Introduction

Assessing the risks and benefits of medications is a complex and multidimensional task. While potential benefits of certain medications are usually apparent, assessment of possible risks is not as straightforward as these are not always captured in randomized clinical trials [1]. There are multiple reasons for this: (1) study design includes participants that do not represent the population that will be exposed to the drug; (2) exposure to medication in clinical trials may be limited, and delayed toxicity may not be evident; (3) accepted level of risk may be skewed, both in public health terms and at the individual level.

Adverse health effects can be related to mechanism of action directly or to off-target effects of a certain drug. Clinicians also deal with potential drug interactions that can result in loss of efficacy or unintended toxicity, especially in polypharmacy, which is common in headache disorders. Nearly 29% of patients with episodic headache have polypharmacy, taking five or more medications [2]. Patients take multiple combinations to treat headache when medications with different mechanisms of action provide additive or synergistic effects. Medications can be prescribed by multiple physicians, or patients self-medicate with over-the-counter drugs. Patients may take medications to treat other disorders. This can lead to adverse effects that require additional treatment, further escalating polypharmacy. Finally, individual genetic traits can increase side effects or have an impact on treatment response of medications that undergo enzymatic transformation or are substrates of membrane transporters [3].

When considering risk, it is prudent to keep in mind that inaction can have its own consequences. Taking unnecessary precautions promotes waste of medical resources, but withholding treatment for fear of side effects also comes with the risks of increasing disability from undertreated migraine and chronification of migraine. When assessing potential risks of acute migraine treatments based on known mechanisms of action, clinicians should consider the risks of alternative options, including withholding treatment.

Medications with established efficacy in the acute treatment of migraine include triptans, ergotamine derivatives, nonsteroidal anti-inflammatory drugs, and opioids such as butorphanol [4••, 5••]. While butorphanol has established efficacy in acute migraine management, opioids are not a recommended treatment option for migraine and will not be reviewed in this paper. The treatment of migraine during pregnancy and lactation was recently reviewed by R. Burch, and we recommend readers to reference that article for more information on those specific safety concerns [6]. In addition, the acute treatment of migraine in children and adolescents has been recently addressed in an updated American Academy of Neurology and American Headache Society practice guideline [7].

Nonsteroidal Anti-inflammatory Drugs (NSAIDs)

NSAIDs such as acetylsalicylic acid, ibuprofen, naproxen, and diclofenac have established (level A) efficacy in acute treatment of migraine. Their analgesic effect is mainly attributed to inhibition of two isoforms of cyclooxygenase enzymes (COX-1 and COX-2) responsible for the conversion of arachidonic acid to several bioactive lipids and prostaglandins. The COX-1 isoform regulates platelet aggregation, thrombosis, gastric cytoprotection, and renal function, explaining why inhibition of COX-1 leads to an increased risk of gastrointestinal complications and bleeding [8]. COX-2 is induced at the site of inflammation, which is believed to be the basis of the anti-inflammatory effect of NSAIDs. This isoform produces prostaglandin I2 in the vascular endothelium, especially at the time of endothelial injury, which prevents platelet aggregation and aids vasodilation. This mechanism is the basis of cardiovascular side effects of NSAIDs, because inhibition of COX-2 impairs the defense mechanism against endothelial injury [9].

Acetylsalicylic acid (ASA) binds irreversibly to both COX enzyme isoforms and inhibits platelet aggregation for the duration of the platelet life. Non-salicylate NSAIDs inhibit both COX isoforms competitively and reversibly but with varying degrees of selectivity [9]. Higher doses (1000 mg) of ASA are required for analgesic effect on headache [10, 11]. Excessively high doses (6–8 g per day) of acetylsalicylic acid can result in neurotoxicity [12], and chronic use can increase the risk of major bleeding, with relative risk ranging from 1.3 for intracranial hemorrhage to 1.58 for gastrointestinal bleeding [13].

Celecoxib is the only selective COX-2 inhibitor available in the USA. Oral solution of low-dose celecoxib (120 mg) was FDA-approved in 2020 for the treatment of migraine with or without aura in adults and is a promising option for patients with high risk for gastrointestinal (GI) complications, although it does carry a warning about cardiovascular and gastrointestinal risks [14,15,16].

Due to some degree of COX-2 inhibition, all NSAIDs now carry black box warnings emphasizing an increased risk of cardiovascular events, although naproxen at higher dose (1000 mg daily) showed lower vascular risk compared to other NSAIDs [8, 17]. In a large Danish population-based study, diclofenac was associated with a 50% increase in major cardiovascular events in patients with prior cardiovascular risk factors [18]. Use of NSAIDs was also associated with an increased risk of hospital admission for heart failure in a large study in four European countries, with odds ratio highest for ketorolac [19]. Addition of aspirin as an adjunct to other NSAIDs to reduce risk of thrombotic events may result in an increased unfavorable interaction and remains debatable; it also further increases risk of GI side effects [20].

Medications with relatively higher COX-1 inhibition include indomethacin, ibuprofen, naproxen, and ketorolac. These are associated with higher risk of gastrointestinal adverse events with highest risk reported for ketorolac [21]. Several strategies have been proposed to reduce risk of GI side effects of NSAIDs with best advice to use medication at the lowest dose for the shortest period [8]. Upper GI bleeding can be prevented with an addition of proton pump inhibitors (PPI) [22, 23]. NSAID-induced lower enteropathy is not related to acid secretion; addition of PPI does not protect against it and may in fact worsen enteropathy by disrupting the intestinal microbiota. Probiotics can reduce the risk of lower gastrointestinal injury with NSAIDs, but there are insufficient data to recommend specific probiotic strains [24].

Nephrotoxicity of NSAIDs is a long-standing concern and a debated topic in nephrology. Risk factors previously associated with nephrotoxicity from NSAIDs included higher doses, long duration of therapy, concomitant use of renin-angiotensin system inhibitors or diuretics, preexisting chronic kidney disease, and advanced age [25]. Avoidance of NSAIDs may lead to use of alternative treatments which can have harmful consequences as well. Opioid use in patients with chronic kidney disease was associated with higher risk of death. Impaired excretion of gabapentinoids in patients with reduced glomerular filtration can lead to encephalopathy, generalized weakness, ataxia, and myoclonus. Use of acetaminophen may also increase risk of renal impairment [26,27,28].

While liver toxicity of acetaminophen and aspirin is well known, hepatotoxicity of NSAIDs is commonly overlooked. The mechanism of liver toxicity associated with typical NSAIDs is likely idiosyncratic rather than due to direct intrinsic toxicity, unlike toxicity associated with aspirin and acetaminophen. The most common pattern of toxicity is hepatocellular presenting with fever, fatigue, and marked elevation of aminotransferases. Less frequent cholestatic injury presents with itching, jaundice, and elevation in alkaline phosphatase and bilirubin levels [29].

In a systematic review by Sriuttha et al., diclofenac was the most common NSAID causing liver toxicity, followed by celecoxib [30]. Ibuprofen was the most common NSAID causing drug-induced liver injury, likely due to higher doses available without prescription [31].

Drug-induced aseptic meningitis (DIAM) is a rare condition observed in association with NSAID use, especially ibuprofen, naproxen, diclofenac, and sulindac. NSAIDs were the second leading cause of DIAM after intravenous immunoglobulins in a French pharmacovigilance database analysis of 329 cases [32]. The likely mechanism of DIAM with oral NSAIDs is an immunological hypersensitivity reaction, which may explain why patients with systemic lupus erythematosus are more susceptible to this condition [33]. Rapid evolution of symptoms and polymorphonuclear pleocytosis in cerebrospinal fluid is more characteristic for DIAM and may help to differentiate it from central nervous lupus activity [34, 35]. Patients who developed DIAM after ingesting one NSAID can usually take another NSAID, as cross-intolerance between NSAIDs in this condition is rare [36].

There is no evidence-based guidance for NSAID use in patients at risk for side effects, but practical recommendations by experts are available. It is advised that prior to the initiation of NSAIDs, the cardiovascular, gastrointestinal, and renal risk profiles of the patient are assessed. Patients with stage 1 to 3 chronic kidney disease have risk of NSAID-associated nephrotoxicity similar to the general population, but in more advanced stages, hypovolemic states, or when used with concurrent renin-angiotensin system inhibitors or diuretics, NSAIDs should be avoided and alternatives should be used instead [37]. Risk of nephrotoxicity is similar between COX-2-selective and non-selective NSAIDs.

In patients with high cardiovascular risk and low GI risk, naproxen, or ibuprofen with PPI for gastric protection, or lower dose (200 mg) celecoxib may be used. Patients with high risk for both gastrointestinal and cardiovascular complications can use low-dose celecoxib with gastric protection, but other NSAIDs are not recommended [38, 39•].

Triptans

Triptans are selective serotonin (5HT) agonists specific to 5HT-1B/D receptors; some triptans also have activity at the 5HT-1F receptor. Cardiovascular safety concerns regarding triptans stem from their vasoconstrictive effect mediated by 5HT-1B receptors. This mechanism of action led to the contraindication for the use of triptans in patients with cardiovascular risks; however, triptans themselves do not increase the risk of cardiovascular events in healthy individuals. In magnetic resonance angiography studies of cranial arteries, sumatriptan prevents migraine-related dilatation and significantly constricts extracerebral middle meningeal arteries, but not cerebral arteries such as middle cerebral, internal carotid, and basilar arteries [40,41,42]. No link between triptan use and risk of stroke was found in epidemiological studies from a UnitedHealthcare database and from a General Practice Research database [43, 44]. A notable point is that population-based studies can be biased, because patients with cardiovascular risk factors are less likely to be prescribed triptans [45]. A number of angiographic studies, and studies on isolated human coronary artery from patients with normal coronary arteries, showed that triptans have minimal vasoconstrictive effect on coronary arteries and are unlikely to cause myocardial infarction in healthy individuals [46, 47]. A population-based study of adults from Southern California also showed no association between triptan use and an increased risk of myocardial infarction, heart failure, or death [48].

An expert panel from the American Headache Society concluded that triptans have an overall favorable cardiovascular risk–benefit profile in the absence of contraindications [49••]. However, in patients with cardiovascular disease, triptans are contraindicated because a small risk of triptan-induced cardiovascular events cannot be excluded. It has been disputed that this warning does not consider the mechanism of action of triptans or pathophysiology of stroke [50•]. Studies of triptan use after a stroke are lacking, and in the current medico-legal environment, prescription of triptans in patients with prior stroke takes a great deal of deliberation despite reassuring literature [51].

Assessment of vascular risk factors may help to decide if additional cardiac workup is necessary to assure that a patient does not have cardiac disease that precludes use of triptans [52, 53]. Tools for risk calculation are available online [54, 55]. In patients with a history of angina, myocardial infarction, and ischemic stroke, alternative therapies for acute treatment of headache should be considered.

Use of triptans in migraine with brainstem aura (formerly known as basilar migraine) is another topic of debate. Basilar migraine was initially described by Bickerstaff as basilar artery migraine due to vascular spasm [56]. Given the lack of evidence of vasospasm in migraine aura, this form of migraine is now called migraine with brainstem aura [57]. Observations of patients with basilar migraine, hemiplegic migraine, or migraine with prolonged aura did not identify any ischemic vascular events [58]. There are no convincing studies to support the belief that migraine with brainstem aura should be treated differently from migraine with typical aura [59].

Triptans have been reported to precipitate reversible cerebral vasoconstriction syndrome; however, patients in case reports had other risk factors [60, 61]. Cases of ischemic colitis associated with triptans were observed, and a high level of suspicion for ischemic colitis should be maintained when a patient taking triptans develops acute abdominal pain and bloody diarrhea in the absence of a known inflammatory bowel disorder [62].

There is a common hesitancy to use triptans and ergot derivatives in patients with cerebral aneurysms because of vasoconstrictive effects. One retrospective review found no complications in 10 pre-coiling and post-coiling patients with cerebral aneurysms who used triptans, suggesting that triptans can be used in this category of patients [63].

Triptans have an FDA warning about potential life-threatening serotonin syndrome when coadministered with selective serotonin reuptake inhibitors (SSRI) and selective norepinephrine reuptake inhibitors (SNRI). This warning was based on 29 case reports, but none of them met validated Hunter criteria for serotonin toxicity [64]. Serotonin syndrome is a drug-induced and dose-dependent toxidrome due to an excess of serotonin in the synaptic cleft [65, 66•, 67]. In experimental studies, serotonin toxicity was mediated by 5HT-2A receptors with some involvement of 5HT-1A receptors, while triptans are agonists with high affinity at 5HT-1B/5HT-1D/5HT-1F and lower affinity to 5HT-1A receptors. Thus, triptans are unlikely to induce this syndrome [68]. In our population-based study of patients who were co-prescribed SSRI or SNRI antidepressants and triptans, we estimated the risk of serotonin toxicity as very low at 2.3 cases per 10,000 person-years of exposure [69]. Another retrospective study of clinical outcomes after intentional triptan or ergotamine overdoses found no definitive cases of serotonin toxicity [70]. Because of low risk of serotonin toxicity, the American Headache Society does not support withholding triptans to treat migraine in patients who take SSRI or SNRI antidepressants [71].

Notable interactions of triptans include eletriptan with CYP3A4 inhibitors, frovatriptan with CYP1A2 inhibitors, sumatriptan and rizatriptan with monoamine oxidase inhibitors, and rizatriptan with propranolol use, which are all listed in product labeling [72].

Ditans

Ditans are selective agonists of 5HT-1F receptors expressed on trigeminal neurons, allowing for the treatment of migraine attacks via a neural mechanism and without the vasoconstrictive effects typically seen with triptans [73]. Lasmiditan is first in class and currently the only ditan available in the USA. It penetrates the blood–brain barrier and can cause CNS-related side effects such as dizziness, fatigue, paresthesia, and sedation [74]. Lasmiditan can impair driving performance despite subjective perception of the ability to drive safely [75]. It is currently classified as a schedule V controlled substance [76].

Since lasmiditan acts on serotonin receptors, there is a theoretical risk of serotonin syndrome, regardless of the presence of other serotonergic medications. In two phase 3 single-migraine-attack studies of lasmiditan, five cases of possible serotonin syndrome were identified, but none met validated Hunter criteria [77]. This is not surprising, as serotonin receptor 1F has not been implicated in the development of serotonin syndrome.

Lasmiditan can interact with a number of medications, including alcohol (resulting in an additive effect) and substrates of P-glycoprotein, such as calcium-channel blockers, cyclosporine, dabigatran etexilate, digoxin, erythromycin, loperamide, protease inhibitors, and tacrolimus [78], but no clinical studies are available to assess clinical significance of these interactions [79].

Lasmiditan can potentiate bradycardia when used concomitantly with heart rate-lowering medications. This was observed in healthy individuals receiving propranolol 80 mg twice daily after one 200-mg dose of lasmiditan, but exact mechanism of interaction remains unclear [80,81,82].

Risk of medication overuse headache (MOH) with lasmiditan remains unknown. One of the hypothesized mechanisms of MOH is related to desensitization and downregulation of the receptors after prolonged exposure to agonists, which includes ditans and triptans [83]. In a pre-clinical rat model of medication overuse headache, lasmiditan induced acute transient cutaneous allodynia in a similar way to sumatriptan, which suggests that lasmiditan may have the capacity to induce MOH [80, 84]

Ergot Alkaloids

Of more than 80 ergot alkaloids, three are used to treat migraine. Ergotamine and dihydroergotamine (DHE) are used for acute treatment, and methysergide is used for prophylaxis [85]. Like triptans, the anti-migraine properties of ergot alkaloids are due to agonism at 5HT-1B and 5HT-1D serotonin receptors. In addition, ergot alkaloids also have high affinity to 5HT-1A and 5HT-2A receptors, dopamine receptor D, and α1/α2-adrenergic receptors, which may contribute to the potential side effects on the one hand, and more clinical efficacy in those with inadequate response to triptans on the other hand [86, 87].

Compared to triptans, ergot alkaloids have more potent vasoconstrictive effects on peripheral arteries, including pulmonary, cerebral, temporal, and coronary arteries. Ergot alkaloids can transiently increase systemic blood pressure for 3 h after parenteral use and should be avoided in patients with uncontrolled hypertension [88]. Ergot alkaloids are also contraindicated in patients with coronary, cerebral, and peripheral vascular disease; arteriosclerosis; in those with clinical symptoms of coronary vasospasm including Prinzmetal’s variant angina; sepsis; and following vascular surgery; and in those with severely impaired hepatic or renal function. Because of oxytocic properties and the ability to cause developmental toxicity, ergot alkaloids are contraindicated in pregnancy and should not be used by nursing mothers [89].

Although both ergotamine and DHE are potent constrictors of venous capacitance vessels, ergotamine is a more potent arterial vasoconstrictor. It has stronger uterotonic effects and causes nausea more often. DHE is better tolerated and is less likely to cause nausea and vomiting, although intravenous DHE can be associated with diarrhea.

Prolonged use of ergotamine can lead to medication overuse headache and can also result in overt ergotism with gangrene, peroneal nerve ischemic neuropathy, anorectal ulcers with chronic use of suppositories, and retroperitoneal, pulmonary, pleural, pericardial, or heart valve fibrosis [90]. Cases of ergotism have been reported when ergotamine was administered with strong CYP3A4 inhibitors, such as macrolide antibiotics; protease inhibitors; and antifungals such as ketoconazole, because both ergotamine and DHE are believed to be a substrate to CYP3A4 [91, 92].

Unlike ergotamine, DHE has minimal risk of MOH and can be used to treat refractory migraine and status migrainosus via continuous infusions over 1–3 days or as repeated intermittent boluses [93, 94]. The bioavailability of ergotamine and DHE highly depends on route of administration because of significant first-pass metabolism. Due to very low oral bioavailability (less than 1%), oral administration of DHE is not useful for acute treatment of migraine [94]. Parenteral formulations of DHE for intravenous and intramuscular administration have 100% bioavailability, and intranasal delivery of DHE using the nasal pump has 40% bioavailability.

The INP104 (DHE administered via Impel’s Proprietary Precision Olfactory Delivery Technology) is another formulation for intranasal use and was approved by the FDA in September 2021 [95] backed by reassuring safety and tolerability study results [96]. Other formulations of DHE for intranasal or orally inhaled formulations are in development with data reassuring for cardiovascular safety of DHE. Orally inhaled DHE MAP0004 was efficacious and well tolerated, but MAP0004 was not approved because of manufacturing issues with the delivery system [87]. A trial of STS101 (dihydroergotamine nasal) powder did not achieve statistical significance on the co-primary endpoints of freedom from pain and freedom from most bothersome symptoms, but a new phase 3 efficacy trial started in 2021 with estimated completion in 2022 [97].

Antiemetics

Antiemetics comprise a diverse group of medications which includes dopamine receptors antagonists, antihistamines, anticholinergic agents, 5HT-3 antagonists, cannabinoids, benzodiazepines, corticosteroids, and neurokinin-1 receptor antagonists [98].

Dopamine antagonists are the most commonly used antiemetics in migraine and are well studied. They include chlorpromazine, prochlorperazine, promethazine, haloperidol, droperidol, and metoclopramide. The antiemetic effect of these medications is mediated by both peripheral (enteric) and central (area postrema) D2 dopamine receptors. Dopamine antagonists can also be used to prevent a migraine attack in patients who have clear prodromal symptoms such as yawning, mood changes, irritability, and fatigue. The atypical neuroleptics olanzapine and quetiapine are used for both acute treatment of prolonged migraine and prophylaxis; however, weight gain and sedation limit their long-term use [99].

These dopamine antagonists are associated with various drug-induced movement disorders. Acute dystonia and acute akathisia can emerge within hours to days of starting a dopamine antagonist. Diphenhydramine reduces risk of akathisia when used with prochlorperazine or high dose (20 mg) metoclopramide [100]. Drug-induced parkinsonism can develop days to weeks after exposure to dopamine antagonists, more commonly with chronic use, with two-thirds of patients recovering within weeks after discontinuation of the offending drug [101]. Tardive syndromes, such as classic tardive dyskinesia, tardive dystonia, tardive akathisia, tardive tremor, and tardive tics, typically appear after months or years of exposure; however, it can be seen with just over 3 months of metoclopramide use [102]. Older women and patients with diabetes, liver or kidney failure, or concomitant antipsychotic drug therapy are at increased risk for this condition [103, 104]. Although development of tardive dyskinesia after an isolated dose of antidopaminergic drug is unlikely, the risk with intermittent use remains unknown and it has been suggested that their use should be limited to 2 days a week [105].

Neuroleptic malignant syndrome is a feared life-threatening reaction to dopamine-blocking drugs reported after exposure to metoclopramide, prochlorperazine, and other dopamine antagonists [101]. This idiosyncratic reaction to therapeutic doses of dopamine antagonists presents with cogwheel rigidity, hyperthermia, autonomic dysfunction, and mental status change typically associated with an elevated creatine phosphokinase and requires an intensive level of care as a neurological emergency [106]. Although there are similarities between neuroleptic malignant syndrome and serotonin toxicity, hyperkinesia and clonus in serotonin syndrome can be distinguished from bradykinesia and lead-pipe rigidity in neuroleptic malignant syndrome [107].

When dopaminergic antiemetics are contraindicated or poorly tolerated, 5HT-3 receptor antagonists can be used. Of the four such antiemetics currently available in the USA, ondansetron is most used. There is a lack of research on this group of medications in migraine, and they are not included in current guidelines for acute migraine treatment.

Prolonged QT is a potential side effect of both dopamine receptor antagonists and serotonergic antiemetics, among many other medications. It increases risk of torsades de pointes (TdP) and sudden cardiac death. Cases are reported after intravenous administration of ondansetron, although arrhythmia was not observed after a single dose of oral ondansetron in healthy individuals [108, 109]. QTc prolongation of more than 500 ms is a contraindication for some neuroleptics, including chlorpromazine, droperidol, and haloperidol [99]. Monitoring of QTc intervals to ensure they remain below 500 ms during treatment is recommended.

Many medications commonly prescribed in patients with migraine can increase risk of QTc prolongation. These include citalopram/escitalopram, venlafaxine, nortriptyline, amitriptyline, imipramine, and tizanidine, among others. An extensive list of medications that prolong QT and induce TdP can be found at the CredibleMeds.org website [110].

CGRP Blocking Agents

Two classes of calcitonin gene-related peptide (CGRP) blocking agents are used to treat migraine: CGRP receptor blocking small molecules (gepants) and CGRP monoclonal antibodies (mAbs) that block either the CGRP receptor (erenumab) or ligand (galcanezumab, fremanezumab, eptinezumab) [111]. Although central activity of gepants was suggested, the blood–brain barrier permeability of gepants and CGRP mAbs is very low, implicating a peripheral site of action for these medications, most likely at the level of the trigeminal ganglion or dura mater and meninges outside the blood–brain barrier [112].

CGRP belongs to a family of neuropeptides and exists in two forms. α-CGRP, most relevant to migraine, is the principal form found in the central and peripheral nervous systems. The β-CGRP isoform is mainly found in the gut; it is formed by a different gene but has 90% homology with α-CGRP. CGRP is contained in perivascular nerves, providing the link with the cardiovascular system [113]. CGRP receptors belong to a group of family B G-protein-coupled receptors that also share structural homology. They are located at multiple sites involved in the pathophysiology of migraine, in both the central and peripheral nervous systems [112].

At this time, two orally administered gepants (ubrogepant and rimegepant) are approved for acute migraine treatment, and intranasal zavegepant (formerly known as vazegepant) is in late-stage development [114]. Another member of this class, atogepant, was approved by the FDA in September 2021 only for the preventive treatment of episodic migraine based on promising results from clinical trials [115]. Rimegepant gained FDA approval for the preventive treatment of episodic migraine treatment as well, making it the first CGRP-specific drug with both acute and prophylactic indications for migraine [116]. CGRP mAbs are currently approved for prevention of migraine, and we further discuss safety of prolonged CGRP antagonists in Part 2: Preventive Treatments.

Concerns for hepatotoxicity halted development of some small‐molecule CGRP receptor antagonists, but both ubrogepant and rimegepant are well tolerated without concerns for hepatotoxicity in healthy individuals, even when used at a high frequency [117, 118]. Rimegepant as a single dose was also well tolerated in healthy individuals, as well as in those with various degrees of liver function impairment. However, maximum observed plasma concentration was increased twofold in persons with severe liver impairment compared to matched healthy individuals [119].

In a long‐term safety evaluation trial, ubrogepant was also well tolerated. Of 1230 participants receiving ubrogepant 50 mg, 100 mg, or usual care, there were only three cases of alanine aminotransferase (ALT) or aspartate aminotransferase (AST) elevations of ≥ 3 times the upper limit of normal that were either possibly or probably related to treatment [120]. Every other day administration of rimegepant for 12 weeks in a phase 2/3 randomized, double-blind placebo-controlled trial showed tolerability similar to placebo with low rates of increased enzymes in both treatment groups [121]. In patients with severe hepatic impairment (Child–Pugh C), rimegepant and atogepant should be avoided; ubrogepant can be used at a reduced dose. In patients with end-stage renal disease (creatinine clearance, CLcr < 15 mL/min) rimegepant and ubrogepant should be avoided; atogepant can be used at the lowest dose 10 mg. Ubrogepant can be used at reduced dose in those with severe renal impairment (CLcr 15–29 mL/min).

The cardiovascular safety of gepants deserves special attention. Gepants were well tolerated in regulatory clinical trials, and this class of drugs has become a welcome addition to the current migraine management armamentarium. These do not have direct vasoconstrictive effects, and neither ubrogepant nor rimegepant has any cardiovascular contraindications listed on the product label. A trial of ubrogepant in study participants with cardiovascular risk factors had no serious cardiovascular adverse events after treatment of a single attack [122]. In long-term studies that included individuals with cardiovascular risk factors, both rimegepant and ubrogepant, used intermittently for up to 1 year, were well tolerated, and there were no cardiovascular safety issues [120, 123]. It is important to note, however, that clinical trials of both rimegepant and ubrogepant excluded patients that had cardiovascular events within 6 months prior to enrollment [122, 124] and use of these medications in patients with recent stroke or myocardial infarction has not been studied.

In animal studies by Mulder et al. [125], olcegepant (not used clinically) and rimegepant worsened ischemic stroke in mice via collateral dysfunction. While these data cannot be directly extrapolated to clinical practice, they raise the question of whether these drugs can worsen an outcome of coincidental stroke or myocardial infarction.

Given that gepants are marketed as medications without cardiovascular concerns, this can lead to the perception that they are safer alternatives to triptans in patients with migraine who experience a recent acute ischemic event. More data on safety with long-term use are needed, but we believe it is reasonable to avoid gepants for at least 6 months after an acute stroke or myocardial infarction, as well as those with unstable disease (such individuals likewise were excluded from clinical trials). Although arbitrary, this time frame seems to be reasonable to assume recovery from an acute vascular event and allow time for risk factors to be adequately addressed prior to initiation of treatment with gepants in this category of patients.

All currently available gepants are subject to multiple drug interactions when coadministered with CYP34A inhibitors or inducers [126, 127]. Medications that may decrease levels of gepants with loss of efficacy include strong CYP3A4 inducers such as phenytoin, barbiturates (including butalbital), rifampin, St. John’s wort, and carbamazepine [92].

Unlike inducers, CYP3A inhibitors increase the exposure to gepants and have the potential to lead to high serum levels. Coadministration of ubrogepant and rimegepant with strong CYP3A4 inhibitors such as ketoconazole, itraconazole, clarithromycin, and protease inhibitors should be avoided. Atogepant can be used with strong CYP3A4 inhibitors at the lowest dose 10 mg daily; no dose adjustment is necessary when used with moderate and weak inhibitors [128]. Limiting ubrogepant dose to 50 mg is advised with concomitant use of moderate CYP3A4 inhibitors such as cyclosporine, ciprofloxacin, fluconazole, fluvoxamine, grapefruit juice, and verapamil. When rimegepant is coadministered with moderate CYP3A4 inhibitors, avoidance of the next dose of rimegepant within 48 h is advised.

Dose reduction of ubrogepant or avoidance of a second dose of rimegepant within 48 h is recommended with concomitant use of inhibitors of P-glycoprotein (verapamil, carvedilol, curcumin, amiodarone, cyclosporine, lapatinib, quinidine, ranolazine, eltrombopag) or breast cancer resistance protein (BCRP) efflux transporter inhibitors (tyrosine kinase inhibitors imatinib, anti-HIV protease inhibitors nelfinavir and ritonavir, antifungal azoles, tamoxifen) [129,130,131].

Dose reduction to 10 mg or 30 mg is recommended for atogepant when used concomitantly with organic anion transporting polypeptide (OATP) inhibitors. OATPs are membrane influx transporters that participate in enzyme-based detoxification. Many drugs can inhibit OATPs causing drug-drug interactions, including gemfibrozil, cyclosporine, leflunomide, teriflunomide, clarithromycin, erythromycin, rifampicin, and anti-HIV protease inhibitors [132].

Conclusion

The acute management of migraine is both fascinating and complex. Clinicians should take into consideration not only adverse health effects directly related to known mechanisms of action, off-target effects, and drug interactions, but also risks associated with withholding the treatment for fear of side effects. With the introduction of migraine-specific therapies such as gepants and ditans, physicians and other healthcare providers have many options to choose from when considering acute treatment for their patients Table 1.

Table 1 Safety of acute migraine treatments
Full size table