Share this article on:

Relation Between Upper and Lower Lids' Meibomian Gland Morphology, Tear Film, and Dry Eye

Pult, Heiko*; Riede-Pult, Britta H.; Nichols, Jason J.

doi: 10.1097/OPX.0b013e318244e487
Original Articles

Purpose. To analyze relations between upper lid (UL) and lower lid (LL) meibomian gland (MG) morphology and tear film and the MG criteria ability to predict dry eye.

Methods. MG, lipid layer, and non-invasive break-up time (NIBUT) were evaluated of the OD of 20 randomly selected subjects (female = 10; median age = 44.5 years, interquartiles = 39.5 to 55 years). Subjects were grouped into nine Ocular Surface Disease Index (OSDI)− and 11 OSDI+ by the OSDI. Non-contact infrared meibography and image analysis were performed to evaluate MG loss, MG thickness, and MG bent angle.

Results. MG loss (Pearson correlation; r = 0.647, p = 0.003) and MG bent angle (r = 0.489, p = 0.027) were significantly correlated between lids, but MG thickness was not (r = −0.059, p = 0.413). MG loss was significantly (t-test; p = 0.048) less in the UL (median = 26.9%; LL = 32.3%), thicker in the LL (p < 0.001) and were more bent in the LL (p = 0.001). MG loss was significantly correlated to lipid-layer thickness (r < −0.597, p < 0.003) and NIBUT (r < −0.453, p < 0.030), whereas MG thickness and bent angle of the UL only were related to NIBUT (r < −0.563, p < 0.018). Combining MG loss of both lids (linear regression analysis) resulted in the best predictive ability of OSDI± (area under the receiver operative characteristic curve = 0.929, p = 0.001).

Conclusions. MG scores between lids were significantly different but correlated. MG loss was significantly correlated to tear film characteristics including lipid layer thickness and stability. MG thickness and bent angle of the UL were related to NIBUT. The combination of both lids' MG loss showed best predictive ability of dry eye.


Dipl Ing (FH)


Dr. Heiko Pult - Optometry and Vision Research, Weinheim, Germany (HP, BHR-P), Contact Lens and Anterior Eye Research Unit (CLAER) (HP), and School of Optometry and Vision Sciences (HP, BHR-P), Cardiff University, Cardiff, Wales, United Kingdom, School of Optometry and Vision Sciences (HP, BHR-P), Cardiff University, Cardiff, Wales, United Kingdom and College of Optometry, The University of Houston, Houston, Texas (JJN).

Received September 2, 2011; accepted November 21, 2011.

Heiko Pult Steingasse 15 69469 Weinheim Germany e-mail:

Meibomian gland dysfunction (MGD) is a chronic and diffuse abnormality of the MG.1 It is one of the most common abnormalities in ophthalmic practice,2 and the major cause of lipid anomaly3 and consequently of the evaporative dry eye.1 Evaporative dry eye is probably the most frequent cause of dry eye.1,3,4 MGD may be accompanied by drop out of the MG,5 which is recommended to be evaluated by meibography.1

To our knowledge, Tapie6 was the first to describe meibography by infrared (IR) photography, and subsequently many other groups have used IR techniques710 in MG observation. However there are two different techniques: transillumination of the everted lid7,8,10 vs. direct illumination (non-contact) meibography.9,1113 In the transillumination method, the lid is everted over a fiber optic probe5 or light source8 whereas non-contact meibography9 consists of a slitlamp equipped with an IR charge-coupled device video camera and an IR transmitting filter9 to allow the observation of the everted lid without contact with the light source. Alternatively, a near adapted IR camera can be used without need of a slitlamp microscope, named the portable non-contact IR meibograph (PNCM).14

Most studies used meibography to classify the MG drop out of the lower lid (LL),7,8,10 whereas only two groups (McCann et al.5 and Arita et al.9,1113) appear to have investigated the upper lid (UL) MG drop out. To our knowledge, all these studies used subjective grading scales and only McCann et al.5 appears to have analyzed relations between MG drop out between lids. Additional criteria like gland thickness and bending were not investigated. Stasis in the gland ducts can be associated with enlargement,3 and we predict that this stasis may also be causing bending of the glands. The hypothesis of this study is that these meibography-based additional criteria are related to tear film and dry eye scores and might be able to improve predictive ability for dry eye.

The aim of this pilot study was to analyze relations between UL and LL MG morphology (MG drop out, gland thickness, and bending of glands) and tear film and dry eye scores using the PNCM and digital image analysis.14

Back to Top | Article Outline


Inclusion and Exclusion Criteria

Twenty subjects (female = 10; median age = 44.5 years, interquartiles = 39.5 to 55 years), were randomly selected from the patient pool of the Horst Riede GmbH, Weinheim, Germany for participation. Subjects were excluded if they had diabetes, recent ocular infections, seasonal allergies, any history of ocular surgery, use of any medication or eye drops known to affect the ocular surface, were a current or prior contact lens wearer, or were pregnant by self-report. All procedures were conducted in accordance with the Declaration of Helsinki (1983), and approval for the study was given by the Cardiff School of Optometry and Vision Sciences Ethics Committee. All subjects provided written informed consent before participating in the study.

Back to Top | Article Outline

Design and Outcomes

This was an observational, cross-sectional study. All objective measures described below were performed on the right eye of each subject. Two masked examiners conducted different portions of the examination. First, the tear film examiner was masked to subjective reports and meibography. Second, the examiner conducting meibography was masked to subjective reports and tear film examination.

Back to Top | Article Outline

Lipid Layer

Lipid layer was observed using a TearScope Plus (Keeler, Windsor, UK) and classified by lipid pattern appearance according to Korb et al.15

Back to Top | Article Outline

Non-Invasive Break-Up Time (NIBUT)

NIBUT was measured non-invasively using a TearScope Plus with a fine grid insert.16 NIBUT was the time measured, in seconds, between the full opening of the eyelids after a complete blink and the first break in the tear film (using the included stop watch of the TearScope Plus). Three consecutive readings were evaluated and the median was noted.

Back to Top | Article Outline


Meibography was performed using the PNCM.14 The PNCM is an IR CCD video camera (802CHA CCD; Shenzhen LYD Technology Co., Shenzhan, China) adapted for near observation. The camera was connected to a computer via a Video-to-FireWire Converter DFG/1394 to 1e (The Image Source Europe GmbH, Bremen, Germany) and photographs were captured by the software (IC Capture 2.0 and IC Imaging Control 3.1; The Image Source Europe GmbH). Photographs were then analyzed by ImageJ 1.42q (Wayne Rasband, National Institute of Health, Bethesda, MD). The area of MG drop out was measured and its relation to the total area was calculated.14 This factor was named MG loss.14 The area of gland drop out was defined by (1) the actual endings of the glands “A,” (2) the width of the area, defined to be between at least from the tear punctum “C,” and the temporal border defined to be to the most well visible tarsal conjunctiva of the everted lid “B,” and (3) the maximal depth of the area was estimated to be where glands would have ended in normal MG morphology “D” (Fig. 1). In the measurement of the total area, this marked area was used as baseline and only the border “A” was moved by mouse clicks to the lid margin. Even though the definition of borders “B” and “D” were less objective, this approach should not have impacted the results because the ratio between dropped-out area and total area was calculated both based on the same unchanged borders “B,” “C,” and “D.”14 Additionally, the width of the most prominent MG (ImageJ units) and the angle (in degrees) of the most bent MG were analyzed (Fig. 2). The criteria of MG thickness and MG bending were based on the morphology of the “worst case” gland. This approach was based on the assumption that the analysis of the “worst case” gland might be representative for the overall MGD in a patient.

Back to Top | Article Outline

Ocular Surface Disease Index (OSDI)

Each subject's subjective reports were evaluated after objective observation using a validated German translation of the OSDI questionnaire.17 Total OSDI scores were calculated as recommended by Schiffman et al.18 Subjects were grouped into two groups (OSDI− and OSDI+) using a cutoff value of 1518:

Back to Top | Article Outline

Statistical Analysis

The data were analyzed using SPSS 15.0 (SPSS, Chicago, IL). Data were tested for normal distribution by the Shapiro-Wilk test. Correlations of MG morphology criteria between lids, tear film, and OSDI were calculated by Pearson correlation, and differences between lids were analyzed by the paired t-test. The ability to discriminate between OSDI groups was analyzed by the receiver operative characteristic curve and area under the receiver operative characteristic curve (AUC). A linear regression analysis (backwards selection; probability of F removal: 0.05; outcome: OSDI; predictors: MG loss, MG thickness and MG bent angle) was conducted to analyze the best combination of measured MG criteria.

Back to Top | Article Outline


MG loss was significantly (t = −2.155, p = 0.048) less in the UL (median = 26.9%) compared with the LL (32.3%). MGs of the LL were significantly (t = −5.461, p < 0.001) wider than of the UL, whereas the MGs of the UL were significantly (t = 4.052, p = 0.001) more bent than that of the LL (Table 1).

Significant correlations were found between UL and LLs in MG loss (Pearson; r = 0.647, p = 0.003, Fig. 3) and bent angle (r = 0.489, p = 0.027) but not in MG thickness (r = −0.059, p = 0.413). MG morphology was moderately related to age, except for MG bent angle of the UL and area of loss of the LL (p < 0.029). No significant relations were found with gender (p > 0.228).

MG loss was significantly correlated to OSDI scores (UL: r = 0.732, p = 0.001; LL: r = 0.841, p < 0.001) (Table 2). MG thickness of the UL was significantly correlated to OSDI scores (r = 0.469, p = 0.033) but not with the LL (OSDI: r = 0.294, p = 0.104). MG bent angle of the LL was significantly related to OSDI scores (r = 0.458, p = 0.021) whereas the relation between the MG bent angle of the UL and the OSDI was significance (r = 0.400, p = 0.062).

MG loss (LL and UL) was significantly correlated to lipid layer thickness (r < −0.597, p < 0.003) and NIBUT (r < −0.453, p < 0.030). MG thickness and bent angle of the UL was related to NIBUT only (r < −0.563, p < 0.018) but not to lipid layer thickness (r = 0.200, p = 0.229 and r = −0.301, p = 0.129; MG thickness and MG bent angle, respectively).

Subjects were grouped into 9 OSDI− and 11 OSDI+ (OSDI−: median score = 2; interquartiles = 0 and 12.2; OSDI+; median scores = 22.5; interquartiles = 20.83 and 29.17). MG loss of the UL and the LL were significant discriminators of OSDI± (UL: AUC = 0.921, p = 0.005; LL: AUC = 0.909, p = 0.002) (Figs. 4 and 5). MG thickness and bent angle were not able to discriminate between groups (AUC ≤0.697, p ≥ 0.093) (Table 3). A regression analysis (backwards) was performed to assess best criteria combination (Fig. 6) as follows.

Back to Top | Article Outline

UL Criteria

Meibomian loss and tension (MGLT) formulas (UL):

Back to Top | Article Outline

LL Criteria

MGLT (LL) formulas:

Back to Top | Article Outline

Criteria of Both Lids (UL and LL)s

MGLT (UL and LL) formulas:

Best combination of MG criteria was the combination of UL and LL MG loss (AUC = 0.929, p = 0.001). The observed power calculated after study completion was (1 − β > 80%).

Back to Top | Article Outline


The area of dropped-out glands was significantly smaller in the UL than of the LL. This is in accordance to the findings of McCann et al.5 However, this group did not found these differences being significant when counting the number of dropped-out glands. Because the number of glands are different between the lids even in normal subjects,19 the computerized measurement of the area of MG drop out might be the better approach as it takes into account the area measured and density of gland within that area.

The MG of the LL was significantly wider than of the UL, which might be due to having less physical space because number of glands is higher in the UL. Interestingly, this is represented by a sketch of MG of Heinrich Meibom in the year 1666, shown in a publication of Knop and coworkers.20 The glands of the UL were more bent than those of the LL, which might be a more anatomical issue than a consequence of any dysfunction of these glands.

However, even though MG morphology between lids were significant different, there was significant relations between the lids in MG loss and bent angle but not in MG thickness. So any change in MG loss and bent angle of one lid should be recognizable in other lid too (LL <> UL).

MG loss of both lids is significantly correlated to lipid layer thickness and NIBUT, but only bent angle and width of the UL glands were significantly correlated to NIBUT. However, correlations between MG morphology and the tear film appears to be reasonable because MGD results in quantitative and/or qualitative changes of the lipid layer.1 The lipid layer protects the tear film aqueous against evaporation and helps stabilize the entire tear film.21

Loss of MGs was moderately related to age, which is assumed to be a risk factor for MGD.13,2225 MG loss observed by meibography might indicate gland atrophy.1,26 Gland atrophy is assumed to be the result of long-term effects of MGD.1,26 As MGD progresses, increased viscosity of the meibomian oil in the gland and/or hyperkeratinization of the gland ductus and orifices results in gland obstruction, stasis, and increased pressure in the gland, with dilation of glands followed by an atrophy of the gland acini.1,26 However, in terms of age, MG loss might be a result of chronic MGD and/or be secondary to other age-related biological effects such as the well-known decline in production of sex-steroid hormones or some other mechanism, all of which have to be determined in larger epidemiological studies.23

Significant correlations between MG morphology and dry eye scores were shown in this study. Additional to the correlations between MG loss of both lids and the OSDI, the LL's MG bent angle and the UL's MG thickness were significantly correlated to OSDI scores.

To assess the ability of the MG morphology to discriminate between normal and dry eye, this cohort of subjects were grouped into OSDI±. Even though some relations between MG morphology and tear film or OSDI were analyzed in this study, only MG loss was able to discriminate between normal and dry eye, when grouped by the OSDI. The median OSDI scores of the dry eye group of 22.5 represents a moderate to severe dry eye.18

Analyzing the best MG morphology criteria combination by regression analysis of the UL or LL always resulted in the need for all criteria (MG loss, bent angle, and thickness). However, the predictive ability of the OSDI was better when using MG loss of the UL or LL or combining both.

Even though it might be obvious to consider gland thickness, gland bending, and changes in the tear film as obvious MGD signs, the greatest association with dry eye was drop out of MG in this cohort. Only a minimal loss of 16.9% of the UL MG and 28.7% of the LL was the threshold to predict dry eye. According to Nichols et al.,10 the cutoff value of 16.9% would be a grade 2, 28.7%, a grade 3 (grade 1 = no partial glands; grade 2 = <25% partial glands; grade 3 = between 25 and 75% partial glands; and grade 4 = >75% partial glands). Transforming our data to the Arita scale9 both 16.9 and 28.7% would be a grade 1 (grade 0 = no loss of MG; grade 1 = area loss was less than one third of the MG area; grade 2 = area loss between one third and two thirds; and grade 3 = area loss more than two thirds). This might signal the importance of computerized grading, because these cutoff values might be not detectable in subjective grading using the published four-grade grading scales.9,10

Back to Top | Article Outline


MG morphology between UL and LL are significantly different but correlated, except MG thickness. The cutoff value of MGD loss to discriminate between OSDI+ and OSDI− was 16.9% at the UL and 28.7% at the LL. The additional MG criteria bent angle and thickness showed only some relations to symptoms and tear film and even though they might be caused by stasis in the MG, the impact on tear film and dry eye appears to be negligible. MG loss showed significant correlation to the tear film, whereas these correlations were better for the UL's MG loss than for the LL. Combining MG loss of the UL and LL improves predictive ability of dry eye.

Heiko Pult

Steingasse 15

69469 Weinheim



Back to Top | Article Outline


1. Nichols KK, Foulks GN, Bron AJ, Glasgow BJ, Dogru M, Tsubota K, Lemp MA, Sullivan DA. The international workshop on meibomian gland dysfunction: executive summary. Invest Ophthalmol Vis Sci 2011;52:1922–9.
2. Foulks GN, Bron AJ. Meibomian gland dysfunction: a clinical scheme for description, diagnosis, classification, and grading. Ocul Surf 2003;1:107–26.
3. Knop E, Knop N, Brewitt H, Pleyer U, Rieck P, Seitz B, Schirra F. [Meibomian glands: part III. Dysfunction—argument for a discrete disease entity and as an important cause of dry eye.] Ophthalmologe 2009;106:966–79.
4. Heiligenhaus A, Koch JM, Kruse FE, Schwarz C, Waubke TN. [Diagnosis and differentiation of dry eye disorders.] Ophthalmologe 1995;92:6–11.
5. McCann LC, Tomlinson A, Pearce EI, Diaper C. Tear and meibomian gland function in blepharitis and normals. Eye Contact Lens 2009;35:203–8.
6. Tapie R. [Bio-microscopical study of meibomian glands]. Ann Oculistique 1977;210:637–48.
7. Mathers WD, Daley T, Verdick R. Video imaging of the meibomian gland. Arch Ophthalmol 1994;112:448–9.
8. Yokoi N, Komuro A, Yamada H, Maruyama K, Kinoshita S. A newly developed video-meibography system featuring a newly designed probe. Jpn J Ophthalmol 2007;51:53–6.
9. Arita R, Itoh K, Inoue K, Amano S. Noncontact infrared meibography to document age-related changes of the meibomian glands in a normal population. Ophthalmology 2008;115:911–5.
10. Nichols JJ, Berntsen DA, Mitchell GL, Nichols KK. An assessment of grading scales for meibography images. Cornea 2005;24:382–8.
11. Arita R, Itoh K, Maeda S, Maeda K, Furuta A, Tomidokoro A, Amano S. Proposed diagnostic criteria for seborrheic meibomian gland dysfunction. Cornea 2010;29:980–4.
12. Arita R, Itoh K, Maeda S, Maeda K, Tomidokoro A, Amano S. Efficacy of diagnostic criteria for the differential diagnosis between obstructive meibomian gland dysfunction and aqueous deficiency dry eye. Jpn J Ophthalmol 2010;54:387–91.
13. Arita R, Itoh K, Inoue K, Kuchiba A, Yamaguchi T, Amano S. Contact lens wear is associated with decrease of meibomian glands. Ophthalmology 2009;116:379–84.
14. Pult H, Riede-Pult BH. Non-contact meibography: keep it simple but effective. Cont Lens Anterior Eye. 2011 . [Epub ahead of print]
15. Korb DR, Baron DF, Herman JP, Finnemore VM, Exford JM, Hermosa JL, Leahy CD, Glonek T, Greiner JV. Tear film lipid layer thickness as a function of blinking. Cornea 1994;13:354–9.
16. Guillon JP. Use of the Tearscope Plus and attachments in the routine examination of the marginal dry eye contact lens patient. Adv Exp Med Biol 1998;438:859–67.
17. Michel M, Sickenberger W, Pult H. The effectiveness of questionnaires in the determination of Contact Lens Induced Dry Eye. Ophthalmic Physiol Opt 2009;29:479–86.
18. Schiffman RM, Christianson MD, Jacobsen G, Hirsch JD, Reis BL. Reliability and validity of the Ocular Surface Disease Index. Arch Ophthalmol 2000;118:615–21.
19. Knop N, Knop E. [Meibomian glands. Part I: anatomy, embryology and histology of the meibomian glands.] Ophthalmologe 2009;106:872–83.
20. Pult H, Korb DR, Blackie CA, Knop E, Marx E. About vital staining of the eye and eyelids. I. The anatomy, physiology, and pathology of the eyelid margins and the lacrimal puncta by E. Marx. 1924. Optom Vis Sci 2010;87:718–24.
21. King-Smith PE, Hinel EA, Nichols JJ. Application of a novel interferometric method to investigate the relation between lipid layer thickness and tear film thinning. Invest Ophthalmol Vis Sci 2009;51:2418–23.
22. Den S, Shimizu K, Ikeda T, Tsubota K, Shimmura S, Shimazaki J. Association between meibomian gland changes and aging, sex, or tear function. Cornea 2006;25:651–5.
23. Schaumberg DA, Nichols JJ, Papas EB, Tong L, Uchino M, Nichols KK. The international workshop on meibomian gland dysfunction: report of the subcommittee on the epidemiology of, and associated risk factors for, MGD. Invest Ophthalmol Vis Sci 2011;52:1994–2005.
24. Hykin PG, Bron AJ. Age-related morphological changes in lid margin and meibomian gland anatomy. Cornea 1992;11:334–42.
25. Sullivan BD, Evans JE, Dana MR, Sullivan DA. Influence of aging on the polar and neutral lipid profiles in human meibomian gland secretions. Arch Ophthalmol 2006;124:1286–92.
26. Tomlinson A, Bron AJ, Korb DR, Amano S, Paugh JR, Pearce EI, Yee R, Yokoi N, Arita R, Dogru M. The international workshop on meibomian gland dysfunction: report of the diagnosis subcommittee. Invest Ophthalmol Vis Sci 2011;52:2006–49.

non-contact meibography; meibomian gland morphology; meibomian gland dysfunction; tear film; dry eye

© 2012 American Academy of Optometry