Yates M.,Norwich University |
Cheong E.,Norwich University |
Luben R.,Strangeways Research Laboratory |
Igali L.,Norwich University |
And 4 more authors.
Digestive Diseases and Sciences | Year: 2014
Background: The timing of the risk factors cigarette smoking, alcohol and obesity in the development of Barrett's esophagus (BE) and esophageal adenocarcinoma (EAC) is unclear. Aims: To investigate these exposures in the aetiology of BE and EAC in the same population. Methods: The cohort included 24,068 men and women, aged 39-79 years, recruited between 1993 and 1997 into the prospective EPIC-Norfolk Study who provided information on anthropometry, smoking and alcohol intake. The cohort was monitored until December 2008 and incident cases identified. Results: One hundred and four participants were diagnosed with BE and 66 with EAC. A body mass index (BMI) above 23 kg/m 2 was associated with a greater risk of BE [BMI ≥23 vs. 18.5 to <23, hazard ratio (HR) 3.73, 95 % CI 1.37-10.16], and within a normal BMI, the risk was greater in the higher category (HR 3.76, 95 % CI 1.30-10.85, BMI 23-25 vs. 18.5 to >23 kg/m2). Neither smoking nor alcohol intake were associated with risk for BE. For EAC, all BMI categories were associated with risk, although statistically significant for only the highest (BMI >35 vs. BMI 18.5 to <23, HR 4.95, 95 % CI 1.11-22.17). The risk was greater in the higher category of a normal BMI (HR 2.73, 95 % CI 0.93-8.00, p = 0.07, BMI 23-25 vs. 18.5 to >23 kg/m2). There was an inverse association with ≥7 units alcohol/week (HR 0.51, 95 % CI 0.29-0.88) and with wine (HR 0.49, 95 % CI 0.23-1.04, p = 0.06, drinkers vs. non-drinkers). Conclusions: Obesity may be involved early in carcinogenesis and the association with EAC and wine should be explored. The data have implications for aetiological investigations and prevention strategies. © 2014 The Author(s).
Walter F.M.,University of Cambridge |
Prevost A.T.,King's College London |
Hall P.N.,University of Western Australia |
Burrows N.P.,University of Cambridge |
And 2 more authors.
British Journal of General Practice | Year: 2013
Background GPs need to recognise significant pigmented skin lesions, given rising UK incidence rates for malignant melanoma. The 7-point checklist (7PCL) has been recommended by NICE (2005) for routine use in UK general practice to identify clinically significant lesions which require urgent referral. Aim To validate the Original and Weighted versions of the 7PCL in the primary care setting. Design and setting Diagnostic validation study, using data from a SIAscopic diagnostic aid randomised controlled trial in eastern England. Method Adults presenting in general practice with a pigmented skin lesion that could not be immediately diagnosed as benign were recruited into the trial. Reference standard diagnoses were histology or dermatology expert opinion; 7PCL scores were calculated blinded to the reference diagnosis. A case was defined as a clinically significant lesion for primary care referral to secondary care (total 1436 lesions: 225 cases, 1211 controls); or melanoma (36). Results For diagnosing clinically significant lesions there was a difference between the performance of the Original and Weighted 7PCLs (respectively, area under curve: 0.66, 0.69, difference = 0.03, P<0.001). For the identification of melanoma, similar differences were found. Increasing the Weighted 7PCL's cut-off score from recommended 3 to 4 improved detection of clinically significant lesions in primary care: sensitivity 73.3%, specificity 57.1%, positive predictive value 24.1%, negative predictive value 92.0%, while maintaining high sensitivity of 91.7% and moderate specificity of 53.4% for melanoma. Conclusion The Original and Weighted 7PCLs both performed well in a primary care setting to identify clinically significant lesions as well as melanoma. The Weighted 7PCL, with a revised cut-off score of 4 from 3, performs slightly better and could be applied in general practice to support the recognition of clinically significant lesions and therefore the early identification of melanoma. © British Journal of General Practice.
Burgess S.,University of Cambridge |
Burgess S.,Strangeways Research Laboratory |
Granell R.,University of Bristol |
Palmer T.M.,University of Warwick |
And 2 more authors.
American Journal of Epidemiology | Year: 2014
A parameter in a statistical model is identified if its value can be uniquely determined from the distribution of the observable data. We consider the context of an instrumental variable analysis with a binary outcome for estimating a causal risk ratio. The semiparametric generalized method of moments and structural mean model frameworks use estimating equations for parameter estimation. In this paper, we demonstrate that lack of identification can occur in either of these frameworks, especially if the instrument is weak. In particular, the estimating equations may have no solution or multiple solutions. We investigate the relationship between the strength of the instrument and the proportion of simulated data sets for which there is a unique solution of the estimating equations. We see that this proportion does not appear to depend greatly on the sample size, particularly for weak instruments (ρ2 ≤ 0.01). Poor identification was observed in a considerable proportion of simulated data sets for instruments explaining up to 10% of the variance in the exposure with sample sizes up to 1 million. In an applied example considering the causal effect of body mass index (weight (kg)/height (m)2) on the probability of early menarche, estimates and standard errors from an automated optimization routine were misleading. © 2014 The Author 2014.
Ederhy S.,Strangeways Research Laboratory |
Di Angelantonio E.,University of Cambridge |
Dufaitre G.,Strangeways Research Laboratory |
Meuleman C.,Strangeways Research Laboratory |
And 4 more authors.
International Journal of Cardiology | Year: 2012
Background: To determine whether C-reactive protein (CRP) in combination with a stroke risk stratification scheme can help in identifying transesophageal echocardiographic (TEE) markers of thromboembolism such as left atrial (LA)/left atrial appendage (LAA) thrombus, severe LA/LAA spontaneous echocardiographic contrast (SEC), and aortic plaque ≥ 4 mm. Methods: Transthoracic echocardiography, TEE, and CRP measurement were performed at admission in 178 patients with non-valvular atrial fibrillation not receiving oral anticoagulant therapy. Patients were classified as at low, moderate, or high risk of thromboembolism based on seven clinical risk stratification schemes (SPAF, CHADS2, Framingham, Birmingham/NICE, ACC/AHA/ESC 2006 guidelines, ACCP 2008, CHA2DS2VASc). Results: Severe LA/LAA SEC, LA/LAA thrombus, and aortic plaque ≥ 4 mm were present in 6.2%, 6.7%, and 10.1% of patients, respectively. The combination of CRP with a cut-off value of 3.4 mg/L with the Birmingham/Nice or ACC/AHA/ESC 2006 risk score, led to a negative predictive value of 100% in low-risk patients and 91% in moderate-risk patients. For the detection of severe LA/LAA SEC or thrombus, a good discrimination (area under curve ≥ 0.70) using only clinical risk markers was observed for all classifications except for the Framingham and CHADS2 risk scores. The addition of CRP did not improve the detection of LA/LAA SEC or thrombus, or of severe LA/LAA SEC, thrombus, or aortic plaque. Conclusion: The combination of clinical risk markers and CRP can help to exclude the presence of the TEE markers LA/LAA SEC or LA/LAA thrombus, particularly in patients classified at low or moderate risk of stroke. © 2011 Elsevier Ireland Ltd. All rights reserved.
Surtees P.G.,Strangeways Research Laboratory |
Wainwright N.W.J.,Strangeways Research Laboratory |
Pooley K.A.,Cancer Research UK Genetic Epidemiology Unit |
Luben R.N.,Strangeways Research Laboratory |
And 3 more authors.
Journals of Gerontology - Series A Biological Sciences and Medical Sciences | Year: 2011
We investigated the association between psychological stress, emotional health, and relative mean telomere length in an ethnically homogeneous population of 4,441 women, aged 41-80 years. Mean telomere length was measured using high-throughput quantitative real-time polymerase chain reaction. Social adversity exposure and emotional health were assessed through questionnaire and covariates through direct measurement and questionnaire. This study found evidence that adverse experiences during childhood may be associated with shorter telomere length. This finding remained after covariate adjustment and showed evidence of a dose-response relationship with increasing number of reported childhood difficulties associated with decreasing relative mean telomere length. No associations were observed for any of the other summary measures of social adversity and emotional health considered. These results extend and provide support for some previous findings concerning the association of adverse experience and emotional health histories with shorter telomere length in adulthood. Replication of these findings in longitudinal studies is now essential. © 2011 The Author.
Burgess S.,Strangeways Research Laboratory |
Thompson S.G.,Strangeways Research Laboratory
Statistical Methods in Medical Research | Year: 2016
Mendelian randomisation is an epidemiological method for estimating causal associations from observational data by using genetic variants as instrumental variables. Typically the genetic variants explain only a small proportion of the variation in the risk factor of interest, and so large sample sizes are required, necessitating data from multiple sources. Meta-analysis based on individual patient data requires synthesis of studies which differ in many aspects. A proposed Bayesian framework is able to estimate a causal effect from each study, and combine these using a hierarchical model. The method is illustrated for data on C-reactive protein and coronary heart disease (CHD) from the C-reactive protein CHD Genetics Collaboration (CCGC). Studies from the CCGC differ in terms of the genetic variants measured, the study design (prospective or retrospective, population-based or case-control), whether C-reactive protein was measured, the time of C-reactive protein measurement (pre- or post-disease), and whether full or tabular data were shared. We show how these data can be combined in an efficient way to give a single estimate of causal association based on the totality of the data available. Compared to a two-stage analysis, the Bayesian method is able to incorporate data on 23% additional participants and 51% more events, leading to a 23-26% gain in efficiency. © SAGE Publications 2012.
Asante E.A.,University College London |
Smidak M.,University College London |
Grimshaw A.,University College London |
Houghton R.,University College London |
And 15 more authors.
Nature | Year: 2015
Mammalian prions, transmissible agents causing lethal neurodegenerative diseases, are composed of assemblies of misfolded cellular prion protein (PrP). A novel PrP variant, G127V, was under positive evolutionary selection during the epidemic of kuru - an acquired prion disease epidemic of the Fore population in Papua New Guinea - and appeared to provide strong protection against disease in the heterozygous state. Here we have investigated the protective role of this variant and its interaction with the common, worldwide M129V PrP polymorphism. V127 was seen exclusively on a M129 PRNP allele. We demonstrate that transgenic mice expressing both variant and wild-type human PrP are completely resistant to both kuru and classical Creutzfeldt-Jakob disease (CJD) prions (which are closely similar) but can be infected with variant CJD prions, a human prion strain resulting from exposure to bovine spongiform encephalopathy prions to which the Fore were not exposed. Notably, mice expressing only PrP V127 were completely resistant to all prion strains, demonstrating a different molecular mechanism to M129V, which provides its relative protection against classical CJD and kuru in the heterozygous state. Indeed, this single amino acid substitution (G-V) at a residue invariant in vertebrate evolution is as protective as deletion of the protein. Further study in transgenic mice expressing different ratios of variant and wild-type PrP indicates that not only is PrP V127 completely refractory to prion conversion but acts as a potent dose-dependent inhibitor of wild-type prion propagation. © 2015 Macmillan Publishers Limited. All rights reserved.
Burgess S.,Strangeways Research Laboratory |
White I.R.,Strangeways Research Laboratory |
Resche-Rigon M.,Strangeways Research Laboratory |
Wood A.M.,Strangeways Research Laboratory
Statistics in Medicine | Year: 2013
Multiple imputation is a strategy for the analysis of incomplete data such that the impact of the missingness on the power and bias of estimates is mitigated. When data from multiple studies are collated, we can propose both within-study and multilevel imputation models to impute missing data on covariates. It is not clear how to choose between imputation models or how to combine imputation and inverse-variance weighted meta-analysis methods. This is especially important as often different studies measure data on different variables, meaning that we may need to impute data on a variable which is systematically missing in a particular study. In this paper, we consider a simulation analysis of sporadically missing data in a single covariate with a linear analysis model and discuss how the results would be applicable to the case of systematically missing data. We find in this context that ensuring the congeniality of the imputation and analysis models is important to give correct standard errors and confidence intervals. For example, if the analysis model allows between-study heterogeneity of a parameter, then we should incorporate this heterogeneity into the imputation model to maintain the congeniality of the two models. In an inverse-variance weighted meta-analysis, we should impute missing data and apply Rubin's rules at the study level prior to meta-analysis, rather than meta-analyzing each of the multiple imputations and then combining the meta-analysis estimates using Rubin's rules. We illustrate the results using data from the Emerging Risk Factors Collaboration.© 2013 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.
Burgess S.,Strangeways Research Laboratory
Statistics in Medicine | Year: 2013
An adjustment for an uncorrelated covariate in a logistic regression changes the true value of an odds ratio for a unit increase in a risk factor. Even when there is no variation due to covariates, the odds ratio for a unit increase in a risk factor also depends on the distribution of the risk factor. We can use an instrumental variable to consistently estimate a causal effect in the presence of arbitrary confounding. With a logistic outcome model, we show that the simple ratio or two-stage instrumental variable estimate is consistent for the odds ratio of an increase in the population distribution of the risk factor equal to the change due to a unit increase in the instrument divided by the average change in the risk factor due to the increase in the instrument. This odds ratio is conditional within the strata of the instrumental variable, but marginal across all other covariates, and is averaged across the population distribution of the risk factor. Where the proportion of variance in the risk factor explained by the instrument is small, this is similar to the odds ratio from a RCT without adjustment for any covariates, where the intervention corresponds to the effect of a change in the population distribution of the risk factor. This implies that the ratio or two-stage instrumental variable method is not biased, as has been suggested, but estimates a different quantity to the conditional odds ratio from an adjusted multiple regression, a quantity that has arguably more relevance to an epidemiologist or a policy maker, especially in the context of Mendelian randomization. © 2013 John Wiley & Sons, Ltd.
Morris T.P.,Hub for Trials Methodology Research |
Morris T.P.,Institute of Public Health |
White I.R.,Institute of Public Health |
Royston P.,Hub for Trials Methodology Research |
And 2 more authors.
Statistics in Medicine | Year: 2014
We are concerned with multiple imputation of the ratio of two variables, which is to be used as a covariate in a regression analysis. If the numerator and denominator are not missing simultaneously, it seems sensible to make use of the observed variable in the imputation model. One such strategy is to impute missing values for the numerator and denominator, or the log-transformed numerator and denominator, and then calculate the ratio of interest; we call this 'passive' imputation. Alternatively, missing ratio values might be imputed directly, with or without the numerator and/or the denominator in the imputation model; we call this 'active' imputation. In two motivating datasets, one involving body mass index as a covariate and the other involving the ratio of total to high-density lipoprotein cholesterol, we assess the sensitivity of results to the choice of imputation model and, as an alternative, explore fully Bayesian joint models for the outcome and incomplete ratio. Fully Bayesian approaches using Winbugs were unusable in both datasets because of computational problems. In our first dataset, multiple imputation results are similar regardless of the imputation model; in the second, results are sensitive to the choice of imputation model. Sensitivity depends strongly on the coefficient of variation of the ratio's denominator. A simulation study demonstrates that passive imputation without transformation is risky because it can lead to downward bias when the coefficient of variation of the ratio's denominator is larger than about 0.1. Active imputation or passive imputation after log-transformation is preferable. © 2013 The Authors. Statistics in Medicine published by John Wiley & Sons, Ltd..