A C T A 0 P H T H A L M 0 L0G IC A
70 (1992)561-569
Long-term kinetic vitreous fluorophotometry Lars L. Knudsen', Thomas Olsenl and Folmer Nielsen-Kudsk2 Department of Ophthalmology', and Institute of Pharmacology2, University of Arhus, Arhus, Denmark
Abstract. Fluorophotometric measurements of vitreous and plasma fluorescence were performed in 14 normal subjects up to 24 h after injection of a single intravenous dose of sodium fluorescein. The data were subjected to a kinetic two-compartmentalanalysis, including the determination of the transfer rate constants between the central and the peripheral compartment (KI2and K2,)as well as between the central and the vitreous compartment (K," and K,,,,). In the central compartment (plasma) a mean terminal disposition rate constant (p) of free fluorescein of 0.23 h-' was found, corresponding to a half-life of 3.01 h. The vitreous fluorescence reached a maximum 2-5 h after the injection and then declined monoexponentially and very slowly (tI/2 = 9.6 h). The rate constant of permeation into the eye (K,,,) was found to be 0.66 h-', while the rate constant of elimination of fluorescein from the vitreous was 0.072 h-l &J. Kin was found to be significantly higher than K,,, presumably indicating an active transport mechanism for fluorescein located at the blood-ocular barrier. &,, was significantly lower than K,,, reflecting a slow vitreous elimination of fluorescein. A permeability index defined as the percentage ratio between the areas under the vitreous and the plasma concentration curves was found to be 3.5%, illustrating the poor penetration of fluorescein into the vitrous. Kinetic long-term fluorophotometry appears to be a promising new tool in the study of the blood-ocular barrier. Key words: vitreous fluorophotometry - fluorescein blood-ocular barrier - pharmacokinetics - permeability.
Vitreous fluorophotometry (VF) was introduced in 1975 by Cunha-Vaz et al. to quantify the breakdown of the blood-ocular barrier in diabetic subjects. According to the original examination proto:I(; Aria
o p l l ~ l l ~ l70.5 l.
col, the method involved the photometric detection of fluorescence in the posterior vitreous 1 h after intravenous injection of fluorescein (F). An increased fluorescence in the vitreous was hypothesized to indicate increased influx of F across the retina, and consequently, an increased permeability of the blood-ocular barrier. The measurement of the 1 h fluorescence in the posterior vitreous has since then been synonymous with VF in numerous studies (Zeimeret al. 1985; Lund-Andersen et al. 1987; Larsen et al. 1983; Chahal et al. 1985; Cunha-Vaz et al. 1978). Although the simplicity of the original VF protocol seemed attractive in a clinical setting, it soon became apparent that the interpretation of the results was not as simple. Numerous sources of error in the measurement of F in the posterior vitrous have been identified: filter characteristics and the importance of subtracting baseline levels (Zeimer et al. 1985), lens hyperfluorescence and absorption (Van Best et al. 1985), vitreous abnormalities (Praeger et al. 1983), contribution of fluorescence from the chorioretinal layer and other artefacts due to the limited resolution of the detection system (Zeimer et al. 1985; Blair et al. 1985; Delori et al. 1985), pH and temperature in the vitreous (Blair et al. 1985; Lakowicz 1983), possible binding of fluorescein in the vitreous (Knudsen 1987) and the presence of an active transport system located in the retina (Blair et al. 1985; Cunha-Vaz & Maurice 1967). In order to quantlfy the transport processes across the blood-ocular barrier it is generally agreed that the concentrationof free F in the blood 561
should be taken into account. Since only the unbound molecule is available for membrane transport, plasma protein binding should also be considered (Lund-Andersen et al. 1982; Blair et al. 1986). A mathematical description of the time course of free plasma F is in this way essential. Unfortunately, the first pass effect in relation to oral administration cannot be taken into account. It is now known that F is rapidly metabolized in the liver to fluorescein glucuronide (FG) which, unfortunately, is a fluorophore also. A selective detection of F and FG can be performed in plasma, but is more complicated in the vitreous (Blair et al. 1986; Chahal et al. 1985; Lund-Andersen et al. 1987). Attempts have been made to calculate a permeability coefficient for F of the blood-ocular barrier (Larsen et al. 1983). Because the vitreous is an unstirred compartment, an elaborate mathematical model including the diffusion of F in the vitreous is necessary. Others have referred to the calculation of a permeability index (Chahal et al. 1985),defined as the ratio between the 1 h vitreous fluorescence and the integrated time course of plasma F concentrations. All of the above studies have been based on the same examination protocol, where the concentration of F is calculated after a fluorophotometric recording at a fixed time after i.v. administration. According to this protocol, the fluorophotometric recording is made from a point varying from near the macula to the center of the vitreous. The point near the macula seems to have been chosen as close as possible to the retina with due regard being paid to the limited spatial resolution of the detection system. The time (60 min) seems to have been chosen because at that time the initial hyperfluorescence from the retina has subsided and yet, at the same time, only minimal redistribution of F within the eye has occurred, so that all the measurable F may be assumed to have come from the retina. Conventional VF is thus designed to describe the penetration of F into the eye, but is based on measurements made at a time of disequilibrium, where a gradient of F concentration exists not only across the blood-ocular barrier, but also within the vitreous. In this system, however, no information can be gathered on the elimination of F from the eye. To study this, an extended time protocol is necessary. 562
Very few studies have dealt with the time course of F concentrations in the vitreous. Zeimer et al. (1983) measured the vitreous F over a 3-h interval and found that the concentration increased over this interval of time. The decay phase was not reached, however. Blair et al. (1986)found maximal posterior viterous fluorescence 10 h after oral administration. Kayazawa & Miyake (1989)measured the vitreous fluorescence up to 48 h after intravenous administration, which enabled them to study differences in vitreous decay constants of fluorescein between normals and diabetic subjects. The present study was initiated in order to investigate the time course of permeation and elimination of F from the vitreous. Compartmental, pharmacokinetic analyses of the time course of free F concentrations in the vitreous and plasma were performed simultaneously in order to obtain data for comparison of kinetic parameters for free fluorescein. Such analyses have not been performed previously for the vitreous compartment, but have been used for several years in the study of aqueous transport mechanisms (Kinsey & Barany 1949; Kinsey & Palm 1955).
Material and Methods 1. Subjects
Fourteen normal volunteers aged between 19-68 years were investigated. They comprised 8 medical students, 3 patients with otological diseases and 3 patients with strabismus. Except for squint and minor refractive errors, all patients were in good ocular health as revealed by a general ocular examination including the measurement of intraocular pressure, biometry, and keratometry. In 5 subjects VF was performed twice, with approximately 1 month between first and second investigation, and in 4 subjects a third time after oral administration of F. 2. Vitreous fluorophotometry (Fig. 1)
The fluorophotometric system was based on a rebuild slit-lamp (Zeiss). The excitation filter passed light below 490 nm. The emitted light was picked up at the end of a fiber optical probe placed in the ocular of the slit-lamp and guided into a photomultiplier (EM1 Thorn 6256s 29954). The barrier filter passed light above 515 nm. An infrared filter was used to block a leakage from the excitation filter within the interval 720-1100 nm.
PEN RECORDER
AMPLIFIER
PHOTOMULTIPLIER
OPTIC F~~~~
SLIT LAMP VIEW
I
ing on a schematic eye model, the point of measurement was found to be approximately 12 mm anterior to the retina, i.e. almost in the center of the eye. Approximately 'I2 h before each vitreous measurement the pupil was dilated (tropicamide 1% and phenylephrine 10%).An antecubital intravenous bolus injection (14 mg/kg bodyweight) of sodium fluorescein was administered. Before administration of F, the background fluorescence from the vitreous and the anterior chamber was measured. The fluorescence in the vitreous and the anterior chamber was measured approximately 2,4,6,8, 10, 12, 14,22 and 24 h after administration of F. After each vitreous measurement a calibration curve was made using known concentrations of F dissolved in a phosphate buffer (0.1 M, pH 7.4). In a separate investigation,linearity between F concentrations and measured fluorescence was found within the range 10-Y-lO-fi M. The lower limit of detection defined as two times vitreous autofluorescence was found to be 1.7 lo-#M (mean).The variation between repeated measurements in the vitreous was about 1.3%.
I
3. Plasma fluorescence
Fig. 1. Schematic diagram of slit-lamp equipment for vitreous fluorophotometry.The eye fixates the incident slit light, which is normal to the corneal surface. The examiner moves the microscope until the fiber optic in the ocular superimposes a point in the vitreous. Two pin-lights mounted on an arm to the slit-lamp are reflected at the corneal surface in order to control the angles of measurement (insert shows slit-lamp view). Matched interference filters are inserted to block long wavelengths (>490 nm) of the incident light and to block short wavelengths (
70 (1992)561-569
Long-term kinetic vitreous fluorophotometry Lars L. Knudsen', Thomas Olsenl and Folmer Nielsen-Kudsk2 Department of Ophthalmology', and Institute of Pharmacology2, University of Arhus, Arhus, Denmark
Abstract. Fluorophotometric measurements of vitreous and plasma fluorescence were performed in 14 normal subjects up to 24 h after injection of a single intravenous dose of sodium fluorescein. The data were subjected to a kinetic two-compartmentalanalysis, including the determination of the transfer rate constants between the central and the peripheral compartment (KI2and K2,)as well as between the central and the vitreous compartment (K," and K,,,,). In the central compartment (plasma) a mean terminal disposition rate constant (p) of free fluorescein of 0.23 h-' was found, corresponding to a half-life of 3.01 h. The vitreous fluorescence reached a maximum 2-5 h after the injection and then declined monoexponentially and very slowly (tI/2 = 9.6 h). The rate constant of permeation into the eye (K,,,) was found to be 0.66 h-', while the rate constant of elimination of fluorescein from the vitreous was 0.072 h-l &J. Kin was found to be significantly higher than K,,, presumably indicating an active transport mechanism for fluorescein located at the blood-ocular barrier. &,, was significantly lower than K,,, reflecting a slow vitreous elimination of fluorescein. A permeability index defined as the percentage ratio between the areas under the vitreous and the plasma concentration curves was found to be 3.5%, illustrating the poor penetration of fluorescein into the vitrous. Kinetic long-term fluorophotometry appears to be a promising new tool in the study of the blood-ocular barrier. Key words: vitreous fluorophotometry - fluorescein blood-ocular barrier - pharmacokinetics - permeability.
Vitreous fluorophotometry (VF) was introduced in 1975 by Cunha-Vaz et al. to quantify the breakdown of the blood-ocular barrier in diabetic subjects. According to the original examination proto:I(; Aria
o p l l ~ l l ~ l70.5 l.
col, the method involved the photometric detection of fluorescence in the posterior vitreous 1 h after intravenous injection of fluorescein (F). An increased fluorescence in the vitreous was hypothesized to indicate increased influx of F across the retina, and consequently, an increased permeability of the blood-ocular barrier. The measurement of the 1 h fluorescence in the posterior vitreous has since then been synonymous with VF in numerous studies (Zeimeret al. 1985; Lund-Andersen et al. 1987; Larsen et al. 1983; Chahal et al. 1985; Cunha-Vaz et al. 1978). Although the simplicity of the original VF protocol seemed attractive in a clinical setting, it soon became apparent that the interpretation of the results was not as simple. Numerous sources of error in the measurement of F in the posterior vitrous have been identified: filter characteristics and the importance of subtracting baseline levels (Zeimer et al. 1985), lens hyperfluorescence and absorption (Van Best et al. 1985), vitreous abnormalities (Praeger et al. 1983), contribution of fluorescence from the chorioretinal layer and other artefacts due to the limited resolution of the detection system (Zeimer et al. 1985; Blair et al. 1985; Delori et al. 1985), pH and temperature in the vitreous (Blair et al. 1985; Lakowicz 1983), possible binding of fluorescein in the vitreous (Knudsen 1987) and the presence of an active transport system located in the retina (Blair et al. 1985; Cunha-Vaz & Maurice 1967). In order to quantlfy the transport processes across the blood-ocular barrier it is generally agreed that the concentrationof free F in the blood 561
should be taken into account. Since only the unbound molecule is available for membrane transport, plasma protein binding should also be considered (Lund-Andersen et al. 1982; Blair et al. 1986). A mathematical description of the time course of free plasma F is in this way essential. Unfortunately, the first pass effect in relation to oral administration cannot be taken into account. It is now known that F is rapidly metabolized in the liver to fluorescein glucuronide (FG) which, unfortunately, is a fluorophore also. A selective detection of F and FG can be performed in plasma, but is more complicated in the vitreous (Blair et al. 1986; Chahal et al. 1985; Lund-Andersen et al. 1987). Attempts have been made to calculate a permeability coefficient for F of the blood-ocular barrier (Larsen et al. 1983). Because the vitreous is an unstirred compartment, an elaborate mathematical model including the diffusion of F in the vitreous is necessary. Others have referred to the calculation of a permeability index (Chahal et al. 1985),defined as the ratio between the 1 h vitreous fluorescence and the integrated time course of plasma F concentrations. All of the above studies have been based on the same examination protocol, where the concentration of F is calculated after a fluorophotometric recording at a fixed time after i.v. administration. According to this protocol, the fluorophotometric recording is made from a point varying from near the macula to the center of the vitreous. The point near the macula seems to have been chosen as close as possible to the retina with due regard being paid to the limited spatial resolution of the detection system. The time (60 min) seems to have been chosen because at that time the initial hyperfluorescence from the retina has subsided and yet, at the same time, only minimal redistribution of F within the eye has occurred, so that all the measurable F may be assumed to have come from the retina. Conventional VF is thus designed to describe the penetration of F into the eye, but is based on measurements made at a time of disequilibrium, where a gradient of F concentration exists not only across the blood-ocular barrier, but also within the vitreous. In this system, however, no information can be gathered on the elimination of F from the eye. To study this, an extended time protocol is necessary. 562
Very few studies have dealt with the time course of F concentrations in the vitreous. Zeimer et al. (1983) measured the vitreous F over a 3-h interval and found that the concentration increased over this interval of time. The decay phase was not reached, however. Blair et al. (1986)found maximal posterior viterous fluorescence 10 h after oral administration. Kayazawa & Miyake (1989)measured the vitreous fluorescence up to 48 h after intravenous administration, which enabled them to study differences in vitreous decay constants of fluorescein between normals and diabetic subjects. The present study was initiated in order to investigate the time course of permeation and elimination of F from the vitreous. Compartmental, pharmacokinetic analyses of the time course of free F concentrations in the vitreous and plasma were performed simultaneously in order to obtain data for comparison of kinetic parameters for free fluorescein. Such analyses have not been performed previously for the vitreous compartment, but have been used for several years in the study of aqueous transport mechanisms (Kinsey & Barany 1949; Kinsey & Palm 1955).
Material and Methods 1. Subjects
Fourteen normal volunteers aged between 19-68 years were investigated. They comprised 8 medical students, 3 patients with otological diseases and 3 patients with strabismus. Except for squint and minor refractive errors, all patients were in good ocular health as revealed by a general ocular examination including the measurement of intraocular pressure, biometry, and keratometry. In 5 subjects VF was performed twice, with approximately 1 month between first and second investigation, and in 4 subjects a third time after oral administration of F. 2. Vitreous fluorophotometry (Fig. 1)
The fluorophotometric system was based on a rebuild slit-lamp (Zeiss). The excitation filter passed light below 490 nm. The emitted light was picked up at the end of a fiber optical probe placed in the ocular of the slit-lamp and guided into a photomultiplier (EM1 Thorn 6256s 29954). The barrier filter passed light above 515 nm. An infrared filter was used to block a leakage from the excitation filter within the interval 720-1100 nm.
PEN RECORDER
AMPLIFIER
PHOTOMULTIPLIER
OPTIC F~~~~
SLIT LAMP VIEW
I
ing on a schematic eye model, the point of measurement was found to be approximately 12 mm anterior to the retina, i.e. almost in the center of the eye. Approximately 'I2 h before each vitreous measurement the pupil was dilated (tropicamide 1% and phenylephrine 10%).An antecubital intravenous bolus injection (14 mg/kg bodyweight) of sodium fluorescein was administered. Before administration of F, the background fluorescence from the vitreous and the anterior chamber was measured. The fluorescence in the vitreous and the anterior chamber was measured approximately 2,4,6,8, 10, 12, 14,22 and 24 h after administration of F. After each vitreous measurement a calibration curve was made using known concentrations of F dissolved in a phosphate buffer (0.1 M, pH 7.4). In a separate investigation,linearity between F concentrations and measured fluorescence was found within the range 10-Y-lO-fi M. The lower limit of detection defined as two times vitreous autofluorescence was found to be 1.7 lo-#M (mean).The variation between repeated measurements in the vitreous was about 1.3%.
I
3. Plasma fluorescence
Fig. 1. Schematic diagram of slit-lamp equipment for vitreous fluorophotometry.The eye fixates the incident slit light, which is normal to the corneal surface. The examiner moves the microscope until the fiber optic in the ocular superimposes a point in the vitreous. Two pin-lights mounted on an arm to the slit-lamp are reflected at the corneal surface in order to control the angles of measurement (insert shows slit-lamp view). Matched interference filters are inserted to block long wavelengths (>490 nm) of the incident light and to block short wavelengths (
