GOBLET CELL RESPONSE TO VITAMIN A TREATMENT FOR CORNEAL XEROPHTHikLMIA ALFRED SOMMER, M.D., AND WILLIAM R. GREEN, M.D. Baltimore, M1 yland

In 11 patients with active corneal , conjunctival bi­ opsies were obtained at various intervals after m&sslve-dose system­ ic vitamin A therapy. Goblet cells began tb repopulate the infero­ nasal quadrant of the conjunctivas after two weeks, and they reached normal concentrations after one month. Corneal healing proceeded much more rapidly than that, suggesting that a full complement of mucus-secreting conjunctival goblet -,Uls is not essential for ,stora­ tion and maintenance of normal corneal appearance.

By its effect on tear filir dynamics, loss guardian, we obtained 3 X 4 mm biopsy of mucus-secreting goblet cells is thought specimens of temporal (and usually l'er­ to contribute to the corneal changes ob- onasal) bulbar conjunz'tiva before therapy 14 ther­ served in vitamin A deficiei-y. In a or at a predetermined interval aftei, recent study ofpatients with mild vitamin apy began, or both. Posttreatment con­ A-responsive conjunctival xerosis, how- junctival specimens were always obtained Spei­ ever,' it often took longer for goblet cells from a previously unbiopsied . to return than it reportedly takes for mens were fixed and processed In routine stain and corneal xerophthalmia to heal." We fashion and stained with Dane's report the goblet cell response to with PAS to Identify goblet cells.' herein vitamin defi­ vitamin A therapy in patients with vita- All subjects had severe A min A-responsive vorneal xerosis and ul- ciency' and all responded to standard7 systemic vitamin A therapy. ceration. high-dose

SUBJECTS AND METHODS RESULTS Children with corneal xerophthpimia We obtained pretreatment conjanctival and conjunctival xerosis of similar s...eri- specimens from 17 patients. The most ty in both were randomly selected severe clinical lesion was ' xerosis for conjunctival biopsy. After recdving in six natients and corneal ulce;ation in informed consents from a parent or 11. Potreatment specimens were ob­ taine,' from nine of these same patients and from two additional patients. Accepted for publication May 13, 1982. We found no goblet cells in any of the From the International Center for Epidemiologic 17 temporal or 15 inferonasal conjunctival and Prventive (Dr. Sommer) and before vitamin A the Eye .,t-,oLaboratory (Dr. Green), Wilmer specimens obtained trsitute, jehas Hoj lns Medical Inbtitutions, B 1i- treatment. more, Ma yland, and 'ale Nutritional Blindness Pre- Goblet cells were not observed in the vention Project, D.ndung, Indonesia (Dr. Sommer). specimens obtained This report was prepared under Co-operative Agree- nasal conjunctival ment AID/DSAN-CA-07-931.045 between the one week after treatment, but were pres­ Ofe of Nutrition, USAID, and ICEPO. ent (inreduced numbers) inone of three M.D., Wil­ Reprint requests to Alfred Sommer, obtained at two weeks wer Institute, 600 N. Wolfe St., Baltimore, MD nasal specimens £t10. and (in normal numbers) in all nasal spec­ 213 @AMERICAN JOURNAL OF OPHTHALMOLOGY 94:213-215, 1982 214 AMERICAN JOURNAL OF OPHTHALMOLOGY AUGUST, 1982

TABLE PRESENCE OF GOBLET CELLS IN CONJUNCTIVAL BIOPSY SPECIMENS FROM PATIENTS WHO HlAD XEROPHTHALMIA AFTER VITAMIN A THERAPY

Weeks Proporton of Specimens Containing Goblet Cells After Temporal Quadrant Inferonasal Quadrant Treatment Present Series Combined Series* Present Series Combined Series* 1 0/1 0/4 0/1 0/2 2 0/3 0/4 1/3 2/4 3-4 1/1 1/2 - 1/2 5-8 3/3 4/4 3/3 4/4 9-12 1/3 1/4 3/3 5/5

*Present series combined with those of Sommer, Green, and Kenyon' and Sullivan, McCulley, and Dohlman.1

imens obtained after the fourh week Unfortunately, the central role that has (Table). The pattern of goblet cell return been assigned to goblet cells in this pro­ was similar, but more erratic, in the cess is inconsistent with the time-course temporal conjunctival specimens. of corneal healing and goblet cell return. Of the 11 cases biopsied after treat- As with the subjects in this study, mild to ment, corneal healing was apparent by moderate manifestations of corneal xe­ day 2 in six cases and by day 4 in all cases. rophthalmia (superficial punctate kera­ In five cases, healing was completed topathy, xerosis, and punched-out ulcera­ within one week; in another four cases it tion) respond rapidly to vitamin A was completed within two weeks. In the therapy. Healing is generally evident remaining two cases, with large areas of within one to four days and is complete necrosis that scarred slowly, the duration within one to two weeks." required for complete healing was not Goblet cells, however, return more ascertained, slowly. Their rate of reappearance in our DISCUSSION patients was consistent with previous case reports of corneal' ," and noncorneal' Biochemical studies and treatment tri- xerophthalmia. As the numbers of cases als indicate that corneal xerophthalmia is were small, we pooled these results for a direct consequence of impaired vitamin more precise analysis (Table). Goblet A metabolism. 7 ,' , The exact mecha- cells begin to reappear in the inferonasal nisms by which corneal xerosis and xe- quadrant between one Pnd two weeks rophthalmic ulceration or necrosis occur, after start of vitamin Atherapy, and they however, remain obscure. It is reason- do not reach their normal complement for able to suspect that loss of mucus- an additional two to three we,-ks, long producing goblet cells early in the dis- after coraeal healing is apparently 4 com­ ease may play an important role by plete. Restoration and maintenance of promoting tear-film break-up that causes normal corneal appearance seemingly localized drying and epithelial loss.'" In- does not require large numbers of con­ deed, some of the preulcerative and early junctival goblet cells. ulcerative corneal changes observed in This does not mean that mucus produc­ xerophthalmia resemble localized expo- tion and tear film stability are unimpor­ sure and dellen formation.1 tant. Possibly small numbers of poorly VOL. 94, NO. 2 VITAMIN A AND XEROPHTHALMIA 215 staining goblet cells, missed on routine 3. Vitamin ADeficiency and Xerophthalmia. Re­ early and participate port of a joint WHO/USAID (World Health Organi­ microscopy, return zation/US Agency for International Development) in the healing process; other sources of meeting. Geneva, World Health Organization, 1976, mucus, such as the lacrimal gland"2 and p. 24. nonbulbar-conjunctival 4. Sommer, A., Green, W. R., and Kenyon, nongoblet or K.R.: Bltot's spots responsive and nonresponsive to cells"8 may provide mucus after goblet vitamin A. Clinicopathologic correlations. Arch. cells disappear or before they return; the Ophthalmol. 99:2014, 1981. Tamba, T.: Vita­ film break-up time ob- 5. Sommer, A., Emran, N., and shortened tear min A-responsive punctate keratopathy in xeroph­ served in corneal xerosis may be a result thalmia. Am. J. Ophthalmol. 87:330, 1979. rather than a cause of corneal keratiniza- 6. Summer, A.: Nutriticnal Blindness. Xeroph­ York, Oxford Uni­ this rule out the potential thalmia and Keratomalacia. New tion. Nor does versity Press, 1982, pp. 68-69, 86-88, ard 188-195. importance of another local factor, aque- 7. Sommer, A., Tarwotjo, I., Muhilal, Djkunaedi, ous tear production, which is reduced E., and Clover, J.: Oral versus intramuscular vitamin Lancet 1:557, early in xerophthalmia and is affected by Ainthe treatment of xerophthalmia. both vitamin A and protein status." 8. Sommer, A., and Tjakrasudjatma, S.: Corneal Conjunctival specinens from the infer- xerophthalmia and keratomalacia. Arch. Ophthalmol. provided a more reliable 100:404, 1982. onasal quadrant 9. Sommer, A., and Muhilal: Nutritional factors ir.t index of goblet cell dynamics than those corneal xerophthalmia and keratomalacia. Arch from the temporal quadrant, probably Ophthalmol. 100:399, 1982. density of goblet 10. Sommer, A., Muhilal, and Tarwotjo, I.: Pro­ cellsebecause o of the presetrgreater inite inferoba tein deficiency and treatment of xerophthalmia. cells normally present in the inferonasal Arch. Ophthalmol. 100:785, 1982. quadrant" and the tendency for xeroph- 11. Sommer, A., Tjakrasudjatma, S., and panocular xe­ thalmic metaplasia to persist in the tern- Djkunaedi, E.: Vitamin A-responsive poral quadrant.' We are now evaluating 9:1630,rophthalmia 1978. in a healthy adult. Arch. Ophthalmol. goblet cell density in the inferonasal 12. Allen, M., Wright, P., and Reid, L.: The histochemical and organ quadrant as a physiologic index of vitamin human lacrimal gland. A A status. nol.culture 88:493, study 1972.of the secretory cells. Arch. Ophthal­ 13. Allansmith, M. R., Baird, R. S., and Greiner, conjunctivitis REFERENCES J. V.: Density of goblet cells in vernal and contact lens-associated giant papillary conjuncti­ 1981. Sullivan, W. R., McCulley, J. P., and Dohi- vitis. Arch. Ophthalmol. 99:884, 1. N.: Tear production man, C. H.: Return of goblet cells after vitamin A 14. Sommer, A., and Emran, xerophthalmia. Am. J. Oplh­ therapy in xerosis of the . Am. J. Ophthal- in vitamin A-responsive 75:720, 1973. thalmol. 93:84, 1982. mol. of the 2. Dohlman, C. H., and Kalevar, V.: Cornea in 15. Kessing, S. V.: Mucus gland system anatomical hypovitaminosis A and protein deficiency. Isr. J. conjunctiva. A quantitiative normal 1968. Med. Sci. 8:1179, 1972. study. Acta Ophthalmol. 95(suppl.):4, FACTORS THAT INFLUENCE THE EFFICACY OF rOPICAL GENTAMICIN PROPHYLAXIS FOR EXPERIMENTAL PSEUDOMONAS KERATITIS KAMAL F. NASSIF, M.D., STARKEY D. DAVIS, M.D., ROBERT A. HYNDIUK, M.D., LAIUE D. SARFF, M.D., AND KATHY S. DIvINE, B.A. Miwaukee, Wiscon~n

We evaluated the efficacy of topical gentamicin prophylaxis for experimental Paeudomonaakeratitis in rabbits. by applying gentamicin in concentrations of 0.3 and 4 mg/100 ml, in both solution and ointment vehicles, one or four hours after a superficial stromal scratch was infected topically with Pseudomonasorganisms. Under these conditions the most effective prophylaxis was that given early, in solution, and in high concentration.

Antibiotic prophylaxis is important in in vitro by agar dilution, with a minimal recently acquired corneal abrasions and inhibitory concentration of 2.5 Rg/ml.3 lacerations. Although much experimental We anesthetized New Zealand white work has been done on the treatment of male rabbits weighing 2 to 3 kg each and keratitis, we are not aware of any study made a 3-mm linear scratch through the that evaluated the factors that influence corneal epithelium into the superficial the efficacy of antibiotic prophylaxis. We stroma with a sharp needle. We then studied the efficacy of a single application infected the topically with three of antibiotic in two concentrations and in drops of undiluted, overnight broth of P. two vehicles using a Pseudomonaskerati- aeruginosastrain 107 grown in Mueller- tis model in rabbits that closely resem- Hinton broth with Mg" + and Ca"+ bled the clinical situation. (1.34 X 1010 organismsml)."1 Treatment consisted of a single topical MATERIAL AND METHODS application of placebo or antibiotic. We Pseudornonas aeruginosa strain 107 applied two drops of solution or a strip of was found to be susceptible to gentamicin ointment, and we treated the rabbits ei­ ther one or four hours after infection. Gentamicin was supplied by the manufac­ turer in a 3-mg/ml solution, a 40-mg/ml solution, a 3-mg/g ointment, and a Accepted for publication May 28, 1982. 40-mg/g ointment. Thep dropper used in From the Departments of Ophthalmology (Drs. t Nassif and Hyndiuk) and Pediatrics (Drs. Davis and the study dispensed 26.7 drops/mi; thus, Sarff and Ms. Divine), Medical College of Wisconsin, two drops of the 40-mg/ml solution con­ Milwaukee, Wisconsin. This study was supported in tained 3 mg of gentamicin, whereas two part by National Institutes of Health Research Grant EY-0186W and Ophthalmic Research Center Grant drops of the 3-mg/ml solution contained EY.01931. 0.22 mg. Twenty strips of ointment, simi­ Reprint requests to Kamal F. Nassf. M.D.. lar to strips used in therapy, had an Cornea-External Disease Unit, Department of Oph­ thalmology, Eye Institute, 8700W. Wisconsin Ave., average weight of 40 mg. Therefore, each Milwaukee, WI =5322. 40-mg/g ointment strip contained about

216 CAMERICAN JOURNAL OF OPHTHALMOLOGY 94:216-219, 1982