To the Editor—The Centers for Disease Control and Prevention (CDC) recently found that adults with positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) test results were twice as likely to report dining at a restaurant than those with negative SARS-CoV-2 test results.^{Reference Fisher, Tenforde and Feldstein1} Sharing of meals in staff break rooms has also been reported as a source of exposure in healthcare personnel with coronavirus disease 2019 (COVID-19).^{Reference Zabarsky, Bhullar and Silva2} Dining in close proximity to others may be a high-risk activity because masks are removed while eating.^{Reference Fisher, Tenforde and Feldstein1–Reference Lentz, Colt and Chen4} Air-conditioned ventilation in restaurants might also facilitate transmission of small aerosolized droplets to distances >1 m.^{Reference Lu, Gu and Li5}

In addition to respiratory droplets, transmission of respiratory viruses may occur due to contact with contaminated surfaces and fomites.^{Reference Doremalen, Bushmaker and Morris6–Reference Boone and Gerba8} To reduce the risk of fomite-mediated transmission, the CDC recommends that restaurants and bars clean shared objects between each use and avoid sharing items such as menus and condiments.³ The effectiveness of these measures in preventing transmission of pathogens is unclear. Here, we evaluated an intervention to reduce person-to-person transmission from contaminated fomites in restaurants.

The study protocol was approved by the Institutional Review Board of the Louis Stokes Cleveland VA Medical Center. We conducted a quasi-experimental study of transmission by fomites contaminated with bacteriophage MS2 in a simulated restaurant setting. Bacteriophage MS2 was propagated in Escherichia coli.^{Reference Tomas, Kundrapu and Thota9} The simulated restaurant consisted of a ~8 m × 8 m room with tables positioned 2 m apart. The study was divided into control simulations and intervention simulations. Research and healthcare personnel participated as customers and wait staff in the simulated restaurant.

Prior to each simulation, an index customer’s hands were contaminated with the benign bacteriophage MS2 by applying a 0.5 mL solution containing the virus. Simulations were conducted with a higher inoculum (10⁶ plaque-forming units [PFU]) intended to mimic a worst-case scenario and a lower inoculum (10³ PFU) intended to be more typical of real-world contamination levels. Three simulations (9 customers) were conducted with the higher inoculum and 4 (12 customers) with the lower inoculum.

For each simulation, wait staff invited the index customer to take a seat and provided a laminated menu. After the menu was reviewed, the wait staff took it to a central counter and returned with a bill. After taking a credit card, the wait staff returned with the final bill and a pen that was used to sign the bill. The index customer’s hands and the table and chair, menu, pen and wait staff’s hands were sampled using premoistened CultureSwabs (Becton Dickinson, San Jose, CA). The swabs were processed as previously described for quantitative culture of virus particles.^{Reference Tomas, Kundrapu and Thota9}

A second customer with no bacteriophage MS2 applied was seated at a second table and the same procedures were followed. The menu and pen were not cleaned between customers. Finally, a third customer was seated at the third table and the same procedures were followed. Wait staff did not perform hand hygiene between customers.

Three simulations (9 customers) with an intervention were conducted using the higher bacteriophage MS2 inoculum. The protocol was identical to the protocol for the initial simulations except disposable paper menus were used, credit cards were inserted into a card reader such that the card was not contacted by the wait staff, the pen was disinfected with a disinfectant wipe after each use, and wait staff used alcohol hand sanitizer between tables. Wait staff did not touch the used menus or non-decontaminated pens. The used pen was picked up using a disinfectant wipe and used menus were placed into a designated receptacle. The index customer’s table and chair were cleaned with an improved hydrogen peroxide disinfectant wipe, and cultures were collected as previously described to assess the effectiveness of disinfection.

The primary outcome tested was the log₁₀PFU bacteriophage MS2 concentrations for customers 2 and 3 for the higher inoculum of MS2 in the control simulations versus the intervention simulation. The Student t test was used to compare the concentrations with versus without the intervention.

Figures 1A and 1B show the log₁₀PFU of bacteriophage MS2 transferred from the index customer’s hands to the environment and hands of wait staff and subsequent customers with no intervention. Figure 1C shows the effectiveness of the intervention in reducing transfer of bacteriophage MS2. For the index customer, the virus was recovered from the table and chair, menu, and pen. After cleaning and disinfection, no bacteriophage MS2 was recovered from the table and chair. For customers 2 and 3, no transfer of bacteriophage MS2 was detected on environmental surfaces or hands (P < .0001 in comparison to transfer without the intervention).

Fig. 1. Transmission of bacteriophage MS2 via fomites in a simulated restaurant setting with a lower (A) and higher (B) inoculum on the index customer’s hands and effectiveness of an intervention in reducing transmission with the higher inoculum (C).

In summary, we found that bacteriophage MS2 inoculated on the hands of an index customer was readily transferred via fomites in a simulated restaurant. A simple intervention involving use of disposable menus and disinfection of pens was effective in preventing transfer to subsequent customers, and the use of disinfectant wipes eliminated the virus from contaminated tables and chairs. These findings provide support for CDC recommendations to reduce surface and fomite-mediated transmission of SARS-CoV-2.³

Our study has some limitations. Simulations cannot mimic all conditions present in a busy restaurant setting and transfer of bacteriophage MS2 may differ from SARS-CoV-2. Additional studies are needed in working restaurants. The index customer in the simulation deposited MS2 from hand contamination rather than via coughing or sneezing. SARS-CoV-2 may survive for several hours on human skin,^{Reference Hirose, Ikegaya and Naito10} but it is not known whether viable SARS-CoV-2 is commonly present on hands of infected patients.