Neural Network Prediction of ICU Length of Stay Following Cardiac Surgery Based on Pre-Incision Variables

Background

Advanced predictive analytical techniques are being increasingly applied to clinical risk assessment. This study compared a neural network model to several other models in predicting the length of stay (LOS) in the cardiac surgical intensive care unit (ICU) based on pre-incision patient characteristics.

Methods

Thirty six variables collected from 185 cardiac surgical patients were analyzed for contribution to ICU LOS. The Automatic Linear Modeling (ALM) module of IBM-SPSS software identified 8 factors with statistically significant associations with ICU LOS; these factors were also analyzed with the Artificial Neural Network (ANN) module of the same software. The weighted contributions of each factor (“trained” data) were then applied to data for a “new” patient to predict ICU LOS for that individual.

Results

Factors identified in the ALM model were: use of an intra-aortic balloon pump; O₂ delivery index; age; use of positive cardiac inotropic agents; hematocrit; serum creatinine ≥ 1.3 mg/deciliter; gender; arterial pCO₂. The r² value for ALM prediction of ICU LOS in the initial (training) model was 0.356, p <0.0001. Cross validation in prediction of a “new” patient yielded r² = 0.200, p <0.0001. The same 8 factors analyzed with ANN yielded a training prediction r² of 0.535 (p <0.0001) and a cross validation prediction r² of 0.410, p <0.0001. Two additional predictive algorithms were studied, but they had lower prediction accuracies. Our validated neural network model identified the upper quartile of ICU LOS with an odds ratio of 9.8(p <0.0001).

Conclusions

ANN demonstrated a 2-fold greater accuracy than ALM in prediction of observed ICU LOS. This greater accuracy would be presumed to result from the capacity of ANN to capture nonlinear effects and higher order interactions. Predictive modeling may be of value in early anticipation of risks of post-operative morbidity and utilization of ICU facilities.     

MeCP2 Affects Skeletal Muscle Growth and Morphology through Non Cell-Autonomous Mechanisms

Rett syndrome (RTT) is an autism spectrum disorder mainly caused by mutations in the X-linked MECP2 gene and affecting roughly 1 out of 10.000 born girls. Symptoms range in severity and include stereotypical movement, lack of spoken language, seizures, ataxia and severe intellectual disability. Notably, muscle tone is generally abnormal in RTT girls and women and the Mecp2-null mouse model constitutively reflects this disease feature. We hypothesized that MeCP2 in muscle might physiologically contribute to its development and/or homeostasis, and conversely its defects in RTT might alter the tissue integrity or function. We show here that a disorganized architecture, with hypotrophic fibres and tissue fibrosis, characterizes skeletal muscles retrieved from Mecp2-null mice. Alterations of the IGF-1/Akt/mTOR pathway accompany the muscle phenotype. A conditional mouse model selectively depleted of Mecp2 in skeletal muscles is characterized by healthy muscles that are morphologically and molecularly indistinguishable from those of wild-type mice raising the possibility that hypotonia in RTT is mainly, if not exclusively, mediated by non-cell autonomous effects. Our results suggest that defects in paracrine/endocrine signaling and, in particular, in the GH/IGF axis appear as the major cause of the observed muscular defects. Remarkably, this is the first study describing the selective deletion of Mecp2 outside the brain. Similar future studies will permit to unambiguously define the direct impact of MeCP2 on tissue dysfunctions.     

Levosimendan Inhibits Peroxidation in Hepatocytes by Modulating Apoptosis/Autophagy Interplay

Background

Levosimendan protects rat liver against peroxidative injuries through mechanisms related to nitric oxide (NO) production and mitochondrial ATP-dependent K (mitoK_ATP) channels opening. However, whether levosimendan could modulate the cross-talk between apoptosis and autophagy in the liver is still a matter of debate. Thus, the aim of this study was to examine the role of levosimendan as a modulator of the apoptosis/autophagy interplay in liver cells subjected to peroxidation and the related involvement of NO and mitoK_ATP.

Methods and Findings

In primary rat hepatocytes that have been subjected to oxidative stress, Western blot was performed to examine endothelial and inducible NO synthase isoforms (eNOS, iNOS) activation, apoptosis/autophagy and survival signalling detection in response to levosimendan. In addition, NO release, cell viability, mitochondrial membrane potential and mitochondrial permeability transition pore opening (MPTP) were examined through specific dyes. Some of those evaluations were also performed in human hepatic stellate cells (HSC). Pre-treatment of hepatocytes with levosimendan dose-dependently counteracted the injuries caused by oxidative stress and reduced NO release by modulating eNOS/iNOS activation. In hepatocytes, while the autophagic inhibition reduced the effects of levosimendan, after the pan-caspases inhibition, cell survival and autophagy in response to levosimendan were increased. Finally, all protective effects were prevented by both mitoK_ATP channels inhibition and NOS blocking. In HSC, levosimendan was able to modulate the oxidative balance and inhibit autophagy without improving cell viability and apoptosis.

Conclusions

Levosimendan protects hepatocytes against oxidative injuries by autophagic-dependent inhibition of apoptosis and the activation of survival signalling. Such effects would involve mitoK_ATP channels opening and the modulation of NO release by the different NOS isoforms. In HSC, levosimendan would also play a role in cell activation and possible evolution toward fibrosis. These findings highlight the potential of levosimendan as a therapeutic agent for the treatment or prevention of liver ischemia/reperfusion injuries.     

Cytokine Profiles at Admission Can Be Related to Outcome in AIDS Patients with Cryptococcal Meningitis

Cryptococcal meningitis (CM) remains as common life-threatening AIDS-defining illness mainly in resource-limited settings. Previous reports suggested that baseline cytokine profiles can be associated to fungal burden and clinical outcome. This study aimed to evaluate the baseline cytokine profiles in AIDS patients with CM and its relation with the outcome at weeks 2 and 10. Thirty AIDS patients with CM diagnosed by cerebrospinal fluid (CSF) Cryptococcus neoformans positive culture, India ink stain and cryptococcal antigen test were prospectively evaluated. As controls, 56 HIV-infected patients without CM and 48 non-HIV individuals were included. Baseline CSF and sera levels of IL-2, IL-4, IL-8, IL-10, IL-12p40, IL-17A, INF-γ and TNF-α were measured by ELISA. Of 30 CM patients, 24 (80%) were male, median age of 38.1. The baseline CSF high fungal burden and positive blood culture were associated with a positive CSF culture at week 2 (p = 0.043 and 0.029). Most CSF and sera cytokines presented higher levels in CM patients than control subjects (p < 0.05). CSF levels of IL-8, IL-12p40, IL-17A, TNF-α, INF-γ and sera TNF-α were significantly higher among survivors at weeks 2 and 10 (p < 0.05). Patients with increased intracranial pression exhibited CSF IL-10 high levels and poor outcome at week 10 (p = 0.032). Otherwise, baseline CSF log10 IFN-γ and IL-17A were negatively correlated with fungal burden (r = -0.47 and -0.50; p = 0.0175 and 0.0094, respectively). The mortality rate was 33% (10/30) at week 2 and 57% (17/30) at week 10. The severity of CM and the advanced immunodeficiency at admission were related to a poor outcome in these patients. Otherwise, the predominant Th1 cytokines profile among survivors confirms its pivotal role to infection control and would be a prognostic marker in cryptococcal meningitis.     

Conformational Changes in a Hyperthermostable Glycoside Hydrolase: Enzymatic Activity Is a Consequence of the Loop Dynamics and Protonation Balance

Endo-β-1, 4-mannanase from Thermotoga petrophila (TpMan) is a modular hyperthermostable enzyme involved in the degradation of mannan-containing polysaccharides. The degradation of these polysaccharides represents a key step for several industrial applications. Here, as part of a continuing investigation of TpMan, the region corresponding to the GH5 domain (TpManGH5) was characterized as a function of pH and temperature. The results indicated that the enzymatic activity of the TpManGH5 is pH-dependent, with its optimum activity occurring at pH 6. At pH 8, the studies demonstrated that TpManGH5 is a molecule with a nearly spherical tightly packed core displaying negligible flexibility in solution, and with size and shape very similar to crystal structure. However, TpManGH5 experiences an increase in radius of gyration in acidic conditions suggesting expansion of the molecule. Furthermore, at acidic pH values, TpManGH5 showed a less globular shape, probably due to a loop region slightly more expanded and flexible in solution (residues Y88 to A105). In addition, molecular dynamics simulations indicated that conformational changes caused by pH variation did not change the core of the TpManGH5, which means that only the above mentioned loop region presents high degree of fluctuations. The results also suggested that conformational changes of the loop region may facilitate polysaccharide and enzyme interaction. Finally, at pH 6 the results indicated that TpManGH5 is slightly more flexible at 65°C when compared to the same enzyme at 20°C. The biophysical characterization presented here is well correlated with the enzymatic activity and provide new insight into the structural basis for the temperature and pH-dependent activity of the TpManGH5. Also, the data suggest a loop region that provides a starting point for a rational design of biotechnological desired features.     

“Once upon a Time in the Mediterranean” Long Term Trends of Mediterranean Fisheries Resources Based on Fishers’ Traditional Ecological Knowledge

We investigate long-term changes in the Mediterranean marine resources driving the trawl fisheries by analysing fishers’ perceptions (Traditional Ecological Knowledge, TEK) throughout the Mediterranean Sea during the last 80 years. To this end, we conducted an extended set of interviews with experienced fishers that enabled us to classify species (or taxa) as ‘decreasing’ or ‘increasing’ both in terms of abundance, as well as average size in the catch. The aspect that most clearly emerged in all the investigated areas over time was the notable increase of fishing capacity indicators, such as engine power and fishing depth range. Atlantic mackerel, poor cod, scorpionfishes, striped seabream, and John Dory demonstrated a decreasing trend in the fishers’ perceived abundance, while Mediterranean parrotfish, common pandora, cuttlefish, blue and red shrimp, and mullets gave indications of an increasing temporal trend. Although, as a rule, trawler captains did not report any cataclysmic changes (e.g. extinctions), when they were invited to estimate total catches, a clear decreasing pattern emerged; this being a notable finding taking into account the steep escalation of fishing efficiency during the past century. The overall deteriorating status of stocks in most Mediterranean regions calls for responsible management and design of rebuilding plans. This should include historical information accounting for past exploitation patterns that could help defining a baseline of fish abundance prior to heavy industrial fisheries exploitation.     

Detection of Peptide-Based Nanoparticles in Blood Plasma by ELISA

AimsThe aim of the current study was to develop a method to detect peptide-linked nanoparticles in blood plasma.ResultsThe ELISA based method for the detection of FITC labeled peptides had a detection limit of 1 ng/mL. We were able to accurately measure peptides bound to pentafluorophenyl methacrylate nanoparticles in blood plasma of rats, and similar results were obtained by LC/MS.ConclusionsWe detected FITC-labeled peptides on pentafluorophenyl methacrylate nanoparticles after injection in vivo. This method can be extended to detect nanoparticles with different chemical compositions.