Indomethacin-induced de novo headache in hemicrania continua—fighting fire with fire?

Introduction

In the current International Classification of Headache Disorders: 2nd edition (ICHD-2), a complete response to therapeutic doses of indomethacin is a crucial feature for the diagnosis of hemicrania continua (HC) (1). Unlike naproxen or ibuprofen, indomethacin exerts its clinical effects by inhibiting nitric oxide-induced dural vasodilation (2). However, at the same time it can provoke de novo headaches as a side effect, which is illustrated by studies dating back to the 1960s in patients with rheumatologic disease (3). The mechanisms of action are largely unidentified and range from a putative medication overuse to infectious meningeal irritation. It is unknown whether patients with HC are more or less susceptible to developing an additional headache while on indomethacin as compared to patients with rheumatologic disease or gout.

Case report

A 34-year-old female patient with a one-year history of continuous daily and initially severe strictly unilateral headache with moderate ipsilateral nasal congestion and ear pain reported to our headache clinic. After a secondary form was excluded by a normal physical exam, cerebral magnetic resonance imaging (MRI) and lumbar puncture, the diagnosis of HC was assumed because of a decrease of headache severity of around 80% to therapeutic doses of indomethacin (50 mg three times a day (t.i.d.)). When increasing the dose to 75 mg t.i.d., she experienced a new onset and previously unknown moderate to severe pulsating bilateral headache with migrainous features (photo- and phonophobia and exacerbation on physical exertion) that occurred several hours per day. She had no personal or family history of migraine or recurrent headaches. Aura symptoms were not present. After the dose was decreased to 150 mg this de novo headache with migrainous features stopped. Several trials to increase the dose to 225 mg resulted in nearly complete pain relief but the migrainous bilateral headache reoccurred. Over the course of the next six months she was able to decrease indomethacin stepwise and the HC headache remitted.

Discussion

Headache induced by indomethacin is a known side effect that was reported in numerous, mostly earlier, publications. Boardman and Hart (3) analyzed a sample of 228 patients taking indomethacin for a mean of 14 weeks mainly because of rheumatologic disease, osteoarthritis and gout, and found headache to be the most frequent side effect (50 of 181 complaints reported in 113 of 228 patients). Stratification of patients into a high- (mean dose of 2.9 mg/kg/day) and low-dose (mean dose of 1.1 mg/kg/day) group revealed that headache remained the dominant complaint in the high-dose group (57.6%) compared to 29.5% in the low-dose group. In another study on patients with rheumatic disease, headaches were seen more often during consumption of indomethacin than during placebo (4). The authors report that headaches often started within an hour after intake and persisted for 24 hours or more even if it was discontinued. Additionally, they observed partial relief from antihistamines, which could point to a neurogenic or allergic pathogenesis.

In Cittadini and Goadsby’s cohort of 39 patients with HC, one patient (#27) with a history of migraine developed a bilateral headache on higher doses of indomethacin (300–500 mg). It remitted once indomethacin was stopped. The authors explain this improvement by stopping medication overuse. However, in the same sample, 28 patients practiced a similar medication overuse but no remission of their headaches was reported in the 21 patients who withdrew (5). The question arises whether the reason for headache in this patient could be something else than overuse.

A rare but well-documented cause of headache following intake of nonsteroidal antirheumatics is aseptic meningitis (6), of which recurrent cases have been reported with up to seven consecutive episodes (7). Interestingly, these cases occurred only in patients with a concomitant autoimmune disease, and immunologic hypersensitivity of type III/IV has been discussed. Although nausea and photophobia were also present in our patient, salient features of aseptic meningitis were missing such as nuchal rigidity and fever. In addition, both the delayed occurrence of aseptic meningitis and the dose-related component challenge this differential diagnosis. It might therefore be possible that above a certain dose threshold headaches are induced with a latency of some days, which may be due to a specific side effect of indomethacin. In addition, there seems to be a dose-related component, as our patient repeatedly suffered from this type of headache when indometacin was increased to 225 mg.

Indomethacin has been associated with other conditions involving secondary headaches. It has been reported in a patient with hypertrophic cranial pachymeningitis after prolonged intake for two years. After discontinuation, it took five months to remit completely (8). In a 68-year-old female, a single dose of 50 mg indomethacin intravenously (i.v.) was the presumed cause of a reversible cerebral vasoconstriction syndrome (9), which is rather unlikely as our patient complained only about headaches without any neurological deficits. Apart from this central vasoconstrictive effect, a reduction of elevated intracranial pressure and a reduction of cerebral blood flow were found in healthy volunteers and could contribute to the formation of a secondary headache.

Although migrainous features can occur during exacerbations of HC (5), the significant reduction of headache intensity at 150 mg of indomethacin per day and the bilateral location argue against a mere exacerbation of HC itself with increasing doses of indomethacin. A history of migraine predisposes to medication-overuse headache (10), which could be similar in indomethacin-induced headache. As our patient is only 34 years old, the highest age prevalence has not yet been reached and she could have an underlying migraine pathophysiology. However, the absence of a personal and family history of migraine or recurrent headaches argues against it.

As the ICHD-2 requires complete response to therapeutic doses of indomethacin (1), the diagnostic criteria are formally not met. Nevertheless, several points support this diagnosis. Firstly, a dramatic improvement was seen with 150 mg of indomethacin per day, improving further with higher doses and phenotypic change to another headache. Secondly, we assume that because of the increasing side effects the daily dose necessary for complete resolution could not be achieved. Thirdly, the necessity of a complete response has been challenged by others who reported patients with “dramatic” improvement but ongoing baseline pain (11,12).

Prakash and Shah reported delayed response to indemethacin (13), which could have explained the possibility of tapering indomethacin without headache recurrence over a period of six months. However, a gradual improvement of the underlying HC cannot be ruled out.

It is striking that only a few patients are reported with indomethacin-induced aggravations of primary headaches as compared to patients with rheumatologic disease or gout, thus case series might just be underpowered to detect meaningful data on its prevalence. It remains an unresolved issue whether patients with HC are less prone to develop secondary headaches as compared to patients with rheumatologic disease or gout—or if it is simply an underdiagnosed entity given that most patients will find it difficult to distinguish between different subtypes of headaches, which would lead to the (false) diagnosis of indomethacin-insensitive headache and hence the diagnosis of HC is (wrongly) dismissed.

Clinical implications/relevance

In patients with suspected hemicrania continua (HC) who develop de novo headache of a different character, reduce the daily indomethacin dose. Some patients with HC may take months to achieve a complete headache response with indomethacin.