The lens and cornea are avascular structures that are supplied nonetheless with nutrients and oxygen by the aqueous humor, which also serves to remove metabolic waste products as it is drained from the eye. Aqueous humor is produced within the 70–80 ciliary processes that are radial folds in the pars plicata of the ciliary body. Both active and passive mechanisms are involved in aqueous production. The two-layered epithelium covering the ciliary processes is sealed together by tight junctions (zonula occludens) forming the blood–aqueous barrier between the posterior chamber and blood vessels that pass through the ciliary body and its processes [35, 36].

Passive mechanisms of aqueous production include diffusion and ultrafiltration:

Active mechanisms of aqueous production involve secretion:

Aqueous production starts in the blood with the final product eventually reaching the anterior chamber. The route is as follows (Fig. 23.1):

Blood plasma ultrafiltrate > capillary endothelial cells > ciliary process interstitial tissue > pigmented epithelium > nonpigmented epithelium (active transport) > posterior chamber > pupil > anterior chamber.

Aqueous humor is a “complex mixture of electrolytes, organic solutes, growth factors, and other proteins that supply nutrients to the nonvascularized tissues of the anterior chamber (i.e., trabecular meshwork (TM), lens, and corneal endothelium).” [37] Aqueous is produced at a rate of 2.0–3.0 μl/min. or 1.5 % of anterior chamber volume/min. The anterior chamber holds approximately 250 μl of aqueous humor, and the posterior chamber about 60 μl of aqueous humor.

All the aqueous is replenished about every 1.5–2 h during the daytime hours. Relative to blood plasma, aqueous has approximately 15 times greater ascorbate and much less protein (0.02 % in aqueous versus 7 % in plasma). Aqueous humor formation decreases with sleep [38] (suppression of 45 ± 20 %) and advancing age [39] (decrease of 2.5 % per decade). Men and women on average have the same rate of aqueous flow [38].

After entering the anterior chamber and angle structures, there are two pathways for the outflow of aqueous humor from the eye [40]. The “conventional” or trabecular-meshwork outflow path is said to be “pressure dependent.” This pathway is usually thought to account for about 80 % of aqueous outflow. The route of aqueous outflow through the trabecular meshwork is as follows:

Anterior chamber > trabecular meshwork > Schlemm’s canal > scleral collector channels > episcleral aqueous veins > anterior ciliary veins > superior ophthalmic veins > venous system

The “unconventional” or uveoscleral outflow path is said to be “pressure independent.” This pathway accounts for the remaining 20 % of aqueous outflow, although there is conjecture that the actual number is higher. The root of aqueous outflow through the uveoscleral pathway is as follows:

Anterior chamber > iris root > ciliary body interstitial tissue > suprachoroidal space (negative pressure) > sclera? vortex veins? along the nerves? > posterior ciliary veins > venous system

The majority of glaucoma medications are eye drops that are placed on the surface of the eye. They enter through the conjunctiva and cornea, with the fornix serving as a reservoir. Usually, drops are applied in great excess, and most of the medication does not enter the eye. The root of absorption into the eye from an eye drop is from the conjunctival cul-de-sac, then becoming part of the tear film (mucin, water, and oil) that is 7–9 μm thick, then passively diffusing into cornea and conjunctiva, and then becoming part of the aqueous humor. Peak concentration of drug in the aqueous humor takes from 30 to 60 min. Glaucoma drugs are generally not uniformly distributed throughout the eye, because of both physical barriers like iris, lens, and ciliary body, as well as circulatory factors like bulk flow of aqueous humor and the blood circulation. In addition, medications often differentially bind to tissues, especially those containing melanin [40].

The great majority of the drop applied to the eye (some estimate this to be 80 % or greater) exits via the lacrimal drainage apparatus and does not enter into the eye at all (see Sidebar 23.3). The cul-de-sac holds at most 30 μL of an eye drop. Most of the drop drains from the cul-de-sac of the eye within the first 5 min. Blinking eliminates about 2 μL of fluid from the cul-de-sac [41]. The route of nasolacrimal duct drainage is conjunctival cul-de-sac > inferior and superior nasal puncta > canaliculi > lacrimal sac > nasolacrimal duct > valve of Hasner > inferior nasal turbinate > then absorption into the vascularized nasopharyngeal mucosa or swallowed [40, 118].

What are the characteristics of an ideal drug for glaucoma therapy? Primarily, both the physician and the patient want a drug that is efficacious in preserving visual acuity and visual field, that is, they want a drug that works! They want drugs that will slow down or halt the progression of glaucoma, both structurally and functionally. Second, glaucoma medications ideally should be extremely “patient friendly,” so that patients will be inclined to use them. The former US Surgeon General, C. Everett Koop, purportedly has commented “Drugs don’t work in patients who don’t take them,” a succinct statement of the importance of medication compliance.

Tables 23.2 and 23.3 summarize both the efficacy- and compliance-related factors desirable in glaucoma medications. Some of these attributes are present in the medications available today; some are on the horizon and some in the more distant future. As we look at the different classes of glaucoma medications and evaluate their pros and cons, keep in mind the various factors listed in these two tables.

Tolerability and safety also are important aspects for choosing and continuing glaucoma medications. While tolerability is often evident based on patent symptoms (e.g., hyperemia, itching), potential safety issues may be substantially more subtle. Often new and established patients complete medical history questionnaires that also contain a “review of systems.” These documents are helpful because they suggest to the treating physician areas of possible systemic side effects with potential glaucoma medications, but they cannot entirely substitute for a thorough interview with the patient by the eye doctor. The danger of increasingly high clinic volumes of patients and shorter patient interviews is that important systemic medical issues will be missed that may result in the patient sustaining side effects. As glaucoma consultants, we have too often seen referred patients who come into our offices on topical beta-blockers, started by their community ophthalmologists who had not inquired about any history of asthma or chronic obstructive pulmonary disease (COPD). While a detailed medical history questionnaire is very helpful, the personal interaction between the patient and the physician results in a better likelihood of uncovering potential medication/medical history interactions and adverse effects.

As explained earlier in this chapter, the rationale for treatment of glaucoma is based on the assumption that lowering intraocular pressure currently is the most important thing we can do for the patient. At the present time, benefits such as improved blood flow or direct neuroprotection of retinal ganglion cells are secondary considerations when choosing primary monotherapy for patients with ocular hypertension or open-angle glaucoma. Pilocarpine was introduced in the late nineteenth century to treat glaucoma. Figure 23.6 shows the timeline of development for glaucoma medications used in the USA.

Medications for the treatment of glaucoma may be divided into six classes (seven classes if we include fixed combinations of products) as shown in Table 23.4. Figure 23.7 depicts the relative potencies of glaucoma agents with respect to lowering intraocular pressure.

A meta-analysis of 28 randomized clinical trials [95] summarizing the relative effectiveness of the more commonly used glaucoma medications for lowering IOP is shown in Table 23.5.

As early as 1967, it was found that the nonselective beta-blocker propranolol reduced IOP when administered intravenously [96]. A paper by Katz et al. in 1976 demonstrated that topical timolol reduced IOP in normal volunteers [97]. From the late 1970s through the beginning of the twenty-first century, beta-adrenergic antagonists (“beta-blockers”) became the mainstay of initial clinical treatment for lowering intraocular pressure in glaucoma patients. Timoptic was introduced by Merck in 1978. This and other nonselective beta-blockers reduce IOP by about 20–25 % by suppression of aqueous humor formation. These drops inhibit cyclic adenosine monophosphate (cAMP) in ciliary epithelium. Carteolol has “intrinsic sympathomimetic activity” (ISA) and subsequently may have less effect on lipid profiles than other nonselective beta-blockers [98–100]. Only the generic version of carteolol is sold in the USA. There have been reports of anterior uveitis in patients using metipranolol, persisting even after its reformulation [101–108]. This drug has fallen out of favor in the USA.

Beta-blockers are available in solutions, gel-forming solutions, and suspensions. In the USA, the following brands are available:

All are available in both 0.25 and 0.50 % concentrations except as noted. Most of the pressure reduction activity of these drugs is mediated through their action on the beta-₂ receptors. Nonselective beta-blockers lower IOP on average 20–25 %. Although most ophthalmologists prescribe the 0.50 % concentration of the nonselective beta-blockers, adequate aqueous concentrations are achieved with lesser amounts of drug, and the 0.25 % commercial preparations are close to the peak of the dose–response curve [109]. These high aqueous concentrations of drug, even with the 0.25 % dose, explain why betaxolol, which is highly selective for the beta-1 receptor, reaches a sufficient concentration to lower intraocular pressure, which is mediated through the beta-2 receptor [110–113]. Betaxolol lowers IOP 15–20 %, but it does not lower pressure as much as the nonselective beta-blockers (20–25 %) [114–117].

All of the listed medications use benzalkonium chloride as the preservative, except Timoptic-XE and Timolol GFS, which use benzododecinium bromide 0.012 % (also a quaternary ammonium compound), and Timolol Ocudose, which is supplied in preservative-free unit-dose 0.2 ml containers. Approved dosing per the package inserts is every 12 h, except for once-per-day dosing (usually in the morning) for Timoptic-XE and Timolol GFS and for Istalol, which contains sorbic acid and is more lipophilic. The greatest activity of the medications occurs between 2 and 4 h after instillation.84 Beta-blockers take about 4 weeks to fully leave the body.

Beta-blockers have little activity during the evening and nighttime hours, when aqueous production by the ciliary body is reduced because of natural circadian factors [38, 63, 118]. Circulating catecholamine concentration (especially epinephrine) is lowest at night, resulting in a deceased sympathetic tone that appears necessary for the production of aqueous. It has been hypothesized that if there is a low sympathetic tone, little aqueous is being produced above a basal level, and thus there is little for the beta-antagonist drugs to “block.” [119] Thus, beta-blockers do not provide effective 24-h control of intraocular pressure [63]. Therefore, drugs that affect aqueous outflow (e.g., hypotensive lipids), rather than aqueous production, may be more effective agents at reducing nocturnal intraocular pressure.

While many beta-blockers in solution form are approved for twice-a-day dosing, most of the benefit of this class of medication can be achieved by a single morning dose [120–122]. Because beta-blockers do not appear to work as well at night, the evening dose is superfluous and increases the risk of side effects. Also, there is some concern that nocturnal effects of beta-blockers may lower ocular perfusion pressure [71, 79, 123–135]. This would be especially worrisome in patients with normal-tension glaucoma. There is some evidence that betaxolol may increase retinal and choroidal blood flow by acting on voltage-dependent calcium channels [132, 136–145]. It may thus have some neuroprotective effect separate from its IOP-lowering properties [146, 147]. Levobunolol has an active metabolite (dihydrobunolol) that may contribute to its longer duration of activity when compared with timolol [148, 149].

Topical beta-blockers have few ocular side effects [150, 151]. Istalol and Betimol sting more than other products, and Timoptic-XE and timolol gel-forming solution (Timolol GFS) may blur vision upon installation. Betoptic S comes as a suspension that helps minimize stinging [152]. All topical beta-blockers cause some corneal anesthesia and punctuate staining [153]. Although their ocular side effects may just be a nuisance, their systemic side effects may be dramatic and even life-threatening in some instances. As described earlier in this chapter, there are two major forms of beta-agonist activity: beta-₁, which largely affects cardiac muscle, and beta-₂, which largely affects pulmonary function. Well-known systemic side effects of beta-blockers limit their use in glaucoma patients who suffer from respiratory ailments. Betaxolol is a relatively selective beta-₁ antagonist and is thought to be safer than the nonselective beta-blockers for use in patients with pulmonary conditions. However, no beta-blocker is safe enough to use in patients with severe reactive airway disease. Fortunately, for these patients, we have a number of good alternative preparations for both primary and adjunctive glaucoma therapies.

Recently, congestive heart failure (CHF) has become only a relative contraindication to the use of oral beta-blockers systemically [154–158] and similarly for the use of beta-blocker eye drops. Nonetheless, it is prudent to check with the patient’s family physician or cardiologist before instituting such therapy in patients with known CHF. Steps taken to limit the systemic side effects of beta-blockers include reduced concentrations (0.25 % compared to 0.5 %) and beta-₁-selective preparations (e.g., betaxolol), along with nasolacrimal duct occlusion. Topical beta-blockers used to treat glaucoma are less effective at lowering IOP in patients already taking oral beta-blockers [159, 160]. This “double dosing” by topical and oral routes significantly increases the risk of systemic side effects. In most cases, patients using oral beta-blockers should not receive topical beta-blockers as their initial glaucoma therapy.

Because of these scientific issues and medical–legal concerns, when prescribing beta-blocker eye drops, physicians must be acutely aware of their potential interaction with the patient’s general medical problems. Before initiating therapy with a topical beta-blocker, the patient’s pulse rate and systemic blood pressure should be checked and recorded in the medical record. This should be repeated at subsequent office visits. Ask about a history of heart block, CHF, bradycardia, and any incidences of fainting [161]. Patients also should be asked about smoking, asthma, and COPD [161]. Patients using topical beta-blockers may exhibit decreased exercise tolerance, less so with beta-₁-selective agents [162–167]. As with oral beta-blockers, lipid profiles may be affected unfavorably with decreased high-density lipoprotein (HDL) and increased triglycerides [168]. The patient’s lipid profile may be less affected with Ocupress, which has intrinsic sympathetic activity (ISA) [98]. Topical beta-blockers can produce central nervous system effects including depression, anxiety, confusion, hallucinations, drowsiness, weakness, memory loss, and impotence, and frail elderly patients can become “weak and dizzy.” [169] Finally, alopecia has been reported with topical beta-blockers (package insert, Timoptic Ophthalmic Solution, Merck & Co., West Point, Pennsylvania).

Beta-blockade from topical preparations may mask symptoms of diabetic-related hypoglycemia [170] and cause impaired glucose tolerance (inhibit lipolysis and glycogenolysis) [171–179]. Beta-blockade can mask symptoms of thyrotoxicosis and abrupt cessation may cause it [180–184].

Patients with dark irises, including African-Americans, may have less intraocular lowering pressure effect from beta-blockers because of binding of the medication to melanin [185–188]. (This is also true for pilocarpine [189] and adrenergic agonists [190] used to treat glaucoma.)

Topical beta-blockers applied unilaterally may exhibit some “crossover” of IOP lowering in the contralateral eye [191–194]. This makes a “one-eyed” trial of a beta-blocker somewhat difficult to evaluate.

These agents generally do not exhibit tachyphylaxis or short-term “escape,” although one early report [66] suggests tachyphylaxis does occur. There is more controversy, however, about long-term drift of IOP with topical beta-blockers [67, 68, 195–199]. A recent well-designed study suggests that long-term drift of beta-blockers is an urban myth among ophthalmologists [69].

Hypotensive lipids (HLs) for treating OAG fall into three subcategories. The prostaglandin analogues (PGAs) include latanoprost (Xalatan 0.005 %) and travoprost (Travatan and Travatan Z, both 0.004 %) and tafluprost (Zioptan 0.0015 %); there is one prostamide – bimatoprost (Lumigan 0.03 %); the docosanoid class is represented by unoprostone isopropyl (Rescula 0.15 %). They are all derivatives of prostaglandin F_2α, based on pioneering work by Bito, Stjernschantz, and Camras [200–203]. This class of glaucoma medications increases both trabecular-meshwork and uveoscleral outflow [204–206] and is less affected by circadian variations in aqueous production than the beta-blockers [91, 207]. Although these drugs have this dual mechanism of action, most of the increased outflow facility can be attributed to their effects on the pressure-independent uveoscleral outflow pathway [206].

Bimatoprost, latanoprost, travoprost, and tafluprost are approved as first-line agents by the FDA in the USA. Unoprostone isopropyl is very infrequently used as either first-line or adjunctive therapy in the USA, although it is popular in Japan where it was developed. It requires twice-per-day dosing (the other hypotensive lipids are dosed once per day) and lowers IOP to a much lesser extent than the other products of this class. Unoprostone was removed from the US market for several years but has been recently (2012) reintroduced. According to the package insert, “Unoprostone also lowers IOP by its effect on FP receptors and may have a local effect on BK (Big Potassium) channels and ClC-2 chloride channels, but the exact mechanism is unknown.”

Unoprostone has a very limited market share in the USA, likely because of its lesser ability to lower IOP and its increased (BID) dosing interval compared to the other hypotensive lipids.

Unless noted otherwise, the phrase “hypotensive lipids” (HLs) subsequently used in this chapter shall refer only to bimatoprost, latanoprost, travoprost (with or without BAK), and tafluprost.

The hypotensive lipids have become accepted by ophthalmologists as first-line therapy in the USA, Europe, and South America for most forms of open-angle glaucoma. Cost considerations and later introductions have limited their use in Asia and Africa, where beta-blockers and miotics are still popular.

Early drugs developed to control glaucoma either decreased aqueous production (beta-blockers) or increased conventional, trabecular-meshwork-mediated outflow (miotics). The major innovation of the HL medications is their ability to increase uveoscleral outflow (with a small to moderate effect on TM function as well). The prototype molecule, prostaglandin F_2α, does itself lower IOP, but it was found to produce too much inflammation and ocular symptoms intolerable to patients and was thus never developed into a commercial product [201]. Molecular engineering has led to derivatives of this parent compound with acceptable efficacy and safety profiles.

Most scientists agree that the mechanism of action for increased uveoscleral outflow involves activation of the FP class of prostaglandin receptors found at the iris root and ciliary body. Stimulation of these receptors results in the upregulation of matrix metalloproteinases (which are similar to collagenases) that in turn result in a remodeling of extracellular matrix and widening of intermuscular spaces in the ciliary body. Aqueous may then more easily exit the eye around the cellular structures [208–211]. A study in FP-receptor-deficient mice, with presumed functional uveoscleral outflow pathways, showed no affect of any of the four hypotensive lipids on reduction of intraocular pressure [212].

How do the commercial preparations differ? As described earlier in this chapter, all of the HLs function as prodrugs. Latanoprost, travoprost, tafluprost, and unoprostone are esterified derivatives of prostaglandin F_2α, while bimatoprost contains an amide group in the same position as the isopropyl ester of the other molecules. The chemical structures of the compounds (except for tafluprost) are shown in Fig. 23.8.

The ester or amide portions of the molecules are cleaved by abundant corneal enzymes, with the free acid forms of the lipids entering into the anterior chamber in sufficient concentration to activate the FP prostaglandin receptors of the ciliary body that is bathed in aqueous. The ester moiety is easily cleaved, allowing the drugs to be manufactured at relatively low concentrations (latanoprost 0.005 %, travoprost 0.004 %, tafluprost 0.0015 %; unoprostone has a higher concentration 0.15 % but is a relatively ineffective drug – we will not consider it in subsequent discussion). The amide group on the bimatoprost molecule is not as easily separated from the free acid [213]. The clinical product used to treat glaucoma has a concentration (0.03 %) six to eight times greater than the other HLs. This increased concentration allows sufficient free acid to enter the eye to effectively activate the FP receptors [214]. Thus, it would appear that bimatoprost lowers IOP by uveoscleral outflow in the same manner as the other HLs.50 [215], However, the manufacturer (Allergan, Irvine, California) has sponsored research suggesting that the entire bimatoprost molecule acts primarily by passing through intact sclera to activate “prostamide” receptors on the ciliary body with subsequent increase in uveoscleral outflow facility [216]. There is one study using a preparation of feline iris sphincter that demonstrates binding of the full bimatoprost molecule to receptors that appear to be distinct from prostaglandin FP receptors [217]. (And recall the murine study described above that showed no reduction in IOP when bimatoprost was applied to the eyes of FP-receptor-deficient mice.186) No published studies have found these receptors in the ciliary body nor has a prostamide receptor been “cloned” using standard biological techniques. Bimatoprost has been used clinically since 2001. While prostamide receptors may exist, clinically bimatoprost seems to function as a prostaglandin analogue, with increased activity of both the conventional and nonconventional outflow pathways. The authors look forward to a research that may demonstrate the presence of the receptors in the relevant tissues and their binding and activation by the nonhydrolyzed version of bimatoprost. From a clinical perspective, in our practices, we consider the three major HLs as belonging to the same class of medication (Fig. 23.9).

Variations in the PGF_2α molecule result in changes in potency and side effects. Latanoprost was the first HL to be developed commercially (by Pharmacia, now Pfizer). It became available in the USA in 1996. In order to reduce the hyperemia associated with PGF_2α, the unsaturated (double) bond between carbons 13 and 14 was saturated. This resulted in some loss of potency but by reducing hyperemia made the drug acceptable to real-world patients. Because there is no major clinical difference in IOP-lowering efficacy whether this class of drugs is dosed daytime or nighttime, it has become customary to prescribe the HLs at bedtime, so that the majority of the immediate hyperemia associated with drug dosing occurs while the patient is asleep. These drugs do have some “chronic” hyperemia that tends to subside over several months of use. Occasionally, patients prefer morning dosing, which is acceptable from an efficacy perspective. Clinical IOP-lowering efficacy of the HLs is better with QD dosing (at the commercially available concentrations) rather than BID dosing [218–220]. Systemic half-life of the drugs is brief (e.g., latanoprost 17 min). The peak plasma concentration of latanoprost is about 10⁻¹⁰ M or less, and there is little if any effect of the drugs on the IOP of the contralateral eye when dosed unilaterally [221].

Both bimatoprost and travoprost came to the US market in March 2001. Recently, two others were approved. Tafluprost (Zioptan) is available in a single-use, preservative-free form. Unoprostone (Rescula) which had been withdrawn from the US market but reintroduced recently (2012) is a significantly weaker drug in the same class of FP-receptor agonists. It is purported to have additional mechanisms of action which may distinguish it from its more powerful competitors in this class of agents. To improve efficacy, Alcon Laboratories modified the PGF_2α molecule to create travoprost by adding a CF3 on the unsaturated benzene ring. This allows for a tighter bonding of the travoprost free acid to the FP receptors [222, 223]. This results in a longer-duration, clinically useful, IOP-lowering effect of both original travoprost and the BAK-free version [91, 92, 224]. This could be important in patients who occasionally miss doses.

Alcon kept the double bond between carbons 13 and 14, increasing efficacy but resulting in more hyperemia for travoprost than found in the commercial latanoprost preparation (the newer travoprost without BAK has slightly less hyperemia than the original product, but both versions produce more clinical hyperemia than latanoprost) [48, 225]. Most of the hyperemia associated with the HLs results from dilated conjunctival vessels in response to direct activation of FP receptors found in the vasculature muscle walls [226]. This is not a toxic or allergic response but is, however, of cosmetic concern to the patient. Allergan, when designing bimatoprost, also kept the unsaturated double bond between carbons 13 and 14 and added the amide group, where the other products have the isopropyl ester. As discussed earlier, bimatoprost has a six- to eightfold greater concentration than the other HLs. This may be related to the clinical observation that bimatoprost causes more red eye than the other two products [225, 227, 228]. A newer version of bimatoprost [229] lowers the IOP with a decreased concentration (by one-third) and increased BAK resulting in less hyperemia likely due to decreased concentration of drug (0.01 % compared to 0.03 for the original concentration) while increasing the penetration of the drug by increasing the BAK fourfold. This is an example of a logical, obvious trade-off to get less side effect with less drug, but it occurs at the expense of using BAK (see Sidebar 23.4). (Latanoprost causes the least conjunctival hyperemia, a bit more with travoprost.) [230]

Additionally, the HLs activate more than just the FP class of prostaglandin receptors. The EP class of prostaglandin receptors mediates inflammatory by-products. Funk in a recent (2001) review [231] of eicosanoid biology said, “Prostaglandins act both at peripheral sensory neurons and at central sites…to evoke hyperalgesia. Recent data support involvement of … EP₁ receptors in pain” [232–234]. Figure 23.10 shows the relative binding affinities for the HLs at both FP and EP₁ receptors. It is desirable for glaucoma drugs to have high binding affinities for the FP receptors (to lower IOP) and low binding affinities to the EP₁ receptors (to minimize inflammation). Here, we see that bimatoprost is a better agonist at the EP₁ receptor than the three other products. Thus, the increased hyperemia seen with bimatoprost may result from both its affinity for EP₁ receptors (inflammatory hyperemia) and its relatively higher concentration (vasodilatory hyperemia in response to FP-receptor activation). From a clinical perspective, most patients find that the hyperemia of all the HLs decreases sufficiently over a month or two to become cosmetically acceptable [235, 236]. We have found that prospectively discussing conjunctival hyperemia with patients before initiating HL therapy greatly decreases their emergent phone calls and “drop-in” office visits.

While conjunctival hyperemia is clinically apparent to the patients, their companions, and the treating physician, other aspects of ocular surface disorders may be less obvious but even more important to chronic disease management. As noted earlier in this chapter, the BAK found in most glaucoma medications may play an important role in causing the signs and symptoms of ocular surface disease and subclinical inflammation (see Sidebar 23.4). The newer formulation of travoprost without BAK, with the preservative SofZia, seems equally effective as the original product but with improved ocular surface health [11, 48]. New glaucoma products without BAK are under development by the pharmaceutical industry.

The three hypotensive lipids lower intraocular pressure on average between 25 and 30 %. This is greater than the nonselective beta-blockers, which lower IOP between 20 and 25 %. The only randomized controlled clinical trial directly comparing the three agents found them to be on average about the same in IOP-lowering efficacy (with a slight, 0.5 mmHg, advantage to bimatoprost) and confirmed the clinical impression that latanoprost had the lowest rate of conjunctival hyperemia [61]. The previously mentioned meta-analysis by van der Valk of 28 clinical studies also shows the HLs to be roughly equivalent in their ability to lower intraocular pressure [95]. A more recent meta-analysis of nine studies that had to compare at least two prostaglandin analogues as monotherapy found that travoprost and bimatoprost lowered IOP −0.98 mmHg (95 % Cl: –2.08; 0.13; p = 0.08) and −1.04 mmHg (95 % Cl: –2.11; 0.04; p = 0.06), respectively, when compared with latanoprost [237].

The HLs have relatively flat IOP curves over 24 h, demonstrating both low circadian IOP fluctuation and, unlike the beta-blockers, effective diurnal and nocturnal IOP control [63, 91, 92, 207, 228]. They do not evidence short-term escape or long-term drift [238–240].

HLs have few systemic side effects. HLs should generally be avoided in women of child-bearing age without adequate birth control or who are pregnant, because although plasma concentrations of topical HLs are low, there is a theoretical risk of abortion due to the action of prostaglandin molecules on uterine smooth muscle [241–243]. There are no references in the ophthalmic literature that this has ever happened. There are some reports of muscle aches, gastrointestinal upset, and influenza-like symptoms with the use of topical HLs to treat glaucoma [244, 245].

There have been case reports of reactivation of clinically dormant herpes simplex virus (HSV) keratitis by patients who begin HLs, confirmed in some cases upon rechallenging the eyes involved. We generally avoid using HLs in patients with a history of HSV keratitis [246–252].

The use of HLs has been associated with cystoid macular edema (CME) especially in patients who have undergone cataract surgery and intraocular lens implantation, more often if the posterior capsule is ruptured [253–255] (some researchers feel that postcataract CME may be related to eye drop preservatives [256]). Sometimes, this CME is clinically apparent with complaints of reduced visual acuity and contrast sensitivity; sometimes, this CME is picked up by fluorescein angiography or optical coherence tomography. In our experience, this HL-related CME almost always improves by withdrawal of the agent and treatment with topical steroids and nonsteroidal anti-inflammatory drugs (NSAIDs). In our practices, we generally try to discontinue HLs several weeks before cataract surgery and try to avoid reinstituting them for about 3 months after surgery. Fortunately, many open-angle glaucoma patients undergoing phacoemulsification cataract surgery have a “honeymoon” period of reduced intraocular pressure lasting several months or more and may require less glaucoma medical therapy during that time [257–259]. Thus, they suffer no harm by temporarily suspending the HL postoperatively. If the posterior capsule is broken, we may avoid the further use of HLs at all. To attempt to prevent CME, many anterior segment surgeons pretreat all patients about to undergo cataract surgery with topical steroids and NSAIDs, especially if they have been using hypotensive lipids [260]. Some cataract surgeons continue to use previously prescribed HLs throughout the perioperative period [261]. The authors do not advise doing so unless IOP control is lost without the use of HLs in the immediate postoperative time frame.

There have been several studies reporting an association of HLs with the development of anterior uveitis or the worsening of preexisting uveitis [262–270]. Because prostaglandins are proinflammatory compounds, most glaucoma experts do not recommend their use in patients with uveitis, uveitic glaucoma, or neovascular glaucoma. There appears to be little role for HLs in managing an attack of acute angle-closure glaucoma, but these agents do appear useful for various forms of chronic angle-closure glaucoma [271, 272]. Finally, the three hypotensive lipids are not additive in their effects, that is, using, for example, bimatoprost in the AM and latanoprost in the PM will result in slightly worse lowering of IOP than using either drug alone [273].

There are several ocular-related side effects of HLs that may be described as cosmetic. Hyperemia, discussed previously, may be intolerable to some patients. Proactive discussion of this issue may improve patient adherence with HLs. A small to moderate percentage of patients (<20 %) using the HLs develop permanent, clinically noticeable changes in iris color, with increasing plumpness of melanocytes (stimulation of melanogenesis) perhaps through an effect on tyrosinase, but no increase in their number. This change in iris color reflects no mutagenesis but is of cosmetic concern to some patients [274] (not unreasonably, many elderly glaucoma patients are unconcerned about eye color changes if the medications will help save their sight). Purely blue eyes are rarely affected, and fully darker brown eyes show no visibly detectable changes. It is the hazel and green eyes that may become noticeably browner. These color changes may occur with greater frequency than generally supposed, as Teus et al. have shown in a 2002 study comparing patients who received latanoprost in only one eye [275].

These color changes are more frequent with latanoprost than bimatoprost or travoprost (per the package inserts). Again, proactive conversation with the patient about this possibility before starting a hypotensive lipid will avoid unhappiness and potential medical–legal complications. Because of concerns about iris heterochromia (and possible lash growth described below), we almost never prescribe the use of HLs when glaucoma medication is needed only in one eye [275]. Figure 23.11 depicts a patient whose irises are different colors after using bimatoprost in only one eye.

Both hypertrichosis and periorbital skin hyperpigmentation are common side effects for patients who use the topical HLs [276–286]. The lashes may become thicker and longer and increase in quantity (Fig. 23.12). They may become darker although poliosis has been reported [287]. Many women find the lash effects desirable [288], and Allergan now has a commercial product (Latisse) for lash enhancement using bimatoprost 0.03 % as the active ingredient. For those patients who are bothered by the changed lashes, there is good news. Lashes naturally fall out and are replaced about every 12 weeks, so once the medication is stopped, the situation reverses itself [289]. (Recall that induced iris color changes are not reversible.) There is a report of hyperpigmentation and hypertrichosis of the vellus hair on the malar region after unilateral treatment with bimatoprost in a Hispanic patient. These phenomena reversed 2 months after the discontinuation of bimatoprost [290]. Hyperpigmentation of the lids and periorbital skin is more common in people with darker skin (Asian, Hispanics, and African-Americans) than very pale skin. The skin pigmentation changes appear to only affect the epidermis, which, similar to eye lashes, is periodically renewed [283]. Most patients find that undesirable skin pigmentation reverses with cessation of the drug. In our clinical experience, bimatoprost is more likely to affect the lashes and periorbital skin than the other two HLs. Some patients find that using petroleum jelly to coat the lash margins and skin below the lower lids, just before taking their HL eye drops, helps reduce these side effects. In addition to pigmentation, periorbital fat loss has now been reported with this class of agents [291, 292]. It remains a curiosity that it took so long to notice this important problem.

The zinc enzymes classified as carbonic anhydrases (CAs) are widespread through the bacteria and plant and animal kingdoms. There are at least 14 known varieties of the alpha-CAs (α-CA) whose main function is the hydration of CO₂ to bicarbonate (HCO^–). Two of these enzymes are important for the production of aqueous humor by the epithelium of the ciliary processes, cytoplasmic CA II and membrane-bound CA IV [293, 294]. Part of aqueous production involving active secretion relies on the formation of bicarbonate by these enzymes to correct the imbalance caused by the ATPase-fueled transport of sodium into the space between the nonpigmented ciliary epithelial cells. More than 99 % of the activity of these enzymes must be inhibited to successfully decrease aqueous production [295].

In 1954, Becker demonstrated that acetazolamide, a sulfonamide-based carbonic anhydrase inhibitor (CAI), could lower intraocular pressure in human subjects [296]. The IOP-lowering effect of CAIs may be enhanced by the metabolic acidosis they can cause [297]. Acetazolamide is more likely to cause clinically relevant systemic acidosis than methazolamide at therapeutic doses [298]. This may be one of the reasons that acetazolamide at full doses lowers IOP better than full doses of methazolamide. CA isoenzymes are present in many tissues in addition to ciliary processes, including corneal endothelium, iris, retinal pigment epithelium, red blood cells, brain, and kidney. Oral CAIs, while extremely effective in lowering IOP, may have profound medicinal and side effects on the body [295]. For this reason, chronic use of oral CAIs is generally reserved as therapy of last resort for patients who have tried topical medications (and perhaps laser treatments) without adequate control and who for one reason or another are not good candidates for glaucoma surgical intervention. Oral CAIs are also used on a temporary basis for emergent situations with acute rise of IOP. Examples of these situations include early postretinal and early postcataract surgeries and acute angle-closure glaucoma until laser therapy becomes available [299, 300].

Before initiating treatment with oral or topical CAIs, the physician must inquire about sulfonamide allergies. Many patients have been treated with sulfonamide-based antibiotics for bladder infections and may consider any gastrointestinal insult caused by these antibiotics as allergic symptoms. Patients should be specifically asked about breathing difficulties and skin reactions [301, 302], which are the most common form of allergic manifestations to sulfonamide antibiotics. Stevens–Johnson syndrome (erythema multiforme major) has been reported with the oral CAIs [303]. This may be more common in Japanese and Korean patients [304, 305]. Fortunately, there are significant structural differences between sulfonamide antibiotics and the CAIs used to treat glaucoma, but prudence suggests avoiding oral CAIs in any patients with true allergies to any sulfonamide medication [306]. On occasion, we have cautiously used topical CAIs with patients who have an ambiguous history of sulfonamide allergy symptoms. We explain the possible risks, have them use the drops while in the office, keep them around for several hours, and tell them that in the rare event of an allergic reaction, we will call for emergency medical services. In our practices, we have never had a patient who has had an acute event under these circumstances.

Oral CAIs are powerful agents for lowering IOP (between 25 and 30 %) [307] and may do so very well when other medical therapies are unable to reach the target IOP in chronic glaucomas or to temporarily bring IOP to safe levels in acute emergent situations. But their systemic side effects are legion and often serious, in addition to the allergic respiratory and skin problems discussed previously. The limited data available suggest that the topical CAIs (TCAIs), while theoretically capable of causing all the same systemic mischief as the oral agents, have mostly ocular and periorbital side effects.

The two common oral CAIs are acetazolamide (Diamox) and methazolamide (Neptazane).

Diamox is available as “regular” tablets (62.5, 125, and 250 mg) and in timed release, 500 mg “Sequels.” Diamox regular is generally prescribed QID as 250-mg tablets for chronic use or better as Diamox Sequels 500 mg Q 12 h. The Sequels produce less GI symptoms and less fluctuation in IOP over 24 h but are considerably more expensive than the standard tablets that are also generically available. Neptazane tablets are prescribed Q 12 h or TID in either 25- or 50-mg dosages. They do come in generic varieties as well.

In a situation that requires acute lowering of IOP, we often prescribe one regular 250-mg Diamox tablet, with relatively quicker onset of action, and at the same time, we have the patient take a Diamox Sequel for its sustained effect. Patients who may have nausea or vomiting from acutely high IOP, as in angle-closure situations, may not tolerate the oral preparations. Acetazolamide is excreted unchanged by the renal system. Methazolamide undergoes first-pass metabolism by the liver. CAs in the kidney promote the absorption of bicarbonate through the renal tubules. CAIs cause alkalinization of the urine along with increased micturition, both day and night, and potassium excretion. Patients prescribed chronic oral CAIs should have their electrolytes monitored, perhaps with the assistance of their family physician, especially if taking other potassium-wasting drugs such as thiazide diuretics and oral corticosteroids [308, 309]. This is especially important in patients taking cardiac glycosides, such as digitalis, because hypokalemia can be lethal in the presence of digitalis. Systemic side effects are generally dose related except for the blood dyscrasias, which will be discussed later [310]. Kidney stone occurrence may be up to 50 times greater in patients taking oral CAIs than prevalent in the general population [311]. Systemic CAIs cause the kidney to secrete less magnesium and citrate and an alkaline urine. This contributes to the formation of calcium phosphate- and oxalate-based stones.

Methazolamide at lower doses is less likely to cause urolithiasis because of less acidosis and no change in concentrations of urinary citrate [298, 312, 313]. If it is necessary to prescribe oral CAIs, methazolamide is preferred to acetazolamide in patients with a history of kidney stones.

Salicylates interact with oral CAIs. Patients taking high-dose aspirin can get tinnitus, increased respiratory rate, and even confusion and coma [314–317]. They should be warned of these symptoms and told to avoid high doses of aspirin if oral CAIs are prescribed. The low dose of 81 mg of aspirin often prescribed to prevent cardiac events by its antiplatelet effects is probably not of significance with respect to interacting with oral CAIs.

Paresthesias of the fingers, toes, and nose are common with oral CAIs, less so with methazolamide at lower doses [318]. Paresthesias may diminish with time [298]. Patients are less likely to be concerned about these symptoms if they are discussed before the drugs are prescribed.

Gastrointestinal symptoms are also common with oral CAIs, more so with QID regular Diamox than Diamox Sequels or methazolamide [319]. Patients may suffer from abdominal cramps, nausea, and in some cases severe diarrhea [320]. Symptoms may improve as time passes, but some patients must discontinue oral CAIs because of these difficulties. The oral CAIs cause a strange metallic taste with foods and carbonated beverages – patients should be warned this is likely to occur [321]. We have had more than one patient return a case of “bad” beer to the store after initiating acetazolamide therapy [322]. (We apologized to them for our lapse in not discussing this side effect of hypogeusia!) Also note that topical CAIs, when drained onto the tongue via the nasolacrimal duct, may produce “funny” taste sensations for a short time after contact. Lower doses of oral CAIs disrupt taste sensations to a lesser degree than do higher doses.

Patients taking oral CAIs, usually after several months, can have an unexpected onset of a malaise-syndrome complex involving – to varying degrees – tiredness, lack of appetite (which may be accompanied by weight loss), and even severe depression [320]. Generalized apathy may ensue. Patients with mild dementia may seem to get worse. Often, neither the patient nor family members attribute the slow onset of this malaise syndrome to oral CAIs that may have been started several months earlier. The symptoms resolve, slowly, with reduced drug dosage or more often with discontinuation of the offending drug. Similarly, decreased libido and even impotency can result from oral CAIs, again often not attributed to the drugs [323, 324]. Patients and physicians must be aware of these possibilities.

Blood dyscrasias from oral CAIs (none has been reported in the literature from topical CAIs as of mid-2009) are extremely serious events. They fall into two categories: one that is sudden and is irreversible, with a high mortality rate (about 50 % for patients with aplastic anemia), and another that is drug dose dependent and may be reversible with termination of the CAI. Fortunately, the incidence of aplastic anemia associated with oral CAIs is low. Keisu et al. in 1990 [325] estimated it is “approximately one in 18,000 patient years.” Patients present with a constellation of symptoms, including fever, easily bruised skin, purpura, sore throat, paleness, nosebleeds, petechiae, and perhaps jaundice [325, 326]. The use of blood tests to monitor for blood dyscrasias is controversial [327]. In an editorial in 2000 [328], Fraunfelder and Bagby weigh the pros and cons of doing routine, periodic blood tests on patients taking oral CAIs, hopefully to catch early any abnormalities and to be able to cease the drug and treat the reversible cases. Because of the serious nature of these dyscrasias and the fact that only about half of the patients have an irreversible, potentially fatal variation, they recommend that “a standard CBC, consisting of white blood count with differential, hemoglobin, hematocrit, and platelet count, should be repeated every 1–2 months during the first 6-month period and then at 6-month intervals.” In our experience, few ophthalmologists ever get diagnostic blood tests on patients beginning or continuing oral CAI therapy [329]. By custom we do not, but we are fortunate not to have had any patients so affected in our practices. Luckily, the odds against this happening seem to have been in our favor. Physicians may feel the inconvenience and cost of these tests outweigh the probability of finding and successfully treating affected patients. Patients should be aware of the untoward symptoms and report them on an emergent basis. Given all of the above problems with oral CAIs, they should generally be used only temporarily in acute situations of very high IOP or when all other means of treatment have failed and surgery is not a good option.