Abstract
Introduction
The clinical characteristics of cluster headache (CH) are so peculiar that its diagnosis is often quite unproblematic. The diagnostic criteria, codified in 1988, can be found in Chapter 3 of the International Headache Society (IHS) classification (1). CH is a primary headache, but there have been cluster-like cases that ended up being secondary.
In a review we recently carried out (2), we found more than 100 cases of symptomatic cluster-like headache. An careful analysis of them highlights how the primary elements of suspicion are a positive neurological examination and an attack that lasts> 180 min. Vascular pathologies end up the most frequent cause of cluster-like headache, together with tumoral and inflammatory ones (2).
As far as we know, no case of cluster-like headache due to drug therapy has ever been reported. We report a case of CH that emerged in close temporal relationship with warfarin treatment, and immediately disappeared after the substitution of the latter with acenocoumarol.
Case report
A 62-year-old man who had been treated with sodium warfarin since he underwent a surgical operation for the substitution of his aortic valve with an artificial one a year before, came to our attention as he suffered from daily painful attacks in his right temporo-orbital region, in association with conjunctival injection, tears and homolateral ptosis. The pain, unbearable and piercing, rose regularly around 13.30 h and spontaneously receded after 3 or 4 h. These episodes started after the second day of anticoagulation therapy with warfarin. Attack therapy (acetaminophene + codeine, tramadol) was ineffective. Alcohol rapidly induced an attack. When the patient came to our observation, 11 months after the appearance of the symptoms, he had been given amitriptyline 30 mg/day, but he discontinued the treatment after 2 months for both inefficacy and somnolence. Neurological examination was negative. Due to previous surgery, cerebral magnetic resonance imaging was not recommended; a cerebral computed tomography scan with and without contrast was normal. In consideration of the age of the patient and of his hypertension and cardiopathy, sumatriptan s.c. was not prescribed for the attack; the inhalation of oxygen (7 L/min for 15 min) was recommended instead, and eventually indomethacin 100 mg. A preventive treatment with verapamil 240 mg/day was started. After 2 months from the first observation, during which the attacks persisted daily and the pain did not respond to oxygen therapy and was just mildly eased by indomethacin, verapamil was discontinued. At the same time – considering the chronological proximity between the warfarin therapy and the beginning of symptoms – we decided to substitute warfarin with another anticoagulant drug, acenocoumarol. From the very next day the patient noticed a clear improvement of his headache, which now presented itself with a helmet-type pain, diffuse but milder, without vegetative phenomena accompanying it, and that had an average duration of about 6 h. Three days after the withdrawal of the warfarin therapy, the patient did not complain of any painful symptom.
Today, after an 8 months long follow-up, the patient is still asymptomatic.
Discussion
Temporal relationship is a basic factor in associating a possible cause with the onset of a new headache. It can sometimes be difficult to correlate the two events: the longer the time between the identification of a possible origin and the onset of the headache, the weaker is the probability of a causal relationship. This concept is also valid for the headaches described in the Chapter 8 of the IHS classification, i.e. headaches caused by taking or suspending exogenous substances. The best way to substantiate a causal relationship between the assumption of an exogenous substance and the beginning of a headache, would be a double-blind vs. placebo investigation, which can confirm the clinical suspect (1). Of course, this is not always feasible, as in the present case.
The possible relationship between headache and oral anticoagulants is very seldom considered, and the data available in the literature are poor (3, 4). Although numerous drugs are available, the oral anticoagulants principally employed are warfarin and acenocoumarol. These two active compounds are marketed with a slightly different composition in their excipients: the warfarin preparation contains stearic acid, magnesium stearate, starch and lactose, while acenocoumarol contains silica precipitate, magnesium stearate, pregelatinized corn starch and lactose.
In the warfarin leaflet headache is given in the list of possible side-effects, though the characteristics and the prevalence of this symptom are not specified. Headache is not present among the acenocumarol side-effects.
Recently, research carried out on 326 patients treated with acenocoumarol highlighted that 166 suffered from headache (66 with migraine without any other nosological specification; 100 non-migraine sufferers). Among the migraineurs, 63% reported a reduction of their attacks after taking acenocoumarol, while only 38% of the non-migraineurs reported an improvement of their headaches (3).
Others anecdotal clues highlighted some improvement in the intensity and frequency of migraine attacks after taking warfarin (4, 5) and acenocumarol (6); in one case, therapy with warfarin, from its beginning, brought a reduction of the attacks (migraine with or without aura) as well as of the episodes of aura without migraine (5).
These observations might indicate an anti-cephalalgic action–in particular, anti-migrainous–both of the acenocoumarol and of the warfarin. The mechanism with which the anticoagulants would provide an improvement of the headache is unknown. A possible action on blood platelet aggregation or a biochemical action, in particular on the production of nitric oxide (NO), has been hypothesized (3).
In contrast with the indications above, in our case the close relationship between treatment with warfarin and the appearance of a cluster-like headache seems to indicate a causal role of the drug in provoking these symptoms: this is strengthened by the fact that the attacks disappeared simultaneously with the discontinuation of the therapy. Another possibility is that the disappearance of the cluster-like headache was due to a therapeutic action of acenocoumarol, as reported by Morales-Asin et al. (3), and not to the warfarin withdrawal, but the close relationship between warfarin administration and the cluster-like headache onset and its prompt remission with warfarin withdrawal makes this hypothesis highly unlikely.
To hypothesize a mechanism is problematic: since the warfarin has been replaced by a molecule that performs the same pharmacological action, i.e. the inhibition of the coagulation factors II, VII, IX and X, as well as of protein C, and the attacks manifested themselves in conjunction only with the warfarin therapy, it seems difficult to attribute a causal role to the intrinsic action of this drug.
Moreover, it is not easy to suspect a causal role of an excipient present in the warfarin pill and absent in the acenocoumarol one, as the only difference between them is the silica precipitate, which in our preparations was present only in the latter. Therefore, the possible mechanisms underlying this clinical picture remain unsolved. Nevertheless, adhering to the criteria described in the IHS classification, the cluster-like headache presented by the patient can be related to the warfarin therapy, and consequently it can be ascribed among the headaches due to exogenous substances, as it possess the requirements described in 8.2 of the same classification (Table 1). With due reservations, this is the first reported case of cluster-like headache apparently subsequent to anticoagulant therapy.
Headache induced by chronic substance use or exposure (8.2): IHS diagnostic criteria (1)
