Does noise (traffic, neighbours, even white noise) harm your sleep and heart?

Why noise reaches you asleep. Your ears stay open all night and the sleeping brain keeps scanning. Pooling sleep-lab studies, each 10 dBA rise in the loudest indoor noise raised the odds of waking by about 35%, for aircraft, road and rail noise alike¹. In surveys that ask about noise, each 10 dB rise in night-time level went with roughly two to three times the odds of being highly sleep-disturbed, but the link was weak when the question did not mention noise¹. People partly adapt. In a lab study of 72 people over 11 nights, effects on sleep structure were subtle, yet heart-rate responses to noise did not fade and people still rated their sleep and recovery worse⁷.

From sleep to heart. Night noise fragments sleep, raises stress hormones and oxidative stress, and can promote vascular dysfunction, inflammation and high blood pressure⁸. In a randomised crossover trial, 70 healthy volunteers slept three nights at home with background noise or with 30 or 60 simulated train passes. Their arteries dilated less after noisy nights (about 8.5% versus 11.2%)⁹. Population data fit: pooling seven long-term studies, each 10 dB rise in average road noise went with an 8% higher risk of ischaemic heart disease, rated high quality². A 2025 update of 53 studies found a smaller effect, about 2 to 3% per 10 dB¹⁰.

Which levels matter? WHO's 2018 guideline recommends road traffic noise below 53 dB averaged over a day (with extra weight for evening and night) and below 45 dB at night, measured outside the most exposed wall⁵. The night figure is based on sleep: 3% of people are highly sleep-disturbed at about that level⁵.

White noise, pink noise and sound machines. A systematic review of 38 articles judged the evidence that continuous noise improves sleep to be very low quality, and noted it might also disrupt sleep or harm hearing³. A 2025 meta-analysis of 12 randomised trials (1,301 people) found better sleep scores in adults, but with large variation between trials, mostly in clinical settings¹¹. In a pilot of 12 volunteers, continuous pink noise at 45 dB softened the effects of traffic noise on sleep¹². As for infant sleep machines, all 14 tested exceeded 50 dBA at 30 cm at maximum volume, and three exceeded 85 dBA⁴.

Practical. Sleep on the quiet side of the home: WHO's reviews found a quiet side reduces annoyance and sleep disturbance⁵. Earplugs with eye masks cut awakenings and sleep-onset time in a trial of 40 healthy volunteers in a simulated hospital⁶. For protecting your ears from loud noise, see Hearing Loss.

Sleep. Basner and McGuire's systematic review for WHO included 74 non-laboratory studies published after 2000¹. In pooled surveys, the odds ratios of being highly sleep disturbed per 10 dB of night-time level were 1.94 (aircraft), 2.13 (road) and 3.06 (rail) when the question referred to noise, but non-significant (1.17, 1.09 and 1.27) when it did not. Pooled polysomnography studies gave an odds ratio for awakening of about 1.35 per 10 dBA of indoor maximum level for all three sources. Evidence quality was moderate for cortical awakenings and for noise-referenced self-reported disturbance, low for movement measures and very low for other outcomes¹. In the laboratory study by Basner and colleagues, 72 healthy people aged 18 to 71 were exposed to 40, 80 or 120 noise events over 11 nights. Effects on sleep structure were subtle because of habituation, but cardiac arousals did not habituate, and sleep quality and recuperation were rated worse. Objective performance was unaffected apart from a 4 ms slower reaction time⁷.

Heart: observational. WHO's cardiovascular review rated the evidence on road traffic noise and ischaemic heart disease as high quality. Seven longitudinal studies gave a relative risk of 1.08 (95% CI 1.01 to 1.15) per 10 dB of average noise², covering 67,224 people, with the lowest exposure at 53 dB⁵. The same review found one study of hypertension, with no increase (RR 0.97, 0.90 to 1.05), and rated evidence for stroke, diabetes and obesity as moderate to very low^2,5. Minkin and colleagues' 2025 update included 53 studies, 87% of them European. For road noise the relative risk of heart disease incidence was 1.02 (1.01 to 1.04) per 10 dB and of heart disease mortality 1.03 (1.01 to 1.05). Aircraft noise gave 1.03 (0.99 to 1.07) for incidence, and 1.00 (0.99 to 1.01) after excluding small-area studies. Risk began between 32 and 45 dB depending on source and outcome, and effects looked larger in older adults¹⁰. The newer estimate is smaller than WHO's, though the direction holds. Observational studies cannot fully exclude confounding by other traffic-related exposures.

Heart: experimental. Herzog and colleagues exposed 70 healthy volunteers, in a randomised crossover design, to background noise or to 30 or 60 simulated train noise events in their own homes for three nights. Flow-mediated dilation of the brachial artery was 11.23% after control nights, 8.71% after 30 events and 8.47% after 60 (p < 0.001). Sleep quality was impaired, and blood proteins shifted toward a pro-inflammatory, pro-thrombotic pattern⁹. These were single nights in young healthy people, so lasting damage is not shown.

White noise and sound masking: conflicting. Riedy's systematic review (38 articles) found wide differences in noise type, sleep measurement, adherence and controls, with results ranging from improved to disrupted sleep. By GRADE criteria the evidence that continuous noise improves sleep was very low quality, and the authors urged more research before promoting it as a sleep aid³. Ding and colleagues' 2025 meta-analysis of 12 randomised trials (1,301 participants, 544 of them critically ill) found that white noise lowered Pittsburgh Sleep Quality Index scores by 3.70 points in adults (-4.90 to -2.50) and 2.71 in older adults, and lengthened 24-hour sleep by 138 minutes in infants and young children but not their nocturnal 12-hour sleep. The authors flagged heterogeneity and methodological limits¹¹. In Vincens and colleagues' pilot, 12 healthy people slept five nights in acoustically isolated rooms. Traffic noise events (45 to 65 dB maximum) fragmented sleep and shifted blood metabolites (leucine, lactic acid, acetone), and continuous 45 dB pink noise attenuated both. The sample was small and homogeneous¹². Together these suggest masking noise may help in some settings, but the evidence is inconsistent, mostly short-term and often hospital-based, and long-term safety is untested.

Infant sleep machines. Hugh and colleagues measured 14 machines at maximum volume at 30, 100 and 200 cm, with corrections for a six-month-old's ear canal. All exceeded 50 dBA at 30 cm, the recommended limit for hospital nurseries, and three exceeded 85 dBA, which over 8 hours exceeds adult occupational limits for noise exposure⁴.

Earplugs. In Huang and colleagues' trial, 40 healthy volunteers were randomised to four simulated-ICU conditions. Earplugs with eye masks produced fewer awakenings and a shorter time to fall asleep, while oral melatonin improved perceived sleep quality more⁶. We found no trial of earplugs against traffic noise in ordinary homes.

Bottom line (confidence 4). Traffic noise at night probably harms sleep and modestly raises the risk of heart disease, with a consistent direction across designs and a plausible mechanism, though the size of the heart effect has shrunk in newer analyses. White noise has mixed, low-quality evidence, and infant machines can be dangerously loud. Reducing noise at the source or in the bedroom is the best-supported step.