Distributional Properties and Criterion Validity of a Shortened Version of the Social Responsiveness Scale: Results from the ECHO Program and Implications for Social Communication Research

Prior work proposed a shortened version of the Social Responsiveness Scale (SRS), a commonly used quantitative measure of social communication traits. We used data from 3031 participants (including 190 ASD cases) from the Environmental Influences on Child Health Outcomes (ECHO) Program to compare distributional properties and criterion validity of 16-item “short” to 65-item “full” SRS scores. Results demonstrated highly overlapping distributions of short and full scores. Both scores separated case from non-case individuals by approximately two standard deviations. ASD prediction was nearly identical for short and full scores (area under the curve values of 0.87, 0.86 respectively). Findings support comparability of shortened and full scores, suggesting opportunities to increase efficiency. Future work should confirm additional psychometric properties of short scores.

     

Correlation in the human aorta of APO B fractions with tissue cholesterol and collagen content.

The amounts of buffer- and Triton-extracted apo B (LDL-protein), as well as the sum of these two fractions, were correlated with the total tissue cholesterol and hydroxyproline content (as a measure of collagen) in grossly normal intima, fatty streaks, and fibrous plaques of human aortas obtained at autopsy. Quantitative values of buffer- and Triton-extracted apo B were obtained by sequentially extracting homogenates of aortic intima with an aqueous buffer and one containing Triton X-100, and measuring the apo B content in each extract by an electroimmunoassay relative to plasma LDL or Triton-treated LDL. Significant positive correlations were obtained between the following: tissue cholesterol and both buffer-extracted and total-extracted apo B in grossly normal intima; tissue cholesterol and Triton-extracted apo B in microdissected fibrotic caps and cores of fibrous plaques, as well as in whole plaques. A positive correlation was also obtained between tissue cholesterol and total-extracted apo B in the necrotic core. A significant negative correlation was found between Triton-extracted apo B and collagen in whole plaques. The calculated mean percent of total tissue cholesterol in the different aortic regions that could be present as part of an intact LDL particle were: 100% in grossly normal intima, 16% in fatty streaks, and 11% in fibrous plaques. The positive correlation between Triton-extracted apo B and cholesterol in plaques suggests one or both of the following: the extracellular pool of cholesterol or some material increasing concurrently with cholesterol interacts with apo B or another part of the LDL particle; or the apo B containing lipoprotein is trapped in the hydrophobic environment of extracellular lipid. Both possibilities would render the particle less soluble in aqueous buffers. The negative correlation between Triton-extracted apo B and tissue collagen and the lack of a significant correlation between buffer-extracted apo B and collagen content suggests that collagen is probably not responsible for apo B retention in the aortic intima.     

Review of processing and analysis methods for DNA methylation array data

The promise of epigenome-wide association studies and cancer-specific somatic DNA methylation changes in improving our understanding of cancer, coupled with the decreasing cost and increasing coverage of DNA methylation microarrays, has brought about a surge in the use of these technologies. Here, we aim to provide both a review of issues encountered in the processing and analysis of array-based DNA methylation data and a summary of the advantages of recent approaches proposed for handling those issues, focusing on approaches publicly available in open-source environments such as R and Bioconductor. We hope that the processing tools and analysis flowchart described herein will facilitate researchers to effectively use these powerful DNA methylation array-based platforms, thereby advancing our understanding of human health and disease.     

Power and sample size calculations for generalized regression models with covariate measurement error

Covariate measurement error is often a feature of scientific data used for regression modelling. The consequences of such errors include a loss of power of tests of significance for the regression parameters corresponding to the true covariates. Power and sample size calculations that ignore covariate measurement error tend to overestimate power and underestimate the actual sample size required to achieve a desired power. In this paper we derive a novel measurement error corrected power function for generalized linear models using a generalized score test based on quasi-likelihood methods. Our power function is flexible in that it is adaptable to designs with a discrete or continuous scalar covariate (exposure) that can be measured with or without error, allows for additional confounding variables and applies to a broad class of generalized regression and measurement error models. A program is described that provides sample size or power for a continuous exposure with a normal measurement error model and a single normal confounder variable in logistic regression. We demonstrate the improved properties of our power calculations with simulations and numerical studies. An example is given from an ongoing study of cancer and exposure to arsenic as measured by toenail concentrations and tap water samples.     

Decreased DNA repair gene expression among individuals exposed to arsenic in United States drinking water

Arsenic is well established as a human carcinogen, but its precise mechanism of action remains unknown. Arsenic does not directly damage DNA, but may act as a carcinogen through inhibition of DNA repair mechanisms, leading indirectly to increased mutations from other DNA damaging agents. The molecular mechanism underlying arsenic inhibition of nucleotide excision repair after UV irradiation (Hartwig et al., Carcinogenesis 1997;18:399-405) is unknown, but could be due to decreased expression of critical genes involved in nucleotide excision repair of damaged DNA. This hypothesis was tested by analyzing expression of repair genes and arsenic exposure in a subset of 16 individuals enrolled in a population based case-control study investigating arsenic exposure and cancer risk in New Hampshire. Toenail arsenic levels were inversely correlated with expression of critical members of the nucleotide excision repair complex, ERCC1 (r(2) = 0.82, p < 0.0001), XPF (r(2) = 0.56, p < 0.002), and XPB (r(2) = 0.75, p < 0.0001). The internal dose marker, toenail arsenic level, was more strongly associated with changes in expression of these genes than drinking water arsenic concentration. Our findings, based on human exposure to arsenic in a US population, show an association between biomarkers of arsenic exposure and expression of DNA repair genes. Although our findings need verification in a larger study group, they are consistent with the hypothesis that inhibition of DNA repair capacity is a potential mechanism for the co-carcinogenic activity of arsenic.     

Non-melanoma skin cancers and glucocorticoid therapy

Non-melanoma skin cancer (NMSC) is an important cause of morbidity and long-term mortality in organ transplant recipients receiving immunosuppressive drugs such as azathioprine and cyclosporin, often combined with adrenocortical steroids (glucocorticoids). At lower doses, glucocorticoids alone are prescribed for other conditions including musculoskeletal, connective tissue and respiratory disorders. Presently, it is unknown whether patients taking glucocorticoids are at an increased risk of skin malignancies. In a population-based case-control study in New Hampshire, USA, we compared use of glucocorticoids in 592 basal cell carcinoma (BCC) and 281 squamous cell carcinoma (SCC) cases and in 532 age and gender matched controls; neither cases nor controls had a history of organ transplantation. Participants underwent a structured personal interview regarding history of medication use and skin cancer risk factors. We used unconditional logistic regression analysis to compute odds ratios associated with glucocorticoid use for 1 month or longer while controlling for potential confounding factors. Risk of SCC was increased among users of oral glucocorticoids (adjusted odds ratio = 2.31; 95% CI = 1.27, 4.18), and risk of BCC was elevated modestly (adjusted odds ratio = 1.49; 95% CI = 0.90, 2.47). In contrast, risk of both SCC and BCC were unrelated to use of inhaled steroids. Our data suggest that use of oral glucocorticoids may increase risk of NMSC, and SCC in particular, among patients other than organ transplant recipients. We hypothesize that immunosuppression induced by oral glucocorticoids may allow these cancers to emerge from immunosurveillance.     

Two-stage methods for the analysis of pooled data.

Epidemiologic studies of disease often produce inconclusive or contradictory results due to small sample sizes or regional variations in the disease incidence or the exposures. To clarify these issues, researchers occasionally pool and reanalyse original data from several large studies. In this paper we explore the use of a two-stage random-effects model for analysing pooled case-control studies and undertake a thorough examination of bias in the pooled estimator under various conditions. The two-stage model analyses each study using the model appropriate to the design with study-specific confounders, and combines the individual study-specific adjusted log-odds ratios using a linear mixed-effects model; it is computationally simple and can incorporate study-level covariates and random effects. Simulations indicate that when the individual studies are large, two-stage methods produce nearly unbiased exposure estimates and standard errors of the exposure estimates from a generalized linear mixed model. By contrast, joint fixed-effects logistic regression produces attenuated exposure estimates and underestimates the standard error when heterogeneity is present. While bias in the pooled regression coefficient increases with interstudy heterogeneity for both models, it is much smaller using the two-stage model. In pooled analyses, where covariates may not be uniformly defined and coded across studies, and occasionally not measured in all studies, a joint model is often not feasible. The two-stage method is shown to be a simple, valid and practical method for the analysis of pooled binary data. The results are applied to a study of reproductive history and cutaneous melanoma risk in women using data from ten large case-control studies.