Penetration of natural prostaglandins and their ester prodrugs and analogs across human ocular tissues in vitro.

The objective of this study was to assess the corneal and scleral permeabilities of natural prostaglandins as well as their prodrugs and analogs through human cornea and sclera in vitro. The "apparent permeability coefficients" (Papp) of natural prostaglandins (PGF2alpha, PGD2 and PGE2), ester prodrugs of PGF2alpha (1-isopropyl PGF2alpha, 11-pivaloyl PGF2alpha and 11,15-dipivaloyl PGF2alpha) and four analogs (16-m-chlorophenoxy tetranor PGF2alpha, 17-phenyl trinor PGF2alpha, 17-phenyl trinor PGE2 and AH 13205) were measured using modified Ussing perfusion chambers and quantitative high performance liquid chromatography. Our results indicate that the corneal penetration of natural prostaglandins (PGs) is poor (the Papp values ranged from 1.65 x 10(-6) to 2.38 x 10(-6) cm/sec), while the PGF2alpha prodrugs showed higher corneal penetration than PGF2alpha. The 11-pivaloyl ester of PGF2alpha penetrated the cornea faster than both 1-isopropyl ester and the lipophilic 11,15-dipivaloyl ester. The PG analogs also showed poor corneal penetration (Papp values ranged from 0.696 x 10(-6) to 1.49 x 10(-6) cm/sec) except for AH 13205. All compounds tested showed good scleral penetration (Papp values ranged from 6.90 x 10(-6) to 17.1 x 10(-6) cm/sec) except PGF2alpha 11,15-dipivaloyl (Papp = 1.22 x 10(-6) cm/sec). The penetration profiles correlated well with tissue uptake ratios (ratio of final tissue concentration to initial dose) for all compounds except 11,15-dipivalate PGF2alpha. All ester prodrugs (but not the PGs and analogs) underwent corneal first-pass metabolism. The study results demonstrate that transcleral absorption may play a significant role in the ocular absorption of these compounds.

Ocular penetration and bioconversion of prostaglandin F2alpha prodrugs in rabbit cornea and conjunctiva.

The objective of this study was to identify prostaglandin F2alpha (PGF2alpha) prodrugs that have an optimal ocular absorption profile and therefore could be potentially useful for the treatment of glaucoma. Rabbit cornea, conjunctiva, and iris/ciliary body were mounted in a flow-through chamber to evaluate the permeability and bioconversion of PGF2alpha and its prodrugs. The prodrugs tested were PGF2alpha 1-isopropyl, 1,11-lactone, 15-acetyl, 15-pivaloyl, 15-valeryl, and 11,15-dipivaloyl esters. After 4 h in the donor or acceptor compartments, the products and formation of PGF2alpha were analyzed by HPLC. Effects on intraocular pressure and ocular surface hyperemia were also determined. All prodrugs penetrated the rabbit cornea faster than PGF2alpha by 4- to 83-fold. All prodrugs penetrated conjunctiva faster than PGF2alpha, except the 15-acetyl ester prodrug, which was equally permeable. No direct correlation between drug lipophilicity and permeability across the cornea or conjunctiva was apparent. The most metabolically stable prodrug was the 1,11-lactone, followed by the 11,15-dipivaloyl, 15-pivaloyl, 15-acetyl, 1-isopropyl, and the 15-valeryl esters, the latter of which was extensively converted to PGF2alpha. A separation index for various prodrugs was calculated from the ratio of the bioavailable PGF2alpha for ocular hypotension to the bioavailable PGF2alpha for hyperemia. The highest separation index was observed for the 1,11-lactone prodrug (2.33), followed by the 11,15-dipivaloyl ester prodrug (1.80). Thus the 1,11-lactone and 11,15-dipivaloyl ester prodrugs appeared to be superior to the others in providing bioavailable PGF2alpha for ocular hypotension, while minimizing hyperemia. The favorable separation index for these compounds appeared to be due to their metabolic stability at the corneal surface and conjunctiva combined with sufficient bioavailability for ocular hypotension.

Characterization of brimonidine metabolism with rat, rabbit, dog, monkey and human liver fractions and rabbit liver aldehyde oxidase.

1. In vitro metabolism of 14C-brimonidine by the rat, rabbit, dog, monkey and human liver fractions was studied to assess any species differences. In vitro metabolism with rabbit liver aldehyde oxidase and human liver slices, and in vivo metabolism in rats were also investigated. The hepatic and urinary metabolites were characterized by liquid chromatography and mass spectrometry. 2. Up to seven, six, 11 and 14 metabolites were detected in rat liver S9 fraction, human liver S9 fraction, human liver slices and rat urine respectively. Rabbit liver aldehyde oxidase catalysed the metabolism of brimonidine to 2-oxobrimonidine and 3-oxobrimonidine, and further oxidation to the 2,3-dioxobrimonidine. Menadione inhibited the liver aldehyde oxidase-mediated oxidation. 3. Hepatic oxidation of brimonidine to 2-oxobrimonidine, 3-oxobrimonidine and 2,3-dioxobrimonidine was a major pathway in all the species studied, except the dog whose prominent metabolites were 4',5'-dehydrobrimonidine and 5-bromo-6-guanidinoquinoxaline. 4. These results indicate extensive hepatic metabolism of brimonidine and provide evidence for aldehyde oxidase involvement in brimonidine metabolism. The species differences in hepatic brimonidine metabolism are likely related to the low activity of dog liver aldehyde oxidase. The principal metabolic pathways of brimonidine are alpha(N)-oxidation to the 2,3-dioxobrimonidine, and oxidative cleavage of the imidazoline ring to 5-bromo-6-guanidinoquinoxaline.

Conjunctival penetration of insulin and peptide drugs in the albino rabbit.

An in vitro model was used to evaluate the conjunctival penetration of three peptides, [D-ala2]metenkephalinamide (YAGFM, MW 647), substance P (MW 1348), and insulin (MW 5778), in comparison with two nonpeptides, atenolol (MW 266) and timolol (MW 433). All three peptides were hydrolyzed to varying extents during penetration across the conjunctiva. The permeability coefficient for intact YAGFM and insulin was 4.5 +/- 0.3 and 4.6 +/- 0.7 microns sec-1, respectively. These values were about two to five times lower than those for atenolol and timolol. No permeability coefficient could be calculated for substance P, since its transconjunctival flux never reached steady state. The conjunctival penetration of YAGFM and insulin was improved by about two and three times, respectively, with the addition of 1% Na glycocholate. Increasing the Na glycocholate concentration was more effective than changing the type of bile salt in improving the conjunctival penetration of insulin. The maximum factor of improvement was 12, as the Na glycocholate concentration was raised to 4%. The way in which Na deoxycholate, glycocholate, and taurocholate affected the conjunctival penetration of atenolol, timolol, and insulin suggests that these three bile salts improved mainly the transcellular penetration of the compounds studied.

Morphologic and physiologic effects of artificial tear formulations on corneal epithelial derived cells.

The biological effects of four commercially available artificial tear formulations were evaluated using sensitive in vitro techniques. Two of the formulations contained ingredients implicated in cell damage; the other two products were not chemically preserved, and their components have not been reported to damage corneal tissue. We assayed the effects of these formulations on viability, morphology, and physiology in corneal cell (SIRC) cultures. Their effect on the hydration of excised rabbit corneas was also determined. In all formulations, cell viability declined with time relative to control cells, but the time course varied significantly. Viability remained at 100% for 6 h in an unpreserved carboxymethylcellulose-based product (CMC-U), and decreased to 50% after greater than 16 hours. Viability decreased to 50% in 3 h for the other unpreserved, polyvinyl alcohol-based product (PVA-U), and in 1 h for a hydroxypropylmethylcellulose formulation (HPMC-P) that contains edetate disodium (EDTA). Cells in a preserved formulation (PVA-P), using polyvinyl alcohol as the polymer and containing EDTA and benzalkonium chloride (BAK), failed to survive even 15 min of treatment. Overall, cells treated with the unpreserved products were nearly indistinguishable from those in the control solution with respect to morphology, electrophysiology, and corneal hydration. Also, the relative ranking from least to most deleterious (control less than CMC-U less than PVA-U less than HPMC-P less than PVA-P) was consistent across several measures. Of the preserved formulations, HPMC-P, which contains the chelating agent EDTA as an additive, was less damaging than was PVA-P, which contains two chemicals, EDTA and BAK, that reportedly damage cells.

Systemic pharmacokinetics of acitretin, etretinate, isotretinoin, and acetylenic retinoids in guinea pigs and obese rats.

Etretinate, a highly lipophilic retinoid, is known to accumulate in the human body with a slow systemic elimination (half-life approximately 100 days) after long-term treatment. Retinoids with high lipophilicity and slow body elimination have the propensity of eliciting teratogenic effects. Therefore, synthetic retinoids with reduced systemic retention are desired. In this study, we evaluated the systemic pharmacokinetics of acitretin, etretinate, isotretinoin, synthetic acetylenic retinoic acids (AGN 190121, AGN 190186, and AGN 190299), and acetylenic retinoates (AGN 190073, AGN 190089, and AGN 190168) in guinea pigs following iv doses. Their pharmacokinetics were also measured in obese rats to probe the effect of body fat on the drug disposition of retinoids. The acetylenic retinoates were hydrolyzed to their corresponding free acids at a much faster rate than etretinate in both animal species. All retinoates showed faster body clearance and larger volume of distribution than their free acids. In the obese rats, longer elimination half-lives and slower body clearance of the retinoids, except isotretinoin, were observed as compared to those in the normal rats. These results suggest that body fat has a significant effect on drug disposition and slows down the systemic clearance of retinoids. Since the synthetic acetylenic retinoates rapidly converted to their less lipophilic free acids after systemic absorption, the potential accumulation of these retinoids, as reported for lipophilic etretinate, were unlikely to occur in humans and animals.

Lipophilicity influence on conjunctival drug penetration in the pigmented rabbit: a comparison with corneal penetration.

The influence of lipophilicity on the conjunctival penetration of beta blockers in the pigmented rabbit was investigated and compared with that on corneal penetration. The beta blockers were hydrophilic sotalol, atenolol, nadolol, pindolol, and acebutolol; lipophilic metoprolol, timolol, oxprenolol, levobunolol, labetalol, and alprenolol; and the very lipophilic propranolol and betaxolol. Drug penetration was evaluated by using the isolated pigmented rabbit conjunctiva and cornea in the modified Ussing chamber and was monitored by reversed phase HPLC. The conjunctiva was more permeable to all the beta blockers than was the cornea. A sigmoidal relationship, rather than the familiar parabolic relationship, best described the influence of lipophilicity on both conjunctival and corneal drug penetration. The ratio of corneal to conjunctival permeability coefficients was most sensitive to changes in log PC within the region of 1.5 and 2.5. Outside of this region, the ratio was relatively independent of changes in lipophilicity. For several beta blockers, their intrinsic sympathomimetic activity may play a minor role in influencing their conjunctival and corneal penetration.

Role of enzymatic lability in the corneal and conjunctival penetration of timolol ester prodrugs in the pigmented rabbit.

The main objective of this study was to investigate how enzymatic lability would affect the extent of corneal and conjunctival penetration of a series of alkyl, cycloalkyl, and aryl ester prodrugs of timolol in the pigmented rabbit. Enzymatic lability of the prodrugs was studied in corneal epithelial and conjunctival homogenates, while their corneal and conjunctival penetration was determined using the isolated tissues in the modified Ussing chamber. The straight-chain alkyl and the unsubstituted cycloalkyl esters were hydrolyzed more rapidly than their corresponding branched-chain and substituted analogues as well as the aryl esters. The corneal and conjunctival penetration of all prodrugs, regardless of enzymatic lability, varied parabolically with lipophilicity. Moreover, the enzymatically more labile straight-chain alkyl esters penetrated the cornea and the conjunctiva more readily than the more stable branched-chain esters of comparable lipophilicity. Enzymatic lability is, therefore, an additional factor that should be considered in designing alkyl ester prodrugs with improved ocular drug delivery characteristics. Enzymatic lability does not, however, play as important a role as lipophilicity in the corneal and conjunctival penetration of cycloalkyl and aryl ester prodrugs.

Detection and identification of endogenous small molecules in ocular tissues by proton nuclear magnetic resonance spectroscopy.

Proton nuclear magnetic resonance (nmr) spectra of rabbit ocular tissue homogenates (corneal epithelium, conjunctiva and iris ciliary body) and aqueous humor have been recorded for the first time by incorporation of a spin-spin relaxation reagent and use of the CPMG pulse sequence. A number of endogenous species such as lactate, glucose, alanine and valine have been observed in these ocular homogenates and significant differences have been noted in the distribution of these small molecules in the ocular tissues studied. This technique has been used to study the hydrolysis of acetylcholine by ocular esterases in the iris ciliary body.

Corneal and conjunctival/scleral penetration of p-aminoclonidine, AGN 190342, and clonidine in rabbit eyes.

The ocular penetration pathways of three alpha 2-adrenergic agents (p-aminoclonidine, AGN 190342, and clonidine) were investigated in rabbits both in vitro and in vivo. The corneal permeabilities of the compounds correlated positively with their octanol/water distribution coefficients. The ocular drug absorption via corneal and conjunctival/scleral penetration routes was evaluated separately after drug perfusion in vivo. In most cases, the corneal route was the major pathway for the intraocular drug absorption. However, the conjunctival/scleral penetration pathway was the predominant pathway for the delivery of p-aminoclonidine, the least lipophilic compound among the three drugs, to the ciliary body. The drug concentration in the iris was contributed mainly by the corneal route and correlated well with drug lipophilicity.

Ocular pharmacokinetics and pharmacodynamics of phenylephrine and phenylephrine oxazolidine in rabbit eyes.

The aqueous humor concentration of phenylephrine and its corresponding mydriatic response were measured over time in New Zealand albino rabbit eyes following a 10-microliters topical instillation of a phenylephrine HCl viscous solution (10%) or a phenylephrine oxazolidine (prodrug) suspension in sesame oil (1 and 10%). The bioavailability of a 1% prodrug suspension in the rabbit eye (AUC of aqueous humor concentration vs time) was 30% lower than that of a 10% phenylephrine solution (P less than 0.1) with the exception that the peak time occurred 34 min earlier with the prodrug. A 10% prodrug suspension increased the aqueous humor bioavailability approximately eightfold but improved the mydriatic activity (AUC of mydriasis vs time) only fourfold. The pharmacokinetic parameters, apparent absorption, and elimination rate constants, of phenylephrine and the prodrug were determined from aqueous humor concentration-time and mydriasis-time profiles. The study showed that the kinetic parameters of phenylephrine estimated from its mydriasis profile do not accurately reflect the kinetics of drug distribution in the iris. These parameters also varied with the instillation of phenylephrine solution or prodrug suspensions. A mydriatic tolerance of the pupil response was apparent after the topical instillation of phenylephrine solution. The mydriatic tolerance may be due to the decrease in receptor number in the iris dilator muscle.

Ocular absorption and disposition of phenylephrine and phenylephrine oxazolidine.

The ocular bioavailability of phenylephrine oxazolidine (PO), a prodrug intended for rapid corneal penetration, was micronized and suspended in sesame oil (1 and 10 per cent) and compared in bioavailability to phenylephrine HC1 (PE) dissolved (10 per cent) in a buffered (pH 5.75), viscous (30 centipoise) vehicle. Cornea and aqueous humor of New Zealand rabbits were measured over time following 10 microliter instillation to the eye. Based upon AUC measurements, corneal and aqueous humor levels were approximately 6 and 8 times greater for 10 per cent PO versus 10 per cent PE, respectively. In addition, the ocular pharmacokinetic values were determined for PE applied in a constant concentration (1 per cent) to the cornea over 180 min to anesthetized rabbits. Cornea and aqueous humor were measured for drug content over time. Using moment analysis and an initial slope method, the absorption rate constant, ka, the steady state volume of distribution in the eye, Vss, and ocular clearance, Qe, were calculated. Values obtained for PE were 4.15 x 10(-5) min-1, 0.423 ml and 14.6 microliter min-1, respectively. The half-life for drug elimination ranged from 63-83 min depending on the tissue or route of administration.

Ocular ketone reductase distribution and its role in the metabolism of ocularly applied levobunolol in the pigmented rabbit.

The distribution of ketone reductase activity in the anterior segment tissues of the pigmented rabbit eye and its influence on the ocular metabolism of topically applied levobunolol were studied. A reversed phase high-performance liquid chromatography procedure was used to assay for this drug and its metabolite, dihydrolevobunolol. Ocular ketone reductase activity was 3 to 4 times more dependent on NADPH than on NADH. The rank order of activity was corneal epithelium greater than iris-ciliary body greater than conjunctiva greater than lens. No activity was detected in the tears, corneal stroma, sclera or aqueous humor. Ketone reductase activity was entirely cytosolic. The pigmented rabbit was significantly less active than the albino rabbit in ketone reductase. Its activity in the corneal epithelium, iris-ciliary body and lens was most sensitive to inhibition by quercetin, whereas that in the conjunctiva was most sensitive to metyrapone. The ketone reductase in the corneal epithelium contributed more to the metabolism of topically applied levobunolol than its counterpart in the iris-ciliary body and lens. Moreover, the extent of levobunolol metabolism both during and after corneal penetration was dose-dependent. Overall, these findings indicate that ocularly applied drugs containing the ketone functional group are subject to varying degrees of metabolism by NADPH-dependent ketone reductases in the corneal epithelium, iris-ciliary body and lens.

Influence of corneal epithelial integrity on the penetration of timolol prodrugs.

The interplay of corneal epithelial integrity and prodrug lipophilicity in the corneal penetration of a homologous series of timolol prodrugs was studied. A reversed phase HPLC procedure was used to assay for timolol and its prodrugs. The optimal distribution coefficient for corneal penetration on a logarithmic basis was found to be 2.5, and for two prodrugs of equal lipophilicity the enzymatically labile one was better absorbed. The corneal permeability coefficient of O-acetyl, propionyl, and butyryl timolol was unaffected by deepithelizing the cornea, whereas that of the other prodrugs was reduced. This suggests that these three prodrugs diffused across the corneal epithelium proper as readily as across its basement membrane, which was not removed by the deepithelizing procedure. In contrast to the in vitro results, deepithelizing the cornea did not affect the aqueous humor concentration of timolol at 20 min afforded by any of the prodrugs except the O-propionyl and octanoyl esters. This suggests that factors other than corneal penetration exerted a more important role in affecting the aqueous humor drug concentration when the cornea was deepithelized.

Relative effectiveness of prodrug and viscous solution approaches in maximizing the ratio of ocular to systemic absorption of topically applied timolol.

The effectiveness of the prodrug approach in maximizing the ratio of ocular to systemic absorption of topically applied timolol was compared with that based on viscous solutions. The pigmented rabbit was the experimental animal. o-Butyryl timolol, a lipophilic ester prodrug of timolol, was found to be twice as effective as hyaluronic acid and poly(vinyl alcohol) solutions in enhancing the ocular absorption of timolol. In addition, it was slightly more effective than retaining the instilled dose in the conjunctival sac for 240 min via nasolacrimal blockade. On the other hand, due to its inability to minimize contact of the instilled dose with the nasal mucosa, o-butyryl timolol was less capable than the viscous solutions in reducing systemic drug absorption. But this could be corrected by reducing the instilled solution volume from 25- to 5 microliter, resulting in a nine-fold reduction in plasma timolol levels while still providing the same amount of drug in the aqueous humor as a five times higher dose of timolol. The net result was a 15-fold improvement in the ratio of ocular to systemic drug absorption, which was also achieved upon retaining the instilled dose in the conjunctival sac for 240 min. It thus appears that the approach based on improved corneal absorption would be as effective as that based on improved retention in the conjunctival sac in maximizing the ratio of ocular to systemic drug absorption, so long as a smaller instilled dose volume of the prodrug is used.

Ophthalmic bioequivalence of steroid/antibiotic combination formulations.

This study compared the relative ocular bioavailability in rabbits of the antibiotic/steroid combination of 0.3 per cent tobramycin and 0.1 per cent fluorometholone acetate to each of the two single-entity products. Two separate studies were conducted, one measuring fluorometholone acetate in cornea and aqueous humour through 240 min and the other measuring tobramycin in the cornea through 120 min. The results for fluorometholone acetate show that the combination product is 15 per cent higher in AUC for the cornea than the single-entity product (0.05 less than p less than 0.1). However, the combination product is only 4.4 per cent higher in AUC when aqueous humour levels are compared (N.S.). The single-entity fluorometholone acetate ointment yielded cornea and aqueous humour levels which were 20 per cent (p less than 0.01) and 6 per cent (N.S.) higher in AUC than the same tissues measured after topical instillation of the suspension. Statistical treatment of the data indicated that the suspension and ointment products, were bioequivalent with respect to fluorometholone acetate in aqueous humour, but not with respect to corneal concentrations. Tobramycin was measured in the cornea following instillation of either the single-entity solution or the combination suspension. Per cent bound (tobramycin to rabbit antibodies determined in vitro from the radioimmunoassay results) was used for statistical analyses of all group means and standard errors. The results indicated that the products were bioinequivalent.

Reduction of phenylephrine drop size in infants achieves equal dilation with decreased systemic absorption.

We studied the effect of reducing eye drop size on the efficacy and systemic absorption of topical 2.5% phenylephrine hydrochloride in neonates and infants. Eleven neonates received an 8-microL drop volume in one eye and a 30-microL drop volume (commercial size) in the fellow eye. Mean pupillary dilation at 60 minutes was equivalent (4.86 mm vs 4.57 mm) for both eyes, respectively. The plasma phenylephrine level was determined for the two drop sizes in a second group of infants. Eight infants received an 8-microL drop volume in both eyes, while nine infants received a 30-microL drop volume in both eyes. The mean phenylephrine level at ten minutes was 0.9 ng/mL for the 8-microL drop group and 1.9 ng/mL for the 30 microL drop group. In neonates and infants, reducing the drop volume of topical phenylephrine may improve the risk-benefit ratio.

Improving the therapeutic index of topical phenylephrine by reducing drop volume.

This study examined the effect of reducing eyedrop volume on the efficacy and systemic absorption of a given quantity of phenylephrine hydrochloride. Aqueous phenylephrine hydrochloride (2.5%) given in the commercially available drop volume of 32 microliters was compared with 10% aqueous phenylephrine given in an 8-microliters volume. Both preparations contained the same total amount of phenylephrine per drop. Ten subjects had both eyes dilated with each of the phenylephrine solutions according to a randomized, double-crossover sequence. The mean final pupillary diameter was nearly 1.0 mm larger for the 8-microliters drop (P = 0.0033). Nine of ten subjects achieved a larger pupillary diameter with the 8-microliters drop. Systemic absorption, as measured by plasma phenylephrine level, was similar for the two drop volumes. Thus, the 8-microliters drop achieved a significantly larger pupillary dilation without an increase in systemic absorption. Phenylephrine may have the most favorable risk-benefit ratio when administered as a high concentration in a small volume.

Improving the ocular absorption of phenylephrine.

An oxazolidine prodrug of phenylephrine and the base form of phenylephrine were synthesized, suspended in sesame oil, and tested for mydriatic activity against phenylephrine HCl. The HCl salt was formulated as a viscous aqueous solution and as a sesame oil suspension. A dosing volume of 10 microliter was instilled into rabbit eyes and the pupillary diameter was measured over time. A 0.045 M prodrug suspension was judged equal in mydriatic activity to a 0.45 M viscous solution of phenylephrine HCl with the exception that the time of maximum response occurred 60 min earlier with the prodrug. When phenylephrine base was suspended in sesame oil at 0.045, 0.12, and 0.45 M, the mydriatic activity was also greater than equimolar suspensions of phenylephrine HCl. The pH of tear fluids was also measured over time and found to rise 1.1, 0.70, and 0.30 pH units for 0.45, 0.12, and 0.045 M suspensions of the base form but remain unchanged when phenylephrine HCl was instilled in the rabbit eye. The greater activity associated with the base form of phenylephrine was judged a result of the change in pH to favour the absorption of phenylephrine. This latter approach should be applicable to either weak acids or weak bases with pKa values outside of the normal pH range (7-8) of the tears and in concentrations greater than 0.045 M suspended in a non-aqueous vehicle.

Aqueous vs viscous phenylephrine. I. Systemic absorption and cardiovascular effects.

We studied 30 patients undergoing vitreoretinal surgery to compare the systemic absorption and cardiovascular effects of 2.5% aqueous and 2.5% viscous (21 cp) ophthalmic solutions of phenylephrine hydrochloride. No significant differences were noted in the plasma levels or changes in blood pressure between the two groups, although there was a tendency toward higher mean plasma levels and blood pressures in groups receiving 2.5% aqueous phenylephrine hydrochloride. Maximum plasma levels were achieved within the first 20 minutes following topical application of phenylephrine eye drops, irrespective of the nature of the vehicle. This finding emphasizes the importance of monitoring these patients, especially those at high risk for any adverse cardiovascular effects during the first 20 to 30 minutes following instillation of phenylephrine eye drops. The patients in our study were supine and under general anesthesia. Therefore, there was no effect by lid blinking, the lacrimal pump, or gravity, which would ordinarily increase absorption by the nasal mucosa through the nasolacrimal system. By eliminating these variable factors, such as lid blinking, the study was performed in a stable and controlled manner, but the results may not be directly applicable to an upright awake patient.