Prevalence of meibomian gland dysfunction at a tertiary centre in Taiwan

ABSTRACT Clinical relevance Meibomian gland dysfunction and dry eye disease are closely related conditions that often coexist and can contribute to the development of each other. Understanding the similarities and differences between these diseases can assist clinicians in implementing effective treatments for both conditions in a clinical setting. Background Dry eye disease is a multifactorial disease of the tears and ocular surface. This study aimed to evaluate the demographic characteristics of patients with meibomian gland dysfunction in Taiwan, investigate the association between meibomian gland dysfunction and dry eye parameters, and estimate the prevalence of meibomian gland dysfunction among patients with dry eye symptoms at a tertiary referral centre. Methods This cross-sectional study enrolled patients aged ≥20 years who complained of dry eye symptoms and visited a tertiary centre between September 2019 and March 2020. The patients completed the Ocular Surface Disease Index and Standard Patient Evaluation of Eye Dryness questionnaires before undergoing ocular examination. The lipid layer thickness and meiboscale scores were recorded. In addition, the study measured tear film break-up time, examined corneal staining, and evaluated the number of meibomian glands yielding liquid secretions using a slit lamp. Results The study evaluated 202 eyes of 202 patients with a mean age of 58.05 ± 13.34 years. The prevalence of meibomian gland dysfunction was 93%. Mean meiboscale score and age were negatively associated, and tear film break-up time was positively associated with the mean number of meibomian glands yielding liquid secretions. Hyperlipidaemia and smoking were the main risk factors for Meibomian gland dysfunction. Conclusions The prevalence of meibomian gland dysfunction among patients with dry eye symptoms was extremely high at the tertiary centre, highlighting the strong relationship between Meibomian gland dysfunction and dry eye disease. Clinicians should consider meibomian gland dysfunction as a possible cause of dry eye.


Introduction
According to the definition of Tear Film & Ocular Surface Society Dry Eye Workshop II, dry eye disease (DED) is a multifactorial disease of the tears and ocular surface characterised by a loss of tear film homoeostasis, visual problems, and eye discomfort. 1 Evaporative dry eye, which accounts for two-thirds of all dry eye cases, is the most common type of DED, followed by aqueous-deficient dry eye.However, these two types of DED can co-occur. 2The Asia Dry Eye Society reported that most Asian patients with dry eye had the short tear film break-up time type of DED caused by an unstable tear film associated with evaporative dry eye, and a combination of symptoms and a short tear film break-up time is sufficient for definitive diagnosis of dry eye. 3eibomian glands are holocrine glands involved in the secretion of oil into the tear film, a process that occurs while blinking and reduces the evaporation rate of tears. 4eibomian gland dysfunction (MGD) is the term commonly used to describe the qualitative and quantitative changes in the levels of meibomian gland expression.It is one of the aetiological factors underlying evaporative dry eye. 4The prevalence of MGD is the highest in Asia, ranging from 31% in India 5 to 62.5% in Taiwan 6 ; in contrast, the prevalence of MGD is 20% in Western countries. 4This discrepancy in the prevalence rates can be attributed to differences in ethnicity and diagnostic criteria for MGD.Associations of subjective dry eye symptoms with objective signs of DED remain unclear, with no appropriate parameters to predict DED. 7Furthermore, many studies have reported that MGD shows no correlation with dry eye symptoms, although MGD is an important cause of evaporative dry eye. 7,8his study aimed to evaluate the demographic characteristics of patients with MGD in Taiwan, a country with a high prevalence of MGD, investigate the association between MGD and dry eye parameters, analyse the relationship between meiboscale scores and MGD, and estimate the prevalence of MGD among patients with dry eye symptoms at a tertiary referral centre.

Participants
This study was approved by the Institutional Review Board of the Tri-Service General Hospital (Institutional Review Board number: C202005070).It was conducted in accordance with the tenets of the Declaration of Helsinki.The requirement for obtaining patient consent was waived by the Institutional Review Board owing to the retrospective nature of this study.
All enrolled patients were at least 20 years of age and only the right eye was included for examination.Patients who visited a tertiary medical centre in Taiwan for dry eye symptoms between September 2019 and March 2020 were included.Patients were excluded if they had active ocular inflammation; had undergone ocular surface and eyelid surgery in the previous 3 months; had a history of ocular trauma, pinguecula, or pterygium; or had used any eye drops within the previous week.

Dry eye assessments
The basic personal and medical data of patients were recorded.Patients completed the Ocular Surface Disease Index and Standard Patient Evaluation of Eye Dryness questionnaires before undergoing ocular examinations.Ocular Surface Disease Index scores range from 0 to 100, and the patients were categorised into normal (0-12 points), mild (13-22 points), moderate (23-32 points), and severe (33-100 points) groups according to previously published guidelines. 9n contrast, the Standard Patient Evaluation of Eye Dryness scores range from 0 to 28, and scores greater than 6 indicate the presence of dry eye. 10

Lipid layer thickness
The lipid layer thickness of each patient was measured using LipiView II (TearScience, Inc., Morrisville, NC, USA), an interferometer introduced in 2011.Lipid layer thickness was recorded in nanometres.

Tear film break-up time
Using a micropipette, 2 μL of 1% fluorescein solution was administered to the inferior fornix of each eye.The patients were then asked to blink several times to ensure that fluorescein covered the entire corneal surface.They were asked to keep their eyes open, and the tear film break-up time was recorded as the number of seconds that elapsed between the last blink and the appearance of the first dry spot in the tear film.The measurement was repeated three times, and the average value was calculated.

Corneal staining
The Oxford grading scale was used to score corneal staining after the instillation of 1% fluorescein dye.Using the reference figure, the examiner graded corneal staining as: 0 = absent, I = minimal, II = mild, III = moderate, IV = marked, and V = severe. 11

Meibomian gland expression and quality
The Meibomian Gland Evaluator (Tearscience) was used to standardise the pressure applied to the eyelid and evaluate the number of meibomian glands yielding liquid secretion under slit-lamp evaluation, as recommended by Korb. 12,13his standardised method can provide an objective measure of meibomian gland activity.
Using the instrument, the three quadrants of the upper and lower eyelids (nasal, central, and temporal) were compressed.Each quadrant represented eight meibomian glands, with a total of 24 glands for each eyelid.Therefore, this study combined the upper and lower eyelid counts and determined the average number of meibomian glands yielding liquid secretion in each eyelid.The quality of the meibomian gland was also evaluated during slit-lamp examination.The inspection was performed by a single individual (Y.M.C) to avoid operational errors.

Meibography
The meibomian gland structure was evaluated with meibography (LipiView II; Tearscience), and the meibography findings were graded according to the criteria proposed by Pult and Riede-Pult. 14The meiboscale has five grades according to the area of meibomian gland loss: 0 = no atrophy, 1 = meibomian gland loss ≤25%, 2 = meibomian gland loss of 26-50%, 3 = meibomian gland loss of 51-75%, and 4 = meibomian gland loss >75%.Subsequently, the upper and lower eyelid scores were added, and the average meiboscale score was determined with a minimum of 0 points and a maximum of 4 points.

Diagnosis of MGD
The modified method proposed by Korb et al. in 2008 to diagnose MGD in this study was used. 13Korb et al. considered the presence of ≤6 meibomian glands yielding liquid secretion to be correlated with dry eye symptoms.Therefore, the definition of MGD in the study was an average of ≤6 meibomian glands yielding liquid secretion in the upper and lower eyelids and any quality problems in the eyelid margin.

Diagnosis of DED
According to the definition of Asia Dry Eye Society, DED was defined as dry eye symptoms (Ocular Surface Disease Index ≥ 13) combined with a tear film break-up time of ≤5 s. 3

Statistical analysis
The data are expressed as mean ± standard deviation and were analysed using SPSS version 23 (SPSS Inc., Chicago, USA).The independent t-test and one-way analysis of variance were used to assess differences in continuous variables between the groups.Pearson correlation was used to evaluate the association between the mean meiboscale score, age, mean number of meibomian glands yielding liquid secretion, and dry eye symptoms and signs.Statistical significance was defined as a P value < 0.05.

Demographic characteristics and medical history of included patients presenting with dry eye complaints
This study enrolled 202 right eyes of 202 patients (81% women and 19% men) whose mean age was 58.05 ± 13.34 years.The mean Standard Patient Evaluation of Eye Dryness and Ocular Surface Disease Index scores were 10.47 ± 5.15 and 31.18 ± 19.69, respectively.
Table 1 shows the demographic characteristics of the included patients.Among the included patients, 53% were aged ≥60 years, 97% had a tear film break-up time ≤ 5 s, and 93% had a mean meibomian glands yielding liquid secretion ≤6.Almost all patients had telangiectasia of the eyelid margin and plugging of the meibomian gland orifices.Thus, the prevalence of MGD in the present study was 93%.Table 2 presents the medical history of the included patients.Owing to retrospective data collection, the medical history of only 124 patients was recorded: 27%, 18%, and 8% had a history of hordeolum, hyperlipidaemia, and smoking, respectively.
Comparisons of demographic characteristics between age groups are also shown in Table 3.There was only a significant difference in the mean meiboscale between the age groups.Older patients had a higher mean meiboscale score in the study.However, although female patients had higher scores in the dry eye questionnaires and shorter tear film break-up time compared to male patients, this difference was not significant.

Association among the mean number of meibomian glands yielding liquid secretion, dry eye parameters, and medical history in patients with MGD
This study investigated the relationship between dry eye parameters and meibomian gland expression in patients with MGD.The mean number of meibomian glands yielding liquid secretion and age showed a significant negative correlation (Pearson r = −0.215,P = .003),indicating that the expression in the meibomian glands decreased with age.Furthermore, the mean meiboscale score was negatively associated with the mean number of meibomian glands yielding liquid secretion (Pearson r = −0.338,P < .001),illustrating reduced meibomian gland loss and high meibum expression.
Moreover, the corneal staining score showed no association with the mean number of meibomian glands yielding liquid secretion (Pearson r = −0.094,P = .201).In contrast, tear film break-up time was significantly positively associated with the mean number of meibomian glands yielding liquid secretion (Pearson r = 0.235, P = .001)(Table 4).Figure 1 shows a scatter plot of mean meibomian glands yielding liquid secretion in mean meiboscale scores, age, and tear film break-up time.
Data regarding medical history were available for only 110 patients.The mean number of meibomian glands yielding liquid secretion was significantly higher in those with hyperlipidaemia than in those without hyperlipidaemia (P = .04).Moreover, patients who did not smoke had a significantly higher mean meibomian glands yielding liquid secretion than those who smoked (P = .005)(Table 5).

Prevalence of DED among all included patients and patients with MGD
Based on the definition of Asia Dry Eye Society, DED was confirmed when the Ocular Surface Disease Index score was ≥13 and the tear film break-up time was ≤5 s; the results are summarised in Table 6.Most enrolled patients had an Ocular Surface Disease Index score of ≥13 (85%) and a tear film break-up time of ≤5 s (97%).The overall prevalence of DED among all patients presenting with symptoms was 82%, and a similar prevalence of DED was noted among patients with MGD (81%).In contrast, the prevalence of MGD in patients with definite DED was 92%.

Prevalence of symptomatic and asymptomatic MGD among patients with MGD
A previous study reported that Standard Patient Evaluation of Eye Dryness scores were correlated with meibomian gland function. 10Therefore, the Standard Patient Evaluation of Eye

Discussion
This study examined the demographic characteristics of patients with MGD in Taiwan to investigate the association between MGD and dry eye parameters.This is thought to be the first to evaluate the prevalence of MGD among patients with dry eye symptoms in Taiwan.The prevalence of MGD varies among different countries; the prevalence is higher in Asian countries than in Western countries. 5,6,15This study showed an extremely high prevalence of MGD (approximately 93%) among patients with dry eye, which is consistent with the results reported by Lemp et al. 2 It was hypothesised that these differences in prevalence were related to the different criteria for MGD used and different subgroups enrolled in different studies.The number of meibomian glands yielding liquid secretion was used to diagnose MGD because it was hypothesised that meibum secretion is a useful parameter for evaluating meibomian gland function.Moreover, the procedure is relatively simple and quick and can be easily performed on all patients.In addition, because the hospital was a tertiary medical centre, patients with relatively severe dry eye symptoms visited the ophthalmology department.
Many cohort studies have discussed the association between MGD and DED. 16,17MGD can influence disturbances in the lipid layer of the tear film, cause excessive tear evaporation, and eventually induce dry eye symptoms. 4It has been suggested that MGD and DED are different diseases with different aetiologies; however, the study showed a high prevalence of MGD among patients with DED.This finding could be attributed to the fact that the criteria of Asia Dry Eye Society were used to diagnose DED.The criteria of Asia Dry Eye Society mainly focus on a short tear film break-up time, which is an essential aetiological factor in Asian patients with dry eye.Moreover, MGD was the main aetiological factor underlying evaporative dry eye and was strongly associated with a short tear film break-up time.Therefore, the present study suggests that MGD is strongly associated with DED in Taiwan.
This study investigated the relationship between dry eye parameters and the number of meibomian glands yielding liquid secretion.A positive correlation was found between the number of meibomian glands yielding liquid secretion and tear film break-up time; a higher number of meibomian glands yielding liquid secretion was associated with a long tear film break-up time.This conclusion was reasonable because meibomian gland expression can increase the lipid component of the tear film and slow down the evaporation time on the corneal surface.However, the corneal staining score, an important parameter of DED, did not show a significant negative association with the number of meibomian glands yielding liquid secretion.
The present findings also support the exclusion of ocular surface damage from the criteria of Asia Dry Eye Society.These results explain how MGD and DED are highly associated, even though some studies have reported them as different diseases.Moreover, lipid layer thickness also showed no significant association with the number of meibomian glands yielding liquid secretion, which was the opposite of that in a previous article. 18Intentionally and forcefully blinking can increase the lipid layer thickness of the tear film.However, blinking force can vary from person to person.Therefore, lipid layer thickness is an objective parameter, but spontaneous blinking in patients is subjective.In addition, the examiner used the Meibomian Gland Evaluator (Tearscience) to standardise the pressure applied on the eyelid, which was more objective.
In contrast, this study found a significant negative association between the number of meibomian glands yielding liquid secretion and meiboscale scores.Thus, the greater the meibomian gland expression, the lower the loss of meibomian glands.Many studies have reported inconsistent results regarding this association.Robin et al. showed that the meiboscale scores of the MGD group were significantly higher than those of the non-MGD group. 14However, the results contradicted those obtained by Blackie et al., who suggested that meibomian glands may appear relatively normal on meibography without obvious atrophy but show low expression, implying non-obvious obstruction. 19These discrepant findings could be related to the differences in the MGD criteria and enrolled groups.
Ageing plays an important role in the pathogenesis of MGD.In an in vitro study, human meibocyte differentiation decreased and meibocyte cell cycling slowed down with ageing; these conditions were related to the development of MGD. 20Arita et al. found that the prevalence of MGD increased with age and was higher in men than in women. 17Similarly, this study found that meibomian gland loss increased with age, and the expression of meibomian glands also decreased with age.These findings suggest that ageing is a vital factor in the pathogenesis of MGD.
These findings also revealed sex-related differences in the incidence of MGD.Several studies have reported high rates of MGD in men 21,22 in contrast, some reported that the incidence of MGD did not show sex-related differences 23,24 or that the prevalence was higher among women. 8These sexrelated differences can be attributed to the effects of hormones such as androgens and oestrogens on the meibomian glands. 4Many experts have suggested that the loss of androgens influences the pathogenesis of MGD in male patients. 1evertheless, the higher prevalence of MGD in female patients may be associated with discrepancies in the definitions of MGD and DED. 8,25his study showed a high prevalence of MGD among women.The study hypothesised that this discrepancy is related to the high prevalence of DED in the enrolled patients since DED is more common in women than in men. 1 Moreover, the study included patients who complained of dry eye, and female patients are more likely to express their symptoms and seek help. 26However, this is just speculation, as the impact of hormones on DED and MGD have yet to be fully determined.
Many studies have reported that asymptomatic MGD is more common than symptomatic MGD. 21,23,27Viso et al. found that the prevalence of asymptomatic MGD was twice that of symptomatic MGD in Spain. 20Furthermore, Amano et al. revealed that the prevalence of asymptomatic MGD was higher than that of symptomatic MGD in Japan. 23However, the present results contradicted their findings; the prevalence of symptomatic MGD was thrice that of asymptomatic MGD.This discrepancy was considered to be related to the differences in the enrolment criteria.Previous studies enrolled participants who did not report any dry eye symptoms, whereas this study only included patients who experienced dry eye symptoms.
Over the last 10 years, several risk factors have been identified for MGD, including the use of contact lenses, 28 hyperlipidaemia, 29 use of eyeliners, 30 and smoking. 31,32herefore, the study also evaluated the correlation between these potential risk factors and the number of meibomian glands yielding liquid secretion.This study found that hyperlipidaemia and smoking were negatively associated with the number of meibomian glands yielding liquid secretion, consistent with the results of previous studies. 29,31,32he pathophysiology of MGD related to dyslipidaemia has been evaluated.The Apolipoprotein E knockout mice revealed the manifestations of MGD, which was associated with inflammation of meibomian gland acinar cells and microenvironment. 33Nevertheless, the mechanisms underlying the associations of MGD with smoking and MGD are not clearly understood, and multiple mechanisms may be involved.Except for hyperlipidaemia and smoking, the other parameters considered risk factors for MGD did not reveal any significant association with MGD.
Despite the detailed analyses, this study has some limitations.This was a retrospective study; therefore, some data may have been missing.The cross-sectional design precludes the identification of causality.Thus, longitudinal analyses are required to explore the association between dry eye parameters and MGD.Because of the inclusion criteria and retrospective design, participant enrolment had a selection bias, and the study population included a higher proportion of female patients than male patients.Therefore, these findings should be interpreted with caution.Because the number of enrolled patients was low, some associations could not be clarified.The study enrolled patients who complained of dry eye symptoms, which can cause bias in the analysis of the percentage of symptomatic and asymptomatic MGD.

Conclusion
This study showed a higher prevalence of MGD than a previous study conducted in Taiwan. 6Doctors should be aware of the possibility of MGD when patients complain of dry eye symptoms.This study also found a high prevalence of DED among patients with MGD; thus, this study highlights the strong relationship between MGD and DED.
Furthermore, meibomian gland expression showed a significant association with meiboscale scores; therefore, this scoring system can be used as a non-invasive tool to evaluate MGD.While risk factors such as hyperlipidaemia and smoking were found to be associated with MGD, the study did not identify other risk factors.Thus, larger studies are needed to clarify the relationship between MGD and known MGD risk factors and to identify other risk factors.

Table 1 .
Demographic characteristics of included patients.

Table 2 .
Medical history of included patients.

Table 3 .
Demographic characteristics of included patients between age groups.

Table 4 .
Association between mean number of MGYLS ≤ 6 and dry eye parameters in patients with MGD (n = 187).
mian glands yielding liquid secretion; OSDI, Ocular Surface Disease Index; SPEED, Standard Patient Evaluation of Eye Dryness; TFBUT, tear film break-up time.Figure 1. Scatter plot of the mean number of meibomian glands yielding liquid secretion and mean meiboscale scores (A), age (B), and the tear film break-up time (C).MGYLS, meibomian glands yielding liquid secretion; TFBUT, tear film break-up time.

Table 5 .
Association between mean number of MGYLS ≤ 6 and medical history parameters in patients with MGD (n = 110).

Table 6 .
Prevalence of DED among all included patients and patients with MGD.

Table 7 .
Prevalence of symptomatic MGD and asymptomatic MGD.