Design: systematic review and random-effects meta-analysis of 11 studies, PROSPERO registered, with leave-one-out sensitivity analysis
Population: 5,568 participants with and without sleep paralysis
Primary outcome: anxiety symptom scores and prevalence of diagnosed anxiety disorder
Effect: anxiety scores SMD 0.35 (95% CI 0.27–0.44), I² = 0%; prevalence RR 1.28 (0.59–2.77), rising to 1.80 (1.08–2.99) after removing one outlier

Sleep paralysis — the transient inability to move at the edge of sleep, often with vivid fear — is usually handled as a curiosity when a patient mentions it, if they mention it at all. A meta-analysis of 11 studies and 5,568 participants asked whether it tracks with anxiety.

It does, on the continuous measure. People with sleep paralysis had higher anxiety scores than controls, a standardised mean difference of 0.35 (95% confidence interval 0.27 to 0.44), and the studies agreed with each other completely: heterogeneity was zero. The categorical analysis was less clean. Pooled risk of a diagnosed anxiety disorder was not significantly raised (risk ratio 1.28, 95% CI 0.59 to 2.77), and only became so when one outlying study was removed (1.80, 1.08 to 2.99).

For clinic, the usable version is the smaller claim: sleep paralysis is a marker of anxiety symptom burden rather than a diagnostic sign. It costs nothing to ask about it when taking a sleep history from an anxious patient, and a patient who has been frightened by episodes for years and has never been told what they are will often be relieved simply by the explanation. Reassurance that the phenomenon is benign, together with attention to sleep timing and sleep deprivation, is a reasonable first response.

Ask directly about episodes of being unable to move on waking or falling asleep — patients rarely volunteer them.
Explain the mechanism: REM-sleep muscle atonia persisting into wakefulness. The explanation itself is therapeutic.
Ask about sleep deprivation and irregular sleep timing, which are the common precipitants.
Screen for narcolepsy if episodes are frequent and accompanied by daytime sleepiness or cataplexy.
Do not treat a single episode as evidence of an anxiety disorder.

Why it matters

A symptom most clinicians treat as an aside carries information about anxiety burden, and costs one question to elicit.

The statistics, in plain English

An I² of 0% is unusual and means the eleven studies found essentially the same effect — the pooled standardised mean difference of 0.35 is about as stable as meta-analysis gets, though 0.35 is a small effect. The categorical result is the opposite: the confidence interval on the risk ratio (0.59 to 2.77) crosses 1.0, so it is compatible with no association at all, and the fact that significance appeared only after removing one study means that result should not be quoted.