Headache and migraine
Everything in the library studied for it, medical and holistic together. Within each kind, what was shown to help comes first, strongest evidence first. What was tested and not shown to help stays on the list.
Strong several reviews agree · Moderate one review or consistent trials · Limited one or a few small studies · Untested no study in people, listed with what it rests on
Medication
Preventing migraine with the antibodies
Shown to reduce migraine by about 1.6 days a month more than placebo, a small effect that a meta-analysis rated high certainty.
Compared with placebo injections or infusions [1][2][3]
The numbers and sources
- Effect
- Across 12 trials, the antibodies reduced monthly migraine days by 1.64 more than placebo (95% CI 1.28 to 1.99), a small effect size (Cohen's d -0.25), with high certainty [1]. A network meta-analysis of 16 trials in episodic migraine found the largest effect for galcanezumab 240 mg (SMD 0.50) and the highest odds of a 50% drop in migraine days for fremanezumab 225 mg (OR 3.17) [2]. In a likelihood to help or harm analysis of phase 3 trials, every antibody dose was more likely to help than to cause stopping for side effects [3].
- Population
- 12 antibody trials in adults [1]; 9,123 participants in 16 trials of episodic migraine [2]; phase 3 trials in episodic and chronic migraine [3].
- Sources
- [1], [2], [3]
- Caveats
- Trials mostly ran 3 to 6 months. The rankings between antibodies come from indirect comparison [2].
Treating a migraine attack with gepants
Shown to leave about 20% pain free 2 hours after rimegepant or ubrogepant, against 11% to 13% on placebo, about 1 extra pain-free person in 10 treated.
Compared with placebo [8][9][10]; triptans, indirectly [10]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Two hours after rimegepant 75 mg, 19.8% were pain free against 10.7% on placebo (RR 1.93, number needed to treat 10), and 14.2% against 6.3% stayed pain free through 24 hours (number needed to treat 11) [8]. On ubrogepant, 20.8% were pain free at 2 hours against 12.6% (RR 1.65), and 37.3% against 27.6% were free of their most bothersome symptom [9]. A network meta-analysis found rimegepant and ubrogepant beat placebo, and that most triptans had higher odds of pain freedom than either [10].
- Population
- 6,734 adults in 7 rimegepant trials [8]; 3,326 adults in 3 ubrogepant trials [9]; 46,442 participants in 64 trials [10].
- Sources
- [8], [9], [10]
- Caveats
- Each trial treated a single attack. The comparison with triptans was indirect [10].
Migraine prevention in adults
Shown to cut migraine frequency by about 1.2 attacks a month against placebo in adults, with about 4 treated for one extra person whose attacks halved.
Compared with placebo [1][2][3]; amitriptyline and other preventive drugs, in network comparisons [2]
The numbers and sources
- Effect
- Across nine trials, topiramate reduced headache frequency by 1.20 attacks per 28 days more than placebo (95% CI -1.59 to -0.80) and doubled the share of people whose attacks halved (RR 2.02, 95% CI 1.57 to 2.60), a number needed to treat of 4 (95% CI 3 to 6). Effects were similar at 50, 100 and 200 mg [1]. A network meta-analysis found 1.1 fewer headaches a month than placebo (95% CI -1.9 to -0.73) and ranked amitriptyline above topiramate [2]. In the two label trials, monthly migraines fell by 2.1 on 100 mg against 0.8 and 1.1 on placebo, from about 5.5 at baseline [3].
- Population
- 1,737 adults in nine trials for frequency and 1,190 for responders [1]; randomized trials of adults with migraine [2]; 937 people aged 12 to 70 in two 26-week trials [3].
- Sources
- [1], [2], [3]
- Caveats
- The network comparison with amitriptyline is indirect, and its abstract calls the evidence for amitriptyline's superiority weak [2].
Stopping a migraine attack
Shown to stop a migraine attack, with 28% pain free 2 hours after oral sumatriptan 50 mg against 11% on placebo.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Across Cochrane reviews of sumatriptan, the oral 50 mg dose left 28% pain free at 2 hours against 11% on placebo, a number needed to treat of 6.1 (95% CI 5.5 to 6.9) in 6,447 participants. For headache relief at 2 hours the number needed to treat was 3.5 (3.2 to 3.7) at oral 100 mg. Treating early, while pain was mild, gave better results than waiting [1][2]. In a meta-analysis of 53 trials, sumatriptan 100 mg gave a headache response at 2 hours in 59%, pain freedom at 2 hours in 29% and sustained pain freedom in 20% [3]. A network meta-analysis of 137 trials found every triptan superior to placebo for pain freedom at 2 hours, with odds ratios from 1.73 (95% CI 1.27 to 2.34) for naratriptan to 5.19 (4.25 to 6.33) for eletriptan [4].
- Population
- 52,236 participants across Cochrane reviews of sumatriptan by all routes [1]; 37,250 participants in 61 studies of oral sumatriptan [2]; 24,089 patients in 53 trials of oral triptans [3]; 89,445 participants in 137 trials of 17 drugs [4].
- Sources
- [1], [2], [3], [4]
- Caveats
- Trials treat a single attack or a few. Twelve of the 53 trials in the earlier meta-analysis were unpublished [3].
Differences between triptans
Shown to differ within the class, with eletriptan and rizatriptan ranking above sumatriptan for pain freedom at 2 hours and naratriptan lowest.
The numbers and sources
- Effect
- In head-to-head comparisons within a network of 137 trials, eletriptan had the highest odds of pain freedom at 2 hours, followed by rizatriptan, sumatriptan and zolmitriptan, and naratriptan had the lowest odds against placebo (OR 1.73) [4]. Against sumatriptan 100 mg, an earlier meta-analysis found better pain relief with rizatriptan 10 mg, better pain relief and lower tolerability with eletriptan 80 mg, and lower pain relief with better tolerability for naratriptan 2.5 mg [3]. A third network analysis found the most favorable results for injected sumatriptan, orally dissolving rizatriptan and zolmitriptan, and eletriptan tablets [6].
- Population
- 89,445 participants in 137 trials [4]; 24,089 patients in 53 trials, 22 of them comparing triptans directly [3]; 133 randomized trials [6].
- Sources
- [4], [3], [6]
- Caveats
- Many rankings rest on indirect comparison through placebo. Confidence in the network estimates ranged from high to very low [4].
Tension-type headache
Shown to leave slightly more adults pain free 2 hours after 1,000 mg for a tension-type headache, with 22 treated for one extra person pain free.
The numbers and sources
- Effect
- For being pain free at 2 hours, the number needed to treat with paracetamol 1,000 mg against placebo was 22 (95% CI 15 to 40) in 8 studies with 5,890 participants, with no difference from placebo at 1 hour. For being pain free or in mild pain at 2 hours it was 10 (95% CI 7.9 to 14). Rescue medication use was lower than on placebo. On limited data, 500 to 650 mg was not superior to placebo, and 1,000 mg did not differ from ketoprofen 25 mg or ibuprofen 400 mg. Adverse events did not differ from placebo (RR 1.1, 95% CI 0.94 to 1.3) [4]. An overview of systematic reviews rated this evidence moderate quality (pain free at 2 hours RR 1.3, 95% CI 1.1 to 1.4) [5].
- Population
- 8,079 adults with frequent episodic tension-type headache and moderate or severe pain at the start, in 23 studies [4]; 36 systematic reviews covering 44 painful conditions [5].
- Sources
- [4], [5]
- Caveats
- The overview draws on the same Cochrane review. Comparisons with ketoprofen and ibuprofen were rated low quality [4].
Acute migraine
Shown to leave 19% of adults pain free 2 hours after 1,000 mg for a migraine, against 10% on placebo.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Taken for moderate to severe migraine pain, paracetamol 1,000 mg left 19% pain free at 2 hours against 10% on placebo (number needed to treat 12). Headache relief at 2 hours was 56% against 36% (5.0) and at 1 hour 39% against 20% (5.2). Paracetamol 1,000 mg plus metoclopramide 10 mg did not differ significantly from oral sumatriptan 100 mg for headache relief at 2 hours [6]. In 3 trials of the over-the-counter combination of acetaminophen, aspirin and caffeine, pain fell to mild or none at 2 hours in 59.3% of 602 treated patients and 32.8% of 618 on placebo, and 50.8% against 23.5% were pain free at 6 hours [7].
- Population
- 2,942 participants and 5,109 attacks in 11 studies [6]; 1,220 people with migraine in 3 trials [7].
- Sources
- [6], [7]
- Caveats
- The combination trials tested acetaminophen with aspirin and caffeine, not acetaminophen alone [7]. The review authors noted the number needed to treat of 12 is higher than for other commonly used pain relievers [6].
Migraine prevention in adults
Shown to reduce migraine frequency against placebo in adults in a network meta-analysis, which ranked it above several other preventive drugs on weak evidence.
Compared with placebo [7]; candesartan, fluoxetine, propranolol, topiramate and valproate, in network comparisons [7]
The numbers and sources
- Effect
- Amitriptyline reduced episodic migraine against placebo (SMD -1.2, 95% CI -1.7 to -0.82). The network analysis found it better than candesartan, fluoxetine, propranolol, topiramate and valproate, and no different from atenolol, flunarizine, clomipramine or metoprolol [7].
- Population
- Adults with migraine in randomized trials, including 11 tricyclic antidepressant trials against placebo [7].
- Sources
- [7]
- Caveats
- The authors describe the evidence for amitriptyline's superiority over other drugs as weak [7].
Tension-type headache
Shown to cut chronic tension-type headaches by about 4.8 a month against placebo with tricyclic antidepressants, which also reduced headaches more than SSRIs.
Compared with placebo, SSRIs and behavioral therapy [9]
The numbers and sources
- Effect
- Tricyclic antidepressants reduced headache frequency against placebo (WMD -4.8 headaches a month, 95% CI -6.63 to -2.95) and painkiller use (WMD -21.0 doses a month). They reduced headaches more than SSRIs and were similar to behavioral therapy in a few trials [9].
- Population
- Adults in 22 randomized trials, 8 comparing a tricyclic with placebo [9].
- Sources
- [9]
- Caveats
- The pooled result is for tricyclics as a group, and the abstract does not separate amitriptyline [9].
Migraine prevention
Shown to lower episodic migraine by about 1.5 headaches a month on propranolol against placebo, with 40% against 23% halving their migraine days in one trial.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Across placebo-controlled trials, propranolol lowered episodic migraine by 1.5 headaches a month at 8 weeks (95% CI -2.3 to -0.65) and made a 50% reduction more likely (RR 1.4, 95% CI 1.1 to 1.7). For chronic migraine the estimate was RR 2.0 (95% CI 1.0 to 4.3) [15]. In a crossover trial, migraine days per 4 weeks were 2.91 on slow-release propranolol 160 mg and 3.53 on placebo, and 40% against 23% had at least a 50% reduction [16].
- Population
- 108 randomized trials of beta-blockers for headache prevention, 50 placebo-controlled [15]; 72 adults with episodic or chronic migraine [16].
- Sources
- [15], [16]
- Caveats
- Most of the evidence is for propranolol. Trials comparing beta-blockers with other preventive drugs were mostly single and underpowered [15].
Preventing migraine with gepants
Shown to reduce monthly migraine days with atogepant, with a 50% responder rate about 1.7 times placebo's. Rimegepant every other day cut 0.8 more days than placebo.
The numbers and sources
- Effect
- Atogepant reduced monthly migraine days (SMD -0.40, 95% CI -0.46 to -0.34) and acute medication days (SMD -0.45), and raised the share with a 50% drop in migraine days (RR 1.66, 95% CI 1.46 to 1.89), with no added benefit at higher doses [5]. Rimegepant 75 mg every other day reduced monthly migraine days by 4.3 against 3.5 on placebo over weeks 9 to 12 [6].
- Population
- 2,813 adults in 4 atogepant trials [5]; 695 adults with episodic migraine analyzed in one 12-week rimegepant trial [6].
- Sources
- [5], [6]
- Caveats
- One meta-analysis for atogepant and one trial for rimegepant prevention.
Migraine returning after emergency treatment
Shown to lower headache recurrence within 72 hours of emergency migraine treatment, with 9 treated to prevent one, using one injected dose of dexamethasone.
The numbers and sources
- Effect
- Added to standard migraine treatment, a single dose of dexamethasone did not change pain relief at the time (difference 0.37, 95% CI -0.20 to 0.94) and lowered recurrence within 72 hours (RR 0.74, 95% CI 0.60 to 0.90), a number needed to treat of 9. Side effects were similar to placebo [15].
- Population
- Adults with acute migraine in 7 randomized trials [15].
- Sources
- [15]
- Caveats
- The trials gave dexamethasone by injection, not by mouth.
Chronic migraine
Shown to cut monthly migraine days modestly against placebo in chronic migraine, with the most dropouts for side effects among the treatments compared.
Compared with placebo, botulinum toxin and acupuncture [4]
The numbers and sources
- Effect
- In a network meta-analysis, topiramate lowered monthly migraine days more than placebo (SMD -0.4, 95% CI -0.75 to -0.04) and did not differ significantly from botulinum toxin. It caused the most treatment-related adverse events and the highest rate of dropouts for adverse events [4].
- Population
- 2,545 people in 15 trials, 582 in the topiramate comparisons [4].
- Sources
- [4]
- Caveats
- One network meta-analysis, with few direct comparisons between treatments [4].
Injected sumatriptan for migraine
Shown to leave 59% pain free 2 hours after a 6 mg sumatriptan injection against 15% on placebo, a larger effect than by mouth.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- A 6 mg injection under the skin left 59% pain free at 2 hours against 15% on placebo, a number needed to treat of 2.3 (95% CI 2.1 to 2.4) in 2,522 participants, and acted faster than other routes [1]. Numbers needed to treat were 2.9 for pain freedom at 1 hour, 2.1 for headache relief at 2 hours and 6.1 for sustained pain freedom at 24 hours. A second injection after a poor response to the first gave no added relief [5].
- Population
- 9,365 participants in 35 studies of injected sumatriptan [5].
- Sources
- [1], [5]
- Caveats
- Adverse events were more frequent with the injection than with other routes [1].
Compared with NSAIDs, acetaminophen and newer migraine drugs
Shown to match or exceed NSAIDs, aspirin and acetaminophen for headache relief at 2 hours, and to outperform lasmiditan, rimegepant and ubrogepant in a network analysis.
The numbers and sources
- Effect
- Standard-dose triptans relieved headache within 2 hours in 42% to 76% of patients, against 46% to 52% for NSAIDs, aspirin and acetaminophen and 38% for ergots [6]. In a network of 137 trials, eletriptan, rizatriptan, sumatriptan and zolmitriptan were each more likely to give pain freedom at 2 hours than lasmiditan, rimegepant or ubrogepant. For sustained pain freedom to 24 hours, eletriptan and ibuprofen ranked highest [4]. Paracetamol 1,000 mg plus metoclopramide 10 mg did not differ significantly from oral sumatriptan 100 mg for headache relief at 2 hours [7].
- Population
- 133 randomized trials [6]; 89,445 participants in 137 trials [4]; 2,942 participants in 11 studies of paracetamol [7].
- Sources
- [6], [4], [7]
- Caveats
- Mostly indirect comparisons across trials with different patients and placebo rates.
Sumatriptan taken with naproxen
Shown to leave 28% pain free at 2 hours on sumatriptan plus naproxen against 8% on placebo, more than either drug alone.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- When starting pain was moderate or severe, sumatriptan 85 mg or 50 mg plus naproxen 500 mg left 28% pain free at 2 hours against 8% on placebo (number needed to treat 4.9, 95% CI 4.3 to 5.7). When starting pain was mild the figures were 50% and 18% (3.1, 95% CI 2.9 to 3.5). Where trials compared directly, the combination did better than sumatriptan alone or naproxen alone, and adverse events were less frequent on naproxen than on sumatriptan [8].
- Population
- 13 studies, with 3,663 participants given the combination, 3,682 placebo, 964 sumatriptan and 982 naproxen [8].
- Sources
- [8]
- Caveats
- Adverse events were more frequent on the combination than on placebo and rarely led to withdrawal [8].
Cluster headache attacks
Shown to leave 48% pain free 15 minutes after a 6 mg sumatriptan injection for a cluster headache attack, against 17% on placebo.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Fifteen minutes after sumatriptan 6 mg under the skin, 48% were pain free and 75% had no or mild pain, against 17% and 32% on placebo (numbers needed to treat 3.3 and 2.4). Zolmitriptan 10 mg by nasal spray left 12% pain free and 28% with no or mild pain, against 3% and 7% on placebo (11 and 4.9) [9].
- Population
- 6 studies of single attacks, with 219 participants given injected sumatriptan, 454 given zolmitriptan and 326 given placebo [9].
- Sources
- [9]
- Caveats
- Six small studies, each treating one attack.
Migraine in children and adolescents
Shown to raise pain freedom in adolescents with migraine, with 6 treated for one extra pain free, and 13 in younger children.
The numbers and sources
- Effect
- Triptans as a class raised pain freedom against placebo in adolescents (RR 1.32, 95% CI 1.19 to 1.47, number needed to treat 6) across 21 studies, and in children (RR 1.67, 95% CI 1.06 to 2.62, number needed to treat 13) across 3 studies. Minor adverse events were more frequent in adolescents on triptans (risk difference 0.13, 95% CI 0.08 to 0.18) [10].
- Population
- 7,026 adolescents in 21 studies and 273 children in 3 studies, from 27 trials with mean ages of 8.2 to 14.7 years [10].
- Sources
- [10]
- Caveats
- The reviewers rated the evidence moderate quality. The sumatriptan tablet label covers adults only [11].
Preventing menstrual migraine
Shown to raise the share of women free of menstrual migraine when frovatriptan was taken around the period, with about 4 to 7 treated for one extra.
The numbers and sources
- Effect
- Taken for 5 to 6 days starting 2 to 3 days before the expected headache, frovatriptan 2.5 mg once a day raised the chance of a period free of menstrual migraine against placebo (RR 1.48, 95% CI 1.27 to 1.72, number needed to treat 7.22), and twice a day more so (RR 1.82, 95% CI 1.58 to 2.09, number needed to treat 3.90). Numbers needed to treat for zolmitriptan 2.5 mg were 4.98 twice a day and 2.52 three times a day [12].
- Population
- Women with menstrual migraine in 6 randomized trials: 1,217 given frovatriptan, 462 naratriptan, 163 zolmitriptan and 1,104 placebo [12].
- Sources
- [12]
- Caveats
- All participants were women, with mean ages of 36 to 38. No trial compared one triptan with another [12]. The sumatriptan label states it is not indicated for preventing migraine [11].
Migraine prevention
One small crossover trial found lisinopril lowered days with migraine by 21% against placebo in adults with 2 to 6 attacks a month.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Over 12-week periods, lisinopril lowered hours with headache by 20% (95% CI 5 to 36), days with headache by 17% (95% CI 5 to 30) and days with migraine by 21% (95% CI 9 to 34) against placebo. Days with migraine fell by at least half in 14 of the 47 participants with complete data [17].
- Population
- 60 adults aged 19 to 59 with 2 to 6 migraine episodes a month, 47 with complete data [17].
- Sources
- [17]
- Caveats
- One small trial from one clinic, of lisinopril only. Not an FDA-approved use [3].
Migraine prevention
Small trials found candesartan lowered migraine days against placebo, with about 40% of people having at least half as many.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Over 12 weeks, days with headache were 13.6 on candesartan 16 mg and 18.5 on placebo, and days with migraine 9.0 against 12.6. Migraine days fell by at least half in 23 of 57 people (40.4%) [21]. In a second crossover trial, migraine days per 4 weeks were 2.95 on candesartan, 2.91 on propranolol and 3.53 on placebo. The share with at least a 50% reduction was 43% on candesartan, 40% on propranolol and 23% on placebo [22].
- Population
- 60 adults with 2 to 6 migraine attacks a month, 57 analyzed [21]; 72 adults with episodic or chronic migraine [22].
- Sources
- [21], [22]
- Caveats
- Two small crossover trials from the same Norwegian group, both of candesartan. Not an FDA-approved use.
Cluster headache prevention
One small trial found verapamil lowered attack frequency in episodic cluster headache against placebo over 14 days.
The numbers and sources
- Effect
- Over 14 days, the verapamil group had a significant reduction in attack frequency and in use of medication to stop attacks, against placebo [13].
- Population
- 30 people with episodic cluster headache, 15 on verapamil 120 mg three times a day and 15 on placebo [13].
- Sources
- [13]
- Caveats
- One small trial, and the abstract gives no effect size. The evidence is for verapamil, a non-dihydropyridine. Not an FDA-approved use.
Atogepant against topiramate
One trial found 64% halved their migraine days on atogepant against 39% on topiramate, and 12% stopped for side effects against 30%.
Compared with topiramate at the highest tolerated dose of 50 to 100 mg a day [7]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Over 24 weeks, 12% on atogepant 60 mg and 30% on topiramate stopped for adverse events (RR 0.4, 95% CI 0.3 to 0.6). A 50% drop in monthly migraine days occurred in 64% against 39% (RR 1.6), and migraine days fell 6.3 against 4.5. One anaphylactic reaction on atogepant was judged drug related [7].
- Population
- 540 adults with at least 4 migraine days a month in 12 countries, 89% women [7].
- Sources
- [7]
- Caveats
- One trial. Topiramate's dose was limited to what each person tolerated [7].
Episodic cluster headache
One trial found galcanezumab 300 mg cut weekly cluster attacks by 8.7 against 5.2 on placebo, with 71% responding against 53%.
Compared with placebo [11]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Over weeks 1 to 3, weekly cluster headache attacks fell by 8.7 on galcanezumab and 5.2 on placebo, from about 17.5 a week. At week 3, 71.4% against 52.6% had at least a 50% drop in attacks [11].
- Population
- 106 adults with episodic cluster headache, 88 men, in an 8-week trial [11].
- Sources
- [11]
- Caveats
- One small trial, described in the label.
Chronic daily headache
One trial found tizanidine lowered a headache index by 54% against 19% on placebo over 12 weeks in adults with chronic daily headache.
The numbers and sources
- Effect
- In the last 4 weeks of treatment, the headache index improved 54% on tizanidine and 19% on placebo, severe headache days 55% against 21%, and headache duration 35% against 19%. Total headache days improved 30% against 22%, which was not significant. Sleepiness was reported by 47% on tizanidine, dizziness by 24% and dry mouth by 23% [11].
- Population
- 134 adults randomized after a placebo run-in, 77% with migraine, of whom 85 completed 12 weeks [11].
- Sources
- [11]
- Caveats
- One trial with many dropouts, at a mean dose of 18 mg a day. Not an FDA-approved use [12].
Acute migraine
One trial found tramadol with acetaminophen relieved migraine pain at 2 hours in 55.8% against 33.8% on placebo, without easing nausea.
The numbers and sources
- Effect
- Treatment response at 2 hours was 55.8% on tramadol with acetaminophen and 33.8% on placebo, and 22.1% against 9.3% were pain free. At 24 hours 52.7% against 37.9% were pain free. Sensitivity to light and sound was less frequent at 2 hours. Migraine-related nausea did not differ (38.5% against 29.4%) [14].
- Population
- 305 adults with migraine, 154 on tramadol with acetaminophen and 151 on placebo [14].
- Sources
- [14]
- Caveats
- One trial of a single attack, of the tramadol and acetaminophen combination only. See the risk block on headache frequency below.
Antibodies against older preventive drugs
Mixed results: in one trial erenumab halved migraine days in 55% against 31% on topiramate, but a network analysis found no significant difference from topiramate or divalproex.
Compared with topiramate [4]; topiramate and divalproex, indirectly through placebo [1]; topiramate, propranolol and onabotulinumtoxinA, indirectly [3]
The numbers and sources
- Effect
- Over 24 weeks, 10.6% stopped erenumab for adverse events against 38.9% on topiramate (OR 0.19, 95% CI 0.13 to 0.27), and 55.4% against 31.2% had at least a 50% drop in monthly migraine days (OR 2.76, 95% CI 2.06 to 3.71) [4]. A network meta-analysis found no significant difference in migraine days between the antibodies and topiramate (WMD -0.19, 95% CI -0.56 to 0.17) or divalproex [1]. The antibodies had higher help-to-harm ratios than propranolol, topiramate or onabotulinumtoxinA [3].
- Population
- 777 adults with at least 4 migraine days a month in Germany [4]; trials of antibodies, topiramate and divalproex [1].
- Sources
- [4], [1], [3]
- Caveats
- One head-to-head trial, in which topiramate was dosed to the level each person tolerated [4]. The network comparison was indirect [1].
Migraine prevention
Mixed results: a Cochrane review found lamotrigine no better than placebo for migraine frequency, while a network meta-analysis listed it among drugs with fewer than three positive trials.
Compared with placebo [10][11]
The numbers and sources
- Effect
- In placebo-controlled trials, lamotrigine did not lower headache frequency per 28 days and did not raise the share of responders [10]. A network meta-analysis of migraine preventive drugs listed lamotrigine among drugs that beat placebo in fewer than three trials, without giving its effect size in the abstract [11].
- Population
- Adults with episodic migraine in 10 trials of nine seizure drugs, lamotrigine among them [10]; adults in randomized trials of migraine preventive drugs [11].
- Sources
- [10], [11]
- Caveats
- Few lamotrigine trials exist, and neither abstract gives their number or size.
Migraine prevention in children and teenagers
Small trials found mixed results in children and teenagers: one label trial found fewer attacks at 100 mg, while a larger trial found no difference from placebo.
Compared with placebo [5][3]; amitriptyline [5]
The numbers and sources
- Effect
- In the trial behind the approval for ages 12 to 17, 100 mg reduced the monthly attack rate by 28% more than placebo, with a mean drop of 3.0 against 1.7 attacks [3]. In a later trial in ages 8 to 17, stopped early for futility, headache days halved in 55% on topiramate, 52% on amitriptyline and 61% on placebo, with no difference in disability [5].
- Population
- 103 adolescents aged 12 to 17 over 16 weeks [3]; 361 children and adolescents aged 8 to 17 over 24 weeks, 328 analyzed [5].
- Sources
- [3], [5]
- Caveats
- The placebo response in the later trial was high. Paresthesia (31% against 8%) and weight loss (8% against 0%) were more common on topiramate [5].
Migraine prevention in children and teenagers
One trial in children and teenagers found amitriptyline no better than placebo for migraine, and it raised fatigue and dry mouth.
Compared with placebo and topiramate [8]
The numbers and sources
- Effect
- Headache days halved in 52% on amitriptyline, 55% on topiramate and 61% on placebo, with no differences in disability or headache days. The trial was stopped early for futility. Fatigue (30% against 14%) and dry mouth (25% against 12%) were more common on amitriptyline [8].
- Population
- 361 children and adolescents aged 8 to 17, 328 analyzed, over 24 weeks [8].
- Sources
- [8]
- Caveats
- One trial, with a high placebo response [8].
Supplements
Migraine frequency
Shown to reduce migraine frequency and attack duration in adults with migraine, across 6 small studies. Severity did not change.
Migraine
Limited evidence: added to a preventive drug, ginger lowered headache severity at 3 months. Reviewers call any acute or preventive benefit too uncertain.
The numbers and sources
- Effect
- A 2026 review found 8 RCTs (594 patients). Added to an active drug for prevention, ginger lowered headache severity at 3 months (MD -3.18, 95% CI -3.94 to -2.42, 2 studies, 183 participants, low certainty). For acute attacks, comparisons rested on single trials and were rated low or very low certainty [11].
- Population
- 594 patients with migraine in 8 RCTs [11].
- Sources
- [11]
- Caveats
- The authors concluded it is too uncertain whether ginger helps as an acute or preventive treatment. Several trials combined ginger with other herbs or drugs [11].
Therapies
Chronic pain across conditions
Shown to ease chronic pain about 0.5 standard deviations more than no acupuncture and about 0.2 more than sham acupuncture, with effects lasting a year.
Compared with sham acupuncture [2][3]; no acupuncture [2][3]
The numbers and sources
- Effect
- In individual patient data from 39 trials, acupuncture was superior to both sham and no acupuncture for each pain condition (all P < .001). The difference was close to 0.5 standard deviations against no acupuncture and close to 0.2 against sham, and about 15% of the effect was lost at 1 year. Effects were smaller in trials whose sham used a penetrating needle [2]. The earlier analysis of 29 trials found pain scores lower than sham by 0.23 SD for back and neck pain, 0.16 SD for osteoarthritis and 0.15 SD for chronic headache, and lower than no acupuncture by 0.55, 0.57 and 0.42 SD [3].
- Population
- 20,827 patients with nonspecific musculoskeletal pain, osteoarthritis, chronic headache or shoulder pain in 39 trials [2]. 17,922 patients in 29 trials [3].
- Sources
- [2], [3]
- Caveats
- The two analyses come from the same collaboration and the later one contains the earlier trials. Patients and acupuncturists cannot be blinded when the comparison is no acupuncture, so the larger effect includes expectation. The authors say factors beyond needling contribute to the effect [3].
Migraine in adults, compared with a waiting list
Shown to reduce migraine frequency and severity against a waiting list in two meta-analyses, with effects lasting more than a year in follow-up.
Compared with waiting list [2]; control conditions the source abstract does not describe [3]
The numbers and sources
- Effect
- Across 55 studies, all biofeedback approaches together gave a medium effect (d 0.58, 95% CI 0.52 to 0.64) that held over an average 17 months of follow-up, and outcomes improved more with biofeedback than with control conditions. Attack frequency improved most, and blood-volume-pulse feedback gave larger effects than skin temperature or muscle feedback [3]. A 2025 review of 9 trials (558 participants) found biofeedback reduced headache frequency and severity against a waiting list [2].
- Population
- 55 studies, including randomized trials and before-and-after studies [3]. Adults with migraine in 9 randomized trials [2].
- Sources
- [3], [2]
- Caveats
- The 2007 analysis included before-and-after studies with no control group [3]. A waiting list cannot control for attention or expectation.
Migraine prevention
Shown to halve migraine frequency in 41% against 17% with no acupuncture. Against sham acupuncture the difference was small, 50% against 41%.
Compared with no acupuncture or routine care [6]; sham acupuncture [6]; preventive migraine drugs [6]
The numbers and sources
- Effect
- After treatment, headache frequency at least halved in 41% with acupuncture and 17% with no acupuncture (RR 2.40, 95% CI 2.08 to 2.76, 4 trials, 2,519 participants, number needed to treat 4). Against sham, it halved in 50% and 41% (RR 1.23, 95% CI 1.11 to 1.36, 14 trials, 1,825 participants, number needed to treat 11), and frequency fell slightly more (SMD -0.18). Against preventive drugs, frequency at least halved in 57% and 46% at 3 months (RR 1.24) and in 59% and 54% at 6 months, a difference that was no longer significant [6].
- Population
- 22 trials with 4,985 adults with episodic migraine, lasting at least 8 weeks [6].
- Sources
- [6]
- Caveats
- Evidence was rated moderate. The no-acupuncture and drug comparisons could not be blinded. Only one trial followed people past treatment against no acupuncture [6].
Tension-type headache
Shown to halve headache frequency more often than routine care, 48% against 19% in the largest trial. Against sham acupuncture the difference was smaller, 51% against 43%.
Compared with routine care or treatment of acute headaches only [7]; sham acupuncture [7]
The numbers and sources
- Effect
- In the largest trial against routine care, headache frequency at least halved in 302 of 629 (48%) with acupuncture and 121 of 636 (19%) without (RR 2.5, 95% CI 2.1 to 3.0). Against sham, it halved in 205 of 391 (51%) and 133 of 312 (43%) (RR 1.3, 95% CI 1.09 to 1.5, 4 trials). One of 420 people receiving acupuncture dropped out for adverse effects [7].
- Population
- 12 trials with 2,349 adults with episodic or chronic tension-type headache [7].
- Sources
- [7]
- Caveats
- Evidence was rated moderate. The two trials against routine care were unblinded and differed in baseline headache frequency. Effects beyond 4 months were not studied against routine care [7].
Migraine in children, compared with a waiting list
Shown to cut migraine attacks by about 2 a month against a waiting list in children and adolescents, in five small trials.
Compared with waiting list [4]
The numbers and sources
- Effect
- Biofeedback reduced migraine frequency (mean difference -1.97, 95% CI -2.72 to -1.21), attack duration (-3.94) and headache intensity (-1.77) compared with a waiting list [4].
- Population
- 137 children and adolescents in 5 trials [4].
- Sources
- [4]
- Caveats
- Only 40% of risk-of-bias judgments were low, and reporting was often incomplete [4].
Tension-type headache
Shown to reduce tension-type headache with a medium to large effect (d 0.73) that lasted about 15 months, more than headache monitoring, placebo or relaxation.
Compared with headache monitoring, placebo, and relaxation therapies [5]
The numbers and sources
- Effect
- Across 53 studies, biofeedback gave an effect size of 0.73 (95% CI 0.61 to 0.84) that held over an average 15 months of follow-up. Headache frequency improved most, with gains in muscle tension, anxiety, depression and pain medication use. Biofeedback combined with relaxation gave the largest effects, and effects were especially large in children and adolescents [5].
- Population
- 53 studies of tension-type headache [5].
- Sources
- [5]
- Caveats
- One meta-analysis from one research group. The abstract does not say how many studies were randomized [5].
Migraine, compared with other control conditions
Shown to reduce migraine days by a small amount in a meta-analysis of 5 trials, a result the authors called preliminary.
Compared with control conditions the source abstract does not describe [8]
The numbers and sources
- Effect
- Spinal manipulation reduced migraine days (Hedges' g -0.35, 95% CI -0.53 to -0.16) and migraine pain or intensity. One trial in chronic migraine was left out of the pooled analysis because it made heterogeneity high [8].
- Population
- 677 people in 6 trials, with treatment lasting 2 to 6 months [8].
- Sources
- [8]
- Caveats
- Several trials were at high or unclear risk of bias for blinding, compliance and missing data [8].
Cervicogenic headache
Limited evidence: two low-quality trials suggested several sessions of neck manipulation eased cervicogenic headache pain more than massage.
Compared with massage [7]
The numbers and sources
Headache
Mixed results: headache pain did not differ from other treatments in the short term, while headache-related disability improved more in tension-type and cervicogenic headache.
Compared with other interventions [16]
The numbers and sources
- Effect
- Short-term headache pain did not differ from other interventions in tension-type headache (SMD -1.27, 95% CI -3.56 to 1.03, 230 participants), cervicogenic headache (SMD -0.41, 95% CI -4.69 to 3.87) or mixed headache including migraine (SMD 0.03, 95% CI -0.42 to 0.48). Disability improved more in tension-type headache (SMD -2.28, 95% CI -2.66 to -1.91) and cervicogenic headache (SMD -0.72, 95% CI -1.09 to -0.34) [16].
- Population
- 11 trials, 9 pooled, of people with tension-type, cervicogenic or migraine headache [16].
- Sources
- [16]
- Caveats
- Only 4 trials were of high quality, and the evidence on pain was rated very low quality [16].
Headache, irritable bowel syndrome and other conditions
Mixed results: an overview of 27 reviews found possible relief in migraine and irritable bowel syndrome, and no relief in acute low back pain, fibromyalgia or other headaches.
Compared with no treatment, sham treatment or other non-osteopathic treatment [16]
The numbers and sources
- Effect
- Osteopathic treatment may relieve symptoms in adults with migraine and irritable bowel syndrome. It did not appear to relieve acute nonspecific low back pain, fibromyalgia, headaches of various causes, lower urinary tract symptoms in women, or cerebral palsy in children. Evidence for vertigo, COPD, high blood pressure and gynecological complaints was insufficient [16]. An earlier overview rated the evidence for children's conditions, headache and irritable bowel syndrome limited and inconclusive [17].
- Population
- 27 systematic reviews covering nine indications [16]. 9 reviews with 55 trials and 3,740 participants [17].
- Sources
- [16], [17]
- Caveats
- Included reviews were rated low or critically low in quality [17], and the trials were heterogeneous and partly of poor quality [16].
Migraine, compared with drugs or other active treatment
Not shown to reduce migraine frequency more than drug treatment or cognitive behavioral therapy in adults, or more than other active treatment in children.
Compared with pharmacotherapy or cognitive behavioral therapy [2]; other active treatment and other behavioral treatment [4]
The numbers and sources
- Effect
- In 3 trials (278 participants), headache frequency did not differ between biofeedback and active treatments such as drugs or CBT [2]. In children, biofeedback added nothing when combined with other behavioral treatment and had no significant advantage over active treatment [4].
- Population
- Adults in 3 trials [2]. Children and adolescents in 5 trials [4].
- Sources
- [2], [4]
- Caveats
- Few trials, and the 2025 review authors also report a possible added benefit when combined with drugs [2].
Migraine, compared with sham manipulation
Not shown to reduce migraine days more than sham manipulation. Migraine days fell in both groups, and the trial authors attributed the effect to a placebo response.
Compared with sham manipulation [9][10]; usual drug treatment [9]
The numbers and sources
- Effect
- In 104 people with migraine, migraine days fell from baseline in the manipulation, sham and usual care groups, and the reduction did not differ between groups. Migraine duration and headache index fell more with manipulation than with usual care by the end of follow-up [9]. A systematic review of high- and acceptable-quality trials found manipulation was not superior to sham for migraine [10].
- Population
- 104 adults with at least one migraine a month in Norway [9]. 6 trials with 534 participants across several conditions [10].
- Sources
- [9], [10]
- Caveats
- One trial of migraine passed the review's quality bar [10]. Blinding was sustained throughout the trial [9].
Practices
Acute migraine
Shown to relieve an acute migraine attack compared with sham in a Cochrane review of 3 small trials, at low quality. It did not prevent attacks.
The numbers and sources
- Effect
- Pooled data from 3 trials (58 participants) found hyperbaric oxygen more likely than sham to relieve a migraine headache (risk ratio 6.21, 95% CI 2.41 to 16.00). There was no evidence that it prevented migraine episodes, reduced nausea and vomiting, or reduced the need for rescue medication [9].
- Population
- 58 participants in 3 trials, within a review of 11 trials and 209 participants covering oxygen therapies for migraine and cluster headache [9].
- Sources
- [9]
- Caveats
- The reviewers rated the evidence low quality, downgraded for small crossover studies with incomplete reporting, and noted the cost and poor availability of chambers [9].
Chronic tension-type headache
One trial found headache intensity fell 1.27 points more on a pain scale with 8 weeks of regular sauna in 37 people. Headache duration did not change.
The numbers and sources
- Effect
- Mean change in headache intensity on a numerical pain rating scale differed between sauna and control by 1.27 points (95% CI 0.48 to 2.07, p = 0.002). Headache duration, sleep disturbance, depression and headache disability did not differ between groups [20].
- Population
- 37 people with chronic tension-type headache, 17 assigned to sauna plus advice and 20 to advice alone, for 8 weeks [20].
- Sources
- [20]
- Caveats
- One small unblinded trial.
Migraine
We haven't sourced research on this yet.