Can CO2 sensors be used to assess COVID-19 transmission risk?

Ventilation and air cleaning are now recognized as important tools for mitigating the risk of COVID-19 indoors. However, many public spaces are either inadequately ventilated, or their ventilation status may be unknown to the users. Although a number of ventilation calculators have been developed to help assess ventilation needs, these tools still require a certain amount of data and technical knowledge to implement.

In addition, the number of people and the nature of their activities within a space can affect air quality and the presence of infectious particles over time. Measuring indoor carbon dioxide (CO2) concentrations using inexpensive digital sensors (typically less than $500) has been proposed as one way to estimate indoor COVID-19 transmission risk, and in fact some jurisdictions have already given directives or guidance regarding the use of CO₂ sensors in settings such as schools. This blog looks at some of the evidence for the use of CO₂ sensors in assessing transmission risk and identifies potential pitfalls for further consideration.

What is the relationship between occupancy and CO₂ levels?

CO₂ monitoring is an established means to assess whether ventilation is adequate for the number of people occupying the space. Although CO₂ levels range from 350 to 450 ppm in the great outdoors, people gathering and respiring inside a building will cause CO₂ to accumulate to much higher levels unless removed through ventilation. The more people that occupy a space and the more intense their physical activity, the more ventilation is required to maintain occupant comfort. CO₂ levels are also a proxy for other difficult-to-detect pollutants (e.g., volatile organic compounds off-gassing from building materials) that may accumulate in the space. Finally, CO₂ sensors can be used to increase energy efficiency, by allowing users to regulate ventilation around a CO₂ setpoint (e.g., 1000 ppm), such that enough fresh air is taken into the building to maintain occupant comfort, but not so much as to waste excessive energy on heating and cooling (i.e., demand control ventilation).

“Comfort”, as mentioned above, is typically defined by the occupants’ ability to perceive “human bioeffluents,” otherwise known as body odours, as well as their thermal comfort. As indoor CO₂ levels rise, a greater proportion of occupants will express dissatisfaction with odours and “stuffiness” if no other air cleaning device is used (e.g., filters that eliminate odours). Many IAQ professionals have adopted a value of 1000 ppm CO₂ as a guideline for acceptable indoor air quality, although this value is not a formal standard (though it is often claimed to be) and the value itself has no particular health significance.

What do CO₂ levels mean for health more generally?

Although popular media and even the scientific literature often assert that indoor CO₂ levels have a negative impact on perceived air quality, cognition, and work performance, the literature examining these relationships is highly inconsistent. A recent review of 10 studies found no clear dose–response relationship between CO₂ levels and cognition, performance, or perceived air quality, and in fact some studies have shown enhanced performance at higher CO₂ levels. Interestingly, Zhang et al. found that exposure to high levels of CO₂ containing human bioeffluents was associated with cognitive impairment and poor perceived air quality, but pure CO₂ at the same level was not. This may suggest that some other factor related to human occupation may be responsible for the ill effects of a “stuffy” indoor environment.

Although we still do not fully understand the relationship between CO₂ levels and human health, it is logical that CO₂ and other indoor air pollutants should not be allowed to accumulate in indoor spaces. CO₂ sensors can be a useful tool for assessing how “fresh” the air is relative to the outdoors, allowing occupants to keep the air as fresh as possible without sacrificing thermal comfort.

Is there a reliable basis for relating CO₂ levels and infectious disease transmission?

Because CO₂ levels increase with the number of occupants in a space, and one or more of those occupants could potentially be infected, it seems logical that respiratory disease risk should scale upwards with indoor CO₂ levels.  In 2003, Rudnick and Milton introduced a model that used CO₂ levels to estimate the risk of airborne transmission of influenza and others respiratory diseases. Since then, numerous studies have examined the relationship between ventilation rates, estimated using CO₂ levels, and disease transmission in various settings, such as school dorms. During the COVID-19 pandemic, researchers have built on this work to estimate SARS-CoV-2 transmission risk in spaces such as nail salons and office buildings, based on occupants’ respiration rates and activities.  

These studies show the value of CO₂ sensing for assessing ventilation as a factor in transmission risk; however, establishing a CO₂ threshold that can reliably signal elevated COVID-19 risk is more complex. As described by Peng and Jimenez, occupants’ activity levels and other factors, like the use of masks, make selecting a single threshold problematic; rather, the authors suggest an activity-based threshold or thresholds. Similarly, a recent white paper from the Environmental Modelling Group of the UK’s Scientific Advisory for Emergencies (SAGE-EMG) identified a number of factors that might affect CO₂ levels but not COVID-19 risk, and vice versa. For example, the presence of pets or combustion devices will increase CO₂ levels, but will not affect COVID-19 risk. Conversely, COVID-19 transmission risk may be enhanced by activities like singing or the presence of a rare super-emitter, which are unlikely to affect CO₂ levels to the same degree.

The SAGE-EMG document also noted that the use of CO₂ as an indicator of ventilation effectiveness would be particularly uncertain in spaces with few occupants, where varying individual contributions would have a greater overall effect on CO₂ fluctuations, and that CO₂ monitoring would not reflect decreases in transmission risk due to measures like air cleaning devices or the use of masks. Overall, the working group concluded that although CO₂ levels are useful to indicate whether enough outside air is being brought into the space, CO₂ levels should not be relied upon to determine whether a space is ventilated enough to mitigate transmission risk.

Challenges in using CO₂ sensors to identify “poorly ventilated” spaces

Although CO₂ levels may be problematic in terms of assessing transmission risk, there is value in using CO₂ levels to identify poorly ventilated spaces or, in the long-term, to prioritize them for remediation. But at what CO₂ level should corrective action be taken, and what should be done when ventilation is inadequate?

In Europe, a simplified “traffic light” scheme is being recommended to help occupants determine whether ventilation in a space is adequate, although the thresholds chosen to represent “green,” “amber”, and “red” levels of CO₂ are somewhat arbitrary. In German schools, CO₂ levels < 1000 ppm are given the green light, levels from 1000 from 2000 ppm are signaled with an amber light, and levels > 2000 ppm are signaled with a red light (REHVA recommends thresholds of 800 pm for amber and 1000 ppm for red). Notably however, these thresholds are not being used to trigger specific actions (e.g., open window at 1000 ppm, evacuate at 2000 ppm). Rather, it is expected that occupants (teachers and students) will be mindful of the sensor throughout the day and try to keep indoor CO₂ levels as low as possible by periodically opening the windows. Abuhegazy et al. found that opening windows was remarkably effective in lowering the concentration of particles <1 um over 15 min in a modelled classroom full of students. If windows cannot be opened, it may be possible to install other ventilation equipment (e.g., exhaust fans) as necessary. Air cleaners may also reduce transmission risk when ventilation is inadequate, but because air cleaners do not remove CO₂, any improvement in air quality would not be apparent via the CO₂ sensor.

Does the use of CO₂ sensors create a technological dependency?

The use of a technological tool like a sensor can also create technological dependency, in which the focus of attention becomes the readout on the sensor, rather than promoting the overall situational awareness that is critical to decreasing transmission risk. COVID-19 transmission risk is dependent on a number of factors, not only poor ventilation as signaled by high CO₂ levels. In fact, ventilation (as the remedy for high CO₂ levels) does nothing to mitigate the much greater risk of close-range interactions with an infected person while one or both people are unmasked. Thus, if CO₂ sensors are to be deployed, it is important to consider how this might be achieved without skewing or narrowing the users’ perception of overall COVID-19 risk.

For example, in the case of German schools, the CO₂ traffic lights described above are not intended as a permanent fixture, but rather as a training device to help occupants learn when to periodically open windows to air out the classroom. Once the habit is entrenched, the CO₂ traffic light should be moved to another classroom. This training tool is useful as it teaches occupants to be continuously aware of indoor air quality; however, the traffic light system is also somewhat misinformative, as the color signals will inevitably be interpreted as an indicator of health risk (e.g., green is good, amber warning light, red means danger). As identified here, there are a number of instances in which the CO₂ traffic light might be green, but COVID-19 transmission risk is high.  Thus, if devices such as CO₂ traffic lights are to be used, they should be implemented while emphasizing the importance of overall risk awareness and the limitations of the technology.

Summary

Overall, the relationship between CO₂ levels in an occupied space and the risk of COVID-19 transmission is weak due to the various factors that can influence each independently. Even so, the low cost of implementing CO₂ monitoring and the use of strategic action levels to trigger ventilation can help to improve indoor air quality overall. It can also help to educate occupants about actively managing indoor air quality. However, although digital CO₂ sensors are easy to read, they are also easy to misinterpret, and jurisdictions considering this technology should also consider how its use affects occupants’ understanding of and actions around COVID-19 transmission risk.