The current study is unique because it examines long-term associations between sleep duration and sleep problems in early childhood and developmental trajectories of UI. Children who had a longer night-time sleep duration at 3½ years were less likely to suffer from UI, especially daytime wetting, at school-age. Preschool children who slept more than 8½ hours per night had a decreased probability of experiencing any type of UI at school-age. Sleep problems in preschool children were associated with daytime wetting alone (difficulty going to sleep, getting up after being put to bed), delayed attainment of daytime and night-time bladder control (night-time waking, breathing problems whilst sleeping), and persistent day and night wetting (night-time waking). Sleep problems were not more common in children with bedwetting alone. Instead, night-time waking and waking up early at age 3½ years were associated with a lower likelihood of children wetting the bed at primary school-age. We found that refusal to go to bed was associated with reduced odds of persistent wetting. This was an unexpected finding that could have been due to chance. It is notable that refusal to go to bed was one of the most common sleep problems in 3½ year-olds in the ALSPAC cohort, which is consistent with other studies [23].

Key strengths of this study include the prospective design based on a large community-based cohort, the availability of data on sleep duration and sleep problems in preschool children, and the use of developmental trajectories of UI from the age of 4–9 years (rather than measurement of incontinence at a single timepoint). We adjusted for a wide range of confounders which have previously been found to have robust associations with sleep problems and UI in children. We have not based our conclusions purely on p value thresholds (e.g. p < 0.05) to determine statistical significance, but instead, we consider the effect estimates alongside the strength of evidence indicated by the p values and confidence intervals in line with current recommendations [24, 25].

We did not restrict our analysis to children who met clinical diagnostic criteria for bedwetting or daytime wetting, therefore, the study provides evidence that sleep duration/problems are associated with UI in a non-clinical sample of children in the community. The prospective cohort design reduces the possibility of reverse causality and recall bias as alternative explanations for our findings. This is important because it has been suggested that parents of enuretic children are more likely than those of continent children to check their child during sleep and, therefore, may be more aware of their child’s sleep problems [6]. This is unlikely in our study because parents provided assessments of their child’s sleep problems before an age at which bedwetting would typically be viewed as abnormal.

Whilst most previous studies focused on sleep and enuresis [3, 4], a strength of our study is the inclusion of daytime wetting, which has been a neglected topic in existing research. A limitation is that the ‘bedwetting alone’ class did not exclude children with other lower urinary tract symptoms such as urgency and frequency irrespective of daytime wetting. Consequently, we were unable to fully ascertain if there are differential associations with monosymptomatic versus non-monosymptomatic enuresis [11]. Another limitation is the reliance on parental reports of sleep problems rather than clinical assessment of sleep disorders. Polysomnography is the gold standard for the assessment of sleep disorders, but this was not feasible in a large community-based birth cohort study. Also, we are unable to generalise our results to minority ethnic groups and less affluent populations because the ALSPAC cohort is predominantly white and affluent.

Previous studies have focused on the comorbidity (co-occurrence at the same time point) of sleep problems and enuresis. Only one previous cohort study examined prospective associations and found that a longer sleep duration in infancy was associated with delayed attainment of night-time bladder control [9]. There are several possible reasons for the difference in findings. The previous study examined only night-time bladder control as the outcome whilst we examined different patterns of bedwetting and daytime wetting. They assessed sleep duration as the typical length of time the child slept throughout a 24-h period (including daytime sleeps) whilst we focused only on night-time sleep duration. Also, the earlier study assessed sleep duration in infancy (age 1–2 years) whilst we focused on sleep duration in preschool children (age 3½ years). We found the strongest association between a longer night-time sleep duration and decreased odds of daytime wetting, whilst there was weaker evidence of an association between sleep duration and bedwetting alone.

A previous study examined associations between sleep problems and both enuresis and daytime UI but did not find evidence for cross-sectional associations [10]. In agreement with this study, we did not find evidence for associations between sleep problems and an increased risk of bedwetting alone. Some previous studies have found that sleep problems are more common in children with enuresis, but these findings are based on small case–control studies of clinically referred children with enuresis [5,6,7,8]. Some of these studies did not exclude children with daytime wetting [5, 6], and some used objective measures of sleep disturbances [5, 8] which might explain the differences in findings. In contrast to the study by von Gontard and colleagues [10], we found evidence that sleep problems (e.g., night-time waking, difficulty going to sleep, getting up after being put to bed) were prospectively associated with incontinence, specifically UI classes that were characterized by daytime wetting alone, or a combination of daytime and night-time wetting at school-age.

There are several possible explanations for the observed associations between sleep duration/ problems in preschool children and UI at primary school-age. There is evidence that insufficient sleep has an impact on brain development including areas that are associated with inhibitory control [26]. Lack of inhibitory control is associated with behaviour problems such as ADHD, which are strongly associated with UI [27]. The development of voluntary control over the bladder is a complex process that involves acquiring the ability to inhibit urination, despite the sensation of bladder fullness, until it is socially appropriate to urinate [28]. Insufficient sleep could affect the development of neural circuits connecting the cerebral cortex and the bladder. Poor sleep in children is also associated with an increased risk of childhood overweight and obesity [29], which are associated with an increased risk of UI [30].

We found that preschool children who had breathing problems whilst sleeping were more likely to experience delays in attaining bladder control. The ‘delayed’ class was characterised by the presence of UI (bedwetting and daytime wetting) up to age 6, followed by a decreasing prevalence of UI up to age 9. Previous research has reported that breathing disorders are more common among children with enuresis [4]. Sleep disturbance due to breathing problems could increase arousal thresholds, which could cause children to fail to respond to signs that they need to urinate [5]. Nocturnal breathing problems are common between ages 2 and 6 years and adenotonsillar hypertrophy is often the primary cause [31]. Resolution of breathing problems through adenotonsillectomy has been found to reduce enuresis [32]. There is also evidence that breathing problems whilst sleeping might also contribute to an increased risk of daytime LUTS (urgency, frequency, urgency UI), and that adenotonsillectomy can decrease these symptoms [33]. Intermittent hypoxia, due to nocturnal sleep problems, can lead to oxidative stress, which has previously been linked to daytime LUTS [34]. Alternatively, the increased levels of daytime sleepiness and hyperactivity in children with nocturnal breathing problems [35], could lead to a loss of concentration and lack of awareness of full bladder signals.

Reverse causality could provide an alternative explanation for some of the observed associations. The prospective design enabled us to show that sleep problems are present in children before incontinence is reported by parents in their school-age children. This does not rule out the possibility that bladder problems are already present in preschool children. Overactive bladder is a common cause of UI and is caused by overactivity of the detrusor (smooth muscle fibres in the bladder wall) [36]. It is, therefore, possible that detrusor overactivity in preschool children causes arousal from sleep and explains the increased occurrence of sleep problems including difficulty going to sleep and getting up after being put to bed (associated with daytime wetting) and night-time waking (associated with persistent [day and night] wetting). UI at school-age could be a result of ongoing detrusor overactivity since the preschool period.

Another explanation is that the observed associations between sleep problems and UI could be due to common causes such as neurodevelopmental disorders. Attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) are associated with a high prevalence of sleep problems [37, 38] and incontinence (especially daytime wetting) [10]. We adjusted our analysis for preschool factors that are associated with ADHD and ASD including developmental delay, difficult child temperament, and child emotional/behaviour problems, but many of the associations between sleep problems and incontinence persisted.

The cross-sectional study by von Gontard and colleagues found evidence that children with enuresis had fewer night-time wakings than their peers who had attained night-time bladder control [10]. Consistent with this finding, we found that pre-schoolers who experienced night-time waking and those who woke up early were less likely to experience bedwetting alone at school-age. Children with enuresis are frequently regarded by their parents as deep sleepers or having difficulty waking in the morning, leading to the suggestion that they have higher arousal thresholds [11]. There is evidence, however, that children with enuresis experience sleep fragmentation and transient arousals which have been detected using polysomnography and may not be observed by parents [39]. It has also been suggested that cortical arousals, sleep fragmentation and poor-quality sleep may affect heart rate and blood pressure, which could cause excess urine production and bladder overactivity [3].