Cellular mechanisms underlying the pharmacological induction of phosphenes

Abstract

Visual sensations evoked by stimuli other than luminance changes are called phosphenes. Phosphenes may be an early symptom in a variety of diseases of the retina or of the visual pathways, but healthy individuals may perceive them as well. Phosphene-like phenomena are perhaps the most common side effect reported in clinical pharmacology. Ivabradine, a novel anti-anginal drug that reduces heart-rate by inhibiting the hyperpolarization activated current expressed in cardiac sinoatrial node cells (I_f) induces phosphenes in some patients. One hypothesis is that ivabradine interacts with the visual system by inhibiting hyperpolarization-activated current in retinal cells (Ih). An I_h current with properties similar to cardiac I_f has been reported in retinal neurones. Under normal circumstances most of the random fluctuations generated within the retinal circuits do not reach the level of conscious perception because they are filtered out. Presumably, filtering occurs mostly within the retina and one serious candidate for this action is the ability of I_h to act as a negative-feedback mechanism. I_h activation in the membrane of visual cells causes dampening of responses to slow noisy inputs thus tuning the visual system to perceptually more relevant signals of higher frequency. I_h inhibition, by altering at the retinal synapses the filtering of signals generated by thermal breakdown of rhodopsin or other fluctuations, is expected to increase the probability of phosphene occurrence. It is the purpose of the present paper to outline and discuss the features of the visual system and the pharmacological conditions relevant to phosphene perception.

Introduction

The visual sensations evoked by stimuli other than luminance changes are called phosphenes (from the Greek phos, light and phainain, to show). Phosphenes can be spontaneous or provoked in a number of ways including a gentle pressure on the eyelids, an electrical or magnetic stimulation of the eye or of the visual cortex (Brindley and Lewin, 1968; Tyler, 1978; Barker et al., 1985). von Helmholtz (1896) gave an early review of the effects of pressure stimulation of the eye; Pflügers (1865) showed that the visual threshold for electric stimuli follows his law of electrotonic stimulation. A more recent quantitative account on the threshold electrical stimulation of the eye and of the visual cortex can be found in Attwell (2003) and Tehovnik et al. (2006), respectively. Phosphene induction by electrical stimulation of the visual pathway with electrodes implanted on the retina or on the visual cortex is currently regarded as a very promising method for making the blind see again (Dobelle et al., 1974; Schmidt et al., 1996; Zrenner, 2002).

In general, phosphenes appear spontaneously when the viewer is subjected to prolonged visual deprivation and it has been argued that this occurrence may be related to an increased cortical excitability to the incoming visual input (see Boroojerdi et al., 2000). Phosphenes may appear with a variety of patterns: they often have a chaotic structure in the form of sparks, sometimes they appear as a glowing circle or part of it, or as a spiral moving in concentric circles (see Walker, 1981). In a substantial fraction of migraine sufferers, phosphenes may appear in more organized forms such as, for instance, the so-called fortification structure (Richards, 1971). More complex and picturesque spontaneous visual phenomena are often indicated as phosphenes, but when associated with emotional factors, drugs, alcohol, stress, fever or psychotic conditions, they should be referred to as visual hallucinations. In general, luminous phenomena such as phosphenes are largely geometric forms non-culturally biased, whereas hallucinations are more complex, iconic forms, culturally controlled (Siegel, 1977).

The perception of phosphenes is very common and often experienced in the absence of an identifiable pathological condition of the retina or of the visual pathways: in a large number of these cases the causes underlying phosphene generation are difficult to assess (Rosenbaum et al., 1987). Phosphenes, mostly occurring upon eye movements may be associated with optic neuritis and in this case they may be caused by mechanical aggravation of a damaged or inflamed optic nerve (Davis et al., 1976). Visual symptoms that also include phosphenes constitute an early sign of optic nerve demyelination in patients that will later develop multiple sclerosis (see Wilkstrom et al., 1980; Levin and Lessel, 2003). Severe myopia, changes in eye pressure and vitreous retraction, which may be a prelude to retinal detachment, stimulate the visual cells mechanically thus generating phosphenes (Forzli el and Brasseur, 1999; Enoch et al., 2003). Spontaneous visual perceptions associated with retinal and optic nerve diseases may overlap with those resulting from retrochiasmal disorders (Murtha and Stasheff, 2003). In all these cases, a number of other subjective symptoms, and objective signs of damage as well, can be detected by conventional neurological and ophthalmic tests. Phosphenes are also a prominent symptom associated with the aura, the warning sensation that preludes the onset of migraine (see Queiroz et al., 1997). Finally, phosphenes and other visual hallucinations are often a collateral effect of a variety of pharmacological agents (see Fraunfelder and Fraunfelder, 2001). Attemping a comprehensive explanation of the mechanism underlying drug-induced phosphene seems impractical because of the variety of molecular and cellular actions that one may encounter. It is perhaps useful, however, examining some of the features of the visual system and the pharmacological conditions relevant to phosphene perception.

Spontaneous activity of the visual system in complete darkness

It is well established that the visual system, as well as other brain regions, is continually active and that, in the dark, retinal ganglion cells (RGC) generate a spike discharge whose rate may fluctuate randomly with frequent irregular bursts (dark discharge, see Levick, 1973). This notion supports the psychophysical theory of the visual threshold proposed by Barlow (1965) according to which the visual system operates in the presence of an intrinsic noise. The perceptual correlate of this noise may be identified with the fuzzy grey sensation experienced in darkness even when all after-images have faded (Eigengrau – subjective grey of Hering, 1878; Eigenlicht – subjective light of von Helmholtz, 1896). In the past, there has been considerable debate on the site of origin of the maintained discharge. It has been proposed, for instance, that the dark discharge is the expression of the autochthonous activity of the RGC (Hughes and Maffei, 1965). This explanation was suggested by the fact that the time course of changes in the discharge rate, upon switching off an adapting light, is much shorter than the change in threshold sensitivity of RGC. Other evidence, based on selectively damaging the photoreceptor cells argues against this conclusion (Rodieck, 1967) and together with indirect arguments (Barlow, 1965) seems to place the source of noise in the photoreceptors. On occasions the firing rate of a RGC is observed to undergo spontaneously a remarkable sequence of oscillations that may even drive the discharge frequency to levels above those achieved by illumination changes (Rodieck and Smith, 1966; Cavaggioni, 1968). The origins for this patterned activity have long been debated, and vascular reflexes (Granit, 1955) and changes of ocular pressure (Collins, 1967) have been charged with being the causes. Further analysis of this patterned activity by multineurone recordings in cat (Mastronarde, 1989) and salamander retinas (Brivanlou et al., 1998) have shown that the sources of noise responsible for spontaneous firing in RGC are distributed at different retinal levels. Fluctuations may be observed as early as in photoreceptors where they reflect features of the photo-transductive machinery. A prominent component of this noise is that caused by spontaneous breakdown of rhodopsin (Baylor et al., 1980, 1984; Rieke and Baylor, 1996; Jones, 1998), which has been shown to be effective in modulating the firing activity in ganglion cells (Brivanlou et al., 1998). Owing to the stability of the rhodopsin molecule, thermal breakdown occurs infrequently: an average life of several 100 years has been estimated for a rhodopsin molecule at 37 °C (see Baylor, 1996). One must consider, however, that because of the large number of molecules (up to 10⁹) packed in the disks at the outer segment of a rod, spontaneous breakdown of rhodopsin may occur in a single rod as frequently as every few tens of seconds. Considering then that the retina contains approximately 10⁶ rods with their uncorrelated spontaneous events, it turns out that the frequency at which these signals impinge on a ganglion cell can be very high and may explain some of the features of the spontaneous firing in the RGC. Patterns of sustained activity, however, have been demonstrated also when all chemical transmission is blocked, suggesting that RGC may produce spontaneous current fluctuations that depolarize their membrane above the firing threshold even in the absence of signals from the distal retina (Brivanlou et al., 1998).

The neuronal trans-membrane voltage needed to produce retinal phosphenes by an externally applied electric field has been recently estimated as 0.6–200 μV (Attwell, 2003). It should be noted that the amplitude of the spontaneous fluctuations in the retina usually exceeds this value (Baylor et al., 1984). Under normal circumstances, the fluctuations that spontaneously arise within the retinal circuits are not perceived because the retina possesses efficient mechanisms for filtering out noisy signals (Sampath and Rieke, 2004).

The functional connections between single RGCs and neurones of the visual cortex are very effective in driving simple cortical cells: because of the considerable divergence between retina and lateral geniculate nucleus (LGN) cortical neurones may be influenced by RGCs through multiple forward pathways (Kara and Reid, 2003). Spontaneous, light-independent, changes in the firing of visual cells have been reported in the LGN (Maffei et al., 1965) and inputs from the brainstem have been shown to modulate the transmission of retinal signals to the visual cortex through the LGN (Ozaki and Kaplan, 2006).

The occurrence or aggravation of spontaneous visual phenomena may therefore be expected in the presence of abnormal stimuli, as a consequence of an increased neuronal excitability or because of an impairment of the filtering properties at one or more stages of the visual pathway.

Abnormal visual excitation of retinal photoreceptors

A variety of pathological conditions of the eye and of the visual pathways may induce visual sensations in the absence of light stimuli. In addition to mechanical stimuli such as changes in the eye pressure, vitreous retraction, compression of the optic nerve, of particular interest is the activation of the phototransductive cascade by unliganded opsin. Inappropriate and constant activation of transduction by high levels of opsin occurs in Lebers congenital amaurosis causing photoreceptor degeneration (see Woodruff et al., 2003). A similar mechanism may also be responsible for the degeneration induced by vitamin A deprivation (Fain and Lisman, 1993). In affected patients continuous abnormal stimulation of visual cells that is caused by unliganded opsin is associated with a persistent luminous sensation and background desensitization (light adaptation), even in the absence of light. A similar condition, however, is hardly comparable with the dynamics of phosphene occurrence, generally referred to as transient flash-like events not entailing significant light desensitization.

Phosphenes of cortical origin

It has long been known that visual sensations may be evoked directly acting on the cortex without using low-level visual pathways. Electrical stimulation of the visual cortex in humans via implanted microelectrodes evokes phosphenes in both sighted and blind subjects (Brindley and Lewin, 1968; Dobelle et al., 1974). They are described as circular spots of white or coloured light (Dobelle and Mladejovsky, 1974; Schmidt et al., 1996) that conform to the receptive field characteristics of the non-human cortical neurones (Hubel and Wiesel, 1977; Hubel and Livingston, 1990). Similarly but less invasively, phosphenes may be elicited by transcranial magnetic stimulation (TMS) of the visual cortex (Barker et al., 1985). TMS is a technique whereby a rapidly changing magnetic field of appropriate strength, applied on the scalp surface, induces an electric current that stimulates the underlying cerebral cortex (see Cowey, 2005). TMS, when directed to excite the occipital cortex, induces phosphene perception in the visual field contralateral to the stimulated hemisphere (Meyer et al., 1991). Phosphene thresholds are a reliable parameter characterizing excitability of the occipital cortex (Stewart et al., 2001). An appropriate criterion for threshold determination is to present the subject with a set of TMS stimuli of randomly intermixed different intensities. The subject is asked to report the presence or absence of phosphenes after each stimulus and a sigmoidal function is then fitted to the measured responses. The stimulus intensity corresponding to 50% positive responses is taken as the phosphene threshold.

Although phosphenes may originate at different levels in the visual system, patient studies have demonstrated that they can be perceived only in the presence of an intact primary visual cortex (V1) (Cowey and Walsh, 2000). Furthermore, in normal observers the activation level of V1 determines whether phosphenes are detected (Silvanto et al., 2005).

Transient or persistent changes in the excitability of neurones, especially in the occipital region of the visual cortex, has received strong consideration and appear associated with brain susceptibility to migraine attacks (Welch et al., 1990). TMS threshold for phosphene generation in the visual cortex was reported to be significantly lower in migraine patients who experienced aura than in normal controls (Wray et al., 1995; Palmer et al., 2000; Aurora et al., 2003; Hall et al., 2004). The issue, however, is somewhat controversial and higher or normal phosphene thresholds have also been reported (see Afra et al., 1998; Bohotin et al., 2003; Antal et al., 2006). Other studies with different paradigms have added consistent data to support hyperexcitability of the visual cortex in migraineurs (Mulleners et al., 2001; Battelli et al., 2002; Young et al., 2004; Chronicle et al., 2006). Further corroboration comes from studies showing that drugs effective in preventing migraine all have the common property of diminishing neuronal excitability (see Buchanan et al., 2004; Linde, 2006). Excitability of the visual cortex may also be modulated by a variety of incoming signals: it has been recently shown that the neural activity giving rise to a phosphene is in competition with the cortical activity elicited by the presentation of a visual stimulus (Rauschecker et al., 2004). All these observations agree that any event apt to modify the excitability of the visual system is bound to enhance or reduce the incidence of phosphene perception. These events include a wide variety of conditions such as the occurrence of visual or non-visual stimuli, sensory deprivation, hypoglycaemia, fever, drug intoxication, psychotic episodes and epilepsy.

Drug-induced phosphenes: a possible consequence of I_h inhibition

Phosphene-like phenomena are perhaps the most common side effect reported in clinical pharmacology and a comprehensive list of the registered pharmacological agents that induce phosphenes is available (Fraunfelder and Fraunfelder, 2001). It is interesting to note that both stimulant and depressant agents can provoke phosphenes. This seemingly paradoxical observation may, in fact, be explained by the complexity of the neuronal circuits whereby the action of depressants differ in distinct brain regions: one should consider, for instance, the hyper-excitability phases that precedes narcotic-induced anaesthesia (see Sloan, 1998). Differences in form, colour and movement of phosphene-like events are associated with distinct inducing agents or, for the same agent, with different subjects. In general, the frequency of phosphene phenomena reported by informed subjects is four times higher than that reported by naïve individuals (Siegel, 1977). Despite all the differences it seems reasonable to assume that a wide variety of causal agents induce phosphene-like events either by a nonspecific direct stimulation of the visual system or by increasing the neuronal excitability or by both.

A newly developed class of heart rate-lowering compounds, whose common mechanism of action is the inhibition of a current responsible for the pacemaker activity of cardiac cells, induces visual symptoms in humans. Table 1 summarizes the present knowledge of the effects of these compounds. It should be noted, however, that these drugs were tested in different experimental conditions which do not allow a direct comparison in terms of affinity for I_f and I_h.

Table 1.

I_f and I_h blocking properties of heart rate reducing molecules

	Alinidine	Cilobradine	Ivabradine	Zatebradine	ZD7288
IC50 for If (μM)	28a	0.021b–0.62c	1.5d	0.066b	0.3 μMe
IC50 for Ih in rods (μM)	NA	NA	2.7f	2.0g	5.9h
IC50 for HCN1 HE (μM)	NA	1.15c	2.05c–0.94i	1.83c	41j
Mechanism of block for If	Use-independenta	Use-dependentb	Use-dependent (current-dependent)d	Use-dependentb	Use-independente
Mechanism of block for Ih in rods	NA	NA	Use-dependentf	Use-dependentg	Use-independenth
Mechanism of block for HCN1 HE	NA	Use-dependentc	Closed channel blocki	Use-dependentc	Open channel blockj
		Use-dependentc
HE HCN1 block reversal by hyperpolarization	NA	NA	NOi	Yesc	Yesj
Visual symptoms	Yesk	NA	Yesl	Yesm	NA
ERG studies	NO	Yesn	Yeso	Yesp	NO

Abbreviations: ERG, Electroretinogram; HCN, Hyperpolarization-activated, Cyclic Nucleotide sensitive; HE, Heterologously expressed; NA, Not available.

^aSnyders and Van Bogaert (1987). Alinidine also blocks other currents.

^bVan Bogaert and Pittors (2003). Zatebradine also blocks other currents.

^cStieber et al. (2006).

^dBucchi et al. (2002).

^eBoSmith et al. (1993).

^fDemontis et al. (2006).

^gSatoh and Yamada (2002).

^hSatoh and Yamada (2000).

ⁱBucchi et al. (2006).

^jShin et al. (2001).

^kCeremuzynski et al. (1987).

^lBorer et al. (2003).

^mFrishman et al. (1995).

ⁿMaccarone et al. (2004).

^oGargini et al. (2006).

^pGargini et al. (1999).

The visual effects, especially those induced by ivabradine, have been extensively investigated in both animal models and humans. Ivabradine is a novel antianginal drug that reduces heart rate by inhibiting the hyperpolarization-activated current expressed in cardiac sino-atrial node cells (I_f) (Di Francesco, 1993, Di Francesco and Camm, 2004). During the non-clinical visual safety program, no toxic damage in any ocular structure has been reported in animal models upon administration of therapeutic doses for humans (EPAR Procoralan, 2005). The visual symptoms induced by ivabradine, reported by patients during the clinical programme, include most commonly phosphenes (14.5% of patients) and less frequently blurred vision. Visula effects appear generally within the first 2 months of treatment and their frequency increases with the dose of ivabradine. Most of these events were reported to occur in conditions of darkness or dim light, when the retinal sensitivity is high. Phosphenes are generally reported to be mild or moderate and to disappear even though treatment continued (77.5% of patients) or after treatment cessation (Borer et al., 2003; Tardiff, 2005, Savelieva and Camm, 2006). The host plausible hypothesis is that ivabradine interacts with the visual system by inhibiting hyperpolarization-activated current in retinal cells (I_h).

An I_h current with properties similar to cardiac (I_f) has been reported in retinal rods (Fain et al., 1978; Owen and Torre, 1983; Bader and Bertrand, 1984; Beech and Barnes, 1989; Demontis et al., 1999, 2002). I_h has also been found in cones (Yagi and MacLeish, 1994), bipolar neurones (Kim et al., 2003; Müller et al., 2003; Cervetto et al., 2005; Cangiano et al., 2006), amacrine (Kozuimi et al., 2004), RGC (Tabata and Ishida, 1996), and it is widely distributed in the cortex, hippocampus and thalamus as well as in peripheral nerves (see Robinson and Siegelbaum, 2003). I_h possesses unusual biophysical properties that allow it to play a multiple role in neuronal excitability. The I_h carrying channel is usually referred to as hyperpolarization-activated, cyclic nucleotide sensitive (HCN) and represents an evolutionary combination between the voltage-gated K⁺ channel and the cyclic nucleotide-gated, non-voltage-gated K⁺ channel. Thus I_h channel possesses a high permeability to K⁺ ions, is voltage gated, but also modulated by intracellular cyclic adenosine monophosphate (cAMP) levels, allowing activity-dependent regulation. More important, the channel has substantial permeability to Na⁺, such that on opening at typical neuronal resting potential, it generates an inward current, causing the cell to depolarize; yet the channel is activated not by depolarization (as with virtually all voltage-gated channels) but by hyperpolarization. Because hyperpolarization produces activation, which in turn leads to depolarization, the HCN channel possesses an inherent negative-feedback property. This negative-feedback principle is evident in the contribution of I_h to neuronal excitability. In a neurone recorded at rest with I_h inactive, a small depolarizing or hyperpolarizing input rapidly produces a steady-state change in voltage. With I_h active, however, a hyperpolarizing input causes slow I_h activation, producing a depolarizing current that returns the membrane potential toward rest. Conversely, a depolarizing input causes deactivation of the I_h that was active at rest; the loss of a tonic depolarizing current causes a hyperpolarization, again returning membrane potential towards rest. Thus I_h tends to stabilize membrane potential near the resting level against either depolarizing or hyperpolarizing inputs. More precisely, I_h diminishes input resistance, thus minimizing the voltage change produced by a given synaptic current (Robinson and Siegelbaum, 2003). In physiological terms, I_h can be either excitatory or inhibitory with respect to its influence on action potential firing. Thus, the HCN embodies two opposing influences on neuronal excitability, preventing simple characterization as either inhibitory or excitatory.

Although all HCN channels possess the fundamental properties described earlier, I_h represents a family of currents with distinct kinetics and tissue distributions. The HCN family of genes, of which four subtypes have been identified (Santoro et al., 1998), encodes four isoforms (HCN1, 2, 3, 4) that when expressed in heterologous cells generate channels with distinct activation kinetics mirroring the properties of native I_h. The predominance of various HCN subtypes varies by location, with HCN1 and HCN2 most prevalent in the retina (Demontis et al., 2002; Cervetto et al., 2005; Gargini et al., 2006). Because the biophysical properties of HCN subtypes vary significantly, the contribution of I_h to neuronal behaviour also varies by both location and neurone type in each region, with individual neurones expressing varying amounts of different HCN isoforms. In addition, because I_h has the potential to affect excitability in a number of ways, modulation of I_h can significantly affect neuronal behaviour. The outcome of I_h activation will therefore depend on: (a) the functional properties of the specific channel subtype, (b) the electrophysiological milieu in which the channel operates (i.e. properties of the additional conductances present on the cell membrane) and (c) the spatial distribution and the nature of the synaptic inputs of the neurone. Recent evidence illustrates two opposite paradigms: in cortical pyramidal neurones I_h seems to take part in the generation of focal paroxysmal activities (Timofeev et al., 2002), by contrast a neurone-stabilizing effect of I_h has been suggested by a study in which I_h was enhanced pharmacologically (Poolos et al., 2002).

In the visual system I_h inhibition has been shown to modify the filtering properties of the retinal processing (Gargini et al., 1999; Mao et al., 2003; Gargini et al., 2006). In addition to the changes in the temporal properties of the light response, other effects that indicate an increased excitability were observed. I_h inhibition by zatebradine has been shown to increase the hyperpolarizing response to light in retinal rods and to induce oscillations during the recovery phase of the membrane potential (Satoh and Yamada, 2002). More recently it has been suggested that I_h inhibits generation of spontaneous action potential by human retinal receptors (Kewai et al., 2005). A current with the properties of I_h has been identified in the amacrine cells of the mouse where it contributes to stabilize the membrane potential (Kozuimi et al., 2004). Patch clamp measurements from retinal cells of mouse show that therapeutic levels of ivabradine are effective in blocking I_h at physiological ranges of membrane potential in rods and that this blockade is activity dependent (Demontis et al., 2006). Ivabradine reversibly reduces the ERG response to periodic stimuli of low temporal frequency, without alterations of morphology, channel distribution and pigment content (Gargini et al., 2006). Under normal circumstances, I_h activation in both the membrane of visual cells (Demontis et al., 1999) and in that of bipolar neurones (Cangiano et al., 2006) causes dampening of responses to slow noisy inputs thus tuning the visual system to perceptually more relevant signals of higher frequency. I_h inhibition, by altering at the rod synapse the filtering out of signals generated by thermal breakdown of rhodopsin or by other fluctuations in the phototransductive cascade and in the spontaneous firing of ganglion cells, is expected to increase the probability of phosphene occurrence. A possible explanation of how I_h inhibition may increase the probability of phosphene occurrence is illustrated in Figure 1. Random fluctuations occurring at the outer segment of visual cells are filtered out by the inner segment (Demontis et al., 1999) and transmitted to the synaptic endings. Further filtering was suggested to occur post-synaptically at the bipolar dendrites (Sampath and Rieke, 2004; Field et al., 2005) where HCN2 channels are in register with the synaptic ribbons and similarly disposed as the glutamate receptors, mGluR6 (Gargini et al., 2006). Suppression of filtering increases the probability of a noisy signal reaching the transmission threshold.

Figure 1.

Example of a hypothetical retinal mechanism of phosphene induction by I_h inhibitors. Retinal rod outer segment (OS). Retinal rod inner segment (IS). Receptor synaptic ending (RS). Rod bipolar cell (RBC).

Impaired signal filtering by I_h inhibition need not be restricted somewhere in between retinal receptors and bipolar neuron dendrites as assumed in this scheme. In principle, there are several other locations within the retina and along the visual pathway that I_h inhibitors may target. Because I_h was also observed in thalamic relay neurons (McCormick and Pape, 1990; Nita et al., 2003), one may argue that phosphenes are possibly caused by I_h inhibition in central neurones. It must be pointed out, however that a central action of ivabradine seems unlikely because the blood–brain barrier is essentially impermeable to this molecule (see Savelieva and Camm, 2006).

Conclusions

It seems reasonable to conclude that a wide variety of causal agents may induce phosphene-like events either by a nonspecific direct stimulation of the visual system or by changing the neuronal excitability or by suppressing noise filtering processes. At variance with many other drugs for which the mechanisms of phosphene induction remain elusive, in the case of I_h inhibitors, experimental data are available to suggest an explanation for the occurrence of the visual phenomena. Recent evidence supports the idea that inhibition of I_h is the most probable cause for the visual symptoms experienced by both healthy volunteers and cardiac patients under ivabradine treatment. Under normal circumstances, most of the random fluctuations generated within the retinal circuits do not reach the level of conscious perception because they are filtered out. Presumably, filtering occurs within the retina at various levels and the ability of I_h to act as a negative-feedback mechanism makes HCN channels likely sites for such actions. It is, however, important to note that not all the findings from patients studies fit easily within the proposed framework. It is baffling to observe that while in nearly all patients ivabradine lowers the heart rate, only a relatively small fraction (∼15%) experiences phosphenes. Somewhat confusing is also the fact that in the majority of cases phosphene occurrence resolve spontaneously during ivabradine treatment, at variance with the slowing of heart rate that seems to persist as long as the drug is given. It is not presently clear whether all this is due to a dose-threshold effect, to a drug target polymorphism or to some other cause. At therapeutic doses for humans ivabradine induces in animal models only small changes in the retinal activity (Gargini et al., 2006). The in vivo minor impact of the drug on the retinal HCN is consistent with the low passage of ivabradine across the blood retinal barrier.

Conflict of interest

The authors have been consultant/scientific advisors for the Institut De Recherches Internationales Servier (IRIS) the company that manifactures Procoralan (ivabradine).