The role of angiogenic and antiangiogenic factors in the second trimester in the prediction of preeclampsia in pregnant women with type 1 diabetes

OBJECTIVE To assess the association between circulating angiogenic and antiangiogenic factors in the second trimester and risk of preeclampsia in women with type 1 diabetes.

RESEARCH DESIGN AND METHODS Maternal plasma concentrations of placental growth factor (PlGF), soluble fms-like tyrosine kinase 1 (sFlt-1), and soluble endoglin (sEng) were available at 26 weeks of gestation in 540 women with type 1 diabetes enrolled in the Diabetes and Preeclampsia Intervention Trial.

RESULTS Preeclampsia developed in 17% of pregnancies (n = 94). At 26 weeks of gestation, women in whom preeclampsia developed later had significantly lower PlGF (median [interquartile range]: 231 pg/mL [120–423] vs. 365 pg/mL [237–582]; P < 0.001), higher sFlt-1 (1,522 pg/mL [1,108–3,393] vs. 1,193 pg/mL [844–1,630] P < 0.001), and higher sEng (6.2 ng/mL [4.9–7.9] vs. 5.1 ng/mL[(4.3–6.2]; P < 0.001) compared with women who did not have preeclampsia. In addition, the ratio of PlGF to sEng was significantly lower (40 [17–71] vs. 71 [44–114]; P < 0.001) and the ratio of sFlt-1 to PlGF was significantly higher (6.3 [3.4–15.7] vs. 3.1 [1.8–5.8]; P < 0.001) in women who later developed preeclampsia. The addition of the ratio of PlGF to sEng or the ratio of sFlt-1 to PlGF to a logistic model containing established risk factors (area under the curve [AUC], 0.813) significantly improved the predictive value (AUC, 0.850 and 0.846, respectively; P < 0.01) and significantly improved reclassification according to the integrated discrimination improvement index (IDI) (IDI scores 0.086 and 0.065, respectively; P < 0.001).

CONCLUSIONS These data suggest that angiogenic and antiangiogenic factors measured during the second trimester are predictive of preeclampsia in women with type 1 diabetes. The addition of the ratio of PlGF to sEng or the ratio of sFlt-1 to PlGF to established clinical risk factors significantly improves the prediction of preeclampsia in women with type 1 diabetes.

Preeclampsia is characterized by the development of hypertension and new-onset proteinuria during the second half of pregnancy (1,2), leading to increased maternal morbidity and mortality (3). Women with type 1 diabetes are at increased risk for development of preeclampsia during pregnancy, with rates being two-times to four-times higher than that of the background maternity population (4,5). Small advances have come from preventive measures, such as low-dose aspirin in women at high risk (6); however, delivery remains the only effective intervention, and preeclampsia is responsible for up to 15% of preterm births and a consequent increase in infant mortality and morbidity (7).

Although the etiology of preeclampsia remains unclear, abnormal placental vascular remodeling and placental ischemia, together with maternal endothelial dysfunction, hemodynamic changes, and renal pathology, contribute to its pathogenesis (8). In addition, over the past decade accumulating evidence has suggested that an imbalance between angiogenic factors, such as placental growth factor (PlGF), and antiangiogenic factors, such as soluble fms-like tyrosine kinase 1 (sFlt-1) and soluble endoglin (sEng), plays a key role in the pathogenesis of preeclampsia (8,9). In women at low risk (10–13) and women at high risk (14,15), concentrations of angiogenic and antiangiogenic factors are significantly different between women who later develop preeclampsia (lower PlGF, higher sFlt-1, and higher sEng levels) compared with women who do not.

Few studies have specifically focused on circulating angiogenic factors and risk of preeclampsia in women with diabetes, and the results have been conflicting. In a small study, higher sFlt-1 and lower PlGF were reported at the time of delivery in women with diabetes who developed preeclampsia (16). In a longitudinal prospective cohort of pregnant women with diabetes, Yu et al. (17) reported increased sFlt-1 and reduced PlGF in the early third trimester as potential predictors of preeclampsia in women with type 1 diabetes, but they did not show any difference in sEng levels in women with preeclampsia compared with women without preeclampsia. By contrast, Powers et al. (18) reported only increased sEng in the second trimester in women with pregestational diabetes who developed preeclampsia.

The aim of this study, which was significantly larger than the previous studies highlighted, was to assess the association between circulating angiogenic (PlGF) and antiangiogenic (sFlt-1 and sEng) factors and the risk of preeclampsia in women with type 1 diabetes. A further aim was to evaluate the added predictive ability and clinical usefulness of angiogenic factors and established risk factors for preeclampsia risk prediction in women with type 1 diabetes.

Root Cause Analysis by a Combined Sparse Classification and Monte Carlo Approach

Classification methods with embedded feature selection capability are very appealing for the analysis of complex processes since they allow the analysis of root causes even when the number of input variables is high. In this work, we investigate the performance of three techniques for classification within a Monte Carlo strategy with the aim of root cause analysis. We consider the naive bayes classifier and the logistic regression model with two different implementations for controlling model complexity, namely, a LASSO-like implementation with a L1 norm regularization and a fully Bayesian implementation of the logistic model, the so called relevance vector machine. Several challenges can arise when estimating such models mainly linked to the characteristics of the data: a large number of input variables, high correlation among subsets of variables, the situation where the number of variables is higher than the number of available data points and the case of unbalanced datasets. Using an ecological and a semiconductor manufacturing dataset, we show advantages and drawbacks of each method, highlighting the superior performance in term of classification accuracy for the relevance vector machine with respect to the other classifiers. Moreover, we show how the combination of the proposed techniques and the Monte Carlo approach can be used to get more robust insights into the problem under analysis when faced with challenging modelling conditions.

Predictors of place of death for those in receipt of home-based palliative care services in Ontario, Canada

Many cancer patients die in institutional settings despite their preference to die at home. A longitudinal, prospective cohort study was conducted to comprehensively assess the determinants of home death for patients receiving home-based palliative care. Data collected from biweekly telephone interviews with caregivers (n=302) and program databases were entered into a multivariate logistic model. Patients with high nursing costs (odds ratio [OR]: 4.3; confidence interval [CI]: 1.8-10.2) and patients with high personal support worker costs (OR: 2.3; CI: 1.1-4.5) were more likely to die at home than those with low costs. Patients who lived alone were less likely to die at home than those who cohabitated (OR: 0.4; CI: 0.2-0.8), and those with a high propensity for a home-death preference were more likely to die at home than those with a low propensity (OR: 5.8; CI: 1.1-31.3). An understanding of the predictors of place of death may contribute to the development of effective interventions that support home death.

The Development and Testing of a New Version of the Cognitive Reflection Test Applying Item Response Theory (IRT)

The cognitive reflection test (CRT) is a short measure of a person's ability to resist intuitive response tendencies and to produce a normatively correct response, which is based on effortful reasoning. Although the CRT is a very popular measure, its psychometric properties have not been extensively investigated. A major limitation of the CRT is the difficulty of the items, which can lead to floor effects in populations other than highly educated adults. The present study aimed at investigating the psychometric properties of the CRT applying item response theory analyses (a two-parameter logistic model) and at developing a new version of the scale (the CRT-long), which is appropriate for participants with both lower and higher levels of cognitive reflection. The results demonstrated the good psychometric properties of the original, as well as the new scale. The validity of the new scale was also assessed by measuring correlations with various indicators of intelligence, numeracy, reasoning and decision-making skills, and thinking dispositions. Moreover, we present evidence for the suitability of the new scale to be used with developmental samples. Finally, by comparing the performance of adolescents and young adults on the CRT and CRT-long, we report the first investigation into the development of cognitive reflection.

Comment on "correlated noise in a logistic growth model"

We argue the results published by Bao-Quan Ai et al [Phys. Rev E 67, 022903 (2003)] on "correlated noise in a logistic growth model " are not correct. Their conclusion that for larger values of the correlation parameter, lambda, the cell population is peaked at x=0, which denotes the high extinction rate is also incorrect. We find the reverse behaviour corresponding to their results, that increasing lambda, promotes the stable growth of tumour cells. In particular, their results for steady-state probability, as a function of cell number, at different correlation strengths, presented in figures 1 and 2 show different behaviour than one would expect from the simple mathematical expression for the steady-state probability. Additionally, their interpretation at small values of cell number that the steady state probability increases as they increase the correlation parameter is also questionable. Another striking feature in their figures (1 and 3) is that for the same values of the parameter lambda and alpha, their simulation produces two different curves both qualitatively and quantitatively.

Discrete Conditional Phase-type model (DC_Ph) for patient waiting time with a logistic regression component to predict patient admission to hospital

Discrete Conditional Phase-type (DC-Ph) models are a family of models which represent skewed survival data conditioned on specific inter-related discrete variables. The survival data is modeled using a Coxian phase-type distribution which is associated with the inter-related variables using a range of possible data mining approaches such as Bayesian networks (BNs), the Naïve Bayes Classification method and classification regression trees. This paper utilizes the Discrete Conditional Phase-type model (DC-Ph) to explore the modeling of patient waiting times in an Accident and Emergency Department of a UK hospital. The resulting DC-Ph model takes on the form of the Coxian phase-type distribution conditioned on the outcome of a logistic regression model.

Nonlinear time series prediction based on a power-law noise model

In this paper we investigate the influence of a power-law noise model, also called noise, on the performance of a feed-forward neural network used to predict time series. We introduce an optimization procedure that optimizes the parameters the neural networks by maximizing the likelihood function based on the power-law model. We show that our optimization procedure minimizes the mean squared leading to an optimal prediction. Further, we present numerical results applying method to time series from the logistic map and the annual number of sunspots demonstrate that a power-law noise model gives better results than a Gaussian model.

Transform both sides model: A parametric approach

A parametric regression model for right-censored data with a log-linear median regression function and a transformation in both response and regression parts, named parametric Transform-Both-Sides (TBS) model, is presented. The TBS model has a parameter that handles data asymmetry while allowing various different distributions for the error, as long as they are unimodal symmetric distributions centered at zero. The discussion is focused on the estimation procedure with five important error distributions (normal, double-exponential, Student's t, Cauchy and logistic) and presents properties, associated functions (that is, survival and hazard functions) and estimation methods based on maximum likelihood and on the Bayesian paradigm. These procedures are implemented in TBSSurvival, an open-source fully documented R package. The use of the package is illustrated and the performance of the model is analyzed using both simulated and real data sets.