Abstract 17285: Metabolite-Derived Network Reveals Cluster of Acylcholine Metabolites Associated With Better Diet Quality and Lower Prevalence of Type 2 Diabetes: Findings From the Boston Puerto Rican Health Study

Introduction: Puerto Rican (PR) adults living on the US mainland are at high risk for developing type 2 diabetes (T2D), and dietary factors may contribute to this increased risk. Network analysis is a data-reduction tool that can identify correlated clusters of co-regulated metabolites that reflect mechanisms underlying diet-T2D associations. Hypothesis: Diet quality will associate with T2D-associated metabolite clusters among PR adults. Methods: We used LC/MS to measure fasting plasma metabolites (>700) among Boston PR Health Study participants, aged 45-75 years, with (n=258) and without (n=421) T2D. We applied an unsupervised correlation network-based method to identify metabolite clusters within a global metabolite network and calculated a score for each cluster using a weighted sum of metabolite concentrations. To estimate diet quality, we calculated a modified version of a previously validated American Heart Association diet score (AHA-DS). Logistic regression was used to assess cross-sectional associations between metabolite clusters and prevalent T2D, and linear regression was used to assess associations between the continuous AHA-DS and T2D-associated metabolite clusters among controls, adjusting for potential confounders and correcting for multiple testing. Results: We identified 7 metabolite clusters that were associated with prevalent T2D ( p <0.05). For every 1-standard deviation (SD) increase in cluster score, the odds ratios for prevalent T2D and 95% confidence intervals were as the follows: acylcholines [0.40 (0.31, 0.50)], aromatic hydrocarbon derivatives [0.33 (0.22, 0.47)], sphingolipids [0.46 (0.33, 0.64)], tricarboxylic acid (TCA) cycle amino acids/peptides [0.39 (0.25, 0.62)], branched-chain amino acid metabolites [4.1 (2.9, 6.0)], acylcarnitines [1.8 (1.3, 2.5)], and TCA cycle/energy metabolites [2.0 (1.4, 3.0)]. The AHA-DS was only significantly associated with the acylcholine metabolites cluster [β (standard error) = 0.01 (0.004) SD increase in cluster score, p=0.02]. Conclusions: In individuals of PR descent, we identified a cluster of acylcholine metabolites where concentrations are higher among those with better diet quality and lower among those with prevalent T2D.

Feasibility Trial of an eHealth Intervention for Health-Related Quality of Life: Implications for Managing Patients with Chronic Pain During the COVID-19 Pandemic

Purpose: This study was conducted to determine the feasibility of providing an eHealth intervention for health-related quality of life (HRQOL) to facilitate patient self-management. Methods: A randomized controlled trial was conducted from 2019–2020 within the Pain Registry for Epidemiological, Clinical, and Interventional Studies and Innovation. Eligible patients included those with chronic low back pain and a SPADE (sleep disturbance, pain interference with activities, anxiety, depression, and low energy/fatigue) cluster score ≥ 55 based on the relevant scales from the Patient-Reported Outcomes Measurement Information System instrument with 29 items (PROMIS-29). Patients were randomized to the eHealth treatment group, which received a tailored HRQOL report and interpretation guide, or to a wait-list control group. The primary outcome was change in the SPADE cluster score, including its five component scales, over 3 months. Secondary outcomes were changes in low back pain intensity and back-related disability. Treatment effects were measured using the standardized mean difference (SMD) in change scores between groups. The eHealth intervention was also assessed by a survey of the experimental treatment group 1 month following randomization. Results: A total of 102 patients were randomized, including 52 in the eHealth treatment group and 50 in the wait-list control group, and 100 (98%) completed the trial. A majority of patients agreed that the HRQOL report was easy to understand (86%), provided new information (79%), and took actions to read or learn more about self-management approaches to improve their HRQOL (77%). Although the eHealth intervention met the criteria for a small treatment effect in improving the overall SPADE cluster score (SMD = 0.24; p= 0.23) and anxiety (SMD = 0.24; p = 0.23), and for a small-to-medium treatment effect in improving depression (SMD = 0.37; p = 0.06) and back-related disability (SMD = 0.36; p = 0.07), none of these results achieved statistical significance because of limited sample size. Conclusion: Given the feasibility of rapid online deployment, low cost, and low risk of adverse events, this eHealth intervention for HRQOL may be useful for patients with chronic pain during the COVID-19 pandemic.

Generating example contexts to help children learn word meaning

This article addresses the problem of generating good example contexts to help children learn vocabulary. We describe VEGEMATIC, a system that constructs such contexts by concatenating overlapping five-grams from Google's N-gram corpus. We propose and operationalize a set of constraints to identify good contexts. VEGEMATIC uses these constraints to filter, cluster, score, and select example contexts. An evaluation experiment compared the resulting contexts against human-authored example contexts (e.g., from children's dictionaries and children's stories). Based on rating by an expert blind to source, their average quality was comparable to story sentences, though not as good as dictionary examples. A second experiment measured the percentage of generated contexts rated by lay judges as acceptable, and how long it took to rate them. They accepted only 28% of the examples, but averaged only 27 seconds to find the first acceptable example for each target word. This result suggests that hand-vetting VEGEMATIC's output may supply example contexts faster than creating them manually.

Switching from Antipsychotic to Quetiapine XR Treatment in Patients with Schizophrenia and High from Low Depression Symptoms

Purpose:Assess the efficacy of once-daily extended release quetiapine fumarate (quetiapine XR) in patients with schizophrenia and depressive symptoms.Methods:12-week, multi-centre, open-label study in adult patients with schizophrenia (D1444C00147). Patients were cross titrated to quetiapine XR over three days (Day 1: 300 mg; Day 2: 600 mg; Days 3-84: 400-800 mg/day [flexible dosing]). Two patient subsets: high depression symptoms (PANSS depression cluster score ≥ 12) and low depression symptoms (PANSS depression cluster score < 12) at baseline. Change from baseline of PANSS total score and PANSS depression cluster score were analysed by subgroup.Results:Mean (SD) change at Day 84 (LOCF) in PANSS depression cluster score was -4.8 (3.3) for patients with high depression (n=109) and -1.1 (2.7) for patients with low depression (n=362). Mean (SD) change from baseline at Day 84 (LOCF) in PANSS total score was -24.3 (18.5) for patients with high depression and -10.5 (18.0) for patients with low depression. at Day 84: in patients switched to quetiapine XR owing to “lack of efficacy”, mean change (SD) from baseline in PANSS depression cluster score was -5.2 (3.2) and -1.3 (2.8) for patients with high and low depression, respectively; in patients switched due to “insufficient tolerability”, mean change (SD) in PANSS depression cluster score was -4.1 (3.5) and -0.6 (2.5) for patients with high and low, respectively.Conclusions:Quetiapine XR showed promising results in patients with schizophrenia and depressive symptoms switched from other antipsychotics due to suboptimal efficacy or tolerability. Further randomised, clinical trials are warranted.

Clustering of Lifestyle CVD Risk Factors and Its Relationship with Biological CVD Risk Factors

The purpose of this study was (a) to investigate whether lifestyle risk factors cluster and (b) to investigate the influence of this clustering on biological CVD risk factors. This study was part of the Amsterdam Growth and Health Study (AGHS), an observational longitudinal study in which 6 repeated measurements were carried out on 181 13-year-old subjects over a period of 15 years. A longitudinal analysis (carried out with generalized estimating equations) showed no significant clustering of lifestyle risk factors at the population level. For each subject at each separate measurement period, lifestyle risk factors were summed to form a cluster score. A longitudinal linear regression analysis showed no significant relationship between the cluster score and biological CVD risk factors, except for a significant inverse relationship with cardiopulmonary fitness. In general, however, the results did not support the assumption that clustering of unhealthy lifestyle is related to biological CVD risk factors.