deo in white clinic coat
Dr. Sapna Deo

Using mass spectrometry techniques, Umer Bakali, Ph.D., who was mentored by Sylvia Daunert, Pharm.D., M.S., Ph.D., the Lucille P. Markey Chair of Biochemistry and Molecular Biology, and Sapna Deo, Ph.D., professor of biochemistry and molecular biology and the director of the Molecular Medicine Pathway of Biochemistry and Molecular Biology, analyzed the filters from the ventilators used by 12 patients for the study. Carbon disulfide, associated with gram-negative bacterial growth, was found in the exhaled breath of 11 patients within three days before clinical suspicion of VAP, suggesting its potential as a biomarker for early detection.

“Our study shows that when you first detect carbon disulfide on a ventilator filter, there is an excellent chance that the patient will soon be manifesting symptoms of pneumonia,” said Dr. Schulman. “This offers us a leg up on giving the patient the right antibiotic at the right time, before the disease process accelerates and conditions worsen.”

First-of-its-Kind Study

The Miller School study is the first to assess filters for detecting VOCs in ventilated patients’ breath. Previous trials collected air samples from patients in tubes attached to ventilators.

“The tube-based collection method is generally used for a short period of time, typically five to 10 minutes,” said Dr. Schulman. “We used filters that were in place for an entire day, allowing sampling over an extended period of time, which gives a broader snapshot of what’s occurring with the patient.”

Another benefit of the study is specificity.

“Other research often examines a panel of VOCs, trying to see how the components change over time,” said Dr. Schulman. “The problem is that many of them are present either in the background or in other disease states, making them less specific for VAP. We found through our research that carbon disulfide is not generally present in other circumstances, giving it a clear advantage as a specific biomarker for VAP.”

Our study shows that when you first detect carbon disulfide on a ventilator filter, there is an excellent chance that the patient will soon be manifesting symptoms of pneumonia. This offers us a leg up on giving the patient the right antibiotic at the right time.
—Dr. Carl Schulman

A third plus is greater efficiency. Carbon disulfide has already been identified as a marker of pneumonia in certain respiratory infections and cystic fibrosis. This allowed the researchers to home in on a known pneumonia correlate, saving the time and expense of looking at all compounds on the filter, including those with no previous association with pneumonia.

Finally, the 24-hour sampling period captures a higher concentration of compounds, many of which may take time to accumulate.

From BMWs to Pneumonia

Dr. Schulman and team did not start the project looking for a way to detect pneumonia. The story begins about a decade ago with automaker BMW Group, with whom they conduct breath-centered research for improving vehicle design and safety. At the time, the team used solid-state sensors to determine if a car occupant was drowsy or fatigued.

The patent for pneumonia-biomarker detection is based on that early work.

“As that project continued, it occurred to us that, since we have this great system to look for things in people’s breath, let’s try and apply it to the intensive care unit and look for actual diseases, as opposed to just fatigue in your car,” Dr. Schulman said.

He will soon travel to Korea to deliver a presentation on his vehicle-based work at an international conference.

The University of Miami Office of Technology Transfer is teaming with Dr. Schulman and team to identify potential partners to license the technology. The aim is to create a new solid-state sensor, or customize an existing one, to attach to a ventilator dashboard for 24/7 data collection. Dr. Daunert and team can identify biomarkers from breath and create or adapt a sensor to measure carbon disulfide.

But it’s costly and time-intensive to approve a new medical device. Licensing the technology to a third party will help make that happen. The Miami model for detecting specific, volatile organic compounds could also be applied to other serious medical conditions, including acute kidney injury and bloodstream infection in mechanically ventilated patients, two potential applications that the research team is examining.

Findings could be applied to both ventilated and non-ventilated patients. In all cases, the aim would be to identify the condition early and intervene with medications, such as antibiotics, or other measures to prevent the progression of the disease and corresponding harm.


Tags: Department of Biochemistry and Molecular Biology, DeWitt Daughtry Family Department of Surgery, Dr. Carl Schulman, Dr. Sapna Deo, Dr. Sylvia Daunert