Medication overuse headache, or MOH, is one of the most burdensome complications of the world’s most disabling neurological conditions. Migraine affects roughly 1.1 billion people and ranks second globally among causes of years lived with disability. Most patients experience it episodically, but in a subset the attacks grow more frequent until the headache-free baseline erodes entirely. Under the International Classification of Headache Disorders, third edition, MOH is diagnosed when a person with a pre-existing primary headache develops headache on 15 or more days per month while overusing acute medications, triptans, ergots, opioids or combination analgesics on 10 or more days per month, or simple analgesics on 15 or more days per month, for more than three months. It affects roughly 1 to 2.5 percent of adults in the general population, but accounts for 30 to 50 percent of patients presenting with chronic daily headache at specialized centers, and women are affected three to four times more often than men.

Sleep complaints are nearly universal in this population. In cohorts of MOH patients assessed with validated insomnia instruments, prevalence ranges from roughly 38 to 60 percent, and comorbid depression or anxiety is present in approximately 40 to 50 percent of cases. When disordered sleep and MOH coexist, both pain severity and treatment failure rates increase, and even after successful detoxification, one-year relapse rates run between 20 and 40 percent. Understanding why requires examining the direction of the relationship, and that is where the review turns to genetics. Using Mendelian randomization, a technique that employs germline genetic variants as instruments to probe causality, a bidirectional two-sample analysis found that genetic liability to insomnia was significantly associated with migraine risk in both a discovery dataset and an independent validation dataset. The reverse pathway, migraine causing insomnia, showed only a nominal effect that did not replicate, with a confidence interval spanning the null.

The authors are careful about what this genetic evidence can and cannot support. The insomnia instruments were derived from UK Biobank’s single self-reported question about trouble falling or staying asleep, capturing broad insomnia liability rather than formally diagnosed chronic insomnia disorder. More importantly, the migraine instruments were built for migraine broadly defined, not for the MOH transition specifically, which involves analgesic overuse behavior, opioid neuroadaptation and pharmacologically driven sleep disruption that genome-wide association data cannot model. Significant pleiotropy was detected in one validation dataset, and sample overlap between the insomnia and migraine genetic datasets was not corrected for. All causal inferences in the review therefore apply to migraine risk as a proxy, not to the MOH phenotype directly, and the authors call for future genetic studies using cohorts phenotyped strictly by ICHD-3 criteria.

In the opposite direction, the pharmacological evidence is more concrete. Each drug class overused in MOH disrupts sleep through mechanisms intrinsic to its own molecular pharmacology. Opioids suppress orexinergic neuronal firing and truncate slow-wave sleep. Barbiturates and benzodiazepines fragment sleep continuity and suppress REM architecture. Triptans may perturb sleep-related serotonergic pathways, though the clinical significance of chronic effects remains poorly characterized. These are not side effects in any conventional sense; they are direct consequences of engaging the same molecular targets that regulate restorative sleep. Withdrawal studies reinforce this direction of causation: in randomized comparisons and prospective cohorts, patients who achieved medication cessation reported improvements in headache frequency, with parallel gains in self-reported sleep quality described after withdrawal, though sleep was not a pre-specified outcome in the controlled trials.

The mechanistic heart of the review lies in three overlapping neurobiological substrates. The first is the orexin system, a cluster of hypothalamic neurons that project to the periaqueductal gray, the trigeminocervical complex, the rostral ventromedial medulla, the locus coeruleus and the dorsal raphe, structures that each play dual roles in wakefulness maintenance and descending pain inhibition. Animal work shows that selective activation of the orexin 1 receptor attenuates trigeminovascular nociceptive signaling, while blockade amplifies it. The only human biomarker data specific to MOH, however, point in an unexpected direction: in a cerebrospinal fluid study of 30 MOH patients, 27 with chronic migraine and 20 controls, orexin-A was significantly elevated in MOH and correlated positively with both monthly drug intake and substance-dependence scores. Whether this elevation reflects a compensatory antinociceptive response, a stress-induced hypothalamic reaction, or a mechanism tied to drug-seeking behavior remains unresolved, and the authors flag this orexinergic paradox as a key translational question.

The second substrate is melatonin and the circadian system. Melatonin’s nocturnal secretion by the pineal gland serves as the primary timing signal for the circadian clock, and it also carries documented analgesic and anti-inflammatory activity independent of its sleep-promoting effects. Altered circadian signaling has been proposed as a contributor to headache chronification, and genetic variants in CK1δ, a regulator of clock timing, associate independently with both sleep disorders and migraine susceptibility, raising the possibility that some patients carry a constitutive circadian vulnerability predisposing them to both conditions. A single randomized trial of nightly melatonin 3 milligrams showed significant reduction in migraine frequency versus placebo, though whether the mechanism is circadian stabilization, direct analgesia, or both is not established. The third substrate is neuroinflammation: even brief sleep restriction upregulates interleukin-1 beta, tumor necrosis factor alpha and interleukin-6, activates microglia and astrocytes, and impairs the glymphatic clearance of neuroactive metabolites, while chronic analgesic overuse adds a second inflammatory wave through toll-like receptor 4 signaling on glial cells. Together these loops may lower the pain threshold, drive further medication use and fragment sleep in a self-reinforcing cycle.

Sleep architecture itself tells part of the story. During slow-wave sleep, the thalamocortical system dampens sensory relay and raises the pain threshold, while during REM sleep monoaminergic descending inhibition drops to near zero, leaving the nervous system exposed to incoming nociceptive signals. Polysomnographic studies in this population document shortened total sleep time, increased nocturnal arousals, reduced slow-wave sleep duration and diminished slow-wave amplitude, a profile that offers a mechanistically coherent explanation for the morning headaches that characterize the condition. Clinically, the picture is recognizable once one knows to look for it: nearly constant dull pressure that periodically escalates, morning headache on waking, and a non-restorative quality of sleep in which patients wake exhausted regardless of hours in bed. Notably, the insomnia often persists even in patients without overt mood disturbance, suggesting that in at least a subset of cases it reflects primary neurobiological dysregulation rather than simple pain-induced wakefulness.

Two comorbidities compound the problem. Obstructive sleep apnea is substantially more prevalent among chronic migraine patients than in headache-free controls, and when apnea and insomnia coincide, the so-called COMISA phenotype, the resulting subgroup is distinctly harder to treat, with recurrent nocturnal hypoxia providing a trigeminovascular stimulus that medication withdrawal cannot eliminate. Untreated anxiety or depression at baseline similarly predicts poorer detoxification response and higher relapse. Against this backdrop, the authors argue that sequential management, detoxify first, treat sleep later, leaves the insomnia-driven component of central sensitization intact during the most fragile phase of recovery, while behavioral sleep treatment attempted during ongoing analgesic overuse is neutralized by the drugs’ pharmacological disruption of sleep. They propose, explicitly as a hypothesis, that concurrent medication withdrawal, cognitive behavioral therapy for insomnia, preventive pharmacotherapy including anti-CGRP monoclonal antibodies, and CPAP where indicated may outperform sequential care, though no randomized trial has ever tested this.

The review is candid about the gaps that temper its claims. Virtually every withdrawal study measured sleep by questionnaire; no polysomnographic evidence exists that sleep architecture objectively recovers after detoxification, and how recovery might differ across drug classes is unknown. Anti-CGRP antibodies have shown benefit in post-hoc analyses of chronic migraine trials with medication overuse, but none pre-specified sleep endpoints, and dual orexin receptor antagonists, which preserve slow-wave and REM sleep unlike benzodiazepines, carry a directionally uncertain rationale given the paradoxically elevated orexin-A in MOH cerebrospinal fluid. The authors’ prescription is a defined research agenda: a randomized trial of concurrent withdrawal, CBT-I and preventive pharmacotherapy versus detoxification alone, with polysomnographic sleep architecture and one-year relapse as co-primary endpoints. Insomnia in MOH, they conclude, is not a symptom waiting passively for the headache to resolve, but a sustaining mechanism in its own right, and treating it from the outset may prove to be what separates durable remission from repeated cycles of detoxification and relapse.

Subject of Research: The reciprocal neurobiological relationship between insomnia and medication overuse headache

Article Title: Insomnia and medication overuse headache: neurobiological links, reciprocal interactions, and integrated therapeutic strategies

Article References: Cotrik, E., Krymchantowski, A. V., Jevoux, C., Caldas Ferreira, I., Monteiro Auler, L., Ferreira Fontoura Sousa, M. C., & Marcussi, A. P. (2026). Insomnia and medication overuse headache: neurobiological links, reciprocal interactions, and integrated therapeutic strategies. Frontiers in Pain Research, 7, Article 1911261. https://doi.org/10.3389/fpain.2026.1911261

Image Credits: AI Generated

DOI: 10.3389/fpain.2026.1911261

Keywords: medication overuse headache, insomnia, migraine, orexin, melatonin, neuroinflammation, CBT-I, CGRP, sleep apnea, detoxification, Mendelian randomization, central sensitization