Laboratory Modeling of SARS-CoV-2 Exposure Reduction Through Physically Distanced Seating in Aircraft Cabins Using Bacteriophage Aerosol

November 2020

Watts L. Dietrich; James S. Bennett, PhD; Byron W. Jones, PhD; Mohammad H. Hosni, PhD

Disclosures

Morbidity and Mortality Weekly Report. 2021;70(16):595-599. 

In This Article

Abstract and Introduction

Introduction

Aircraft can hold large numbers of persons in close proximity for long periods, which can increase the risk for transmission of infectious disease.* Current CDC guidelines recommend against travel for persons who have not been vaccinated against COVID-19, and a January 2021 CDC order requires masking for all persons while on airplanes.†,§ Research suggests that seating proximity on aircraft is associated with increased risk for infection with SARS-CoV-2, the virus that causes COVID-19.[1,2] However, studies quantifying the benefit of specific distancing strategies to prevent transmission, such as keeping aircraft cabin middle seats vacant, are limited. Using bacteriophage MS2 virus as a surrogate for airborne SARS-CoV-2, CDC and Kansas State University (KSU) modeled the relationship between SARS-CoV-2 exposure and aircraft seating proximity, including full occupancy and vacant middle seat occupancy scenarios. Compared with exposures in full occupancy scenarios, relative exposure in vacant middle seat scenarios was reduced by 23% to 57% depending upon the modeling approach. A 23% exposure reduction was observed for a single passenger who was in the same row and two seats away from the SARS-COV-2 source, rather than in an adjacent middle seat. When quantifying exposure reduction to a full 120-passenger cabin rather than to a single person, exposure reductions ranging from 35.0% to 39.4% were predicted. A 57% exposure reduction was observed under the vacant middle seat condition in a scenario involving a three-row section that contained a mix of SARS-CoV-2 sources and other passengers. Based on this laboratory model, a vacant middle seat reduces risk for exposure to SARS-CoV-2 from nearby passengers. These data suggest that increasing physical distance between passengers and lowering passenger density could help reduce potential COVID-19 exposures during air travel. Physical distancing of airplane passengers, including through policies such as middle seat vacancy, could provide additional reductions in SARS-CoV-2 exposure risk.

The study consisted of three components. The first involved analysis of data on virus aerosol dispersion in aircraft cabin mock-ups from a previous study conducted at KSU during July–August 2017 as part of a pandemic influenza research initiative.[3] Next, these data were used to create a regression model to estimate the reduction in aerosol concentration as distance from a source increased. Finally, these regression models were applied to conceptual aircraft seating scenarios to simulate the reduction in exposure resulting from vacant middle seats in an aircraft cabin. Laboratory experiments were performed with bacteriophage MS2 virus obtained from the American Type Culture Collection. Bacteriophage MS2 has frequently been used as a surrogate for pathogenic viruses in aerosolization studies[4] and was used to approximate the airborne dispersion of SARS-CoV-2. During the aerosol dispersion study at KSU, mannequins with realistic passenger heat emission were seated in the cabin mock-ups, and then MS2 aerosol was introduced from a source location and collected at six different sample locations in the cabin. This process was repeated four times: twice in a single-aisle cabin and twice in a twin-aisle cabin (Figure 1), resulting in 24 total samples.** Because these data were collected before the COVID-19 pandemic, the effects of passengers wearing masks on the aerosol dispersion behavior were not measured. These viral aerosol data were then used to create a nonlinear regression model†† which assesses the association between the number of plaque-forming units (PFUs) (evidence of the presence of viable virus) and the distance between source and sample locations. For both single-aisle and twin-aisle scenarios, findings from the nonlinear regression model indicate that the number of PFUs declined exponentially with increasing distance (Supplementary Figure, https://stacks.cdc.gov/view/cdc/104935).

Figure 1.

Diagram of aircraft cabin configurations and source and sampling locations to assess exposure to aerosolized bacteriophage MS2 virus as a surrogate for airborne SARS-CoV-2 exposure in single-aisle and twin-aisle cabins* — Kansas State University, July–August 2017
Source: Modified with permission from Lynch JA, Bennett JS, Jones B, Hosni MH. Viral particle dispersion and viability in commercial aircraft cabins. In: 2018 American Society of Heating, Refrigerating and Air-Conditioning Engineers Annual Conference Proceedings; June 23–27, 2018; Houston, TX; 2018.
Abbreviations: aft = back of the plane; fore = front of the plane; source = aerosol source; X = sampling location.
*For both single-aisle and twin-aisle cabin scenarios, two different source configurations were assessed for placement of infectious passengers: one with the source at the front of the plane in an aisle and one with the source in a seat. Each configuration consisted of six total sampling locations, for a total of 24 samples.
Data were collected at Kansas State University during July–August 2017 as part of a pandemic influenza research initiative.

In November 2020, CDC applied this data-driven model to simulate the protective effect of a vacant middle seat versus full aircraft occupancy. Two analytical approaches were used. Both approaches analyzed reductions in relative exposures (the number of PFUs divided by the maximum predicted value) rather than absolute exposure.

The first approach considered only the extra distance between passengers created by the vacant middle seat. The regression model estimated exposure as a function of distance to assess the exposure reduction of moving an adjacent passenger one seat further away from an infectious passenger, leaving an empty middle seat between them. The distance effect was explored further to simulate the total exposure reduction for groups of passengers up to and including a full simulated cabin of 120 seats.§§ A total of 300 simulations were tested using Monte Carlo methods, where the number (one to three) and placement of infectious passengers were varied. The total exposure reduction for all passengers in the cabin was predicted by placing a source at an arbitrary seat location and applying the regression model to calculate relative exposure at all other seat locations, which were summed to obtain a total exposure for the cabin.

The second approach combined the distance effect predicted by the regression model and the reduced occupancy effect predicted by simple probability estimation, as these are inseparable in realistic arrangements of infectious passengers and other passengers. When simply defining exposure risk as reduced occupancy, a vacant middle seat reduced exposure by an estimated 33% compared with full occupancy, in single-aisle, three-seats-per-side cabins, because there are 33% fewer potentially infectious passengers.

The first approach predicted a 23% exposure reduction by moving an adjacent passenger one seat further away from an infectious passenger. The total reduction in relative exposure for a full 120-seat cabin yielded reduction of 35.0%–36.4%, 35.1%–38.2%, and 35.9%–39.4% for one, two, and three infectious passengers, respectively, depending on their seating pattern. All sources were placed in window or aisle seats such that the potential reduction in number of infectious passengers onboard from vacant middle seating was not considered (Figure 2). The second approach was applied to a cluster of nine infectious passengers (including three in middle seats) among 18 total passengers in three rows (Figure 3). When the infectious and other passengers who would have had middle seats were removed, leaving six infectious passengers out of 12 total passengers remaining in the window and aisle seats, a 57% exposure reduction was observed.

Figure 2.

Estimated reduction in relative exposure to aerosolized bacteriophage MS2 as a surrogate for SARS-CoV-2 through physically distanced seating in a single-aisle, 20-row simulated aircraft cabin* — November 2020
*A total of 300 simulations in which the number (one, two, or three) and placement of infectious passengers varied were tested using Monte Carlo methods. The simulated cabin had a single aisle, with 20 rows and three seats per side (120 total seats); the distance between rows was 3 ft (0.9 m); the distance between adjacent seats was 1.6 ft (0.5 m). In the source configuration shown here, the total reduction in exposure with vacant middle seats was calculated to be 35.4%.

Figure 3.

Relative estimated reduction in exposure to bacteriophage MS2 as a surrogate for SARS-CoV-2 through physically distanced seating in a single-aisle, three-row section of an aircraft cabin with full occupancy* compared with vacant middle seats — November 2020
*A cluster of nine infectious passengers (including three in middle seats) among 18 total passengers in three rows. Removing the infectious and other passengers who would have had middle seats, leaving six infectious passengers out of 12 total passengers remaining in the window and aisle seats, resulted in an estimated 57% reduction in SARS-CoV-2 exposure. Exposures were as follows: 21.1 for six noninfectious passengers in the configuration with no middle seats; 48.7 for the nine noninfectious passengers with full occupancy.
The local prevalence (the percentage of passengers in the three rows who are infectious) was held constant in the comparison to reasonably account for the fact that keeping middle seats vacant is expected to prevent both infectious and noninfectious passengers from sitting there.

*https://www.who.int/whr/2007/en/
https://www.cdc.gov/coronavirus/2019-ncov/travelers/travel-during-covid19.html
§ https://www.cdc.gov/quarantine/pdf/Mask-Order-CDC_GMTF_01-29-21-p.pdf
https://www.atcc.org
**The laboratory cabin environments were a five-row section of an actual Boeing 737 fuselage and an 11-row section of a geometrically accurate mock-up, including original equipment seats, of a Boeing 767 cabin. Ventilation systems used actual Boeing 737 (single aisle) and 767 (twin aisle) equipment where practical, but with no jet engine, so that supply ventilation came from conditioned air in the laboratory rather than from a jet engine. Bacteriophage MS2 virus aerosol was introduced by spray bottle from a source location using three closely timed, consecutive sprays, totaling approximately 3.7 mL. For both single-aisle and twin-aisle cabin scenarios, two different source configurations were assessed: one with the source at the front of the plane and one with the source at the back. Each configuration consisted of six total seat locations; a total of 24 samples were evaluated. Viral aerosol samples were collected in bioaerosol liquid (10% glycerol in 1× phosphate buffered saline) impingers. These devices collected aerosol by drawing air though a small volume of liquid for 5 minutes; samplers were held in place with clamp and stand assemblies on seats, with the impinger intake facing forward to approximate seated head-height for an average adult. Collection fluid aliquots were evaluated for viral particle presence by plaque assay, and the number of plaque-forming units (PFUs) was considered proportional to the airborne concentration references of viable virus.
††The virion data were fitted to exponential regression equations of the form, C = AeBD, with number of PFUs (C) as the dependent variable and the distance between spray and sampling locations (D) as the independent variable. A and B are constants determined by regression analysis, as was the coefficient of determination, R2.
§§This simulated cabin was single-aisle, with 20 rows and three seats per side, for a total of 120 seats; the distance between rows was 3 ft (0.9 m); the distance between adjacent seat centers was 1.6 ft (0.5 m). This number of rows is on the lower end of a typical Boeing 737 coach cabin, depending on the specific design. The total exposure reduction for all passengers in the cabin was predicted by placing a source at an arbitrary seat location and applying the regression model to calculate relative exposure at all other seat locations, which were summed to obtain a total exposure for the cabin.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....