Digoxin-Induced Reversible Dysfunction of the Cone Photoreceptors in Monkeys

Abstract

Purpose.: To investigate functional alteration of the retina induced by digoxin in monkeys.

Methods.: Digoxin was intravenously administered to cynomolgus monkeys and standard full-field electroretinograms (ERGs) were serially recorded. In other digoxin-treated monkeys, the rod and cone a-waves to high-intensity flashes were obtained and analyzed by the a-wave fitting model (a-wave analysis). The following responses were also recorded: dark- and light-adapted responses to flashes of different intensities (dark- and light-adapted luminance responses), photopic ERG elicited by long-duration stimulus (ON-OFF response), and the photopic negative response (PhNR).

Results.: Delayed b-wave was observed in all responses of the standard full-field ERGs; amplitude of the b-wave was increased in the rod response, but was decreased in the single-flash cone response and the 30-Hz flicker. These changes recovered gradually after elimination of digoxin from the blood. Digoxin enhanced and delayed the b-wave in the dark-adapted luminance-response analysis regardless of stimulus intensity. In the light-adapted luminance-response analysis, digoxin attenuated the a- and b-waves only at high and middle stimulus intensity, respectively. The a-wave analysis revealed selective decrease in the maximum response parameter (R_max ) in the cone a-wave. Both the b- and d-waves of the ON-OFF response were delayed.

Conclusions.: The selectively reduced R_max in the cone a-wave indicated dysfunction of the cone photoreceptors in digoxin-treated monkeys. Meanwhile, the enhanced and delayed rod response suggested alteration of retinal components other than the cone photoreceptors. These results may contribute to the understanding of digoxin-induced visual disturbances in humans. It is suggested that the cone function is markedly, but not exclusively, affected in the retina of such patients.

Introduction

Digoxin, an Na⁺/K⁺ ATPase inhibitor, is widely used for the treatment of congestive heart failure. It is well known that digoxin can produce alterations in the visual system of patients, such as reduced visual acuity, photophobia, and blurred or yellow vision. ^1–5 Reversible abnormalities in electroretinograms (ERGs), indicating functional impairment of the retina, have been reported in digoxin-treated patients who complained of these visual disturbances. ^1,3–5 The site of the digoxin-induced visual defects, therefore, is generally considered to be the retina. 

It is known that the inner segment of the photoreceptors contains Na⁺/K⁺ ATPase, which plays an important role to maintain a dark current along the photoreceptors. Furthermore, it has been reported that digoxin, an Na⁺/K⁺ ATPase inhibitor, induced concentration-dependent reductions in the light response of isolated photoreceptors from amphibians. ⁴ Thus, it is speculated that the reversible visual disturbances in digoxin-treated patients are associated with inhibition of Na⁺/K⁺ ATPase in the inner segment of the photoreceptors. 

In case reports, Weleber et al. ¹ and Nagai et al. ⁵ recorded rod- and cone-driven ERGs separately in patients during clinical digoxin toxicity and found subnormal amplitude and prolonged implicit time for the cone-driven ERGs. On the other hand, the rod-driven ERGs obtained simultaneously from those patients were within normal limits. These results indicated that the cone pathway was more affected than the rod pathway in the retina of patients with digoxin intoxication. However, those clinical ERGs in patients seem to not be sufficient to accurately identify the functional impairment of the photoreceptors. This could be attributable to the following limitations of ERGs in clinical practice or studies in patients: (1) there are no waveform components that originate almost exclusively from the photoreceptors in ERGs as generally assessed in clinical practice (i.e., it has been shown that the a-wave of such ERGs includes not even a little contribution from postreceptoral components of the retina) ^6,7 , and (2) digoxin intoxication is an life-threatening condition, which makes it difficult to obtain investigative ERGs with extended protocols, and a burden to patients. 

One efficient means to examine digoxin-induced alteration of retinal function would be to assess the ERGs with extended protocols not in patients with digoxin intoxication but in animals treated with digoxin that exceeds clinical doses. However, there are very few reports that describe digoxin-provoked changes in the ERG in experimental animals. To our knowledge, no studies have examined ERGs in experimental animals with blood digoxin levels comparable with those reported in patients with visual disturbances; although, Maehara et al. ⁸ assessed ERGs in dogs with plasma digoxin levels within the therapeutic range. 

Taking these findings together, detailed functional abnormalities of the retina induced by a high level of digoxin are unknown in patients or animals, and therefore the exact mechanisms of digoxin-induced visual disturbances need to be elucidated. Thus, the purpose of this study was to intensively investigate digoxin-induced changes of the retinal function in experimental animals. For this purpose, we first administered digoxin at a high-dosage level to monkeys, a species in which the physiology and anatomical structure of the retina is widely known to be similar to that in humans. The clinically-relevant ERGs were serially recorded to confirm that monkeys are an appropriate experimental animal to investigate mechanism of digoxin-induced visual disturbances and to find the peak of the ERG change. Thereafter, we evaluated the digoxin-induced retinal toxicity in detail by analysis of extended-protocol ERGs recorded at the time of the peak effect. 

Methods

Animals

A total of 14 cynomolgus monkeys ( Macaca fascicularis ) between 3 and 5 years of age were used in this study. The animals were housed individually in stainless steel cages (W 60 cm × D 68 cm × H 75 cm) in an animal study room where the environmental condition was set as follows: room temperature, 24°C; relative humidity, 60%; illumination, 12-hour lighting (7 AM to 7 PM) at 300 luces. The animals were fed 100 g/animal/day of pellet food for monkeys (PS; Oriental Yeast Co., Ltd., Tokyo, Japan). Tap water from a feed-water nozzle was supplied ad libitum to the animals. All experimental procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and were approved by the Institutional Animal Care Committee of Daiichi Sankyo Co., Ltd. 

Study Design

Digoxin was administered to six animals, and standard full-field ERGs were serially recorded (immediately after, 24 hours, 7, 14, and 28 days after dosing) as described below. Six days before the dosing of digoxin, physiological saline (Otsuka normal saline; Otsuka Pharmaceutical Factory, Inc., Tokushima, Japan) was administered to the six animals in the same manner, and the standard full-field ERGs were recorded 24 hours after dosing. Based on the results of the standard full-field ERGs, a definitive study was conducted in eight other animals. They were assigned to the digoxin- or vehicle-treated groups (4 animals/group) so that the group means and variances for ERG parameters obtained prior to the start of this experiment were as nearly equal as possible. Digoxin or physiological saline was administered to these animals. From these eight animals, extended-protocol ERGs were recorded before and 24 hours after dosing as described below. 

Drug Administration

An injectable formulation of digoxin (DIGOSIN; Chugai Pharmaceutical Co., Ltd., Tokyo, Japan) was diluted with physiological saline (Otsuka normal saline; Otsuka Pharmaceutical Factory, Inc.). The drug was administered intravenously for 40 minutes to animals at a dose of 0.1 mg/kg. The dose was expected to produce exposure in monkeys dozens of times higher than that in human patients, based on the recommended oral dosage (0.125–0.25 mg/d) ⁹ and bioavailability (66%) ¹⁰ in humans. 

Clinical Observation

For the six digoxin-treated animals in which the standard full-field ERGs were serially recorded, clinical signs were observed every day up to 28 days after dosing. 

Animal Preparation for ERG Recording

The animals were anesthetized with intramuscular injection of ketamine hydrochloride (Ketalar Intramuscular 500 mg; Daiichi Sankyo Co., Ltd., Tokyo, Japan) (10 mg/kg initial dose, 5–10 mg/kg/h maintenance dose) and 0.6 mg/kg xylazine hydrochloride (Celactal; Bayer Medical Ltd., Osaka, Japan). The pupils were dilated with topical 0.5% tropicamide and 0.5% phenylephrine hydrochloride (Mydrin-P ophthalmic solution; Santen Pharmaceutical Co., Ltd., Osaka, Japan); the corneas were anesthetized with topical 0.4% oxybuprocaine hydrochloride (Benoxil ophthalmic solution 0.4%; Santen Pharmaceutical Co., Ltd.) and protected with topical hydroxyethylcellulose (Scopisol solution for eye; Takeda Chemical Industries, Ltd., Osaka, Japan). 

Visual Stimulation

Three systems for visual stimulation were used in this study. One was for obtaining the standard full-field ERGs, and the other two systems were for obtaining the extended-protocol ERGs (i.e., responses for a-wave analysis, and for dark- or light-adapted luminance-response analysis, ON-OFF response, and single-flash cone response [R/B]). The intensities of flashes generated in these systems were measured by a calibrated photometer (IL1700; International Light Technologies, Inc., Peabody, MA) and an optical detector (SED033/Y/R; International Light Technologies, Inc.). 

The standard full-field ERGs were elicited by white xenon photostrobe flashes with a nominal flash duration of approximately 30 μs that were generated by a Ganzfeld stimulator (Ganzfeld Stimulator Model 2503S; LKC Technologies, Inc., Gaithersburg, MD). The maximum flash intensity measured in the dome was 6.3 cd·s/m². The stimulus intensity was attenuated with neutral-density gelatin filters (Wratten; Eastman Kodak Company, Rochester, NY). 

The ON-OFF response was elicited by a long flash of white light that was generated with a light-emitting diode (LED). The LEDs were built into contact lens recording electrodes (Contact lens electrode with built-in LED; Mayo Corporation, Aichi, Japan). The luminance and duration of the flashes and the background luminance were controlled by a LED stimulator (LS-C; Mayo Corporation). 

The responses for the a-wave analysis and for the dark- or light-adapted luminance-response analysis, and the single-flash cone response (R/B) were elicited with another Ganzfeld stimulator (BigShot Ganzfeld; LKC Technologies, Inc.) that was equipped with two different types of light source: xenon photostrobe and LEDs. White light flashes at an intensity of 60 cd·s/m² or more were generated with the xenon photostrobe. The maximum intensity of the white flash generated with the xenon photostrobe was 826.7 phot cd·s/m². The stimulus intensity was altered by varying the capacitance and applied voltage of a capacitor using software (Ganzfeld control panel; LKC Technologies, Inc.) installed in a personal computer (Dell, Inc., Austin, TX). The white light flashes that were below 60 cd·s/m², colored-light flashes and steady background illumination were generated with the following LEDs housed in the stimulator; red (λ_max = 627 nm), green (λ_max = 530 nm), and blue (λ_max = 470 nm). The white flashes were produced by combining the output from these three LEDs. The duration of all stimuli was less than 5 ms. The maximum intensity of white flash generated with the LEDs in the stimulator was 27.3 phot cd·s/m². The stimulus intensity and the background luminance were altered by varying the LED pulse duration using software (Ganzfeld control panel; LKC Technologies, Inc.) installed in a personal computer (Dell, Inc.). 

Recording and Analysis of the Standard Full-Field ERGs

Standard full-field ERGs were evaluated according to the guidelines ¹¹ of the International Society for Clinical Electrophysiology of Vision (ISCEV). A bipolar contact lens electrode (H6515NFC; Mayo Corporation) was placed on the corneal surface of the left eye. A ground electrode (TN208-016; Unique Medical Co., Ltd., Tokyo, Japan) was attached to the parietal region of the scalp. Following 40 minutes or more of dark adaptation, the rod-driven response (the rod response) and the rod- and cone-driven response (the combined rod–cone response) were elicited by white light flashes at an intensity of 0.007 and 2.7 phot cd·s/m², respectively. Subsequently, after 10 minutes of light adaptation with a white background light at 40 phot cd/m², the single-flash cone response and the 30-Hz flicker were elicited by white light flashes at an intensity of 2.7 phot cd·s/m² under the white background light. Responses were amplified 10,000 times and were filtered with a band pass from 0.5 to 1000 Hz. The amplified signals were stored in the evoked potential test equipment (MEB-9104; Nihon Kohden Corporation, Tokyo, Japan). A limited number of waveforms (3–10) for each response were averaged. In the waveform analysis, the amplitude and implicit time of the a-wave were measured from baseline to the a-wave trough and from stimulus onset to the a-wave trough, respectively, for the combined rod–cone response and the single-flash cone response; the amplitude and implicit time of the b-wave were measured from the a-wave trough to the b-wave peak and from stimulus onset to the b-wave peak, respectively, for all the responses. 

Recording and Analysis of the Extended-Protocol ERGs

Prior to ERG recording, pupil diameter of the left eye was measured to calculate retinal illuminance (in trolands [td]) for the a-wave analysis. The ERGs were recorded from the left eye of each animal with the same bipolar contact lens electrode as in the recording of the standard full-field ERGs, except for recording the ON-OFF response, which was recorded with a contact lens electrode with built-in LED. All responses were amplified, filtered, and stored in the same way as that in recording the standard full-field ERGs. 

Dark-Adapted Luminance-Response Analysis.

Following 40 minutes or more of dark adaptation, scotopic ERGs were elicited with white flashes of increasing intensity ranging from −3.2 to 1.4 log phot cd s/m². Three responses to stimuli of the same intensity were averaged. For the waveform analysis, amplitude of the a- and b-waves and implicit time of the b-wave were measured in the same way as that in the standard full-field ERGs. 

A-Wave Analysis.

Three steps of stimuli ranging from 3.7 to 4.3 log scot td-s were used to elicit the rod- and cone-driven response in the fully dark-adapted state. For each stimulus intensity, three flashes were given with 30-second intertrial intervals and then the three responses elicited were averaged. Thereafter, the cone-driven response in the fully dark-adapted state were elicited by means of a transient rod-saturation procedure similar in principle to that used by Friedburg et al., ¹² Robson et al., ⁷ and Jeffrey et al. ¹³ In our procedure, the white test stimulus was delivered 500 ms after the conditioning white flash of 4.0 log scot td-s. We conducted a preliminary experiment in cynomolgus monkeys (not illustrated), in which pairs of identical conditioning flashes of 4.0 log scot td-s were delivered at a range of interflash intervals and the time course of recovery of the amplitude of the a-wave elicited by the second flash was examined as a function of interflash interval. We found an initial rapid but partial recovery during the interflash intervals from 50 to 300 ms that was followed by slower complete recovery during the interflash intervals from 700 ms to 20 seconds. Based on this result, we assumed that during the time between these two recovery phases the cones would have fully recovered but the rods would still be fully saturated. To elicit the cone-driven response, three steps of test stimuli ranging from 3.4 to 4.0 log phot td-s, which were identical to those used to elicit the rod- and cone-driven response, were used. For each stimulus intensity, six pairs of flashes were given with 3 seconds intertrial interval and then the six responses elicited were averaged. Rod-isolated responses were obtained by subtracting the cone-driven responses from the rod- and cone-driven responses. To evaluate rod function, the leading edge of the a-wave of the rod-isolated response (i.e., the rod a-wave) was fitted to a curve by the Hood and Birch modification ¹⁴ of the Lamb and Pugh model ¹⁵ :  

 

 

In Equation 1, I is the flash intensity (log scot td-s); t_d is the time delay (ms); t is the time after the flash onset (ms); S is the sensitivity (s⁻²(td-s)⁻¹); and R_max is the maximum response amplitude (μV). The value of t_d was fixed at 3.1 ms, which is the mean value from age-matched healthy monkeys, and S and R_max were varied for the best fit. To evaluate cone function, the leading edge of the a-wave of the cone-driven response (i.e., the cone a-wave) was fitted to a Michaelis-Menten version ¹⁶ of Equation 1 combined with an exponential filter:  

 

 

In Equation 2, I is the flash intensity (log phot td-s); t_d is the time delay (ms); t is the time after the flash onset (ms); S is the sensitivity (s⁻³(td-s)⁻¹); and R_max is the maximum response amplitude (μV). The value of t_d was fixed at 1.8 ms, which is the mean value from age-matched healthy monkeys, and S and R_max were varied for the best fit. 

Light-Adapted Luminance-Response Analysis.

Following 10 minutes of light adaptation with a white background light at 29.0 phot cd/m², photopic ERGs were elicited with white flashes of increasing intensity ranging from −1.3 to 2.6 log phot cd s/m². Ten responses to stimuli of the same intensity were averaged. For the waveform analysis, amplitude of the a- and b-waves and implicit time of the b-wave were measured in the same way as that in the standard full-field ERGs. 

ON-OFF Response.

A photopic ERG was elicited by a 200-ms white-light stimulus at luminance of 63.0 phot cd/m² under a white background light at 32 phot cd/m². Ten to 30 responses were averaged. For the waveform analysis, amplitude of the b- and d-waves were measured from the a-wave trough to the b-wave peak and from the baseline at the time point of stimulus offset to the d-wave peak, respectively. In addition, implicit times of the b- and d-waves were measured from stimulus onset to the b-wave peak and from stimulus offset to the d-wave peak, respectively. 

Single-Flash Cone Response (R/B).

A photopic ERG was elicited by red light flash at an intensity of 2.9 phot cd s/m² under blue background light at 6.9 phot cd/m². Ten responses were averaged. For the waveform analysis, amplitude of the b-wave and the photopic negative response (PhNR) were measured from the a-wave trough to the b-wave peak and from baseline to the PhNR trough, respectively. Additionally, the PhNR/b-wave amplitude ratio was calculated as the ratio of the PhNR amplitude to the b-wave amplitude. 

Ophthalmoscopy

After recording the standard full-field ERGs at each time point until 28 days after dosing, the fundi of both eyes were inspected with a binocular indirect ophthalmoscope (HEINE OMEGA 500; HEINE Optotechnik GmbH & Co., KG, Herrsching, Germany). 

Toxicokinetics

A volume of approximately 0.6 mL of blood was collected from the femoral vein immediately after the recording of the ERGs to measure the serum digoxin concentration. In the standard full-field ERGs study, blood was collected just after recording the ERGs on the day of dosing and 24 hours after dosing (1.15–1.25 and 25.23–25.45 hours after starting the infusion of digoxin, respectively). Since serum digoxin was below the lower limit of quantitation (0.3 ng/mL) 7 days after dosing in all these animals, blood samples were not collected thereafter. In the extended-protocol ERGs study, blood was collected just after recording the ERGs 24 hours after dosing (24.58–24.72 hours after starting the infusion of digoxin). Ten days after recording the extended-protocol ERGs from four monkeys given digoxin, 0.1 mg/kg of digoxin was administered again to the same animals in the same manner. Then, blood was collected in the same manner at before dosing, 10, 20, and 40 minutes, and 1, 2, 4, 7, 24, 48, and 72 hours after the start of dosing. The serum was prepared from the blood samples by letting the samples sit at room temperature for 20 minutes followed by centrifugation at 1200g for 10 minutes. The serum was then stored at −80°C until measurement. The serum concentration of digoxin was determined by enzyme-multiplied immunoassay (Emit 2000 Digoxin Assay; Dade Behring, Inc., Cupertino, CA). 

Statistics

For statistical analysis of the standard full-field ERGs, the paired t-test with the Bonferroni correction was used to assess the difference between the values before and after dosing. For statistical analysis of the dark- and light-adapted luminance-response data, a two-way ANOVA was used to assess the differences between the vehicle- and digoxin-treated groups and between the values before and after dosing. The two-way ANOVA was followed by a Student's t-test or a paired t-test only if statistical significance was observed not only on the group comparison but also on the interaction between response and stimulus intensity. For statistical analysis of the ERGs other than those mentioned above, the Student's t-test was used to assess the significance between the vehicle- and digoxin-treated groups and the paired t-test was used to assess the significance between the values before and after dosing. The differences were considered to be significant when P was less than 0.05. 

Results

Clinical Observation

Vomitus including whitish froth was observed in five out of six animals within 24 hours after dosing. These five animals were anorexic on the day after dosing. Subsequently, no treatment-related findings were observed in any animals until the end of the study period (i.e., 28 days after dosing). 

Electroretinogram

Standard Full-Field ERGs.

Typical waveforms of the standard full-field ERGs at baseline and those recorded after digoxin dosing are shown in Figure 1. The amplitude and implicit time of the ERG components are summarized in Table 1. The rod response was enhanced and delayed: significant increases in the amplitude and implicit time of the b-wave after dosing were detected. The combined rod–cone response also was delayed: the implicit time of the a- and b-waves significantly increased after dosing. The single-flash cone response and the 30-Hz flicker were delayed and attenuated: significant increases in the b-wave implicit time after dosing were detected. Five out of six animals showed obviously decreased b-wave amplitude after dosing in both the single-flash cone response and the 30-Hz flicker, although a statistically significant difference was not detected. The changes in the standard full-field ERGs mentioned above were most prominent at 24 hours after dosing. These changes gradually but fully recovered within 28 days after dosing. 

Figure 1

 

Typical waveforms of the standard full-field ERGs in a digoxin-treated monkey. Digoxin at a dose of 0.1 mg/kg was administered, and the standard full-field ERGs were serially obtained as described in the text. Arrowheads indicate onset of the light flashes. The responses at baseline (gray trace) are superimposed on those obtained after dosing of digoxin (black trace). Each trace represents an average of 3 to 10 responses.

Table 1.

 

Effects of Digoxin on the Standard Full-Field ERGs

	Times After Dosing
	Baseline*	Immediately	24 h	7 d	14 d	28 d
Rod response
Implicit time, ms
b-wave	83.5 ± 2.82	87.0 ± 2.27†	91.3 ± 4.49‡	89.5 ± 5.42	84.2 ± 1.95	84.8 ± 2.52
Amplitude, μV
b-wave	49.0 ± 6.60	58.9 ± 16.58	72.3 ± 12.32†	59.8 ± 20.06	47.5 ± 14.40	51.3 ± 9.48
Combined rod–cone response
Implicit time, ms
a-wave	17.3 ± 0.48	17.4 ± 0.30	18.4 ± 0.64‡	18.0 ± 0.69‡	17.8 ± 0.78	17.3 ± 0.44
b-wave	38.0 ± 1.81	39.4 ± 1.66‡	42.9 ± 2.56†	40.9 ± 3.29	38.7 ± 1.81	38.2 ± 1.41
Amplitude, μV
a-wave	58.4 ± 8.44	58.6 ± 6.89	57.5 ± 8.09	57.9 ± 9.16	66.8 ± 15.83	65.0 ± 9.59
b-wave	141.5 ± 16.36	138.6 ± 19.59	148.2 ± 11.19	168.6 ± 23.35	158.0 ± 31.08	159.0 ± 23.94
Single-flash cone response
Implicit time, ms
a-wave	11.6 ± 0.43	11.8 ± 1.08	12.4 ± 1.09	12.6 ± 0.88	12.0 ± 0.42	11.9 ± 0.34
b-wave	23.9 ± 0.71	24.7 ± 1.00	26.2 ± 1.00‡	26.1 ± 1.11‡	25.1 ± 0.61‡	24.6 ± 0.69
Amplitude, μV
a-wave	14.4 ± 2.64	12.3 ± 1.10	13.7 ± 2.45	14.7 ± 4.09	14.8 ± 1.70	13.5 ± 2.81
b-wave	57.7 ± 10.90	53.0 ± 7.41	47.9 ± 7.85	53.7 ± 14.71	60.5 ± 7.31	59.5 ± 10.69‡
30-Hz flicker
Implicit time, ms
b-wave	25.7 ± 0.67	26.3 ± 0.69	27.1 ± 0.69‡	26.9 ± 0.77†	26.7 ± 0.88	26.2 ± 0.59
Amplitude, μV
b-wave	57.8 ± 6.77	54.5 ± 9.51	47.8 ± 9.96	54.5 ± 15.36	61.1 ± 8.57	61.5 ± 6.55

 Data are expressed as the mean ± SD of six animals.

*  24 hours after dosing of physiological saline.

†   P < 0.05; significantly different in comparison with the baseline value by the paired t-test with Bonferroni correction.

‡   P < 0.01; significantly different in comparison with the baseline value by the paired t-test with Bonferroni correction.

Extended-Protocol ERG: Dark-Adapted Luminance-Response Analysis.

Typical waveforms of the dark-adapted response to flashes of increasing intensity in vehicle- and digoxin-treated monkeys are shown in Figure 2, and intensity-response functions of the dark-adapted response in vehicle- and digoxin-treated monkeys are shown in Figure 3. Digoxin enhanced and delayed the b-wave. In the digoxin-treated group, the amplitude and implicit time of the b-wave were significantly increased in comparison with both the vehicle-control values and the predosing values: no interaction was detected between these changes and stimulus intensity. 

Figure 2

 

Typical waveforms of the dark-adapted response to flashes of increasing intensity in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the dark-adapted responses to flashes of increasing intensity were obtained as described in the text. The responses obtained before dosing (gray trace) are superimposed on those obtained 24 hours after dosing (black trace) in the right panels. Each trace represents an average of three responses.

Figure 3

 

Intensity-response functions of the dark-adapted response in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered. The dark-adapted response to flashes of increasing intensity were recorded before (open circles) and after (closed circles) dosing, and the amplitude of the a- and b-waves (A) and the implicit time of the b-wave (B) were measured as described in the text. Data are expressed as mean ± SD of four animals. The values in the digoxin-treated group were significantly different in comparison with the values in the vehicle-control group (†P < 0.01) and the predosing values (*P < 0.01) by the two-way ANOVA. In the digoxin-treated group, significant differences were detected in comparison with the vehicle-control value (†P < 0.01) and the predosing value (*P < 0.01) by the two-way ANOVA.

Extended-Protocol ERG: Light-Adapted Luminance-Response Analysis.

Typical waveforms of the light-adapted response to flashes of increasing intensity in vehicle- and digoxin-treated monkeys are shown in Figure 4, and intensity-response functions of the light-adapted response in vehicle- and digoxin-treated monkeys are shown in Figure 5. Digoxin remarkably attenuated the a-wave. In the digoxin-treated group, the a-wave amplitude was significantly decreased in comparison with both the vehicle-control values and the predosing values. The interaction between response and stimulus intensity of the a-wave amplitude in the digoxin-treated group was significant: the attenuated a-wave compared with the predosing values was detected only at high stimulus intensities. The b-wave amplitude before dosing in the digoxin-treated group was significantly higher than that in the vehicle-treated group. The b-wave amplitude after dosing in the digoxin- and vehicle-treated groups did not differ, since the b-wave amplitude significantly decreased in the digoxin-treated group after dosing. The interaction between response and stimulus intensity of the b-wave amplitude in the digoxin-treated group was significant: significant attenuation in the b-wave was detected only at middle stimulus intensities. The implicit time of the b-wave in the digoxin-treated group was significantly increased after dosing in comparison with both the vehicle-control values and the predosing values: no interactions were detected between these changes and stimulus intensity. 

Figure 4

 

Typical waveforms of the light-adapted response to flashes of increasing intensity in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the light-adapted responses to flashes of increasing intensity were obtained as described in the text. The responses obtained before dosing (gray trace) are superimposed on those obtained 24 hours after dosing (black trace) in the right panels. Each trace represents an average of 10 responses.

Figure 5

 

Intensity-response functions of the light-adapted response in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered. The light-adapted response to flashes of increasing intensity were recorded before (open circles) and after (closed circles) dosing, and the amplitude of the a- and b-waves (A) and the implicit time of the b-wave (B) were measured as described in the text. Data are expressed as mean ± SD of four animals. The values in the digoxin-treated group were significantly different in comparison with the values in the vehicle-control group (†P < 0.01) and the predosing values (*P < 0.01) by the two-way ANOVA. Significant difference in the b-wave amplitude was also detected between the vehicle- and digoxin-treated groups only before dosing (§P < 0.01). The interaction between response and stimulus intensity of the a- and b-waves amplitude was significant (P < 0.01) in the digoxin-treated group, and significant differences in comparison with the predosing value were detected by the paired t-test (#P < 0.05, ##P < 0.01).

Extended-Protocol ERG: A-Wave Analysis.

Typical waveforms of the rod a-wave and the cone a-wave in vehicle- and digoxin-treated monkeys are shown in Figures 6 and 7, respectively. Model parameters of the a-wave analysis are summarized in Figure 8 and Table 2. In the rod a-wave, no marked changes were found in the waveform or in the model parameters. In the cone a-wave, the R_max in the digoxin-treated group was significantly reduced in comparison with both the vehicle-control values and the predosing values; the mean R_max after dosing in the digoxin-treated group was lower than that in the vehicle-treated group by 0.23 log units (41% reduction). On the other hand, no change was detected in the S. 

Figure 6

 

Typical waveforms of the rod a-wave in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the rod a-waves were derived by subtracting cone responses from combined rod-cone responses as described in the text. The dotted lines signify the curves fit from Equation 1 in the text. The responses obtained before dosing (gray trace) are superimposed on those obtained 24 hours after dosing (black trace) in the right panels.

Figure 7

 

Typical waveforms of the cone a-wave in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the dark-adapted cone responses were elicited as described in the text. The dotted lines signify the curves fit from Equation 2 in the text. The responses obtained before dosing (gray trace) are superimposed on those obtained 24 hours after dosing (black trace) in the right panels. Each trace represents an average of six responses.

Figure 8

 

The effect of digoxin on the a-wave in monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the model parameters (S, R_max ) were determined in the rod- and cone a-waves before and 24 hours after dosing as described in the text. Data are expressed as the mean ± SD of four animals. Ranges indicated by gray areas signify the 95% confidence intervals based on the values from age-matched healthy cynomolgus monkeys (N = 92). Significant differences in the log R_max in the cone a-wave were detected by the paired t-test (*P < 0.01) and the Student's t-test (†P < 0.05).

Table 2.

 

Effects of Digoxin on the ERG Parameters of Extended Protocols

	Vehicle-Treated Group		Digoxin-Treated Group
	Before	24 h After Dosing	Before	24 h After Dosing
A-wave analysis
Rod a-wave
Log S, s−2(td-s)−1	1.14 ± 0.195	1.11 ± 0.138	1.18 ± 0.095	1.08 ± 0.176
Log Rmax , μV	2.00 ± 0.153	2.00 ± 0.161	1.97 ± 0.137	2.00 ± 0.158
Cone a-wave
Log S, s−3(td-s)−1	3.74 ± 0.114	3.75 ± 0.108	3.88 ± 0.120	3.86 ± 0.228
Log Rmax , μV	1.63 ± 0.099	1.62 ± 0.122	1.63 ± 0.124	1.39 ± 0.130*‡
ON-OFF response
Implicit time, ms
b-wave	28.6 ± 1.94	29.2 ± 1.60	27.5 ± 1.86	34.4 ± 3.30*†
d-wave	20.1 ± 1.82	20.5 ± 1.96	20.2 ± 1.62	24.3 ± 3.59†
Amplitude, μV
b-wave	49.9 ± 21.26	47.4 ± 19.16	69.4 ± 6.51	60.1 ± 16.15
d-wave	24.8 ± 12.57	24.0 ± 10.68	31.7 ± 1.64	38.0 ± 14.84
Single-flash cone response (R/B)
Amplitude, μV
b-wave	36.6 ± 11.09	33.1 ± 9.51	44.5 ± 23.22	43.3 ± 17.80
PhNR	12.3 ± 5.46	12.8 ± 8.28	14.1 ± 2.66	13.1 ± 2.13
PhNR/b-wave ratio	0.38 ± 0.210	0.40 ± 0.246	0.37 ± 0.149	0.35 ± 0.157

 Data are expressed as the mean ± SD of four animals.

*   P < 0.05; significantly different in comparison with the vehicle-treated group by the Student's t-test.

†   P < 0.05; significantly different in comparison with the baseline value by the paired t-test.

‡   P < 0.01; significantly different in comparison with the baseline value by the paired t-test.

Extended-Protocol ERG: ON-OFF Response.

Typical waveforms of the ON-OFF response in vehicle- and digoxin-treated monkeys are shown in Figure 9. Electroretinogram parameters are summarized in Table 2. Digoxin induced a general delay in the ON-OFF response. In the digoxin-treated group, the implicit time of the b-wave was significantly increased in comparison with both the vehicle-control values and the predosing values. Meanwhile in the d-wave implicit time, significant increase compared with the predosing values was detected in the digoxin-treated group. 

Figure 9

 

Typical waveforms of the ON-OFF response in vehicle- and digoxin-treated monkeys. Vehicle or digoxin at a dose of 0.1 mg/kg was administered, and the ON-OFF responses were obtained as described in the text. The responses obtained before dosing (gray trace) are superimposed on those obtained 24 hours after dosing (black trace) in the right panel. Each trace represents an average of 10 to 30 responses. The horizontal line at the bottom of each trace represents turning the stimulus on and off. Note delayed b- and d-waves after digoxin treatment.

Extended-Protocol ERG: Single-Flash Cone Response (R/B).

Electroretinogram parameters for this response are summarized in Table 2. No change was noted in the amplitude of the b-wave or the PhNR, or the PhNR/b-wave amplitude ratio. 

Ophthalmoscopy

No apparent changes in fundus oculi were observed in any animals treated with digoxin. 

Toxicokinetics

Serum digoxin concentrations at the time of ERG recording are shown in Table 3, and serum concentrations and toxicokinetic parameters of digoxin after single dosing are shown in Table 4. The serum concentration of digoxin at the time of recording the standard full-field ERGs immediately after dosing was 20.8 ± 3.65 ng/mL. The area under the serum concentration-time curve up to infinity (AUC_0-inf) of digoxin after single dosing was 334.4 ± 65.43 ng·h/mL. 

Table 3.

 

Serum Digoxin Concentrations at the Time of ERG Recording

	Dose, mg/kg	Number of Animals	Time Point (After Dosing)	Serum Concentration, ng/mL
Standard full-field ERGs	0.1	6	Immediately	20.8 ± 3.65
		6	24 h	2.8 ± 0.30
		6	7 d	NC*
Extended- protocol ERGs	0.1	4	24 h	4.5 ± 0.93

  Data are expressed as the mean ± SD. NC, not calculated.

*  All individual values were below the lower limit of quantitation (0.3 ng/mL).

Table 4.

 

Serum Concentrations and Toxicokinetic Parameters After Digoxin Dosing

Dose, mg/kg	Number of Animals	Serum Concentration, ng/mL										AUC0-inf, ng·h/mL	Cmax, ng/mL	t1/2, h
		Time After the Start of 40-min Infusion
		Pre	20 min	40 min	1 h	2 h	4 h	7 h	24 h	48 h	72 h
0.1	4	NC*	41.5 ± 11.44	52.1 ± 24.97	22.5 ± 4.47	9.7 ± 2.19	5.5 ± 0.94	4.8 ± 0.62	2.7 ± 0.38	2.0 ± 0.45	1.4 ± 0.24	334.4 ± 65.43	52.7 ± 24.61	44.8 ± 25.00

 Data are expressed as the mean ± SD.

*  Three individual values were below the lower limit of quantitation (0.3 ng/mL).

Discussion

Comparison of Digoxin Exposures in Monkeys and in Humans

The blood digoxin levels in patients with visual disturbances ranges from 2.2 to 13.1 ng/mL. ^1–5 In this study, the mean serum concentration of digoxin at the time of recording standard full-field ERGs was 20.8 ± 3.65 ng/mL. Thus, the blood digoxin level in monkeys showing altered ERGs substantially exceeded that reported in the patients. 

The clinical oral dosage of digoxin is 0.125 to 0.25 mg/d. ⁹ In a randomized crossover study of the pharmacokinetics of 0.25 mg/d digoxin in healthy subjects, the AUC at steady-state over a uniform dosing interval τ (AUC_τ,ss) was 15.4 ng·h/mL. ¹⁷ In this study, the AUC_0-inf of digoxin after single dosing was 334.4 ± 65.43 ng·h/mL. Therefore, the AUC of digoxin in the monkeys also substantially exceeded that reported in humans. 

Comparison of ERG Changes in Monkeys and in Humans

In digoxin-treated patients with visual disturbances, serial recording of ERG revealed reversible attenuation in the cone-driven ERGs, ^1,5 suggesting retinal dysfunction in the cone pathway. Meanwhile, alterations in the rod-driven ERGs and the rod- and cone-driven ERGs from these patients were mild and generally within normal limits. It was also reported in patients with digoxin intoxication that, whereas digoxin blood levels returned to normal within days, abnormal ERGs and visual symptoms took weeks to recover. ^4,5 In the standard full-field ERGs recorded from digoxin-treated monkeys, attenuated responses were observed only in the photopic ERGs. None of these ERGs had fully recovered by 7 days after dosing by that time serum digoxin levels were already undetectable in all monkeys. These changes in ERGs finally recovered by 28 days after dosing. Based on the similarities in property and time course of the changes in ERGs, the cynomolgus monkey is considered to be an appropriate animal model for investigating the mechanism of digoxin-induced visual disturbances. 

Retinal Function

Cone Pathway.

In this study, the photopic a-wave was attenuated by digoxin only at high stimulus intensities in the light-adapted luminance-response analysis. The origin of the a-wave in the photopic ERG from monkeys has been shown to be the OFF-pathway postsynaptic to the cone photoreceptors as well as the cone photoreceptors themselves. ⁶ It has also been shown that the cone photoreceptors contribute significantly to the total a-wave for brighter stimuli. ⁶ Therefore, our data mentioned above suggest that the function of the cone photoreceptors was affected in the monkeys given digoxin. 

For a more in-depth investigation into the functional effects of digoxin on the photoreceptors, an a-wave analysis based on the Hood and Birch methods ^14,16 was conducted in digoxin-treated monkeys. The R_max of the cone a-wave was reduced, whereas no apparent change in the S was noted. Decreased R_max seems attributable to the loss of large sections of the photoreceptors and/or the shortening of their outer segments. ^16,18 However, it is obvious that the cause of the decreased R_max in this study was not morphologic abnormalities in the photoreceptors since the ERG changes were reversible. Hood et al. ¹⁸ have shown that steady background illumination induced a marked but reversible decrease in R_max , while S was substantially unchanged in healthy humans. They considered that the reduction in the R_max was caused by the polarization of the membrane of the photoreceptors (i.e., the dark current was decreased) as a result of the steady field. Hence, similarity in the characteristic change of the a-wave parameters (i.e., selectively and reversibly reduced R_max ) between digoxin-treated monkeys and human subjects suggests that the dark current of the photoreceptors was decreased in digoxin-treated monkeys. This speculation is supported by the observation that digoxin inhibits Na⁺/K⁺ ATPase, an enzyme that greatly contributes to maintenance of the dark current of the photoreceptors. The reduced R_max in this study, furthermore, is consistent with the previous report that digoxin reduced the light response in isolated photoreceptors from tiger salamanders. ⁴  

In this study, digoxin delayed the photopic b-waves from low to high stimulus intensities in the light-adapted luminance-response analysis. In the ON-OFF response assessment, an increased implicit time was identified both in the b- and d-waves, indicating that the delay occurred both in the cone ON- and OFF-pathways in the retina. However, the mechanisms by which the delay in the cone pathway is produced are unclear. 

Interestingly, digoxin attenuated the photopic b-wave only at middle stimulus intensities in the light-adapted luminance-response analysis. Although the exact mechanism of this phenomenon remains to be elucidated it could be explained based on the greater delay in the b-wave than in the d-wave (Fig. 9). The b-wave of the photopic flash ERG in primates has been shown to originate mainly from the ON-bipolar cells and to be shaped by the interaction between the ON- and OFF-bipolar cells (the push–pull model). ¹⁹ On the other hand, Ueno et al. ²⁰ have shown in the photopic ERG from monkeys that the OFF-component was markedly delayed at higher intensity in comparison with the ON-component. This result indicates that the OFF-bipolar cells contribute little to the amplitude of the b-wave of the photopic flash ERG elicited with higher stimulus intensity in monkeys. Therefore, at middle stimulus intensity the digoxin-induced increase in phase lag between two positive components derived from each type of bipolar cells would lead to the amplitude reduction. Meanwhile, at high stimulus intensity the increased phase lag might have little effect on the response amplitude. 

Rod Pathway.

Digoxin induced no apparent change in either of the model parameters (R_max or S) of the rod a-wave in the a-wave analysis in the present study. Additionally, in the dark-adapted luminance-response analysis no obvious attenuation in the a-wave was observed after digoxin dosing. These results suggested that the function of the rod photoreceptors was generally preserved. In isolated photoreceptors, cones were approximately 50-fold more sensitive to digoxin than rods. ⁴ From this difference, it is speculated that the retinal digoxin level was below that at which obvious dysfunction in the rod photoreceptors would be induced in monkeys in this study even at the peak of the ERG changes. Although the reasons for this difference between rods and cones are unclear, it is also speculated that Na⁺/K⁺ ATPase isotype in the cones might be more sensitive to digoxin than that in the rods. Further investigations are needed to test this hypothesis. 

In digoxin-treated monkeys, the scotopic b-waves were enhanced and delayed in the dark-adapted luminance-response analysis. These changes were observed even at stimulus intensities that were subthreshold for the cone photoreceptors (−2.1 log phot cd·s/m² or lower), and therefore were considered not to be directly related to the cone dysfunction indicated by the reduced R_max in the cone a-wave. One possible explanation of the enhanced and delayed b-wave is that the alteration in the b-wave was secondary to the digoxin-induced alteration in the RPE. Mice given intravenous sodium iodate, a RPE toxicant, showed increased maximum amplitude and delayed peak latency in the b-wave of the scotopic ERG accompanied by histopathologic damage in the RPEs. ^21,22 On the other hand, one interesting feature of the RPEs is the predominant localization of Na⁺/K⁺ ATPase in their apical membrane. ²³ In addition, inhibition of Na⁺/K⁺ ATPase by ouabain decreased transepithelial electrical resistance in cultures of human RPEs. ²⁴ Taking these observations together, it is speculated that the enhanced and delayed b-wave in the present study might be related to decreased RPE resistance, resulting from inhibition of Na⁺/K⁺ ATPase in the RPEs by digoxin. 

Inner Retina.

It has been reported that cardiac glycosides could cause a toxic effect directly upon the optic nerve. ²⁵ Additionally, radioisotope distribution studies revealed that digoxin was concentrated in the retinal ganglion cells (RGCs), ²⁶ the axons of which constitute the optic nerve. These reports implied that digoxin has the potential to affect RGC. To address this question in digoxin-treated monkeys, we assessed the PhNR, a waveform component of the ERG that has been shown to originate mainly from the RGCs in monkeys. ²⁷ As a result, no obvious change was detected in the PhNR/b-wave amplitude ratio, an indicator of selective PhNR loss. ²⁸ Our data indicate that the function of the inner retina including the RGCs was generally preserved in the monkeys given digoxin. 

Conclusions

The selectively reduced R_max in the cone a-wave indicated dysfunction of the cone photoreceptors in digoxin-treated monkeys. Meanwhile, the enhanced and delayed rod response suggested alteration of retinal components other than the cone photoreceptors. These results may contribute to the understanding of digoxin-induced visual disturbances in humans. It is suggested that the cone function is markedly, but not exclusively, affected in the retina of such patients. 

Acknowledgments

The authors thank Eiichiro Nagasaka and Hidetaka Kudo of Mayo Corporation for technical assistance. 

Disclosure: J. Kinoshita, None; N. Iwata, None; T. Kimotsuki, None; M. Yasuda, None