When Insomnia Is a Symptom: The Medical Conditions Behind Persistent Sleeplessness

The clinical question worth asking first is not which sleep aid to try. It is whether the insomnia is the problem or the presenting sign of a problem. Those lead to very different plans.

Insomnia and medical illness also run in both directions. Poor sleep worsens pain sensitivity, glycemic control, and mood, which in turn worsen sleep. That loop is one reason chronic insomnia rarely resolves when only one end of it is addressed.

Chronic insomnia disorder is defined by difficulty initiating or maintaining sleep on at least three nights per week for at least three months, with daytime consequences. Nothing in that definition specifies a cause, which means the diagnosis can be correctly applied to a patient whose underlying condition has never been looked for.

Treating the sleep complaint without evaluating for a medical driver can delay recognition of untreated sleep apnea, heart failure, reflux disease, or an unrecognized movement disorder. In some of those cases, sedation is not a neutral intervention.

The most cited community-level estimate comes from a study of 772 men and women aged 20 to 98, published in Sleep in 2007 by Taylor and colleagues. Participants completed self-report measures of sleep, health, depression, and anxiety, and the analysis compared insomnia rates across medical conditions in both directions.

Looked at from the illness side, the proportions are striking. Chronic insomnia was reported by 66.7 percent of people with neurologic disease versus 24.3 percent of those without, 59.6 percent of those with breathing problems versus 21.4 percent, 55.4 percent with gastrointestinal problems versus 20.0 percent, 48.6 percent with chronic pain versus 17.2 percent, 44.1 percent with heart disease versus 22.8 percent, 44.0 percent with high blood pressure versus 19.3 percent, 41.5 percent with urinary problems versus 23.3 percent, and 41.4 percent with cancer versus 24.6 percent.

When all medical problems were entered together, only hypertension, breathing problems, urinary problems, chronic pain, and gastrointestinal problems remained independently associated with higher insomnia rates. That adjustment matters: it suggests some of the single-condition associations reflect clustering of illness rather than separate causal pathways.

The mechanisms are not interchangeable, and knowing which one is operating tells you where to intervene. Respiratory conditions fragment sleep through repeated arousals and nocturnal desaturation. Pain conditions work through nociceptive arousal compounded by the hyperalgesia that sleep loss itself produces. Cardiac conditions interrupt sleep through orthopnea, paroxysmal nocturnal dyspnea, and nocturnal diuresis.

Gastrointestinal conditions, particularly reflux disease, produce arousals tied to recumbency. Urologic conditions act through nocturia, whether from prostatic obstruction, overactive bladder, diuretic timing, or chronic kidney disease. Neurologic conditions are the most heterogeneous group, encompassing restless legs syndrome, REM sleep behavior disorder, circadian disruption in dementia, and the sleep architecture changes of Parkinson disease.

Hormonal contributors follow their own pattern. In 3,045 women aged 42 to 52 followed across seven annual assessments in the Study of Women’s Health Across the Nation, more frequent vasomotor symptoms were associated with higher odds of difficulty falling asleep, difficulty staying asleep, and early morning awakening, and adjusted odds for the first two rose through the menopausal transition.

Restless legs syndrome is the clearest example of a treatable condition that reaches clinicians as a sleep complaint. In a survey of 15,391 completed questionnaires from adults aged 18 and over, symptoms occurring at least twice weekly and rated moderately or severely distressing were present in 2.7 percent. Of that group, 81.0 percent had discussed their symptoms with a primary care physician, and only 6.2 percent had received a diagnosis of restless legs syndrome.

Obstructive sleep apnea presents similarly. Patients frequently describe the problem as insomnia rather than as snoring or daytime sleepiness, particularly women and older adults, and the sleep maintenance pattern can look identical to primary insomnia from the history alone.

The practical implication is that a sleep complaint deserves a history that reaches past the sleep itself: what wakes the person, what position they sleep in, what happens to their legs in the evening, what medications they take and when, and what their bladder, chest, joints, and stomach are doing overnight.

The American College of Physicians recommends cognitive behavioral therapy for insomnia as the initial treatment for chronic insomnia disorder in all adult patients, a strong recommendation based on moderate-quality evidence, with pharmacotherapy considered through shared decision-making only when that approach has not succeeded.

The American Academy of Sleep Medicine’s 2017 pharmacologic guideline is notable for what it advises against. Using GRADE methodology, the task force suggested that clinicians not use trazodone, diphenhydramine, melatonin, tryptophan, or valerian for sleep onset or sleep maintenance insomnia. Several of those are the exact agents patients reach for without a prescription.

None of that argues against treating the sleep complaint. It argues that sedation is a weak tool applied to a symptom whose cause has not been characterized, and that the agents most readily available over the counter have the least support behind them.

The honest summary of the cannabinoid sleep literature is that it is mixed and methodologically limited. A systematic review of 39 publications examining cannabinoid administration and quantitative sleep measures in humans reported varied effects across sleep outcomes and concluded that methodological problems in the majority of studies precluded any definitive conclusion.

A widely cited retrospective chart review of 72 adults at a psychiatric clinic found sleep scores improved within the first month in 66.7 percent of patients but fluctuated over time. A retrospective case series with no control group and a fluctuating outcome is hypothesis-generating, and it is regularly quoted as though it were a trial.

For obstructive sleep apnea specifically, the position is explicit. A phase II randomized trial of 73 adults with moderate or severe apnea found that dronabinol reduced the apnea-hypopnea index in a dose-dependent fashion, but the American Academy of Sleep Medicine subsequently issued a position statement that medical cannabis and its synthetic extracts should not be used to treat obstructive sleep apnea, citing unreliable delivery methods and insufficient evidence on effectiveness, tolerability, and safety, and recommending that apnea be excluded from state medical cannabis program condition lists.

Anyone using cannabis for sleep who has not been evaluated for apnea is in exactly the situation that position statement was written about.

A sleep history that specifies the pattern matters more than a global complaint of poor sleep. Difficulty falling asleep, difficulty staying asleep, and early morning awakening point in different directions, and the answer to what wakes the person is often the diagnosis.

A medication and substance review belongs in the same visit, covering timing as well as agents: diuretic dosing, stimulant dosing, corticosteroids, and alcohol used as a sleep aid are common and modifiable contributors.

Depending on that history, evaluation may reasonably extend to screening for sleep-disordered breathing, assessment for restless legs syndrome, thyroid function testing, review of reflux and urologic symptoms, and screening for depression and anxiety, which are both common comorbidities and independent predictors. A pooled analysis of 21 longitudinal epidemiological studies found that people with insomnia and no depression at baseline had roughly twice the odds of developing depression at follow-up, with a random-effects odds ratio of 2.60.

None of this is a protocol a patient should work through alone. It is the shape of a conversation worth having with a clinician who will take the overnight details seriously.

The comorbidity estimates come largely from cross-sectional community surveys using self-report. That design is appropriate for establishing that insomnia and medical illness travel together, and it establishes that clearly and consistently. It cannot establish which came first.

The prognostic and longitudinal material is stronger. The cardiovascular association used objective polysomnography with a median 11.4 years of follow-up, and the depression association pooled 21 longitudinal studies. Those designs support directional claims that the cross-sectional work cannot.

The headline comorbidity percentages are frequently quoted without the multivariate adjustment that accompanied them. When all medical problems were considered together, only hypertension, breathing problems, urinary problems, chronic pain, and gastrointestinal problems retained an independent association. Quoting the unadjusted figure for cancer or neurologic disease overstates what the analysis showed.

Self-reported medical history and self-reported sleep also share measurement error in a predictable direction: people who feel unwell report both more illness and worse sleep. The cardiovascular finding is more robust precisely because sleep duration there was measured objectively rather than reported.

This literature does not show that treating an underlying medical condition resolves the insomnia. That is a reasonable clinical expectation in many cases and it has been tested in relatively few. The authors of the comorbidity study said as much, noting that the efficacy of treating insomnia in most of these comorbid conditions has not been tested.

It also does not support ranking these conditions by how often they cause insomnia in an individual patient. Community prevalence is not diagnostic probability, and a patient’s own history carries far more weight than a population percentage.

Sleep medicine spent decades separating primary from secondary insomnia, then largely abandoned the distinction because it proved unreliable and because it encouraged clinicians to defer treating sleep until the other condition resolved. Current guidelines treat insomnia as a comorbid disorder deserving its own attention.

That shift should not be read as permission to skip the evaluation. The contemporary position is that both deserve treatment, not that the underlying cause stopped mattering. The conditions most commonly missed, apnea and restless legs syndrome, are also among the most treatable.

The question I ask that changes the visit most often is not how long it takes someone to fall asleep. It is what wakes them up. Patients describe it precisely when asked: a full bladder, a coughing fit, reflux burning, a leg that will not settle, or waking with a start and no idea why. Each of those points somewhere different.

I see a lot of people who have been using cannabis for sleep for years and have never had an apnea evaluation. That combination worries me. The sleep specialty organizations have been clear that cannabis is not a treatment for apnea, and a product that makes the nights feel shorter can delay recognition of something that is quietly damaging the heart.

None of this means sleep should be left untreated while a workup proceeds. It means both threads get pulled at once, which is what current guidelines actually say.

Chronic insomnia frequently accompanies an identifiable medical condition, and the mechanism differs enough between conditions that the differential changes the plan. Ask what wakes the patient, review medication timing, and consider evaluation for sleep-disordered breathing and restless legs syndrome before settling on a sedating intervention. Treating the sleep and investigating the cause are parallel tasks, not sequential ones.

Carry forward the mechanism framework and the two conditions most often missed, sleep apnea and restless legs syndrome. Carry forward the adjusted rather than the unadjusted comorbidity findings. Leave behind any published table assigning confidence intervals to cannabinoid effects on individual sleep stages. Those numbers circulate widely and are not derived from measured data.