Mask use by patients in the context of COVID-19 can increase the risk of postinjection endophthalmitis.

Abstract

The close proximity puts ophthalmologists at particular risk for contracting COVID-19 from patients and vice versa. To reduce the risk of postinjection endophthalmitis (PIE), clinicians commonly wear masks when performing intravitreal injections (IVI). In the context of COVID-19, it may seem intuitive to extend this practice to include the patient wearing a mask. However, observations raise the question as to whether patient masks could paradoxically increase the risk of PIE; masks can leak exhaled air towards the eyes and contaminate the injection site with aerosolized droplets (Hadayer et al. 2020; Raevis et al. 2021). The suggestion that mask use by patients puts them in peril of PIE is alarming. Consequently, we have compared the risk of PIE at Oslo University Hospital (OUH) since the beginning of the COVID-19 pandemic with the pre-pandemic risk. We defined the pandemic period as March 1, 2020 through January 31, 2021. Mask use by patients was influenced by rules and recommendations from the Norwegian Institute of Public Health and the City of Oslo in particular and occurred sporadically until August. It then gradually became common, and ultimately a majority of our patients wore masks during the IVI procedure. In both the pre-pandemic and pandemic period, the IVI were generally performed ambulatory and took place in positive air pressure cleanrooms. Povidone iodine 5% was used as antiseptic (no antibiotic prophylaxis was used). The standard procedure was to drape the patient after the initial application of povidone iodine, but in the pandemic period clinicians temporarily chose to first drape patients not wearing a mask. There were five cases of PIE of 68.150 IVI with compounded syringes at OUH in the pre-pandemic period (Blom et al. 2020). A total of 25 904 IVI were given in the pandemic period, 14 649 of these from August 2020 through January 2021. Seven cases of PIE were identified, six of these after August 2020. Microbiological analyses were positive for pathogens in most of the cases; the bacteria all belonged to the microbiota of the skin and upper respiratory tract. Table 1 displays a summary of the PIE cases. The relative risk (RR) of PIE in the pandemic period compared to the pre-pandemic period was 3.68 (95% CI 1.17–11.60; p = 0.026). From August 2020 through January 2021, when mask use by patients gradually became common, the RR of PIE was 5.58 (95% CI 1.70–18.29; p = 0.005). Pre-pandemic period n = 68 150 Pandemic period n = 25 904 Our findings support the hypothesis that patient masks could redirect exhaled air towards the eyes and contaminate the injection site. Additional aspects may also contribute to an increased risk of PIE. First, the widespread use of masks by the general public has led to a new form of dry eye disease: mask-associated dry eye (MADE). A healthy tear film possesses antimicrobial properties, and air leakage from poorly fitting masks that dries the ocular surface may disrupt the eye’s innate immunity and increase its susceptibility to infections (McDermott 2013). Second, a normal ocular microbiota helps prevent pathogenic species from colonizing the ocular surface (Petrillo et al. 2020). Regular mask use may alter the microbiota by contaminating the surface with microbes of higher pathogenic potential. An increased RR notwithstanding, PIE remains a rare occurrence at OUH. The risk of contracting COVID-19, on the other hand, is still imminent. Accordingly, we do not encourage patients to stop using masks but instead advise clinicians to be vigilant about sealing the injection site and counsel patient about properly adjusting their masks and treatment for dry eyes. Ultimately, COVID-19 vaccination would mitigate the risk of transmission in the setting of IVI and allow for safe reimplementation of a procedure for which patients do not wear masks.