A Simplification and Implementation of Random-effects Meta-analyses Based on the Exact Distribution of Cochran’s Q

Background: The random-effects (RE) model is the standard choice for meta-analysis in the presence of heterogeneity, and the stand ard RE method is the DerSimonian and Laird (DSL) approach, where the degree of heterogeneity is estimated using a moment-estimator. The DSL approach does not take into account the variability of the estimated heterogeneity variance in the estimation of Cochran’s Q. Biggerstaff and Jackson derived the exact cumulative distribution function (CDF) of Q to account for the variability of Ť 2.

Objectives: The first objective is to show that the explicit numerical computation of the density function of Cochran’s Q is not required. The second objective is to develop an R package with the possibility to easily calculate the classical RE method and the new exact RE method.

Methods: The novel approach was validated in extensive simulation studies. The different approaches used in the simulation studies, including the exact weights RE meta-analysis, the I 2 and T 2 estimates together with their confidence intervals were implemented in the R package metaxa.

Results: The comparison with the classical DSL method showed that the exact weights RE meta-analysis kept the nominal type I error level better and that it had greater power in case of many small studies and a single large study. The Hedges RE approach had inflated type I error levels. Another advantage of the exact weights RE meta-analysis is that an exact confidence interval for T 2is readily available. The exact weights RE approach had greater power in case of few studies, while the restricted maximum likelihood (REML) approach was superior in case of a large number of studies. Differences between the exact weights RE meta-analysis and the DSL approach were observed in the re-analysis of real data sets. Application of the exact weights RE meta-analysis, REML, and the DSL approach to real data sets showed that conclusions between these methods differed.

Conclusions: The simplification does not require the calculation of the density of Cochran’s Q, but only the calculation of the cumulative distribution function, while the previous approach required the computation of both the density and the cumulative distribution function. It thus reduces computation time, improves numerical stability, and reduces the approximation error in meta-analysis. The different approaches, including the exact weights RE meta-analysis, the I 2 and T 2estimates together with their confidence intervals are available in the R package metaxa, which can be used in applications.

Characteristics of rapid vs slow progression to type 1 diabetes in multiple islet autoantibody-positive children

AIMS/HYPOTHESIS

Islet autoantibody-positive children progress to type 1 diabetes at variable rates. In our study, we asked whether characteristic autoantibody and/or gene profiles could be defined for phenotypes showing extreme progression.

METHODS

Autoantibodies to insulin (IAA), GAD (GADA), insulinoma-associated antigen-2 (IA-2A) and zinc transporter 8 (ZnT8A) were measured in follow-up sera, and genotyping for type 1 diabetes susceptibility genes (HLA-DR/HLA-DQ, INS variable number of tandem repeats [VNTR] and single nucleotide polymorphisms at PTPN22, PTPN2, ERBB3, IL2, SH2B3, CTLA4, IFIH1, KIAA0350 [also known as CLEC16A], CD25, IL18RAP, IL10, COBL) was performed on the DNA samples of children born to a parent with type 1 diabetes and prospectively followed from birth for up to 22 years.

RESULTS

Of the 1,650 children followed, 23 developed multiple autoantibodies and progressed to diabetes within 3 years (rapid progressors), while 24 children developed multiple autoantibodies and remained non-diabetic for more than 10 years from seroconversion (slow progressors). Rapid and slow progressors were similar with respect to HLA-DR/HLA-DQ genotypes, development of IAA, GADA and ZnT8A, and progression to multiple autoantibodies. In contrast, IA-2A development was considerably delayed in the slow progressors. Furthermore, both groups were effectively distinguished by the combined presence or absence of type 1 diabetes susceptibility alleles of non-HLA genes, most notably IL2, CD25, INS VNTR, IL18RAP, IL10, IFIH1 and PTPN22, and discrimination was improved among children carrying high-risk HLA-DR/HLA-DQ genotypes.

CONCLUSIONS/INTERPRETATION

Our data suggest that genotypes of non-HLA type 1 diabetes susceptibility genes influence the likelihood or rate of diabetes progression among children with multiple islet autoantibodies.

Challenges in planning and conducting diagnostic studies with molecular biomarkers

Biomarkers are of increasing importance for personalized medicine in many areas of application, such as diagnosis, prognosis, or the selection of targeted therapies. In many molecular biomarker studies, intensity values are obtained from large scale ‑omics experiments. These intensity values, such as protein concentrations, are often compared between at least two groups of subjects to determine the diagnostic ability of the molecular biomarker. Various prospective or retrospective study designs are available for molecular biomarker studies, and the biomarker used may be univariate or may even consist in a multimarker rule. In this work, several challenges are discussed for the planning and conduct of biomarker studies. The phases of diagnostic biomarker studies are closely related to levels of evidence in diagnosis, and they are therefore discussed upfront. Different study designs for molecular biomarker studies are discussed, and they primarily differ in the way subjects are selected. Using two systematic reviews from the literature, common sources of bias of molecular diagnostic studies are illustrated. The extreme selection of patients and controls and verification bias are specifically discussed. The pre-analytical and technical variability of biomarker measurements is usually expressed in terms of the coefficient of variation, and is of great importance for subsequent validation studies for molecular biomarkers. It is finally shown that the required sample size for biomarker validation quadratically increases with the coefficient of variation, and the effect is illustrated using real data from different laboratory technologies.

Anti-JC virus antibodies in a large German natalizumab-treated multiple sclerosis cohort

OBJECTIVE

To investigate the rate of seropositivity of anti-JC virus (JCV) antibodies in a German multiple sclerosis (MS) cohort treated with natalizumab in the postmarketing setting and to assess anti-JCV serostatus in samples obtained before diagnosis of progressive multifocal leukoencephalopathy (PML).

METHODS

This was a blinded, retrospective cross-sectional and longitudinal analysis for anti-JCV antibodies using a confirmatory 2-step ELISA on 2,782 blood samples obtained from 2,253 patients nationwide for routine testing for anti-natalizumab antibodies during open-label treatment between 2007 and 2010.

RESULTS

Of the natalizumab-treated patients with MS, 58.8% tested positive for anti-JCV antibodies. The rate of seropositivity was higher in males and increased with age, with a plateau between age intervals 20-29 and 30-39 years. In longitudinal analyses, 19 of 194 (9.8%) patients converted from anti-JCV antibody-negative to seropositive status over 7.7 months; 4.7% reverted from antibody-positive to seronegative status over 7.9 months. Antibody levels, especially in the latter group, were low, indicating fluctuations around the lower cut point of the assay. Neither anti-JCV serostatus nor antibody levels were associated with immunosuppressive pretreatment, duration of natalizumab treatment, or anti-natalizumab antibodies. All samples obtained from 10 patients who developed PML were seropositive (13 samples before PML diagnosis [2.0-37.6 months]; 2 samples at diagnosis). Antibody levels in these samples were higher than those in samples from seropositive patients who did not develop PML.

CONCLUSIONS

These data argue for the potential clinical utility of JCV serology for PML risk stratification. However, further investigations of fluctuations in serostatus and of antibody levels for a more precise understanding of the predictive value are warranted.

Data analysis and data mining: current issues in biomedical informatics

BACKGROUND

Medicine and biomedical sciences have become data-intensive fields, which, at the same time, enable the application of data-driven approaches and require sophisticated data analysis and data mining methods. Biomedical informatics provides a proper interdisciplinary context to integrate data and knowledge when processing available information, with the aim of giving effective decision-making support in clinics and translational research.

OBJECTIVES

To reflect on different perspectives related to the role of data analysis and data mining in biomedical informatics.

METHODS

On the occasion of the 50th year of Methods of Information in Medicine a symposium was organized, which reflected on opportunities, challenges and priorities of organizing, representing and analysing data, information and knowledge in biomedicine and health care. The contributions of experts with a variety of backgrounds in the area of biomedical data analysis have been collected as one outcome of this symposium, in order to provide a broad, though coherent, overview of some of the most interesting aspects of the field.

RESULTS

The paper presents sections on data accumulation and data-driven approaches in medical informatics, data and knowledge integration, statistical issues for the evaluation of data mining models, translational bioinformatics and bioinformatics aspects of genetic epidemiology.

CONCLUSIONS

Biomedical informatics represents a natural framework to properly and effectively apply data analysis and data mining methods in a decision-making context. In the future, it will be necessary to preserve the inclusive nature of the field and to foster an increasing sharing of data and methods between researchers.