Side effects of topical ophthalmic therapy
`with anti-inflammatory steroids and [3-blockers
`
`Ion R. Polansky, MD
`
`Cellular Pharmacology Laboratories, University of
`California Medical Center, San Francisco, California, USA
`
`Anti-inflammatory corticosteroids and B-blockers remain two of the
`most widely prescribed classes of drugs in ophthalmology despite
`their well-known ocular and systemic side effects. For corticosteroids,
`studies of ocular drug levels and corneal anti-inflammatory assays
`have been used to compare the currently available therapeutic
`choices. The safety and efficacy of different corticosteroid formulations
`are being explored, including direct injection of corticosteroids into
`the vitreous and other locations, and the use of col|agen—shie|d
`drug delivery approaches. With the demonstration of effective
`antimicrobials
`for herpes keratitis,
`the clinical
`requirement
`for
`corticosteroids is being examined because of the concern for steroid
`complications. For ophthalmic B-blockers, adverse systemic side
`effects and possible ocular complications remain important
`issues.
`Different doses and formulations of B-blockers have been assessed,
`with lower-dose regimens and combinations of antiglaucoma agents
`under evaluation as means of maximizing efficacy while minimizing
`complications. Useful quantitative data regarding the potential of
`ophthalmic B-blockers for producing systemic side effects have been
`provided by B]- and B3-adrenergic receptor binding assays, whereas
`different mechanisms appear to be required to explain intraocular
`pressure effects. Additional laboratory and clinical evaluations should
`help in making rational therapeutic choices and in designing new
`products.
`
`Current Opinion in Ophthalmology 1992, 3259-272
`
`in evaluating the current literature concerning the side
`effects of ophthalmic c0I‘tic()steroicls and B—blockers.
`it was important to consider the evidence and opinv
`ions regarding efficacy versus safety concerns as well
`as potential mechanisms for the different responses.
`Defining the properties ofthe individual drugs may en-
`able rational choices concerning the use of one prod-
`uct over another in a given clinical situation. Articles
`that provide quantitative data on drug evaluations of
`pharmacokinetics and proposed mechanisms are also
`reviewed as a means of gaining an improved knowl-
`edge of drug effects in ocular tissues.
`
`reviews (including two of the publications reviewing
`ophthalmic corticosteroids and B-blockers l1,2"]) and
`citation information for them is given in the references
`and recommended reading list.
`
`Ophthalmic corticosteroids
`
`There has been considerable clinical and experimen-
`tal experience with the use of anti—inflammatory cor-
`ticosteroids in ophthalmology, which we reviewed in
`1984 I1]. At that time. we conducted a number of ba-
`
`This review primarily focuses on recent publications.
`and original articles in the earlier literature are usually
`not discussed in detail. However, the earlier articles are
`
`discussed in conjunction with the recent papers and
`
`sic studies on various ophthalmic steroids available to
`help explain differences in efficacy or side effects as
`mentioned in the review. Although some questions re-
`mained, this approach yielded useful information on
`
`Abbreviations
`GCMS—gas chromatography—mass spectrometry; HTM—human trabecular meshwork; IOP—intraocu|ar pressure.
`
`© 1992 Current Science ISSN 1040-8738
`
`259
`
`SENJU—M|TSUB|SH| 2019
`
`

`
`260
`
`Ocular manifestations of systemic disease
`
`observed steroid effects in the eye. Recently. papers
`have appeared that require comment because of the
`data and opinions presented l3'.4.S'.6.7,8l.
`
`their own complications. Topical eyedrop administra-
`tion is the major route considered in this review, al-
`though data on injectable preparations and new routes
`of delivery are considered briefly.
`
`Deumethasone
`1 I -phosphate
`
`21 -alr nltol
`
`Dexarnetllasone
`
`Fig. 1. Ophthalmic corticosteroids evaluated in g|ucocorti-
`coid receptor competitive binding assays. Human skin fi-
`broblasts were incubated with 5 nM of 3H-dexamethasone
`and varying concentrations of unlabeled steroid for 60 min-
`utes at 37'C, using Dulbecco's modified Eagle's medium in
`a 10% carbon dioxide incubator. Counts are expressed as
`the percentage of control binding (5 nM of 3H-dexametha-
`sone alone) for dexamethasone 21-alcohol compared with
`rlexamethasnne )1-phosphate. (From Polansky and Weinreb
`If I; with permission.)
`
`Current evaluations of ophthalmic steroids have also
`involved administration of available steroid formula-
`
`tions by injection or by using drug delivery systetns
`as a means of producing a greater effect, reducing the
`amount of steroid required. or both. Corticosteroid use
`with nonsteroidal agents and more effective antimicro-
`bial agents has also received attention as a means of
`producing effective therapy while potentially limiting
`ocular steroid complications.
`
`Elevated intraocular pressure (IOP), cataract formation,
`and the spread or masking of infections are some of the
`well-established side effects of these drugs. Informa-
`tion regarding potential mechanisms for steroid glau-
`coma and cataract is being uncovered using trabecular
`meshwork and lens cell cultures, which could help to
`explain prior clinical observations and to understand
`new therapeutic approaches. Other corticosteroid com-
`plications of potential concem are a delay in wound
`healing, a “rebound” reaction and corneal ulceration
`if topical steroids are stopped abruptly, and systemic
`effects with frequent dosing of corticosteroid eyedrops
`in children. Periocular steroids and direct injection into
`ocular tissues are considered for conditions that do not
`
`seem to respond to topical or systemic steroid therapy.
`but these uses have been shown to be associated with
`
`Efficacy
`Corticosteroids are temted glucocorticoids because of
`their mechanism of action through the glucocorticoid
`receptor to produce anti-inflammatory and other ther-
`apeutic effects. Measures of glucocorticoid receptor
`binding provide a means of comparing different cor-
`ticosteroids for potency when the drug is applied di-
`rectly to ocular tissues. As long as the corticosteroid
`reaches the tissue it is affecting in an active form, its
`ability to produce therapeutic effects (or side effects)
`is determined by its receptor-binding activity and by
`the specific influences it has on inflammatory cells and
`resident cell types in different ocular tissues. Pharma-
`cokinetic information and assays of inflammation in ex-
`perimental animal models help in evaluating the active
`drug which reaches the required site in the eye from
`different routes of administration.
`
`Suspensions versus solutions
`
`is appropriate to use an ophthalmic eye-
`Whether it
`drop as a solution rather than a suspension has been
`a long-standing area of controversy. Potential advan-
`tages of solutions over suspensions. including a greater
`reproducibility in the amount of drug administered and
`the desire of patients to use clear solutions have been
`emphasized in the past. The potential for overdos-
`ing or underdosing if the suspensions are not shaken
`has probably been lessened with small uniform par-
`ticles but remains a reasonable concern for some pa-
`tients. However. the efficacy of corticosteroid suspen-
`sions over solutions has been the major issue. The
`overriding clinical impression. bolstered by the quan-
`titative studies of corneal inflammation by Leibowitz
`and Kupferman l5l conducted several years ago, has
`been that the suspensions of 1% prednisolone acetate
`or 0.1% dexamethasone eyedrops (rather than the so-
`lutions made through 21-phosphate additions) repre-
`sented the best choice when maximal ocular anti-in-
`
`flammatory therapy was required.
`
`The possible choice of 1% prednisolone acetate sus-
`pensions over 1% prednisolone phosphate solutions
`was recently raised by Sousa l4°], who concluded that
`the use of the latter may well be justified over the
`former. He argues against overreliance on the early
`anti-inflammatory data of Leibowitz and Kupferrnan.
`which shows superiority of prednisolone acetate for-
`mulations, and uses pharmacokinetic evaluations con-
`ducted over the past few years by Olejnik and Weis-
`becker ISI and by Musson er al. [6°l as the primary ba-
`sis for his argument. Although these investigators pro-
`vide potentially interesting information, their data do
`
`

`
`Side effects of topical ophthalmic therapy with anti-inflammatory steroids and B-blockers Polansky
`
`261
`
`not appear to change the essential clinical or basic un-
`derstanding of the greater potency of the ophthalmic
`suspensions.
`
`In considering potential mechanisms for the reduced
`effect of phosphate derivatives of both dexamethasone
`and prednisone reported in the cornea. we examined
`whether this substitution was reducing the gluc0coni-
`coid activity of the steroids by using receptor binding
`as a direct measure of anti-inflammatory drug activity.
`Figure 1 presents receptor binding of the 21-phosphate
`derivative of dexamethasone versus binding of the 21-
`alcohol compound from our prior review Ill. The bind-
`ing of the 21-alcohol compound to the glucocorticoid
`receptor shows the expected high atfinity binding for
`this compound reported in other systems. which typi-
`cally shows an equilibrium dissociation constant near
`3 nM. '|'he 2.1-phosphate compound shows an activity
`approximately 50-fold less. Similar findings were ob-
`tained when the prednisolone phosphate versus alco-
`hol derivatives were tested (ie. the phosphate showed
`much lower binding affinity than the acetate). These
`data help to explain the decrease in topical effective-
`ness of the solutions in corneal inflammation. Given
`
`appropriate phosphatase activity in ocular tissues. the
`phosphate group probably can be removed effectively
`as the drug passes through the comea and into the
`deeper tissues of the anterior segment. producing the
`fully active alcohol compound.
`
`Musson 01 ul. l6°] did show that mainly prednisolone
`and a smaller amount of phosphate derivative are
`found in the anterior chamber after prednisolone phos-
`phate eyedrop administration. The aqueous humor lev-
`els of prednisolone found after multiple-drop admin-
`istration of 1% prednisolone phosphate and Wu pred-
`nisolone acetate products were interpreted as being
`similar. with the fluxes they described appearing rea-
`sonable. These findings were in agreement with prior
`in ritro cornea permeability studies of Hull 0! al. [7],
`which the authors cited. However. the findings do not
`appear to warrant a change in the recommendations
`made based on the Leibowitz and l\'upferI11an corneal
`inflammatory model. Musson el aI.'s study showed evi-
`dence of increased flux of prednisolone into the aque-
`ous humor with the acetate versus the phosphate coin-
`pound, and the presence of the prednisolone phos-
`phate in the aqueous humor certainly suggests that not
`all of the steroid is convened to the more active form in
`
`the cornea. Also. the single-eyedrop study by Schoen-
`wald and Boltntlik l8|. quoted by Musson er a1.
`|6'l.
`showed approximately 20% larger peak concentrations
`in the aqueous humor with the 1% prednisolone ac-
`etate formulation. which agrees with preliminary stud-
`ies conducted in our laboratory using glucocorticoid ta-
`dioreceptor assays. Although the area under the curves
`appears similar for both drugs. the higher peak level
`of prednisolone in the anterior chamber with 1% pred-
`
`nisolone acetate may help the drug reach more inac-
`cessible areas of inflammation in uveitis. However. be-
`
`cause of the questions raised regarding the clove oil
`model [6'|. it might be worthwhile to examine steroid
`effects in other models of corneal inflammation.
`
`Because the current pharmacokinetic models for oph-
`thalmic steroids. such as those presented by Oleinik
`and Weisbecker I5]. do not readily explain the observed
`increased pemieability of the prednisolone acetate sus-
`pension. additional pharmacokinetic factors probably
`need to be considered. Possible mechanisms for the
`
`effect could involve the persistence of steroid particles
`in the eye. which could provide surface driving force
`for drug to penetrate into the tissue. or the preservation
`of a saturated solution for a somewhat longer period in
`the tears. Within the comea. the relative anti-inflamma-
`tory effect could be more dramatic if the acetate group
`exists fora period or ifthere is a delay in the removal of
`phosphate as the drugs penetrate the cornea. We have
`observed that the 21-acetate derivatives of steroids ap-
`pear reproducibly higher in receptor-binding activity
`than the alcohol derivatives in preliminary studies; this
`higher activity may be an additional reason for the in-
`creased anti-inflammatory effectiveness of these drugs
`in the clove oil keratitis model.
`
`For now. there appears to be no reason for reconsider-
`ing the use of 1% prednisolone acetate formulations as
`the most effective anti-inflammatory therapy for acute
`inflammatory keratitis. Suspensions of prednisolone or
`dexamethasone may also be most appropriate for pro-
`viding maximum anti-inflammatory therapy for uveitis.
`but this has not yet been determined. It would be most
`useful if quantitative data systematically comparing dif-
`ferent steroid treatments in experimental uveitis were
`obtained and published. Olejnik and Weisbecker l5]
`and Musson at al.'s l6'l data could be used to support
`the use of l% prednisolone phosphate solutions for
`some cases of intraocular inflammation-if a patient has
`a problem with suspensions. but this should be done
`with sotne caution until tnore is known about the con-
`
`version of the phosphate derivative into the active al-
`cohol form in different pathologic conditions.
`
`Last year. McGhee 0! al. l9l published a gas chroma-
`tography—mass spectrometry (GCMS) study showing
`much higher prednisolone levels in the aqueous hu-
`mor after administration of 1% prednisolone acetate
`than after administration of other ophthalmic steroids.
`The data are difficult to interpret because some of
`the steroid aqueous humor levels do not appear to
`make sense based on comparisons with other stud-
`ies using different methods of steroid detection. The
`levels of prednisolone following topical administra-
`tion of 1% prednisolone acetate are in general agree-
`ment with studies conducted by others using radioac-
`tivity and by Ole]nik's group using high-performance
`
`

`
`262
`
`Ocular manifestations of systemic disease
`
`liquid chromatography methods (as well as glucocor-
`ticoid radioreceptor assays conducted in our labora-
`tory). However, the levels of prednisolone phosphate,
`dexamethasone, and fluorometholone are much lower
`than those detected with other assays. It
`is doubtful
`that the human and rabbit corneas would show such
`marked differences in steroid penetration, suggesting
`that alternative explanations are needed. One possibil-
`ity is that cenain metabolites of the active steroid are
`not detected for some reason by the GCMS approach
`employed. Perhaps more likely is that the standard-
`ization may have varied between steroid molecules.
`in any case.
`it would be important to correlate the
`different methods of ophthalmic research to benefit
`from the excellent sensitivity of the GCMS for aqueous
`humor steroid measurements. Perhaps rabbit studies
`can be used to standardize the assays in aqueous hu-
`mor samples following administration of prednisolone.
`dexamethasone. and fluorometholone ophthalmic eye-
`drops. If metabolism of lliiorometliolone in ocular tis-
`sues is occurring more rapidly in humans than in rab-
`bits. or if some other difference exists between this
`compound and other steroids, it could be quite impor-
`tant. Nevenheless. questions regarding standardization
`of G_CMS must be resolved before the evidence of the
`low aqueous humor levels of the different steroids re-
`ported by the authors is considered relevant.
`
`Steroid metabolism of fluorometholone versus
`dexamethasone
`
`Metabolism of steroids in the systemic tissues as well
`as the eye appears to be a major determinant of steroid
`effects and side effects [1,10l. In fact, we emphasized
`previously that a major part of the increased potency of
`steroids such as dexamethasone (and betamethasone,
`
`2 I -phosphate
`
`Denmcthasone
`
`Fig. 2. Aqueous humor glucocorticoid activity after ocular
`administration of two 25-iiL doses of 0.1% fluorometholone
`or 0.1% dexamethasone phosphate in albino rabbits. At the
`times indicated, 50-iii. samples of aqueous were removed
`and evaluated through a series of dilutions in a receptor-
`binding competition assay. (From Polansky and Alvarado
`[10]; with permission.)
`
`its equipotent stereoisomer) systemically is due to sub-
`stitutions onto the cortisol molecule. some of which
`decrease degradation and others of which increase
`receptor binding. By contrast, certain steroids were
`synthesized using progesterone rather than cortisol as
`their foundation (ie. fluorometholone and medrysone),
`and these agents appeared particularly susceptible to
`degradation. In the eye, both medrysone and fluo-
`rometholone were shown to have a lower propensity
`to raise IOP than expected on the basis of anti-inflam-
`matory effectiveness, but clinically they were not con-
`sidered to be as good as other ophthalmic steroids for
`uveitis, probably because of their inactivation due to
`metabolism.
`
`Fluorometholone has long been known to have a high
`topical activity in both the eye and the skin. We l1,l0]
`confirmed this difference in activity using g|ucocor-
`ticoid receptor-binding assays in our laboratory,
`in
`which fluorometholone binding approximated that of
`dexamethasone, which had an approximately 10-fold
`higher affinity than medrysone and cortisol. In addition
`to showing fluorometholone‘s higher receptor bind-
`ing, we also demonstrated that fluorometholone could
`exert a full glucocorticoid effect
`in human trabecu-
`lar meshwork cells [10]. These studies reinforced the
`view that susceptibility of the drug to metabolism was
`likely to be the major basis for its favorable “split” be-
`tween topical potency and a reduced propensity for
`intraocular complications. The earlier suggestion that
`fluorometholone would not penetrate into the aque-
`ous humor was disproved by the studies of Hull et al.
`[7]. in which fluorometholone was shown to penetrate
`the cornea as well or better than other steroids used in
`ophthalmic practice.
`
`Glucocoiticoid iadioreceptor assays on aqueous hu-
`mor show that fluorometholone penetrated well and
`actually produced higher active drug levels than did
`dexamethasone phosphate (Fig. 2); also, as shown,
`dexamethasone phosphate demonstrated a substan-
`tially longer half-life in the aqueous humor compared
`with fluorometholone. Because of the short half-life
`of fluorometholone and its rapid metabolism, fluo-
`rometholone would be expected to be less effective
`than other potent steroids in the treatment of intraoc-
`ular inflammation, but this limitation might be over-
`come if active drug levels were sustained through re-
`peated dosing or drug delivery approaches. Medrysone
`is also expected to show a favorable “split” in topi-
`cal anti-inflammatory activity versus intraocular side ef-
`fects based on its structure, although in viva and In
`ultra studies of its metabolism are yet to be reported
`in ocular tissues.
`
`On the other hand, it is possible that fluorometholone
`17-acetate derivative (fluorometholone acetate), which
`is being introduced into ophthalmic practice as a
`highly potent ophthalmic steroid for corneal inflamma-
`
`

`
`Side effects of topical ophthalmic therapy with anti-inflammatory steroids and |3-blockers Polansky
`
`263
`
`tion [11°], may not show the same decreased side ef-
`fects as fluorometholone. This possibility is based on
`metabolism studies we conducted [12] several years
`ago including comparisons of fluorometholone with
`fluorometholone acetate for In vtrro degradation (Un-
`published data). The acetate substitution is in the 17
`position for fluorometholone acetate (because steroids
`such as fluorometholone which are based on proges-
`terone are missing the 21 position). We do not yet
`know the effect of the substitution on potency, but sus-
`pect that it will inhibit fluorometholone degradation.
`This possibility needs to be explored in greater de-
`tail experimentally in Mao and in vltro and in head-
`to-head IOP raising studies comparing available flu-
`orometholone formulations with fluorometholone ac-
`etate. The relative advantage for IOP response reported
`for 0.1% fluorornetholone acetate compared with 0.1%
`dexamethasone phosphate by Stewart 9! al. [131 is cer-
`tainly not sufficient as it stands to draw conclusions rel-
`ative to fluorometholone. although the raw IOP data
`from the different patients the study reported on could
`be useful if they were presented. Until additional stud-
`ies are conducted to evaluate the advantages and dis-
`advantages of this compound in ophthalmology. fluo-
`rometholone acetate should probably be regarded as a
`potent steroid in which safety issues are unresolved.
`
`Metabolism may also be an important factor in the use
`of dexamethasone versus prednisolone suspensions in
`certain clinical settings. Prednisolone acetate (1%) is
`formulated at 10 fold higher concentrations than dex-
`amethasone alcohol (0.1%). whereas in receptor assays
`the difference is only fourfold to fivefold. In addition
`to having higher receptor binding. dexamethasone is
`expected to be substantially more resistant to degra-
`dation than prednisolone, and could have a relatively
`longer half—life in ocular tissues. This suggests that topi-
`cal dexamethasone might be useful when attempting to
`achieve maximal suppression of intraocular inflamma-
`tion in deeper ophthalmic tissues, in which metabolism
`could play a greater role. For similar reasons, topical
`dexamethasone derivates may have a proportionately
`greater effect on IOP and may be more likely to pro-
`duce other steroid-induced complications than pred-
`nisolone derivates. as mentioned earlier.
`
`Drug delivery approaches
`In our prior review, we emphasized that future devel-
`opments in drug delivery offered the possibility of im-
`proved efficacy, reduced side effects. or both for spe-
`cific ophthalmic corticosteroid indications [1]. At that
`time, there was already an awareness that new meth-
`ods for delivering ophthalmic steroids to specific ocular
`tissues could make a large impact in reducing the dose
`required, increasing therapeutic effectiveness, or both.
`Ointments and gels that prolonged Contact time of the
`drug with the cornea were under development to in-
`crease duration of topical corticosteroids. Controlled
`
`release systems such as the hydrocortisone insert and
`other methods were being considered to prevent the
`transient overdoses and underdoses that accompany
`normal eyedrop use. as explained in an earlier consid-
`eration of ophthalmic drug delivery by Shell [14].
`
`It was also clear that the type of vehicle as well as the
`drug employed could make a substantial difference, as
`borne out in the release profiles from topical ointments
`and gels.
`in addition, the vehicle appeared to have
`an influence on injectable steroid preparations. Water-
`soluble corticosteroid preparations for injections were
`short acting because they diffused from the depot more
`rapidly than did other steroid preparations, even when
`compounds such as dexamethasone were used, which
`were regarded as long acting on the basis of their re-
`sistance to metabolism when delivered as topical eye-
`drops. Less soluble steroid preparations had a longer
`duration of action, with triamcinolone acetonide show-
`ing very long-lasting effects. The duration of action of
`periocular steroids was known to vary depending on
`the anatomic site of the depot iniection and distribution
`into surrounding tissues.
`
`The idea of novel drug delivery systems may only now
`be making an impact, as discussed in a recent article
`by Lee [15“l concerned with the optimization of oc-
`ular drug delivery. Controlled release systems such as
`in situ gel-fonning systems could help to expand the
`clinical value of topical steroid therapy. With regard
`to the corneal route. the collagen corneal shield, origi-
`nally developed for use as a surgical dressing has been
`shown to be an effective means of enhancing the de-
`livery of corticosteroids and other compounds to tis-
`sues inside the eye [16,17]. Although Lee [15"] points
`out that not all compounds benefit from this approach
`and that the shield is not a true sustained release sys-
`tem, the collagen shield does appear to facilitate coni-
`costeroid drug delivery to intraocular tissues. Sawusch
`el al. [16] provided evidence of a substantial improve-
`ment in drug delivery effects when collagen shields
`were used for a variety of therapeutic agents, including
`prednisolone acetate. Hwang et al. [17] showed that the
`collagen shield can enhance delivery of dexametha-
`sone to a substantial degree; their data showed that
`the corneal and the aqueous humor levels achieved
`with the shield were significantly greater than those
`achieved with repeated use of topical eyedrops. Their
`data also emphasize the potential use of the collagen
`shield to administer steroids to less accessible areas, in-
`cluding the posterior tissues, as a means of minimizing
`the need for injections.
`
`Further development of drug delivery systems may
`provide new entries into practical ocular corticosteroid
`therapy. In formulating the products to be used in
`these systems, an understanding of the individual cor-
`ticosteroid properties (receptor activity, pharmacoki—
`
`

`
`264
`
`Ocular manifestations of systemic disease
`
`netics and pharmacodynamics. and susceptibility to
`metabolism) will play an important role.
`
`Combination protocols
`
`The use of steroids with antimicrobial agents has been
`a long-standing interest of a number of investigators
`in the treatment of ocular diseases, especially because
`of the need to maintain the integrity of ocular tissues.
`which steroids provide. while not increasing the spread
`of infections. As we reviewed previously ll]. fungal in-
`fections were one of the first systems in which the joint
`use of antimicrobials and steroids was demonstrated
`
`in experimental animal systems. Currently, there is a
`very large interest in examining antiherpetic agents for
`corneal inflammation, and the question has been raised
`as to the proper use of conicosteroids in view of the
`improved antimicrobials that are being developed and
`the concern for steroid side effects l18—2ll. For herpes
`zoster infections, the effectiveness of the agents em-
`ployed has been proposed to obviate the need for cor-
`ticosteroid therapy |22l. Studies are undenmly to help
`resolve these issues for individual clinical indications.
`
`Another approach to combination therapy involves the
`use of steroids with nonsteroidal agents, because these
`two types of agents act on different pans of the path-
`way for tissue injury, as reviewed recently by Flach
`l25'l. Nonsteroidal agents may work in cases where
`steroids do not. and the use of nonsteroidal agents
`could minimize the steroid dose required. This may
`provide a means of limiting the serious ocular com-
`plications due to high steroid doses. In addition, com-
`bined use of potent steroids and nonsteroidal agents
`might show improved efficacy.
`
`Ocular complications
`Local effects
`
`Steroid injections could also cause problems because
`of the preservatives used and because of the direct tox-
`icity of high steroid concentrations on cell types within
`the eye that may not be as resistant as the corneal
`and conjunctival epithelia [24—Z6,27’,28l. In the recent
`study by Kwak and D‘Amico l27°] that established a
`safety profile for dexamethasone phosphate injections
`into the vitreous, the authors did not consider the direct
`solvent effect of steroids at high concentrations. This
`effect appears a more likely cause of the toxicity than
`the glucocotticoid receptor-mediated responses they
`cite. The repeated injections that may be required, par-
`ticularly when short-acting preparations are employed,
`may produce patient discomfort and inflammatory in-
`sults at the injection site. There is also a risk of inad-
`vertent intraocular injection.
`
`Moderately soluble respiratory corticosteroids have
`previously been reported to raise IOP and induce
`
`steroid glaucoma, and the use of sparingly solu-
`ble preparations, such as triamcinolone acetonide,
`may have even more serious complications. Smith
`and Nozik I29] indicated that periocular methylpred-
`nisolone preparations could show steroid effects for
`longer than is generally appreciated. perhaps caused
`by inflammatory reactions at the site of injection due
`to commercially available formulations rather than by
`a true difference in steroid molecules.
`
`Elevated intraocular pressure and steroid glaucoma
`
`Conicosteroid-induced elevation of IOP, eventually de-
`veloping into a “steroid glaucoma." is a major con-
`cem, especially when suppression of ocular inflamma-
`tion requires long-term corticosteroid administration.
`Following initial case reports of elevated IOP during
`the course of steroid therapy for uveitis, the general-
`ized effect of these drugs to raise IOP was confirmed
`in population studies using 0.1% dexamethasone phos-
`phate (or betamethasone phosphate) eyedrops for sev-
`eral weeks. Certain subjects received special attention
`because they showed dramatic IOP elevations, some-
`times within 1 week. (“High responders" were reported
`with much greater frequency among patients with pri-
`mary open-angle glaucoma than among the normal
`population.) Longer durations of steroid treatment and
`higher concentrations of the drug appear to produce
`larger responses to a given steroid, as we reviewed pre-
`viously Ill.
`
`The ability to obtain quantitative information con-
`cerning the IOP response has enabled the evalua-
`tion of possible differences between the various oph-
`thalmic steroid preparations for their propensity to
`produce steroid glaucoma. Among the initial stud-
`ies were reports that 0.1% fluorometholone and 1%
`medrysone had significantly less effect on the aver-
`age IOP response and substantially lower concen-
`trations than 0.1% dexamethasone phosphate. Fluo-
`rometholone (0.1%) and medrysone (1%) showed 36%
`and 5% of the effect of 0.1% dexamethasone phos-
`phate on IOP. respectively. It was of interest that 1°/o
`prednisolone acetate also appeared to raise IOP signif-
`icantly less than 0.1% dexamethasone. This result goes
`along with the concept that prednisolone may be me-
`tabolized more readily than dexamethasone in ocular
`tissues. but it requires further confimtation. The recent
`paper by Leibowitz er al. I11°], which considered fluo-
`rometholone (0.1% and 0.25%) versus fluorometholone
`
`acetate (0.1%) and other steroids, suggested that fluo-
`rometholone acetate could have an advantageous split
`between benefit and side effects based on the authors‘
`anti—inflammatory model in the cornea. The efficacy of
`the fluorometholone acetate could be of value even if it
`is not as safe as fluorometholone, especially because its
`susceptibility to metabolism could be better than that
`of prednisolone.
`
`

`
`Side effects of topical ophthalmic therapy with anti-inflammatorl steroids and B-blockers Polansky
`
`265
`
`sparingly soluble lnjectable
`As mentioned earlier,
`preparations may release drug over a long period,
`which could present a substantial problem by produc-
`ing sustained IOP elevations. These elevations have
`been reported to occur even if the patients did not
`show a high response in topical eyedrop tests by Her-
`schler [30]. The use of triamcinolone acetate may be a
`particular problem because this steroid has a very high
`potency and is relatively inert to metabolism. Recently.
`we analyzed a steroid sample from a subject who re-
`quired surgical removal of the tissue at the site of tri-
`amcinolone injection because of persistently elevated
`IOP months after the use of the product. Although it
`has not been proven. a drug with such long-term re-
`lease has the potential for causing progressive catarac-
`tous changes.
`
`We have emphasized that the lower the effective con-
`centration of corticosteroid that is expected to reach
`the trabecular meshwork (and the greater the steroid's
`ability to be metabolized). the greater the chance of
`avoiding ocular side effects. The employment of more
`dilute steroid formulations when an inflammatory con-
`dition is under control is a potential means of reducing
`complications. To avoid complications, relatively weak
`steroids (eg, medrysone) may be considered for mild
`external inflammatory conditions, and the dilution of
`dexamethasone to a concentration of 0.001% is used
`
`clinically by some as another method of reducing cor-
`ticosteroid side effects. In some cases. application of
`steroids daily or even week