Face Masks and Bacterial Dispersion Toward the Periocular Area.

During the coronavirus disease 2019 pandemic, patients wear face masks, creating unique airflow toward the ocular surface that may complicate ocular procedures such as intravitreal injections. It has been hypothesized that oropharyngeal droplets from either the patient or physician increase the risk of postinjection infectious endophthalmitis. Prior studies have demonstrated reduced bacterial growth on culture media with simulated injections when the physician uses a face mask , and when implementing a no-talking policy. In this study, we set out to determine how a patient wearing a mask affects bacterial dispersal in the direction of the eye because this may be a factor for intravitreal injection-related endophthalmitis.

Using schlieren imaging, we qualitatively evaluated air currents with a face mask on. When focusing on the superior aspect of a participant wearing a mask, schlieren imaging showed air escaping from the superior aspect of the mask. The same mask with tape covering the superior aspect showed no significant airflow toward the eye (Fig 1 ).

Figure 1

Schlieren imaging showing postimaging processing of a participant wearing (A, C) a facemask without tape and (B, D) a facemask with tape (outlined by dashed red lines) covering the superior aspect of the mask. The control with (A, B) no breathing and (C, D) breathing are shown. C, Wearing a mask without tape redirects air movement towards the eye. D, Having the mask taped prevents air from being directed toward the eye, and air is seen passing through the face mask away from the eye (arrow).

Next, to determine if the redirected air toward the ocular surface contained more bacteria, we used blood agar plates and counted the colony-forming units (CFUs) with different methods of mask wearing. Institutional review board approval was obtained through the University of Wisconsin, and all research conducted adhered to the tenets of the Declaration of Helsinki. Consenting participants were all older than 18 years.

Blood agar plates with 5% sheep blood in tryptic soy agar base were incubated at 37° C in 5% carbon dioxide for 48 hours. Plates were brought to room temperature from their storage temperature of 4° C to eliminate variability in organism recovery rate based on initial incubation temperatures. Participants used a standard face mask with elastic earloops and a wire-containing nasal bridge (3M Company).

Control plates were held uncovered and perpendicular to the floor for 2 minutes away from the individual. In the subsequent groups, a blood agar plate was placed at each inferior orbital rim perpendicular to the floor. We enrolled 54 participants, and each performed the 5 scenarios below, in which they were instructed to speak or count aloud for 2 minutes (Fig S2, available at www.aaojournal.org): (1) control; (2) no face mask worn; (3) face mask fully covering the mouth, but placed just below the nose (inappropriate use); (4) face mask covering the mouth and nose (recommended use); and (5) face mask covering the mouth and nose, with paper adhesive tape applied to seal the superior portion of the mask.

The mean CFUs for each group were as follows: control, 0.24 (95% confidence interval [CI], 0.14–0.42); no mask, 1.93 (95% CI, 0.54–6.86); mask below the nose, 0.67 (95% CI, 0.34–1.30); mask appropriately worn, 0.35 (95% CI, 0.16–0.78); and taped mask group, 0.13 (95% CI, 0.06–0.29; Table S1 and Fig S3, available at www.aaojournal.org).

The taped mask group showed 81% (95% CI, 48%–93%; P = 0.001) fewer CFUs than the group wearing a mask inappropriately below the nose. Fewer CFUs also were observed when the group appropriately wearing masks was compared with the group wearing masks below the nose (47% reduction; incidence rate ratio, 0.53; 95% CI, 0.32–0.87; P = 0.011). Some suggestion (P = 0.08) was found that taped masks showed a lower mean CFU compared with when masks are worn normally (incidence rate ratio, 0.37; 95% CI, 0.12–1.13). Taped masks also showed 73% fewer CFUs (95% CI, 26%–90% fewer; P = 0.011) than the average CFUs for other nontaped forms of wearing a mask (appropriate and inappropriately worn grouped together).

We did not find taping the superior aspect of a mask to decrease bacterial dispersal toward the ocular surface when compared with an appropriately worn mask. Even more importantly, this study looked at only bacterial dispersal toward the ocular surface and did not evaluate for endophthalmitis; this would be extremely difficult owing to its low prevalence. We hypothesize that taping the superior aspect of a mask decreases dispersal of bacteria toward the eye based on the statistical trend and schlieren imaging taken together. Further studies are needed to evaluate bacterial dispersal.

During the planning stage of the study, we anticipated greater separation between these groups. Based on current best estimates, and still intending to detect at least a 2-fold separation between mean CFUs, a replication study involving only these 2 groups would need 72 participants to have an approximately 81% chance of identifying such an effect at the 0.05 level, assuming separation of at least that size genuinely exists.

Many recommend a no-talk policy, by both the physician and patient, during intravitreal injections to help prevent excessive bacterial dispersal toward the ocular surface. After a small pilot study, we chose to have the patients speak for 2 minutes because this did improve culture yields. But for our intravitreal injections performed in clinic, we have everyone in the room maintain a no-talk policy unless essential information needs to be relayed. Additional limitations of this study include not controlling for facial shape, anatomic features, or hair. We also did not determine what species of bacteria grew on these plates.

Other benefits may exist for taping face masks, because condensation often forms on glasses during acuity checks and on examination lenses during slit-lamp and indirect examination. Also, when the mask is taped to the patient’s face, it reduces the temptation for the patient to lower or remove the mask during the visit.

Overall, we did not show less bacteria being dispersed toward the ocular surface when comparing appropriately worn masks with masks with the superior edge taped. However, inappropriately worn masks direct more bacteria toward the ocular surface, and taping the superior aspect of masks redirects air away from the eye, as shown with schlieren imaging, but the clinical significance could not be determined in this study. These data should serve for hypothesis generation and should help to guide future directions for examining how taping the superior aspect of a mask affects bacterial dispersal.