Neil Shephard

Whole-Genome Scan, in a Complex Disease, Using 11,245 Single-Nucleotide Polymorphisms: Comparison with Microsatellites

Despite the theoretical evidence of the utility of single-nucleotide polymorphisms (SNPs) for linkage analysis, no whole-genome scans of a complex disease have yet been published to directly compare SNPs with microsatellites. Here, we describe a whole-genome screen of 157 families with multiple cases of rheumatoid arthritis (RA), performed using 11,245 genomewide SNPs. The results were compared with those from a 10-cM microsatellite scan in the same cohort. The SNP analysis detected HLA*DRB1, the major RA susceptibility locus (P=.00004), with a linkage interval of 31 cM, compared with a 50-cM linkage interval detected by the microsatellite scan. In addition, four loci were detected at a nominal significance level (P<.05) in the SNP linkage analysis; these were not observed in the microsatellite scan. We demonstrate that variation in information content was the main factor contributing to observed differences in the two scans, with the SNPs providing significantly higher information content than the microsatellites. Reducing the number of SNPs in the marker set to 3,300 (1-cM spacing) caused several loci to drop below nominal significance levels, suggesting that decreases in information content can have significant effects on linkage results. In contrast, differences in maps employed in the analysis, the low detectable rate of genotyping error, and the presence of moderate linkage disequilibrium between markers did not significantly affect the results. We have demonstrated the utility of a dense SNP map for performing linkage analysis in a late-age-at-onset disease, where DNA from parents is not always available. The high SNP density allows loci to be defined more precisely and provides a partial scaffold for association studies, substantially reducing the resource requirement for gene-mapping studies.

In adult onset myositis, the presence of interstitial lung disease and myositis specific/associated antibodies are governed by HLA class II haplotype, rather than by myositis subtype.

The aim of this study was to investigate HLA class II associations in polymyositis (PM) and dermatomyositis (DM), and to determine how these associations influence clinical and serological differences. DNA samples were obtained from 225 UK Caucasian idiopathic inflammatory myopathy patients (PM = 117, DM = 108) and compared with 537 randomly selected UK Caucasian controls. All cases had also been assessed for the presence of related malignancy and interstitial lung disease (ILD), and a number of myositis-specific/myositis-associated antibodies (MSAs/MAAs). Subjects were genotyped for HLA-DRB1, DQA1 and DQB1. HLA-DRB1*03, DQA1*05 and DQB1*02 were associated with an increased risk for both PM and DM. The HLA-DRB1*03-DQA1*05-DQB1*02 haplotype demonstrated strong association with ILD, irrespective of myositis subtype or presence of anti-aminoacyl-transfer RNA synthetase antibodies. The HLA-DRB1*07-DQA1*02-DQB1*02 haplotype was associated with risk for anti-Mi-2 antibodies, and discriminated PM from DM (odds ratio 0.3, 95% confidence interval 0.1–0.6), even in anti-Mi-2 negative patients. Other MSA/MAAs showed specific associations with other HLA class II haplotypes, irrespective of myositis subtype. There were no genotype, haplotype or serological associations with malignancy. The HLA-DRB1*03-DQA1*05-DQB1*02 haplotype associations appear to not only govern disease susceptibility in Caucasian PM/DM patients, but also phenotypic features common to PM/DM. Though strongly associated with anti-Mi-2 antibodies, the HLA-DRB1*07-DQA1*02-DQB1*02 haplotype shows differential associations with PM/DM disease susceptibility. In conclusion, these findings support the notion that myositis patients with differing myositis serology have different immunogenetic profiles, and that these profiles may define specific myositis subtypes.

Sample size requirements to estimate key design parameters from external pilot randomised controlled trials: a simulation study

BackgroundExternal pilot or feasibility studies can be used to estimate key unknown parameters to inform the design of the definitive randomised controlled trial (RCT). However, there is little consensus on how large pilot studies need to be, and some suggest inflating estimates to adjust for the lack of precision when planning the definitive RCT.MethodsWe use a simulation approach to illustrate the sampling distribution of the standard deviation for continuous outcomes and the event rate for binary outcomes. We present the impact of increasing the pilot sample size on the precision and bias of these estimates, and predicted power under three realistic scenarios. We also illustrate the consequences of using a confidence interval argument to inflate estimates so the required power is achieved with a pre-specified level of confidence. We limit our attention to external pilot and feasibility studies prior to a two-parallel-balanced-group superiority RCT.ResultsFor normally distributed outcomes, the relative gain in precision of the pooled standard deviation (SDp) is less than 10% (for each five subjects added per group) once the total sample size is 70. For true proportions between 0.1 and 0.5, we find the gain in precision for each five subjects added to the pilot sample is less than 5% once the sample size is 60. Adjusting the required sample sizes for the imprecision in the pilot study estimates can result in excessively large definitive RCTs and also requires a pilot sample size of 60 to 90 for the true effect sizes considered here.ConclusionsWe recommend that an external pilot study has at least 70 measured subjects (35 per group) when estimating the SDp for a continuous outcome. If the event rate in an intervention group needs to be estimated by the pilot then a total of 60 to 100 subjects is required. Hence if the primary outcome is binary a total of at least 120 subjects (60 in each group) may be required in the pilot trial. It is very much more efficient to use a larger pilot study, than to guard against the lack of precision by using inflated estimates.

Identification of a novel susceptibility locus for juvenile idiopathic arthritis by genome-wide association analysis

Objective Juvenile idiopathic arthritis (JIA) is a chronic rheumatic disease of childhood. Two well-established genetic factors known to contribute to JIA susceptibility, HLA and PTPN22, account for less than half of the genetic susceptibility to disease; therefore, additional genetic factors have yet to be identified. The purpose of this study was to perform a systematic search of the genome to identify novel susceptibility loci for JIA. Methods A genome-wide association study using Affymetrix GeneChip 100K arrays was performed in a discovery cohort (279 cases and 184 controls). Single-nucleotide polymorphisms (SNPs) showing the most significant differences between cases and controls were then genotyped in a validation sample of cases (n = 321) and controls, combined with control data from the 1958 UK birth cohort (n = 2,024). In one region in which association was confirmed, fine-mapping was performed (654 cases and 1,847 controls). Results Of the 112 SNPs that were significantly associated with JIA in the discovery cohort, 6 SNPs were associated with JIA in the independent validation cohort. The most strongly associated SNP mapped to the HLA region, while the second strongest association was with a SNP within the VTCN1 gene. Fine-mapping of that gene was performed, and 10 SNPs were found to be associated with JIA. Conclusion This study is the first to successfully apply a SNP-based genome-wide association approach to the investigation of JIA. The replicated association with markers in the VTCN1 gene defined an additional susceptibility locus for JIA and implicates a novel pathway in the pathogenesis of this chronic disease of childhood.

Haemorrhoidal artery ligation versus rubber band ligation for the management of symptomatic second-degree and third-degree haemorrhoids (HubBLe): a multicentre, open-label, randomised controlled trial

Summary Background Optimum surgical intervention for low-grade haemorrhoids is unknown. Haemorrhoidal artery ligation (HAL) has been proposed as an efficacious, safe therapy while rubber band ligation (RBL) is a commonly used outpatient treatment. We compared recurrence after HAL versus RBL in patients with grade II–III haemorrhoids. Methods This multicentre, open-label, parallel group, randomised controlled trial included patients from 17 acute UK NHS trusts. We screened patients aged 18 years or older presenting with grade II–III haemorrhoids. We excluded patients who had previously received any haemorrhoid surgery, more than one injection treatment for haemorrhoids, or more than one RBL procedure within 3 years before recruitment. Eligible patients were randomly assigned (in a 1:1 ratio) to either RBL or HAL with Doppler. Randomisation was computer-generated and stratified by centre with blocks of random sizes. Allocation concealment was achieved using a web-based system. The study was open-label with no masking of participants, clinicians, or research staff. The primary outcome was recurrence at 1 year, derived from the patients self-reported assessment in combination with resource use from their general practitioner and hospital records. Recurrence was analysed in patients who had undergone one of the interventions and been followed up for at least 1 year. This study is registered with the ISRCTN registry, ISRCTN41394716. Findings From Sept 9, 2012, to May 6, 2014, of 969 patients screened, 185 were randomly assigned to the HAL group and 187 to the RBL group. Of these participants, 337 had primary outcome data (176 in the RBL group and 161 in the HAL group). At 1 year post-procedure, 87 (49%) of 176 patients in the RBL group and 48 (30%) of 161 patients in the HAL group had haemorrhoid recurrence (adjusted odds ratio [aOR] 2·23, 95% CI 1·42–3·51; p=0·0005). The main reason for this difference was the number of extra procedures required to achieve improvement (57 [32%] participants in the RBL group and 23 [14%] participants in the HAL group had a subsequent procedure for haemorrhoids). The mean pain 1 day after procedure was 3·4 (SD 2·8) in the RBL group and 4·6 (2·8) in the HAL group (difference −1·2, 95% CI −1·8 to −0·5; p=0·0002); at day 7 the scores were 1·6 (2·3) in the RBL group and 3·1 (2·4) in the HAL group (difference −1·5, −2·0 to −1·0; p<0·0001). Pain scores did not differ between groups at 21 days and 6 weeks. 15 individuals reported serious adverse events requiring hospital admission. One patient in the RBL group had a pre-existing rectal tumour. Of the remaining 14 serious adverse events, 12 (7%) were among participants treated with HAL and two (1%) were in those treated with RBL. Six patients had pain (one treated with RBL, five treated with HAL), three had bleeding not requiring transfusion (one treated with RBL, two treated with HAL), two in the HAL group had urinary retention, two in the HAL group had vasovagal upset, and one in the HAL group had possible sepsis (treated with antibiotics). Interpretation Although recurrence after HAL was lower than a single RBL, HAL was more painful than RBL. The difference in recurrence was due to the need for repeat bandings in the RBL group. Patients (and health commissioners) might prefer such a course of RBL to the more invasive HAL. Funding NIHR Health Technology Assessment programme.

Data for Genetic Analysis Workshop (GAW) 15 Problem 2, genetic causes of rheumatoid arthritis and associated traits

For Genetic Analysis Workshop 15 Problem 2, we organized data from several ongoing studies designed to identify genetic and environmental risk factors for rheumatoid arthritis. Data were derived from the North American Rheumatoid Arthritis Consortium (NARAC), collaboration among Canadian researchers, the European Consortium on Rheumatoid Arthritis Families (ECRAF), and investigators from Manchester, England. All groups used a common standard for defining rheumatoid arthritis, but NARAC also further selected for a more severe phenotype in the probands. Genotyping and family structures for microsatellite-based linkage analysis were provided from all centers. In addition, all centers but ECRAF have genotyped families for linkage analysis using SNPs and these data were additionally provided. NARAC also had additional data from a dense genotyping analysis of a region of chromosome 18 and results from candidate gene studies, which were provided. Finally, smoking influences risk for rheumatoid arthritis, and data were provided from the NARAC study on this behavior as well as some additional phenotypes measuring severity. Several questions could be evaluated using the data that were provided. These include comparing linkage analysis using single-nucleotide polymorphisms versus microsatellites and identifying credible regions of linkage outside the HLA region on chromosome 6p13, which has been extensively documented; evaluating the joint effects of smoking with genetic factors; and identifying more homogenous subsets of families for whom genetic susceptibility might be stronger, so that linkage and association studies may be more efficiently conducted.

Derivation and validation of a risk adjustment model for predicting seven day mortality in emergency medical admissions: mixed prospective and retrospective cohort study

Objectives To derive and validate a risk adjustment model for predicting seven day mortality in emergency medical admissions, to test the value of including physiology and blood parameters, and to explore the constancy of the risk associated with each model variable across a range of settings. Design Mixed prospective and retrospective cohort study. Setting Nine acute hospitals (n=3 derivation, n=9 validation) and associated ambulance services in England, Australia, and Hong Kong. Participants Adults with medical emergencies (n=5644 derivation, n=13 762 validation) who were alive and not in cardiac arrest when attended by an ambulance and either were admitted to hospital or died in the ambulance or emergency department. Interventions Data were either collected prospectively or retrospectively from routine sources and extraction from ambulance and emergency department records. Main outcome measure Mortality up to seven days after hospital admission. Results In the derivation phase, age, ICD-10 code, active malignancy, Glasgow coma score, respiratory rate, peripheral oxygen saturation, temperature, white cell count, and potassium and urea concentrations were independent predictors of seven day mortality. A model based on age and ICD-10 code alone had a C statistic of 0.80 (95% confidence interval 0.78 to 0.83), which increased to 0.81 (0.79 to 0.84) with the addition of active malignancy. This was markedly improved only when physiological variables (C statistic 0.87, 0.85 to 0.89), blood variables (0.87, 0.84 to 0.89), or both (0.90, 0.88 to 0.92) were added. In the validation phase, the models with physiology variables (physiology model) and all variables (full model) were tested in nine hospitals. Overall, the C statistics ranged across centres from 0.80 to 0.91 for the physiology model and from 0.83 to 0.93 for the full model. The rank order of hospitals based on adjusted mortality differed markedly from the rank order based on crude mortality. ICD-10 code, Glasgow coma score, respiratory rate, systolic blood pressure, oxygen saturation, haemoglobin concentration, white cell count, and potassium, urea, creatinine, and glucose concentrations all had statistically significant interactions with hospital. Conclusion A risk adjustment model for emergency medical admissions based on age, ICD-10 code, active malignancy, and routinely recorded physiological and blood variables can provide excellent discriminant value for seven day mortality across a range of settings. Using risk adjustment markedly changed hospitals’ rankings. However, evidence was found that the association between key model variables and mortality were not constant. Supplementary data appendix

Will the real disease gene please stand up

A common dilemma arising in linkage studies of complex genetic diseases is the selection of positive signals, their follow-up with association studies and discrimination between true and false positive results. Several strategies for overcoming these issues have been devised. Using the Genetic Analysis Workshop 14 simulated dataset, we aimed to apply different analytical approaches and evaluate their performance in discerning real associations. We considered a) haplotype analyses, b) different methods adjusting for multiple testing, c) replication in a second dataset, and d) exhaustive genotyping of all markers in a sufficiently powered, large sample group. We found that haplotype-based analyses did not substantially improve over single-point analysis, although this may reflect the low levels of linkage disequilibrium simulated in the datasets provided. Multiple testing correction methods were in general found to be over-conservative. Replication of nominally positive results in a second dataset appears to be less stringent, resulting in the follow-up of false positives. Performing a comprehensive assay of all markers in a large, well-powered dataset appears to be the most effective strategy for complex disease gene identification.

Fine mapping of genes within the IDDM8 region in rheumatoid arthritis

The IDDM8 region on chromosome 6q27, first identified as a susceptibility locus for type 1 diabetes, has previously been linked and associated with rheumatoid arthritis (RA). The region contains a number of potential candidate genes, including programmed cell death 2 (PDCD2), the proteosome subunit beta type 1 (PSMB1), delta-like ligand 1 (DLL-1) and TATA box-binding protein (TBP) amongst others. The aim of this study was to fine map the IDDM8 region on chromosome 6q27, focusing on the genes in the region, to identify polymorphisms that may contribute to susceptibility to RA and potentially to other autoimmune diseases. Validated single nucleotide polymorphisms (SNPs; n = 65) were selected from public databases from the 330 kb region of IDDM8. These were genotyped using Sequenom MassArray genotyping technology in two datasets; the test dataset comprised 180 RA cases and 180 controls. We tested 50 SNPs for association with RA and any significant associations were genotyped in a second dataset of 174 RA cases and 192 controls, and the datasets were combined before analysis. Association analysis was performed by chi-square test implemented in Stata software and linkage disequilibrium and haplotype analysis was performed using Helix tree version 4.1. There was initial weak evidence of association, with RA, of a number of SNPs around the loc154449 putative gene and within the KIAA1838 gene; however, these associations were not significant in the combined dataset. Our study has failed to detect evidence of association with any of the known genes mapping to the IDDM8 locus with RA.

Linkage analysis of cross-sectional and longitudinally derived phenotypic measures to identify loci influencing blood pressure

BackgroundThe design of appropriate strategies to analyze and interpret linkage results for complex human diseases constitutes a challenge. Parameters such as power, definition of phenotype, and replicability have to be taken into account in order to reach meaningful conclusions. Incorporating data on repeated phenotypic measures may increase the power to detect linkage but requires sophisticated analysis methods. Using the simulated Genetic Analysis Workshop 13 data set, we have estimated a variety of systolic blood pressure (SBP) phenotypic measures and examined their performance with respect to consistency among replicates and to true and false positive linkage signals.ResultsThe whole-genome scan conducted on a dichotomous hypertension phenotype indicated the involvement of few true loci with nominal significance and gave rise to a high rate of false positives. Analysis of a cross-sectional quantitative SBP measure performed better, although genome-wide significance was again not reached. Additional phenotypic measures were derived from the longitudinal data using random effects modelling for censored data with varying levels of covariate adjustment. These models provided evidence for significant linkage to most genes influencing SBP and produced few false positive results. Overall, replicability of results was poor for loci, representing weak effects.ConclusionLongitudinally derived phenotypes performed better than cross-sectional measures in linkage analyses. Bearing in mind the sample design and size of these data, linkage results that fail to replicate should not be dismissed; instead, different lines of evidence derived from complementary analysis methods should be combined to prioritize follow up.