Digitalis-Induced Visual Disturbances with Therapeutic Serum Digitalis Concentrations

Methods

Six patients (aged 66 to 85 years) were referred to ophthalmologists for evaluation of photopsia (five patients) or decreased visual acuity (one patient). Findings are summarized in Table 1 and Table 2. Serum digoxin concentrations were measured by the Becton Dickinson Affinity enzyme immunoassay method (Becton Dickinson, Orangeburg, New York). Electroretinography was done by placing bipolar Burian-Allen contact lens electrodes on the anesthetized cornea. Pupils were dilated to 7 to 8 mm in diameter with 1% neosynephrine and 1% cyclopentolate. Cone electroretinograms were obtained by Ganzfeld flashes of white light at 5.1 Hz on a relatively bright Ganzfeld (white) background of 5 to 10 × 10³ photopic trolands [14]. To optimize the signal:noise ratio, about 50 to 100 responses were averaged by a Nicolet Viking II signal averager (Nicolet, Madison, Wisconsin).

Results and Discussion

Visual Symptoms

The six patients described in Table 1 and Table 2 had visual symptoms characteristic of digitalis intoxication [1, 3-13]. Two distinct clinical patterns were observed (Table 1 and Table 2): Five patients had photopsia or related symptoms, and one patient had decreased visual acuity. In the former group, four patients had photopsia, and one patient reported brightness and glare (which may be a variant of photopsia); the visual acuity of each of these five patients was either normal or at the patient's previous baseline. The remaining patient (patient 6) developed poor central visual acuity and centrocecal scotomas, a recognized toxic effect of cardiac glycoside therapy [3, 7, 10, 11, 13]. He also noted brightness-induced glare, which was worse out-doors and when he awoke, together with “snow on the grass and trees,” which may be a variant subjective interpretation of photopsia.

The patients with photopsia described “white lights” or “snow.” Photopsia was variously described as flickering, flashing, sparkling, or swirling lights or as “light passing through the gaps in vertical blinds” and was characteristically described as occurring peripherally, although in one patient it also occurred centrally. The most distinctive characteristics of the photopsia were the multiplicity of lights and their dependence on external illumination. A similar luminance dependence has been described by other investigators [6, 10, 12], who reported that photopsia and related symptoms in elderly patients who received digitalis were subjectively worse when the patients awoke or when they looked outdoors into daylight. We suspect that the luminance dependence of the visual symptoms may reflect digitalis-induced hyperactivity of retinal photoreceptors. This luminance dependence, the predominance of the photopsia in the peripheral visual fields, its continuous appearance, and the innumerable quantity of lights are distinct qualities that tend to distinguish digitalis-induced photopsia from other photopsias, such as those associated with migraine, occipital epilepsy, retinal detachment, vitreous traction, and retinal emboli. For example, patient 3 had a long history of classic migraine, which had been acephalalgic for 20 years and at the time of the study had consisted primarily of 20-minute episodes of scintillating scotomas without pain. She reported that the new visual phenomena differed considerably from her migrainous scintillations: The lights were more numerous; they were continuously present; they covered the entire visual field; and there were no transient scotomas.

Photopsia varied in severity, depending on the patient. Patient 3 insisted that her physicians “do something” or she would “go crazy,” whereas patient 4 (in whom digoxin therapy was not discontinued) was able to tolerate his photopsia after learning that his symptoms did not reflect serious ocular or neurologic disease.

Evidence for Digitalis Toxicity

All six patients had serum digoxin concentrations within or below the usual therapeutic range (Table 1 and Table 2), and no patient had clinical evidence of cardiac or gastrointestinal manifestations of digitalis intoxication. Patient 6 had noted yellow vision, but it had resolved before our examination, after discontinuation of digoxin therapy. Only one other patient (patient 3) reported any disturbance of color vision (yellow photopsia, along with white photopsia). No patient had evidence of any clinical abnormality known to predispose to digitalis toxicity [15, 16]: There was no laboratory evidence of azotemia or hypokalemia, and there was no clinical evidence of hypothyroidism or of hypomagnesemia. No patient was receiving quinidine, which is known to elevate serum digoxin levels [17] and is associated with the development of visual symptoms in patients who receive digitalis [3, 10, 12, 13, 18]. In some instances, a diagnosis of possible digitalis intoxication had previously been considered by the patients' primary physicians, but, in all six patients, the diagnosis was made by a neuro-ophthalmologist. In the five patients in whom discontinuation of digoxin therapy was possible, the photopsia resolved gradually within days and disappeared within 1 to 2 weeks (restoration of baseline visual acuity in patient 6 took several weeks); reinstitution of digoxin therapy in patient 1 caused recurrence of photopsia in 4 days, which cleared when therapy was again discontinued. Resolution of visual symptoms was associated with normalization of the cone electroretinogram in the four patients so examined (see below). The resolution of visual symptoms after discontinuation of digoxin therapy, together with the normalization of the cone electroretinogram, supported the presumptive diagnosis of digoxin intoxication.

Flickering before the eyes may be a very early sign of digitalis intoxication in some patients [3]. However, our patients had no clinical evidence of cardiac or gastrointestinal manifestations of digitalis intoxication for as long as 12 to 14 months after the appearance of visual symptoms (Table 1 and Table 2); furthermore, no evidence of permanent retinal damage was noted after digoxin therapy was discontinued. These findings suggest that it may be possible for selected patients with digoxin-induced photopsia to continue to receive digoxin therapy if they have no other clinical evidence of digitalis intoxication and if the indications for continuation of digitalis therapy are sufficiently strong. In patients in whom the digitalis dosage is not reduced, careful and prolonged observation (including ambulatory electrocardiographic monitoring and periodic electroretinography) is advisable.

Electroretinographic Studies

Five patients (all patients except patient 3) had cone electroretinographic studies. Four of these patients (patients 2, 4, 5, and 6) had prolonged (> 32.8 ms) b-wave implicit times, which returned to the normal range in the three patients (patients 2, 5, and 6) in whom digoxin therapy was discontinued (Table 1 and Table 2; Figure 1). The fifth patient (patient 1) had b-wave implicit times within the normal range while receiving digoxin therapy; these times were shortened after digoxin withdrawal. Prolongation of the b-wave implicit time during digitalis therapy, with shortening of this time after digitalis withdrawal, suggests a digitalis effect [12]. Cone b-wave amplitudes (normal range, 28 to 232 µ V in persons older than 60 years) were normal or mildly decreased in four of the five electroretinographic studies, but patient 6 (who had poor visual acuity and centrocecal scotomas) had low b-wave amplitudes that returned to the normal range (Figure 1) in conjunction with improvement in visual acuity after discontinuation of digoxin therapy. The increase in b-wave amplitude implicated cone dysfunction as the cause of this patient's decreased visual acuity, as suggested by Weleber and Shults [12].

Our results suggest that prolongation of the b-wave implicit time of the cone electroretinogram may be used to monitor the effect of digitalis. The effect of digitalis on the b-wave implicit time in patients without visual symptoms is not known, and thus prolongation of this interval does not necessarily reflect a toxic retinal effect of digitalis. Because the b-wave implicit time has considerable individual variability, measurements in a group of patients before and during digitalis therapy are required to ascertain whether prolongation is associated with a toxic retinal effect of digitalis.

Mechanisms of Digitalis Effects

Most laboratory studies have suggested a retinal origin for the visual manifestations of digitalis intoxication [9, 12, 19, 20]. The likelihood of a retinal origin is also supported by the fact that several retinal cell types contain abundant quantities of digitalis-sensitive isoforms of sodium- and potassium-dependent adenosine triphosphatase [Na+, K+-ATPase] [21-23], which actively pumps sodium out of, and potassium into, most mammalian cells. The Na+, K+-ATPase pumps of photoreceptor cells, glial Muller cells, and retinal pigment epithelial cells prevent excessive retinal extracellular K+ accumulation in an area of intense neuronal activity [24-26]; inhibition of these pumps by digitalis is associated with alterations in retinal electrical properties [9, 12, 20, 27].

A possible but unproven contributory factor in the development of visual disturbances in our patients was age: Patients tended to be elderly [average age, 77 years; range, 66 to 85 years]. The basis for the sensitivity of these patients to visual manifestations of digitalis toxicity is unknown but may reflect 1) age-related loss of photoreceptor cells [28, 29] or retinal pigment epithelial cells [29] with a loss of Na+, K+-ATPase potassium pumping capacity [26] or 2) an increase in retinal expression of one or both of the more digitalis-sensitive Na+, K+-ATPase isoforms.

Clinical Implications

The small size of this patient group and self-selection for prominence of visual symptoms do not permit us to assess the frequency with which significant digitalis-induced visual disturbances occur in the absence of other clinical features suggesting digitalis toxicity. Nevertheless, our findings indicate the importance of considering the possibility of digitalis intoxication in patients (especially elderly patients) who have new visual symptoms while receiving digitalis, even if serum digitalis levels are within the normal therapeutic range. Such patients often report new visual symptoms to a physician (usually an ophthalmologist) other than the one who prescribed digitalis, and the association may be overlooked. Our experience underscores the continued validity of the warning by Smith [16], who more than two decades ago first pointed out that digitalis-induced arrhythmias and cardiac conduction disturbances could occur in patients with serum digoxin concentrations in the therapeutic range [30] and concluded that “clearly, quantitative technics in current use in no way lessen the necessity for frequent, detailed observation of the digitalized patient” [16].

Appendix: Glossary

Cone b-wave, b-wave implicit time, and b-wave amplitude: properties of the cone electroretinogram (described in the legend to Figure 1).

Ganzfeld: refers to a flash stimulus to the full visual field (as distinguished from point stimulation).

Photopic: refers to the fact that the luminosity of the light is calculated for daylight vision.

Photopsia: subjective sensation of seeing lights that are not present in the environment; these lights are ordinarily white, unless otherwise indicated.

Troland: an international unit that allows pupillary size to be considered in determining retinal illumination.

Drs. Odel, Behrens, and Gouras: Edward S. Harkness Eye Institute, 635 West 165th Street, New York, NY 10032.

Dr. Rath: Department of Ophthalmology, Western Galilee Regional Hospital—Nahariya, Nahariya 2100, Israel.

Dr. Wolin: 1655 East Greenville Street, Anderson, SC 29622.

Dr. Martin: Wake Forest University Eye Center, Bowman-Gray School of Medicine, Winston-Salem, NC 27157.

Dr. Kardon: Department of Ophthalmology, University of Iowa, Iowa City, IA 52242.