Association Between Eyelid Laxity and Obstructive Sleep Apnea

Key Points

Question  Is there an association between obstructive sleep apnea and eyelid laxity when using quantitative eyelid measurements?

Findings  This cross-sectional study that included 201 patients from a sleep clinic found no statistically significant association between obstructive sleep apnea severity and markers of eyelid laxity or secondary ocular surface disease. Subset regression analysis revealed several potential confounding variables.

Meaning  There appears to be no association between the presence or severity of obstructive sleep apnea and markers of eyelid laxity or secondary ocular surface disease.

Importance  While much has been reported on the relationship between floppy eyelid syndrome and obstructive sleep apnea (OSA), the diagnostic criteria of floppy eyelid syndrome are often subjective and vague.

Objective  To evaluate the association between OSA and quantitative markers of eyelid laxity or secondary ocular surface disease in a sleep clinic population.

Design, Setting, and Participants  This investigation was a cross-sectional observational study at the Center for Sleep Medicine at Icahn School of Medicine at Mount Sinai. Participants were individuals referred for overnight polysomnography from March 1 to August 30, 2015.

Main Outcomes and Measures  Eyelid laxity and ocular surface disease were assessed on bedside ophthalmologic examination. The presence and severity of OSA were determined from polysomnography results. Initial correlation between OSA and ocular surface and eyelid markers was calculated through bivariate linear regression analysis, and the association between ocular symptoms was obtained through bivariate ordered logistic regression. Analysis was repeated adjusting for known associations between OSA and sex, age, body mass index, and medical comorbidities through multivariable analysis.

Results  In total, 201 individuals (402 eyes) were enrolled in the study. Their mean (SD) age was 53.2 (13.5) years, 43.3% (n = 87) were female, 56.7% (n = 114) were of white race/ethnicity, 26.9% (n = 54) were black/African American, 4.0% (n = 8) were Asian, 8.0% (n = 16) were multiracial or other, and 4.5% (n = 9) were of unknown race/ethnicity, with 21.9% (n = 44) of all individuals self-identifying as Hispanic and 75.1% (n = 151) self-identifying as non-Hispanic. After adjustment, no association was observed between OSA severity and an eyelid laxity score (regression coefficient, 0.85; 95% CI, −0.33 to 0.62; P = .40) or an ocular surface score (regression coefficient, 1.09; 95% CI, −0.32 to 0.29; P = .93). Through subset analysis, male sex was associated with a higher ocular surface score, while older age and diabetes were associated with a higher eyelid laxity score. Only one patient (0.5%) exhibited findings of floppy eyelid syndrome.

Conclusions and Relevance  Among individuals referred for overnight polysomnography, quantitative markers of eyelid laxity were not associated with the presence or severity of OSA. Subset analysis suggests that prior studies may have been limited by confounding variables or the technique of identifying eyelid laxity.

Obstructive sleep apnea (OSA) is a disorder characterized by recurrent disruptions in breathing and decreased oxygen saturation when sleeping due to partial or complete physical obstruction of the pharynx and upper airway, resulting in disturbed sleep and daytime somnolence. Patients with OSA are at increased risk for impaired neurocognitive performance and adverse medical outcomes secondary to hypoxemia and frequent arousals during sleep.^1,2

Quiz Ref IDFloppy eyelid syndrome (FES), often reported to be associated with OSA, was first described in 1981 by Culbertson and Ostler³ in a case series of 11 patients. They described easily everted, floppy upper eyelids with lax rubbery tarsal plates and secondary chronic papillary conjunctivitis in obese, middle-aged men. However, subsequent definitions did not include the original criteria of sex, age, and altered tarsal plate consistency.^4-6 The inconsistent definitions, combined with discovery of FES findings in patients outside of those described by the classic definition,^7,8 led van den Bosch and Lemij⁹ to establish a broader clinical entity of lax eyelid syndrome (LES), a disorder characterized by lax upper eyelids, chronic papillary conjunctivitis, punctate epithelial keratitis, and ocular discharge secondary to eyelid laxity. They argue that FES is a distinct subset of LES, distinguished by symptoms and its association with male sex and obesity. Although some researchers have acknowledged this definition of LES, the term itself has not achieved recognition.^10-12

Because FES remains the more well-known entity, much of what has been reported about the relationship between disorders of the eye and OSA describes FES instead of LES.^7,13-16 The prevalence of FES among the general population is not well established but appears to range from 2.3% to 3.8%.¹⁷ Some reports have used vague or subjective definitions of FES, whereas other reports have attempted to use quantitative measures of eyelid laxity.^13,18-23 There are also inconsistencies in methods of diagnosing OSA, ranging from sleep questionnaires to polysomnography.^17,21,22 In addition, some studies have a small study size, creating both sample bias and difficulty with proper statistical analysis to control for potential confounding variables.^18-23

Owing to the subjective nature and lack of standardized criteria in diagnosing FES, we sought to assess the relationship between the presence and severity of OSA and markers of eyelid laxity in a cohort of patients referred for overnight polysomnography testing. This study also presents a validated, quantitative, and reproducible scale of measuring eyelid laxity. It represents the largest study to date examining the association between LES and OSA.

This investigation was a prospective cross-sectional observational study that was approved by the institutional review board of Icahn School of Medicine at Mount Sinai. The research was conducted in accord with the principles expressed in the Declaration of Helsinki and the Health Insurance Portability and Accountability Act of 1996 regulations. Written informed consent was obtained from all participants.

Individuals were recruited from the Center for Sleep Medicine at Icahn School of Medicine at Mount Sinai from March 1 to August 30, 2015. Before undergoing overnight polysomnography testing, they were voluntarily enrolled in the study. Persons younger than 18 years, those whose primary language was not English, prisoners, and individuals unable to provide informed consent were excluded.

Detailed ophthalmic history and bedside ophthalmologic testing were performed before patients underwent polysomnography. Examiners (T.P.F., J.A.S., A.C.C., and F.P.P.-N.) were masked to the reason for polysomnography, as well as the individuals’ medical history and Epworth Sleepiness Scale (ESS) score. All examiners were trained together by one of us who is a board-certified oculoplastic specialist (A.Y.W.) for 2 months before examining patients individually.

Individuals were questioned on the presence of chronic ocular surface irritation, including discomfort, tearing, itching, and redness. They were given a numeric score based on symptom severity as follows: 0 for none, 1 for mild to moderate, and 2 for severe. Participants were also asked whether they or a sleeping partner noted spontaneous eyelid eversion within the past year. A score of 1 was given for a positive spontaneous eyelid eversion history.

Bedside examination included the following: measurement of corrected visual acuity documented in logMAR; measurement of eyelid laxity, including horizontal eyelid distraction, upper eyelid traction, and presence of eyelash ptosis; and measurement of ocular surface disease, including palpebral conjunctival reaction and corneal staining with fluorescein. Horizontal eyelid distraction was performed by measuring the distance in millimeters from the upper eyelid margin to the corneal apex, while the upper eyelid was distracted away from the globe with gentle pulling on the eyelashes as previously described (Figure, A).^24,25 The measurement in millimeters was converted to a 0 to 4+ scale for statistical purposes. Upper eyelid traction maneuver was performed by lifting the lateral third of the upper eyelid with a single finger and evaluating eyelid elasticity and the presence of induced eyelid eversion (Figure, B). Upper eyelid traction was graded on a scale of 0 (normal) to 4+ (very lax with induced eyelid eversion). Eyelash ptosis was evaluated by observing the position of the eyelashes relative to the eyelid margin. Eyelash ptosis was graded on a scale of 0 (normal) to 4+ (severe eyelash ptosis with encroachment into the visual axis) (Figure, C).

To evaluate for palpebral conjunctival reaction, the upper and lower eyelids were everted and observed for gross papillary conjunctivitis, again graded on a scale of 0 (normal) to 4+ (severe papillary conjunctivitis). Finally, corneal staining was performed using fluorescein strips and evaluating for punctate corneal staining with a cobalt blue light, again graded on a scale of 0 (no staining) to 4+ (severe punctate epithelial erosions). Each eye was tested and recorded separately.

After study closure on August 30, 2015, all participants’ electronic medical records and polysomnography results were reviewed. The presence of medical comorbidities, including diabetes, hypertension, dyslipidemia, thyroid disease, and coronary artery disease, was recorded. Body mass index (BMI) was recorded from the night of each individual’s sleep study.

Data were recorded from polysomnography results. They included the number of apneas and hypopneas per individual, the apnea-hypopnea index (AHI), the minimum oxygen saturation observed, the mean oxygen saturation observed, the oxygen desaturation index (number of 4% desaturations per hour), and the percentage of sleep period each participant spent with an oxygen saturation less than 90%.

The ESS was administered by the Division of Pulmonary, Critical Care, and Sleep Medicine of the Department of Medicine at Icahn School of Medicine at Mount Sinai to individuals before undergoing polysomnography. The ESS is a common tool used to assess sleep propensity in 8 typical situations on a scale of 0 to 3. A higher score indicates a greater degree of sleepiness. A score less than 10 is generally found in healthy individuals.²⁶ The examiners were masked to ESS scores, which were retrospectively analyzed.

Bedside ophthalmologic testing was grouped into the following 3 scores: the eye symptoms score, the eyelid laxity score, and the ocular surface score (Table 1). The eye symptoms score was calculated by adding an individual’s ocular surface irritation score (range, 0-2) to the spontaneous eyelid eversion score (range, 0-1), with a maximum score of 3.

The eyelid laxity score included horizontal eyelid distraction, vertical traction maneuver, and grade of eyelash ptosis as previously described in the Testing subsection of the Methods section. Each eye was measured separately. Horizontal eyelid distraction was measured in millimeters and converted to a 0 to 4+ graded scale as follows: 0 is less than 6 mm, 1 is 6 to 8 mm, 2 is 9 to 11 mm, 3 is 12 to 15 mm, and 4+ is greater than 15 mm. With all 3 measurements being graded on a 0 to 4+ scale and individual eyes averaged together, the maximum eyelid laxity score was 12. The ocular surface score was a summation of the amount of palpebral conjunctival reaction and corneal staining with fluorescein, both of which were graded on a 0 to 4+ scale. Individual eyes were averaged together, creating a maximum ocular surface score of 8. For both the eyelid laxity score and the ocular surface score, individual eyes (in addition to the mean scores per patient) were also analyzed.

Obstructive sleep apnea severity was determined by an individual’s AHI, which measures the frequency of respiratory disruptions (apneas and hypopneas) during sleep. An apnea was defined as cessation in breathing lasting more than 10 seconds, and a hypopnea was defined as a decrease in airflow of at least 30% with an arousal or oxygen desaturation of 3% or more. Obstructive sleep apnea severity was defined as mild for an AHI of at least 5 but less than 15, moderate for an AHI of at least 15 but 30 or less, and severe for an AHI exceeding 30 per hour.²⁷

Study data were analyzed through bivariate and multivariable regressions. All 3 outcomes (the eye symptoms score, the eyelid laxity score, and the ocular surface score) were analyzed at the patient level; however, the effect of OSA severity on the eyelid laxity score and the ocular surface score was also analyzed with individual eyes. Unadjusted statistical analysis of OSA severity and participant demographics was performed using bivariate ordered logistic regression. A t test was used to explore the unadjusted association between the presence of OSA and the patient characteristics. Two-sided significance level was set at P < .05. Bivariate linear regression was used to explore correlations between OSA severity and the eyelid laxity score, whereas bivariate ordered logistic regression was used to examine correlations between OSA severity and the eye symptoms score and the ocular surface score. These models were replicated in multivariable analysis to control for possible confounding comorbidities between OSA and sex, age, BMI, and medical comorbidities, which were performed both on the total score per patient, as well as for each individual eye for the eyelid laxity score and the ocular surface score. Analysis was repeated exploring a potential correlation between the ESS score and the eyelid laxity score and the ocular surface score. All analyses were performed using statistical software (Stata, version 13.0; StataCorp LP).

In total, 201 patients of 265 possible participants met the study inclusion criteria and voluntarily enrolled. All 201 participants (402 eyes) underwent bedside ophthalmologic testing and polysomnography. Thirty-five participants had normal polysomnography results, with 33 patients classified as having mild OSA, 70 patients as having moderate OSA, and 63 patients as having severe OSA. The mean (SD) age of participants was 53.2 (13.5) years, 43.3% (n = 87) were female, 56.7% (n = 114) were of white race/ethnicity, 26.9% (n = 54) were black/African American, 4.0% (n = 8) were Asian, 8.0% (n = 16) were multiracial or other, and 4.5% (n = 9) were of unknown race/ethnicity, with 21.9% (n = 44) of all individuals self-identifying as Hispanic and 75.1% (n = 151) self-identifying as non-Hispanic. Analyzing participant demographics (Table 2), there was an unadjusted association noted between OSA severity and male sex, BMI, older age, and history of hypertension.

Quiz Ref IDUnadjusted analysis of the presence or severity of OSA and the eye symptoms score, the eyelid laxity score, and the ocular surface score found no significant association. After adjustment through multivariable regression analysis for sex, age, BMI, and medical comorbidities, no association was observed between OSA severity and the mean eye symptoms score (odds ratio [OR], 0.92; 95% CI, −0.69 to 1.22; P = .57), eyelid laxity score (regression coefficient [RC], 0.85; 95% CI. −0.33 to 0.62; P = .40), or ocular surface score (RC, 1.09; 95% CI, −0.32 to 0.29; P = .93). Mean values of eye symptom, eyelid laxity, and ocular surface scores are provided in Table 3. Analyzing individual eyes also found no association between the eyelid laxity score (RC, 0.07; 95% CI, −0.08 to 0.22; P = .38) or the ocular surface score (RC, 0.09; 95% CI, −0.10 to 0.28; P = .36) or and OSA severity.

Quiz Ref IDAnalyzing patient demographics and comorbidities as independent variables, there was a positive association between the eye symptoms score and male sex (mean [SD] RC, 0.53 [0.16]; P = .04). Analysis with spontaneous eyelid eversion as the independent variable demonstrated no significant association between OSA severity and demographic or medical comorbidity variables. The eyelid laxity score was associated with older age (mean [SD] RC, 0.18 [0.06]; P = .005) and the presence of diabetes (mean [SD] RC, 0.82 [0.31]; P = .008). The ocular surface score was associated with male sex (mean [SD] RC, 0.30 [0.15]; P = .048) (Table 4). When analyzing individual eyes, there was no association between OSA severity and the eyelid laxity score or the ocular surface score; however, there was positive association between the eyelid laxity score and age (mean [SD] RC, 0.16 [0.06]; P = .004), BMI (mean [SD] RC, −0.03 [0.01]; P = .008), and the presence of diabetes (mean [SD] RC, 0.86 [0.22]; P < .001). There was also a correlation between the ocular surface score and BMI (mean [SD] RC, 0.04 [0.02]; P = .03) and hypertension (mean [SD] RC, 0.57 [0.29]; P = .01). Epworth Sleepiness Scale scores were available for 189 patients. No association was observed between ESS scores and the eyelid laxity score (regression coefficient, 0.01; 95% CI, −0.09 to 0.11; P = .41), the ocular surface score (regression coefficient, 0.02; 95% CI, −0.04 to 0.09; P = .67), or ocular surface symptoms (OR, 0.98; 95% CI, −0.92 to 1.03; P = .71).

The primary outcome measure of the study was to evaluate measures of eyelid laxity rather than to diagnose patients as having FES. However, there was only one patient (0.5%) identified who fit the classic description of FES outlined in the initial report by Culbertson and Ostler.³

Our results suggest no association between patients with known OSA or ESS score and eyelid laxity or LES. No association was found between the presence or severity of OSA and markers of eyelid laxity, ocular surface disease, or symptoms after controlling for sex, age, BMI, and medical comorbidities. Adjusted subset analysis identified possible confounding variables. Male sex was found to be associated with a higher eye symptoms score and ocular surface score, and older age and the presence of diabetes were associated with a higher eyelid laxity score. While many prior relevant reports have controlled for age and BMI, only one study²² controlled for sex. No previous studies on this subject to date controlled for medical comorbidities.

In our study, we found one patient (0.5%) who exhibited the classic description of FES. Our 0.5% prevalence of FES is lower than that in previous reports, where FES prevalence among OSA populations ranges from 2% to 34%.^{17-20,22,23,28-30} This discrepancy may reflect variability in study methods, particularly with how OSA and FES were diagnosed. Many of the prior studies were also limited by small sample size. Floppy eyelid syndrome likely remains an uncommon diagnosis, and our finding of low FES incidence may ultimately reflect the rarity of the disease. Larger cohort sizes might more accurately detect rates of this condition. While the prevalence of FES among the general OSA population is low, as in our study, the prevalence of OSA among patients with known FES appears to be much higher.¹⁴

Of the 10 published studies^{17-20,22,23,28-31} to date evaluating the prevalence of FES among OSA populations, only half used quantitative measurements to evaluate eyelid laxity.^17-20,23 We sought to increase the specificity of our examination results by using several previously reported maneuvers to measure markers of LES. Horizontal eyelid distraction was described by Iyengar and Khan²⁴ and has been used by Kadyan et al,¹⁷ Mojon et al,¹⁹ Ezra et al,²¹ and Muniesa et al.²³ Mojon et al¹⁹ found signs of FES in 13 of 44 (29.5%) total patients with OSA and reported on the varying prevalence across OSA severity (22.2% in those with mild OSA, 16.7% in those with moderate OSA, and 39.1% in those with severe OSA). Although Kadyan et al¹⁷ measured horizontal eyelid distraction, they ultimately based the diagnosis of FES on subjective assessment of eyelids that easily evert, the presence of papillary conjunctivitis, and symptoms of ocular irritation, finding FES in 28 of 89 (31.5%) patients with OSA. Muniesa et al²³ defined FES as a positive eyelid laxity measurement with lids that easily evert and papillary conjunctivitis in the same upper eyelid. They reported that 60.7% of patients with OSA had indications of hyperlaxity and found signs of FES in 14 of 89 (15.7%) patients with OSA, with FES prevalences of 11.1% among those with mild OSA, 8.3% among those with moderate OSA, and 19.6% among those with severe OSA. A case-control study by Ezra et al²¹ found an association between FES and OSA (OR, 12.5; 95% CI, 3.1-108.9, P < .001) and other measures of eyelid laxity, including upper eyelid medial canthal laxity, upper eyelid distraction, and palpebral aperture.

Vertical traction of the upper eyelid associated with FES and OSA has previously been reported by McNab,¹⁸ who evaluated 20 patients with established OSA for eyelid abnormalities by measuring vertical excursion of the upper lid. He found little overlap between patients with FES (measurements of 15-25 mm) and those without FES (measurements of 7-12 mm). Only one of 20 patients (5%) was found to have classic FES, although no statistical analysis was performed.¹⁸ Lash ptosis has also been associated with eyelid laxity and was initially reported as being associated with FES by McNab,¹⁸ with further data by Langford and Linford³² validating its association shortly thereafter. Other quantitative measures have been used to diagnose LES. For instance, Karger et al²⁰ used an electronic strain gauge device and found no significant association between subjectively easy eyelid eversion or the force required to displace the upper eyelid and OSA status after controlling for age and BMI.

The wide variability in the literature may also in part stem from inconsistent methods in which OSA is diagnosed. Overnight polysomnography in a sleep laboratory remains the criterion standard in diagnosing and classifying OSA.^1,2 Other studies^18,28 also had small sample sizes and failed to control for potential confounding variables. Some controlled for age and BMI, but most studies failed to adjust for medical comorbidities. The variability also may represent bias of examiners to overreport FES in patients with known OSA.³³

Quiz Ref IDTo our knowledge, this study represents the largest study of its type to date to examine all patients undergoing polysomnography at a sleep clinic with bedside ophthalmologic evaluation. Study weaknesses include the limitation of our bedside examination: slitlamps were not accessible, thus precluding evaluation of subtle ocular findings. While all examiners were trained for 2 months by the same physician, variance between them was not measured. The examiners were unable to be masked to the sex, age, and approximate BMI of the participants, which may have introduced possible bias. In addition, because all individuals were referred for overnight polysomnography, the results herein may not be generalizable.

Study strengths include rigorous and validated measurements of eyelid laxity performed by a limited number of examiners. Our use of a quantitative scale also allowed for an objective and reproducible method to grade eyelid laxity through bedside examination, and we believe that use of the eyelid laxity scale increased the specificity of our examination findings for LES. Our large sample size allowed not only for decreased type II errors and sample bias but also for complex regression analysis to adjust for known and potential associations of participant demographics and medical comorbidities. Finally, all of our patients underwent criterion standard overnight polysomnography testing in a sleep laboratory, which enabled more detailed analysis of sleep dynamics.

Our study found no association between measures of eyelid laxity and either the presence or severity of OSA among individuals referred for overnight polysomnography. We also found a low prevalence of FES among our study population. Our study methods allowed for a validated, objective way to approach LES. The correlation observed between male sex and older age with OSA and FES status suggests that smaller studies reporting positive associations may have been limited by variability in examination techniques, OSA and LES or FES diagnostic criteria, and confounding variables.

Accepted for Publication: July 15, 2017.

Corresponding Author: Albert Y. Wu, MD, PhD, Department of Ophthalmology, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, PO Box 1183, New York, NY 10029 (albert.wu@mssm.edu).

Published Online: September 7, 2017. doi:10.1001/jamaophthalmol.2017.3263

Author Contributions: Drs Fox and Wu had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Fox, Chang, Feinsilver, Wu.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Fox, Yim, Feinsilver, Wu.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Fox, Schwartz, Wu.

Obtained funding: Wu.

Administrative, technical, or material support: Fox, Chang, Yim, Feinsilver, Wu.

Study supervision: Fox, Feinsilver, Wu.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported.

Additional Contributions: Christopher M. Smith, MA, CMI (Mount Sinai Health System) composed the illustrations. Desh Nandedkar, MD (Department of Sleep Medicine, Icahn School of Medicine at Mount Sinai) assisted with interpreting Epworth Sleepiness Scale data. No compensation was received.