Abstract
Objective
It has been reported that patients with antiphospholipid antibodies (aPL) and refractory migraine may experience symptomatic improvement with antithrombotic therapy, but this phenomenon has not been well studied. This study was undertaken to detail the response to trials of antithrombotic therapy in these patients.
Methods
This is a retrospective study of 75 patients with refractory migraine and aPL who were given a 2–4 week trial of aspirin, clopidogrel and/or anticoagulation. Major response was defined as 50–100% improvement in frequency and/or severity of migraine; minor response: 25–49% improvement; no response: <25% improvement.
Results
66 patients were given a trial of aspirin: 47% responded (21% major); 60 patients were given a trial of clopidogrel: 83% responded (67% major); and 34 patients were given a trial of anticoagulation (usually apixaban): 94% responded (85% major). The response rate to any anti-thrombotic therapy was 89% (83% major). Many patients also noted improvement in non-headache symptoms. No patient experienced stroke. There was no major bleeding during any 2–4 week treatment trial and only 3 of 69 patients maintained on an antithrombotic regimen for a median follow up of 29.9 months (5–100) experienced major bleeding.
Conclusions
There was a high rate of symptomatic response to antithrombotic therapy in this context and long-term follow up suggested an individualized symptom-derived antithrombotic regimen may be associated with a low bleeding risk. Our data support consideration of a 2–4 week trial of antithrombotic therapy, usually starting with antiplatelet therapy, in aPL-positive patients with refractory migraine, particularly if other treatment options have been exhausted. As a retrospective study, our data provide only Class IV level of evidence, but they suggest randomized controlled trials are warranted to validate these encouraging findings.
Introduction
Headache (usually migrainous) is the most common symptom in patients with the antiphospholipid syndrome (APS).1–3 APS is a systemic autoimmune disease characterized by the presence of antiphospholipid antibodies (aPL). aPL are associated with an increased risk of arterial and venous thrombosis–the hallmark of the syndrome 4 and it has been suggested that aPL may provide the missing link between migraine and stroke. 5
The Sydney criteria for a diagnosis of “definite APS” were intended for research purposes and are recognized to have diagnostic limitations.4,6 The criteria require a clinical event (thrombosis or pregnancy morbidity) and persistent moderate or high titer aPL (greater than assay cut-offs). “Criteria” aPL include anticardiolipin and anti-beta-2 glycoprotein-1 antibodies and the lupus anticoagulant. There are also “non-criteria” aPL7,8 and “non-criteria” clinical manifestations which span every specialty of medicine.4–9
It was noted historically that APS patients treated with anticoagulation for a thrombotic event often reported improvement or resolution of headaches. 1 In 2001, a 2-week “heparin therapeutic trial” was published in 21 patients with aPL and severe headaches who had not had a thrombotic event. 10 Headaches were aborted or improved in all patients and 9 had some response to aspirin, but few clinical details and no follow-up was provided. Similar findings were subsequently published in case reports.11,12 The “heparin therapeutic trial,” however, never became a widespread practice. The present study was undertaken to detail the response to trials of antithrombotic therapy in 75 patients with refractory migraine and aPL.
Methods
Patients
This is a retrospective chart review study of 75 consecutive patients meeting inclusion criteria. All patients were evaluated and followed by JRS between January 2012 and April 2020.
Inclusion criteria
Patients meeting International Classification of Headache Disorders, 3rd edition (ICHD-3) criteria for a diagnosis of episodic or chronic migraine with or without aura.
13
Patients with refractory migraine which was defined broadly as failure to manage headaches despite consultation with a neurologist and trial of preventive agents. This was a retrospective study and unfortunately details about which and how many preventive and abortive agents were failed or partially successful and continued or discontinued was not available for all patients. Many patients were treated prior to the availability of the calcitonin gene related peptide inhibitors. MRI of the brain within the preceding 5 years that was either unremarkable or showed white matter lesions that may be seen in patients with migraine and/or APS. Positivity for >= one aPL (any titer) more than once >=12 weeks apart, including criteria aPL (anticardiolipin IgG, IgM, anti-beta-2 glycoprotein IgG, IgM, lupus anticoagulant) and/or non-criteria aPL (anticardiolipin IgA, anti-beta-2 glycoprotein IgA, anti-phosphatidylserine-prothrombin IgG, IgM, antiphosphatidylserine IgG, IgM, IgA, anti-prothrombin IgG, anti-annexin V IgG, IgM, and antiphosphatidylethanolamine IgG, IgM). Competency (in the opinion of the investigator) to assess response to a therapeutic trial and comply with the recommendation to not make other medication changes during therapeutic trials. No contraindication(s) to antithrombotic therapy.
Treatment trials
Treatment trials ranged from 2 to 4 weeks depending on headache frequency as determined by patient report. If there was not a definite significant response, treatment was stopped.
Seventy two of 75 (96%) patients were initially offered a trial of aspirin or clopidogrel. The initial trial was with aspirin unless there was a contraindication to aspirin therapy (usually gastrointestinal). Patients with a suboptimal response were offered a trial of the alternative antiplatelet agent, unless it was contraindicated. Patients who responded to antiplatelet therapy, but who did not have complete headache resolution were offered a trial of add-on therapy with anticoagulation. A trial of anticoagulation was also considered in patients who didn’t respond to antiplatelet therapy. Only 34 (45%) patients were given a trial of anticoagulation. Patients who experienced an optimal response to antiplatelet therapy did not try anticoagulation and some patients received chronic anticoagulation for a thrombotic history so the effect of anticoagulation on migraine was not assessed in these patients. For 30 of 34 patients (88%) in whom a trial of anticoagulation was undertaken, it was added to antiplatelet therapy due to a positive but incomplete response. The other 4 patients were given an up-front trial of anticoagulation due to a history of VTE or active leg ulcers. Apixaban 5 mg twice daily was used as the first-line anticoagulant. The dose was adjusted up or down in a subset of patients with the long-term dose determined by symptomatic response.
Response
Major response was defined as 50-100% improvement in the frequency and/or severity of headaches; minor response: 25-49% improvement; no response: <25% improvement.
Statistical analysis
Data analysis was performed using R version 3.6.2. A paired t-test or Chi-Square test was used to assess the significant differences between groups. Statistical significance was set at p < 0.05.
Standard protocol approvals, registrations, and patient consents
The study was exempted by an independent institutional review board due to its retrospective nature and informed consent was waived.
Data availability
All data not presented in this manuscript will be made available in a trusted data repository or shared at the request of other investigators for purposes of replicating procedures and results.
Results
Patient characteristics
Clinical and laboratory characteristics of the patients are summarized in Table 1. Migraine with aura occurred in 49%; most had chronic migraine, many with daily headaches.
Clinical characteristics of the patients.
Response to antiplatelet therapy
Response to antiplatelet therapy is shown in Figure 1(a) and (b). Clopidogrel was significantly more effective than aspirin (p < 0.001).
Response to antithrombotic therapy trials.
Response to anticoagulation
Figure 1(c) shows response to anticoagulation. Most patients given an anticoagulation trial were initially treated with apixaban 5 mg bid; the dose providing optimal symptomatic response ranged from 2.5 to 12.5 mg bid. Three patients had a better response to warfarin or enoxaparin than to apixaban.
Response to antiplatelet therapy added to anticoagulation
Eleven patients had migraine despite chronic anticoagulation for a prior thrombotic event; they were offered a trial of add-on antiplatelet therapy with responses shown in Figure 1(d).
Response to any antithrombotic therapy
Response to any antithrombotic therapy is shown in Figure 2, for all patients (A), patients meeting (B) and not meeting Sydney criteria (C) and patients with only non-criteria aPL (D). There was not a statistically significant difference between any of the classification criteria subgroups and no clinical manifestation predicted response.
Response to any antithrombotic therapy.
Improvement in non-headache symptoms
Many responsive patients also reported improvement in non-headache symptoms (see Table 2).
Non-headache symptoms that improved in some patients in response to antithrombotic therapy.
Bleeding
As shown in Table 3, there was no major bleeding during any of the 2–4 week treatment trials. One patient developed significant bruising during clopidogrel administration prompting its discontinuation. Major bleeding episodes requiring emergency care or hospitalization occurred in 3 of 69 (4.4%) patients maintained on a long-term antithrombotic regimen for an average of 29.9 months (range 5 to 100) are listed in Table 3. Mild to moderate bruising was common in patients maintained on single and dual antithrombotic regimens.
Follow-up and bleeding.
Special patient populations
Adolescents
Eight of 9 adolescents (all female) had a major response and all 8 experienced complete resolution of headaches; the 9th patient had a partial response. All responded to anti-platelet therapy alone except one patient who noted a decrease in daily headaches when starting apixaban 2.5 mg bid preventively following PICC line placement. She developed DVT despite apixaban and the dose was increased to 5 mg bid, resulting in a further decrease in headaches to 2/week. A trial of adding clopidogrel for frequent line clotting resulted in complete resolution of headaches and requirement for alteplase line declotting.
Long term therapy
Three of 8 non-responders continued antiplatelet therapy preventively. The others maintained the regimen listed in Table 3 because it provided meaningful symptomatic improvement.
Loss of efficacy
Most responders continued to respond for the duration of follow-up ranging from 5 to 100 months (average 29.9). Three of 52 (6%) patients who initially had a major response to antiplatelet therapy developed increasing/recurrent headaches after a responsive period of 5 to 18 months; headaches resolved in each case by adding anticoagulation. Three of 29 (10%) patients lost efficacy to dual therapy after 23 to 86 months. One patient was free of headaches on apixaban 5 mg bid/clopidogrel for 30 months when she experienced recurrence. She responded to an increase in apixaban to 7.5 mg bid, but echocardiogram revealed worsening APS valvular disease and she was changed to warfarin (target INR of 2.5 to 3.5) which was also effective symptomatically. The second patient experienced multiple prior arterial thrombotic events, daily headache and frequent TIA symptoms on warfarin. She was stable without headache, TIA symptoms or thrombosis for 23 months on apixaban 7.5 mg bid/clopidogrel when she developed increasing headaches, vertigo and memory loss. There was symptomatic improvement with an incremental increase in apixaban to 12.5 mg bid, but given mild to moderate persistence of symptoms and multiple prior arterial events, rituximab was added resulting in further symptomatic improvement. She has remained stable on apixaban 12.5 mg bid/clopidogrel/rituximab for 11 months. The third patient had been stable without migraine for 86 months on apixaban 5 mg bid/clopidogrel. Daily headaches recurred and did not respond to a trial of increasing apixaban. Initially, headaches resolved with the addition of gabapentin; they recurred after 6 months but were rapidly aborted by adding galcanezumab. This patient has a germline B-cell mutation and migraine recurred concomitant with worsening autonomic neuropathy requiring increasing doses of intravenous immunoglobulin and rituximab.
Discussion
Migraine is a common problem, affecting up to 12% of the population. 14 Clinical clues that should raise suspicion for APS/aPL in patients with migraine headaches, include refractoriness to several preventive and/or abortive agents, a history of arterial or venous thrombosis, miscarriage, pre-eclampsia and/or intrauterine growth restriction, memory loss, livedo reticularis, Raynaud’s phenomenon, thrombocytopenia, leukopenia, CNS white matter change, seizures, dysautonomia, valvular thickening, cardiac syndrome X, stress fractures, avascular necrosis and a personal or family history of autoimmune disease (see Table 4). We propose that migraine patients with two or more of these clinical features, should be tested for the criteria and non-criteria aPL
Clinical clues in migraine patients suggesting possible APS/aPL.
High response rate
Our data show a response rate of 90% (83% major) to anti-thrombotic therapy in patients with aPL and refractory migraine. Many patients also noted improvement in non-headache symptoms (see Table 2) and it was not infrequent for patients to use descriptors such as “life changing” when describing their response. It has been hypothesized that migraine and memory loss in APS/aPL might occur due to aPL-induced cerebral vascular sludging/microthrombosis. 1 This hypothesis is supported by the present data and also by human autopsy and animal studies. Leach et al. reported widespread cerebral microthrombosis in an autopsy study of a young APS patient with complex partial seizures and transient ischemic attacks. 15 Similarly, in a well-established APS mouse model, mice develop cognitive/behavioral changes and microthrombosis of capillaries visible by electron microscopy, but not by routine histology. 16
It has been suggested that aPL may provide the missing link between migraine and stroke. 5 Many APS patients with TIA or stroke report a history of migraine 5 and a prior study showed frequent headaches in young aPL-positive patients may be followed by stroke an average of 29 (range 9-72) months after migraine onset. 17 Additionally, a meta-analysis of 43 studies found the presence of aPL was associated with a 5.5-fold increased risk for stroke or TIA. 18 No patient in our study developed stroke during an average follow up of 29.9 months, suggesting detection of aPL in patients with migraine and a symptomatic response to antithrombotic therapy might reduce stroke risk, an important area for prospective trials.
Individualized symptom-derived antithrombotic regimen
Our experience suggests symptomatic response to antithrombotic therapy may be used to safely guide treatment in this context where presumed hypercoagulability exists. This phenomenon has been reported previously by Cuadrado et. al. who described APS patients on warfarin knowing when their INR had fallen as “their headaches, mental fuzziness and even dysarthria would return”. 3 We hypothesize that symptomatic response to antithrombotic therapy in aPL positive patients may reflect normalization/near normalization of presumed prothrombotic imbalance which might account for the few bleeding events seen in our patients.
Our data suggest different APS/aPL patients respond variably to different antithrombotic agents and trial-and-error—guided by symptoms—to determine an individualized regimen may be preferable to a one-size-fits-all approach. In addition to selecting the antithrombotic regimen based on headache frequency/severity, three patients had recurrent TIA-like episodes that resolved after adjusting their regimen. Therapeutic manipulations driven by persistent symptoms included switching to a different antiplatelet agent or anticoagulant, intensification of a given anticoagulant, or using a dual antithrombotic regimen with both antiplatelet and anticoagulant agents. A subset of patients manipulated their regimen with physician guidance for various reasons, e.g. one patient transiently increased her apixaban dose when symptoms returned as estrogen levels rose during ovarian stimulation for in vitro fertilization, another increased her apixaban dose for recurrent headaches when reaching a certain estrogen dose started for severe perimenopausal symptoms and 3 patients found symptomatic benefit by transiently increasing their apixaban dose when experiencing an occasional “flare” of symptoms. Some patients also used signs such as increased bruising, bleeding longer (e.g. post-phlebotomy) or recurrent epistaxis to de-escalate their regimen. For example, one patient had mild intermittent epistaxis when taking clopidogrel 75 mg daily but not 4–5 days weekly, despite the same migraine control.
Sydney criteria
Alijotas-Reig et al. showed that in 1640 obstetric patients testing positive for aPL, those meeting and not meeting the Sydney criteria had similar treatment outcomes. 8 We likewise did not find a statistically significant difference in response in patients meeting and not meeting Sydney criteria, including those with only non-criteria aPL.
The use of apixaban in this context
A few early studies on direct oral anticoagulant (DOAC) use in APS suggested they may be inferior to warfarin, especially in triple antibody positive patients and those with arterial thrombosis.19,20 Platelets play a central role in the pathogenesis of APS. 21 Our data suggest antiplatelet therapy may be important in APS/aPL patients with migraine and that response to therapy is heterogeneous with different patients responding differently to various antithrombotic agents. We have shown that some patients require higher doses of DOAC to control APS manifestations just as some require warfarin to an INR > 3 or even >4 to control overt thrombotic events or symptoms.1,22 In addition, many studies used rivaroxaban; its once-daily dosing despite a half-life of 9-12 hours may make it a suboptimal choice for patients with APS analogous to the inferiority of once- versus twice-daily low molecular weight heparin in patients with potent thrombophilias. 23 Three of our patients tried both apixaban and rivaroxaban (for insurance reasons) and each found apixaban provided significantly better symptom control.
Loss of efficacy
Most responders experienced a sustained response to their individualized antithrombotic regimen. Our experience suggests a trial of adding anticoagulation should be considered for the occasional patient who develops recurrent/increasing headaches after initially responding to antiplatelet therapy. Loss of efficacy to dual antithrombotic therapy should prompt evaluation for APS valvular disease and consideration of immune modulatory therapy.
Proposed management in APS/aPL patients with refractory migraine who have exhausted other treatment options
As a retrospective study, our data provide only Class IV level of evidence, but they suggest randomized controlled trials are warranted to validate these encouraging findings. Given the thrombotic risk, the disabling nature of refractory migraine, and lack of major bleeding in any of the treatment trials in the present study, we recommend considering a 2-4 week trial (depending on migraine frequency) of aspirin or clopidogrel in APS/aPL patients with migraine and two or more of the clinical features outlined in Table 4 who have exhausted other treatment options. Testing both criteria and non-criteria aPL may increase identification of patients who might benefit from a trial of antithrombotic therapy; 13% of patients had only non-criteria aPL yet they experienced a similar response to treatment. For patients with more than one aPL or high titer positivity of a single aPL, we would consider a trial of antiplatelet therapy while waiting to perform confirmatory testing in 3 months; whereas patients with only a single low titer aPL, we would recommend confirming persistent positivity in 3 months before considering a trial of antiplatelet therapy.
Both aspirin and clopidogrel irreversibly inhibit platelet activation and aggregation, but they act by different mechanisms. Our data showed response to clopidogrel was significantly greater than response to aspirin, so it is reasonable to consider a trial first of clopidogrel in most patients. Both agents take 4-5 days to achieve maximal platelet inhibition. Clopidogrel is considered as safe as aspirin and is not associated with gastrointestinal toxicity. 24 The most highly responsive patients with daily migraine noticed improvement in 2–3 days. If there is no response to aspirin or clopidogrel, we recommend switching to the other agent. Aspirin is usually started at 81 mg daily, but if no response after 1–2 weeks, we increase the dose to 325 mg daily as a few patients only responded to the higher dose. Clopidogrel is usually dosed at 75 mg daily.
If there is a definite, but incomplete response to antiplatelet therapy, we consider a 2 to 4-week trial of adding apixaban 5 mg bid. Apixaban reversibly inhibits factor Xa, does not require laboratory monitoring and has few drug interactions. It is associated with a significantly lower risk for major, fatal and intracranial bleeding than warfarin.23,25 It has a rapid onset of action (peak effect by 2 hours) and highly responsive patients with daily migraine may improve after one day. Patients with a history of venous thrombosis (for whom anticoagulation was not maintained) might consider a trial of anticoagulation first. We recommend maintaining the least aggressive, most effective regimen. If there is not a definite significant response, treatment should be stopped.
It is well recognized that there is a significantly increased risk for ischemic stroke in patients with migraine with aura (O.R. 4.3 in women; O.R. 3.6 in men in the most recent study published in 2020). 26 There may also be an increased risk for hemorrhagic stroke in patients with migraine, however. A large study published in 2013 27 found an increased risk (adjusted H.R. of 2.13 for any hemorrhagic stroke in all patients with migraine) while another large study published in 2014 28 found a small increased risk of intracerebral hemorrhage (but not subarachnoid hemorrhage) only in women with migraine with aura (O.R. 1.6). The decision whether to continue an antithrombotic regimen long-term in responsive patients must be individualized, weighing the potential risks and benefits. In general, we hypothesize that patients with aPL whose migraine headaches are clearly responsive to antithrombotic therapy are likely at greater risk for clotting than bleeding, but there is always a risk of bleeding in any patient treated with antithrombotic therapy.
Study limitations
The present study has several limitations, most importantly its retrospective and single provider design. Assessment of treatment response was based on patient report not headache diaries, but many patients had daily or near daily headaches and treatment trials were short, so recall bias was likely small. Additionally, 74% of patients had dysautonomia and might be different from other aPL-positive populations. Our data, however, are similar to those reported by prior investigators 10 and the antithrombotic response was not different in patients with/without autonomic disease.
Conclusions
Our data confirm and expand upon older reports suggesting a high rate of sustained symptomatic response to antithrombotic therapy in patients with refractory migraine and aPL, including those who do not meet Sydney criteria for a diagnosis of “definite APS” and those with only non-criteria aPL. For APS/aPL patients with refractory migraine who have exhausted other treatment options, a 2-4 week trial of antithrombotic therapy might be considered, usually starting with antiplatelet therapy. In addition to the disabling nature of refractory migraine, these patients are at an increased thrombotic risk, there were no bleeding events during any of the 2-4 week treatment trials in this study and our follow up data suggest an individualized symptom-derived antithrombotic regimen can be maintained in responsive patients with a low bleeding risk. As a retrospective study, our data provide only Class IV level of evidence, but they suggest randomized controlled trials are warranted to validate these encouraging findings.
Footnotes
Declaration of conflicting interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: J.R. Schofield received an honorarium from Inova Laboratories to create an educational video on the antiphospholipid syndrome. M. Birlea, H.N. Hughes and K.L. Hassell have nothing to disclose.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
