The effects of fish oil plus vitamin D3 intervention on non-alcoholic fatty liver disease: a randomized controlled trial.

PURPOSE: The present study aimed to investigate fish oil plus vitamin D3 (FO + D) supplementation on biomarkers of non-alcoholic fatty liver disease (NAFLD). METHODS: In a 3-month randomized controlled trial, 111 subjects with NAFLD, aged 56.0 ± 15.9 y, were randomized into FO + D group (n = 37), fish oil group (FO, n = 37) or corn oil group (CO, n = 37). The subjects consumed the following capsules (3 g/day), which provided 2.34 g/day of eicosapentaenoic acid (EPA) + docosahexaenoic acid (DHA) + 1680 IU vitamin D3 (FO + D group), or 2.34 g/day of EPA + DHA (FO group), or 1.70 g/d linoleic acid (CO group). RESULTS: Using multivariable-adjusted general linear model, there were significant net reductions in serum alanine aminotransferase (ALT), and triacylglycerol (TAG) and TNF-α levels in the FO + D and FO groups, compared with the control group (P < 0.05). The supplemental FO + D also showed significant reductions in insulin (- 1.58 ± 2.00 mU/L vs. - 0.63 ± 1.55 mU/L, P = 0.050) and IL-1ß (- 6.92 ± 7.29 ng/L vs. 1.06 ± 5.83 ng/L, P < 0.001) in comparison with control group. Although there were no significant differences between FO + D and FO groups regarding biochemical parameters, supplemental FO + D showed decreases in ALT (from 26.2 ± 13.5 U/L to 21.4 ± 9.6 U/L, P = 0.007), aspartate aminotransferase (AST, from 22.5 ± 7.0 U/L to 20.2 ± 4.0 U/L, P = 0.029), HOMA-IR (from 3.69 ± 1.22 to 3.38 ± 1.10, P = 0.047), and TNF-α (from 0.43 ± 0.38 ng/L to 0.25 ± 0.42 ng/L, P < 0.001) levels following the intervention. CONCLUSION: The present study demonstrated that groups supplemented with FO + D and FO had similar beneficial effects on biomarkers of hepatocellular damage and plasma TAG levels in subjects with NAFLD, while in the FO + D group, there were some suggestive additional benefits compared with FO group on insulin levels and inflammation. TRIAL REGISTRATION: ChiCTR1900024866.

Representation learning for clinical time series prediction tasks in electronic health records.

BACKGROUND: Electronic health records (EHRs) provide possibilities to improve patient care and facilitate clinical research. However, there are many challenges faced by the applications of EHRs, such as temporality, high dimensionality, sparseness, noise, random error and systematic bias. In particular, temporal information is difficult to effectively use by traditional machine learning methods while the sequential information of EHRs is very useful. METHOD: In this paper, we propose a general-purpose patient representation learning approach to summarize sequential EHRs. Specifically, a recurrent neural network based denoising autoencoder (RNN-DAE) is employed to encode inhospital records of each patient into a low dimensional dense vector. RESULTS: Based on EHR data collected from Shuguang Hospital affiliated to Shanghai University of Traditional Chinese Medicine, we experimentally evaluate our proposed RNN-DAE method on both mortality prediction task and comorbidity prediction task. Extensive experimental results show that our proposed RNN-DAE method outperforms existing methods. In addition, we apply the "Deep Feature" represented by our proposed RNN-DAE method to track similar patients with t-SNE, which also achieves some interesting observations. CONCLUSION: We propose an effective unsupervised RNN-DAE method to summarize patient sequential information in EHR data. Our proposed RNN-DAE method is useful on both mortality prediction task and comorbidity prediction task.