Assessment of dry eye disease in N95 versus surgical face mask wearers during COVID-19.

Purpose: Investigating the effect of different face masks on dry eye disease (DED) among healthcare workers during the COVID-19 pandemic.
Methods: This was a comparative, cross-sectional study. Participants were included into two groups: group 1 (n = 30) wore surgical masks, and group 2 (n = 30) wore N95 masks with face shields. Demographic and ocular surface disease index questionnaires (OSDI) were performed. In addition, Tear break-up time (TBUT), corneal and conjunctival fluorescein staining, and meibography to assess meibomian gland loss (MGL) were performed on all participants. Independent T-test was used to compare continuous parameters and Chi-square test for categorical variables. The relationship between continuous variables was tested using bivariate Pearson correlation.
Results: Sixty healthcare workers participated in this study (36 females and 24 males). The mean (±SD) age of the surgical mask and N95 groups was 35.33 (±14.94) and 36.63 (±10.64) years, respectively. Both masks caused dryness according to TBUT, MGL, and OSDI scores. DED per DEWS II definition was observed in 14 (46.7%) and 16 (53.3%) patients in groups 1 and 2, respectively. Comparing the two groups, N95 mask caused significantly more dryness according to TBUT (P = 0.042) and fluorescein staining (P = 0.038 for the right eye and P = 0.015 for the left eye).
Conclusion: Physicians should be aware of the potential dry eye signs secondary to face mask wear during the COVID-19 pandemic. Further attention should be taken in patients who suffer from preexisting dry eye syndrome and in patients who undergo intraocular operations.

Since January 2020, a highly contagious RNA virus called coronavirus has attacked the whole world and caused atypical pneumonia that was termed as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).[1] One of the World Health Organization recommendations was the use of face masks among the healthcare workers and the general public as the coronavirus disease (COVID-19) is still causing a high number of deaths worldwide. Healthcare workers who work in COVID-19 high-risk areas should be face-protected by N95 masks and face shields while attending the patients, whereas the rest by surgical facemasks. Israeli healthcare workers were required to wear face masks since March 2020 when the COVID-19 pandemic started in the country.

Face masks can cause de novo headache or worsening of a previously existing headache and cutaneous irritation that might be secondary to mechanical factors and hypercarbia.[23]

Two previous letters published the possibility of dry eye and irritation due to the widespread use of face masks recently.[45] Therefore, this study was conducted to investigate deeply the effect of different face masks on dry eye disease (DED) among healthcare workers during the COVID-19 pandemic. To our knowledge, this is the first study in the literature to evaluate this correlation.

Methods

Study population

This comparative, cross-sectional study included healthcare workers between September and November 2020. The study was approved by the local institutional review board and adhered to the principles of the Declaration of Helsinki. All participants signed a consent form. The study included two groups. Group 1 included healthcare workers who do not work in the COVID-19 department; they use surgical face masks. Group 2 included healthcare workers in the COVID-19 department; they use N95 masks and face shields while attending COVID-19 patients.

Workers who smoke, use contact lenses, or suffer from connective tissue disease that can cause dry eye were not included in this study.

All participants filled two questionnaires with one investigator (AN): demographic data, role, department, and the mean daily and weekly hours wearing the masks. The second questionnaire was the ocular surface disease index (OSDI). It is a 12-item questionnaire that assesses symptoms of ocular irritation and measures the subjective severity of DED. The participants were familiar with the language of the OSDI questionnaire. The score of OSDI is on a scale of 0–100. A higher OSDI score represents more severe dryness.[6]

Objective dry eye measurements were examined by a different investigator (SHA) in the following order: tear break-up time (TBUT, this is the timing of the breakup of the precorneal tear film after the initial blink; a TBUT of <10 s is considered abnormal),[7] corneal and conjunctival fluorescein staining with a yellow barrier filter (it was graded according to a modified Oxford scheme: absent, mild, moderate, and severe),[8] and Meibomian gland loss (MGL; this was performed by capturing infrared images with the BG-4M noncontact meibography system (SBM Sistemi, Turin, Italy)). Images were digitally analyzed using Integrated Complete Platform (ICP) software. MGL is defined as the percentage of the gland loss in relation to the total tarsal area of the everted lid that ranges from 0% to 100%).[9] DED was defined according to the TFOS DEWS II definition as ocular surface symptoms presence and other signs of DED (OSDI score more than 13 and any 1 sign of abnormal TBUT or abnormal ocular surface staining).[10]

In addition, Schirmer test was performed (without anesthesia using Schirmer tear test strips (Med Devices Lifesciences, London, UK). For further information regarding a detailed description of the objective dry eye measurements, an additional supplemental file is provided.

All the questionnaires and the tests were performed in two adjacent examining rooms at the ophthalmology department. The rooms’ temperature was normal and centrally controlled (22°C–24°C) and humidity up to 50%. The participants were examined during their duty, post 6 hours since starting their shift for consistency.

Sample size

A priori sample size calculation was not performed. During the time of the current study, the only people in the country who used N95 on daily basis were healthcare workers at the COVID-19 departments. In our tertiary hospital, there were around 32 workers, of whom 30 accepted to participate in the study. Thus, to compare two equal sample groups, we sampled 30 participants in the other group. In addition, 30 is often considered sufficient for central limit theorem to hold. Thus, participants in the N95 group were chosen based on availability (not randomly), whereas surgical mask group participants were randomly chosen from the hospital database of the employees that was exported to an Excel file. The built-in rand function was used to select the subjects for the N95 group.

Statistical analysis

Categorical variables were described using frequencies and percentages; continuous variables were described using means and standard deviations. Tests assumptions of normal distribution were assessed based on the guidelines suggested by Kim, showing no significant departure from normality in all continuous variables within the groups.[11] Independent T-test was used to compare continuous outcomes and Chi-square for categorical variables between the two groups. The relationship between continuous variables was tested using bivariate Pearson correlation. P < 0.05 was considered as statistically significant. All analyses were performed using R software (version 4.0.2).

Results

Thirty participants were included in each group. In group 1, the surgical mask group, there were 7 men; in group 2, the N 95 group, 17 men. This difference was significant (P = 0.008). However, sex was not associated with any of the study outcomes. The mean age ± standard deviation in groups 1 and 2 was 35.33 ± 14.94 (range: 18–61) and 36.63 ± 10.64 years (range: 18–64), respectively. In group 1, 8/30 (26.6%) were physicians and the rest were nurses, technicians, and cleaners, compared to 11/30 physicians in group 2. There were no significant differences in age, eyeglasses wearing, and role between the groups. However, the surgical mask group used the mask for significantly more weekly hours than the N95 mask group (43.37 hours vs. 25.67, P < 0.001). The demographic data are summarized in Table 1.

Table 1

Summary table of the demographic data of the study

	Surgical mask (n=30)	N95 (n=30)	Total (n=60)	P
Sex (male)	7 (23.3%)	17 (56.7%)	24 (40.0%)	0.0082
Age - M (SD)	35.33 (14.94)	36.63 (10.64)	35.98 (12.88)	0.6991
Role				0.4052
Non-Physician	22 (73.3%)	19 (63.3%)	41 (68.3%)
Physician	8 (26.7%)	11 (36.7%)	19 (31.7%)
Weekly hours wearing a mask - M (SD)	43.37 (10.67)	25.67 (11.13)	34.52 (14.02)	<0.0011

1t-test for unpaired groups. 2Pearson’s Chi-squared test

Dry eye assessments

All participants had normal Schirmer test results. Both masks caused dryness according to TBUT, MGL, and OSDI scores [Table 2]. According to the DEWS II definition, 14 (46.7%) and 16 (53.3%) patients had DED in groups 1 and 2, respectively (P = 0.606). In addition, there was a significantly higher MGL in the upper lid compared to the lower lid in all participants (22.7 vs. 7.4, P < 0.001, Fig. 1). Comparing the two groups [Table 2], TBUT (P = 0.042) and fluorescein staining (P = 0.038 for right eye and P = 0.015 for left eye) were significantly higher in the N95 mask group. In contrast, there was no significant difference regarding upper lid and lower lid MGL and OSDI score between the groups (P = 0.903, P = 0.936, P = 0.879). A multiple linear regression analysis was performed with mean MGL as a dependent variable predicted by total weekly hours of mask-wearing, type of mask, sex, and age [Table 3]. There was no significant association between mean MGL and the mean total weekly hours of mask-wearing (b = 0.02, P = 0.805); only age was significant in predicting MGL (b = 0.15, P = 0.033). Next, a multiple linear regression analysis was performed to predict TBUT as a dependent variable by total weekly hours of mask-wearing, type of mask, sex, and age [Table 4]. It was found that low TBUT was significantly associated with longer time of wearing a mask (b = −0.06, P = 0.015), older age (b = −0.05, P = 0.010), wearing N95 mask (b = −2.47, P = 0.001), and being a female (b = −1.14, P = 0.039). N95 mask was significantly inversely correlated with TBUT (P = 0.001), indicating that N95 masks are associated with lower TBUT and more dryness than surgical masks. Finally, a logistic regression to predict DED based on total weekly hours of mask-wearing, age, sex, and type of mask revealed that no significant association between all independent variables and DED [Table 5]. However, the type of mask tended to be significant (P = 0.054), where wearing an N95 mask was associated with more DED (OR = 4.66).

Table 2

Comparison of dry eye variables by type of mask

	Surgical mask (n=30)	N95 (n=30)	Total (n=60)	P
Right Fluorescein staining				0.0382
Absent	26 (86.7%)	19 (63.3%)	45 (75.0%)
Mild	4 (13.3%)	11 (36.7%)	15 (25.0%)
Left Fluorescein staining				0.0152
Absent	27 (90.0%)	19 (63.3%)	46 (76.7%)
Mild	3 (10.0%)	11 (36.7%)	14 (23.3%)
OSDI M* (SD)	14.96 (18.80)	14.27 (15.84)	14.61 (17.24)	0.8791
DED	14 (46.7%)	16 (53.3%)	30 (50%)	0.6062
Upper eyelid MGL** - M (SD)	22.88 (9.92)	22.58 (8.96)	22.73 (9.37)	0.9031
Lower eyelid MGL - M (SD)	7.50 (5.68)	7.35 (8.53)	7.42 (7.18)	0.9361
Tear break-up time - M (SD)	5.25 (1.91)	4.13 (2.23)	4.69 (2.14)	0.0421

1t-test for unpaired groups. 2Pearson’s Chi-squared test. *OSDI=ocular surface disease index, **MGL=meibomian gland loss

Figure 1

The difference in meibomian gland loss (MGL) between the upper and lower eyelids in all participants. Paired samples t-test showed that there was a significant difference between upper and lower eyelid (t (59) = 118, P < 0.001, delta = 15.3). As demonstrated in this figure, the mean (SD) of MGL for the lower eyelid was significantly lower than that of the upper eyelid (7.4 (7.2) vs. 22.7 (9.4)). ***P <.001

Table 3

Multiple linear regression analysis for predicting mean MGL* based on total weekly hours of mask wearing, age, sex, and type of mask (n=60)

Dependent: MGL Predictors	Coefficient	CI	t	P
Constant	8.35	−2.07-18.77	1.61	0.114
Weekly time wearing a mask (hours)	0.02	−0.15-0.19	0.25	0.805
Age	0.15	0.01-0.29	2.19	0.033
Sex [female]	0.77	−3.00-4.53	0.41	0.685
Type of mask [N95]	0.20	−4.55-4.94	0.08	0.934
R2/R2 adjusted	0.088/0.022

*MGL=meibomian gland loss

Table 4

Multiple linear regression analysis for predicting TBUT* based on weekly hours of mask wearing, age, sex, and type of mask

Dependent: TBUT Predictors	Total sample		Surgical mask		N95
	Coefficient	P	Coefficient	P	Coefficient	P
Constant	10.55	<0.001	10.35	<0.001	7.66	0.001
Weekly time wearing a mask (hours)	−0.06	0.015	−0.06	0.054	−0.06	0.151
Age	−0.05	0.010	−0.06	0.016	−0.04	0.377
Sex [female]	−1.14	0.039	−0.61	0.420	−1.62	0.061
Type of mask [N95]	−2.47	0.001
n	60		30		30
R2/R2 adjusted	0.273/0.220		0.274/0.190		0.202/0.110

*TBUT=Tear break-up time; 1st regression model included the total sample, 2nd model included only surgical mask group, and the 3rd model included only N95 group. Since a lower TBUT would indicate more dry eye then a significant negative correlation indicates more dry eye.

Table 5

Logistic regression analysis for predicting DED* based on total weekly hours of mask wearing, age, sex, and type of mask

Dependent: DED Predictors	Total sample			Surgical mask			N95
	OR	CI	P	OR	CI	P	OR	CI	P
Constant	0.04	0.00-1.19	0.077	0.00	0.00-0.21	0.034	1.62	0.03-95.36	0.812
Weekly time wearing a mask (hours)	1.05	1.00-1.13	0.070	1.10	1.01-1.26	0.071	1.02	0.94-1.11	0.665
Age	1.00	0.96-1.04	0.945	1.02	0.96-1.08	0.587	0.98	0.90-1.06	0.544
Sex [female]	2.59	0.81-9.08	0.119	12.08	1.32-384.77	0.064	1.34	0.28-6.93	0.712
Type of mask [N95]	4.66	1.04-25.11	0.054
n	60			30			30
R2 Tjur	0.095			0.247			0.027

*DED=Dry eye disease

Discussion

DED is one of the most common ocular surface diseases that causes ocular irritation and affects the quality of life of the patients.[12] It is classified into two subtypes: aqueous deficient dry eye (ADDE) and evaporative dry eye (EDE).[13] Meibomian glands secrete the external lipid layer of the tear film, which prevents tear evaporation and reduces the surface tension of the tear film to maintain a healthy ocular surface.[14] Therefore, any condition that affects the meibomian glands can cause EDE. It has not been determined yet if the use of face masks increases the risk of dry eye disease during the COVID-19 pandemic. In this study, we investigated the objective and subjective signs of dry eye secondary to the use of two different face masks amongst healthcare workers. We found that when comparing between the groups, there was a significant increase of the objective signs of dry eye in N95 mask group versus surgical mask group; however, when increasing the dry eye diagnosis threshold, N95 mask was not associated significantly with more DED. When evaluating the results by multiple regression analysis, age was the only independent variable related to mean MGL, which is consistent with a previous study by Pult.[15] Interestingly, correlations between TBUT and N95 masks were stronger than correlations between TBUT and age. In addition, sex was excluded by multiple regression analysis, and there were no significant differences in DED between men and women. The OSDI score did not differ significantly between the two groups. This finding is not surprising because it is known that ocular surface symptoms may have low correlations with clinical signs.[16]

A literature review was conducted to find studies that investigated DED secondary to face mask use. A commentary paper by Moshirfar et al.[4] published that they observed an increase in ocular irritation among face mask users during the current COVID-19 epidemic. The finding has an important consequence on the health of the ocular surface. In addition, a letter to the editor by Giannaccare et al.[5] was published recently regarding dry eye in the COVID-19 era. They proposed two mechanisms that could be responsible for DED: the increase of smart schooling/working and the use of face mask. Those observations were noted in our findings too. None of the mentioned publications performed the tests that were done in this study. Moreover, this study included only healthcare workers as they are compliant and use masks almost full-time. We postulate that one of the mechanisms of developing DED secondary to face mask users is due to the escape of the exhaled warm air from the mask; this causes tear evaporation and therefore reduces TBUT as was observed in this study. Previous studies showed that devices that blow air around the face mechanically may increase dry eye symptoms, such as powered air-purifying respirators and continuous positive airway pressure masks.[1718] The same mechanism may be valid for face masks. Our results showed that wearing N95 mask was associated with more dry eye signs (TBUT and corneal staining) compared to surgical mask users, although it was used for significantly fewer weekly hours. That is a surprising finding as N95 masks are more air sealed compared to the surgical masks and there is no difference in the temperature and humidity between COVID-19 and regular departments. This can be explained by the nose wire fitting which may cause displacement of the lower lid that results in lagophthalmos or lower lid dragging and therefore affects the eye blinking process.

In the current study, we observed a significant effect on MGL in the upper lid compared to the lower lid in both groups which may be secondary to the masks or morphological diversity between the two eyelids. Our finding is consistent with Dogan et al.,[19] who stated that the upper lid might be the preferred lid to make an evaluation of detecting MGL. Previous studies showed that MGL at the upper eyelid is significantly smaller than that of the lower eyelid.[1520] This finding differs from ours. During the previous studies,[1520] masks were not used as it is nowadays due to the ongoing COVID-19 pandemic. Therefore, masks may play a role in the discrepancy. It may be of interests to evaluate MGL in both eyelids for masks and non-mask users in the future.

To our knowledge, this is the first study that examined the DED secondary to face masks during COVID-19. We cannot comment, based on the findings of this study, whether the masks themselves induce dry eye disease as there was not a third arm of non-mask wearers. Furthermore, though there were significantly more dry eye signs in the N95 group, there were no significant differences in symptoms. We speculate that if the dry eye signs continue for a long time, it may lead to dry eye disease, including worsening in dry eye symptoms. In the future, this should be further investigated in a prospective study design. The findings of this study are important during this crisis as dry eye signs can cause irritation and affect the ocular surface health. Moreover, it can lead the person to touch his eyes more frequently, which may increase the virus transmission.[21]

In the current COVID-19 climate, wearing masks is mandatory indoors and in public places and especially on hospital grounds. As such, we could not, from an ethical standpoint, have a third group of non-mask wearers. Therefore, our study is limited by not including a control group. In addition, the sample size is relatively small; however, to the best of our knowledge, it is the largest in the literature to date. In addition, comparing the effect of masks with and without tapes on DED is of interests to evaluate it in further investigation that may give a better idea of why the masks cause dryness. Another limitation is that it was not a double-blinded study, but the results of the questionnaires were not seen by the clinician who performed the ophthalmic examinations, which should reduce any potential researcher bias. In addition, eyelid blinking rate was not estimated during our study which can affect ocular dryness.

Conclusion

Clinicians should be aware of the potential dry eye signs secondary to face mask use among all the population during the COVID-19 pandemic. Further attention should be emphasized for patients who suffer from preexisting ocular dryness, contact lens users, and during postoperative period of ophthalmic procedures as those patients are at increased risk of ocular surface alterations and infections. They should be examined carefully and treated for ocular dryness during the follow-up period. Our recommendations are to fit the mask properly, mainly when there is a sensation of an escape of air into the eyes or impaired eyelid closure. In addition, patients and healthcare workers should use non-preservative eye lubricants more frequently during the pandemic and to allow the eye to rest from face mask more often.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.