How Should Migraine Treatment Be Chosen? The Limits of Evidence for Acupuncture, Pain Relievers, and Melatonin

Evidence, Limitations, and Safety in Migraine Treatment

Bottom Line

Migraine treatment should balance potential benefit, uncertainty in the evidence, and the individual’s safety risks. Acupuncture may help some patients, but it cannot be claimed to work invariably or to cause no adverse effects. Nor can a drug be presumed safer merely because it “contains no aspirin.”（PMID: 28241154；PMID: 41219786；PMID: 42245938）

Acupuncture: A Positive Signal Is Not a Settled Verdict

Professor Ching-Shun Lin previously discussed a randomized controlled trial involving migraine without aura. The study compared true acupuncture, sham acupuncture, and a waiting-list group, enrolling 249 participants and scheduling 20 treatment sessions. Its design, sample size, and central conclusion—that the true-acupuncture group might achieve a more durable reduction in recurrence—were all consistent with the professor’s account at the time.（PMID: 28241154）

The result deserves serious consideration, but it should not be reduced to the claim that “acupuncture has been proven effective for every migraine.” A single trial answers a question about particular participants, a particular intervention, and particular comparison conditions. The benefit an individual patient actually receives may still depend on migraine subtype, how the treatment is delivered, and variation in personal response. By citing an editorial comment and emphasizing the need for more persuasive evidence, Professor Lin was appropriately cautious about the strength of the evidence; he was not declaring that acupuncture had no effect.（PMID: 28241154）

The literature has since been updated. Later meta-analyses have continued to find that acupuncture may reduce migraine days and attack frequency compared with sham acupuncture. This gives us more reason to regard acupuncture as an option worth discussing. Nevertheless, a pooled result cannot erase differences in the quality of the included studies, uncertainty in effect size, or variation in treatment technique. The more accurate formulation is: “It may help, but the evidence must still be interpreted in light of its limitations.”（PMID: 41219786）

Acupuncture Safety: Avoid the Absolute Claim of “No Side Effects at All”

Viewed against today’s evidence, the unqualified proposition that “acupuncture has no adverse side effects” does not align with the literature. Studies have explicitly documented mild, temporary adverse reactions. Although the cited safety data did not come from a migraine study, they are sufficient to show that acupuncture cannot be described as risk-free.（PMID: 42245938）

This is a calibration of language made possible by an updated literature and the evolution of science. It preserves the historical context of the professor’s original article, which encouraged readers to consider trying acupuncture after understanding its limitations, while adding the safety information that can now be established. For patients, the practical decision involves more than whether to try it. It also includes who will perform the procedure, how discomfort or other reactions will be managed, and whether trying an adjunctive therapy could delay established medical care.

Acetaminophen: It Relieves Pain and Fever, but Overdose-Related Liver Injury Is a Major Warning

Acetaminophen is expressly identified as an analgesic and antipyretic, and the literature directly addresses acetaminophen-induced liver injury. The professor’s warning—that this common drug can injure the liver when taken in excess—is therefore consistent with the available information.（PMID: 32009106；PMID: 36034775）

However, the literature collected by ECU directly verifies the drug’s pain-relieving and fever-reducing uses and its risk of liver injury. Those titles alone cannot establish the full pharmacological proposition that it “has no anti-inflammatory effect,” nor are they sufficient to construct additional risk strata for people who consume alcohol or already have liver disease. In other words, familiarity does not justify ignoring dosage or personal medical history. At the same time, precise thresholds or risk multipliers should not be added without support from the materials at hand.

NSAIDs: The Direction of Gastrointestinal Risk Is Supported, but Precise Comparisons Require Restraint

Professor Lin classified aspirin and ibuprofen as first-generation NSAIDs and noted that both may cause gastric ulcers. Systematic reviews and meta-analyses do treat NSAID use and gastrointestinal bleeding risk as part of the same important safety issue. The general warning that gastrointestinal injury must be considered during use is therefore supported by the literature.（PMID: 40915652）

The title of the article included in ECU, however, does not independently verify the COX-1 mechanism or the specific endpoint of gastric ulcer, and it does not directly demonstrate that aspirin is necessarily more harmful than ibuprofen. The most defensible conclusion here is to acknowledge the gastrointestinal risks of NSAIDs without presenting the relative severity among drugs as a ranking established by this particular source.（PMID: 40915652）

Selective COX-2 Drugs: Gastrointestinal and Cardiovascular Risks Must Be Considered Together

For selective COX-2 drugs such as celecoxib, the evidence supports evaluating gastrointestinal and cardiovascular effects together. A separate multicenter randomized controlled trial specifically examined the cardiovascular safety of celecoxib. These data support the core direction of the professor’s warning: users should not focus solely on gastric tolerability but should also consider cardiovascular safety.（PMID: 16168077；PMID: 27959716）

The statement that these drugs “do not cause gastric ulcers,” however, is too absolute. The sources listed by ECU also provide no effect size with which to state the magnitude of heart attack or stroke risk. A more appropriate evidence-based formulation is that different drugs may have different safety profiles, and that gastrointestinal and cardiovascular factors should both be weighed when choosing among them. No drug class should simply be treated as free of gastrointestinal risk.（PMID: 16168077；PMID: 27959716）

Melatonin: Do Not Extrapolate from Limited Uses to a Universal Therapy

Melatonin may be effective for certain aspects of sleep, but it does not work for everyone and cannot be regarded as a treatment definitively superior across the board to first-line CBT-I for insomnia. This is consistent with Professor Lin’s observation that it may be used for sleep or jet lag, but that its effects vary from person to person.（PMID: 32580450）

The same evidentiary boundary applies to cancer-related claims. The latest systematic review found insufficient data on key therapeutic outcomes in patients with cancer and a high risk of bias. At present, the evidence does not support describing melatonin as an established cancer treatment or as a therapy that broadly improves quality of life and sleep—still less as a miracle cure.（PMID: 40304216）

Turning Evidence into a Practical Choice

For migraine, the most reliable question is not merely whether a treatment “works or does not work.” It is what outcome the treatment may improve, what it was compared against, how stable the evidence is, and whether adverse effects were fully reported. Acupuncture may be an option after informed discussion, but it should not be presented as risk-free. Common pain relievers may have clearly defined uses, but their hepatic, gastrointestinal, and cardiovascular safety boundaries still matter. Melatonin should not be extrapolated into a universal treatment merely because it is perceived as “natural.”（PMID: 41219786；PMID: 42245938；PMID: 36034775；PMID: 40915652；PMID: 27959716；PMID: 40304216）