A quarter of Chileans screen positive for sleep apnoea risk — on questionnaires alone

A systematic review and meta-analysis of obstructive sleep apnoea in Chile, published in Revista Médica de Chile, pooled 15 studies covering 13,157 participants and arrived at a pooled prevalence of moderate-to-high OSA risk of 25% (95% CI 16–34%) [s1]. The number carries a large asterisk. That 25% is a screening-questionnaire result, defined by a STOP-Bang score above 3 — and of the 15 studies included, only three used respiratory polygraphy [s1].

What the Chilean review found

Sleep apnoea is a common breathing disorder that the review notes affects nearly a billion people worldwide, and that remains underdiagnosed and undertreated, particularly in middle-income countries [s1]. The review searched seven databases for studies reporting the prevalence, risk, or symptoms of OSA in Chile [s1]. The STOP-Bang questionnaire and the Epworth Sleepiness Scale were the instruments used most often; the Berlin Questionnaire and various sleep quality indices appeared less frequently [s1]. The most commonly reported symptoms were snoring, excessive daytime sleepiness, and witnessed apnoeas [s1].

The heterogeneity statistic tells its own story: I² = 98.4% [s1]. At that level, the pooled estimate is best read as a summary of a very inconsistent literature rather than as a national prevalence figure. The confidence interval spanning 16% to 34% is the honest version of the finding.

The authors' conclusion is framed around the evidence gap rather than the number: the findings reveal a substantial burden of OSA risk and related symptoms in the Chilean population, based chiefly on screening questionnaires, with objective diagnostic methods limited — and they call for population-based research using standardised, objective assessment tools [s1].

Why questionnaire screening is not prevalence

The Chilean estimate is explicitly a pooled prevalence of moderate-to-high risk of OSA, defined by a STOP-Bang score above 3 [s1]. Risk-stratification instruments of that kind exist to decide who should be sent for testing, not to establish who has the condition. Using one to estimate prevalence conflates "at elevated risk" with "affected," and the size of the resulting error depends on how the underlying population is distributed across the demographic characteristics such instruments score.

The Chilean review is explicit that this is what its estimate rests on [s1], which is the correct way to report it. But it means the widely quotable figure — a quarter of Chileans — describes screening positivity, not disease.

The Peruvian counterpoint

A retrospective analysis published in February examined 350 adults evaluated at a specialised sleep centre in Lima, Peru, using the Epworth Sleepiness Scale for daytime sleepiness and overnight polysomnography for diagnosis [s2]. Because this is a referred clinical population rather than a community sample, its prevalence figures describe who gets sent to a sleep laboratory, not who has apnoea in Peru.

Within that population, excessive daytime sleepiness was present in 84.6% (296 of 350), with Epworth levels II and III predominating [s2]. Polysomnography found no evidence of sleep apnoea in 8.0% (28 patients), central sleep apnoea in 49.1% (172), and obstructive sleep apnoea in 42.9% (150) [s2].

Factors significantly associated with sleep apnoea included male sex (OR 5.32; 95% CI 2.01–14.06), daytime sleepiness (OR 3.51; 95% CI 1.52–8.10), difficulty falling asleep (OR 5.20; 95% CI 2.32–11.64), waking with sleepiness (OR 35.11; 95% CI 8.16–151.07), and a choking sensation on waking (OR 15.62; 95% CI 2.10–116.42) [s2]. The last two intervals are extremely wide, which is what small numbers in one cell of a contingency table produce; those odds ratios should be read as "associated," not as effect sizes anyone should quantify.

The finding that matters most

The Peruvian study built a decision tree in which excessive daytime sleepiness emerged as the main predictor of apnoea type, and the model achieved an overall accuracy of 58.5% [s2].

An overall accuracy of 58.5% for sorting patients into no apnoea, central apnoea or obstructive apnoea is modest, and the authors treat it that way. Their stated conclusion is that the model's moderate predictive accuracy reinforces the importance of polysomnography as the gold-standard diagnosis [s2] — while also arguing that screening for daytime sleepiness in primary care remains a practical way to identify people who warrant referral [s2].

Read alongside the Chilean review, that is the same message from two directions. Symptom-based tools are useful for deciding who to test and poor at deciding who is affected. In Chile the consequence is a prevalence literature built on the wrong instrument [s1]; in Peru it is a reminder of why the right instrument cannot be skipped [s2].

What neither study establishes

Neither is a population-based epidemiological study with objective diagnosis, which is precisely the gap both identify. The Chilean pooled estimate cannot be converted into a number of Chileans with OSA. The Peruvian figures — including the striking finding that central sleep apnoea outnumbered obstructive apnoea in that clinic [s2] — reflect referral patterns at a single specialised centre in Lima and should not be generalised to Peru, or to any other setting.

Neither study reports treatment access, which is the question that follows a prevalence estimate and which the region's literature does not appear to answer.

What to watch

Whether any South American country runs a population-based sleep study with home or laboratory testing rather than questionnaires. Until one does, statements about how common sleep apnoea is on the continent rest on instruments designed for a different job.

Sources

• Epidemiology of Obstructive Sleep Apnea in Chile: A Systematic Review and Meta-Analysis — Revista Médica de Chile, June 2026
• Excessive Daytime Sleepiness and Risk of Sleep Apnea: A Retrospective Analysis of Peruvian Adults — Clinical Practice & Epidemiology in Mental Health, 25 February 2026