Abstract
Recent studies suggest that nitric oxide (NO) plays an important role in nitrate‐induced headache and in spontaneous migraine attacks. Organic nitrates act as prodrugs for NO and headache is a predominant adverse effect of nitrates but often disappears during continuous treatment. Insight into tolerance to headache could lead to insight into vascular headache mechanisms in general. The specific aim of the present study was therefore to characterize the headache and accompanying symptoms during continuous nitrate administration until a state of tolerance to headache had developed. 5‐isosorbide‐mononitrate (5‐ISMN) 30 mg three times daily was administered orally for 7 days in 11 healthy subjects in a double‐blind, randomized placebo controlled cross‐over design. Wash‐out between periods was 14 days or more. Haemodynamic data from the present study were compared to the observed changes of headache over time. Headache during 5‐ISMN was longer lasting and more severe compared to placebo (P < 0.004). In 10 subjects the headache fulfilled the pain sub‐criteria for migraine and in five subjects all diagnostic criteria for migraine without aura were fulfilled. Conversely, 20 min of intravenous infusion of glyceryl trinitrate caused a milder headache and no migraine. The present results therefore suggest that NO may elicit a migraine attack in many healthy subjects if a high enough dose is given for several hours. A close temporal association between the disappearance of headache and the attenuation of the 5‐ISMN induced dilatation of the superficial temporal artery was observed. In contrast, tolerance in the middle cerebral artery already appeared after 24 h, which was earlier than the development of tolerance to headache. If vasodilatation is the cause of headache the results point to extracerebral arteries. However, cytotoxic and pain modulating central nervous system effects of NO, the time courses of which are unknown, may also play a role, involving both intra‐ and extracranial arteries.
The organic nitrates, such as glyceryl trinitrate (GTN), isosorbide‐dinitrate (ISDN) and isosorbide‐5‐mononitrate (5‐ISMN), are metabolized to nitric oxide (NO) (1) which induces headache in healthy individuals and migraine in sufferers of migraine (2). Headache is the most prominent adverse effect of nitrate therapy and varies from mild or absent to severe. It may even prohibit the use of nitrates in sufferers of cardiovascular diseases (3). However, during long‐term nitrate treatment, headache often abates over a period of time. These observations have most often been made retrospectively, and by physicians with no particular interest in headache. Consequently, a more detailed description of nitrate‐induced headache and its characteristics during continuous nitrate therapy is warranted. Furthermore, tolerance to the headache‐inducing effects may elucidate the basic mechanisms of NO‐induced headache. The specific aim of the present study was to characterize prospectively and accurately, the headache and accompanying symptoms in healthy subjects during oral administration of a long‐acting nitrate (5‐ISMN).
Methods
Material, design and measurements
Sixteen healthy subjects, 15 females and one male, aged 30–51 years (mean 38.7 years) participated in the study. Excluded were subjects with a personal history of migraine (a family history of migraine was not investigated) and subjects with other headaches more than 1 day/month, subjects on daily medication other than contraceptives, and pregnant and nursing women. Prior to inclusion in the study all subjects had experienced headache but no hypotension after a sublingual test dose of 0.5 mg nitroglycerin (NTG). An interval of at least 3 days separated the study period from the day of test dose given. The study followed a double blind, randomized, placebo controlled crossover design. A 5‐ISMN dose of 30 mg was chosen because a previously performed dose–response study showed that this dose induced an acceptable intensity of headache but no major decrease in blood pressure or other severe adverse effect (4). 5‐ISMN 30 mg × 3 or placebo was administered orally during two separate periods of 7 days, each period being separated from the other by at least a fortnight of wash‐out.
Headache severity was scored on a numerical verbal rating scale from to 0–10 (1 presenting a pre‐pain, including a feeling of pressure or pulsation, 5 a moderate headache and 10 the worst possible headache) (5). Headache characteristics (pain location, quality), its aggravation by physical activity and the presence of accompanying symptoms (nausea/vomiting, phono‐/photo‐phobia) were simultaneously recorded and compared with the diagnostic criteria for migraine of the International Headache Society (IHS) (6). Eight hours prior to the investigation, weak analgesics (paracetamol, acetylsalicylic acid), antiemetic (metoclopramid), tea, coffee, alcohol or tobacco were not allowed, neither were vasoactive migraine drugs 24 h before the study. The study followed the Helsinki II declaration and was approved by the National Board of Health and by the Ethics Committee of the County of Copenhagen. All subjects gave their informed consent and were free to withdraw at any time and to receive treatment if needed.
Procedure
The subjects were free from headache on arrival in the laboratory at 09.00 after a breakfast without coffee or tea. The first dose of either 5‐ISMN(30 mg) or placebo was administered. The subjects were then allowed to leave the hospital but instructed to return 4 hours later for an interview about headache and for practising filling‐in the headache questionnaire. Afterwards the subjects were discharged from hospital and each subject was offered paracetamol/acetyl‐salicylicacid for immediate or later use, if headache became unacceptable. In order to test compliance by capsule count and review of diary record, the subjects were asked to return to the hospital the next day and after 3 and 7 days at 11.00. During each visit haemodynamic measurements of the middle cerebral, temporal and radial arteries were performed, using transcranial Doppler and high frequency ultrasound (as yet unpublished).
5‐ISMN or placebo was administered for 7 days by the subjects themselves at home at 07.00, 15.00 and 22.00. During and for a fortnight after 5‐ISMN/placebo a questionnaire concerning headache intensity and characteristics, adverse effects other than headache and the use of the offered medication, was filled out every day at home at 07.00, 15.00, 18.00 and 22.00.
Statistical analysis
Headache scores were evaluated as maximal headache scores and as summed headache scores. Differences in total headache score (total summed headache score) between the 5‐ISMN period and placebo period were analysed using Wilcoxon test for paired‐data. Differences in headache response over time (sum of headache scores/24 h) during 5‐ISMN administration and placebo were analysed separately using a non‐parametric two‐way analysis of variance, Friedman's test. Significant changes from baseline were located by a multi‐comparison procedure (Dunn's test) (7).
For all analyses, a P‐value < 0.05 was considered statistically significant.
Results
Five subjects withdrew within 24 h from start of the 5‐ISMN period because of adverse effects: vertigo (three female subjects), unacceptable headache (one female subject), nausea (one female subject). Data from these subjects were excluded from statistical analysis. Median headache scores during 5‐ISMN and placebo administration are presented in Fig. 1(A,B). Median peak headache score and accompanying symptoms are presented in Table 1.
Headache characteristics during 5‐ISMN
A and B Headache during treatment with 5‐ISMN 30 mg × 3 (A) or placebo (B) for 7 days. Individual responses (n = 11, thin lines) are shown and the bold line represents median headache intensity of 11 subjects. Headache was most severe during the first 3 days of 5‐ISMN therapy, followed by a gradual attenuation until a state of tolerance of headache developed after 5 days of therapy. During placebo no significant changes from baseline were observed.
Headache intensity and tolerance of headache
All subjects were headache free before medication. During placebo five subjects experienced short episodes of headache (non‐migrainous) which probably were related to other conditions (influenza/menstrual period). 5‐ISMN induced more headache than placebo, thus median summed headache score was 67 (range 3–200) vs. 0 (range 0–61) (P < 0.004, Wilcoxon). Headache intensity (peak, median) was also higher during 5‐ISMN administration (whole period), 5 (range 3–8) vs. 2 (range 0–7) during placebo (P < 0.012, Wilcoxon). Headache changed significantly over time during 5‐ISMN administration (P < 0.001, Friedman) while no significant changes over time were observed during placebo (P = 0.30, Friedman). Headache was experienced by eight of 11 subjects at 4 h after the first dose of 5‐ISMN (median 2, range 0–6), and after the second dose all subjects experienced a headache, the intensity of which had increased to a median of 4 (range 2–7). Headache was still present (median 3, range 0–8) in all but one subject on the fourth day. Thereafter a gradual attenuation was observed and on the sixth day significant changes from baseline were no longer present (P > 0.05, Dunn's test). However, two subjects experienced an increase in headache intensity during the end of the first week while a mild headache/pre‐pain was still present in four subjects (median 1, range 0–5) and did not disappear until 2 days after the end of administration of 5‐ISMN (Table 1).
Headache characteristics and accompanying symptoms
In most subjects the headache was pressing (seven subjects), bilateral (eight subjects) and of frontal/temporal location (11 subjects) throughout the headache period (5‐ISMN). In all but one subject the headache occasionally was aggravated by routine physical activity. This changed the headache from pressing to pulsating or from bilateral to unilateral at some time points. In one subject the headache was pulsating and unilateral throughout the headache period. Eight subjects were nauseated and five vomited. In five subjects the headache fulfilled the diagnostic criteria for migraine without aura (6). In these subjects nausea and vomiting probably were due to the development of a migraine attack, which occurred at a mean time of 9 h (range 4–12 h) after the initial dose of 5‐ISMN. Phono‐ and photo‐phobia were predominantly reported by the subjects experiencing migraine attacks. The duration of headache fulfilling migraine criteria was 1, 2, 3, 5 and 6 days, respectively. The first two subjects had recurrence of migraine after 4 and 5 days, respectively.
No adverse effects other than headache and nausea/vomiting were reported from subjects completing the 5‐ISMN period. Besides headache no adverse events were reported during the placebo period. Eight subjects took the offered paracetamol/acetylsalicylic acid within 24 h after the initial 5‐ISMN dose; during the following days 10 subjects occasionally took the ’over the counter’ (OTC) analgesics. During placebo only two subjects took OTC analgesics.
Discussion
Incidence and characteristics of nitrate‐induced headache
Almost a century ago widespread clinical and industrial use of glyceryl trinitrate (GTN) (nitroglycerin) began. Within a few years it became evident that headache was a prominent adverse effect of nitrate exposure (8). In an overall evaluation of the incidence of nitrate‐induced headache in recent double‐blind studies, approximately one‐third of studied patients experienced headache (9, 10). In a large study of quality of life, including more than 400 men with chronic stable angina during continuous transdermal GTN therapy, the majority of withdrawals were associated with headache which, in addition, had a detrimental effect on quality of life (11). A common feature, however, characterizing both early reports on headache and recent cardiological studies of the effects of organic nitrates, is that they do not include prospective systematic description of headache intensity, pain characteristics and accompanying symptoms (3, 12). Headache following nitrate therapy may turn out to be more severe, disabling and frequent than hitherto reported, as demonstrated in the present study.
In our laboratories an experimental model of human vascular headache has been developed. Using continuous intravenous infusion of GTN for 20 min, a mild to moderate headache is induced in most healthy non‐migrainous subjects. The headache resembles migraine by being pulsating and aggravated by physical activity but is generally less intense and accompanying symptoms are most often absent (5). Migraine sufferers are more sensitive to the headache‐inducing effect of GTN. This manifests in two ways; migraine sufferers experience a stronger headache during GTN‐infusion (immediate GTN‐induced headache) and at a variable time thereafter a migraine attack fulfilling the operational diagnostic criteria for migraine without aura is induced in 80% of patients (13). 5‐ISMN t.i.d. was used in the present study because its long half‐life enhances the development of tolerance when given three times daily. Like GTN it exerts its effect through liberation of nitric oxide (NO) (1).
All participants (non‐migrainous) in the present study experienced headache compared to one‐third to two‐thirds of patients treated with 5‐ISMN in other studies (12, 14–16). The difference may be due to the inclusion in the present study only of subjects responding to a test dose of NTG. However, this procedure excluded only one (female) subject. The specific attention given to headache in the present study or the lack of slow titration to optimal daily dose as recommended by cardiologists (9) are more likely explanations. In 10 of 11 subjects the headache was aggravated by physical activity. It was moderate in intensity but frequently required the use of analgesics. Thus, the pain criteria for migraine of the IHS were fulfilled in the majority of subjects. The additional criteria for accompanying symptoms were met in five subjects but only several hours after the administration of the initial dose of 5‐ISMN.
In a previous study, focusing especially on the headache inducing effect of 5‐ISMN, it was also observed that some healthy non‐migrainous participants developed migraine after 5‐ISMN (4). The same was seen during 7 h of GTN‐infusion (17). In contrast, when GTN was infused for 20 min only migraineurs experienced migraine attacks. This occurred one to several hours after the infusion (18). Thus, the effect of 5‐ISMN differs from that of a short lasting infusion of GTN in healthy non‐migrainous individuals. 5‐ISMN exerts a stronger effect and may cause a headache fulfilling migraine criteria. This is most likely caused by the much longer duration of exposure to NO after 5‐ISMN. Conversely, the short‐lasting availability of NO after GTN may cause an attack only in individuals with an (unknown) migrainous genetic predisposition. The subjects who experienced migraine (five of 11 subjects) in the present study were not interviewed about a family predisposition to migraine; it seems, however, unlikely that almost 50% of investigated subjects would have a family history of migraine.
Time profile of tolerance to headache
Tolerance is defined as an attenuation or loss of one or more effects during chronic administration. Vast evidence is available regarding the haemodynamic changes during continuous nitrate therapy, whereas knowledge of tolerance of headache in large part is limited to the clinical experience, that nitrate‐induced headache is transient. However, specific attention to the disappearance of nitrate‐induced headache was given in recent studies during treatment with 5‐ISMN over 1–2 years. During the initial 2 weeks some degree of headache was reported by one‐third of the patient population, but thereafter headache was only occasionally reported by a few patients (14, 15).
A large study using 24‐hour application of transdermal GTN had similar results; after 4 weeks headache reporting was similar in placebo‐treated and GTN‐treated patients (19). Quality of life during continuous nitrate therapy improved after 2 weeks as headache disappeared (11). During 7 h of continuous GTN‐infusion tolerance to headache was not observed (17). In the present study headache was most severe during the first 3 days after 5‐ISMN administration, followed by a gradual attenuation, until a state of tolerance to headache had developed on the sixth day. Earlier disappearance of headache in the present study compared to previous cardiological studies is probably due to a different administration regimen. The present study deliberately used three times daily administration in order to achieve maximal development of tolerance, whereas in cardiological use, 5‐ISMN is administered once daily in order to avoid cardiovascular tolerance.
Mechanisms of nitrate tolerance
All organic nitrates including GTN and 5‐ISMN exert their effects through biodegradation to NO, which increases the activity of guanylate cyclase and thereby increases the production of cyclic guanosine monophosphate (20). The latter induces vasodilatation by reducing the availability of ionized calcium to the contractile proteins (21). During chronic nitrate administration, the haemodynamic and anti‐ischaemic efficacy is rapidly blunted due to the development of tolerance, the time course of which varies widely in different sections of the vascular bed. In general, there are two proposed mechanisms of tolerance (22, 23). One is nitrate specific and relates to the development of an inability of the vascular smooth muscle to convert organic nitrates to NO, particularly caused by pharmacokinetic changes or intracellular sulfhydryl depletion. The other suggested mechanism is nitrate non‐specific and relates to systemic and physiological responses to prolonged vasodilatation, i.e. enhanced responses of the sympathoadrenal axis and of the renin‐angiotensin system, that counteracts the effects of the drug.
Mechanisms of nitrate‐induced headache and tolerance of headache
A possible mechanism whereby NO could cause headache is by dilatation of cerebral and extracerebral blood vessels. This concept is supported by studies in which arterial dilatation during and after infusion of GTN or oral administration of 5‐ISMN was consistently observed (4, 24, 25). Furthermore the haemodynamic responses during nitrate‐induced headache resemble to some extent spontaneous migraine attacks with unchanged cerebral blood flow and dilatation of both intra‐ and extracerebral large arteries (26–28). Haemodynamic data from the present study demonstrated a significant dilatation of both cephalic and limb arteries at 4 h after administration of the first dose of 5‐ISMN; in the middle cerebral artery tolerance was complete within 24 h after the first dose, whereas tolerance was only partial in the temporal and radial arteries even after 7 days (unpublished). The time course of tolerance to headache is similar to the time course in the temporal artery but considerably slower than tolerance in the middle cerebral artery, indicating that NO may cause headache via an effect on extracerebral arteries (Fig. 2). In cardiological studies, headache often disappears while the anti‐anginal effect remains, suggesting tolerance in cephalic but not in extracephalic arteries. A study of these parameters, using the recommended once daily administration of 5‐ISMN, would be highly interesting.
A close temporal association between the disappearance of headache and attenuation of the dilatory potency of 5‐ISMN in the temporal artery was observed. The intracranial artery (middle cerebral artery) was already tolerant within 24 h. The diameter of the temporal artery was measured directly, using high resolution ultrasound scanning. Middle cerebral blood velocity response was measured, using transcranial Doppler. The vessel area can be reliably estimated as cerebral blood flow (CBF) remains unchanged during nitrate administration (CBF = velocity × vessel area). ▪▪= headache score; –▪–= temporal artery (TA); –•–= middle cerebral artery (MCA).
It is likely that migraine is caused by other mechanisms in addition to arterial dilatation. Even nitrate‐induced headache may involve other mechanisms, since vasodilatation after 5‐ISMN or GTN persists longer than headache (24, 25). NO is a noxious molecule which may sensitize perivascular sensory nerve endings via the formation of free radicals or via the initiation of perivascular neurogenic inflammation (29–31). NO also causes the release of calcitonin gene‐related peptide from sensory nerve terminals (32). Furthermore, via interaction with NMDA‐receptors in the central nervous system, NO may give rise to hyperalgesia (33).
In conclusion, tolerance to nitrate‐induced headache reflects an attenuation of the sum of a variety of pain generating effects of NO, including vasodilatation, toxicity, inflammation and pain modulation. Which of those are the most strongly involved is uncertain, but if vasodilatation is the cause of headache, the vascular component most likely arises from extracerebral arteries.
Footnotes
Acknowledgements
The study was supported financially by the University of Copenhagen.