These authors contributed equally to the work.
Floppy eyelid syndrome is associated with obstructive sleep apnoea: a prospective study on 127 patients
Version of Record online: 11 OCT 2011
© 2011 European Sleep Research Society
Journal of Sleep Research
Volume 21, Issue 3, pages 308–315, June 2012
How to Cite
CHAMBE, J., LAIB, S., HUBBARD, J., ERHARDT, C., RUPPERT, E., SCHRODER, C., MALAN, A., BOURCIER, T. and BOURGIN, P. (2012), Floppy eyelid syndrome is associated with obstructive sleep apnoea: a prospective study on 127 patients. Journal of Sleep Research, 21: 308–315. doi: 10.1111/j.1365-2869.2011.00968.x
- Issue online: 19 MAY 2012
- Version of Record online: 11 OCT 2011
- Accepted in revised form 21 August 2011; received 28 February 2011
- palpebral laxity;
- sleep apnoea
A few investigations have raised the question of a possible relationship between obstructive sleep apnoea syndrome (OSAS) and floppy eyelid syndrome (FES). FES is an easily inverted floppy eyelid with papillary conjunctivis, and is a subset of the general pathology, lax eyelid syndrome. The aim of the current study is to determine whether OSAS severity is associated with FES. One hundred and 27 consecutive subjects (aged 25–75 years) referred to the Strasbourg University Sleep Clinic with suspicion of OSAS were included. All patients underwent overnight ambulatory respiratory polygraphy, comprehensive ophthalmological examination and completed standard sleep questionnaires. OSAS severity was defined based on the patient’s obstructive apnoea–hypopnoea index (AHI). As expected, age, body mass index (BMI) and the proportion of males increased with OSAS severity. FES was observed in 15.8% of the subjects without OSAS, 25.8% of the total OSAS population and the frequency was significantly increased (40%) in patients with severe OSAS (AHI > 30 h−1). A significant correlation between OSAS severity and FES was found after adjustment for age, sex and BMI, using a principal component analysis (PCA). The multivariate analysis included clinical, polygraphic and comorbidity data and was followed by logistic regressions for the main components extracted from the PCA. In summary, our findings show an association between OSAS severity and FES and suggest that severe OSAS might be an independent risk factor for FES. These two disorders may share common biological determinants, such as tissue elasticity. Finally, clinicians should be aware of this association so that underlying OSAS or FES can be detected.
Recent investigations have suggested a relationship between obstructive sleep apnoea syndrome (OSAS) and different ophthalmological disorders, such as chronic open angle glaucoma, non-arteritic anterior ischaemic optic neuropathy, keratoconus and floppy eyelid syndrome (FES) (Abdal et al., 2006; Ezra et al., 2010; Kadyan et al., 2010; McNab, 2007; Pepin et al., 2010; Waller et al., 2008). FES was first described by Culberston and Ostler in 1981 as an easily everted eyelid with papillary conjunctivis (van den Bosch and Lemij, 1994; Culbertson and Ostler, 1981) and corresponds to a specific subset of the more general pathology, lax eyelid syndrome (LES) (Fowler and Dutton, 2010). The prevalence of FES within the general population is not well established, but seems to vary within a range of 2.3 and 3.8% (Kadyan et al., 2010). FES has been associated with keratoconus (Culbertson and Tseng, 1994), although the underlying pathophysiological mechanisms remain poorly understood (Culbertson and Tseng, 1994; Kadyan et al., 2010; Mojon et al., 1999). An association with sleep has been suggested because FES is commonly more severe on the side the patient is used to sleeping on (McNab, 2007). A few investigations have raised the question of a possible relationship between FES and OSAS. The frequency of OSAS in patients affected with FES is unknown, shown by the fact that two different studies reported it as 96 (McNab, 2007) and 31% (Ezra et al., 2010). Conversely, studies looking at FES in patients with OSAS found a prevalence ranging from 1.5 to 31.5% (Ezra et al., 2010; Karger et al., 2006; McNab, 2007; Robert et al., 1997). The occurrence of eyelid laxity in OSAS patients is much higher, but given previous estimations ranging from 29.5 to 65% (De Groot, 2009), a consensus has clearly not been reached. The inconsistencies between these studies can be explained due to low patient sample size or methodological considerations regarding diagnostic criteria for FES or OSAS. In most studies the criteria for FES is based on the association of lax eyelid and papillary conjunctivitis; however, conclusions remain unclear, as one study associates it with lash ptosis (Karger et al., 2006) and another with symptomatic ocular irritation (Kadyan et al., 2010).
The international criteria for OSAS diagnosis are based on a respiratory events index during sleep and require polygraphic or polysomnographic recording (PSG) at night (Iber et al., 2007). In the previously mentioned studies, the methodologies used are inconsistent with one another. Specifically, several used only oxymetry (Kadyan et al., 2010) or an Epworth questionnaire (Ezra et al., 2010), both of which are methods not considered valid for OSAS diagnosis (Iber et al., 2007; Silber et al., 2007). Among the three studies which used polysomnographic recordings, one had a very low patient sample size (McNab, 1997) and the second lacked any adjustments for confounding factors (Mojon et al., 1999). Finally, the only study using PSG for OSAS diagnosis, also taking these factors into account (Karger et al., 2006), found no correlation between FES and OSAS after adjusting for age and body mass index (BMI). Thus, FES prevalence in the OSAS population remains unknown and the association between FES and OSAS is yet to be demonstrated completely.
In the present study, we looked at FES and eyelid laxity in patients referred to the Strasbourg University Sleep Clinic and evaluated for OSAS. The primary goal of the study was to investigate the association between FES and OSAS, while controlling for possible confounding factors such as gender, age and BMI. The secondary aim was to evaluate the prevalence of FES and eyelid laxity in OSAS patients.
We included one hundred and 27 (a sample size of 25 necessary patients per AHI category group was calculated, based on Kadyan et al., 2010) consecutive, unrelated Caucasian subjects, referred for the first time to the Strasbourg University Sleep Clinic with a suspicion of OSAS (51 women, 75 men, mean age 52.8 ± 2.1 years, age range 25–75 years). The time-period for patient inclusion was February 2009–December 2010. Patients diagnosed or treated previously for sleep apnoea and patients with morbid obesity were excluded (BMI > 40). All procedures complied with the guidelines for clinical research as stated in the Helsinki Declaration. All included patients were informed and gave written consent for taking part in this study. All patients underwent a physical examination, a measure of neck circumference (only taken in 80 patients), sleep interviews, a full overnight ambulatory respiratory polygraphy, ophthalmological examination and completed standard sleep questionnaires (Epworth Sleepiness Scale, Pittsburgh Sleep Questionnaire and Pichot Fatigue Scale). The ophthalmological examination and polygraphy were scored independently by two investigators (S.L. and J.C. or B.K., respectively), each of them blind to the outcome of the other. Information pertaining to age, sex, BMI, diabetes, hypertension, dyslipidaemia, tabagism and atopy were assessed systematically.
Overnight respiratory polygraphy (EMBLETTA© portable sleep apnoea monitoring device; Resmed, Paris, France) included measurements of oronasal airflow (nasal canula plus thermistor), chest and abdominal movements, oxygen saturation (finger pulse oxymetry), body position and snoring (microphone). The polygraphic recordings were scored manually according to the American Academy of Sleep Medicine (AASM) criteria for apnoea and hypopnoea (Iber et al., 2007; Silber et al., 2007) by two independent experienced evaluators who were blind to any information pertaining to the study (with a final consensus between scorers in case of discrepancies). In the absence of electroencephalogram (EEG) to define arousal, inspiratory airflow limitation was defined as a flattening of the nasal pressure waveform lasting at least 10 s in conjunction with heart rate acceleration concomitant to regaining normal respiratory function. Polygraphy was repeated if the total scoring time lasted less than 240 min. The scoring period, corresponding to an estimation of the total sleep time, excluded the time-periods indicated by the patient as being awake or with artefacts obviously corresponding to waking.
The apnoea/hypopnoea index (AHI) was calculated as the number of apnoea and hypopnoea episodes per hour of scoring time. The respiratory events index (REI) was calculated by adding AHI to the number of inspiratory airflow limitation. Other parameters for analysis were also calculated over the scoring period, including snoring time (S), mean oxygen saturation (mSaO2) and oxygen desaturation index (ODI) calculated as the number of desaturation drops (>4%) from baseline SaO2 per hour. OSAS severity was defined according to the international recommendations of Sleep Disorders-2 criteria (AASM, 2005) and patients were separated into four groups: no OSAS (AHI ≤ 5), mild OSAS (AHI > 5, ≤15), moderate OSAS (AHI >15, ≤30) and severe OSAS (AHI > 30).
The ophthalmological examination included measurements of visual acuity, slit-lamp examination with anterior segment analysis and ophthalmoscopy, intraocular pressure and corneal thickness measurements as well as corneal topography and visual field recording via a computerized device. An optical coherence tomography (OCT) was also performed for macular profile and optic fibre analysis (retinal nerve fibre layer analysis programme). Eyelid examination was conducted specifically to evaluate eyelid laxity and obtain a FES diagnosis. In the absence of standardized evaluation, ophthalmologists involved with the study chose consensus criteria upon which to score the horizontal laxity of the upper eyelid in five groups of severity: grade 0 was normal laxity; grade 1 asymptomatic upper eyelid hyperlaxity: the clinical definition of lax eyelid syndrome (LES); grade 2: FES defined as papillar conjonctivis with eyelid hyperlaxity, which is the clinical definition of FES; grade 3: grade 2 + tarsal eversion when the eyelid is horizontally tracted; and grade 4: grade 3 + persisting tarsal eversion. FES was defined starting at grade 2. The examination was performed by the same ophthalmologist to limit the variability of the measure and who was blind to the results of polygraphy.
A double-blind scoring for a random sample of 24 subjects was performed with a second scorer and confirmed the high reproducibility of the evaluation (kappa coefficient = 0.91).
Statistical analyses were performed by biostatisticians (A.M., J.C.) using Statistica version 8 (StatSoft, Tulsan, OK, USA).
Values for each of the parameters are presented as mean ± standard error of the mean (SEM). Each continuous variable was tested for normality with a Kolmogorov–Smirnov test, and graphically with a qq-plot. BMI, sex, age and mSaO2 were normally distributed. AHI, REI and ODI were not normally distributed. Therefore, we used a log-transformation of those variables to include them in the principal component analysis model (PCA) model (see below).
To compare the frequencies (FES, laxity) between AHI category groups, we used Pearson’s chi-squared tests for dichotomous variables and one-way analysis of variance (anova) followed by a Bonferroni correction for continuous variables. We selected the relevant variables to include in the multivariate analysis from the correlation matrix. We included sleep factors (AHI, REI, mSaO2, ODI), age, gender and BMI. No significant correlation was observed between FES or LES and hypertension, dyslipidaemia, diabetes, tabagism or atopy. To examine the relation between OSAS severity and FES/LES we performed a logistic regression first with FES, and then LES as the dependent factor, including sleep factors, age, gender and BMI. However, the model met fitting difficulties because of collinearity between factors. Therefore, we used a PCA followed by correlation with the most informative components. This multivariate analysis was employed to circumvent the problem of collinearity between confounding factors and between sleep parameters and the PCA included several variables pertaining to OSAS severity such as AHI, REI, mSaO2 and ODI and confounding factors such as age, sex and BMI. This analysis allowed us to reduce the large number of factors to three orthogonal axes (principal components).
A logistic regression was then performed for the first three principal components with FES, then LES, and a multiple linear regression for the first three principal components with eyelid laxity severity as the dependent variable. These correlations did not intend to show an association with FES or LES of a specific variable, but with OSAS as a whole. As stated in the results below, component 1 of the PCA was a nearly exclusive combination of sleep variables.
Within the group of 127 subjects, all cases of SAS were due to obstructive mechanism with no central sleep apnoea syndrome observed. Data on demographics and comorbidities of the population are shown in Table 1, according to OSAS severity. As expected, subjects with greater OSAS severity were significantly older and had a higher BMI. Additionally, the prevalence of patients with high blood pressure, diabetes and of male gender significantly increased with OSAS severity.
|AHI||<5||≥5; <15||≥15; <30||≥30||P|
|Age||45.6 ± 1.9||53.8 ± 1.9||56.8 ± 2.0||59.2 ± 1.4||<0.001|
|BMI||25.7 ± 0.7||28.8 ± 0.7||27.5 ± 0.8||30.0 ± 0.9||0.002|
|Epworth||8.6 ± 0.8||10.5 ± 0.7||10.2 ± 1||10.9 ± 1||0.27|
|Pichot||15.3 ± 1.3||15.4 ± 1.1||13.8 ± 1.5||12.5 ± 1.9||0.47|
|Neck circumference||38.4 ± 0.6||40.0 ± 0.8||40.1 ± 0.9||43.1 ± 1.6||0.026|
The Epworth Sleepiness Scale showed a non-significant trend towards an increase with OSAS severity. Neck circumference increased significantly with OSAS severity (P = 0.026). No significant differences in dyslipidaemia, tabagism or atopy were observed between the groups (Table 1).
Floppy eyelid syndrome (eyelid laxity grades 2–4) was observed in 22.8% of the patient population and its prevalence was significantly higher in severe OSAS than in non-OSAS patients (P = 0.041). More specifically, FES was present in 15.8% of subjects without apnoeas (AHI < 5 h−1) and in 25.8% of the patients affected with OSAS (AHI > 5 h−1) (not significant). However, this percentage reached 40% in patients with severe OSAS (AHI > 30 h−1) and was significantly higher than those observed in the less severe groups (P = 0.047).
LES (grades 1–4) was observed in 44.1% of the population and its prevalence increased significantly with OSAS severity (P = 0.015). Laxity was present in 31.6% of subjects without apnoeas (AHI < 5 h−1) and in 49.4% of all patients affected with SAS (AHI > 5 h−1) (not significant), yet the frequency reached 75% in severe OSAS (AHI > 30 h−1) (P = 0.003). Results are shown in Fig. 1.
The difference of eyelid laxity severity between groups did not reach statistical significance. Regarding other ocular disorders, the frequency of chronic open-angle glaucomas in our population sample was 6%, distributed as follows: three non-OSAS patients, five mild OSAS patients and none in the moderate or severe groups. No cases of keratoconus were observed.
The correlation matrix showed collinearities between AHI, REI, mSaO2, ODI, BMI, age and gender, which was confirmed by multivariate regression and explained the problems with fitting the models. Therefore, we used PCA and subsequent multiple regressions, the results of which are summarized in Fig. 2. In the PCA statistical model, we merged the variables of interest from the previous bivariate correlation study, that is the OSAS indicators AHI, REI, mSaO2 and ODI, as well as possible confounding factors: age, sex and BMI. This analysis enabled us to group all variables into a few major orthogonal factors or principal components that were linear combinations of the original variables (Fig. 2a,b). The first three principal components were retained, as they contained 82% of the total information. The first was the most indicative, explaining 53% of the variance (Fig. 2c). This factor was made up of all OSAS parameters analysed (Fig. 2d) and, to a minor extent, BMI. The two remaining factors corresponded mainly to the gender (factor 2) and to age and BMI (factor 3) (Fig. 2d) (see also Fig. 1).
Considering FES, the logistic regression analysis revealed that a statistically significant proportion of the variance was accounted for by OSAS severity (first principal component; P = 0.03) (Table 2). LES was also correlated with OSAS severity (first component: P < 0.001) and with age and BMI (third component, P = 0.003) (Table 2). The eyelid laxity severity grade was correlated with OSAS severity (P = 0.047) and age and BMI (P = 0.002).
|Eyelid laxity severity||Comp 1||0.17||0.08||2||0.047|
Neck circumference only showed a trend to be larger in patients with FES (P = 0.08), but this latter variable was not included in the PCA model due to a high number of missing values (47) (Table 2).
This is one of the first studies to demonstrate the association between OSAS severity and eyelid hyperlaxity including FES. The current findings are of particular value, as our study is the only one with larger sample size of consecutive non-treated patients and with a reference method for OSAS diagnosis while controlling for confounding factors. Previous reports suggesting this association are less conclusive due to the small number of patients (Ezra et al., 2010; McNab, 2007; Mojon et al., 1999; Waller et al., 2008), inclusion of continuous positive airway pressure (CPAP)-treated patients (Kadyan et al., 2010; Mojon et al., 1999) or lack of systematic objective assessment of sleep apnoea syndrome (Ezra et al., 2010; Kadyan et al., 2010).
FES was observed in 26% of our patients affected with OSAS, as defined by an AHI more than 5 h−1. This result is comparable to the frequency (31.5%; AHI cut-off of 10 h−1) reported in the only other study performed on a comparable sample size, although this latter work included OSAS patients under CPAP treatment (Kadyan et al., 2010). Regarding LES, the prevalence in patients without OSAS was relatively high, although comparable to other studies for the OSAS group (44%, for an estimation between 29.5 and 65%) (De Groot, 2009). More interestingly, the occurrence in our population of FES and LES increased with AHI severity, and severe OSAS seems to be critical in determining higher FES and LES prevalence. This observation is strengthened by the result from the multivariate analysis showing that FES, LES and eyelid laxity grade correlate with OSAS severity. The FES and LES prevalence observed in our subjects without OSAS seemed to be high in comparison with the few reports available from the literature (Kadyan et al., 2010). The absence of a control group in our study raises the question of whether our data in the non-OSAS patients are representative of the general population, in which FES prevalence remains undetermined. In our study, gender, age, BMI, hypertension and diabetes also increased with OSAS severity within a range suggesting that our subject sample is a good reflection of what would be observed in the general population. The subjects were not age-, sex- or BMI-matched with controls. They were recruited from the consultation service for sleep disorders, including a high proportion of snorers, even though snoring and FES have not been shown to be associated. Moreover, the FES prevalence was not different between OSAS and non-OSAS patients based on an AHI cut-off of 5 h−1. However, a cut-off of five is very low, and has little medical relevance. It is generally acknowledged that severe patients with an AHI >30 h−1 have a higher risk of functional repercussions (He et al., 1988) and should be treated. Interestingly, FES and LES prevalence are higher in the severe group than in subjects without OSAS or in patients with an AHI <30 h−1. These results suggest that the risk for developing FES and eyelid laxity increases primarily in the case of severe OSAS.
Another limitation to the present study is the lack of quantitative measurements, although it included semi-quantitative clinical criteria. The prevalence of LES in our non-OSAS group is therefore not comparable with other studies that used quantitative measures (Karger et al., 2006), pointing to the need for a consensus on a standardized method for measuring eyelid laxity. This limitation, however, does not seem to influence significantly the assessment of eyelid laxity as the prevalence of LES in our OSAS group stays within the same range of those published previously (van den Bosch and Lemij, 1994; De Groot, 2009; Robert et al., 1997; Shah-Desai et al., 2004). Loss of eyelid elasticity is also associated with OSAS in the present study, as we show an association between severities in both, independently of confounding factors.
Little is known about the risk factors that may account for the occurrence of FES. Gender is one of the first described, as it is observed more frequently in men (Culbertson and Ostler, 1981; Dhillon et al., 2007; Robert et al., 1997). Our study is inconclusive in this way, as only four women had severe OSAS, and none display FES. Keratoconus is also known to be associated with FES (Burkat and Lemke, 2005; Culbertson and Tseng, 1994; Fowler and Dutton, 2010; Mojon et al., 1999), but was not observed in our population, in accordance with the low prevalence of the disease. Other associations described previously for FES were observed in our population, such as with BMI and age, and there was a correlation between gender and eyelid laxity grade (Culbertson and Ostler, 1981). The statistical model we used to analyse risks factors for FES took into account confounding factors and collinearities between variables. Previous publications have reported on the association between FES and a single OSAS indicator, such as AHI (Karger et al., 2006; Mojon et al., 1999) or oxygen desaturation index (Kadyan et al., 2010). The methodology used here allows us to show an association between FES or LES and OSAS as a whole syndrome, and not only with a specific indicator. We also added to the model classical parameters for characterizing respiratory events or related-hypoxic consequences, even though polysomnographic evaluations, including arousal, would have given more precision. Finally, the comprehensive statistical analysis conducted in our study confirms the correlation between FES and OSAS severity, and suggests that OSAS might be an independent risk factor for eyelid hyperlaxity, including FES.
The pathophysiological mechanisms underlying this extremely high frequency of FES in severe OSAS remain unclear, and may be a more general problem of overall tissue elasticity. In this regard, further studies should include specific evaluation of floppy uvula to look for correlation with FES.
More specifically, a lack of elasticity produces FES and may contribute to the occlusive mechanism of the oropharynx during apnoeas. Elastic fibres are a key component in providing these tissues with specific mechanical properties such as resilience and deformability. A reduced amount of tarsal elastin in FES patients has also been reported (Netland et al., 1994). Moreover, Schlötzer-Schrehardt et al. (2005) reported that patients with FES had an up-regulation of metalloproteinase, enzymes that are known to degrade extracellular tissue, especially elastic fibres. The activity of those enzymes is known to be influenced by hypoxic stress and could thus be modified in OSAS patients (Tazaki et al., 2004). Therefore, hypoxia caused by the respiratory events could eventually lead to an increased metalloproteinase activity and a subsequent loss of elastic fibres. This hypothesis might explain the possibility of improvement of FES under CPAP treatment (McNab, 2000) and is in line with our observation that oxygen desaturation is a major factor in the PCA analysis, whereas snoring alone showed no correlation with FES. Systematic realization of biopsies in future studies may greatly improve the understanding of the tissues changes associated commonly with FES and OSAS. Furthermore, additional mechanisms may exist, including the degeneration of elastic fibres, as age appears to be an important risk factor for OSAS and FES. The role of amyloidosis in FES pathogenesis has also been suggested previously (Carbone et al., 1985; Kiuru et al., 1999), although no subjects in our study exhibited this problem. Further research is needed to improve our understanding of the pathogenesis of both disorders.
Given the strength of the association between FES and OSAS severity, the current findings raise the questions of: (i) whether OSAS screening should be performed systematically in patients affected with FES, and conversely, (ii) whether eyelid laxity should be evaluated in patients with severe OSAS. In any case, sleep physicians, primary care physicians and ophthalmologists should all be aware of this association and thus test for OSAS or FES while continuing to direct the management and treatment of their patients. Searching for FES might be especially critical in OSAS patients treated with CPAP and who display ocular side effects. Indeed, an airflow leak directed to an incomplete nocturnal palpebral occlusion may lead to conjunctival oedema. Moreover, previous reports (McNab, 2000) have suggested the possibility of eyelid laxity improvement under CPAP chronic treatment. Thus, the question should be addressed further with the extension of our study to a larger set of subjects, including a follow-up under CPAP treatment.
The present findings provide evidence for a significant association between OSAS severity and FES and suggest that severe OSAS may be an independent risk factor for FES and LES. These results raise the possibility that a common vulnerability of elastic tissues underlies both disorders. Further studies on larger samples and with longitudinal design are required to elucidate the relationship between both conditions. Ultimately, treatment of OSAS may reduce eyelid laxity, and thus have beneficial effects on chronic papillary conjunctivitis outcome and CPAP tolerance.
The authors are particularly grateful to Bernadette Kowalski for the double scoring of each respiratory polygraphy. We acknowledge Resmed for the provision of the Embletta© portable diagnostic system during the study.
- The eye in sleep apnea syndrome. Sleep Med., 2006, 7: 107–115. , and
- American Academy of Sleep Medicine. International Classification of Sleep Disorders: Diagnostic and Coding Manual, 2nd edn. American Academy of Sleep Medicine, Westchester, IL, 2005.
- The lax eyelid syndrome. Br. J. Ophthalmol., 1994, 78: 666–670. and
- Acquired lax eyelid syndrome: an unrecognized cause of the chronically irritated eye. Ophthal. Plast. Reconstr. Surg., 2005, 21: 52–58. and
- Sleep apnea in amyloidosis. Chest, 1985, 87: 401–403. , and
- The floppy eyelid syndrome. Am. J. Ophthalmol., 1981, 92: 568–575. and
- Corneal disorders in floppy eyelid syndrome. Cornea, 1994, 13: 33–42. and
- Eye diseases in patients with sleep apnea syndrome: a review. Bull. Soc. Belge Ophtalmol., 2009, 312: 43–51.
- Sleep-disordered breathing and effects on ocular health. Can. J. Ophthalmol., 2007, 42: 238–243. , and
- The associations of floppy eyelid syndrome: a case control study. Ophthalmology, 2010, 117: 831–838. , , , , and
- Floppy eyelid syndrome as a subset of lax eyelid conditions: relationships and clinical relevance (an ASOPRS thesis). Ophthal. Plast. Reconstr. Surg., 2010, 26: 195–204. and
- Mortality and apnea index in obstructive sleep apnea. Experience in 385 male patients. Chest, 1988, 94: 9–14. , , , and
- for the American Academy of Sleep Medicine. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications, 1st edn. Academy of Sleep Medicine, Westchester, IL, 2007. , , , and
- Ocular findings in sleep apnoea patients using continuous positive airway pressure. Eye (Lond), 2010, 24: 843–850. , , and
- Prevalence of floppy eyelid syndrome in obstructive sleep apnea-hypopnea syndrome. Ophthalmology, 2006, 113: 1669–1674. , , et al.
- Obstructive sleep apnoea in heretitary gelsolin-related amylosis. J. Sleep Res., 1999, 8: 143–149. , and
- Floppy eyelid syndrome and obstructive sleep apnea. Ophthal. Plast. Reconstr. Surg., 1997, 13: 98–114.
- Reversal of floppy eyelid syndrome with treatment of obstructive sleep apnoea. Clin. Exp. Ophthalmol., 2000, 28: 125–126.
- The eye and sleep apnea. Sleep Med. Rev., 2007, 11: 269–276.
- Eyelid, conjunctival, and corneal findings in sleep apnea syndrome. Ophthalmology, 1999, 106: 1182–1185. , , et al.
- Histopathologic features of the floppy eyelid syndrome. Involvement of tarsal elastin. Ophthalmology, 1994, 101: 174–181. , , and
- Frequent loss of nyctohemeral rhythm of intraocular pressure restored by nCPAP treatment in patients with severe apnea. Arch. Ophthalmol., 2010, 128: 1257–1263. , , , , and
- Eyelid hyperlaxity and obstructive sleep apnea (O.S.A.) syndrome. Eur. J. Ophthalmol., 1997, 7: 211–215. , , and
- The pathogenesis of floppy eyelid syndrome: involvement of matrix metalloproteinases in elastic fiber degradation. Ophthalmology, 2005, 112: 694–704. , , , , and
- Lax eyelid syndrome or ‘progeria’ of eyelid tissues. Orbit, 2004, 23: 3–12. , and
- The visual scoring of sleep in adults. J. Clin. Sleep Med., 2007, 3: 121–131. , and et al.
- Increased levels and activity of matrix metalloproteinase-9 in obstructive sleep apnea syndrome. Am. J. Respir. Crit. Care Med., 2004, 170: 1354–1359. , , et al.
- Sleep disorders and the eye. Mayo Clin. Proc., 2008, 83: 1251–1261. , and