Risk and mitigation of aerosolisation from lung function testing: results from the AERATOR study

Ethics

This study was performed as part of the wider AERATOR study to assess the risk of aerosolised transmission of SARS-CoV-2 in healthcare settings. Ethical approval was given by the North West Research Ethics Committee (Ref: 20/NW/0393, HRA Approved 18/9/20).

Environmental set up, recruitment, and procedures

Full technical methods are in the supplementary appendix (S1), and have been reported elsewhere.⁴ In brief, healthy volunteers were recruited to allow collection of aerosol emission in ultra-clean, laminar flow operating theatres. Participants underwent a protocolised set of testing (breathing, speaking, and coughing) followed by lung function testing designed to represent standard clinical practice. Time stamps were used to indicate timing and duration of tidal breathing and forced manoeuvres including formal spirometry (as per ATS/ERS guideline(14)), peak flow measurement, and FENO measurement. All testing was performed by an accredited lung function technician using The spirometers systems used single use bacterial viral filters (BVF), peak flow meters were tested both with and without BVF (see supplement). A proof-of-concept test of mitigation strategy to reduce the risk of procedure induced cough aerosolisation was undertaken using a re-usable full face cardiopulmonary exercise facemask with BVF in 3 subjects. The sampling funnel was positioned over the point of greatest exhalation flow.

Measurements of aerosol were taken simultaneously by two separate devices, the Aerodynamic Particle Sizer (APS), and Optical Particle Sizer (OPS), both manufactured by TSI Ltd), via a 3D printed funnel and through 0.45m of non-conductive tubing. Measurements of aerosol were taken at the exit port of spirometer (for spirometry), the end of the peak flow device (for peak flow), and at various points around the device (for FENO), within 15cm of the source of aerosol..

For brevity, we report only the APS figures as we have previously shown correlation between the devices to be excellent (r> 0.95) (15).

Demographics

33 healthy volunteers were recruited; with demographics and lung function results in Table 1. They were young (median age 32), of normal weight (median BMI 23.6) and had normal lung function; as would be expected from this cohort. 16 (48%) were male, with 17 being female (52%).

Spirometry

Table 2 describes the geometric mean and standard deviation for FEV1, peak flow, and FENO, in comparison with speaking, breathing and coughing, with Figure 1 describing this on a boxplot. As all of those activities apart from breathing and speaking are sporadic activities, we report the max particles per cm³ per sample. Of all the activities, cough produces by far the most, with an average of 1.61 particles/cm³/sample, with unfiltered peak flow approximately half (0.76 particles/cm³/sample). For brevity, we only discuss the APS results as correlation between the devices is so high.

• Download figure
• Open in new tab

Figure 1:

Concentrations of aerosol detected from spirometric activities (a: APS, b: OPS)

Usage of a filter on the peak flow device reduced aerosol emission to a peak emission comparable with the average emission produced while speaking (0.09 particles /cm³/sample vs 0.104 particles/cm³/ sample). Performing an FEV1 had similar peak aerosol emission (0.11 particles /cm³/ sample).

The FENO device does not have a clear exhalation port and the manufacturer does not make clear where exhaled breath leaves the device. We interrogated the device and measured aerosol at all possible exhalation ports (see image in supplementary appendix). Regardless, we did not find any significant aerosol emission from the FENO device.

Usage of a CPET mask to reduce aerosol emission was tested on 3 participants, with large reductions in aerosol emission during breathing, speaking and coughing (results in Table 2). This mask can be worn before, during, and after spirometry and may be a potential mitigation technique as induced coughing is common during lung function.

In summary, formal lung function testing using a spirometer with a filter did not generate significant aerosol in comparison to a cough; with peak levels similar to average emissions during speaking.

Peak flow with a filter

To ensure peak flow results were reliable with a filter, we compared peak flow readings both with and without the protective filter for 9 participants. This is shown in figure 2, showing strong correlation with a R of 0.966. Therefore, although there may be a slight reduction in recorded FEV1 with the peak flow monitor when used with a filter, this is likely clinically insignificant.

• Download figure
• Open in new tab

Figure 2:

Correlation between unfiltered and filtered peak flow

Aerosol emission and relation to spirometric function

Finally, we investigated the relationship between predicted and total FEV1 and FVC with aerosol emission during speaking, breathing and coughing, and found no clear relationship (Figure 3, r <0.4 for all comparisons.)

• Download figure
• Open in new tab

Figure 3:

FVC and FEV1 compared to aerosol emission when breathing, speaking and coughing.

Strengths and weaknesses

This study has many strengths. Firstly, we performed spirometry as per standard UK practice in an ultraclean, laminar flow theatre with very little background aerosol concentration. This allows us to make confident conclusions about the source and level of aerosol produced. Secondly, we were able to test mitigation strategies allowing for simple practical guidance on risks. Thirdly, we used two simultaneous measurement techniques (OPS and APS), that ensured robustness of measured results. Finally, we were able to compare to a known reference (cough), as a comparison for a procedure that does generate significant aerosol.

Common to all volunteer studies, we are limited in that our volunteers have largely normal spirometry. Reassuringly, however, we found no relationship between aerosol emission and FEV1 or FVC, and there is no evidence available from the current literature to suggest that patients with lung disease are more likely to generate aerosol. The AERATOR study will include further aerosol emission data from patients with underlying lung disease and increase the number of healthy volunteers wearing CPET mask in the near future which will improve sample size and address this question. Also, none of the volunteers were known to be suffering from viral infections which have been shown to increase aerosol production (16, 17). However, while the available aerosol data from patients with active COVID-19 is limited they do suggest comparable emissions to healthy volunteers (15). Finally, as activities such as coughing or peak flow are forceful and short lived, these are observed as a rapid rise in the reported number of aerosol concentration, then a decay away over a few sample measurements (typically equivalent to 10-15 s for a cough) as the aerosol dissipates from the sampling funnel and is diluted by the clean room air. This means the comparison between these forceful events and more continuous events (such as breathing, or speaking) in total aerosol emission is difficult.

Comparisons with previous literature

There is a lack of high-quality experimental data to reconcile current conflicting guidelines. We identified three recent papers that have attempted to address the issue. Helgeson et al. undertook a small-scale study that identified aerosol generation during spirometry performed by five healthy volunteers (18). It detected aerosol production despite the inclusion of viral filters, though this was not evident when samples 1.5 foot from the exhalation port. It was performed in a room with a good ventilation (12.8 air changes/hr (ACH)). The sampling method used in this study did not detect any increase in aerosol production with speaking bringing into questions the sensitivity of the methodology. Furthermore, data was discarded if a participant coughed. We consider this an important and relevant reference for other tested AGPs. Lastly, their small sample size limits generalisability.

The second study (19) included 33 healthy volunteers who performed different respiratory manoeuvres outside of ultra clean, environment. Like our data they found coughing was the highest aerosol-producing manoeuvre and interestingly they found that increased expiratory flow rates did not increase aerosol production. The study was performed without viral filters and thus does not provide data relevant to this simple mitigation factor. Furthermore, the authors acknowledge that their PFTs did not meet ERS/ATS spirometry criteria, and 22 of the volunteers performed nothing more than tidal breathing.

The third paper (20) analysed aerosol presence whilst a variety of pulmonary function tests (PFTs) were performed by 28 patients with viral filters in rooms with three different ACH rates. Just one measure of aerosol (OPS) was performed during PFTs and for 30-60 minutes following the procedure after the patients and researchers had left the room. The paper concludes that PFTs should be considered an AGP, however on examination of the presented data, during the PFT procedures, aerosol levels remain steady or fall. The peak aerosol production occurs when the patients move away from the apparatus and are breathing, speaking and coughing without a mask on. There are no data presented to show baseline aerosol levels with patients in the room without masks on, and thus we can only conclude that the data shows the relatively low level of aerosol production during PFTs and re-emphasises the importance of mask-wearing and social distancing during normal interactions. It also highlights the potential benefits of the use of measures such as a CPET mask in preventing aerosol production in the time between procedural manoeuvres.

Significance of this study

This study has demonstrated that peak flow measurement and spirometry performed with viral filters, and standard FENO do not promote significant aerosol production. It supports their removal from the designated list of AGPs. This should allow community and hospital-based lung physiology teams to increased capacity for these essential respiratory diagnostics by reviewing the necessity for restrictive measures such as community PCR testing, patient self-isolation prior to attendance, the use of ‘full PPE’ during procedures, and prolonged fallow periods during which time further procedures cannot be performed.