JOEL S. MINDEL
Table Of Contents
MECHANISM OF ACTION|
SIDE EFFECTS AND TOXICITIES
|Hydrocortisone (cortisol) is the major naturally occurring glucocorticoid
in humans. The blood level of hydrocortisone is determined by the rate
of pituitary adrenocorticotrophic hormone (ACTH) secretion, which
varies in a diurnal fashion. The daily rate of hydrocortisone secretion
by the adrenal cortex is approximately 20 to 25 mg/day. The highest
level is at approximately 6:00 AM, and the lowest is at approximately
midnight. Plasma levels of hydrocortisone fluctuate from a high of approximately 16 μg/100 ml
to a low of approximately 4 μg/100 ml.|
Patients prescribed a corticosteroid secrete less ACTH. This feedback suppression of the hypothalamus and pituitary gland reduces the pharmacologic effect from the exogenous corticosteroid until the dose exceeds that equivalent to the physiologic hydrocortisone secretion (i.e., exceeds 20 to 25 mg/day). A long-acting corticosteroid, such as dexamethasone, in a dose as low as 0.5 mg suppresses ACTH secretion and the plasma hydrocortisone level for 24 hours.
Hydrocortisone circulates bound to blood proteins. Transcortin is a corticosteroid-binding globulin with a high affinity, but there is relatively little of it. Albumin has a low affinity, but it is in such large quantity that its total binding capacity is high. Only that portion of the corticosteroid dose not bound to these proteins produces a physiologic or pharmacologic effect. As the dose increases, saturating the binding sites, the drug's effects are enhanced.
|MECHANISM OF ACTION|
|Ophthalmologists use corticosteroids primarily for their anti-inflammatory
and immunosuppressive activities. Glucocorticoids can inhibit many
aspects of inflammation both by increasing the transcription of anti-inflammatory
genes and by decreasing the transcription of proinflammatory
genes.1 For example, corticosteroids can reduce the synthesis of proinflammatory
agents such as prostaglandins. This can be achieved2,3 by inhibiting the transcription of the genes for phospholipase A2, the enzyme that cleaves arachidonic acid from phospholipids, and COX-2 and
by inhibiting the activation of phospholipase A2. Corticosteroids are capable of promoting the release of the phospholipase
A2 inhibitor, lipocortin, from leukocytes. It is unclear, however, how important
this mechanism is as the serum and plasma levels of metabolites
of prostaglandins are not reduced by treatment with large doses of corticosteroids.4 Other mechanisms that have been proposed, for which there is some evidence:|
|A common feature of the anti-inflammatory corticosteroids is an 11β-hydroxy
group. Prednisone does not have this structure and has no anti-inflammatory
activity until it is converted in the liver, with approximately 80% efficiency, to prednisolone.9 Patients with hepatic disease may have impaired conversion of prednisone
to prednisolone.10 Because of the hydroxylation requirement of prednisone, the permilligram
clinical efficacy of oral prednisone is three to four times that of
hydrocortisone, whereas the per-milligram clinical efficacy of prednisolone
is four to five times that of hydrocortisone.11 It also is why anti-inflammatory eyedrop preparations may contain prednisolone
but do not contain prednisone.|
Prednisolone may be unique among synthetic corticosteroids in that it competes with endogenous hydrocortisone for protein-binding sites on transcortin.12,13 Prednisolone binds with approximately 2.5 times greater affinity than cortisol and can displace it. Although prednisolone binds albumin with approximately 300 times stronger affinity than hydrocortisone, prednisolone does not displace cortisol. These considerations are important at low prednisone/prednisolone doses because it is only the unbound drug that is pharmacologic active. At low-prednisolone doses, the active unbound drug may be the weaker endogenous corticosteroid, hydrocortisone, that has been displaced from transcortin.
Many corticosteroids have low water solubility. Hence, they are prepared and administered in the form of water-soluble esters such as phosphates and hemisuccinates. In a real sense, these esters are prodrugs and must be hydrolyzed to their free alcohols for their full potency. Dexamethasone disodium phosphate ester is an example.14 The half-life of dexamethasone disodium phosphate ester in blood is approximately 5.4 minutes. The maximum plasma concentration of the free (alcohol) is achieved approximately 10 minutes after intravenous injection.15 The dexamethasone elimination half-life is approximately 210 minutes.
The clearance of corticosteroids from the blood is determined primarily by liver metabolism. This metabolism can be increased by drugs that induce liver enzymes.16 Chronic treatment with diphenylhydantoin can increase prednisolone metabolism by 77% and speed its clearance half-time by 45%. Diphenylhydantoin use can, by similar mechanisms, increase the metabolic inactivation rates of hydrocortisone, dexamethasone, and methylprednisolone by 15%, 51%, and 56%, respectively. In turn, corticosteroids can selectively enhance or decrease liver enzyme syntheses (e.g., the synthesis of transaminases involved in gluconeogenesis are increased, whereas synthesis of butyrylcholinestrase [pseudocholinesterase] is decreased). By day 12, patients receiving prednisone, 50 to 100 mg daily, have a 50% decrease in circulating butyrylcholinestrase activity.17 If the prednisone dosage is reduced to 10 to 15 mg daily, the butyrylcholinestrase level gradually recovers, becoming normal in 30 to 60 days. The activity of erythrocyte acetylcholinesterase is unaffected by corticosteroids. Immunoglobulin synthesis by the liver is variably affected: serum immunoglobulin G (IgG) is reduced, sera immunoglobulin A (IgA) and immunoglobulin M (IgM) are little changed, and serum immunoglobulin E (IgE) is elevated initially but, by 3 weeks of pharmacologic doses, is reduced.18
The plasma concentrations of free and total prednisolone show a diurnal variation when the drug is administered orally.19 Volunteers were given 2 mg prednisolone/10 kg body weight at 6 A.M., noon, 6:00 P.M., and midnight. Depending on the time of day, differences were found in prednisolone bioavailability, plasma concentrations, and clearance rates. The two extremes occurred at noon and 6:00 P.M. The 6:00 P.M. plasma concentration was 60% higher than during the rest of the day. This may explain in part the observation that there is less suppression of adrenal function when corticosteroids are administered in the morning than in the evening.20
The secretion of ACTH and hydrocortisone, the plasma levels of therapeutic corticosteroids, and the variations in intraocular pressure (IOP) all show diurnal phenomena. In addition, plasma glucocorticoid levels maintained above normal may cause an increase in the IOP. Therefore, it is logical to suspect that the normal diurnal secretion of hydrocortisone is responsible for the diurnal fluctuations in IOP. Glucocorticoid receptors have been detected in cultures of human trabecular cells.21 When adrenalectomized patients have had their blood levels of corticosteroids maintained at a constant level, their diurnal variations in ocular pressure did not occur.22
Long-term administration of corticosteroids above physiologic doses results in adrenal gland atrophy and a sustained suppression of the hypothalmic-pituitary-adrenal axis. If withdrawal from corticosteroid treatment is too sudden, the patient may exhibit weakness, fatigue, orthostatic hypotension, hypoglycemia, nausea, arthralgia, and dyspnea; deaths have been attributed to adrenal insufficiency.23 The degree of pituitary-adrenal suppression is not predictable in a given patient. Although certain generalizations can be made relating to the degree and duration of suppression with the dose and duration of corticosteroid treatment, there are many exceptions found when the degree of recovery actually is tested by the administration of ACTH-releasing hormone.24 In general, dosages of prednisone, 40 mg once daily for less than a week, do not result in significant adrenal suppression.25
Once adrenal suppression occurs, full recovery of hypothalamic-pituitary-adrenal function may take as long as 9 months.26 Weaning the patient away from corticosteroids may be required and can begin with rapid reduction to approximately physiologic doses of corticosteroid. For prednisone and prednisolone, the upper limits of doses having physiologic equivalence are approximately 5 mg/day; for dexamethasone, the upper limits of doses are approximately 0.75 mg/day. Tapering down below these doses is best done with a short-acting corticosteroid (e.g., hydrocortisone is better than prednisone or prednisolone, although the latter two drugs frequently are used). Dexamethasone, because of its long action, is best avoided as a weaning drug. An example of a withdrawal program would be to give the weaning drug daily at a dose equivalent to the normal physiologic secretion rate for 1 week and then to decrease the daily dose, at weekly intervals, by one-tenth steps of that dose.
Alternate-day therapy with doses less than 40 mg every other day tends not to produce suppression, even after long periods of treatment.27 In alternate-day therapy, the 2-day total dosage is given as a single dose (e.g., 25 mg prednisone, four times a day, becomes 200 mg prednisone every other day). However, the benefits of every-other-day therapy are lost if long-acting preparations (e.g., dexamethasone) are used. Prolonged alternate-day corticosteroid use eventually may cause the other side effects of daily therapy (e.g., when 25 patients taking daily prednisone for a mean duration of approximately 5 years were compared with 25 patients taking alternate-day therapy, drug-induced osteoporosis was present to a comparable degree in both groups).28 Surprisingly, long-term (i.e., mean 5-year duration) treatment of temporal arteritis may not induce osteoporosis.29,30
There is some evidence that alternate-day oral therapy may, in part, have less toxicity and be less efficacious because there is less drug absorption when large doses are used.31 When volunteers were given oral prednisone 0.2 mg/kg or 0.8 mg/kg, the total plasma prednisolone, the unbound plasma prednisolone, and the protein-bound prednisolone were much less than four times as much for the 0.8 mg/kg dose.
Another approach to avoiding the complications of prolonged corticosteroid use has been pulsed therapy. Experience gained by rheumatologists in treating autoimmune diseases, such as systemic lupus erythematosus, has been applied to ocular inflammatory diseases.32,33 In pulse therapy, a large dose of corticosteroid is given intravenously (e.g., methylprednisolone 1 g) over a short period (e.g., less than 1 hour) and repeated for several days (e.g., for 3 days). The patient then goes for a relatively long period, weeks or months, receiving little or no corticosteroid medication until the pulse therapy is repeated.
Corticosteroids usually are palliative. They attack the results of the disease process and not the causes. The correct corticosteroid dose in a given patient depends on the severity of the disease process and is based on both the clinician's experience and by a trial-and-error process.
Oral corticosteroids and subconjunctival, sub-Tenon's, and retrobulbar injections of corticosteroids have been used primarily for treating retrolenticular pathologies. Eyedrops and ointments have been used primarily for conditions of the lids, conjunctiva, cornea, and iris-ciliary body. Parenteral depot preparations may contain carriers, such as polyethylene glycol/myristyl picolinium, that provide prolonged release characteristics. An injection of one of these can give a therapeutic response for 6 or more weeks. Allergic reactions to the myristyl picolinium in a methylprednisolone preparation have been reported.34 The data available on the duration of drug presence in the orbit and its intraocular penetration are largely from animal studies (i.e., quantitative human data are relatively few). When a 5-mg dose of dexamethasone disodium phosphate, a nondepot formulation, was injected peribulbarly in 61 patients, a mean vitreous drug peak of 13 ng dexamethasone/ml was achieved in 6 to 7 hours. The mean serum drug peak of approximately 60 ng/ml was reached 20 to 30 minutes after the injection.35 Oral administration of dexamethasone, 7.5 mg, produced peak vitreous concentrations ranging from 1.7 to 23.4 ng/ml (median, 5.2 ng/ml) within 4 to 10 hours; the peak serum concentrations ranged from 2.5 to 98.1 ng/ml (median, 61.6 ng/ml) and occurred between 1 and 3 hours after ingestion.36
Early studies indicated that topical corticosteroids readily penetrated the cornea. However, as the assay techniques improved, the aqueous humor levels were revised downward. When 50 μl of a 0.1% betamethasone sodium phosphate solution were applied to the cornea, the mean peak level, detected 1.5 to 2 hours later, was 7.7 ng/ml (i.e., 1/13,000 of the original concentration).37 When subjects were given a single 50-μl drop of a commercially available preparation of dexamethasone alcohol 0.1% suspension,38 fluorometholone alcohol 0.1% suspension, prednisolone acetate 1% suspension,39 or prednisolone sodium phosphate 0.5% solution40 just before cataract surgery, the aqueous humor levels of dexamethasone, fluorometholone, and prednisolone were measured. The results are listed in Table 1.
Commercially available preparations of corticosteroid eyedrops exist as solutions or suspensions. The drug concentration that is dissolved in the supernatant of a suspension is not altered by shaking it (unless the temperature of the solution becomes elevated). However, the number of drug-containing particles in the supernatant is transiently affected by shaking.41 Different brands, depending on such variables as particle size and particle number, can behave differently. The relative amounts of prednisolone acetate after shaking were compared using drops from two different brands; the respective percent amounts of drug after 0, 10, 20, and 40 shakes (numerator), when compared with the amounts after 20 minutes of shaking on a mechanical rotator (denominator), were for one brand 23%, 53%, 69%, and 82%, respectively, and were for the other brand 5%, 14%, 13%, and 22%, respectively. Because the drug-containing particles can be trapped under the lids and act as a reservoir, prior shaking of the bottle alters the effective dose of a corticosteroid suspension.
Giant Cell Arteritis
Corticosteroids have been the primary treatment for giant cell arteritis. The dose size and frequency have been monitored using relief of symptoms, such as headache, jaw claudication, musculoskeletal pain and malaise, and/or an improvement in laboratory data, such as reduced Westergren erythrocyte sedimentation rate, C-reactive protein level and reactive thrombocytosis, or recovery from anemia.42–44 Initially, large doses of corticosteroid are given orally (e.g., 100 to 150 mg/day prednisone) or intravenously (e.g., 1 g [or more] prednisolone disodium phosphate per day). Once the symptoms are reduced and the laboratory study results approach normal levels, the dose of corticosteroid is lowered.
A major goal of corticosteroid treatment is to prevent ischemic infarction of the optic nerves. Vision loss can occur even after the initiation of intravenous corticosteroids,45 but it is rare after 96 hours (i.e., 4 days) of treatment.46 The resultant loss of vision rarely is reversible,47 and in those patients claiming improvement, the course of the disease often was atypical (i.e., one patient had an uncharacteristically slow [months] progression of visual loss to light perception and 20/400 visual acuity before responding to a pulse dose of methylprednisolone 80 mg intravenously followed by 1 month of oral prednisone 100 mg/day).48 However, aggressive therapy can be recommended, if not to salvage the affected eye, then to protect the unaffected one.
The need for continued corticosteroid treatment once the initial signs and symptoms disappear is quite variable, from as short as 1 month's time to indefinitely.49,50 Relapses frequently occur after treatment is stopped. In one study, one third of patients relapsed,51 75% of these occurring within 3 months; one patient relapsed more than 10 years after stopping treatment. The percentage of CD8 cells has been correlated with disease control: patients with giant cell arteritis treated 6 months with corticosteroids whose CD8 cell counts were lower than one standard deviation from those of control subjects required a higher dose of prednisone, a longer duration of treatment, and were more likely to relapse.52
Because the initial therapeutic response to large doses of corticosteroids is dramatically rapid, usually within 24 hours, some have advocated using this as a diagnostic test rather than the temporal artery biopsy. According to this line of reasoning, the literature indicates that temporal artery biopsy specimen results are negative in 60% to 70% of temporal arteritis suspects.53,54 Because only a minority of patients have a positive biopsy specimen result, the procedure should be reserved for those in whom either the corticosteroid response is rapid but prolonged treatment is medically contraindicated or there is failure to rapidly respond to corticosteroid treatment55 but the physician believes that giant cell arteritis is present.
However, the alternative argument seems more cogent: the temporal artery biopsy results, whether positive or negative, seem correct approximately 95% of the time. Most clinicians would prefer to have a positive biopsy specimen result before committing their patients to the potential complications of long-term corticosteroid treatment53,54 or a negative biopsy specimen result before withholding vision-preserving therapy. When 109 patients with giant cell arteritis or polymyalgia rheumatica or both were followed prospectively for a mean ± standard error period of 68.5 ± 5.4 weeks, 10 patients had fractures develop, five of which were vertebral, and four patients had peptic ulcers develop, two of which perforated.56 But here, too, exists a counter argument. Despite the frequently stated fear of corticosteroid complications, many investigators have found that temporal arteritis and its treatment does not negatively affect longevity.49,51,57
Methotrexate has been used to reduce the need for corticosteroids.58 The investigators' goal was to control the disease after an initial 2 weeks of large-dose prednisone, with 10 mg prednisone per day or less. Methotrexate was added as a single weekly 10-mg dose at the beginning of treatment. A mean length of 14 weeks was needed for 11 newly diagnosed patients with giant cell arteritis to achieve control on 10-mg prednisone per day and a mean length of 30 weeks before corticosteroids could be withdrawn completely.
Alternate-day therapy also minimizes the side effects of treatment. However, control of the disease is compromised. Three groups of 20 patients with biopsy-proven temporal arteritis were treated after the acute phase of their disease as follows:
Group A: Prednisone 15 mg every 8 hours
After 4 weeks of following this therapy, the patients found the subsequent results in their respective groups:
Group A: Eighteen of 20 patients were asymptomatic with mean hemoglobin (in
g/dl) having increased from 11.4 to 14.1 and mean sedimentation rate (mm/hr) having
decreased from 96 to 14.
Although statistical analysis confirmed that alternate-day therapy was not as effective, none of these patients experienced an ischemic optic neuropathy or other vasculopathy. The question remains whether some loss of disease control is worth the potential reduction in drug-induced morbidity. This question is best answered on a patient-by-patient basis.
The question of “Can a temporal artery biopsy be positive after prolonged systemic corticosteroid therapy?” is not the same as the more important question of “Will the frequency of finding a positive temporal artery biopsy result be reduced by prolonged corticosteroid therapy?” The answer to both questions is yes. It is possible for the results of the temporal artery biopsy specimen to remain positive (i.e., show signs of inflammatory disease including giant cells) for long periods after initiation of corticosteroid treatment. Positive biopsy specimen results have been reported after 1 month of prednisone 60 mg daily,60 6 weeks of prednisone 30 to 40 mg daily,61 and 6 months of prednisone 30 to 60 mg daily.62 However, there also is evidence to suggest that corticosteroid use results in a rapid reduction in the incidence of positive results of biopsy specimens. When 132 patients were analyzed who were clinically diagnosed as having temporal arteritis and in whom 84 had a positive biopsy specimen result, it was found that the incidence of a positive biopsy result before corticosteroid treatment was 82% but fell to 60% with less than 1 week of therapy and was 10% after 1 week of treatment.63
There are surprisingly few controlled investigations in the literature regarding corticosteroid treatment of optic neuritis.64 Those that exist have had relatively few subjects and have failed to convincingly show efficacy (e.g., retrobulbar injections of a single dose of triamcinolone,65 40 mg, in patients with optic neuritis of less than 10 days' duration did not result in a significant difference in visual acuity or visual field recovery at 6 months after treatment). Uncontrolled studies have tended to be more enthusiastic.66
The visual loss from a single attack of optic neuritis usually resolves spontaneously. When permanent visual loss occurs, it often is the result of the cumulative effects of recurrent attacks. If a patient is given corticosteroids for his initial attack, he tends to attribute his recovery to the drug. He becomes “hooked” and the physician is under pressure to treat each successive attack with corticosteroids. If the physician does not and the visual recovery is worse than that of the preceding attack, the physician can be accused of withholding a beneficial therapy. However, repeated treatments with corticosteroids, especially if the attacks are close together, can lead to the side effects associated with this class of drugs.
To determine the efficacy of corticosteroid treatment in optic neuritis, a multicenter National Eye Institute-funded investigation was performed by the Optic Neuritis Study Group.67–73 Unfortunately, there has been a strong tendency to attribute its findings to the different routes of corticosteroid administration (e.g., statements are made such as “Oral steroids are contraindicated in optic neuritis”) rather than to the differences in doses and rates of drug administration. This discussion will focus on the latter two.
Approximately 450 subjects were entered into the study. Subjects were between 18 and 46 years of age and were seen within 8 days of their first known attack of optic neuritis in the affected eye; 31 of 448 subjects were known to have prior optic neuritis in the contralateral eye. None of the subjects were considered to have “probable MS” or “definite MS” and none had had previous treatment with systemic corticosteroids. The investigators focused on the effect of systemic corticosteroids on the rate of recovery of visual function, the final level of visual function recovery, the frequency of optic neuritis recurrences, and the frequency of progression to probable MS and definite MS. The patients were entered into one of three groups. There was a high-dose corticosteroid treatment group that received pulse therapy of intravenous methylprednisolone sodium succinate (molecular weight = 496.53) 250 mg every 6 hours for 3 days followed by oral prednisone 1 mg/kg for 11 more days; a moderate-dose corticosteroid treatment group that received oral prednisone 1 mg/kg/day for 14 days; and an oral placebo-group. The first two treatment groups also received a short oral taper of prednisone 20 mg on day 15 and prednisone 10 mg on days 16 and 18. Assuming that the liver's conversion rate of prednisone (molecular weight = 348.43) to the active form of the drug, prednisolone (molecular weight = 364.43), occurs with 80% efficacy and that the patients weighed no less than 60 lb or more than 250 lb, the prednisolone-equivalent amount of corticosteroid received by the high-dose/pulsed group during the first 3 days of treatment was 2202 mg, whereas that of the moderate-dose group was 68 to 285 mg (i.e., the high-dose pulsed group received between 8 and 32 times more prednisolone).
The use of corticosteroids increased the early rate of recovery of visual function (e.g., on day 4, the mean visual acuity in the high-dose pulsed prednisolone group had improved from 20/80 at entry to 20/25, whereas the acuity in the placebo-treated group had not changed). However, the use of corticosteroids did not improve the final level of acuity compared with that achieved by spontaneous recovery. Six months after treatment, the acuity in eyes receiving the high-dose pulsed treatment was only minimally better than placebo-treated eyes; by 1 year, there was no significant difference. Moderate dose but neither high-dose pulsed corticosteroid treatment nor placebo treatment significantly increased the risk of a future attack of optic neuritis in the same or the contralateral eye. During the first 2 posttreatment years, 30% of patients taking the moderate dose (68 to 285 mg prednisolone equivalent for the first 3 days) had a second attack compared with 14% taking the high-dose pulsed treatment (2202 mg prednisolone equivalent for the first 3 days) and 16% taking the oral placebo. This significant increase remained at the end of 5 years being, respectively, 41% moderate dose and 25% for both pulsed high-dose and placebo treatments. With regard to the frequency of progression to probable MS or definite MS, high-dose pulsed therapy showed a benefit over both moderate-dose therapy and placebo that persisted for 2 years but was no longer statistically significant by 5 years posttreatment. At 2 years, 7% of the high-dose pulsed corticosteroid treatment group had converted to definite MS, whereas the corresponding figures for placebo and moderate-dose treatments were 16.7% and 14.7%, respectively. Most of the benefit from high-dose pulsed treatment came in those subjects with two or more brain magnetic resonance imaging signal abnormalities; the rate of progression to definite MS in patients with normal scans was too low to be evaluated. However, by the third year posttreatment, there was no significant difference in the incidence of definite MS among the three groups. At 5 years posttreatment, 36% of the 388 study patients who, at entry, were considered to have neither probable MS nor definite MS had converted to the former (27%) or latter (9%). There were no statistically significant differences between the conversion rates of the three treatment groups. These findings suggest an overall beneficial effect that lasts for 2 years with 3 days of high-dose pulsed corticosteroid treatment and raise the possibility that high-dose pulsed treatment should be repeated every 2 years in those patients having all three of these characteristics: optic neuritis, no diagnosis of probable MS or definite MS, and two or more brain plaques seen on magnetic resonance imaging. The complication rate from corticosteroids in these subjects, ages 18 to 46 years, was small. Only two of the 150 subjects receiving high-dose pulsed corticosteroid therapy had possible drug-related significant side effects: the first patient became psychotic and the second patient had pancreatitis develop.
The methodology of the Optic Neuritis Study Group has been criticized. The high-dose pulsed corticosteroid treatment group was not masked or controlled because it was the only one to receive its initial medication intravenously. The one control group in the study received an oral placebo. Thus, only the moderate treatment group, which received its medication orally, was masked and controlled adequately. Patient or physician bias or both could have affected the results reported for the high-dose pulsed group. Another criticism is the failure of the study group to consider a prior attack of optic neuritis in the contralateral eye or a recurrence of optic neuritis in the same or contralateral eye as evidence of probable MS or definite MS. Many, if not most, neurologists and neuro-ophthalmologists would consider a second attack of optic neuritis as justifying the word multiple in multiple sclerosis. The statistical effect of a reclassification of these subjects is not clear.
Retrospective and prospective studies have assessed the value of the IOP response to corticosteroids as a predictor for developing glaucoma. In one study,74 29 of 134 patients had glaucoma develop 5 to 15 years after being challenged with corticosteroid drops. Nine (26%) of the 29 patients were from the group of 34 patients who had had a high response (16 mmHg or higher) to topical dexamethasone 0.1% drops; 13 (20%) of the 29 were from the group of 66 patients who had had an intermediate response (6 to 15 mmHg), and 7 (21%) of the 29 were from the group of 34 patients who had had a low response (5 mmHg or lower). The author concluded that the test was of little predictive value. However, the overall incidence of glaucoma developing in this study population was so inexplicably high (21.6%) that some problem in the methodology used is suspected. In a retrospective study of 788 subjects who had had their IOP response assessed 5 or more years previously using dexamethasone 0.1%, four times a day for 6 weeks and had been classified by Becker's criteria, 13% of the high responders had glaucoma develop (i.e., elevated pressures and visual field loss) and an additional 64% had ocular hypertension develop.75 The corresponding figures for intermediate and low responders were as follows: 3% and 0% had glaucoma develop, respectively, and 46% and 2% had ocular hypertension develop, respectively. In another study of 22 normotensive subjects who were high responders (i.e., subjects whose IOPs increased 16 mmHg or greater) and who were followed up 10 or more years, five had IOPs greater than 21 mmHg develop. Two of these five subjects had visual field defects develop.76
The IOP response to corticosteroids may have some value in predicting who will have open-angle glaucoma develop, but there are too many false-positives to justify routine prospective testing using it (i.e., too few of the high responders had glaucoma develop).
Thyroid (Graves) Ophthalmopathy
This is an autoimmune condition in which the orbital tissues exhibit lymphocytic infiltration, increased mucopolysaccharide content, and edema. The extraocular muscles initially become inflamed and swollen and later fibrosis occurs. Diplopia, proptosis, and a compressive optic neuropathy can result. Large doses and prolonged use of systemic corticosteroids can arrest the process. However, once scar tissue forms or optic atrophy has occurred, drug therapy is of no benefit in reversing these problems. The active phase of thyroid ophthalmopathy is self-limited. Functional sequelae can be prevented if the process is controlled by corticosteroids until the active phase burns itself out. This usually takes months to, rarely, years of treatment.
One early report77 of 10 patients with thyroid optic neuropathy found a return of visual acuity to 20/30 or better in all subjects. They were treated with daily oral prednisone, 20 to 120 mg for 1 to 54 months. Subsequent authors have not had as good success. Only 10 of 21 eyes with thyroid optic neuropathy responded, even though doses of prednisone up to 100 mg were used daily for at least 30 days78; these authors noted that if benefit were to occur, visual function began to improve within 1 week of starting treatment. Patients with severe exophthalmos, if acute, also have responded to high-dose systemic corticosteroids. Administering oral prednisone (e.g., 80 mg or more a day) improves approximately half the patients treated.79,80 If patients do not respond, it is either because the dose of corticosteroid is too low or secondary changes have occurred that prevent improvement. However, not all clinicians accept this explanation. The enigma (to some) of why patients do not always respond has led to several immunologic investigations.81
In one study, most patients with acute thyroid ophthalmopathy were found to have a decreased number of thymus-derived lymphocytes that, in vitro, formed rosettes with sheep erythrocytes; these patients responded well to corticosteroids. Those patients who did not respond to treatment were from the minority that did not have a decrease in circulating thymus-derived rosette-forming lymphocytes. In another study, the presence of human leukocyte antigen (HLA)-DR4 was significantly associated with a good response to corticosteroid therapy. Fourteen of 37 patients responding to corticosteroid treatment were HLA-DR4 positive. None of the 20 patients who did not respond to corticosteroids were HLA-DR4 positive.82 Somatostatin receptors occur on lymphocyte membranes and octreotide binds to them. Orbital scintigraphy with (111I n-diethylenetriamine penta-acetic acid-D-Phe1) octreotide has been used to predict which patients will respond to corticosteroids.83 Presumably, there are more lymphocytes present in the orbit during active disease when the condition is responsive to corticosteroids.
Rapid control of the various forms of thyroid ophthalmopathy has been achieved using pulse therapy followed by oral therapy (e.g., 500 mg methylprednisolone intravenously within 30 minutes on days 1 and 2 followed by 40 mg prednisolone daily).84–86 Because of potential irreversibility once optic atrophy occurs, even larger initial pulse doses have been advocated for thyroid optic neuropathy (e.g., 1000 mg methylprednisolone daily for 3 days).87
Radiation therapy for thyroid ophthalmopathy has been claimed superior to corticosteroid therapy short term88 but less effective long term.89 In prospective, randomized trials, radiation therapy and oral prednisone were found equivalent 24 weeks posttreatment,90 and combined use of radiation therapy and corticosteroids was found more effective than radiation alone at 6 to 9 months posttreatment.91 Oral corticosteroids appear to be as effective as intravenous immunoglobulin therapy 6 months after initiation of treatment92 and slightly more effective than subcutaneous somatostatin treatment, three times a day, 3 months after initiation of therapy.93
Although controversial, there is some basis to believe that radioactive iodine treatment of hyperthyroidism is associated with an increased risk of developing or exacerbating thyroid ophthalmopathy. In one randomized, prospective, but not masked, study, ophthalmopathy developed or worsened in four of 38 medically (methimazole) treated patients, six of 37 surgically (subtotal thyroidectomy) treated patients, and 13 of 39 iodine-treated patients (p = .02 for the risk of this last group compared with the other two combined).94 Assuming that radioactive iodine results in the release of antigenic agents from damaged cells and that these antigenic agents are the cause of the increased thyroid ophthalmopathy, systemic corticosteroids have been used to try to prevent or attenuate postradioiodine thyroid ophthalmopathy. Prednisone, 0.4 to 0.5 mg/kg/day starting 2 to 3 days after radioiodine and continuing for 1 month, followed by a 2-month taper was associated with improvement or no progression of ophthalmopathy in 145 patients. In 150 patients receiving radioiodine but not oral corticosteroids, 23 (15%) had worsening ophthalmopathy or ophthalmopathy develop 2 to 6 months after treatment.95
ALLERGIC CONJUNCTIVITIS. Systemic corticosteroids do not alter the early phase of allergic conjunctivitis (first hour, neutrophil response) but do markedly limit the late-phase (eosinophil) response.96 Topical corticosteroids are effective in the control of conjunctival allergic symptoms.97–99
CHALAZIA. Intralesional corticosteroid injection of lid chalazia usually is effective, causing them to resolve and reducing the need for their incision and drainage.100–102
CORNEAL GRAFT REJECTION. Corneal transplant and graft rejections can be treated with topical,103 subconjunctival, and/or systemic104 corticosteroids. The success rate is variably reported to be in the 50% to 75% range.105
EOSINOPHILIC GRANULOMA. Eosinophilic granulomas of the orbit have been treated successfully with intralesional corticosteroid106 injections.
ERYTHEMA MULTIFORME. Systemic corticosteroids do not prevent or limit the ocular involvement in acute erythema multiforme (Stevens-Johnson syndrome).107
GLAUCOMA-FILTERING SURGERY. Prednisolone acetate eyedrops given during the first 20 days after trabeculectomy surgery for glaucoma in patients who did not receive antimetabolites improved the intraocular control when these patients were evaluated 10 years after surgery.108
HEMANGIOMAS OF THE LIDS. Systemic corticosteroids reduced the sizes of cutaneous hemangiomas.109,110 Subsequently, local intralesional injections of corticosteroids and a topical cream111 of clobetasol propionate 0.05% were used successfully to treat lid hemangiomas.112,113 Necrosis of lid tissue, fat atrophy, and adrenal suppression have been reported as complications of intralesional injections of hemangiomas.114–116
MYASTHENIA GRAVIS. The functional deficits of the eyelids and extraocular muscles found in myasthenia gravis will respond to corticosteroid therapy. However, acute initial worsening of peripheral muscular weakness, including respiratory distress, can occur. Therefore, corticosteroid therapy of myasthenia gravis is best left to the neurologist. If an exception exists, it is the purely ocular form of myasthenia gravis documented as such by an electromyogram. Ocular myasthenia gravis will respond to relatively low doses of corticosteroid. An initial dose of approximately 20 mg/day prednisone is maintained until symptoms stabilize, usually in approximately 3 weeks. A daily dose of 5 to 10 mg/day prednisone usually can then be used to maintain the improvement.
ORBITAL PSEUDOTUMOR. Inflammatory pseudotumors usually respond to systemic corticosteroids,120 but relapses are frequent.121 In one study,122 approximately 80% of 32 patients showed an initial response; approximately half of these patients were cured but the pseudotumors of the other half recurred. Perhaps some of the recalcitrant cases reported in the past have represented lymphoid-lymphoma tumors rather than inflammatory processes.123
POSTLASER PRESSURE SPIKE. Pretreatment with prednisolone acetate 1% eyedrops, one drop every 6 hours for 36 hours, does not prevent the elevation in IOP that occurs after argon laser trabeculoplasty.124
POSTOPERATIVE INFLAMMATION. Iritis. Topical corticosteroid treatment usually125 has been effective in reducing intraocular inflammation after cataract surgery. Corticosteroid use may result in an increase in complications such as unintentional conjunctival filtering blebs and iris prolapse (i.e., complications resulting from reduced wound strength).126 A single subconjunctival injection of a nondepot corticosteroid preparation given at the conclusion of cataract surgery may be of some short-term (e.g., 24 hours) benefit127 but does not seem to add any long-term advantage to several weeks of daily topical corticosteroids.128
Cystoid Macular Edema. Chronic inflammation after pseudophakic surgery can lead to cystoid macular edema. Systemic corticosteroid therapy can be helpful in treating this condition, but prolonged drug use often is required.129 Success has been reported using a single retrobulbar injection or multiple sub-Tenon's injections of corticosteroid.130
SCLERITIS. Subconjunctival triamcinolone has been of value in treating nonnecrotizing scleritis.133
SJO¯GREN'S SYNDROME. Topical corticosteroid drops have been used successfully to treat the keratoconjunctivitis sicca of Sjogren's syndrome.134
SYMPATHETIC OPHTHALMIA. Therapy of sympathetic ophthalmia with systemic corticosteroids and early enucleation of the exciting eye had a beneficial effect in maintaining useful vision in the sympathizing eye.135,136
TRAUMATIC OPTIC NEUROPATHY. Case reports have suggested a role for high-dose intravenous corticosteroids in the treatment of traumatic optic neuropathies,137 including those from lightning strikes.138 However, no adequate prospective, controlled studies have been performed.
THYGESON'S KERATOPATHY. Topical corticosteroids have been used successfully to treat Thygeson's superficial punctate keratopathy.139
|SIDE EFFECTS AND TOXICITIES|
Bacterial and Fungal
Corticosteroids limit the inflammatory and immune responses to infections.140 Therefore, an argument can be made for avoiding corticosteroid use in ocular infections. However, the inflammatory response can irreparably damage corneal clarity and the integrity of the delicate intraocular structures. Therefore, a counter argument can be made for corticosteroid use early in the treatment of corneal stromal and intraocular infections. One approach is to treat the infection with antibiotics alone for 24 to 48 hours and, if the eye improves, to assume that the antibiotic choice is correct. Corticosteroids could then be given with some sense of relative safety. However, the clinical reality is that it often is difficult to ascertain in 24 to 48 hours whether the endophthalmitis or keratitis is improving.
HERPES SIMPLEX. Corneal epithelia infected with Herpes simplex may heal despite corticosteroid use. But many, if not most, patients will have their disease progress, often at an accelerated rate.141 Corneal perforation can result from continued use. A general consensus is that in active corneal epithelial disease, corticosteroids are to be avoided. In stromal (metaherpetic) disease, many clinicians give corticosteroids but usually only if the patient will apply antiviral medications simultaneously. In a masked study,142 patients were assigned to 10 weeks of either tapering doses of prednisolone phosphate eyedrops plus trifluridine eyedrops or placebo eyedrops plus trifluridine eyedrops. Both regimens had similar adverse events attributable to Herpes activation (i.e., the use of corticosteroids did not increase corneal epithelial disease: the prednisolone group had, in 57 patients, four instances of reactivation of viral epithelial disease and one instance of development of an epithelial defect of more than 1-mm diameter; for the 49 placebo-treated patients, one instance of reactivation and three instances of enlarged epithelial defects occurred). The prednisolone phosphate group had a milder and shortened course of stromal keratitis. When placebo-treated patients who were not improving had prednisolone eyedrops added to their regimens, they improved to a similar level of acuity as did those patients initially receiving topical corticosteroids (i.e., the delay did not have adverse consequences).
HERPES ZOSTER. In the era before effective systemic anti-Herpetic medications (e.g., acyclovir, famciclovir), the use of systemic corticosteroids or ACTH in the early (prevesicular or localized vesicular) stages of Herpes zoster infections occasionally was associated with the development of a severe, generalized vesicular eruption.143
The strength of a healing wound resides in the degree of proliferation of connective tissue. For the cornea, this means that the stroma, rather than the epithelium or endothelium, is the key tissue. Corticosteroids not only slow healing by decreasing scar formation, but they will promote the removal of formed collagen (e.g., corticosteroids are used to reduce the size of keloids). Dexamethasone 0.1% eye ointment applied during the first 3 weeks after intraocular cataract surgery resulted in an 8.7% incidence of filtering blebs versus a 1.6% incidence in untreated eyes.144
MYDRIASIS AND PTOSIS
There are many reports in the literature that topical corticosteroids produce ptosis and mydriasis. There is no correlation with the degree of these effects and the amount of IOP elevation. Animal studies implicating the vehicle rather than the corticosteroid have used injections rather than topical administration.145
Relative mydriasis was observable in the treated eyes of 12 of 21 patients who received unilateral topical corticosteroids.146 Using pupillography, all patients were found to have mydriasis in the unilaterally treated eye.147 Several subjects also had up to 3 mm of ptosis develop. The ptosis and mydriasis have been attributed to a myopathic-wasting effect of the corticosteroid.
Chronic corticosteroid use, either topically or systematically, can result in posterior subcapsular cataract formation. However, surprisingly, Cushing's syndrome, which results from an overproduction of endogenous corticosteroids, has not been associated with a high incidence of cataract formation. In 60 patients having Cushing's syndrome from 1 to 20 years (mean ± standard deviation: 5.5 ± 3.7 years), only two patients had posterior subcapsular cataracts; these were bilateral and symmetric.148
In 1960, a 23% incidence of posterior subcapsular cataracts was reported in patients with rheumatoid arthritis receiving daily doses of 10 to 16 mg prednisone orally for 1 or more years and a 75% incidence for those receiving more than 16 mg/day.149 Adults with renal transplants and treated with corticosteroids had almost a 50% incidence of posterior subcapsular cataracts.150
When 48 children with nephrotic syndrome who were diagnosed between 1.5 and 12 years of age were treated with oral prednisone, one third developed posterior subcapsular cataracts.151 The children with cataracts had ingested prednisone for at least 1 year. When 39 of these children were re-examined at an average interval of 16 months later, nine of the 15 with cataracts no longer had them. Six of these nine children had not received corticosteroids between the two visits, but three children had. Therefore, cessation of corticosteroid treatment could not explain all those instances of cataract reversibility.152 Reversal of posterior subcapsular opacities in children, when the systemic dosage has been reduced or changed to topical therapy, has been reconfirmed.153
In 30 children with chronic asthma treated with alternate-day oral prednisone for 3 to 11 years (mean duration, 4.3 years), only one was found to have cataracts. This child had bilateral posterior subcapsular cataracts.154 In 18 asthmatic children receiving oral prednisone for 2 or more years, five (45%) of 11 receiving daily prednisone had posterior subcapsular cataracts, whereas two (29%) of seven receiving alternate-day prednisone had posterior subcapsular cataracts; in both groups, the average dose was the same in those with and without cataracts.155 Thus, the incidence of cataracts during chronic corticosteroid use is quite variable.
Inhaled corticosteroids are a mainstay of asthma therapy. The agents used (e.g., beclomethasone, budesonide, and fluticasone) were chosen because they would be minimally absorbed into the circulation. The literature as to their cataractogenic effect is conflicting. Perhaps some are more cataractogenic than others, and those patients most likely to require frequent use of inhaled corticosteroids also would be most likely to require supplemental systemic corticosteroids. Several studies have concluded that there is little or no risk of cataracts from inhaled corticosteroids,156–158 whereas other studies find a dose-related risk.159–161
Corticosteroid treatment clearly is associated with cataract formation, but there has been no clear-cut correlation found with their formation and total dosage, weekly dosage, duration of dosage, or patient age.162 Other factors163 may determine the likelihood of cataract development such as HLA status (e.g., CW3).
The widespread use of corticosteroids in the 1950s led to isolated reports of individual patients who had elevated IOP.164,165 However, several negative studies166 found no effect in patients receiving corticosteroids for years. In 1962, Goldmann167 described five patients with corticosteroid-induced glaucoma. Many other reports followed. The frequency of this hypertensive effect may have been missed earlier because corticosteroids reduce scleral rigidity, lowering the IOP as recorded with a Schiotz tonometer. It was not until the early 1960s that applanation tonometry came into widespread use. At one time, it was suspected that this “steroid glaucoma” was produced by a mineralocorticoid action, perhaps causing increased sodium transport by the ciliary processes. However, treatment with potent mineralocorticoids (e.g., aldosterone) did not reproduce the phenomenon.
Not all patients treated with corticosteroids will have IOPs of more than 21 mmHg develop. Conversely, those patients who are susceptible will have high IOPs develop irrespective of the method of administration (e.g., inhaled, oral)168–170 or age,171 provided that the drug is given long enough and in a high-enough dosage. Those who will not have markedly elevated pressures develop by one method of administration will not have it develop by another.172 The magnitude of the elevation is variable from patient to patient and usually reaches a maximum in 4 to 8 weeks. In populations of normotensive subjects treated unilaterally with 0.1% dexamethasone or betamethasone eyedrops, different investigators have reported average increases of 4.6 to 6.6 mmHg in 4 to 9 weeks. However, some individually will respond more dramatically than others.
The corticosteroid-induced elevation in IOP has raised several questions about its relation to chronic simple glaucoma. Is there a genetic component involved in the response? If there is, then is this same genetic component a factor in the familial tendency to develop chronic simple glaucoma? Patients with glaucoma reportedly have significantly greater amounts of plasma glucocorticoid activity than normotensive patients.173 Although almost all of the glucocorticoid activity in normotensives was attributable to hydrocortisone, significant amounts of the activity in subjects with glaucoma was caused by other glucocorticoids. An uncommon form of autosomal-dominant juvenile open-angle glaucoma174 is linked to a segment of chromosome 1 and has been studied in two families,175 both of which showed a mutation in the trabecular-meshworkinducible-glucocorticoid-response gene (TIGR) mapping to chromosome 1. Cultured human trabecular meshwork cells exposed to corticosteroids have shown changes in nuclear size and DNA content, metalloproteinase activity, cytoskeletal organization, glycoprotein secretion, and sodium-potassium-chloride transport.176–181 Enzyme homogenates from trabecular meshwork cell cultures from patients with primary open-angle glaucoma seem to metabolize cortisol in a different manner than homogenates from nonglaucomatous patients.182
There are similarities between steroid glaucoma and chronic simple glaucoma: open filtration angles, decreased tonographic outflow facility,183 asymptomatic elevation of IOP, arcuate pattern of visual field loss, cupping of the optic disc, and therapeutic response to the same medications. Further, patients with open-angle glaucoma tend to respond with greater IOP elevations than do normotensives. For example, patients with glaucoma,184 despite continued use of their ocular medications, who were treated with corticosteroid eyedrops showed an average increase IOP of 16 mmHg. In another study, only 80% of patients with glaucoma responded to corticosteroid eyedrops, but these patients had an average increase of 13.5 mmHg.185
Among the theories proposed to explain the corticosteroid-induced pressure response are increased synthesis of trabecular meshwork mucopolysaccharides (e.g., hyaluronic acid), increased episcleral venous pressure from corticosteroid-induced vasoconstriction, increased aqueous humor osmolarity from altered electrolyte secretion by the ciliary processes, and a change in the actin organization in trabecular meshwork cells.186 When human eyes obtained within 24 hours of death were perfused with dexamethasone, approximately 30% had ocular hypertension develop associated with thickening of the trabecular beams and juxtacanalicular tissue and increased amounts of an amorphous extracellular material.187 Compared was ocular tissue from five subjects diagnosed as having corticosteroid-induced glaucoma, six subjects with chronic simple glaucoma treated with corticosteroids for months to years, and seven subjects with chronic simple glaucoma not treated with corticosteroids. A material resembling basement membrane was found in the eyes with corticosteroid-induced glaucoma; little or none of this material was found in chronic simple glaucoma eyes treated with corticosteroids and none was found in the untreated chronic simple glaucoma eyes.188 The authors speculated that a preexisting defective status of the trabecular meshwork in patients with chronic simple glaucoma might explain their exaggerated response to relatively small amounts of material being deposited in response to corticosteroids.
Human trabecular meshwork cells from patients with chronic simple glaucoma have been found to metabolize hydrocortisone at different rates and produce different metabolites than found from those with nonglaucomatous tissues.189,190
In general, the corticosteroid-induced IOP elevation is associated with altered trabecular meshwork resistance to aqueous humor outflow. When 24 human subjects were evaluated before and after treatment with topical dexamethasone phosphate 0.1% drops four times daily, there was a significant increase in IOP but no significant change in aqueous humor flow.191 This is consistent with the finding by multiple investigators that a corticosteroid-induced reduction in the tonographically measured trabecular outflow facility “C” value147,192 has occurred. For example, 10 normotensive volunteers given unilateral corticosteroid eyedrops responded with IOP elevations of 2 to 13 mmHg and reduced outflow facilities; the greatest reductions in C value were in those eyes with the greatest pressure responses. Similar results were found in another study of chronic simple glaucoma and glaucoma suspect subjects,151 but less than 25% of the mean increase in IOP, 8.6 mmHg, could be explained by the reduction in C values. A puzzling finding has been the discovery of significant changes in the outflow facilities of the untreated contralateral eyes.193 Inconsistencies have been reported as well. For example, in 50% of subjects receiving corticosteroid eyedrops, outflow facility was reduced; in 20%, there was no effect; in 15%, outflow facility increased; and in 15%, the response, with time, was biphasic.194
The high frequency of corticosteroid-hypertensive responses in patients with glaucoma might only represent the effect of trabecular damage caused by the glaucoma. Patients with unilateral secondary glaucoma with closed angles or traumatic angle recessions were evaluated.195 Bilateral corticosteroid eyedrops were given. A positive corticosteroid response was considered as either a greater than one-third reduction in outflow facility or an increase in IOP of more than 6 mmHg. The overall incidence of positive responses in these subjects was 32% to 36%, approximately the same as found in the normal population. Further, both eyes responded in the same way. It was concluded that the corticosteroid response was not caused by local trabecular damage.
The cornea does not appear to play a significant role in the corticosteroid pressure response. When 103 subjects without ocular disease were administered dexamethasone phosphate, four times a day for 4 weeks to one eye, there was no significant change in corneal thickness when each eye was compared to itself despite a significant increase in IOP.196
There have been three main interpretations of the corticosteroid-induced pressure response and its relation to chronic simple glaucoma: Becker's interpretation,197,198 Armaly's interpretation,199–203 and Francois' interpretation.185
Before the 1960s, chronic simple glaucoma generally was assumed to be caused by a single dominant gene with variable expressivity. However, because of the corticosteroid response, Becker proposed that a single codominant gene was responsible for both chronic simple glaucoma and the corticosteroid ocular pressure response. Because approximately 4% of the general population older than 40 years had IOPs of more than 21 mmHg, Becker predicted that the incidence of the glaucoma causative gene (G) was 20% and that of the normotensive gene (N) was 80%. This led to Becker predicting the following incidences of genotypes: (0.2G + 0.8N)2 = 0.04GG + 0.32NG + 0.64NN. Becker believed each of these genotypes would respond differently to corticosteroids. Patients were tested using corticosteroid eyedrops for several weeks. He found he could divide the responses into three groups of approximately the right size he had predicted if he used the following assumptions:
In addition, Becker found that all patients with chronic simple glaucoma given corticosteroid eyedrops had a marked elevation in IOP. In addition, 87% of all offspring of patients with chronic simple glaucoma had a moderate elevation in IOP.
Armaly was able to divide patient responses to corticosteroids into three groups with roughly the same relative proportions as did Becker. Armaly used a different criterion: the amount of elevation rather than the final pressure. Armaly's three groups were divided into elevations of less than 6 mmHg, 6 to 15 mmHg, and more than 15 mmHg. However, Armaly found a poor correlation between corticosteroid responsiveness and chronic simple glaucoma. Only 46% of the latter were high pressure steroid responders. Armaly proposed that the phenotypic patient with chronic simple glaucoma was the result of having several abnormal genes. One of these genes, having a high but not mandatory association with glaucoma, was the steroid responsive gene. Chronic simple glaucoma occurred 101 times more frequently in the high pressure corticosteroid-response population and 18 times more frequently in the moderate pressure corticosteroidresponse population.
Francois did not believe that the corticosteroid response was inherited. Several predictions should follow from either Becker's or Armaly's theory or both, and these did not stand up when tested204:
Francois continued to believe that chronic simple glaucoma was an autosomal-dominant trait with variable penetration. Other investigators also found that the steroid responsiveness of monozygotic twins was much less congruent than predicted by either Becker's or Armaly's theory (e.g., in monozygotic twins, only 62% of twin pairs had similar IOP responses to corticosteroid eyedrops; this percent was not statistically superior to that of dizygotic twins).205 Those coming to the defense of a genetic basis for the corticosteroid response then tested the reproducibility of the corticosteroid response in the same eye and in the contralateral eye.206 Using Becker's criteria for NN, NG, and GG responses, these investigators found the following:
Patients with diabetes are a subgroup developing a high frequency of markedly elevated IOPs after prolonged corticosteroid use. Both groups of patients with juvenile and adult-onset diabetes have higher baseline IOPs than the general population. Therefore, it has been argued that, not surprisingly, these patients have a higher incidence of elevated IOPs after corticosteroid treatment.207,208
Children younger than age 10, just as adults, also show a distribution of IOP responses. However, they may be slightly less likely than adults to have a marked pressure response209 (e.g., 89% of 44 children, ages 4 to 19 years, without a family history of glaucoma and who were treated with dexamethasone 0.1% eyedrops four times daily for 6 weeks had elevated IOPs of 5 mmHg or less).210
Patients exist whose elevations in IOP do not fully reverse when corticosteroids are discontinued. Therefore, some have differentiated a “markedly positive corticosteroid response” from “corticosteroid-induced glaucoma.” A markedly positive corticosteroid response is the typical reversible increase in IOP. Usually, within 1 week of discontinuing corticosteroids, these patients' IOPs have returned to baseline. In contrast, corticosteroid-induced glaucoma would be reserved for an elevation in IOP that was absent before the corticosteroid was given, was induced by the corticosteroid, persisted after the corticosteroid was discontinued, responded to those drugs used to treat chronic simple glaucoma (e.g., pilocarpine), and recurred when these therapeutic drugs were discontinued. One investigator211 found that corticosteroid-induced glaucoma occurred in 3% of his subjects with a markedly positive corticosteroid response. Other investigators212 reported that the length of corticosteroid treatment was a factor (e.g., patients who used corticosteroid drops more than 4 years did not return to baseline after discontinuation).
These effects of exogenous corticosteroids on IOP have led to speculation that endogenous corticosteroids might produce the diurnal IOP variation and elevated IOP of chronic simple glaucoma. Skeptics claim that the diurnal variation occurs with too short a cycle time to be explained by the rising and falling levels of endogenous corticosteroids (i.e., exogenous corticosteroids can take days to weeks to produce an effect on the IOP of the same magnitude as that which occurs in some patients during their diurnal variations). If corticosteroids were responsible for the diurnal curve, then the most likely candidate would seem to be an effect on the rate of aqueous humor secretion. However, the diurnal rhythm of aqueous humor flow does not seem to be corticosteroid dependent (i.e., the aqueous humor flow rate, which is twice that in the awake state as when sleeping, is not altered by oral dexamethasone 0.5 mg every 6 hours).213
However, arguments of this type have not prevented attempts to link endogenous corticosteroid secretions with the magnitude and variability of the IOP. Diverse pieces of evidence are cited by advocates. High IOP corticosteroid responders given dexamethasone have less suppression of their plasma corticosteroid levels than do intermediate or low responders.214 There is less suppression of corticosteroid secretion by dexamethasone in patients with ocular hypertension with visual field loss than in those with ocular hypertension without field loss.215,216 The IOP response to corticosteroids can at times be rapid. When four patients with open-angle glaucoma not taking their medications were given 3 mg of oral dexamethasone, their IOPs rose significantly, approximately 2 mmHg in the 0- to 4-hour period and 5.5 mmHg in the 4- to 8-hour period compared with those of control day pressures.217 Metyrapone inhibits the synthesis of endogenous cortisol. A single administration of metyrapone, 2 to 3 g, to 14 subjects with elevated IOPs in a placebo-controlled crossover trial resulted in a bilateral significant decrease in IOPs218 at 1 hour after ingestion.
How dependent is the magnitude of the corticosteroid-induced intraocular response on the drug used? In one study,219 compliance was controlled by having nurses administer all the corticosteroid eyedrops to in-patients or hospital employees with initial IOPs under 20 mmHg.21 Each of the two eyes of the volunteers received a different potent corticosteroid for 6 weeks: prednisolone acetate 1%, dexamethasone phosphate 0.1%, dexamethasone 0.1%, or prednisolone sodium phosphate 1%. All four drops gave equivalent responses. By Becker's criteria, for high, intermediate, and low responders, 6% were high responders, 18% were intermediate responders, and 76% were low or nonresponders. Another group of investigators220 used 10 high responders (by Becker's criteria) in a crossover study. Their mean millimeters of mercury ± standard error IOP rise after an average duration of 4.6 weeks was as follows: dexamethasone phosphate 0.1%, 22.0 ± 2.9; prednisolone acetate 1%, 10.0 ± 1.7; dexamethasone phosphate 0.005%, 8.2 ± 1.7; fluorometholone 0.1%, 6.1 ± 1.4; hydrocortisone 0.5%, 3.2 ± 1.0; tetrahydrotriamcinolone 0.25%, 1.8 ± 1.3; and medrysone 1%, 1.0 ± 1.3.
Long-lasting parenteral corticosteroid preparations, such as triamcinolone acetonide, triamcinolone diacetate, and methylprednisolone acetate, injected subconjunctivally, sub-Tenon's or retrobulbarly to control intraocular and intraorbital pathology have produced ipsilateral marked elevations of the IOP. One report was of 12 patients whose IOPs were 21 mmHg or less before injection.221 After injection, one patient reached 70 mmHg, two patients reached 50 mmHg, and eight patients reached 40 mmHg. Reports of attempts to shorten the effects of long-acting corticosteroid action by surgically excising them usually have been successful.222 Seven patients treated for cystoid macular edema with prednisolone acetate 1% eyedrops for 4 or more weeks did not exhibit an elevated IOP until triamcinolone acetonide, 20 to 80 mg, was injected subconjunctivally.223 Their IOPs rose to 28 to 44 mmHg. Surgical excision of the triamcinolone was performed 3 to 13 months after the injection. In six of the seven patients, the IOPs became normal within 1 week without the need for hypotensive medications. Assay of the excised material suggests great variability between individuals in their rates of triamcinolone disappearance.
Corticosteroids have been developed that, when given topically, do not produce a marked IOP elevation.224 Unfortunately, their therapeutic potency also seems to be reduced. If this observation is correct, then the same benefit can be achieved simply by diluting existing preparations. A dose-response relationship for the IOP response using dexamethasone 0.1%, 0.005%, and 0.001% has been reported.225 Armaly found that patients with a marked IOP response (greater than 15 mmHg) to 0.1% dexamethasone would have no more than a 5-mmHg increase in IOP if 0.001% dexamethasone were used. Similar results have been reported for betamethasone and triamcinolone.226 When compliance was controlled by using in-patients treated by nurses for 6 weeks four times a day,227 for every 1 mmHg increase in IOP caused by medrysone 1%, a 4-mmHg increase was created by fluorometholone 0.1% and a 7.6-mmHg increase was created by dexamethasone phosphate 0.1%. Others225 have reported that medrysone 1% increased the IOPs in only two of 24 high responders, that these two elevations were less than 15 mmHg, and that 0.001% dexamethasone produced the same elevation in IOP and had the same anti-inflammatory effect. Clobetasone butyrate 0.1%, another corticosteroid developed to have a reduced ocular-hypertensive effect, was significantly less effective than betamethasone phosphate 0.1% in improving objective signs of anterior uveitis inflammation.228 However, other studies have found the two drugs equivalent in treating anterior uveitis,229 episcleritis,230 and postcataract surgery intraocular inflammation.231 Despite some inconsistencies found in clobetasone's efficacy, all studies have indicated that clobetasone has less of an ocular-hypertensive effect. Rimexolone, commercially available as a 1% suspension, and fluorometholone 0.1% appear near equal in their ability to elevate the IOP.232 In 33 subjects who had responded to topical corticosteroids (either dexamethasone sodium phosphate or prednisolone acetate) with IOP elevations of at least 10 mmHg, 10 subjects responded with a 10-mmHg or more increase in IOP to rimexolone 1% four times a day for 6 weeks and seven subjects responded to fluorometholone 0.1% given in the same manner.
These weaker corticosteroids provide only relative safety with regard to eliciting a hypertensive response. Both the magnitude and rate of the IOP elevation tend to be less, but this generalization does not always hold for the individual patient. Thirty subjects who responded to 0.1% dexamethasone sodium phosphate drops four times daily with an IOP rise of 5 to 15 mmHg and 13 subjects who responded with an IOP rise of more than 15 mmHg were rechallenged with fluorometholone 0.1% four times a day for 3 to 11 weeks.233 Seventeen had IOP elevations of less than 5 mmHg, but 23 had IOP elevations of 5 to 15 mmHg and three had intraocular elevations of more than 15 mmHg. When fluorometholone 0.25% was instilled in one eye and dexamethasone sodium phosphate 0.1% was instilled in the contralateral eye of 14 corticosteroid responders, both drugs elevated the IOP but fluorometholone did significantly less so.234 The mean ± standard deviation respective fluorometholone 0.25% versus dexamethasone 0.1% mmHg elevations at 2, 4, and 6 weeks were 4.4 ± 4.8 versus 8.1 ± 7.1, 6.6 ± 5.0 versus 10.0 ± 5.8, and 8.1 ± 5.2 versus 11.6 ± 5.6, respectively. The rate at which the IOP rises is slowed when the weaker corticosteroids are used; in subjects requiring a mean response time of 3 weeks to develop a given IOP elevation from dexamethasone sodium phosphate 0.1%, a statistically significant longer time, 4 weeks, was needed when fluorometholone 0.1% was applied.235
Rapidly inactivated moieties (so-called soft drugs) have been developed to minimize the toxicities and side effects of corticosteroids. One of these is loteprednol etabonate. Drugs of this type are intended to be active only at the site to which they are directly applied and are intended for topical application. Loteprednol etabonate has more than four times dexamethasone's affinity for glucocorticoid receptors, but it contains a 17 betachloromethyl ester, which allows it to undergo rapid hydrolysis into inactive components.236 The drug has good stability at pH 5.9. Presumably, the drug undergoes inactivation hydrolysis by tissue esterases, but spontaneous hydrolysis due to elevations in pH and temperature also may play a role when the drug is applied. Two hundred ninety-three volunteers were treated with a 0.5% preparation four times a day for 6 weeks or with placebo.97 None of the drug-treated subjects had an IOP rise of 10 mmHg develop; 10% had an increase of 6 mmHg or more. In another study, 14 subjects who had developed an IOP elevation of 6 mmHg or greater when treated 6 weeks or less with topical dexamethasone 0.1% or prednisolone acetate 1% were given loteprednol etabonate 0.5% four times a day for 6 weeks. Loteprednol etabonate produced a pressure increase of 6 mmHg or greater in four subjects, with one of these subjects having an elevation greater than 15 mmHg.237 In a study of 0.2% loteprednol etabonate given four times a day,99 7% of subjects demonstrated a 10-mmHg or greater increase in IOP. It is not clear whether this drug possesses any real advantage over the other agents designed to minimize the IOP elevation despite the attractiveness of the concept behind its development.
Anterior uveitis can develop in subjects while they are receiving or just after discontinuing corticosteroid eyedrops. This phenomenon was first described in subjects receiving corticosteroid eyedrops to assess their IOP responses.238 Another group of investigators found that two (4%) of 54 normotensive volunteers had acute iritis develop within 72 hours of discontinuing corticosteroid eyedrops given for prolonged (i.e., weeks) periods.219 One subject in whom iritis developed was rechallenged with a different corticosteroid preparation and had a recurrence239; another subject, challenged only with the vehicles, did not. The phenomenon occurred primarily in blacks and was not associated with either any specific corticosteroid or with any specific type of IOP response. A familial form of this response has been reported240; when seven black siblings (six sisters, one brother) were tested, four sisters had iridocyclitis develop.
MELTING OF THE CONJUNCTIVA AND CORNEA
Subconjunctival injections of triamcinolone have been associated with thinning241 and melting242 of the conjunctiva. Corticosteroids, given to treat corneal alkali burns, may be associated with corneal melting; concomitant use of drops of potassium ascorbate (vitamin C) 10% and oral vitamin C, 1 g/day, may protect against this complication.243
DEPIGMENTATION OF THE DERMIS
Injection of corticosteroids into the lids has been associated with depigmentation.244
RETINOPATHY OF PREMATURITY
Antenatal maternal use of dexamethasone has been associated with a decreased severity of retinopathy of prematurity.245 Postnatal use of corticosteroids has been associated with development of a more severe retinopathy.246
Chronic application of corticosteroid eyedrops rarely has been claimed to cause a Cushingoid-like condition in children.247 However, the proof is not strong and the evidence that systemic absorption is sufficient to produce pituitary suppression is lacking.248
4. Sinzinger H, Virgolini I, Ravscha F et al: Glucocorticoid- treatment does not influence the synthesis of thromboxane B2 and bicyclo-PGE2 in humans. Prostaglandins Leukot Essent Fatty Acids 35:135, 1989
10. Ui T, Mitsunaga M, Tanaka T et al: Determination of prednisone and prednisolone in human serum by highperformance liquid chromatography-especially on impaired conversion of corticosteroids in patients with chronic liver disease. J Chromatography 239:711, 1982
35. Weijtens O, van der Sluijs FA, Schoemaker RC et al: Peribulbar corticosteroid injection: Vitreal and serum concentrations after dexamethasone disodium phosphate injection. Am J Ophthalmol 123:358, 1997
44. Kyle V, Cawston TE, Hazleman BL: Erythrocyte sedimentation rate and C-reactive protein in the assessment of polymyalgia rheumatica/giant cell arteritis on presentation and during follow-up. Ann Rheum Dis 48:667, 1989
47. Matzkin DC, Slamovits TL, Sachs R et al: Visual recovery in two patients after intravenous methyl prednisolone treatment of central retinal artery occlusion secondary to giant-cell arteritis. Ophthalmology 99:68, 1992
52. Salvarani C, Boiardi L, Macchioni P et al. Role of peripheral CD8 lymphocytes and soluble TL-2 receptor on predicting the duration of corticosteroid treatment in polymyalgia rheumatica and giant cell arteritis. Ann Rheum Dis 5: 640, 1995
83. Calao A, Lastoria S, Ferone D et al: Orbital scintigraphy with (111I n-diethylenetriamine pentaacetic acid-D-Phe1) octreotide predicts the clinical response to the corticosteroid therapy in patients with Graves' ophthalmopathy. J Clin Endocrin Metab 83:3790, 1998
91. Marcocci C, Bartalena L, Bogazzi F et al: Orbital radiotherapy combined with high dose systemic glucocorticoids for Graves' ophthalmopathy is more effective than radiotherapy alone: Results of a prospective randomized study. J Endocrinol Invest 14:853, 1991
97. Dell SJ, Shulman DG, Lowry GM et al: A controlled evaluation of the efficacy and safety of loteprednol etabonate in the prophylactic treatment of seasonal conjunctivitis. Am J Ophthalmol 123:791, 1997
98. Shulman DG, Lothringer LL, Rubin JN et al: A ramdomized, double-masked, placebo controlled parallel study of loteprednol etabonate 0.2% in patients with seasonal allergic conjunctivitis. Ophthalmology 106:362, 1999
99. Friedlaender MH, Howes J: A double-masked, placebo-controlled evaluation of the efficacy and safety of loteprednol etabonate in the treatment of giant papillary conjunctivitis. Am J Ophthalmol 123:455, 1997
107. Power WJ, Ghoraishi M, Merayo-Lloves J: Analysis of the acute ophthalmic manifestations of the erythema multiforme/Stevens- Johnson/toxic epidermal necrolysis disease spectrum. Ophthalmology 102:1669, 1995
128. Shaw SM, McHugh JD, Spalton DJ: The effects of subconjunctival betamethasone on the blood aqueous barrier following cataract surgery: A double-blind randomised prospective study. Br J Ophthalmol 76:475, 1992
131. Gartry DS, Kerr Muir MG, Lohmann CP et al: The effect of topical corticosteroids on refractive outcome and corneal haze after photorefractive keratectomy: A prospective, randomized, double-blind trial. Arch Ophthalmol 110:944, 1992
132. O'Bart DPS, Lohmann CP, Klonos G et al: The effects of topical corticosteroids and plasmin inhibitors on refractive outcome, haze, and visual performance after photorefractive keratectomy. Ophthalmology 101:1565, 1994
189. Southren, AL, Gordon GG, Munnangi PR et al: Altered cortisol metabolism in cells cultured from trabecular meshwork specimens obtained from patients with primary open-angle glaucoma. Invest Ophthalmol Vis Sci 24:1413, 1983
231. Ramsell TG, Bartholomew RS, Walker SR: Clinical evaluation of clobetasone butyrate: A comparative study of its effects in post-operative inflammation and on intraocular pressure. Br J Ophthalmol 64:43, 1980