Introduction: Postpartum weight management is difficult for many mothers due to the demands of parenthood. Women have highlighted a need for support but experience barriers to engaging with lifestyle interventions hence more adaptable approaches are required. This work examined participants’ engagement with a 12-month, theory-based, automated, text message (SMS) delivered, intervention supporting postpartum weight management.

Methods: SMS content was informed by: 1) the ‘Health Action Process Approach’ (HAPA)1; 2) behaviour change techniques associated with effectiveness in weight management interventions2; 3) women's accounts of postpartum weight-related experiences; and 4) personal and public involvement. A two-arm pilot RCT recruited women within two years postpartum, with a BMI ≥ 25 kg/m2, through community sources and social media. Women were randomised via a secure remote system to receive the intervention or an active control delivering child development messages. Participants received 353 messages during the 12 month intervention. Two-way messages were used to assess engagement: 50 messages prompted women to respond with their weight; 36 interactive messages requested participants’ to respond ‘Yes/No’ to a question which then triggered a feedback message. Participant engagement with two-way messages was calculated as a percentage of replies sent by women and was categorised as ‘high’ or ‘low’ according to the median number of replies sent. Weight was measured at 0, 3, 6, 9 and 12 months.

Results: 51 of 100 women recruited were randomised to receive the intervention. In months 0–6, (47%) and (95%) of participants responded to the weight messages and the ‘Yes/No’ messages respectively. In months 7–12, the responses were (77%) and (86%) respectively. Participants who were high engagers with weight messages had greater mean weight loss compared with low engagers at all time points: at 12 months high engagers (n = 18) lost -2.66 kg and had a reduction in waist circumference at 12 months of -8.9 cm, compared to changes in low engagers (n = 18) of -0.84 kg and -3.6 cm. Likewise, high engagers with ‘Yes/No’ messages had greater mean weight loss compared with low engagers at all time points: at 12 months high engagers (n = 16) lost -2.87 kg and had a reduction in waist circumference of -9.4 cm, compared to changes in low engagers (n = 20) of -0.86 kg and -3.6 cm.

Discussion: The use of two-way text messages was a useful way to encourage engagement with this SMS-delivered intervention. Higher engagement resulted in more weight loss compared to low engagement.

Food Based Dietary Guidelines (FBDG) are a set of recommendations that describe the quantity and types of foods to consume to promote healthy eating and prevent diseases such as obesity. However, when assessing compliance with FBDG, calculating contributions from composite dishes is challenging, since the specific recipe ingredients are often unknown. This project aims to establish proportional contribution of composite dishes to food groups defined by the Irish and UK FBDG. This will facilitate automated assessment of compliance with FBDG for users of novel technology such as Nutritics, a suite of integrated nutrition analysis software tools for healthcare professionals, educators and industries. Territory specific food composition data (n = 3291) for Ireland and the UK were downloaded from the 2015 McCance & Widdowson Composition of Foods Integrated Dataset (CoFIDs). Recipes were identified and classified into groups broadly aligned with the food groups defined in the Healthy Food for Life Ireland and the Eatwell Guide UK. Supplementary recipe details were accessed from McCance & Widdowson 7th Edition book and online resources. Recipes were categorised by recipe type and ingredients were categorised into food groups. Percentage contribution of each food group to the total recipe was calculated. Of the 3,291 foods, 1,108 were classified as recipes, details were available for 138 of these. Of the 138 recipes, there were fruit & vegetable dishes (n = 20), breads, potatoes, pasta and rice dishes (n = 12), meat, fish and alternative protein sources dishes (n = 40), dairy dishes (n = 23), high fat/oil, sugar, salt dishes (n = 43). For fruit & vegetable dishes, the median percentage contribution to the fruit & vegetable food group was 67% (15% min/ 100% max). For breads, potatoes, pasta and rice dishes the median percentage contribution to the breads, potatoes, pasta and rice food group was 53% (17% min/ 89% max). For meat, fish and alternative protein sources dishes, the median percentage contribution to meat, fish and alternative protein sources food group was 55% (16% min/ 85% max), for dairy dishes, the median percentage contribution to the dairy food group was 90% (53% min/ 96% max). For high fat/oil, sugar, salt dishes the median percentage contribution to the high fat/oil, sugar, salt food group was 22% (3% min/ 97% max). The calculation of recipes into percentage contribution to food groups can support assessment of adherence to FBDG when using reference recipes. This is a useful tool to support healthcare professionals when assessing dietary intake where specific recipe components are unknown.

Reduced inflammatory signaling (IL-1RI-/-) alters metabolic responses to dietary challenges (1). Inflammasome deficiency (e.g. IL-18-/-, Asc-/-) can modify gut microbiota concomitant with hepatosteatosis; an effect that was transferable to wild-type (WT) mice by co-housing (2). Taken together, this evidence suggests that links between diet, microbiota and IL-1RI-signaling can influence metabolic health. Our aim was to determine whether IL-1RI-mediated signaling interacted with the gut microbiome to impact metabolic tissue functionality in a diet-specific fashion. Male WT (C57BL/J6) and IL-1RI-/- mice were fed either high-fat diet (HFD; 45% kcal) or low-fat diet (LFD; 10% kcal) for 24 weeks and were housed i) separately by genotype or ii) with genotypes co-housed together (i.e. isolated vs shared microbial environment; n = 8–10 mice per group). Glucose tolerance and insulin secretion response (1.5 g/kg i.p.), gut microbiota composition and caecal short-chain fatty acids (SCFA) were assessed. Liver and adipose tissue were harvested and examined for triacylglycerol (TAG) formation, cholesterol and metabolic markers (Fasn, Cpt1α, Pparg, Scd1, Dgat1/2), using histology, gas-chromatography and RT-PCR, respectively. Statistical analysis included 1-way or 2-way ANOVA, where appropriate, with Bonferroni post-hoc correction. Co-housing significantly affected gut microbiota composition, illustrated by clustering in PCoA (unweighted UniFrac distance) of co-housed mice but not their single-housed counterparts, on both HFD and LFD. The taxa driving these differences were primarily from Lachnospiraceae and Ruminococcaceae families. Single-housed WT had lower hepatic weight, TAG, cholesterol levels and Fasn despite HFD, an effect lost in their co-housed counterparts, who aligned more to IL-1RI-/- hepatic lipid status. Hepatic Cpt1α was lowest in co-housed WT. Adipose from IL-1RI-/- groups on HFD displayed increased adipocyte size and reduced adipocyte number compared to WT groups, but greater lipogenic potential (Pparg, Scd1, Dgat2) alongside a blunted IL-6 response to pro-inflammatory stimuli (~32%, P = 0.025). Whilst caecal SCFA concentrations were not different between groups, single-housed IL-1RI-/- adipocytes showed greatest sensitivity to SCFA-induced lipogenesis. Interestingly, differences in tissue functionality and gut microbiome occurred despite unaltered glucose tolerance; although there was a trend for phenotypic transfer of body weight via co-housing. For all endpoints examined, similar genotype/co-housing effects were observed for both HFD and LFD with the greatest impacts seen in HFD-fed mice. In conclusion, while the gut microbiome may be an important consideration in dietary interventions, these results question the magnitude of its impact in relation to the IL-1RI-dependent immunometabolism-glucose homeostasis axis.

Leukocyte DNA methylation patterns associated with habitual diet may reveal molecular mechanisms involved in the pathogenesis of diet-related chronic diseases and highlight targets for prevention and treatment. We aimed to examine peripheral blood derived leukocyte DNA methylation signatures associated with diet quality. We meta-analyzed epigenome-wide associations between diet quality and DNA methylation levels at over 400,000 cytosine-guanine dinucleotides (CpGs). We conducted analysis primarily in 6,662 European ancestry (EA) participants and secondarily in a group additionally including 3,062 participants of non-European ancestry from five population-based cohort studies. DNA methylation profiles were measured in whole blood, CD4 + T-cells, or CD14 + monocytes. We used food frequency questionnaires to assess habitual intake and constructed two diet quality scores: the Mediterranean-style diet score (MDS) and Alternative Healthy Eating Index (AHEI). Our primary analysis identified 32 diet-associated CpGs, 12 CpGs for MDS and 24 CpGs for AHEI (at FDR < 0.05, corresponding p-values = 1.2×10-6 and 3.1×10-6, respectively) in EA participants. Four of these CpGs were associated with both MDS and AHEI. In addition, Mendelian randomization analysis indicated that seven diet-associated CpGs were causally linked to at least one of the CVD risk factors. For example, hypermethylation of cg11250194 (FADS2), which was associated with higher diet quality scores, was also associated with lower fasting triglycerides concentrations (p-value = 1.5×10-14) and higher high-density lipoprotein cholesterol concentrations (p-value = 1.7×10-8). Transethnic meta-analysis identified nine additional CpGs associated with diet quality (either MDS or AHEI) at FDR < 0.05. Overall quality of habitual diet was associated with differential peripheral leukocyte DNA methylation levels of 32 CpGs in EA participants. The diet-associated CpGs may serve as biomarkers and targets for preventive measures in CVD health. Future studies are warranted to examine diet-associated DNA methylation patterns in larger, ethnically diverse study samples.