A preference for eveningness (being a "night owl") and preterm birth (<37 weeks of gestation) are associated with similar adversities, such as elevated blood pressure, impaired glucose regulation, poorer physical fitness, and lower mood. Yet, it remains unclear if and how preterm birth is associated with circadian preference. The aim of this study was to assess this association across the whole gestation range, using both objective and subjective measurements of circadian preference. Circadian preference was measured among 594 young adults (mean age 24.3 years, SD 1.3) from two cohorts: the ESTER study and the Arvo Ylppö Longitudinal Study. We compared 83 participants born early preterm (<34 weeks) and 165 late preterm (34 to <37 weeks) with those born at term (≥37 weeks, n = 346). We also compared very low birth weight (VLBW, <1500 g) participants with term-born controls. We obtained objective sleep data with actigraphs that were worn for a mean period of 6.8 (SD 1.4) nights. Our primary outcome was sleep midpoint during weekdays and weekend. The sleep midpoint is the half-way time between falling asleep and waking up, and it represents sleep timing. We also investigated subjective chronotype with the Morningness-Eveningness Questionnaire (MEQ) in 688 (n = 138/221/329) ESTER participants. The MEQ consists of 19 questions, which estimates the respondent to be of a "morning", "evening," or "intermediate" chronotype, based on the Morningness-Eveningness Score (MES). We analyzed the data from the actigraphs and the MES with three linear regression models, and analyzed distribution of the chronotype class with Pearson χ2. There were no consistent differences across the study groups in sleep midpoint. As compared with those born at term, the mean differences in minutes:seconds and 95% confidence intervals for the sleep midpoint were: early preterm weekdays 11:47 (-8:34 to 32:08), early preterm weekend 4:14 (-19:45 to 28:13), late preterm weekdays -10:28 (-26:16 to 5:21), and late preterm weekend -1:29 (-20:36 to 17:37). There was no difference in sleep timing between VLBW-participants and controls either. The distribution of chronotype in the MEQ among all participants was 12.4% morningness, 65.4% intermediate, and 22.2% eveningness. The distribution of the subjective chronotype class did not differ between the three gestational age groups (p = 0.98). The linear regression models did not show any influence of gestational age group or VLBW status on the MES (all p > 0.5). We found no consistent differences between adults born early or late preterm and those born at term in circadian preference. The earlier circadian preference previously observed in those born smallest is unlikely to extend across the whole range of preterm birth.
