RESUMO
BACKGROUND: Enteral nutrition may affect risks of gastrointestinal bleeding, pneumonia and mortality in critically ill patients and may also modify the effects of pharmacological stress ulcer prophylaxis. We undertook post hoc analyses of the stress ulcer prophylaxis in the intensive care unit trial to assess for any associations and interactions between enteral nutrition and pantoprazole. METHODS: Extended Cox models with time-varying co-variates and competing events were used to assess potential associations, adjusted for baseline severity of illness. Potential interactions between daily enteral nutrition and allocation to pantoprazole on outcomes were similarly assessed. RESULTS: Enteral nutrition was associated with lower risk of clinically important gastrointestinal bleeding (cause-specific hazard ratio [HR]: 0.29, 95% confidence interval: [CI] 0.19-0.44, p < .001), higher risk of pneumonia (HR: 1.44, 95% CI: 1.14-1.82, p = .003), and lower risk of all-cause mortality (HR: 0.22, 95% CI: 0.18-0.27, p < .001). Enteral nutrition with allocation to pantoprazole was associated with a lower risk of mortality (HR: 0.27, 95% CI: 0.21-0.35, p < .001), similar to enteral nutrition with allocation to placebo (HR: 0.17, 95% CI: 0.13-0.23, p < .001). Allocation to pantoprazole with no enteral nutrition had little effect on mortality (HR: 0.83, 95% CI: 0.63-1.09, p = .179), whilst allocation to pantoprazole and receipt of enteral nutrition was mostly compatible with increased all-cause mortality (HR: 1.27, 95% CI: 0.99-1.64, p = .061). The test of interaction between enteral nutrition and pantoprazole treatment allocation for all-cause mortality was statistically significant (p = .024). CONCLUSIONS: Enteral nutrition was associated with an increased risk of pneumonia and a reduced risk of gastrointestinal bleeding. The interaction between pantoprazole and enteral nutrition suggesting an increased risk of mortality requires further study.
RESUMO
We validated a Deep Embedded Clustering (DEC) model and its adaptation for integrating mixed datatypes (in this study, numerical and categorical variables). Deep Embedded Clustering (DEC) is a promising technique capable of managing extensive sets of variables and non-linear relationships. Nevertheless, DEC cannot adequately handle mixed datatypes. Therefore, we adapted DEC by replacing the autoencoder with an X-shaped variational autoencoder (XVAE) and optimising hyperparameters for cluster stability. We call this model "X-DEC". We compared DEC and X-DEC by reproducing a previous study that used DEC to identify clusters in a population of intensive care patients. We assessed internal validity based on cluster stability on the development dataset. Since generalisability of clustering models has insufficiently been validated on external populations, we assessed external validity by investigating cluster generalisability onto an external validation dataset. We concluded that both DEC and X-DEC resulted in clinically recognisable and generalisable clusters, but X-DEC produced much more stable clusters.