Background Dietary advice plays an important role in the management of gestational diabetes (GDM), yet eating behaviours are susceptible to a range of social and environmental factors capable of undermining the utility of dietary advice. The aim of the present study was to establish the direction and strength of any associations between one such factor (sleep) and eating behaviours amongst pregnant women with GDM, and thereby explore the potential benefit of sleep assessment/modification.

Methods Women attending a specialist clinic for the management of GDM were consented to complete the Pittsburgh Sleep Quality Index (PSQI) and Dutch Eating Behaviour questionnaire. Sleep duration, total PSQI score, and each PSQI sub-score were recoded as binary exposure variables by dividing each at/closest to the median value observed. In a similar way data, for the ‘external’, ‘emotional’ and ‘restrained’ eating scores were also recoded as binary outcome variables. Multivariable logistic regression analyses were conducted on n = 52–62 women with complete data, using STATA IC14 to assess the strength and direction of any association(s) between all eight sleep characteristics and each of the three eating behaviours, before and after adjustment for a number of key potential confounders (age, ethnicity, BMI, oral glucose test result and clinical treatment). Given the number of separate analyses this entailed (n = 24) on data generated from a single administration of both questionnaires, a Bonferroni correction was applied to establish a revised α of 0.002.

Results Seven of the 24 analytical models found a statistically significant association between at least one sleep characteristic and each of the three eating behaviour scores, with significance levels ranging from p = 0.048−0.005. This is a far higher number of significant associations than might have been expected on the basis of chance alone ([24,100]’5 = 1.2), and although none of these associations achieved the level of significance recommended following Bonferroni correction (α = 0.002), the strength and consistency of the trends observed do suggest a number of potentially meaningful relationships: women reporting poorer sleep quality, extended sleep latency and worse sleep efficiency having 6.55 (95% CI: 1.76, 24.45), 5.84 (95% CI: 1.41, 24.12) and 3.88 (95% CI: 1.05, 14.29) the odds of being ‘restrained’ eaters, respectively; while women reporting poorer quality sleep also had 4.68 (95% CI: 1.15, 19.01) the odds of being ‘emotional’ eaters. In contrast, women reporting shorter sleep durations and more disturbed sleep had only 0.25 (95% CI: 0.06, 0.99) and 0.20 (95% CI: 0.05, 0.74) the odds of being ‘external’ eaters, respectively

Conclusion These findings suggest that there may be a role for sleep assessment/modification in identifying/supporting eating behaviours conducive to improved glucose control in GDM.