Purpose
To evaluate the association of children’s daily electronic screen use with severe meibomian gland atrophy (MGA).
Design
Retrospective cross-sectional study.
Methods
Children (aged 6-17years) presenting at clinical practice December 2016 – October 2017 were evaluated for ≥grade 2 MGA vs age-matched controls with insignificant atrophy (<grade 1 atrophy). Questionnaires assessed dry eye symptoms, daily electronic screen use hours, diet, and outdoor time. Meibography imaging assessed for severe meibomian gland atrophy (≥grade 2 atrophy; ≥1 eyelid on validated, 4-point, ImageJ scale: 0 [normal] – 3 [severe]). Autoimmune disease biomarker positivity was assessed in 16 severe meibomian gland atrophy cases after being found relevant in firstcase.
Results
A total of 172 children were evaluated. Patients with known meibomian gland atrophy causes or poor-quality meibographies were excluded. Forty-one met inclusion criteria (mean age, 11 years; 49% female): 17 cases had severe meibomian gland atrophy; 24 controls had insignificant gland atrophy. All severe meibomian gland atrophy cases had ocular symptoms/signs of dry eye disease including corneal neovascularization (29%), best-corrected visual acuity loss (41%), and central corneal neovascularization (14%). No controls had significant dry eye symptoms/signs. Controls had lower/“better” meibogrades vs cases ( P < .01). In severe meibomian gland atrophy cases, 86% reported ≥4 hours of daily electronic screen use; 50% reported ≥8 hours. No controls exceeded 2 hours. Increased electronic screen use was positively associated with increased/“worse” meibogrades (odds ratio: 2.74; 95% confidence interval, 1.39-5.41). In 16 severe meibomian gland atrophy cases, 62.5% tested positive for autoimmune biomarker(s), though none had systemic symptoms: 18.8% rheumatoid factor; 6.25% SS-A/SS-B; 31.3% early Sjögren syndrome biomarkers; 6.25% ANA-positive/RF-negative. Autoimmune disease biomarker positivity was not significantly associated with severe meibomian gland atrophy vs controls ( P = .34, right-eye; P = .71, left-eye).
Conclusions
Children’s excessive electronic screen use is associated with severe meibomian gland atrophy. Further research is needed to establish formal electronic screen use limits based on meibography grade and evaluate correlation of autoimmune disease biomarker positivity in children with severe meibomian gland-atrophy.
Global lockdowns owing to the coronavirus disease 2019 (COVID-19) pandemic have necessitated increased electronic screen use (ESU) for work and schoolwork, particularly in children, which have surged to all-time highs. , Additionally, the recent increase in the search term “eye pain” is likely a secondary effect representing the rise in global ESU rather than of a direct effect from COVID-19. ,
“Digital addiction” describes the global concern of excessive ESU, though its definition remains unclear. Viewing electronic devices have been associated with numerous ailments, including attention deficit/hyperactivity disorder, anxiety, depression, decreased outdoor activity, obesity, “transient smartphone blindness,” headaches, neck/back pain (ie, computer vision syndrome [CVS]), sleep disturbances, and dry eye disease owing to meibomian gland dysfunction/atrophy (MGA) , (ie, owing to decreased blink rates during ESU). Of these, dry eye disease (DED) and MGA are particularly concerning, as DED patients’ quality-of-life scores are comparable to those of patients with moderate/severe angina in adults, and DED has been associated with suicidal ideation. As tools are developed to diagnose and treat these electronic screen–induced ailments in adults and legislation is introduced to mitigate digital addiction, digital screens continue to enter more homes, schools, and preschools, which is affecting young children, particularly during the COVID-19 pandemic.
Dry eye symptoms have been correlated with ESU ≥4 hours/day in adults owing to decreased blink rates. , Each normal blink forces meibum oil from the meibomian glands and pushes ocular surface debris into the eyelid fornix and out the nasolacrimal duct. This cycle is essential for proper tear production, meibomian gland secretion, and tear distribution/drainage. Decreased or incomplete blinks (as occurs with ESU via cerebral pathways involving multiple neural sources and altered by mood, reading, ESU, and other attention-requiring visual tasks) can lead to rapid tear evaporation, an unstable tear film, increased ocular surface inflammatory factors, meibomian gland orifice scarring, and chronic eye pain. , In children, dry eye symptom risk factors include spending longer durations on phones/screens and shorter durations outdoors.
Meibography is a noninvasive technology that objectively visualizes MGA, the key cause of dry eye symptoms (Supplemental Fig. 1 , available at AJO.com). Severe meibomian gland atrophy, defined as ≥33% meibomian gland atrophy (ie, ≥grade 2 atrophy) on a validated 4-point ImageJ scale, can lead to a lifetime of chronic eye redness, irritation, and pain (Supplemental Fig. 2 , available at AJO.com). Meibography may help define a critical threshold of excessive ESU as it impacts one’s meibomian glands, particularly in children.
In adults, other severe MGA causes include Sjögren syndrome, rosacea, hormone use, radiation, chemotherapy, demodex, contact lens use, allergies, and autoimmune diseases (ie, rheumatoid arthritis [RA], juvenile idiopathic arthritis [JIA], and lupus). There are no published papers to date demonstrating associations between severe MGA on meibography and RA/JIA or lupus in children or demonstrating the potential additive effect of ESU on MGA of a patient with autoimmune disease biomarker positivity (ADBP). This cross-sectional study uses meibography to report an association between ESU hours and MGA severity. Additionally, we evaluate ADBP as a potential independent risk factor for severe MGA.
Methods
Participants and Procedures
All children aged 6-17 years, presenting to our office between December 2016 and October 2017, who met inclusion criteria and had high-quality infrared meibography imaging on ≥2 eyelids, were included in this cross-sectional study. Inclusion criteria included patients with severe MGA with no known history of contact lens use, rosacea, isotretinoin/Retin-A/hormone use, refractive surgery, ocular surgery, eye trauma, pterygium, ocular herpes, viral conjunctivitis, autoimmune disease, cancer/chemotherapy/radiation, or any past medical history including personal/family history of autoimmune disease, or systemic symptoms (ie, joint pain, dry mouth). Cases were diagnosed as severe MGA (ie, ≥grade 2 meibography scores [range: 0-3] on at least 1 eyelid); controls had <grade 1 MGA at the same time period as cases were assessed. This study adhered to the Declaration of Helsinki’s principles. Georgetown University Hospital’s Institutional Review Board approved retrospective and prospective data collection for this study. Patients’ verbal and parents’ informed written consent was obtained.
Children had an ophthalmologic eye examination including uncorrected visual acuity, best-corrected visual acuity, refraction, anterior segment examination before and after fluorescein drops, dilated fundus examination, and infrared eyelid meibography (LipiScan; TearScience, Morrisville, North Carolina, USA). Meibography scores were assessed by 2 masked graders using a validated, 4-point ImageJ scale (http://imagej.nih.gov/ij/; NIH, Bethesda, Maryland, USA): grade 0 = no meibomian gland loss (MGL); grade 1 = <33% MGL; grade 2 = 33%-66% MGL; grade 3 = >66% MGL ; severe MGA ≥33% MGA , (Supplemental Fig. 1 and 2 ).
Questionnaires assessed systemic symptoms/findings, including the Ocular Surface Disease Index (validated tool assessing dry eye symptoms/vision impact), cumulative hours of ESU per day (CHESUD), daily diet choices (ie, vegetable and omega-3 consumption), age, resident ZIP codes (ie, median household income measure), and outdoor time (ODT). Additionally, categorical variables (ie, race, sex, allergy history, demodex presence) were collected.
An unexpected finding of a positive rheumatoid factor (RF), repeated twice, with no initial presenting systemic complaints in the first child with severe MGA, prompted full evaluations in the remaining cases. Of severe MGA cases, 16 underwent ADBP tests including complete blood count, RF, antinuclear antibody (ANA), erythrocyte sedimentation rate, SS-A, SS-B, C-reactive protein, and early Sjögren biomarkers (salivary protein-1, parotid secretory protein antibodies, carbonic anhydrase VI) (Sjö; Nicox, Inc, Durham, NC). Children with positive results were referred to rheumatology.
Statistical Analysis
Categorical variables (ie, race, sex, allergy, vegetable consumption, demodex, meibography scores) were summarized using counts and percentages by cases and controls and were compared using either Pearson χ 2 test or Fisher exact test. Age, ODT, median household income, and CHESUD variables were summarized using mean (± standard deviation) and were compared using a Wilcoxon signed rank test. Unadjusted ordinal logistic regression was performed to assess the associations between each predictor as an independent variable and meibography scores of left/right/both eyes as a dependent variable, respectively. Adjusted ordinal logistic regression models were followed to assess the association between CHESUD and meibography scores of left/right/both eyes, respectively, with adjustment for covariates with univariate association P value< .1. We also dichotomized the CHESUD variable as excessive (≥4 hours, 32%) and mild exposure (≤3 hours, 68%) and repeated the same analyses. To assess the association between ADBP and severe MGA, we used the Wilcoxon signed rank test where each eyelid’s total gland loss area (%) was a surrogate measure of severe MGA. A 2-sided P value < .05 was considered statistically significant. R software (ver 3.4.0) was used for statistical analyses.
Post hoc power analysis was conducted based on 1,000 computer simulations with a 2-sided type I error of 5%. Considering the distribution of combined meibography scores (n = 41) in Supplemental Table 1 (Supplemental Material available at AJO.com), the power to evaluate the effect size of clinically meaningful odds ratio (OR) > 2.0 (or 2.7 in our study) in the ordinal logistic regression model was >99%.
Results
A total of 172 children were evaluated; 41 met inclusion criteria. Seventeen cases had severe MGA (ie, ≥grade 2 meibography scores [range: 0-3] on at least 1 eyelid); 24 controls had <grade 1 MGA (Supplemental Fig. 3 , available at AJO.com). Supplemental Table 1 presents initial findings. Cases had significantly increased CHESUD compared to controls (6.29 ± 3.93 hours, cases vs 74 ± 0.14 hours, controls, P = .02).
In severe MGA cases, 86% reported ≥4 hours CHESUD; 50% reported ≥8 hours CHESUD. There was no age difference between cases and controls ( P = .30). Cases were mostly female (n = 12, 70.6%); controls were mostly male (n = 16, 66.7%) ( P = .02). Race, allergy history ( P = .11), demodex infection ( P = .19), and median household income ( P = .96) were not statistically significant between cases and controls. Vegetable consumption frequencies were significantly different ( P < .01): 57% of severe MGA cases reported never/rarely eating green leafy vegetables and 79% reported never/rarely eating omega-3-rich foods. Significantly more outdoor time was observed for controls ( P < .01); children in the control group spent on average 2.25 hours per day outdoors, compared to 0.92 hours per day in the case group. In severe MGA cases, 70% reported spending ≤1hour ODT during the school week and 14% reported spending zero hours ODT during the week. Significantly higher meibography scores (ie, worse meibomian gland disease) were observed for both left and right eyes in severe MGA cases ( P < .01, respectively). Both-eye combined meibography scores were also higher in severe MGA cases ( P < .01).
Cases’ initial presentation are noted in Supplemental Table 1.
All cases had severe MGA on meibography scores of ≥grade 2 (range: 0-3) on at least 1 eyelid (Supplemental Fig. 3 ), and 21% needed future corneal transplant.
CHESUD appeared to be positively associated with the increase in combined meibography scores in both unadjusted and adjusted analyses (OR: 2.81; 95% confidence interval [CI], 1.66 to 4.77; OR: 2.74; 95% CI, 2.39 to 5.41, respectively) (Supplemental Table 2, available at AJO.com). We observed similar associations in left and right eyes (left eye OR, 1.39; 95% CI, 0.84 to 2.31; and right eye OR, 4.75; 95% CI, 1.19 to 18.91), respectively (Supplemental Table 2). ODT shows a marginally negative association with the increase in combined meibography scores ( P = .10 [OR, 0.34; 95% CI, 0.10 to 1.19]). This weak association remains for the left eye (OR, 0.30; 95% CI, 0.09 to 1.05) but not for the right eye (OR, 0.18; 95% CI, 0.21 to 3.22). Allergy was not associated with elevated meibography scores in all models, with P > .05. With binary CHESUD categories (>4 hours vs ≤4 hours), excessive electronics exposure is very strongly associated (OR > 1,000) with the increased combined meibography scores ( P < .01).
In the 16 tested cases, 62.5% tested positive for an autoimmune marker: 18.8% rheumatoid factor positive, 6.25% SS-A, SS-B positive, 31.3% early Sjögren syndrome biomarker(s) positive, 6.25% ANA positive/RF negative. Six patients (37.5%) had negative serologies. There was no significant association of ADBP to total gland area loss of the left eye and the right eye ( P = .71 and P = .34, respectively), and there was no significant effect of modifications of ESU by the presence of ADBP ( P = .16, right eye; P = .18, left eye) Fig. 3
Discussion
Increased electronic-screen viewing has been associated with increased dry eye symptoms. Blinking rates decrease (from 15.5 blinks/minute to 6.5 blinks/minute, median) and incomplete blinks increase when viewing digital devices. We present the first study using a new, objective tool to assess for severe MGA in children as a potential diagnostic indicator of excessive ESU (ie, “digital addiction”), the severity of CVS, and possible autoimmune disease risk.
We found that 86% of cases reported >4 hours of CHESUD, which has been reported to increase DED risk in adults. Disturbingly, 50% of cases reported ≥8 hours CHESUD. Children with electronic screen-induced dry eye disease (SIDED) and severe MGA have an increased risk of chronic eye pain, corneal scarring, and vision loss. This can translate into a lifetime of eye drops, expensive eyelid treatments, and procedures. Two of our patients treated with amniotic membrane grafts will need corneal transplantation to decrease amblyopia risk.
The term SIDED is preferable to the term CVS, as any electronic screen has the potential to impact blink rates. Also, screen-induced systemic syndrome may be more appropriate for systemic sequelae, such as headaches, neck pain, and insomnia.
While excessive reading can potentially cause severe MGA, page turning may force eye blinks. Also, few children sit for many hours at a time to read a book, whereas many view screens/video games for hours. Given complex cortical pathways involved in spontaneous and reflex blink in adults and children, future randomized controlled studies are needed to compare excessive reading to excessive screen use on long-term MGA in children Tables 1 and 2 .
