Arousal and the pupil: why diazepam-induced sedation is not accompanied by miosis
- First Online:
- Cite this article as:
- Hou, R.H., Samuels, E.R., Langley, R.W. et al. Psychopharmacology (2007) 195: 41. doi:10.1007/s00213-007-0884-y
- 217 Views
There is a close relationship between arousal and pupil diameter, decrease in the level of arousal being accompanied by constriction of the pupil (miosis), probably reflecting the attenuation of sympathetic outflow as sedation sets in. Paradoxically, sedation induced by benzodiazepines is not accompanied by miosis.
The objective of this study was to examine the hypothesis that diazepam may attenuate both the sympathetic and the opposing parasympathetic outflow to the iris, which may mask the miosis. Dapiprazole (sympatholytic) and tropicamide (parasympatholytic) were applied topically, together with the cold pressor test (CPT), to manipulate the sympathetic/parasympathetic balance.
Materials and methods
Sixteen healthy male volunteers participated in four weekly sessions according to a balanced double-blind protocol. Diazepam 10 mg (two sessions) and placebo (two sessions), associated with either 0.01% tropicamide or 0.5% dapiprazole eyedrops, were administered orally. Pupil diameter, light and darkness reflexes and pupillary sleepiness waves were recorded with infrared video pupillometry, alertness was measured by critical flicker fusion frequency (CFFF) and visual analogue scales (VAS), blood pressure and heart rate by conventional methods. CPT was applied after post-treatment testing. Data were analysed by analysis of variance, with multiple comparisons.
Diazepam caused sedation (reduction in VAS alertness scores and CFFF, increase in sleepiness waves), dapiprazole had a sympatholytic and tropicamide a parasympatholytic effect on the pupil. Diazepam had no effect on pupil diameter and reflexes or their modifications by the antagonists. CPT increased pupil diameter, blood pressure and heart rate, and the increase only in systolic blood pressure was attenuated by diazepam.
Diazepam-induced sedation is not accompanied by any change in either the sympathetic or parasympathetic influence on the iris.
KeywordsDiazepamSedationMiosisDapiprazoleTropicamideCold pressor testPupillary reflexesBlood pressure
It is generally recognised that there is a close relationship between the level of arousal and the function of the pupil: a decrease in the level of arousal is accompanied by constriction of the pupil (miosis; Loewenfeld 1993a) and an increase in spontaneous pupillary oscillations in darkness (sleepiness waves; Wilhelm et al. 1998a). The miosis has been attributed to the weakening of the central sympathetic inhibition of the parasympathetic pupillomotor nucleus (Edinger–Westphal nucleus), and the sleepiness waves to the instability of this central inhibition (Barbur 2004).
The assessment of pupil diameter is routinely used by anaesthetists when gauging the depth of anaesthesia (Aitkenhead et al. 2001). Although most sedative drugs “decrease pupillary diameter in proportion to their sedative–hypnotic effects”, there are exceptions to this general rule. Whilst both the tricyclic antidepressant amitriptyline and the antiparkinsonian drug pramipexole are highly sedative, the sedation evoked by neither drug is accompanied by miosis: Amitriptyline causes little change in pupil diameter (Szabadi et al. 1980; Phillips et al. 2000a), whereas pramipexole actually increases it (Samuels et al. 2006a, b). In the case of both drugs, the dissociation between sedation and miosis can be explained by the interaction of these drugs with different aspects of the pupil control mechanism. Amitriptyline is known to block the reuptake of noradrenaline and muscarinic cholinoceptors in the iris: These two effects are likely to lead to pupil dilatation, which can mask the miosis resulting from central sedation and peripheral α1-adrenoceptor blockade (Szabadi and Bradshaw 1986). In the case of pramipexole, the pupil dilatation is likely to be the result of the central parasympatholytic effect of the drug which may supersede the miosis resulting from central sedation (Samuels et al. 2006a, b). Finally, although the sedative first generation antihistamine diphenhydramine causes miosis, this miosis is relatively weak, as it is likely to be attenuated by the mydriasis resulting from the anticholinergic property of the drug (Hou et al. 2006b, 2007).
The benzodiazepines are also highly sedative drugs which, again, fail to cause miosis. In fact, it is well documented that benzodiazepine-induced sedation is not accompanied by any change in pupil diameter (Karnoil et al. 1976; Safran 1984; Walser et al. 1987; Loewenfeld 1993a; Bitsios et al. 1998; Huron et al. 2002; Hou et al. 2006b). Interestingly, according to one report, whilst sedation induced by the benzodiazepine lorazepam was not accompanied by miosis, it was associated with an increase in pupillary sleepiness waves (Huron et al. 2002). A possible explanation for the absence of miosis in benzodiazepine-induced sedation is that these drugs may attenuate the opposing sympathetic and parasympathetic inputs to the iris in equal measure, and thus the miosis due to sympathetic inhibition may be masked by mydriasis resulting from parasympathetic inhibition. As benzodiazepines have little action in the periphery and almost all their effects are due to the potentiation of GABAergic inhibition in the central nervous system (Charney et al. 2001), a possible modification of the sympathetic and/or parasympathetic influence on the iris is likely to be of central origin.
In this respect, it is of interest that there have been some early studies investigating the effects of the benzodiazepine diazepam on the sympathetic and parasympathetic innervations of the iris. These studies by E.B. Sigg and his colleagues (Sigg and Sigg 1969; Sigg et al. 1971; Sigg and Sigg 1973) were carried out in anaesthetised cats, and thus no inferences can be made about the relationship between sedation and pupillary function. Diazepam had no effect on either baseline sympathetic outflow to the iris or the increase in sympathetic activity and pupil diameter evoked by hypothalamic stimulation (Sigg and Sigg 1969; Sigg et al. 1971). On the other hand, diazepam reduced the size of the pupillary light reflex response, suggesting a central parasympatholytic effect (Sigg and Sigg 1973).
Materials and methods
Sixteen healthy drug-free male volunteers aged 21.0 ± 0.6 years (mean ± SEM) and weighing 72.0 ± 2.1 kg (mean ± SEM) were recruited for this experiment. Subjects with a history of eye disease or any abnormal finding on ocular examination (funduscopy, visual inspection of the anterior chamber) were excluded. All subjects gave their written informed consent after a verbal explanation of the study and after reading a detailed information sheet. Each subject completed a brief medical history and underwent a physical examination, including a full neurological status, before inclusion in the study. All subjects denied any history of illicit drug use or excessive consumption of alcohol in general and agreed to remain free of prescribed illicit drugs during the period of the study. Furthermore, they agreed not to drink any alcohol during the 48 h before each visit. The caffeine intake was less than six cups of coffee or tea per day in general. The subjects were required not to consume any caffeine-containing drinks during the 24 h before each visit and during the period of the study. All subjects were non-smokers or light smokers (i.e. less than five cigarettes per day). All subjects reported compliance with these requirements on each test day and stated that they were well rested at the start of each experimental session, having slept at least 7 h during the previous night. The study protocol was approved by the University of Nottingham Medical School Ethics Committee.
Treatment combinations administered in the four experimental sessions
Each subject participated in four experimental sessions at weekly intervals (returning to the laboratory at the same time each week for each session). Subjects were allocated to treatments and sessions according to a double-blind, balanced crossover design.
Tests and apparatus
Static pupillometry, with a binocular infrared video pupillometer (Procyon, London, UK) was used to obtain resting pupil diameter readings under four luminance levels (darkness, 6, 91 and 360 cd m−2) using a calibrated internal light source within the pupillometer. Measurements were carried out in a light-sealed darkened room. Pupil diameter was first recorded in darkness and then under each of the increasing luminance levels for 2 s at 4 Hz and stored to disk for off-line analysis (Phillips et al. 2000b). Measurements were repeated four times at each level of luminance, and mean of the four measurements was used for future analysis.
Dynamic pupillometry, with an infrared binocular television pupillometer (TVC 1015B Applied Science Laboratories, Waltham, MA, USA) was used to record light and darkness reflex responses. The sampling rate of the pupillometer was 60 Hz, and the detection accuracy was better than 0.05 mm.
Light reflex responses were evoked in the dark with four light flashes (green, 565 nm peak wavelength) of 200-ms duration and of incremental luminance (measured in the plane of cornea): 5.2, 41, 320 and 2,050 cd m−2. The light flashes were delivered at 25-s intervals via a light emitting diode positioned 1 cm from the cornea of the subject’s right eye, providing “full face” light stimulation (Bitsios et al. 2004). Four measurements were taken at each stimulus intensity and were averaged for future analysis. The amplitude of light reflex response (i.e. the difference between the initial and the minimal diameters of a pupillary response to a light flash, mm) was measured (Bitsios et al. 1999b, 2004).
Darkness reflex responses were evoked in the following way. Pupil diameter was continuously recorded during alternating periods of darkness and bright ambient illumination. After a period of dark adaptation, an illuminated screen positioned 1.5 m in front of the subject’s eyes was switched on for 10 s. The luminance of the screen was 1,370 cd m−2. The illumination of the screen was switched to darkness for 20 s (which was a sufficient length of time for the dilatory response to darkness to reach a plateau). The cycle was repeated, and the means of the parameters of two darkness responses were calculated. The parameters of the darkness response were initial velocity and amplitude. Initial velocity (mm s−1) was defined in terms of the time required to obtain 25% of the maximum response measured from the onset of the response. Amplitude was defined as the difference between the dark- and light-adapted pupil diameters (Bitsios et al. 1996).
Spontaneous pupillaryfluctuations in darkness were measured by a monocular infrared television pupillometer [pupillographic sleepiness test (PST); setup version 1.20: AMTech, Weinheim, Germany] over 11 min at a frequency of 25 Hz (for details, see Ludtke et al. 1998; Phillips et al. 2000a). Software-driven artefact rejection was applied to the raw data to remove blinks, and high frequencies of ≤0.8 Hz were captured and subjected to fast Fourier transform. Absolute changes in pupil diameter were also recorded and subjected to low pass filtering. This was achieved by sampling for 0.64-s periods (16 data points at 25 Hz); the cumulative difference between successive 0.64-s samples over 1 min yields the pupillary unrest index PUI (mm min−1). For further details, see Lüdtke et al. (1998). Mean pupil diameter obtained during the 11- min recording period was also computed.
Non-pupillary measures of alertness
The level of arousal was assessed using critical flicker fusion frequency (CFFF) which is defined as the frequency at which a flickering light appears to be continuous (Smith and Misiak 1976). The Leeds psychomotor tester (Psychopharma, Surrey, UK) was used to collect measurements of the threshold. Recordings were taken in a room with constant illumination: The luminance in the vicinity of the equipment was 12 cd m−2 (recorded by a Spectra Spotmeter PR-1500; Photo Research, Chatsworth, CA, USA); the constancy of the luminance was checked at weekly intervals. The subjects were asked to view with both eyes an assembly of four red light emitting diodes. Eight measurements of the threshold were taken, four with increasing frequencies and four with decreasing frequencies. The mean of the eight measurements was taken as the CFFF for each testing session (Abduljawad et al. 1997).
Subjects rated their subjective state of mood using a computerised version of the visual analogue scales (VAS) developed by Norris (1971). The ratings on the 16 scales were grouped under the headings of “alertness” (alert/drowsy; strong/feeble; muzzy/clear headed; well coordinated/clumsy; lethargic/energetic; mentally slow/quick witted; attentive/dreamy; incompetent/proficient), “calmness” (calm/excited; tense/relaxed) and “contentedness” (contented/discontented; troubled/tranquil; happy/sad; withdrawn/gregarious) based on a factor analysis carried out by Bond and Lader (1974).
Non-pupillary autonomic functions
Blood pressure and heart rate recordings were taken in the sitting position using an electroaneroid sphygmomanometer. Salivation was measured by placing three cotton wool dental rolls in the mouth (two buccally and one sublingually) and recording the increase in their weights (Peck 1959; Arya et al. 1997) over a 1-min period (Szabadi and Tavernor 1999).
Cold pressor test
Cold pressor test was used to elicit sympathetic activation (Tavernor et al. 2000). The test consisted of immersing the left hand up to the wrist crease in a bucket containing crushed ice and water (4°C) for a period of 90 s and recording sympathetically mediated responses (rise in blood pressure and heart rate, pupillary dilatation).
A two-way analysis of variance (ANOVA) was carried out to examine whether there was any difference between the effects of two topical placebo conditions (Pp1 and Pp2; Dp1 and Dp2) at each of the three testing occasions [pre-treatment, post-treatment(1) and post-treatment(2)] on static pupil diameter, light reflex amplitude, darkness reflex amplitude and initial velocity. As no significant treatment effect was revealed in any of these comparisons [F(1,15) < 2, NS], the means of the measures obtained in the two topical placebo conditions, designated as Pp and Dp, were used in all further analysis.
In the case of static pupillometry, absolute pupil diameter and anisocoria (the difference between the diameters of the left and right pupils) was calculated. Anisocoria data, based on using the untreated eye as a control, were included as a reliable index of the effectiveness of topically applied mydriatic or miotic drugs (Kardon 1998). The absolute pupil diameter and the anisocoria data on three testing occasions [pre-treatment, post-treatment(1) and post-treatment(2)] were analysed separately using three-way ANOVA, with luminance, systemic and topical treatment being factors. Effects of the cold pressor test on the untreated, tropicamide-treated and dapiprazole-treated pupil were analysed separately using three-way ANOVA, with testing occasion, systemic and topical treatment being factors. In the case of the light reflex response, the absolute value of amplitude obtained after the three topical treatments (tropicamide [left eye], dapiprazole [left eye], artificial tear [right eye]) was measured, and the difference between the amplitudes obtained in the left and the right pupils (between-eye difference of amplitudes) was calculated. The “between eye difference of amplitude”, using the consensual light reflex response in the untreated eye as a control, was included as an index of the effectiveness of the topically applied mydriatic or miotic drug in modifying light reflex response amplitude (Hou et al. 2007) on the basis that the consensual light reflex response is approximately of the same size as the direct one (Loewenfeld 1993b, pp 238). Absolute amplitude values and between-eye differences of amplitudes, obtained on two testing occasions [pre- and post-treatment (1)], were subjected to three-way and two-way ANOVA, respectively, the factors being luminance, systemic and topical treatment. The parameters (amplitude, velocity) of the darkness reflex response were subjected to two-way ANOVA, the factors being systemic and topical treatment. When a significant effect was found, multiple comparisons were made between different treatments.
In the case of simple measures such as PST (and also non-pupillary measures of alertness and autonomic functions, see below) the pre/post-treatment changes were taken, and paired t tests were used to compare the effects of placebo and diazepam.
Non-pupillary measures of alertness and autonomic functions
Paired t tests were used to compare the effects of placebo and diazepam on pre/post-treatment(1) changes in the case of non-pupillary measures of alertness (CFFF and VAS) and autonomic functions (systolic and diastolic blood pressure, heart rate, salivation) and also on post-treatment(1)/post-treatment(2) changes in the case of cardiovascular parameters (systolic, diastolic blood pressure, heart rate).
Measures of alertness
Pupil diameter: effect of treatments
The pupil diameter data are shown in Fig. 5, and the results of the analysis of variance in Table 2. On the pre-treatment occasion, there was a significant effect of luminance on pupil diameter, pupil diameter decreasing as a function of luminance. There was no anisocoria in darkness or at any level of luminance. Pre-treatment pupil diameter did not differ in the four experimental sessions associated with different treatments. On the post-treatment(1) occasion, comparison of each active topical treatment condition in darkness and at each luminance level with the topical placebo condition showed that the miotic effect of dapiprazole and the mydriatic effect of tropicamide were significant in darkness and at the three luminance levels (P < 0.05 in all cases). The effects of the topical drugs were also reflected in a considerable degree of anisocoria, tropicamide causing a change (left minus right pupil) in the positive direction and dapiprazole in the negative direction. Comparison with anisocoria in the pre-treatment condition showed that the anisocoria evoked by each topical drug was significant at each luminance level (P < 0.05). On the post-treatment(2) occasion, i.e., after the application of the cold pressor test, comparison of each active topical treatment condition in darkness and at each luminance level with the topical placebo condition showed that the miotic effect of dapiprazole and the mydriatic effect of tropicamide were significant in darkness and at the three luminance levels (P < 0.05 in all cases). The anisocoria evoked by tropicamide was significant at 6, 91 and 360 cd m−2, but not in darkness, and that evoked by dapiprazole was significant in darkness and at each level of luminance (P < 0.05). Systemic treatment had no effect on pupil diameter in the untreated eye and failed to modify the changes in pupil diameter evoked by the topical drugs.
Pupil diameter: effect of cold pressor test
The effect of the cold pressor test on pupil diameter was also analysed by comparing the post-treatment(1) and post-treatment(2) data (see Fig. 6 and Table 3) In the case of placebo-treated pupil diameter (right eye), comparison of the effects of the two testing occasions revealed that the cold pressor test significantly increased pupil diameter in darkness and at all three levels of luminance (P < 0.05 in each case). In the case of tropicamide-treated pupil diameter, comparison of the effects of the two testing occasions revealed that the cold pressor test significantly increased pupil diameter in the tropicamide-treated eye in darkness and at each of the three luminance levels (P < 0.05). However, there was no significant effect of the cold pressor test on the degree of anisocoria, indicating that the cold pressor test dilated both the untreated and the tropicamide-treated pupils in equal measures. In the case of the dapiprazole-treated pupil diameter, comparison of the two testing occasions revealed that the cold pressor test significantly increased pupil diameter in the dapiprazole-treated eye in darkness and at the 6 and 91 cd m−2 luminance levels (P < 0.05). The cold pressor test significantly increased the anisocoria at 6, 91 and 360 cd m−2, reflecting the attenuation of dapiprazole-induced miosis by the cold pressor test. Systemic treatment failed to modify the effect of the cold pressor test on pupil diameter in any of the three topical treatment (artificial tear, tropicamide, dapiprazole) conditions.
Light reflex response
The light reflex response amplitude and the between-eye difference of amplitudes (see “Materials and methods”) are shown in Fig. 7 and Table 4. Multiple comparisons revealed that the post-treatment topical treatment effect was due to a reduction in light reflex amplitude caused by tropicamide at each light intensity both when absolute amplitude and the between-eye difference of amplitudes were analysed (P < 0.05). There was no effect of systemic treatment on light reflex amplitude either after the topical application of placebo or tropicamide or dapiprazole.
Darkness reflex response
The darkness reflex amplitude on the post-treatment(1) testing occasion is shown in Fig. 7. Pre-treatment there were no significant effects of either topical or systemic treatment, whereas post-treatment there was a significant effect of topical treatment [F(2,30) = 37.56, P < 0.001], but not of systemic treatment [F(1,15) = 1.85, NS]. Multiple comparison showed that the topical treatment effect was due to a reduction in darkness reflex amplitude caused by dapiprazole (P < 0.05).
The initial velocity of the darkness reflex on the post-treatment(1) testing occasion is shown in Fig. 7. There was no significant effect of either systemic or topical treatment on initial velocity (in all cases, F < 1).
Effects of treatments on pupil diameter: results of analysis of variance
Luminance (TT × L)
Topical treatment (ST)
Interaction treatment (TT)
F(3,45) = 325.18 P < 0.001
F < 1 NS
F < 1 NS
F < 1 NS
F < 1 NS
F < 1 NS
F(3,45) = 290.14 P < 0.001
F < 1 NS
F(2,30) = 125.12 P = 0.001
F(6,90) = 16.30 P < 0.001
F(3,45) = 54.93 P < 0.001
F < 1 NS
F(1,15) = 125.12 P < 0.001
F(3,45) = 18.89 P < 0.001
F(3,45) = 223.36 P < 0.001
F < 1 NS
F(2,30) = 110.97 P < 0.001
F(6,90) = 6.84 P < 0.001
F(3,45) = 5.68 P = 0.01
F < 1 NS
F(1,15) = 195.63 P < 0.001
F(3,45) = 7.09 P < 0.01
Effects of the cold pressor test on pupil diameter: results of analysis of variance
Luminance (TO × L)
Testing treatment (ST)
F(3,45) = 304.21 P < 0.001
F < 1 NS
F(1,15) = 140.00 P < 0.001
F(3,45) = 9.41 P < 0.001
F(3,45) = 229.00 P < 0.001
F < 1 NS
F(1,15) = 53.00 P < 0.001
F(3,45) = 10.19 P < 0.001
F < 1 NS
F < 1 NS
F < 1 NS
F(3,45) = 203.10 P < 0.001
F < 1 NS
F(1,15) = 75.68 P < 0.001
F(3,45) = 5.17 P < 0.01
F(3,45) = 28.46 P < 0.001
F < 1 NS
F(1,15) = 66.95 P < 0.001
F(3,45) = 18.47P < 0.001
Effects of treatments on light reflex response amplitude: results of analysis of variance
F(3,45) = 154.71 P < 0.001
F < 1 NS
F < 1 NS
F < 1 NS
F < 1 NS
F < 1 NS
F(3,45) = 189.71 P < 0.001
F < 1 NS
F(2,30) = 35.22 P < 0.001
F < 1 NS
F < 1 NS
F(1,15) = 52.17 P < 0.001
Non-pupillary autonomic functions
Figure 8 shows the effects of the treatments [pre-/post-treatment(1) change] and modification of the effect of treatment by the cold pressor test[post-treatment(1)/post-treatment(2) change] on systolic and diastolic blood pressure and heart rate. In the post-treatment(1) condition, diazepam had no effect on either measure. The cold pressor test caused a significant increase in all three measures both after treatment with placebo and after treatment with diazepam (P < 0.05 in all cases). In the post-treatment(2) condition, diazepam reduced the cold pressor test-evoked increase in systolic blood pressure (P < 0.05), but had no effect on the cold pressor test induced increases in diastolic blood pressure and heart rate (P > 0.05 in both cases).
The pre-/post-treatment(1) change in salivary output was not affected by diazepam (P > 0.05). The pre-/post-treatment(1) changes in salivary output (mean ± SEM) were placebo (−0.02 ± 0.01) and diazepam (−0.17 ± 0.09); there was no significant difference between placebo and diazepam (P > 0.05).
Effects of diazepam on level of alertness
The single dose of diazepam (10 mg) had robust sedative effects as shown by reductions in subjectively related alertness and CFFF (Fig. 4). As it is known that CFFF is dependent upon pupil diameter (Landis 1954), the reduction in CFFF induced by a sedative drug may be contaminated by miosis. However, as diazepam did not cause any change in pupil diameter in the present experiment (see below), it is unlikely that such contamination played any role in the CFFF data obtained. Diazepam also increased the indices of sleepiness (pupillary unrest index, total power of pupillary fluctuations) in the PST (Fig. 3).This observation is in agreement with numerous reports demonstrating the sedative effects of single doses of diazepam on subjectively rated alertness (Mattila 1988; Mattila et al. 1993; Bitsios et al. 1998, 1999a; Abduljawad et al. 2001; Scaife et al. 2005; Hou et al. 2006b), on CFFF (Mattila 1988; Mattila et al. 1993; Scaife et al. 2005; Hou et al. 2006b) and a number of tests of psychomotor performance (Mattila 1988; Mattila et al. 1993). However, this is the first report of the sedative effect of diazepam on the parameters of pupillary sleepiness waves which were increased by diazepam, similar to other sedative drugs (clonidine: Phillips et al. 2000c; Hou et al. 2005, amitriptyline Phillips et al. 2000a, diphenhydramine: Szabadi et al. 2006; Hou et al. 2006b; Hou et al. 2007). This property of diazepam is likely to be shared by other benzodiazepines, as it has been reported that lorazepam also increases pupillary sleepiness waves (Huron et al. 2002).
Effects of diazepam on pupil diameter modified by tropicamide and dapiprazole
As reported previously (Bitsios et al. 1998; Phillips et al. 2000b; Hou et al. 2005; Samuels et al. 2006a), there was a near-linear negative relationship between the logarithm of the level of ambient luminance and resting pupil diameter. The luminance level/pupil diameter curves were displaced upwards (pupil dilatation) by tropicamide and downwards (pupil constriction) by dapiprazole. The mydriatic effect of tropicamide and the miotic effect of dapiprazole were also reflected in the degrees of anisocoria between the treated (left) and untreated (right) pupils. The relationship between luminance level and pupil diameter remained largely intact after the administration of the eye drops. This is reflected in the relative constancy of the degree of anisocoria across the luminance range studied, with the exception of some decrease in the dapiprazole-evoked anisocoria at the two higher luminance levels, probably reflecting a “floor effect” (Szabadi 1977). The mydriatic effect of tropicamide (Scinto et al. 1994; Steinhauer et al. 2004; Hou et al. 2006a) and the miotic effect of dapiprazole (Steinhauer et al. 2004; Giakoumaki et al. 2005) are well documented in the literature. Although Steinhauer et al. (2004) measured responses to tropicamide and dapiprazole both in the dark and in the light, with the exception of one study (Hou et al. 2006a), there have been no other investigations of the effects of increasing luminance levels on the size of the responses to the topical drugs.
Diazepam had no effect on pupil diameter in darkness, as also indicated by the average pupil diameter obtained in the PST, or at any of the three luminance levels, either in the untreated right or in the left eye after treatment with tropicamide or dapiprazole. On the basis of our hypothesis, we predicted that diazepam would enhance both tropicamide-induced mydriasis and dapiprazole-induced miosis, as the relative impact of a putative reduction in both the parasympathetic and sympathetic outflows is expected to be greater on the parasympathetic contribution to pupil diameter after it has been weakened by tropicamide and on the sympathetic contribution after its attenuation by dapiprazole. However, obviously, this was not the case. It is unlikely that this was due to a flawed pharmacological model, as previous workers have used tropicamide (Steinhauer et al. 2004) and dapiprazole (Steinhauer et al. 2004; Giakoumaki et al. 2005) successfully to study the impact of the sympathetic and parasympathetic outflows on the iris in relative isolation. Therefore, we have to conclude that the inability of the eyedrops to unmask the postulated effect of diazepam on the central pupil control system is likely to be due to the fact that neither the sympathetic nor the parasympathetic outflow to the iris is affected by benzodiazepines. The lack of effect of diazepam on sympathetic outflow to the iris has been demonstrated by Sigg and his colleagues in experimental animals (Sigg and Sigg 1969; Sigg et al. 1971). It should be noted, however, that the same group of investigators have found some evidence consistent with a reduction in parasympathetic outflow in response to diazepam indicated by a reduction in light reflex responses recorded from the short ciliary nerves in anaesthetised animals (Sigg and Sigg 1973). However, our experiment could not confirm the operation of a similar mechanism in human subjects, as diazepam failed to cause mydriasis or influence light reflex responses (see below), as would be expected after the attenuation of parasympathetic outflow.
The lack of effect of diazepam on the central pupil control mechanism is an unexpected finding, as benzodiazepines are known to exert their action by enhancing the effectiveness of inhibitory GABAA receptors, and these receptors are very widely distributed in the CNS (Pirker et al. 2000; Mohler et al. 2002). Indeed, it has been shown that GABA receptors are associated with pre-autonomic neurones both in the paraventricular nucleus (PVN) of the hypothalamus (Kalsbeek et al. 2000; Haywood et al. 2001; Han et al. 2002; Uschakov et al. 2006) and the locus coeruleus (LC; Kaur et al. 1997; Nitz and Siegel 1997; Chen et al. 1999), nuclei known to project to the pre-ganglionic sympathetic neurones innervating the iris and pre-ganglionic parasympathetic neurones in the EWN, the source of the parasympathetic output to the iris (Loewy 1990a; Szabadi and Bradshaw 1996). It is of interest that it has been reported that benzodiazepines may be unable to modulate GABAergic neurotransmission at some sites in the brain stem, either due to the absence of GABA type A signalling, as shown in medial habenular neurones (Wang et al. 2006), or due to insensitivity of the GABAA receptors to diazepam, as demonstrated in the LC (Chen et al. 1999). It is an intriguing possibility that as in the LC, GABAA receptors in the EWN may also be insensitive to benzodiazepines. Therefore, the modulation of GABAA receptor activity by diazepam in the LC is unlikely to contribute to the sedative effect of the drug or underlie the increase in pupillary sleepiness waves caused by the drug.
As mentioned above (see “Introduction”), although many sedative drugs cause miosis, this is far from a universal feature of this class of drug, as there are a number of exceptions (e.g. tricyclic antidepressants, pramipexole) to this general rule. In all these cases, it has been shown that the sedative drug in question has a direct action at some sites of the pupil control mechanism. In these cases, it can be assumed that the translation of central sedation into a change in pupil diameter was masked by some effect of the drug beyond the “arousal/pupil control interface”. This interface is likely to be the noradrenergic LC which is known to occupy a prominent position in the sleep/arousal neuronal circuitry (Nelson et al. 2002; Hou et al. 2005) and also in the autonomic control of the pupil (Szabadi and Bradshaw 1996; Hou et al. 2005). Indeed, it has been reported that fluctuations in the firing rates of LC neurones are paralleled by fluctuations in pupil diameter, increases in firing rates corresponding to increases in pupil diameter (Aston-Jones and Cohen 2005). Therefore, it is an intriguing possibility that fluctuations in pupil diameter associated with sleepiness (pupillary sleepiness waves) reflect fluctuations in LC activity due to instability between wakefulness-promoting and sleep-promoting inputs to the LC. Interestingly, all the sedative drugs studied by us caused an increase in pupillary fatigue waves, as detected by the PST, including drugs which cause miosis (e.g. clonidine: Phillips et al. 2000b; Hou et al. 2005), those which do not (e.g. amitriptyline: Szabadi et al. 1980; Phillips et al. 2000a) and even those causing mydriasis (e.g. pramipexole: Samuels et al. 2006a, b). Interestingly, diazepam also seems to share this property: In the present experiment, whilst failing to alter pupil diameter, it evoked an increase in pupillary sleepiness waves.
Apart from destabilising the relationship between wakefulness- and sleep-promoting influences on the LC, reflected in an increase in pupillary sleepiness waves, sedative drugs are also expected to induce an overall decrease in LC activity, leading to a reduction in sympathetic outflow and miosis. Indeed, it has been reported that benzodiazepines reduce the firing rate of LC neurones (Laurent et al. 1983; Sanghera and German 1983; Simson and Weiss 1989). Paradoxically, however, this reduction in firing rate is not translated into a sympatholytic effect, as shown in the present experiment (see lack of effect of diazepam on pupil diameter after tropicamide treatment, above, and on parameters of the darkness reflex response, below). Therefore, we have to assume that the reduction in LC activity is “intercepted” downstream from the LC, most likely at the level of the spinal cord where preganglionic sympathetic neurones integrate descending inputs not only from the LC but also from a number of other brain stem nuclei and also from an intricate network of interneurones (Gilbey 1997). We suggest that a potential reduction in preganglionic sympathetic neuronal activity, resulting from a reduction in LC activity due to sedation, may be masked by a potential increase in preganglionic sympathetic neuronal activity due to the disinhibition of these neurones by diazepam. The disinhibition of the preganglionic sympathetic neurones may result from the potentiation of the inhibitory action of GABA either on inhibitory bulbospinal neurones (Dembowsky et al. 1989) or on excitatory bulbospinal neurones mediating inhibition via inhibitory interneurones (Gilbey 1997) or on inhibitory interneurones themselves. Therefore, although a reduction in the sympathoexcitatory effect of the LC on the pupil may be masked by an effect of diazepam downstream from the LC, the ability of the LC to transmit the instability between wakefulness-promoting and sleep-promoting inputs to the LC, as sleepiness sets in, is likely to remain unimpaired.
Effects of diazepam on pupillary reflexes
In agreement with numerous previous reports (Flett et al. 1992; Leung et al. 1992; Theofilopoulos et al. 1995; Bitsios et al. 1996, 1999b; Prettyman et al. 1997), there was a near-linear positive relationship between the logarithm of intensity of the light stimulus and the amplitude of the response (Fig. 7). After the administration of tropicamide, the stimulus/response curve was displaced downwards, consistent with a reduction in light reflex amplitude by the drug, whilst the curve obtained after dapiprazole did not differ significantly from that obtained after placebo. Tropicamide affected the amplitude only of the direct response: evoked from the stimulated left eye: The sparing of the consensual response is reflected in a significant difference between the amplitudes of the direct and consensual responses (between eye difference of amplitude), the direct responses being attenuated compared to the consensual responses at each luminance level. The antagonism of the light reflex response by tropicamide is in agreement with the results of previous experiments (Loewenfeld 1993a; Hou et al. 2006a) and consistent with the view that the amplitude of the light reflex response is largely controlled by parasympathetic (cholinergic) activity (Smith 1992; Loewenfeld 1993b). The predominant role of the cholinergic input to the iris in determining light reflex amplitude is further strengthened by our observation that the same dosage of the α1-adrenoceptor antagonist dapiprazole which effectively reduced pupil diameter was without any effect on the amplitude of the light reflex response.
As light reflex amplitude is selectively controlled by parasympathetic activity (see above), it was expected that if diazepam had any effect on central parasympathetic activity, which may be masked by an effect on sympathetic activity, this could be revealed by analysing light reflex responses. However, diazepam failed to modify light reflex amplitude either in the untreated eye or after tropicamide treatment when the influence of the parasympathetic activity had been attenuated. Our observation of the lack of effect of diazepam on pupillary light reflex responses is in agreement with a previous report in humans (Bitsios et al. 1998) and at variance with the result of an early study in cats showing that diazepam reduced the size of the pupillary light reflex response, suggesting a central parasympatholytic effect (Sigg and Sigg 1973). It is not clear whether these discrepant findings reflect a genuine species difference or differences in experimental procedure (i.e. recording changes in pupil diameter in unanaesthesized human subjects after the ingestion of a single dose of 10 mg diazepam vs recording from the short ciliary nerves, without monitoring pupil diameter, in anaesthetised cats after intravenous injection of 0.1–3 mg/kg diazepam). Indeed, a central parasympatholytic effect by diazepam would be expected to result in mydriasis and the attenuation of the light reflex response which is not the case when diazepam is administered to human subjects (Karnoil et al. 1976; Safran 1984; Walser et al. 1987; Loewenfeld 1993a; Bitsios et al. 1998; Hou et al. 2006b; see also present results).
Two parameters of the darkness reflex response, amplitude and initial velocity were also recorded (Fig. 7), which are regarded as relatively pure indices of the sympathetic influence on the pupil (Smith 1992; Loewenfeld 1993c). The amplitude of the darkness reflex response was reduced by dapiprazole consistent with the mediation of this response parameter by α1-adrenoceptors responsible for transmitting the effect of sympathetic activation to the iris. Tropicamide had no effect on the parameters of the darkness reflex response, indicating that the parasympathetic innervation plays little role in determining the size of these parameters. It should be noted, however, that in a previous study, we found that tropicamide caused a slight increase in both the amplitude and initial velocity of the darkness reflex response (Hou et al. 2006a), suggesting that the parasympathetically mediated tone of the pupil sphincter muscle may exert an attenuating influence on the manifestation of the sympathetically mediated darkness reflex response in its full size. Diazepam failed to influence the parameters of the darkness reflex response either in the untreated eye or after the administration of tropicamide or dapiprazole. This observation argues against an inhibitory influence of diazepam on sympathetic outflow to the iris and is consistent with the lack of effect of diazepam on pupil diameter.
Effects of diazepam on pupil diameter modified by the cold pressor test
It was predicted, on the basis of our hypothesis, that a pupillary effect of diazepam may be revealed by altering the relationship between the functionally antagonistic sympathetic and parasympathetic outflows to the iris. Therefore, the cold pressor test was used to increase the influence of the sympathetic outflow on pupil diameter without affecting the parasympathetic influence. Plunging one hand in ice-cold water is a powerful noxious stimulus which is known to evoke increases in blood pressure, heart rate and pupil diameter, consistent with general sympathetic activation (Lovallo 1975; Stratton et al. 1983; Marriott et al. 1990; Seals 1990; Yarnitsky and Ochoa 1990; Tassorelli et al. 1995). We examined the effects of diazepam on the pressor, cardio-acceleratory and mydriatic responses evoked by the cold pressor test, predicting that a sympatholytic effect of diazepam may manifest as antagonism of these responses. This prediction was strengthened by previous reports that the sympatholytic effect of diazepam could be detected only after sympathetic activity had been increased. Thus, diazepam was shown to be without any effect on baseline blood pressure in experimental animals, whereas it effectively antagonised the pressor response evoked by hypothalamic stimulation (Sigg and Sigg 1969; Sigg et al. 1971). Similarly, it was reported that diazepam did not have any effect on baseline sweat gland activity in human subjects, while it was effective in antagonizing the increase in sweat gland activity evoked by a heat stressor (Banjar et al. 1987) or noxious (electric) stimulation (Scaife et al. 2005).
The effect of the cold pressor test on pupillary parameters is illustrated in Figs. 5 (post-treatment 2) and 6. Figure 5 shows that the ambient luminance/pupil diameter relationship was qualitatively the same after the application of the cold pressor test (post-treatment 2) as before it (post-treatment 1), with the exception that somewhat higher absolute pupil diameter values were obtained, both in the untreated right eye and after the prior modification of pupil diameter by tropicamide or dapiprazole in the left eye. The mydriatic effect of tropicamide and the miotic effect of dapiprazole were retained after the application of the cold pressor test as shown by the anisocoria recorded, whilst the size of the miotic effect of dapiprazole was enhanced by the cold pressor test (Fig. 6). The potentiation of dapiprazole-evoked miosis by the cold pressor test is likely to be due to the attenuation of the cold pressor test-evoked mydriasis by the local sympatholytic effect of the drug. Our results are in agreement with previous reports that the cold pressor test evokes a mydriatic response in human subjects (Tassorelli et al. 1995; Tavernor et al. 2000). Furthermore, Tassorelli et al. (1995) found that the cold pressor test-evoked mydriasis was attenuated by the topical application of thymoxamine, an α1-adrenoceptor antagonist, consistent with our observation using dapiprazole. It is clear from Figs. 5 and 6 and Table 3 that there was no difference between the values recorded in the placebo and diazepam conditions indicating that the cold pressor test, whilst increasing pupil diameter, failed to unmask any “hidden” effect of diazepam. The lack of effect of diazepam on the mydriatic response to the cold pressor test is in agreement with previous reports from our laboratory showing that the mydriatic response to the threat of an electric shock in a laboratory model of human anxiety is resistant to antagonism by diazepam, in contrast with the susceptibility of another pupillary effect of anxiety, the attenuation of the pupillary light reflex response (Bitsios et al. 1998).
The effects of diazepam on cardiovascular functions, both at baseline and modified by the cold pressor test are summarised in Fig. 8. The cold pressor test increased systolic and diastolic blood pressure and heart rate, consistent with numerous previous reports (Lovallo 1975; Cummings et al. 1983; Allen et al. 1992; Grosse et al. 1993; Tavernor et al. 2000). Diazepam had no effect on cardiovascular functions at baseline; however, it significantly reduced the cold pressor test-induced increase in systolic blood pressure. As it is generally accepted that the pressor response to the noxious cold stimulus is due to sympathetic activation (Lovallo 1975; Marriott et al. 1990; Seals 1990), our result indicates a sympatholytic effect of diazepam. Although there is little information in the literature on the effects of diazepam on cardiovascular changes evoked by the cold pressor test, it is of interest that it has been reported that diazepam attenuates the increase in cardiac output evoked by the cold pressor test in a subgroup of hypertensive patients (Murakami et al. 1980). This observation is also consistent with a central sympatholytic effect of the drug, although it should be noted that the authors attributed it rather to the parasympathetic modulation of cardiac activity.
Our results with the cold pressor test demonstrate a dissociation between the pupillary and cardiovascular effects of diazepam: Whilst the pressor response was susceptible to the drug, the mydriatic response was resistant to it. It is of interest that this observation is congruent with earlier reports of Sigg and his colleagues (Sigg and Sigg 1969; Sigg et al. 1971) who described a similar dissociation: Whilst the pressor response to hypothalamic stimulation was reduced by diazepam, the mydriatic response was impervious to it. These authors concluded that the sympathetic control of different organs is not homogeneous, different sections showing different pharmacological sensitivities.
The mode of sympathetic stimulation may be also relevant for the observation of a central sympatholytic effect of diazepam. Thus, whilst it has been reported that an increase in sweat gland activity evoked by a heat stressor (Banjar et al. 1987) or noxious (electric) stimulation (Scaife et al. 2005) is amenable to antagonism by diazepam, the increase in skin conductance evoked by the threat of an electric shock is resistant to it (Scaife et al. 2005), probably reflecting the involvement of different neuronal networks upstream from the sympathetic output system.
The differential sensitivities of sympathetic responses to diazepam may reflect the distribution of benzodiazepine-sensitive and benzodiazepine-insensitive GABAA receptors in the pre-autonomic circuits of the hypothalamus and the brain stem. The two major pre-autonomic sympathetic nuclei are the PVN of the hypothalamus and the noradrenergic nuclei of the brain stem (LC, A5 area; Loewy 1990b; Pacak and Palkovits 2001). We would like to suggest that the sympathetic control of the pupil is mainly via the LC, which also acts as an interface between regulation of arousal and control of pupillary function (Szabadi and Bradshaw 1996; Hou et al. 2005). On the other hand, sympathetic outflow to the cardiovascular system may be mainly under the direct influence of the PVN (Badoer 2001; Mueller et al. 2003; Coote 2005). The cold pressor test activates both the LC and the PVN (Pacak and Palkovits 2001), leading to pupil dilatation and an increase in blood pressure. As the GABAA receptors on the LC neurones are insensitive to diazepam (see above), diazepam is unable to modify the mydriatic response to the cold pressor test. On the other hand, it is well documented that neurones in the PVN are richly endowed with GABAA receptors (Kalsbeek et al. 2000; Haywood et al. 2001; Han et al. 2002), and these receptors are likely to be susceptible to modulation by diazepam. Thus, diazepam would reduce the pressor response to the cold pressor test due to the augmentation of the GABAergic inhibition of the PVN. This model may also explain the differential effects of diazepam on the heat stressor-evoked and threat-evoked sweat gland responses, as whilst it is known that heat stressor activates the PVN (Cham et al. 2006), anxiety, including that associated with conditioned fear, has been shown to lead to the activation of the LC both in experimental animals (Ishida et al. 2002; Liu et al. 2003) and humans (Liddell et al. 2005). The importance of the distribution of the benzodiazepine sensitive and insensitive GABAA receptors in the pre-autonomic sympathetic circuitry is highlighted by the fact that the barbiturates, which, like the benzodiazepines, act by modulating GABAA receptors (Rudolph 2004) and also potentiate GABAergic responses in the LC (Chen et al. 1999), are effective in antagonising the sympathetic activation of the pupil (Sigg and Sigg 1969, 1973).
In conclusion, we have confirmed the dissociation between the sedative and miotic effects of diazepam. We have shown that the absence of miosis is unlikely to be due to the equal inhibition of both the sympathetic and parasympathetic outputs to the iris, as diazepam does not attenuate the influence of either outflow on the iris. We suggest that a potential sympatholytic effect of diazepam on the iris, which is expected to accompany sedation, may be covered up by a sympathoexcitatory effect resulting from disinhibition of preganglionic sympathetic neurones. Furthermore, we have shown that diazepam, like other sedative drugs, increases pupillary sleepiness waves, which are likely to reflect the fluctuations in the firing frequency of central sympathetic neurones in the LC. Finally, the differential effect of diazepam on different sympathetic responses may reflect the distribution of diazepam sensitive and diazepam insensitive GABAA receptors in the pre-autonomic circuits of the hypothalamus and brain stem.