Night eating and night eating syndrome: associations with dysfunctional eating behaviors, mental health and quality-of-life measures in Australian adults

Night Eating (NE) is defined in the Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5) as excessive food consumption after the evening meal (evening hyperphagia) or eating after awakening from sleep (nocturnal ingestions) and it is the key clinical feature of NE Syndrome (NES). NES is a type of Other Specified Feeding or Eating Disorder (OSFED) in the DSM-5 [1].

The majority of research on NES has been conducted within clinical and high weight settings, where frequent comorbidity with other eating disorders and physical health problems have been reported [2]. Mitchison et al. [3] has reported NES to be particularly common (estimated prevalence 2.7%) in a representative sample of adolescents. They also found it was the most common OSFED in adolescent boys, more so than for girls. A representative study of adults, however, found NES to be more prevalent in women than in men [4]. This study also reported high levels of socioeconomic disadvantage, sleep problems, body weight, and poorer physical health-related quality of life (PHRQoL) relative to others in the general population including people with other eating disorders.

However, there has been a paucity of research in representative adult general population samples examining the relationships between NE or NES and dysfunctional eating disorders, mental health and quality-of-life measures. In addition, there is an incomplete understanding of the relationships between NE or NES and other eating disorders, such as binge eating and use of laxatives, and avoidant or restrictive feeding intake disorders (ARFID), such as restrictive dieting to control shape/weight and use of drugs to reduce appetite/overeating, as well as food avoidance or restriction for other reasons as is found in Avoidant/Restrictive Food Intake Disorder (ARFID). Furthermore, the majority of literature is from clinical samples, and that the relationship between NES and other key eating disorders and eating disorder features remains to be elucidated for a representative sample of adult population. These behaviours particularly binge eating and restrictive dieting are known to be increasing in a representative adult population sample, when co-morbid with high weight, and are a public health concern [5].Studies using clinical samples have found subjects diagnosed with NES more likely to have a comorbid eating disorder, whose prevalence ranges from 5% to 44% [6–8]. In another clinical study, NES was detected in 10.3% of patients with AN, in 34.9% of patients with BN and in 51.7% of patients with BED [9]. Regarding the association between NES and health-related quality of life (HRQoL) using a representative general adult population sample, one study [10] has found a negative association between the two. The present study thus examined the associations of these clinical features (perceived subjective anxiety or depression, mental health-related quality of life (MHRQoL), PHRQoL, binge eating, restrictive dieting to control shape/weight, use of laxatives, use of drugs to reduce appetite/overeating and restrictive dieting to control shape/weight), with NE or NES, as an outcome variable in a representative adult population. For the purpose of the study, the DSM-5 descriptor of NES was used as it was impracticable to include all the proposed criteria by Allison et al. [11] in a large community epidemiological sample. These included questions on whether in the past 3 months, the respondents had any episodes of waking from sleep and eating, or episodes of eating a very large amount of food after evening meal excluding any such events at social gatherings or travelling overseas on a night flight or because of work shifts. NES was defined by at least weekly episodes of NE with ‘a lot’ of distress.

One hundred and sixty-seven (5.75%, 95% CI 4.9–6.6%) adults (aged 18–99 years) had NES. As DSM-5 does not specify exclusion criteria for OSFED, the estimated prevalence of OSFED—NES in this sample was 49(1.69%, 95% CI 1.7–41.7%).

Correlations between the episodes of NE and other symptoms and HRQoL are found in Additional file 2 (Table S1). Table1, also presents using ordinal logistic regression, the associations of MHRQoL, PHRQoL, binge-eating in the last 3 months, purging in the last 3 months, lifetime use of drugs to reduce appetite/overeating, diet/fasting to control shape in the last 3 months and perceived subjective anxiety/depression, with the odds of episodes of NE (n = 167), after controlling for age, sex and body weight. Compared to no NE at all, there were 1.756 times significantly higher odds (OR = 1.756, 95% CI 1.527–2.020, p < 0.001) of NE, with binge-eating episodes in the last 3 months. There were 5.2% and 2.4% significantly lower odds of NE in association with MHRQoL (OR = 0.948, 95% CI 0.921–0.975, p < 0.001) and PHRQoL (OR = 0.976, 95% CI 0.966–0.986, p < 0.001) indicating that the association was stronger with MHRQoL in terms of the effect size.

The present study investigated associations of NE and NES with binge-eating, weight control behavior, including use of purging, dietary restriction and medication, mood symptoms and HRQoL. Binge eating, use of purging, dietary restriction, medication, mood symptoms and HRQoL were captured by the following variables: episodes of binge-eating in the last 3 months, use of purging in the last 3 months, strict dieting/fasting for weight/shape control in the last 3 months, lifetime use of medications to reduce eating/weight, perceived subjective anxiety/depression, MHRQoL and PHRQoL. When adjusting for age, sex and body weight, multivariate ordinal logistic regression analyses found there were significantly higher odds of an increase in episodes of NE, in association with binge-eating episodes and significantly lower odds of increase in episodes of NE, in association with increases in MHRQoL and PHRQoL. With similar adjustments for age, sex and body weight, multivariate binary logistic regression analyses found there were significantly higher odds of having NES, in association with binge eating episodes in the last 3 months and restrictive dieting in the last 3 months and significantly lower odds of NES, in association with increases in MHRQoL. These results support the Muscatello et al. [2] review noting associations and overlap between NES and other eating disorders characterized by recurrent binge-eating, and the reported associations with disorders of restrictive eating. As both NE and binge-eating are symptoms of over or excessive eating, this was not unexpected. However, the findings at a diagnostic level in this study did also support research that has found overlap between NES and disorders characterized by restrictive eating and/or purging behaviors. By “at a diagnostic level” we meant at a level associated with marked distress, the DSM [1] requires either functional impairment or psychological distress to be present as a defining feature of any mental health disorder and distress is a defining feature of NES. Refinements of diagnostic criteria may address this by closer alignment of the descriptors of overeating behaviors and clarification of the boundaries between NES and other eating disorders.

Mental rather than physical health-related quality of life is associated with NES. More research is required to confirm this result and it does not negate the clinical importance of consideration of physical health status of people with NES [2, 4, 15]. This study did not find that those who have anxiety/depression have significantly higher odds of NE, as well as NES. This may have been because the present study did not have an assessment of depression or anxiety using a validated instrument, but rather a broad self-reported experience of current anxiety and/or depression. Other studies have been also more often conduced in clinical populations which may be expected to have higher rates of mental health comorbidities [2] than this general population sample. However, often self-reported anxiety is close to population prevalence rates and is, therefore, appropriate as a quick assessment of likely anxiety. While the use of self-reported depression (SRD) in health surveys and cohorts may be appropriate as a brief assessment of possible depression, due to its moderate–low sensitivity in reference to a validated tool like PHQ-8 for depression [16]. Importantly, SRD has high global agreement (92.7%) with PHQ-8; still the prevalence of depression was higher when assessed with self-reported measurement (7.7%) than with PHQ-8 (5.9%) [16]. The association between perceived subjective anxiety/depression, and NE/NES, may thus have been biased in this study due to the use of self-reported measurement for anxiety/depression.

Although this study included a moderate sized, community-based, adult sample, there were limited participants with diagnostic level NES and thus analyses were also done at the level of NE as a symptom. The potential bias of self-reported data could not be excluded; however, the validity of self-reported data has been supported by several studies and is commonly utilized [17, 18]. Trained lay persons rather than clinicians conducted the interview as is usual in large epidemiological surveys; however, 10% of interviews were audited to ensure fidelity to the interview questions. (This was thus a respondent-based interview not investigator based). Finally, as the study is cross-sectional, causality and bidirectionality could not be investigated.

Notwithstanding these limitations, the present findings have clinical relevance. NES and NE should be considered and asked about in people presenting with other eating disorders, particularly but not limited to disorders of recurrent binge-eating and in those with poorer mental and/or physical health-related quality of life. As the nature of overeating is defined more broadly in NES than in other eating disorders, it is important to explore all forms of overeating, including whether there is loss of control, and not assume NE is absent in people who do not report binge eating.

Those with a history of BE have a higher likelihood of experiencing NE and NES. The likelihood of NE is also increased in those with poorer mental and physical health-related quality of life. The same is found for the association between NES and mental health-related quality of life. Longitudinal research is required to further investigate the associations identified in this paper, particularly whether they are bi- or uni-directional.

Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 2.