Tear Film and Ocular Surface Changes after Closure of the Meibomian Gland Orifices in the Rabbit JEFFREY P. GILBARD, MD, SCOTI R. ROSSI, MS, KATHLEEN GRAY HEYDA, BA

Abstract: To determine whether meibomian gland dysfunction can increase tear film osmolarity and produce ocular surface changes analogous to those seen with lacrimal gland disease (keratoconjunctivitis sicca [KCS]), the authors closed the meibomian gland orifices in the right eyes of 11 rabbits by light cautery and studied the changes for 20 weeks. Tear film osmolarity was in­creased throughout the observation period. Conjunctival goblet cell density and corneal epithelial glycogen levels declined progressively. Closure of the mei­bomian gland orifices thus increased tear film osmolarity in the presence of normal lacrimal gland function and caused ocular surface abnormalities similar to KCS. Ophthalmology 96:1180-1186, 1989

In our previous rabbit models for keratoconjunctivitis sicca (KCS), 1•

2 ocular surface disease, as indicated by de­creases in conjunctival goblet cell density and corneal ep­ithelial glycogen, was directly proportional to increases in tear film osmolarity and time from creation of disease. The data suggested that ocular surface disease was sec­ondary to increases in tear film osmolarity or to the ab­sence of factors normally delivered to the ocular surface from the orbital lacrimal gland. Mishima and Maurice3

studied the tear film evaporation rate in rabbit eyes by following changes in corneal thickness, and they con­cluded that closure of the meibomian gland duct orifices by light cautery increases the evaporation rate.

To determine whether closure of the meibomian gland orifice could increase tear film osmolarity and whether the elevated osmolarity could produce the ocular surface disease of our KCS rabbit models in the presence of nor­mal lacrimal gland function, we closed the meibomian

Originally received: August 29, 1988. Revision accepted: March 10, 1989.

From the Cornea Unit, Eye Research Institute of Retina Foundation, and the Department of Ophthalmology, Harvard Medical School, Boston.

Supported in part by grant R01 EY03373 from the National Institutes of Health.

Reprint requests to Jeffrey P. Gilbard, MD, Eye Research Institute, 20 Staniford St, Boston, MA 02114.

1180

gland orifices by light cautery and studied the tear film and ocular surface as a function of time.

MATERIALS AND METHODS

The protocols described were approved by the Eye Re­search Institute's Animal Care and Use Committee. Eleven New Zealand white rabbits of either sex, weighing 2.0 to 2.5 kg, were anesthetized with intramuscular ket­amine ( 100 mgfkg) and xylazine ( 10 mg/kg). With mi­croscopic visualization the meibomian gland orifices in their right eyes were individually closed by applying an ACCU temp disposable cautery (Concept, Inc, Clearwater, FL). Heat was applied to each orifice for approximately 1 second. Orifice closure was confirmed with slit-lamp biomicroscopy by the absence of visible meibomian gland orifices, the absence of oil on the lid margin and in the tear film, and inability to express oil from the gland with digital pressure. The unoperated left eye in each rabbit served as a paired control.

Tear film osmolarity was measured weekly for 12 weeks and then essentially every other week for an additional 8 weeks. Slit-lamp examinations and Schirmer tests with proparacaine were done essentially weekly between 4 and 10 weeks postoperatively and then usually every other week until 20 weeks postoperatively. Five rabbits were euthanatized at 12 weeks and four rabbits at 20 weeks.

GILBARD et al • MEIBOMIAN GLAND ORIFICE CLOSURE

Fig 1. Rose bengal staining of the cornea. Left. 4 weeks after meibomian gland orifice closure, the right cornea, including the inferotemporal sector, stained diffusely. Right, contralateral control eye shows normal pattern for rabbit cornea.

15

r

__.,..... Me•bom•an gland onf•ces closed ··~·· Control

P<0.01 P< 0.05

·········+···j ········· ···· ... .. A··t··?· ········· · ··+

Time Post-op (weeks)

Fig 2. Schirmer test results with proparacaine anesthesia after meibomian gland orifice closure and in contralateral control eyes. Vertical bars in­dicate mean ± standard error of the mean.

Corneal epithelial glycogen levels and conjunctival goblet cell density were measured as described previously. 1

Briefly, the central and midperipheral corneal epithelium was removed for glycogen measurements, and 5-mm tre­phine biopsy specimens were removed from four con­junctival quadrants for goblet cell density counts. Then the deepithelialized cornea was removed with an 11-mm trephine, and peripheral epithelialized cornea and adja­cent inferobulbar conjunctiva with sclera were removed with sharp dissection, fixed in Karnovsky's fixative (2.5% glutaraldehyde and 2.0% paraformaldehyde in cacodylate buffer), stained with alkaline Giemsa, and studied by light microscopy. In addition, the lids were removed, fixed in Karnovsky's fixative, stained with alkaline Giemsa or he­matoxylin and eosin, and studied by light microscopy. An additional two rabbits were euthanatized at 4 weeks primarily to determine the presence or absence of lid in­flammation at this time. Corneal epithelial glycogen and

conjunctival goblet cell density were also studied in these rabbits.

All data are expressed as mean ± standard error of the mean (n = number) and were analyzed using Student's two-tailed t test for paired or unpaired data.

RESULTS

For the first 3 postoperative days, eschar typically was present at the cautery site. By 4 days the eschar had cleared, and, aside from closure of the meibomian gland orifices, slit-lamp examination was unremarkable. By 4 weeks postoperatively, however, slit-lamp examination showed abnormal rose Bengal staining of the cornea in­ferotemporally (Fig 1). By 8 weeks, dilation of the inferior tarsal meibomian glands and cyst formation were evident. By 20 weeks, meibomian gland cysts were visible in the superior tarsus also. These cysts had a predilection for the meibomian glands located temporally.

At 4, 5, 6, and 9 weeks postoperatively, mean Schirmer tests with proparacaine were significantly higher in the operated eyes than in the contralateral control eyes; from 10 weeks on, Schirmer test values in operated and un­operated eyes were not significantly different (Fig 2). Over the entire 20-week period, tear film osmolarity in the ex­perimental eyes was significantly higher than in the con­tralateral control eyes (Fig 3).

Conjunctival goblet cell density was decreased 5.1 ± 0.4% (n = 2), 8.6 ± 0.9% (n = 5), and 15.5 ± 0.5% (n = 4) relative to contralateral control eyes at 4, 12, and 20 weeks, respectively. The decrease was statistically signif­icant at both 12 and 20 weeks (P < 0.01) and was most prominent in the temporal quadrants (Fig 4). There was also a progressive decrease in corneal epithelial glycogen levels compared with contralateral controls; at 4 weeks glycogen levels had decreased 2.3 ± 0.2% (n = 2); at 12 weeks: 11.4 ± 0.6% (n = 5); and at 20 weeks: 21.3 ± 0.8% (n = 4). The decrease was statistically significant at both 12 and 20 weeks (P < 0.01).

1181

OPHTHALMOLOGY • AUGUST 1989 • VOLUME 96 • NUMBER 8

At 4 weeks postoperatively, meibomian gland ducts were not inflamed and only mildly dilated. At 12 weeks postoperatively, light microscopy showed many dilated meibomian gland ducts (Fig 5). Within the dilated glands and surrounding the meibomian glands, inflammatory cells could be seen (Fig 6). At 12 weeks, inflammatory cells extended into the tarsal conjunctiva (Fig 7) but were not seen in the bulbar conjunctiva. By 20 weeks, inflam­matory cells were also seen within the bulbar conjunctiva (Fig 8). At 12 weeks there was swelling of superficial tarsal conjunctival epithelial cells and some swelling of super­ficial bulbar conjunctival and corneal cells. The superficial epithelial abnormalities were slightly more pronounced at 20 weeks in both cornea (Fig 9) and conjunctiva.

DISCUSSION

We found that meibomian gland orifice closure in­creases tear film osmolarity in the rabbit in the presence of normal lacrimal gland secretion. Mishima and Maurice3 concluded that the tear film evaporation rate increases after meibomian gland orifice closure, and we attribute our observed increase in tear film osmolarity to increased evaporation. Rabbits with closed meibomian gland orifices, normal lacrimal gland secretion, normal or elevated Schirmer test results, and elevated tear film osmolarity had decreases in conjunctival goblet cell den­sity and corneal epithelial glycogen analogous to those seen in our two rabbit models for KCS. 1

•2 In those two

models, as in our current model with meibomian gland

110

100 ------ ---e c

90 0 (,.)

0 <f. 80

>. - 70 (/)

c Q)

0

"ai 60

(,.)

Qi 50 :0 0

C)

40

30 12 weeks

Time Post-op

1182

320

315

310

._.....,_ Me•bom•an gland onf•ces closed ··~· • Control

•• P< 0.01 ~ . ~~

kf.. l'J:. J, P.. .I .... 2" ··§·++·{ -r •• r· .. /{ ····.y.-···1 ··········?

3oo+-~~~,-~~~~~~T-~~~~~ 1 2 3 4 5 6 7 8 9 10 11 12 14 16 20

Time Post-op (weeks)

Fig 3. Tear film osmolarity after meibomian gland orifice closure and in contralateral controls. Vertical bars indicate mean ± standard error of the mean.

orifice closure, the average decreases in goblet cell density and in corneal epithelial glycogen level correlated with the average increases in tear film osmolarity and in time from creation of disease. The rabbits with meibomian gland closure had the smallest average increase in os­molarity and decrease in goblet cell density and corneal glycogen of the three models.

-Mean of four quadrants Quadrants: c:::::J Supero-nasal ~ lnfero-nasal

~ lnfero-temporal !ZZI Supero-temporal

20 weeks

Fig 4. Conjunctival goblet cell density after meibomian gland duct closure (MGDC) relative to contralateral con­trols.

GILBARD et al • MEIBOMIAN GLAND ORIFICE CLOSURE

Fig 5. Massively dilated mei­bomian gland duct on the left and moderately dilated duct on the right, 12 weeks after meibomian gland orifice clo­sure. Cellular debris can be seen within the ducts (alka­line Giemsa; original mag­nification, X 118).

Given the presence of normal lacrimal gland secretion in the rabbits with closed meibomian gland orifices, we now consider it unlikely that the decrease in goblet cell density or corneal glycogen in our models resulted from the absence of factors normally delivered to the ocular surface by lacrimal gland fluid. The data suggest that the

Fig 6. Inflammatory cells (arrows) within dilated mei­bomian gland duct and ad­jacent to meibomian gland parenchyma 12 weeks after meibomian gland orifice clo­sure (alkaline Giemsa; origi­nal magnification, X300).

ocular surface disease in KCS is secondary to increases in tear film osmolarity.

We do not know why goblet cell loss was most prom­inent in the temporal conjunctiva after closure of the meibomian gland orifices. Perhaps it is related to the ob­servation that the meibomian glands of the temporal lids

1183

OPHTHALMOLOGY • AUGUST 1989 • VOLUME 96 • NUMBER 8

became more bloated than those of the nasal lid, sug­gesting that they normally produce more oil.

Four weeks after orifice closure rose Bengal staining of the cornea was abnormal, and by 12 weeks, morphologic abnormalities were seen in the cornea by light microscopy. In our full KCS model, we did not see abnormal rose

1184

Fig 7. Inflammatory cells (arrows) within tarsal con­junctival epithelium 12 weeks after meibomian orifice closure (alkaline Giemsa; original magnifica­tion, X 1185).

Bengal staining until 48 weeks postoperatively and ab­normal corneal morphology until 52 weeks. Given the absence of inflammation in the lids 4 weeks postopera­tively, we cannot attribute this early staining to inflam­mation. We know from our earlier work with rabbit mod­els of aqueous tear deficiency that the staining at 4 weeks

Fig 8. Inflammatory cells (arrows) within the bulbar conjunctival epithelium and swelling of the superficial ep­ithelium 20 weeks after mei­bomian orifice closure (al­kaline Giemsa; original mag­nification, X 1185).

GILBARD et al • MEIBOMIAN GLAND ORIFICE CLOSURE

Fig 9. Corneal epithelium 20 weeks after meibomian ori­fice closure. The superficial epithelial cells are swollen. The epithelium is free of in­flammatory cells (alkaline Giemsa; original magnifica­tion, X 1185).

cannot be attributed to the effect of elevated tear film osmolarity. The basis for the inferotemporal staining that began 4 weeks postoperatively is unclear to us currently.

Meibomian gland orifice closure increases tear film os­molarity and, by 12 weeks postoperatively, causes a slowly progressive inflammatory process that is presumably due to retention of meibum within the glands. We cannot exclude a role for this inflammation in the ocular surface disease that develops in these rabbits, and we consider these rabbits a model for both meibomian gland dys­function and meibomitis.

Our study may shed light on the pathogenesis of mei­bomitis in humans. Previous thinking attributed mei­bomitis to bacterial infection, specifically from Staphy­lococcus aureus.4 Histopathologic and clinical studies in­dicated that meibomian gland duct orifice closure is a characteristic feature in meibomitis in humans. 5•

6 Our study suggests that orifice closure, with the resultant stasis of meibum within the gland, may account for the inflam­mation in the meibomian glands and conjunctiva in this disease.7 Infection with bacterial agents is not requisite for the development of this condition, explaining why Seal and co-workers8 were unable to detect any bacteri­ologic abnormalities in their meibomitis patients. Inves­tigators have· not shown basic differences in fatty acid composition in meibum from meibomitis patients and controls.9 The cause of meibomian orifice closure by epithelium in humans is unknown.

Our data also suggest that closure of meibomian gland orifices in humans may be a pathway for the development of increased tear film osmolarity in these patients, pro­viding a possible explanation for the frequent association

between meibomian gland disease and ocular surface dis­ease resembling that seen with lacrimal gland disease (KCS). 10 We have found increased tear film osmolarity in patients with meibomian gland orifice closure (unpub­lished data).

At 4 weeks postoperatively, before the onset of inflam­mation and after healing of the lid margin, and for the first 9 weeks, Schirmer test results in this study were higher in eyes with closed meibomian gland orifices than in con­trol eyes. Oil could not be seen on the lid margin, although it was always seen in normal rabbits. The oil on the lid margin acts as a barrier to aqueous tears. We postulate that in normal eyes, oil absorbed from lid margin onto the Schirmer strip as well as oil remaining on the lid mar­gin may retard the absorption of tear fluid by the Schirmer strip. This would account for the initial increase in Schirmer test results after meibomian orifice closure.

After 13 weeks Schirmer tests in operated eyes decreased and normalized, whereas inflammation and tear osmo­larity in these eyes increased. The increase in tear film osmolarity after 13 weeks suggests that tear secretion may have decreased at that time, perhaps from damage to ac­cessory lacrimal gland tissue located within the inflamed tarsal conjunctiva or perhaps from "fatigue" of a reflex mechanism. A decrease in tear secretion after this time would explain the changes in Schirmer test results. We do not suspect that the tear film evaporation rate increased after 13 weeks postoperatively.

The Schirmer test results may be influenced by two variables-tear secretion rate and meibum production. Decreases in tear secretion rate may decrease Schirmer test values, but decreases in meibum production may in-

1185

OPHTHALMOLOGY • AUGUST 1989 • VOLUME 96 • NUMBER 8

crease these values. This could account in part for the relatively poor reliability of the Schirmer test in predicting which patients have dry eye based on symptoms, 11 in­creased tear film osmolarity, 12 and ocular surface disease. 13

Other investigators have proposed that meibomian gland dysfunction can lead to dry-eye disease. 14

•15 We be­

lieve that our study provides the first objective evidence in support of this hypothesis. It suggests that meibomian gland dysfunction may be a common cause for dry-eye disease, and that increased tear film osmolarity may be the final common pathway by which decreased tear se­cretion or increased tear film evaporation, or both, results in the ocular surface disease of KCS.

REFERENCES

1. Gilbard JP, Rossi SR, Gray KL. A new rabbit model for keratocon· junctivitis sicca. Invest Ophthalmol Vis Sci 1987; 28:225-8.

2. Gilbard JP, Rossi SR, Gray KL, et al. Tear film osmolarity and ocular surface disease in two rabbit models for keratoconjunctivitis sicca. Invest Ophthalmol Vis Sci 1988; 29:374-8.

3. Mishima S, Maurice DM. The oily layer of the tear film and evaporation from the corneal surface. Exp Eye Res 1961; 1 :39-45.

4. Baum JL. Ocular infections. N Engl J Med 1978; 299:28-31.

1186

5. Gutgesell VJ, Stern GA, Hood Cl. Histopathology of meibomian gland dysfunction. Arn J Ophthalmol1982; 94:383-7.

6. Robin JB, Jester JV, Nobe J, et al. In vivo transillumination biomi· croscopy and photography of meibornian gland dysfunction: a clinical study. Ophthalmology 1985; 92:1423-6.

7. McCulley JP, Sciallis GF. Meibomian keratoconjunctivitis: oculo-dermal correlates. CLAO J 1983; 9:130-2.

8. Seal DV, McGill Jl, Jacobs P, et al. Microbial and immunological in­vestigations of chronic non-ulcerative blepharitis and meibornianitis. Br J Ophthalrnol 1985; 69:604-11.

9. Nicolaides N, Santos EC, Robin J, Smith RE. Meiburn lipids in rosacea blepharitis, and chalazia. ARVO Abstracts. Invest Ophthalmol Vis Sci 1983; 24(Suppl):78.

10. McCulley JP, Sciallis GF. Meibomian keratoconjunctivitis. Am J Ophthalmol1977; 84:788-93.

11. Rolando M, Refojo MF, Kenyon KR. Increased tear evaporation in eyes with keratoconjunctivitis sicca. Arch Ophthalmol 1983; 101: 557-8.

12. Gilbard JP, Farris RL. Tear osmolarity and ocular surface disease in keratoconjunctivitis sicca. Arch Ophthalmol1979; 97:1642-6.

13. van Bijsterveld OP. Diagnostic tests in the sicca syndrome. Arch Ophthalmol1969; 82:10-4.

14. Tiffany JM. The role of meibomian secretion in the tears. Trans Ophthalmol Soc UK 1985; 104:396-401.

15. Bron AJ, Mengher LS. Congenital deficiency of meibomian glands. Br J Ophthalmol1987; 71:312-4.