FormalPara Key Summary Points

Why carry out this study?

Individual differences in pharmacokinetics affect patient responsiveness to drugs.

The factors related to the ocular penetration of drugs after the administration of eye drops in humans have not been examined in detail.

We investigated the influence of patient factors on the intraocular penetration of eye drops via post hoc pooled analysis of clinical studies that determined drug concentrations in the aqueous and vitreous humor after the topical instillation of brimonidine-related eye drops.

What was learned from the study?

A positive correlation between age and the concentrations of brimonidine and brinzolamide in the aqueous humor was found and that between age and timolol concentration in the aqueous humor showed a trend in the same direction.

The findings of this study emphasize the necessity of considering individual differences in ocular pharmacokinetics for drug therapy (formulation design of the eye drops and dose regimen).

Introduction

Eye drops are the most commonly used drug formulation in ophthalmology for the treatment of many eye conditions, including glaucoma, conjunctivitis, ocular inflammation and dry eye disease [1, 2].

Significant individual differences, such as the degree of the effect of the drug and the occurrence of side effects, may be observed in the patient’s response to the drug. This inter-individual variability occurs because of genetic and environmental factors affecting pharmacodynamics and pharmacokinetics. In particular, the causes of variable pharmacokinetics have been investigated extensively [3].

After topical administration, the drug is absorbed through the cornea, conjunctiva and sclera and penetrates the ocular tissues. The posterior segment of the eye may absorb drugs from topically applied eyedrops by topical absorption into the eye or systemic absorption through the bloodstream [4,5,6].

In various animal experiments, biological factors such as tear volume, histological factors such as corneal thickness and physiological factors such as blink frequency and the retina-blood barrier can affect the ocular pharmacokinetics of eye drops [7]. These factors may also have a clinical impact.

However, it is difficult to sample human eye tissue from a living human being, and only a limited number of studies have reported drug concentrations in ocular tissue after instillation in humans. Thus, the factors that affect ocular penetration of drugs after administration of eye drops in humans have not been examined in detail.

In this study, we investigated the influence of patient-related factors on the intraocular penetration of eye drops by post hoc pooled analysis of clinical studies that determined drug concentrations in the aqueous and vitreous humor after the topical instillation of eye drops containing brimonidine.

Methods

This post hoc pooled analysis integrated and reanalyzed data sourced from three clinical studies that determined brimonidine concentrations in the aqueous and vitreous humor after the topical application of the following eye drops containing brimonidine: 0.1% brimonidine tartrate ophthalmic solution (Aiphagan®; Senju Pharmaceutical Co., Ltd., Osaka, Japan), 0.1% brimonidine tartrate and 0.5% timolol (equivalent to 0.68% timolol maleate) fixed-combination ophthalmic solution (Aibeta®; Senju Pharmaceutical Co., Ltd.) and 0.1% brimonidine tartrate and 1% brinzolamide fixed-combination suspension (Ailamide®; Senju Pharmaceutical Co., Ltd.) [8,9,10] (Tables 1 and 2).

Table 1 Patient background
Full size table
Table 2 Ocular pharmacokinetics and patient factors according to the brimonidine-related studied
Full size table

The Aiphagan study was approved by the Institutional Review Board of the University of Fukui, while the Aibeta and Ailamide studies were approved by the Certified Review Board of the University of Fukui. These previous studies complied with the tenets of the Declaration of Helsinki. Since the current study was a post hoc pooled analysis of data from previous studies, online registration as a clinical trial in Japan was not required.

In these previous studies, each brimonidine-related eye drop was topically administered twice daily for 1 week to patients scheduled for pars plana vitrectomy to treat idiopathic epiretinal membrane or macular hole until the day before surgery. On the day of the surgery, the eye drops were administered in the morning and 2 h before the surgery. The aqueous and vitreous humor was collected before vitrectomy, and the brimonidine concentrations were measured using liquid chromatography-tandem spectrometry (LC/MS/MS). Furthermore, the timolol and brinzolamide concentrations were also measured in the aqueous and vitreous humor samples obtained from patients that had been treated with Aibeta and Ailamide.

Sex, the presence or absence of lens (phakia/pseudophakia), age, corneal thickness, corneal endothelial cell density, tear secretion, eye axial length, height, weight and body mass index (BMI) were examined as patient background factors. Data on the corneal thickness, corneal endothelial cell density, tear secretion or eye axial length were not collected in the Aiphagan study. Data regarding sex, the presence or absence of the lens and age were obtained by interviewing the patients. The corneal thickness and eye axial length were measured using ophthalmic ultrasound imaging equipment (IOLMaster 700, Carl Zeiss Meditec AG, Jena, Germany). The corneal endothelial cell density was measured using a specular microscope (Konan Specular Microscope XI (FA-3709P), Konan Medical, Inc., Hyogo, Japan). Tear secretion was measured using Schirmer’s test without anesthesia. Weight and height were measured using a digital weighing scale. BMI was calculated by dividing the weight in kilograms by the square of the height in meters.

Statistical analyses were performed using the JMP 15 software (SAS Institute, Inc., Cary, NC, USA). The Wilcoxon rank-sum test was used to analyze the statistical differences between the sexes and between the presence or absence of lens in the concentrations of brimonidine, timolol and brinzolamide in the aqueous and vitreous humor. Correlations between patient-related factors (age, corneal thickness, corneal endothelial cell density, tear secretion, eye axial length, height, weight and BMI) and the concentrations of brimonidine, timolol or brinzolamide in the aqueous and vitreous humor were estimated using the Spearman’s rank correlation coefficient method. For all statistical tests, statistical significance was set at p < 0.05.

Results

Patient Characteristics

Data were collected from 42 participants, comprising 15 men and 27 women (Table 1). Twenty-four, eight and ten participants received Aiphagan, Aibeta and Ailamide, respectively. The number of men and women in the Aiphagan, Aibeta and Ailamide studies was 5 and 19, 5 and 3, and 5 and 5, respectively. The median age of all participants was 67.0 years. There was no difference in the median age among the studies: Aiphagan, 67.0 years; Aibeta, 67.5 years; Ailamide, 69.5 years. The median concentration (interquartile range [IQR]) of drugs in the aqueous humor was 294 (160–542) nM of brimonidine, 3220 (1870–4620) nM of timolol and 840 (521–1710) nM of brinzolamide (brimonidine, n = 40; timolol, n = 8; brinzolamide, n = 10; Table 2). Furthermore, the median concentration (IQR) of drugs in the vitreous humor was 3.69 (2.70–6.11) nM of brimonidine, 48.6 (21.8–108) nM of timolol and 8.23 (4.94–10.8) nM of brinzolamide (brimonidine, n = 42; timolol, n = 8; brinzolamide, n = 10). There was no difference in the median concentration (IQR) of brimonidine in the aqueous humor among the studies: Aiphagan, 229 (120–579) nM; Aibeta, 294 (181–515) nM; Ailamide, 432 (172–521) nM. There was no difference in the median concentration (IQR) of brimonidine in the vitreous humor among the studies: Aiphagan, 3.48 (2.81–5.94) nM; Aibeta, 3.77 (1.47–8.75) nM; Ailamide, 5.15 (3.19–6.00) nM. Because of the similarity of the data among the three studies, it was determined that the data were appropriate to be used in the present analysis.

Sex Differences

The median concentration (IQR) of brimonidine in the aqueous humor was 425 (167–555) nM in men and 293 (149–517) nM in women. The median concentration (IQR) of timolol in the aqueous humor was 3860 (1520–5100) nM in men and 2830 (1710–3610) nM in women. The median concentration (IQR) of brinzolamide in the aqueous humor was 1700 (726–2350) nM in men and 693 (370–840) nM in women. No correlation was found between sex and the concentration of each drug in the aqueous humor (brimonidine, p = 0.49; timolol, p = 0.55; brinzolamide, p = 0.075; Fig. 1).

Fig. 1
figure 1

Correlation between sex and brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor. a Brimonidine; n = 15 and 25, timolol; n = 5 and 3, brinzolamide; n = 5 and 5 in men and women, respectively. Brimonidine, p = 0.49; timolol, p = 0.55; brinzolamide, p = 0.075. b Brimonidine; n = 15 and 27, timolol; n = 5 and 3, brinzolamide; n = 5 and 5 in men and women, respectively. Brimonidine, p = 0.61; timolol, p = 0.23; brinzolamide, p = 0.53

Full size image

The median concentration (IQR) of brimonidine in the vitreous humor was 3.72 (1.61–5.75) nM in men and 3.66 (2.73–6.66) nM in women. The median concentration (IQR) of timolol in the vitreous humor was 37.7 (17.1–69.5) nM in men and 117 (23.4–174) nM in women. The median concentration (IQR) of brinzolamide in the vitreous humor was 8.96 (5.60–14.4) nM in men and 7.81 (4.88–11.6) nM in women. No correlation was found between sex and the concentration of each drug in the vitreous humor (brimonidine, p = 0.61; timolol, p = 0.23; brinzolamide, p = 0.53).

Effects of Presence or Absence of Lens

The median concentration (IQR) of brimonidine in the aqueous humor was 295 (160–538) nM in the phakic eyes and 267 (112–555) nM in the pseudophakic eyes. The median concentration (IQR) of brimonidine in the vitreous humor was 3.66 (2.60–6.07) nM in the phakic eyes and 3.77 (3.05–8.14) nM in the pseudophakic eye. Brimonidine concentrations in the aqueous and vitreous humor were not significantly different between the phakia and pseudophakia groups (aqueous humor, p = 0.89; vitreous humor, p = 0.88; Fig. 2). Evaluation of timolol and brinzolamide concentrations was not possible as there was only one case of pseudophakia.

Fig. 2
figure 2

Correlation between the presence or absence of lens and brimonidine concentrations in the aqueous (a) and vitreous (b) humor. a n = 33 and 7 in the phakic and pseudophakic eyes, p = 0.89. b n = 35 and 7 in the phakic and pseudophakic eyes, p = 0.88

Full size image

Effects of Age

A weak positive correlation was observed between age and brimonidine concentration in the aqueous humor (r = 0.3948, p = 0.012; Fig. 3), and the 95% confidence interval (CI) of the coefficient of correlation ranged between 0.0918 and 0.6270. A positive correlation was observed between age and brinzolamide concentration in the aqueous humor (r = 0.6809, p = 0.030), and the 95% CI of coefficient of correlation ranged between − 0.0805 and 0.8855. There was a trend towards a significant correlation between age and timolol concentration in the aqueous humor (r = 0.6425, p = 0.086), and the 95% CI of the coefficient of the correlation ranged between − 0.0708 and 0.9331. No correlation was found between age and the concentration of each drug in the vitreous humor.

Fig. 3
figure 3

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by age. A positive correlation between age and the concentrations of brimonidine and brinzolamide in the aqueous humor was found, and that between age and timolol concentration in the aqueous humor showed a trend in the same direction. a Brimonidine; n = 40, r = 0.3948 (95% CI 0.0918–0.6270), p = 0.012, timolol; n = 8, r = 0.6425 (95% CI − 0.0708 to 0.9331), p = 0.086, brinzolamide; n = 10, r = 0.6809 (95% CI − 0.0805 to 0.8855), p = 0.030. b Brimonidine; n = 42, r = 0.1066, p = 0.50, timolol; n = 8, r = 0.5092, p = 0.20, brinzolamide; n = 10, r = 0.0303, p = 0.93. CI confidence interval

Full size image

Effects of Weight

A negative correlation was observed between weight and the concentration of timolol in the vitreous humor (r =  − 0.8333, p = 0.010; Fig. 4), and the 95% CI of the coefficient of correlation ranged between − 0.9606 and − 0.1975. In contrast, no significant correlation was found between weight and timolol concentration in the aqueous humor or between weight and the concentration of brimonidine and brinzolamide in the aqueous and vitreous humor.

Fig. 4
figure 4

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by weight. A negative correlation was observed between weight and the vitreous concentration of timolol. a Brimonidine; n = 40, r =  − 0.1440, p = 0.38, timolol; n = 8, r =  − 0.3571, p = 0.39, brinzolamide; n = 10, r =  − 0.0182, p = 0.96. b Brimonidine; n = 42, r =  − 0.1550, p = 0.33, timolol; n = 8, r =  − 0.8333 (95% confidence interval: − 0.9606 to − 0.1975), p = 0.010, brinzolamide; n = 10, r = 0.2727, p = 0.45

Full size image

Effects of Other Factors

No correlation was found among brimonidine, timolol or brinzolamide concentrations in the aqueous and vitreous humor and corneal thickness, corneal endothelial cell density, tear secretion (Schirmer’s test), eye axial length, height or BMI (Figs. 5, 6, 7, 8, 9, 10).

Fig. 5
figure 5

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by corneal thickness. a Brimonidine; n = 18, r =  − 0.1901, p = 0.45, timolol; n = 8, r =  − 0.5476, p = 0.16, brinzolamide; n = 10, r = 0.0000, p = 1.0. b Brimonidine; n = 18, r =  − 0.2247, p = 0.37, timolol; n = 8, r = 0.1905, p = 0.65, brinzolamide; n = 10, r =  − 0.3697, p = 0.29

Full size image
Fig. 6
figure 6

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by corneal endothelial cell density. a Brimonidine; n = 18, r = 0.1084, p = 0.67, timolol; n = 8, r = 0.0952, p = 0.82, brinzolamide; n = 10, r = 0.0426, p = 0.91. b Brimonidine; n = 18, r =  − 0.0661, p = 0.79, timolol; n = 8, r = 0.0952, p = 0.82, brinzolamide; n = 10, r =  − 0.0545, p = 0.88

Full size image
Fig. 7
figure 7

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by tear secretion. a Brimonidine; n = 17, r = 0.0358, p = 0.89, timolol; n = 7, r = 0.1802, p = 0.70, brinzolamide; n = 10, r = 0.0948, p = 0.79. b Brimonidine; n = 17, r = 0.4392, p = 0.078, timolol; n = 7, r = 0.0180, p = 0.97, brinzolamide; n = 10, r = 0.5061, p = 0.14

Full size image
Fig. 8
figure 8

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by eye axial length. a Brimonidine; n = 18, r =  − 0.2986, p = 0.23, timolol; n = 8, r =  − 0.2994, p = 0.47, brinzolamide; n = 10, r =  − 0.6140, p = 0.059. b Brimonidine; n = 18, r = 0.0077, p = 0.98, timolol; n = 8, r =  − 0.4791, p = 0.23, brinzolamide; n = 10, r = 0.1394, p = 0.70

Full size image
Fig. 9
figure 9

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by height. a Brimonidine; n = 40, r =  − 0.3591, p = 0.23, timolol; n = 8, r = 0.0719, p = 0.87, brinzolamide; n = 10, r =  − 0.1785, p = 0.62. b Brimonidine; n = 42, r =  − 0.1233, p = 0.44, timolol; n = 8, r =  − 0.6587, p = 0.076, brinzolamide; n = 10, r = 0.3252, p = 0.36

Full size image
Fig. 10
figure 10

Scatterplot of brimonidine, timolol or brinzolamide concentrations in the aqueous (a) and vitreous (b) humor by BMI. a Brimonidine; n = 40, r = 0.0966, p = 0.55, timolol; n = 8, r =  − 0.5238, p = 0.18, brinzolamide; n = 10, r =  − 0.1885, p = 0.60. b Brimonidine; n = 42, r =  − 0.0615, p = 0.70, timolol; n = 8, r =  − 0.5714, p = 0.14, brinzolamide; n = 10, r = 0.2000, p = 0.58. BMI body mass index

Full size image

Discussion

Drug therapy plays a crucial role in medical care, and it is important to identify the optimal drug therapy for each individual in terms of efficacy and safety. Large individual differences are present in the pharmacokinetics of drugs, and it is necessary to clarify the factors influencing these fluctuations and select a drug therapy suitable for each individual. Eye drops are no exception. This study aimed to investigate the patient-related factors influencing the intraocular penetration of instilled drugs. We evaluated the aqueous and vitreous humor concentrations of drugs in humans after the instillation of eye drops containing brimonidine in three clinical studies and found that age affects drug penetration into the aqueous humor.

There were no differences between the sexes in the concentrations of brimonidine, timolol and brinzolamide in the aqueous or vitreous humor. To the best of our knowledge, there are no reports on sex differences in the intraocular penetration of eye drops in humans. However, there have been reports on the shape and size of the eye in relation to sex [11,12,13,14]. Most ocular tissues have sex steroid hormone receptors, and sex hormones influence the anatomy and function of the eye [15,16,17,18,19]. In addition, various drug-metabolizing enzymes and transporters are present in the ocular tissues, and their involvement in drug metabolism and transport has been suggested [20, 21]. Furthermore, sex differences in drug-metabolizing enzymes have been reported in rat ocular tissues [22, 23], suggesting that there may be sex differences in humans as well. As this study was limited by the number to drugs, further investigation on potential sex differences in intraocular penetration is required.

Brimonidine concentration in the aqueous and vitreous humor did not differ significantly between the phakia and pseudophakia groups. A previous clinical study using 0.2% brimonidine tartrate ophthalmic solution showed higher vitreous humor concentrations of brimonidine in pseudophakic eyes compared with those in phakic eyes, but with no significant difference [24]. Previous studies have also reported that the penetration of brimonidine into the posterior ocular tissues occurs mainly via the conjunctival-scleral route after ocular administration [25, 26]. A previous study using nipradilol suggested that diffusion from the posterior periocular tissues across the posterior sclera, and not through the anterior chamber, might be the main route for the local penetration of the instilled drug into the ocular posterior segment [27]. In addition, the concentrations of brimonidine, timolol and brinzolamide in the lens after the administration of eye drops in rabbits were much lower than those in the conjunctiva and sclera, which makes the lens an unlikely major penetration route to the posterior eye segment after the administration of eye drops [28,29,30]. Therefore, the presence or absence of lens may not affect vitreous humor penetration in humans.

A positive correlation between age and the concentrations of brimonidine and brinzolamide in the aqueous humor was found, and that between age and timolol concentration in the aqueous humor showed a trend in the same direction. A previous clinical study using a 0.5% timolol ophthalmic solution also found a weak positive correlation between age and timolol concentration in the aqueous humor [31]. Age-related structural and functional changes occur in all ocular tissues [32, 33]. Interestingly, age did not affect the penetration of eye drops into the vitreous humor in this study. Therefore, although the detailed mechanism is unknown, age-related changes in the cornea [34] may affect the penetration of the drugs into the aqueous humor.

No correlation was found between the corneal thickness and the concentrations of brimonidine, timolol or brinzolamide in the aqueous or the vitreous humor. Corneal endothelial cell density showed similar results. Several clinical studies have reported no significant correlation between the corneal thickness and drug concentration in the aqueous humor [35, 36]. A positive correlation was reported between corneal thickness and timolol concentration in the aqueous humor, but the correlation was weak [31]. The cornea comprises the corneal epithelium (the main barrier for drug absorption into the eye), stroma and endothelium, with relative thicknesses of approximately 0.1:1:0.01 [37,38,39]. Therefore, it is reasonable that there is no correlation between the intraocular penetration of drugs and corneal thickness, which predominantly reflects the thickness of the corneal stroma.

No correlation was found between tear secretion (Schirmer’s test) and the concentrations of brimonidine, timolol or brinzolamide in the aqueous or vitreous humor. No significant correlation was found between tear secretion and timolol concentration in the aqueous humor in a previous clinical study using 0.5% timolol ophthalmic solution [31]. Dilution by the tear film and subsequent drainage (lacrimal and eyelid movement) reduces the bioavailability of all topically administered drugs [40]. The volume of brimonidine-related eye drops and the tear secretion of participants in this study ranged from 30 to 35 µl and 0 to 36 µl, respectively, suggesting that the effect of tear dilution was low. In addition, Schirmer’s test can only detect tear secretion; however, the ability to drain tears, rather than tear secretion, may be more closely associated with intraocular penetration of eye drops in humans.

No correlation was found between the axial length and the concentrations of brimonidine, timolol or brinzolamide in the aqueous or vitreous humor. Generally, drugs penetrate the eye via diffusion and are distributed to each ocular tissue. Therefore, we expected to find a correlation between the eye axial length and drug concentration in the vitreous humor; however, no such correlation was observed. The average eye axial length is approximately 24 mm, and an eye axial length ≥ 26 mm is defined as excessive myopia [41]. The width of the eye axial length in this study, which included participants with excessive myopia, was 22.3–31.5 mm; however, no correlation was observed. Therefore, axial length may not be relevant to the penetration of drugs into the vitreous in humans.

No correlation was found between height and the concentrations of brimonidine, timolol or brinzolamide in the aqueous or vitreous humor. BMI showed similar results. Only timolol concentration in the vitreous humor showed a trend toward a negative correlation with weight. A study on rabbits reported that systemic absorption was responsible for about half of the drug levels in the vitreous humor and retina following eyedrop instillation [4,5,6]. Therefore, it is possible that the pharmacokinetics of the whole body may affect the pharmacokinetics of the posterior segment of the eye. In participants with obesity, the changes associated with significant weight gain in drug distribution volume and function or blood flow to the liver and kidneys affect systemic pharmacokinetics [42]. The volume of distribution for β-adrenoceptor blockers (especially lipophilic agents) is consistently increased by approximately 30% in patients with obesity compared with those without obesity [43]. Moreover, clearance increases or decreases owing to pathological changes in the liver, such as the development of cirrhosis and changes in the expression of drug-metabolizing enzymes [43]. Although timolol is reportedly metabolized by CYP2D6 [44], there are no reports on the effect of obesity on CYP2D6 expression and activity. Although not reported with timolol, obesity can modify the pharmacokinetics of β-adrenoceptor blockers [42,43,44,45]. The reason why the correlation was only evident for timolol in this study is unknown, and further studies are needed on the correlation between the vitreous penetration of the drug and weight.

The present study has some limitations. A limited number of compounds and eye drop formulations were included in this study. The ocular disposition and elimination of a therapeutic agent are dependent upon its physicochemical properties, including lipophilicity, water solubility and molecular size and formulation compositions of the eye drops [46, 47]. The three compounds included in this study are low-molecular compounds. In the Biopharmaceutics Classification System (BCS) classification, brimonidine and timolol are classified as Class 1 (high solubility, high membrane permeability), and brinzolamide is classified as Class 2 (low solubility, high membrane permeability) [48, 49]. Therefore, for high-molecular compounds and compounds with low membrane permeability, such as Class 3 (high solubility, low membrane permeability) and Class 4 (low solubility, low membrane permeability), patient background factors that affect intraocular penetration may differ. Furthermore, this study is one of the few reports describing the intraocular penetration of eye drops in humans. Rabbits are commonly used for pharmacokinetic studies of eye drops; however, species differences between rabbits and humans have not yet been elucidated. In this study, although there were differences in the formulation compositions among Aiphagan, Aibeta and Ailamide, the brimonidine concentrations in the aqueous and vitreous humor were comparable among the three eye drops containing brimonidine. A similar tendency was observed in studies on rabbits [50, 51]; we consider that there are no species differences in the effects of formulation properties on ocular pharmacokinetics.

Conclusions

In conclusion, this study examined the influence of several different patient-related factors on the intraocular penetration of drugs following topical administration. The findings of this study emphasize the necessity of considering individual differences in ocular pharmacokinetics during drug therapy (formulation design of the eye drops and dose regimen). In particular, drug therapy with eye drops that takes age into account may be important in the context of personalized medicine.