Changes in Sleep Duration and Sleep Timing in the General Population from before to during the First COVID-19 Lockdown: A Systematic Review and Meta-Analysis

Abstract

Background: The COVID-19 lockdown had a profound effect on everyday life, including sleep health. This systematic review and meta-analysis evaluated changes in quantitative sleep parameters during the first lockdown compared with pre-lockdown in the general population.

Methods: A search in scientific databases was performed to identify eligible observational studies from inception to 8 February 2023. We performed a random effects meta-analysis of those studies reporting (a) means of sleep duration, time in bed (TIB), and sleep timing (bedtime and wake-up time); (b) the percentages of atypical sleep duration before and during the lockdown; (c) the percentages of change in sleep duration and sleep timing.

Results: A total of 154 studies were included. A small increase in sleep duration (0.25 standardized mean difference, 95% CI 0.180-0.315) was found, with 55.0% of the individuals reporting changes, predominantly an increase (35.2%). The pooled relative risk for sleeping more than 8/9 h per night was 3.31 (95% IC 2.60-4.21). There was a moderately significant delay in sleep timing and a surge in napping.

Conclusion: An increase in sleep duration and napping, and delayed sleep timing were observed. High-quality studies should evaluate whether these parameters have now become chronic or have returned to pre-lockdown values.

Keywords: COVID-19 lockdown; bedtime; general population; napping; sleep duration; sleep timing; systematic review; wake-up time.

Figure 1

PRISMA flow diagram of study selection.

Figure 2

Forest plot showing pooled changes in sleep duration (hours) between before and during the lockdown. Caption: error bars = 95% confidence interval; square boxes = individual study point estimates; diamond box = pooled point estimates. Aishworiya R., 2021a: female; Aishworiya R., 2021b: male; Asensio-Cuesta S., 2021a: male; Asensio-Cuesta S., 2021b: female; Bann D., 2021a: 1946; Bann D., 2021b: 1958; Bann D., 2021c: 1970; Bann D., 2021d: 1990; Bann D., 2021e: 2001; Cellini N., 2021a: Belgian regular workers, female; Cellini N., 2021b: Belgian regular workers, male; Cellini N., 2021c: Belgian remote workers, female; Cellini N., 2021d: Belgian remote workers, male; Cellini N., 2021e: Belgian students, female; Cellini N., 2021f: Belgian students, male; Cellini N., 2021g: Belgian unemployed/retired, female; Cellini N., 2021h: Belgian unemployed/retired, male; Cellini N., 2021i: Italian regular workers, female; Cellini N., 2021j: Italian regular workers, male; Cellini N., 2021k: Italian remote workers, female; Cellini N., 2021l: Italian remote workers, male; Cellini N., 2021m: Italian Stop working, female; Cellini N., 2021n: Italian Stop working, male; Cellini N., 2021o: Italian students, female; Cellini N., 2021p: Italian students, male; Cellini N., 2021q: Italian unemployed/retired, female; Cellini N., 2021r: Italian unemployed/retired, male; Mandelkorn U., 2021a: Study 1; Mandelkorn U., 2021b: Study 2; Ramos Socarras L., 2021a: 18–21-years old; Ramos Socarras L., 2021b: 22–25-years old.

Figure 3

Forest plot showing a pooled percentage of changes in sleep duration from before to during the lockdown: changes (a), decrease (b), and increase (c). Caption: error bars = 95% confidence interval; square boxes = individual point estimates; diamond box = pooled point estimates.

Figure 4

Forest plot showing pooled changes in atypical sleep duration from before to during the lockdown: atypical (a), decrease (b), increase (c). Caption: error bars = 95% confidence interval; square boxes = individual study point estimates; diamond box = pooled point estimates.

Figure 5

Forest plot showing pooled changes in TIB from before to during the lockdown: increase, decrease. Caption: error bars = 95% confidence interval; square boxes = individual point estimates; diamond box = pooled point estimates. Cellini N., 2020a: student; Cellini N., 2020b: worker; Cellini N., 2021a: Belgian regular workers, female; Cellini N., 2021b: Belgian regular workers, male; Cellini N., 2021c: Belgian remote workers, female; Cellini N., 2021d: Belgian remote workers, male; Cellini N., 2021e: Belgian students, female; Cellini N., 2021f: Belgian students, male; Cellini N., 2021g: Belgian unemployed/retired, female; Cellini N., 2021h: Belgian unemployed/retired, male.

Figure 6

Forest plot showing pooled changes in bedtime from before to during the lockdown. Caption: error bars = 95% confidence interval; square boxes = individual study point estimates; diamond box = pooled point estimates. Cellini N., 2020a: student; Cellini N., 2020b: worker; Cellini N., 2021a: Belgian regular workers, female; Cellini N., 2021b: Belgian regular workers, male; Cellini N., 2021c: Belgian remote workers, female; Cellini N., 2021d: Belgian remote workers, male; Cellini N., 2021e: Belgian students, female; Cellini N., 2021f: Belgian students, male; Cellini N., 2021g: Belgian unemployed/retired, female; Cellini N., 2021h: Belgian unemployed/retired, male; Ramos Socarras L., 2021a: 18–21-years old; Ramos Socarras L., 2021b: 22–25-years old.

Figure 7

Forest plot showing pooled changes in bedtime from before to during the lockdown: changes (a), delayed (b), earlier (c). Caption: error bars = 95% confidence interval; square boxes = individual point estimates; diamond box = pooled point estimates.

Figure 8

Forest plot showing pooled changes in wake-up time from before to during the lockdown. Caption: error bars = 95% confidence interval; square boxes = individual study point estimates; diamond box = pooled point estimates. Cellini N., 2020a: student; Cellini N., 2020b: worker; Cellini N., 2021a: Belgian regular workers, female; Cellini N., 2021b: Belgian regular workers, male; Cellini N., 2021c: Belgian remote workers, female; Cellini N., 2021d: Belgian remote workers, male; Cellini N., 2021e: Belgian students, female; Cellini N., 2021f: Belgian students, male; Cellini N., 2021g: Belgian unemployed/retired, female; Cellini N., 2021h: Belgian unemployed/retired, male; Ramos Socarras L., 2021a: 18–21-years old; Ramos Socarras L., 2021b: 22–25-years old.

Figure 9

Forest plot showing pooled changes in wake-up time from before to during the lockdown: changes (a), delayed (b), earlier (c). Caption: error bars = 95% confidence interval; square boxes = individual point estimates; diamond box = pooled point estimates.