Modeling Sage data with a truncated gamma-Poisson model

Background  

Serial Analysis of Gene Expressions (SAGE) produces gene expression measurements on a discrete scale, due to the finite number of molecules in the sample. This means that part of the variance in SAGE data should be understood as the sampling error in a binomial or Poisson distribution, whereas other variance sources, in particular biological variance, should be modeled using a continuous distribution function, i.e. a prior on the intensity of the Poisson distribution. One challenge is that such a model predicts a large number of genes with zero counts, which cannot be observed.     

The effect of losartan therapy on ventricular function in Marfan patients with haploinsufficient or dominant negative <Emphasis Type="Italic">FBN1</Emphasis> mutations

Background

Mild biventricular dysfunction is often present in patients with Marfan syndrome. Losartan has been shown to reduce aortic dilatation in patients with Marfan syndrome. This study assesses the effect of losartan on ventricular volume and function in genetically classified subgroups of asymptomatic Marfan patients without significant valvular regurgitation.

Methods

In this predefined substudy of the COMPARE study, Marfan patients were classified based on the effect of their FBN1 mutation on fibrillin-1 protein, categorised as haploinsufficient or dominant negative. Patients were randomised to a daily dose of losartan 100 mg or no additional treatment. Ventricular volumes and function were measured by magnetic resonance imaging at baseline and after 3 years of follow-up.

Results

Changes in biventricular dimensions were assessed in 163 Marfan patients (48?% female; mean age 38 ± 13 years). In patients with a haploinsufficient FBN1 mutation (n = 43), losartan therapy (n = 19) increased both biventricular end diastolic volume (EDV) and stroke volume (SV) when compared with no additional losartan (n = 24): left ventricular EDV: 9 ± 26 ml vs. ?8 ± 24 ml, p = 0.035 and right ventricular EDV 12 ± 23 ml vs. ?18 ± 24 ml; p < 0.001 and for left ventricle SV: 6 ± 16 ml vs. ?8 ± 17 ml; p = 0.009 and right ventricle SV: 8 ± 16 ml vs. ?7 ± 19 ml; p = 0.009, respectively. No effect was observed in patients with a dominant negative FBN1 mutation (n = 92), or without an FBN1 mutation (n = 28).

Conclusion

Losartan therapy in haploinsufficient Marfan patients increases biventricular end diastolic volume and stroke volume, furthermore, losartan also appears to ameliorate biventricular filling properties.

     

Cardiovascular drug trials: how to examine interaction, and why so

Background: In practice the benefit of cardiovascular medicines is less consistent than it is in clinical trials. This is due to multiple uncontrolled factors that co-determine the efficacy of the new treatment. In statistical terms, they interact with the new treatment. Interaction effects are rarely assessed in cardiovascular trials. Objective: To review (1) important factors that may interact with the treatment efficacy, (2) how to examine such factors, and (3) why so. Results: Important factors include (a) possible risk factors such as specific patient characteristics, and concomitant medications, and (b) study-specific aspects such as heterogeneities of investigators, health centres, and individual patients including patient compliance. Such factors can be assessed by comparing subgroups. A common but incorrect approach is the comparison of the significances of difference between treatment modalities in either subgroup. Instead, a direct comparison of effect sizes relative to the standard errors is adequate. As an alternative, regression modelling is adequate and convenient. Results of interaction assessments are post-hoc and, therefore, of an exploratory and unconfirmed nature. So, why should they be performed? In cardiovascular research the effects of patient characteristics and drug-drug interactions on drug efficacies are numerous. It is valuable to account at least post-hoc for such mechanisms. Second, current cardiovascular trials involve heterogeneous health centres, investigators, and patient groups. Accounting for these heterogeneities can be helpful to better predict individual responses in future patients. Conclusion: Cardiovascular trials enrolling patient groups at risk for heterogeneity should include at least a post-hoc assessment for interaction. Correct and incorrect methods for that purpose are described. Interaction assessments are helpful to better predict the efficacy/safety of new cardiovascular medicines in the future treatment of subgroups of patients. (Neth Heart J 2007;15:61-6.)     

PREVENTION-ACHD: PRospEctiVE study on implaNTable cardioverter-defibrillator therapy and suddeN cardiac death in Adults with Congenital Heart Disease; Rationale and Design

Background

Many adult congenital heart disease (ACHD) patients are at risk of sudden cardiac death (SCD). An implantable cardioverter-defibrillator (ICD) may prevent SCD, but the evidence for primary prevention indications is still unsatisfactory.

Study Design

PREVENTION-ACHD is a prospective study with which we aim to prospectively validate a new risk score model for primary prevention of SCD in ACHD patients, as well as the currently existing guideline recommendations. Patients are screened using a novel risk score to predict SCD as well as current ICD indications according to an international Consensus Statement. Patients are followed up for two years. The primary endpoint is the occurrence of SCD and sustained ventricular arrhythmias. The Study was registered at ClinicalTrials.gov (NCT03957824).

Conclusion

PREVENTION-ACHD is the first prospective study on SCD in ACHD patients. In the light of a growing and aging population of patients with more severe congenital heart defects, more robust clinical evidence on primary prevention of SCD is urgently needed.

     

A joint model for repeated events of different types and multiple longitudinal outcomes with application to a follow‐up study of patients after kidney transplant

This paper presents an extension of the joint modeling strategy for the case of multiple longitudinal outcomes and repeated infections of different types over time, motivated by postkidney transplantation data. Our model comprises two parts linked by shared latent terms. On the one hand is a multivariate mixed linear model with random effects, where a low‐rank thin‐plate spline function is incorporated to collect the nonlinear behavior of the different profiles over time. On the other hand is an infection‐specific Cox model, where the dependence between different types of infections and the related times of infection is through a random effect associated with each infection type to catch the within dependence and a shared frailty parameter to capture the dependence between infection types. We implemented the parameterization used in joint models which uses the fitted longitudinal measurements as time‐dependent covariates in a relative risk model. Our proposed model was implemented in OpenBUGS using the MCMC approach.     

The Impact of Allergic Rhinitis and Asthma on Human Nasal and Bronchial Epithelial Gene Expression

Background

The link between upper and lower airways in patients with both asthma and allergic rhinitis is still poorly understood. As the biological complexity of these disorders can be captured by gene expression profiling we hypothesized that the clinical expression of rhinitis and/or asthma is related to differential gene expression between upper and lower airways epithelium.

Objective

Defining gene expression profiles of primary nasal and bronchial epithelial cells from the same individuals and examining the impact of allergic rhinitis with and without concomitant allergic asthma on expression profiles.

Methods

This cross-sectional study included 18 subjects (6 allergic asthma and allergic rhinitis; 6 allergic rhinitis; 6 healthy controls). The estimated false discovery rate comparing 6 subjects per group was approximately 5%. RNA was extracted from isolated and cultured epithelial cells from bronchial brushings and nasal biopsies, and analyzed by microarray (Affymetrix U133+ PM Genechip Array). Data were analysed using R and Bioconductor Limma package. For gene ontology GeneSpring GX12 was used.

Results

The study was successfully completed by 17 subjects (6 allergic asthma and allergic rhinitis; 5 allergic rhinitis; 6 healthy controls). Using correction for multiple testing, 1988 genes were differentially expressed between healthy lower and upper airway epithelium, whereas in allergic rhinitis with or without asthma this was only 40 and 301 genes, respectively. Genes influenced by allergic rhinitis with or without asthma were linked to lung development, remodeling, regulation of peptidases and normal epithelial barrier functions.

Conclusions

Differences in epithelial gene expression between the upper and lower airway epithelium, as observed in healthy subjects, largely disappear in patients with allergic rhinitis with or without asthma, whilst new differences emerge. The present data identify several pathways and genes that might be potential targets for future drug development.     

Methods for analysing cardiovascular studies with repeated measures

Background. Repeated measurements in a single subject are generally more similar than unrepeated measurements in different subjects. Unrepeated analyses of repeated data cause underestimation of the treatment effects. Objective. To review methods adequate for the analysis of cardiovascular studies with repeated measures. Results. (1) For between-subjects comparisons, summary measures and random-effects mixedlinear models are possible. Examples of summary measures include the area under the curve of drug time-concentration and time-efficacy curves, maximal values, mean values, and changes from baseline. A problem is that precision is lost because averages, rather than individual data, are applied. Random-effects mixed-linear models, available in SPSS statistical software and other software programmes, provide better precision for that purpose. (2) For within-subjects comparisons, repeated-measures ANOVAs are available in SPSS and other software programmes. Subgroup factors such as gender differences and age class can be included. Discussion. For non-Gaussian data, Wilcoxon's and Friedman's tests are available, for binary data McNemar's tests can be used in case of two repeated observations. No standard methods are available for repeated binary measures with more than two observations. The purpose of this review was not to present a complete report but, rather, to underline that ample efforts should be made to account for the special nature of repeated measures. (Neth Heart J 2009;17:429–33.)     

Association between Infancy BMI Peak and Body Composition and Blood Pressure at Age 5–6 Years

Introduction

The development of overweight is often measured with the body mass index (BMI). During childhood the BMI curve has two characteristic points: the adiposity rebound at 6 years and the BMI peak at 9 months of age. In this study, the associations between the BMI peak and body composition measures and blood pressure at age 5–6 years were investigated.

Methods

Measurements from the Amsterdam Born Children and their Development (ABCD) study were available for this study. Blood pressure (systolic and diastolic) and body composition measures (BMI, waist-to-height ratio, fat percentage) were gathered during a health check at about 6 years of age (n = 2822). All children had multiple BMI measurements between the 0–4 years of age. For boys and girls separately, child-specific BMI peaks were extracted from mixed effect models. Associations between the estimated BMI peak and the health check measurements were analysed with linear models. In addition, we investigated the potential use of the BMI at 9 months as a surrogate measure for the magnitude of the BMI peak.

Results

After correction for the confounding effect of fetal growth, both timing and magnitude of the BMI peak were significantly and positively associated (p<0.001) with all body composition measures at the age of 5–6 years. The BMI peak showed no direct association with blood pressure at the age 5–6 year, but was mediated by the current BMI. The correlation between the magnitude of the BMI peak and BMI at 9 months was approximately 0.93 and similar associations with the measures at 5–6 years were found.

Conclusion

The magnitude of the BMI peak was associated with body composition measures at 5–6 years of age. Moreover, the BMI at 9 months could be used as surrogate measure for the magnitude of the BMI peak.     

Modelling the correlation between the activities of adjacent genes in drosophila

Background  

Correlation between the expression levels of genes which are located close to each other on the genome has been found in various organisms, including yeast, drosophila and humans. Since such a correlation could be explained by several biochemical, evolutionary, genetic and technological factors, there is a need for statistical models that correspond to specific biological models for the correlation structure.     

Comparing transformation methods for DNA microarray data

Background  

When DNA microarray data are used for gene clustering, genotype/phenotype correlation studies, or tissue classification the signal intensities are usually transformed and normalized in several steps in order to improve comparability and signal/noise ratio. These steps may include subtraction of an estimated background signal, subtracting the reference signal, smoothing (to account for nonlinear measurement effects), and more. Different authors use different approaches, and it is generally not clear to users which method they should prefer.