Dynamic inference in general nested case‐control designs

Nested case‐control designs are attractive in studies with a time‐to‐event endpoint if the outcome is rare or if interest lies in evaluating expensive covariates. The appeal is that these designs restrict to small subsets of all patients at risk just prior to the observed event times. Only these small subsets need to be evaluated. Typically, the controls are selected at random and methods for time‐simultaneous inference have been proposed in the literature. However, the martingale structure behind nested case‐control designs allows for more powerful and flexible non‐standard sampling designs. We exploit that structure to find simultaneous confidence bands based on wild bootstrap resampling procedures within this general class of designs. We show in a simulation study that the intended coverage probability is obtained for confidence bands for cumulative baseline hazard functions. We apply our methods to observational data about hospital‐acquired infections.     

Global DNA hypomethylation is associated with NTD‐affected pregnancy: A case‐control study

BACKGROUND: Neural tube defects are severe, common birth defects that result from failure of neural tube closure. They are considered to be a multifactorial disorder, and our knowledge of causal mechanisms remains limited. We hypothesized that abnormal DNA methylation occurs in NTD‐affected fetuses. The correlations of global DNA methylation levels with complexity of NTDs and known risk factors of NTDs, MTHFR genotype and fever, were analyzed. METHODS: A hospital‐based case‐control study was performed. Epidemiologic data, pathologic diagnosis, and methylenetetrahydrofolate reductase (MTHFR) genotype analysis were completed. Array comparative genomic hybridization was used to exclude cytogenetic abnormalities. Global DNA methylation statuses were determined for both brain and skin tissue. RESULTS: Sixty‐five NTD‐affected fetuses and 65 normal controls matched for gestational and maternal ages were collected. In brain tissue, global DNA methylation levels were significantly decreased in cases compared with controls (4.12 vs. 4.99%; p < 0.001). DNA hypomethylation (<4.35%) resulted in a significant 5.736‐fold increased risk for NTDs (95% confidence interval, 1.731–19.009; p = 0.004). Nonisolated NTDs had lower levels of global DNA methylation than did isolated NTDs (3.77 vs. 4.70%; p = 0.022). After stratifying subjects by MTHFR genotype, we observed a skewed distribution of global DNA methylation levels. For genotype C/C, global DNA methylation status was the same in the two groups (4.51 vs. 4.72%; p = 0.687). For T/T, cases had significantly lower global methylation levels than did controls (5.23 vs. 3.79%; p < 0.001). CONCLUSIONS: Global DNA hypomethylation in fetal brain tissue was associated with NTD‐affected pregnancy. DNA methylation levels were correlated with NTD complexity. The MTHFR genotype contributed to global DNA hypomethylation. Birth Defects Research (Part A), 2010. © 2010 Wiley‐Liss, Inc.     

LMO1 polymorphisms and the risk of neuroblastoma: Assessment of meta‐analysis of case‐control studies

Neuroblastoma (NB), a neuroendocrine tumour, is one of the most prevalent cancers in children. The link between LMO1 polymorphisms and NB has been investigated by several groups, rendering inconclusive results. Here, with this comprehensive systematic review and up‐to‐date meta‐analysis, we aim to distinctively elucidate the possible correlation between LMO1 polymorphisms and NB susceptibility. Eligible studies were systematically researched and identified using PubMed, Web of Science and Scopus databases up to 10 February 2019. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to assess the strength of the associations. Our findings revealed that rs110419 and rs2168101 polymorphisms were significantly associated with a decreased risk of NB in all genetic models. In addition, the rs4758051 variant appeared protective against NB in homozygous, dominant and allele genetic models, whereas the rs10840002 variant markedly decreased the risk of NB in the allele model. In contrast, the rs204938 polymorphism showed a positive association with NB susceptibility in allele genetic models. In summary, our meta‐analysis is the first to provide clear evidence of an association between specific polymorphisms of LMO1 and susceptibility to NB. Of note, additional larger well‐designed studies would be helpful to further evaluate and confirm this association.     

LIN28A gene polymorphisms modify neuroblastoma susceptibility: A four‐centre case‐control study

Neuroblastoma ranks the most common seen solid tumour in childhood. Overexpression of LIN28A gene has been linked to the development of multiple human malignancies, but the relationship between LIN28A single nucleotide polymorphisms (SNPs) and neuroblastoma susceptibility is still under debate. Herein, we evaluated the correlation of four potentially functional LIN28A SNPs (rs3811464 G>A, rs3811463 T>C, rs34787247 G>A, and rs11247957 G>A) and neuroblastoma susceptibility in 505 neuroblastoma patients and 1070 controls from four independent hospitals in China. The correlation strengths were determined by using odds ratios (ORs) and corresponding 95% confidence intervals (CIs). Among these SNPs, rs34787247 G>A exhibited a significant association with increased susceptibility in neuroblastoma (GA vs GG: adjusted OR = 1.30, 95% CI = 1.03‐1.64; AA vs GG: adjusted OR = 2.51, 95% CI = 1.36‐4.64, AA/GA vs GG: adjusted OR = 1.42, 95% CI = 1.12‐1.80, AA vs GG/GA: adjusted OR = 2.39, 95% CI = 1.29‐4.42). Furthermore, the combined analysis of risk genotypes revealed that subjects carrying three risk genotypes (adjusted OR = 1.64, 95% CI = 1.02‐2.63) are more inclined to develop neuroblastoma than those without risk genotype, and so do carriers of 1‐4 risk genotypes (adjusted OR = 1.26, 95% CI = 1.01‐1.56). Stratification analysis further revealed risk effect of rs3811464 G>A, rs34787247 G>A and 1‐4 risk genotypes in some subgroups. Haplotype analysis of these four SNPs yields two haplotypes significantly correlated with increased neuroblastoma susceptibility. Overall, our finding indicated that LIN28A SNPs, especially rs34787247 G>A, may increase neuroblastoma risk.     

Association between METTL3 gene polymorphisms and neuroblastoma susceptibility: A nine‐centre case‐control study

Neuroblastoma ranks as the most commonly seen and deadly solid tumour in infancy. The aberrant activity of m⁶A‐RNA methyltransferase METTL3 is involved in human cancers. Therefore, functional genetic variants in the METTL3 gene may contribute to neuroblastoma risk. In the current nine‐centre case‐control study, we aimed to analyse the association between the METTL3 gene single nucleotide polymorphisms (SNPs) and neuroblastoma susceptibility. We genotyped four METTL3 gene SNPs (rs1061026 T>G, rs1061027 C>A, rs1139130 A>G, and rs1263801 G>C) in 968 neuroblastoma patients and 1814 controls in China. We found significant associations between these SNPs and neuroblastoma risk in neither single‐locus nor combined analyses. Interestingly, in the stratified analysis, we observed a significant risk association with rs1061027 AA in subgroups of children ≤ 18 months of age (adjusted OR = 1.87, 95% CI = 1.03‐3.41, P = .040) and females (adjusted OR = 1.86, 95% CI = 1.07‐3.24, P = .028). Overall, we identified a significant association between METTL3 gene rs1061027 C>A polymorphism and neuroblastoma risk in children ≤18 months of age and females. Our findings provide novel insights into the genetic determinants of neuroblastoma.     

Variable selection in Bayesian generalized linear‐mixed models: An illustration using candidate gene case‐control association studies

The problem of variable selection in the generalized linear‐mixed models (GLMMs) is pervasive in statistical practice. For the purpose of variable selection, many methodologies for determining the best subset of explanatory variables currently exist according to the model complexity and differences between applications. In this paper, we develop a “higher posterior probability model with bootstrap” (HPMB) approach to select explanatory variables without fitting all possible GLMMs involving a small or moderate number of explanatory variables. Furthermore, to save computational load, we propose an efficient approximation approach with Laplace's method and Taylor's expansion to approximate intractable integrals in GLMMs. Simulation studies and an application of HapMap data provide evidence that this selection approach is computationally feasible and reliable for exploring true candidate genes and gene–gene associations, after adjusting for complex structures among clusters.