Sage Journals: Discover world-class research

Abstract

Adenosine is an endogenous neurotransmitter that is released from the brain during hypoxia and relaxes isolated human cerebral arteries. Many cerebral artery dilators cause migraine attacks. However, the effect of intravenous adenosine on headache and cerebral artery diameter has not previously been investigated in man and reports regarding the effect of intravenous adenosine on cerebral blood flow are conflicting. Twelve healthy participants received adenosine 80, 120 μg kg^-1 min^-1 and placebo intravenously for 20 min, in a double-blind, three-way, crossover, randomized design. Headache was rated on a verbal scale (0-10). Regional cerebral blood flow (rCBF) with ¹³³Xe inhalation and single-photon emission computed tomography (SPECT) and MCA flow velocity (V_MCA) with transcranial Doppler, were measured in direct sequence. Six participants developed headache during 80 μg kg^-1 min^-1 and six during 120 μg kg^-1 min^-1 compared with none on placebo (P = 0.006). The headache was very mild and predominantly described as a pressing sensation. When correcting data for adenosine-induced hyperventilation, no significant changes in rCBF (P = 0.22) or V_MCA (P = 0.16) were found between treatments. A significant dilation of the superficial temporal artery (STA) was seen (P < 0.001). These results show that circulating adenosine has no effect on rCBF or V_MCA, while it dilates the STA and causes very mild headache.

Keywords

Adenosine cerebral blood flow migraine transcranial Doppler

Introduction

Adenosine is a ubiquitous endogenous nucleoside with a wide range of biological effects throughout the human body, playing a role in both vascular and nociceptive systems (1, 2). It is released from many tissues in response to metabolic stress. In the brain, it has been suggested to be a mediator between cerebral metabolism and blood flow (3). Adenosine is released from neurons during hypotension, hypoxia, seizures and cortical spreading depression (4, 5), a neurophysiological phenomenon, which may underlie the migrainous aura. The effect of infused adenosine on human cerebral blood flow (CBF) is controversial, which may reflect differences in adenosine dose or methodology used (6–8).

Evidence suggests that adenosine plays a role in migraine. Plasma levels of adenosine may be elevated in migraineurs during attacks, while interictal levels are similar to healthy controls (9). Furthermore, there is anecdotal evidence that adenosine may precipitate migraine with aura (10). Finally, the non-selective adenosine antagonist caffeine is generally assumed to have an adjuvant effect on migraine and caffeine withdrawal has been suggested as a trigger of migraine (11).

We have previously shown that dilation of the middle cerebral artery (MCA) is usually associated with headache or migraine (12). Adenosine dilates human pial arteries in vitro (13), while the effect of circulating adenosine in large cerebral arteries is unknown. The effect of a specific compound may be different in cerebral arterioles and the large cerebral arteries, possibly explained by differences in blood–brain barrier function. We hypothesized that if adenosine is a mediator involved in the migraine attack, infused adenosine would cause headache, possibly by dilating the MCA. We therefore decided to investigate the effect of two different doses of infused adenosine on headache and MCA calibre changes in normal subjects, using a modification of the headache provocation model that has previously been shown to accurately predict the migraine-generating potential for several drugs and messenger molecules (12, 14, 15).

Methods

Design and participants

A double-blind, three-way cross-over design was used. Placebo (0.9% NaCl), adenosine 80 µg kg⁻¹ min⁻¹ and 120 µg kg⁻¹ min⁻¹ were administered intravenously in a balanced randomized order on one single study day. Syringes with placebo or adenosine (5 mg/ml; Item Development AB, Stocksund, Sweden) diluted with 0.9% sterile saline were prepared and marked 1–3 representing the order of infusion by a nurse not otherwise participating in the study. The final concentration in the syringes reflected the weight of the participant, which allowed the same rate of infusion to be used at all times.

Twelve healthy participants, six men and six women, not suffering from migraine, with a median age of 25 ± 3 years were included (16). Inclusion criteria were: healthy individuals 18–35 years old and weighing 50–100 kg. Fertile women were included only if using either oral contraceptives or intrauterine devices. Exclusion criteria were: history of migraine, close relatives with migraine, episodic tension-type headache more than once a week, history or clinical findings suggesting neurological or cardiovascular disorders, asthma or bronchospasm and daily intake of any kind of medicine except oral contraceptives. A full medical history and examination was performed, and an ECG was recorded to detect pre-existing atrioventricular conductance disturbances of the heart, including Wolf–Parkinson–White syndrome. All participants gave their written informed consent before inclusion. The study was approved by the ethics committee of Copenhagen and the Danish Medicines Agency, and was conducted according to the Helsinki II declaration.

Experimental protocol

The participants arrived at the clinic at 08.00 h, free from headache. The intake of coffee, tea, cocoa or other methylxanthine-containing foods or beverages was not allowed at least 8 h before the start of the study, nor was the use of any medication, except oral contraceptives, for at least four times the plasma half-life of the drug. On arrival, the participants were placed in the supine position, and catheters [Optiva∗2 (18 G); Johnson & Johnson, Ethicon, Italy] were inserted into the right and left antecubital vein for blood sampling and drug infusion, respectively. After completing these procedures, the participant rested 30 min before starting the first round of infusion, using a time and volume controlled infusion pump (Braun Perfusor, Melsungen, Germany). Each infusion lasted 20 min followed by a 40-min wash-out. In each round, baseline values were obtained at time point − 5 min (T₋₅) for headache and adverse events, middle cerebral artery flow velocity (V_MCA), superficial temporal artery diameter, vital signs and plasma adenosine. Infusion of adenosine or placebo was initiated at T₀ and maintained for 20 min. V_MCA, superficial temporal artery diameter and vital signs were recorded at T₅, T₁₅, T₃₀, T₄₅ and T₆₀ after each dose, plasma adenosine at T₁₈, and CBF at T₁₀₋₁₅. Corresponding V_MCA values at T₁₅ were measured immediately after completing the CBF acquisition protocol. After completing the last round, the participants were discharged from the clinic, and asked to complete a headache diary every hour until 10 h after discharge. They were allowed to take rescue medication of their own choice at any time.

Adverse events

Headache was recorded on a verbal scale from 1 to 10. One represented a very mild headache, including sensation of pressing or throbbing or otherwise altered sensation in the head, 5 a headache of moderate intensity, and 10 the worst possible headache imaginable by the participant (14). Headache characteristics and accompanying symptoms were recorded according to the International Headache Society (16), using a structured questionnaire.

CBF measurement

The examination was performed with the subject in the supine position, in quiet surroundings with eyes closed and ears unplugged. Four markers were drawn on the skin to ensure accurate positioning in each acquisition. End-tidal partial pressure of CO₂ (P_etCO₂) was measured during each examination (Datex Normocap 200, Roedovre, Denmark). CBF was measured with ¹³³Xe inhalation (Ceretronix XAS SM 32C, Randers, Denmark) and single-photon emission computed tomography (SPECT), using a brain-dedicated gamma camera (Ceraspect, DSI, Waltham, MA, USA). The system uses a stationary annular NaI crystal and a fast rotating collimator. A dynamic protocol of ¹³³Xe inhalation using the Kanno-Lassen algorithm was used (17, 18). Calculation of flow in the perfusion territories of the major cerebral arteries was performed by fitting standard vascular regions of interest (ROI). Flow in the territory of the MCA (rCBF_MCA) was calculated as mean of the right and left side, since there was no difference between the sides. If significant changes in P_etCO₂ were found, values were corrected with 2% for each mmHg change (19).

Cerebral artery blood flow velocity

Blood flow velocity in the middle cerebral artery (V_MCA) was recorded bilaterally using hand-held 2-MHz probes and transcranial Doppler (TCD; Multidop X, DWL, Sipplingen, Germany). We used hand-held probes, since fixed probes cause headache and discomfort on their own (20). A time-averaged mean over four cycles each comprising approximately four cardiac cycles or 4 s was used as final measure for each time point. Identification of the MCA and marking reproducible fix points were done as previously described (18, 21). P_etCO₂ was recorded simultaneously with the transcranial Doppler measurements using an open mask that caused no respiratory resistance. If changes in P_etCO₂ were found, V_MCA was corrected with e^0.034 for each mmHg change in P_etCO₂ (22). V_MCA measurements at T₁₅ were performed immediately after CBF measurements. Since rCBF_MCA equals the product of V_MCA and cross-sectional area of the MCA, changes in MCA diameter (Δd) between treatment a and b were calculated as Δd = √[(rCBF_MCA(b) × V_MCA(b) ⁻¹ × (V_MCA(a)× rCBF_MCA(a))]⁻¹ (23).

Diameter of superficial arteries

Diameter of the frontal branch of the superficial artery was measured using high-resolution ultrasonography, 20 MHz, bandwidth 15 MHz (Dermascan C, Hadsund, Denmark) (24), as previously described (25).

Vital signs

Heart rate and blood pressure were measured using an auto-inflatable cuff (Omega 1400, Orlando, FL, USA). ECG (Cardiofax V, Nihon-Cohden, Japan) was monitored on an LCD screen and recorded on paper at every 15 min.

Biochemistry

Sample collection has been described (26). Briefly, blood collected from the antecubital vein at baseline and at T₁₈ (8 ml/sample) was drawn into vacuum tubes on ice containing a stopping solution [0.2 mm dipyridamole, 4.2 mm Na₂EDTA, 5 mm EHNA, 79 mm AOPCP, heparin sulphate 1 IU/ml, deoxycoformycin (200 µg/ml), and 0.9% NaCl], which prevents degradation and uptake of adenosine. Supernatants were deproteinized by the addition of 100 µl of perchloric acid (6 m) and then centrifuged (1500 g for 10 min). Samples were lyophilized before being chromatographed.

The assay technique has been described (26). Briefly, a Hewlett Packard HP 1100 modular system was used, with a diode array detector. Lyophilized samples (500 µl) were mixed with 1 ml of phosphate buffer (NaH₂PO₄/Na₂HPO4, pH 4), injected in a 1-ml loop, and then eluted with a methanol gradient on a Merck LIChrospher C18 column (0% for 3 min, then 10–25% methanol for 15 min). The intra- and interassay coefficient of variation for nucleosides ranged between 1 and 3%. The limit of detection at 254 nm was 1 pmol in 1 ml of plasma matrix injected. For identification and quantification, retention times and spectra were compared with those of exogenous adenosine and metabolites. Quantifications were made by comparing areas obtained for samples with areas of known quantities of nucleosides.

Reagents

Adenosine (crystallized, 99% pure), dipyridamole, a,b methylene-adenosine-5′-diphosphate (AOPCP), adenosine triphosphate (ATP), 9-erythro(2-hydroxy-3-nonyl) adenine (EHNA), were from Sigma (Saint Quentin Fallavier, France). Deoxycoformycin was from Lederle Laboratories (Paris, France). Heparin was from Sanofi Winthrop (Gentilly, France). Na₂-ethylenediaminetetracetic acid (Na₂-EDTA) was from Sigma. Bovine serum albumin (BSA) was from Johnson and Johnson Clinical Chemistry (Rochester, NY, USA). The reversed-phase chromatography column (Merck LIChrospher C18, and RP8250 × 4 mm) and other reagents were from Merck (Darmstadt, Germany).

Statistical analysis

Baseline was defined as T₋₅ before the start of infusion of each dose. Data are presented as mean ± SD, except headache score, which is listed as medians (range). Significant differences between headache scores (peak and AUC) were analysed with a non-parametric analysis of variance (Friedman test). Binary categorical data were analysed with Cochran Q-test, followed by McNemar test, if overall changes were found. Continuous variables were analysed for changes over time for each dose separately with univariate analysis of variance with the fixed factors volunteer and time. If overall differences were found, Dunnett's test was applied in order to characterize which time points were different from baseline. Analysis for differences between treatments groups were performed using the area under the response curve (AUC) as summary measure with a univariate analysis of variance with the fixed factors treatment group and subject number. If overall significance was found, differences were located with the least significant difference method (LSD).

Due to the very short lasting effect of adenosine, we did not expect any carryover effect, but period effects may occur. Study design did not allow simultaneous testing for carryover effect and period effect. Assuming no carryover effect, period effect was tested with univariate analysis of variance with the factors: number of infusion and volunteer. Test variables were baseline and response (AUC) for analysing period effects. Post hoc analysis was performed with LSD. Assuming no period effect, carryover effect was analysed using the test variables: ratio between baselines for each adenosine dose and placebo, and ratio between AUC for each adenosine dose and placebo. Post hoc analysis was performed with LSD. All analyses were performed with SPSS for Windows 11.5 (Chicago, IL, USA). P < 0.05 was considered significant.

Results

Six participants reported a headache score of more than zero during infusion of adenosine 80 µg kg⁻¹ min⁻¹ and six during 120 µg kg⁻¹ min⁻¹ compared with none during placebo. Five participants experienced headache on both doses. Median headache score by those experiencing headache was 1 during both adenosine doses. There was a significant difference between groups in peak headache score (P = 0.007) and area under the headache curve (P = 0.011). Post hoc analysis (Wilcoxon signed rank test) demonstrated that the significant changes were located between adenosine 80 µg kg⁻¹ min⁻¹ and placebo and between 120 µg kg⁻¹ min⁻¹ and placebo, but not between adenosine doses. However, only two participants experienced manifest headache on either dose. The remaining experienced an altered, pressing or throbbing, but not painful sensation in the head. Maximum headache score observed was 2, experienced by one participant at one time point. Headache was exclusively reported during adenosine infusion, suggesting that there was no confounding carryover effect on headache response. Nausea occurred in one participant during adenosine 80 µg kg⁻¹ min⁻¹ and in five during 120 µg kg⁻¹ min⁻¹ (P = 0.015). Palpitations were experienced by two during infusion of placebo, four during 80 µg kg⁻¹ min⁻¹ and nine during 120 µg kg⁻¹ min⁻¹ (P = 0.008). Subjective dyspnoea was reported by one on placebo, and six and five on adenosine doses, respectively (P = 0.006). Oppression or tightness in the chest was not seen on placebo, but experienced by six during adenosine 80 µg kg⁻¹ min⁻¹ and nine during 120 µg kg⁻¹ min⁻¹ (P = 0.002).

After discharge from the clinic, six participants developed a mild, bilateral and pressing headache (median for the affected group was 1.5, range 1–6); of these, three took rescue medication. The headache started from 3.5 to 8.5 h after the first active dose. Four of the six participants with late occurring headache had experienced immediate headache during the infusion of adenosine 80 µg kg⁻¹ min⁻¹ and/or 120 µg kg⁻¹ min⁻¹. The two participants with the most pronounced headache (peak headache score of 3 and 6, respectively) did not experience headache during the infusion of either adenosine dose.

Global CBF was 52.14 ± 7.01 ml 100 g⁻¹ min⁻¹ during infusion of placebo, 47.04 ± 5.75 ml 100 g⁻¹ min⁻¹ during adenosine 80 µg kg⁻¹ min⁻¹ and 49.65 ± 11.60 ml 100 g⁻¹ min⁻¹ during adenosine 120 µg kg⁻¹ min⁻¹. There was no significant difference between treatment groups (P = 0.15). rCBF_MCA was 51.24 ± 6.69 ml 100 g⁻¹ min⁻¹, 46.63 ± 5.65 ml 100 g⁻¹ min⁻¹ and 49.44 ± 11.39 ml 100 g⁻¹ min⁻¹, respectively. There was no difference between doses (P = 0.197). P_etCO₂ measured during CBF acquisition was 38.1 ± 3.6 mmHg and decreased during infusion of adenosine 80 µg kg⁻¹ min⁻¹ (34.8 ± 3.4 mmHg) and 120 µg kg⁻¹ min⁻¹ (33.8 ± 2.4 mmHg), P < 0.001. After correction for changes in P_etCO₂, there were still no changes in gCBF (P = 0.278) or rCBF_MCA (P = 0.22). There were similarly no changes between doses with or without correction for P_etCO₂ in all other tested vascular ROIs.

V_MCA decreased significantly over time after both adenosine 80 and 120 µg kg⁻¹ min⁻¹ (P < 0.001), but not after placebo (P = 0.19) (see Fig. 1). Baseline values are listed in Table 1. Post hoc analysis located the changes to T₅ and T₁₅. Peak decrease in V_MCA was 10.7 ± 12.4% after 80 µg kg⁻¹ min⁻¹ and 13.6 ± 10.9% after 120 µg kg⁻¹ min⁻¹. However, the difference between groups was not quite significant (P = 0.06). P_etCO₂ decreased significantly over time after both doses (P < 0.001), but not after placebo (P = 0.49), and a significant difference between groups was seen (P = 0.008). At T₁₅ the decrease in P_etCO₂ was 13.0 ± 7.2% from baseline during adenosine 80 µg kg⁻¹ min⁻¹ and 16.3 ± 7.2% during 120 µg kg⁻¹ min⁻¹. After correction for changes in P_etCO₂, there was no significant difference between groups (P = 0.16) (Fig. 1b). Calculated changes in MCA based on corresponding CBF and V_MCA measurements (23) were 0.4 ± 6.5% during infusion of adenosine 80 µg kg⁻¹ min⁻¹ and 1.8 ± 4.7% during 120 µg kg⁻¹ min⁻¹, compared with baseline, and did not significantly differ from no change (P = 0.82 and P = 0.25, respectively, one-sample t-test).

Figure 1

(a) Mean flow velocity in the middle cerebral artery (V_MCA) ± SD after placebo (•), adenosine 80 µg kg⁻¹ min⁻¹ (▪) or 120 µg kg⁻¹ min⁻¹ (▵). Values not corrected for changes in end-tidal partial pressure of CO₂. Changes were significant over time for both adenosine doses (P < 0.001), but not between treatments, though a trend was seen (P = 0.064). (b) V_MCA ± SD after placebo (•), adenosine 80 µg kg⁻¹ min⁻¹ (▪) or 120 µg kg⁻¹ min⁻¹ (▵) after correction for changes in P_etCO₂. Significant increases were seen after adenosine 80 µg kg⁻¹ min⁻¹ (P = 0.02) and 120 µg kg⁻¹ min⁻¹ (P = 0.006), but not placebo, but there were no changes between groups (P = 0.16).

Table 1

Baseline values ± SD

	Placebo	Adenosine80 µg kg⁻¹ min⁻¹	Adenosine 120 µg kg⁻¹ min⁻¹
V_MCA (cm s⁻¹)	73 ± 8	72 ± 9	70 ± 8
Temporal art diameter (mm)	1.43 ± 0.34	1.43 ± 0.34	1.43 ± 0.33
P_etCO₂ (mmHg)	37.7 ± 1.0	38.6 ± 0.9	37.9 ± 0.8
BP, systolic (mmHg)	117 ± 19	118 ± 8	118 ± 15
BP, diastolic (mmHg)	67 ± 17	68 ± 12	66 ± 12
Heart rate (min⁻¹)	59 ± 8	58 ± 7	58 ± 11
Plasma adenosine (µm)	0.64 ± 0.75	0.96 ± 1.28	0.66 ± 0.65

Baseline values for middle cerebral artery flow velocity (V_MCA), end-tidal partial pressure of CO₂ (P_etCO₂), blood pressure (BP) and other variables before the infusion of placebo, adenosine 80 µg kg⁻¹ min⁻¹ or 120 µg kg⁻¹ min⁻¹.

The diameter of the superficial temporal artery increased significantly over time after both adenosine doses (P < 0.001), but not after placebo (P = 0.691). The difference was highly significant between doses (P < 0.001). Peak increase was seen at T₁₅ and was 9.5 ± 5.3% on the lower dose, and 20.6 ± 16.2% on the higher, compared with baseline (Fig. 2).

Figure 2

Superficial temporal artery diameter ± SD placebo (•), adenosine 80 µg kg⁻¹ min⁻¹ (▪) or 120 µg kg⁻¹ min⁻¹ (▵) measured with high-resolution ultrasonography. A non-dose-dependent increase in diameter was seen.

Regarding vital signs, there were no significant changes over time or between doses for systolic or mean blood pressure. Baseline values are listed in Table 1. For diastolic BP, there were no changes between doses or over time after placebo (P = 0.08) or adenosine 80 µg kg⁻¹ min⁻¹ (P = 0.51), but a 9.1 ± 10.6% significant increase over time (P = 0.04) was seen after adenosine 120 µg kg⁻¹ min⁻¹ located at T₆₀, compared with baseline. The heart rate increased over time after both doses of adenosine (P < 0.001), but not after placebo (P = 0.243). Peak increase was 33.6 ± 28.2% after 80 µg kg⁻¹ min⁻¹ and 52.6 ± 37.4% after 120 µg kg⁻¹ min⁻¹ compared with baseline, and there was a significant difference between groups (P = 0.004).

Baseline values of plasma adenosine levels are listed in Table 1. There were no significant changes between the ratio between plasma levels during infusion of placebo, adenosine 80 µg kg⁻¹ min⁻¹ or 120 µg kg⁻¹ min⁻¹, and baseline (P = 0.228) (Fig. 3).

Figure 3

Mean plasma adenosine concentrations ± SD. Baseline compared with response during the infusion of placebo, adenosine 80 µg kg⁻¹ min⁻¹ (80) and 120 µg kg⁻¹ min⁻¹ (120). Measurements were reproducible, but revealed no changes in adenosine due to the short half-life.

There were no significant period or carryover effects for baseline or response for V_MCA, superficial temporal artery diameter, P_etCO₂, blood pressure or heart rate. There was no carryover effect for CBF or rCBF_MCA, but period effects could not be evaluated due to the nature of the data. Carryover effect on plasma adenosine could not be evaluated because of insufficient data, but there was no period effect on baseline.

Discussion

Several vasodilator arteries have been demonstrated to cause headache and migraine concomitant with MCA dilation in both normal participants and migraine patients (12, 18, 27–29), though the role of large cerebral dilation in experimental headache models is presently not clear (15, 25). No previous study has evaluated the effect of adenosine on headache response and diameter of the large cerebral arteries.

TCD measurements of velocity as a stand-alone are difficult to interpret, because it is dependent on both the diameter and rCBF, but can be used to estimate changes in the cerebral arteries in combination with quantitative rCBF. We have investigated the MCA because measurements of this artery are the most precise and reproducible using TCD (21). A previous TCD study showed a ∼ 15% decrease in V_MCA using adenosine 140 µg kg⁻¹ min⁻¹ i.v. (30), similar to our findings, but no rCBF was measured. In the same study, mean P_etCO₂ decreased from 41 to 31 mmHg, which may account for the observed decrease in V_MCA (22). In the present study, adenosine likewise caused a marked decrease in P_etCO_2, which is expected to reduce CBF and consequently V_MCA, if the MCA diameter remains unchanged. A correlation between P_etCO₂ and V_MCA has previously been established, allowing the calculation of P_etCO₂-corrected V_MCA values (22). After using this correction, there were no significant changes in V_MCA. However, it can be argued that P_etCO₂ does not always reflect P_aCO₂, which is the true determinant of CBF. This is the case if lung perfusion and ventilation do not match, e.g. due to bronchospasm or a sudden change in cardiac output. It can also be argued that adenosine potentially has several such effects. However, it has previously been established that i.v. adenosine markedly increases minute ventilation, possibly due to activation of carotid body chemoceptors (30–33), and adenosine 140 µg kg⁻¹ min⁻¹ i.v. decreases mean P_aCO₂ from 41 to 31 mmHg (31), similar to the observed change in P_etCO₂ in the present study. We therefore suggest that the observed changes in P_etCO₂ reflect changes in arterial blood gases caused by hyperventilation, and that the observed changes in V_MCA are explained by changes in P_aCO₂, not MCA dilation.

Data on the effect of adenosine on rCBF are conflicting. In a magnetic resonance imaging and positron emission tomography (PET) study, nine healthy participants received adenosine at a rate of 60–100 µg kg⁻¹ min⁻¹ i.v. causing no significant effect on CBF (6). In another PET study, seven patients scheduled for cerebral arteriovenous malformation surgery were examined during general anaesthesia and controlled ventilation. In six of the patients, P_etCO₂ was kept as close as possible to 40 mmHg, and very large doses of adenosine (200–500 µg kg⁻¹ min⁻¹) were administered via a central venous catheter. This resulted in a 55% (range 23–85%) mean increase in global CBF, though there was no clear correlation between adenosine dose and CBF changes (7). A ^99mtechnetium hexamethylpropyleneamine oxine (Tc-HMPAO) SPECT study, including six healthy participants and six patients with occlusive carotid disease, showed a marked increase of approximately 30% in the cortex and 70–80% in the thalamus and basal ganglia after adenosine 140 µg kg⁻¹ min⁻¹ i.v. (8), but this tracer is only suitable for measuring the relative distribution of rCBF, not absolute rCBF changes (34). The present results support the PET study (6) and together this strongly suggests that doses < 120 µg kg⁻¹ min⁻¹ i.v. do not increase CBF in awake, spontaneously breathing participants. In contrast, a 6–10%, but not statistically significant (P = 0.15), decrease in global CBF was seen, possibly explained by hypocapnia. Using these rCBF_MCA measurements, there was no change in calculated MCA diameter (23) in the present study.

There are at least two possible explanations to why adenosine does not dilate cerebral arteries and arterioles in man, despite the presence of vasodilatory A₂ receptors in cerebral artery smooth muscle cells and a dilatory effect on human cerebral arteries in vitro (1, 13, 35). Firstly, due to the very short plasma half-life of < 10 s (36), adenosine might be cleared from the blood before reaching the brain arteries. However the superficial temporal artery dilated significantly in the present study, and P_etCO₂ decreased (32, 33). Adenosine therefore must have reached the cranial vasculature. More likely, circulating adenosine may not reach the vascular receptors. In vitro experiments, with a model that allows selective intra- and extraluminal application of drugs to rat cerebral arterioles, support this hypothesis, since adenosine and adenosine analogues are much less potent when applied into the lumen than when applied extraluminally (37). These findings, together with the present data, also seem to rule out the presence of endothelial adenosine receptors. In addition, studies using radiolabelled adenosine infused into the bloodstream or the cerebrospinal fluid in the dog show that very little adenosine crosses the blood–brain barrier (BBB) in either direction (3). Considering the high density of nucleoside transporters in human cerebral arterioles (35), it is surprising that adenosine does not cross the BBB except possibly in very high concentrations. The explanation is probably that adenosine is rapidly degraded in the vascular endothelium, which thus constitutes a ‘biochemical BBB’ (35). If the previously reported elevated adenosine concentrations during migraine attacks prove to be a reproducible finding, the adenosine probably originates from extracerebral sites, e.g. platelets or vascular tissue.

The role of adenosine in headache generation

In a large open-label multicentre study, adenosine 140 µg kg⁻¹ min⁻¹ over 6 min caused headache in 14.2% of more than 9256 patients undergoing cardiac stress testing (38). This study has limited value because there was no control group, and headache is a frequently reported adverse event in both placebo and active groups in most drug trials. The present study used a randomized, double-blind, cross-over design with frequent and systematic scoring of headache. We found significantly more headache after adenosine than after placebo. The headache induced during the infusion was very mild and merely reported to be an altered, pressing or throbbing sensation in the head by four of six on both adenosine doses. Since there was no indication of a dose–response relationship in the present study, it is doubtful that higher doses would have caused significant degrees of headache. Due to need to maintain the infusion for 20 min and the pronounced discomfort experienced by all participants, we estimated that 120 µg kg⁻¹ min⁻¹ was the highest testable dose. Considering the marked side-effects it seems improbable that even higher plasma adenosine levels would occur naturally during a migraine attack. Migraineurs are more sensitive to headache-provoking substances than healthy participants, but in our experience substances that provoke migraine in migraine patients cause a much greater headache response in normal participants than seen in the present study (mean peak headache 2.5–3 vs. 0.5) (15, 25, 39).

We found a surprisingly high number of participants who developed delayed headache several hours after the experiment. For this, however, there was no placebo control, since the study was not designed to address this problem. With other infused substances such as histamine, calcitonin gene-related peptide (CGRP) and glyceryl trinitrate (GTN), a biphasic headache response has been observed in patients with migraine without aura, i.e. an immediate response observed during infusion followed by a secondary headache response 3–8 h later, long after elimination of the test substances (28, 29, 39). This phenomenon has also been demonstrated in normal participants after GTN (39), but less frequently. There was, however, no evident correlation between immediate and delayed headache response, suggesting that the headache observed after finishing the study may not be true delayed headache. Indeed, the patients reporting the highest headache score of 3 and 6, respectively, in the observation period, experienced no headache during adenosine infusion. In summary, circulating adenosine is not a likely mediator of migraine attacks. The question remains, whether the migraine attack could be caused by adenosine released from perivascular nerve terminals or from the brain itself. In this context, it is interesting that intrathecal adenosine had a general analgesic effect, but nevertheless caused mild headache in three of six normal subjects at a dose of 2 mg (40).

In conclusion, intravenous adenosine in the maximal tolerated dose has no dilator effects on the MCA or cerebral resistance vessels, but dilates extracranial arteries. We found little evidence supporting a role for circulating adenosine in the pathogenesis of migraine or headache, but repeating the study with migraine patients may be necessary to preclude an effect in particularly sensitive subjects. It is likewise possible that adenosine released from perivascular nerve terminals or from the brain itself may cause headache.

Footnotes

Acknowledgements

The authors thank Boehringer Ingelheim Pharma KG (Ingelheim am Rhein, Germany) for financing the study, MSc Joergen Holm Petersen, Department of Medical Statistics, University of Copenhagen, for expert statistical advice, and lab technicians Lene Elkjaer and Kirsten Brunsgaard for their technical assistance. S.B. receives a grant from the County of Copenhagen.

References

Tabrizchi

Bedi

. Pharmacology of adenosine receptors in the vasculature. Pharmacol Ther 2001; 91: 133–47.

Sawynok

. Adenosine receptor activation and nociception. Eur J Pharmacol 1998; 347: 1–11.

Berne

Rubio

Curnish

. Effect on cerebral vascular resistance and incorporation into cerebral adenine nucleotides. Circ Res 1974; 35: 262–71.

Winn

Rubio

Berne

. The role of adenosine in the regulation of cerebral blood flow. J Cereb Blood Flow Metab 1981; 1: 239–44.

Kaku

Hada

Hayashi

. Endogenous adenosine exerts inhibitory effects upon the development of spreading depression and glutamate release induced by microdialysis with high K+ in rat hippocampus. Brain Res 1994; 658: 39–48.

Stange

Greitz

Ingvar

Hindmarsh

Sollevi

. Global cerebral blood flow during infusion of adenosine in humans: assessment by magnetic resonance imaging and positron emission tomography. Acta Physiol Scand 1997; 160: 117–22.

Sollevi

Ericson

Eriksson

Lindqvist

Lagerkranser

Stone-Elander

. Effect of adenosine on human cerebral blood flow as determined by positron emission tomography. J Cereb Blood Flow Metab 1987; 7: 673–8.

Soricelli

Postiglione

Cuocolo

De Chiara

Ruocco

Brunetti

Effect of adenosine on cerebral blood flow as evaluated by single-photon emission computed tomography in normal subjects and in patients with occlusive carotid disease. A comparison with acetazolamide. Stroke 1995; 26: 1572–6.

Guieu

Devaux

Henry

Bechis

Pouget

Mallet

Adenosine and migraine. Can J Neurol Sci 1998; 25: 55–8.

10.

Brown

Waterer

. Migraine precipitated by adenosine. Med J Aust 1995; 162: 391.

11.

Couturier

Hering

Steiner

. Weekend attacks in migraine patients: caused by caffeine withdrawal? Cephalalgia 1992; 12: 99–100.

12.

Thomsen

Olesen

. Nitric oxide theory of migraine. Clin Neurosci 1998; 5: 28–33.

13.

Hardebo

Edvinsson

. Adenine compounds: cerebrovascular effects in vitro with reference to their possible involvement in migraine. Stroke 1979; 10: 58–62.

14.

Iversen

Olesen

Tfelt-Hansen

. Intravenous nitroglycerin as an experimental model of vascular headache. Basic characteristics. Pain 1989; 38: 17–24.

15.

Kruuse

Thomsen

Jacobsen

Olesen

. The phosphodiesterase 5 inhibitor sildenafil has no effect on cerebral blood flow or blood velocity, but nevertheless induces headache in healthy subjects. J Cereb Blood Flow Metab 2002; 22: 1124–31.

16.

Headache Classification Committee of the International Headache Society. Classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988; 8: 1–96.

17.

Kanno

Lassen

. Two methods for calculating regional cerebral blood flow from emission computed tomography of inert gas concentrations. J Comput Assist Tomogr 1979; 3: 71–6.

18.

Kruuse

Jacobsen

Lassen

Thomsen

Hasselbalch

Dige-Petersen

Dipyridamole dilates large cerebral arteries concomitant to headache induction in healthy subjects. J Cereb Blood Flow Metab 2000; 20: 1372–9.

19.

Shirahata

Henriksen

Vorstrup

Holm

Lauritzen

Paulson

Regional cerebral blood flow assessed by 133Xe inhalation and emission tomography: normal values. J Comput Assist Tomogr 1985; 9: 861–6.

20.

Thomsen

Ahlgren

Iversen

Olesen

. Middle cerebral artery blood flow velocity increases during experimental pressure pain in the head. In: Olesen

Moskiwitz

, editors. Experimental headache models. New York: Raven Press, 1995: 337–9.

21.

Thomsen

Iversen

. Experimental and biological variation of three-dimensional transcranial Doppler measurements. J Appl Physiol 1993; 75: 2805–10.

22.

Markwalder

Grolimund

Seiler

Roth

Aaslid

. Dependency of blood flow velocity in the middle cerebral artery on end-tidal carbon dioxide partial pressure—a transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 1984; 4: 368–72.

23.

Dahl

Russell

Nyberg-Hansen

Rootwelt

. Effect of nitroglycerin on cerebral circulation measured by transcranial Doppler and SPECT. Stroke 1989; 20: 1733–6.

24.

Nielsen

Iversen

Tfelt-Hansen

Olesen

. Small arteries can be accurately studied in vivo, using high frequency ultrasound. Ultrasound Med Biol 1993; 19: 717–25.

25.

Kruuse

Thomsen

Birk

Olesen

. Migraine can be induced by sildenafil without changes in middle cerebral artery diameter. Brain 2003; 126: 241–7.

26.

Guieu

Peragut

Roussel

Hassani

Sampieri

Bechis

Adenosine and neuropathic pain. Pain 1996; 68: 271–4.

27.

Kruuse

Jacobsen

Thomsen

Hasselbalch

Frandsen

Dige-Petersen

Effects of the non-selective phosphodiesterase inhibitor pentoxifylline on regional cerebral blood flow and large arteries in healthy subjects. Eur J Neurol 2000; 7: 629–38.

28.

Lassen

Haderslev

Jacobsen

Iversen

Sperling

Olesen

. CGRP may play a causative role in migraine. Cephalalgia 2002; 22: 54–61.

29.

Lassen

Thomsen

Olesen

. Histamine induces migraine via the H1-receptor. Support for the NO hypothesis of migraine. Neuroreport 1995; 6: 1475–9.

30.

Zanferrari

Razumovsky

Lavados

Sen

Oppenheimer

. Effect of adenosine on cerebral blood flow velocity. J Neuroimaging 2001; 11: 272–9.

31.

Biaggioni

Olafsson

Robertson

Hollister

Robertson

. Cardiovascular and respiratory effects of adenosine in conscious man. Evidence for chemoreceptor activation. Circ Res 1987; 61: 779–86.

32.

Watt

Reid

Stephens

Routledge

. Adenosine-induced respiratory stimulation in man depends on site of infusion. Evidence for an action on the carotid body? Br J Clin Pharmacol 1987; 23: 486–90.

33.

Reid

Watt

Penny

Newby

Smith

Routledge

. Plasma adenosine concentrations during adenosine-induced respiratory stimulation in man. Eur J Clin Pharmacol 1991; 40: 175–80.

34.

Edvinsson

MacKenzie

McCulloch

. Fundamental responses of the cerebral circulation. In: Edvinsson

Krause

, editors. Cerebral blood flow and metabolism, 1st edn. New York: Raven Press, 1993: 497–523.

35.

Kalaria

Harik

. Adenosine receptors and the nucleoside transporter in human brain vasculature. J Cereb Blood Flow Metab 1988; 8: 32–9.

36.

Klabunde

. Dipyridamole inhibition of adenosine metabolism in human blood. Eur J Pharmacol 1983; 93: 21–6.

37.

Ngai

Winn

. Effects of adenosine and its analogues on isolated intracerebral arterioles. Extraluminal and intraluminal application. Circ Res 1993; 73: 448–57.

38.

Cerqueira

Verani

Schwaiger

Heo

Iskandrian

. Safety profile of adenosine stress perfusion imaging: results from the Adenoscan Multicenter Trial Registry. J Am Coll Cardiol 1994; 23: 384–9.

39.

Olesen

Iversen

Thomsen

. Nitric oxide supersensitivity: a possible molecular mechanism of migraine pain. Neuroreport 1993; 4: 1027–30.

40.

Eisenach

Curry

Hood

. Dose–response of intrathecal adenosine in experimental pain and allodynia. Anesthesiology 2002; 97: 938–42.

The Effect of Circulating Adenosine on Cerebral Haemodynamics and Headache Generation in Healthy Subjects

Abstract

Keywords

Introduction

Methods

Design and participants

Experimental protocol

Adverse events

CBF measurement

Cerebral artery blood flow velocity

Diameter of superficial arteries

Vital signs

Biochemistry

Reagents

Statistical analysis

Results

Discussion

The role of adenosine in headache generation

Footnotes

Acknowledgements

References