Guangyong Zou

Toward Using Confidence Intervals to Compare Correlations.

Confidence intervals are widely accepted as a preferred way to present study results. They encompass significance tests and provide an estimate of the magnitude of the effect. However, comparisons of correlations still rely heavily on significance testing. The persistence of this practice is caused primarily by the lack of simple yet accurate procedures that can maintain coverage at the nominal level in a nonlopsided manner. The purpose of this article is to present a general approach to constructing approximate confidence intervals for differences between (a) 2 independent correlations, (b) 2 overlapping correlations, (c) 2 nonoverlapping correlations, and (d) 2 independent R2s. The distinctive feature of this approach is its acknowledgment of the asymmetry of sampling distributions for single correlations. This approach requires only the availability of confidence limits for the separate correlations and, for correlated correlations, a method for taking into account the dependency between correlations. These closed-form procedures are shown by simulation studies to provide very satisfactory results in small to moderate sample sizes. The proposed approach is illustrated with worked examples.

Omega-3 Free Fatty Acids for the Maintenance of Remission in Crohn Disease: The EPIC Randomized Controlled Trials

CONTEXT Maintenance therapy for Crohn disease features the use of immunosuppressive drugs, which are associated with an increased risk of infection. Identification of safe and effective maintenance strategies is a priority. OBJECTIVE To determine whether the oral administration of omega-3 free fatty acids is more effective than placebo for prevention of relapse of Crohn disease. DESIGN, SETTING, AND PATIENTS Two randomized, double-blind, placebo-controlled studies (Epanova Program in Crohns Study 1 [EPIC-1] and EPIC-2) conducted between January 2003 and February 2007 at 98 centers in Canada, Europe, Israel, and the United States. Data from 363 and 375 patients with quiescent Crohn disease were evaluated in EPIC-1 and EPIC-2, respectively. INTERVENTIONS Patients with a Crohns Disease Activity Index (CDAI) score of less than 150 were randomly assigned to receive either 4 g/d of omega-3 free fatty acids or placebo for up to 58 weeks. No other treatments for Crohn disease were permitted. MAIN OUTCOME MEASURE Clinical relapse, as defined by a CDAI score of 150 points or greater and an increase of more than 70 points from the baseline value, or initiation of treatment for active Crohn disease. RESULTS For EPIC-1, 188 patients were assigned to receive omega-3 free fatty acids and 186 patients to receive placebo. Corresponding numbers for EPIC-2 were 189 and 190 patients, respectively. The rate of relapse at 1 year in EPIC-1 was 31.6% in patients who received omega-3 free fatty acids and 35.7% in those who received placebo (hazard ratio, 0.82; 95% confidence interval, 0.51-1.19; P = .30). Corresponding values for EPIC-2 were 47.8% and 48.8% (hazard ratio, 0.90; 95% confidence interval, 0.67-1.21; P = .48). Serious adverse events were uncommon and mostly related to Crohn disease. CONCLUSION In these trials, treatment with omega-3 free fatty acids was not effective for the prevention of relapse in Crohn disease. TRIAL REGISTRATION clinicaltrials.gov Identifiers: EPIC-1: NCT00613197, EPIC-2: NCT00074542.

On the Estimation of Additive Interaction by Use of the Four-by-two Table and Beyond

A four-by-two table with its four rows representing the presence and absence of gene and environmental factors has been suggested as the fundamental unit in the assessment of gene-environment interaction. For such a table to be more meaningful from a public health perspective, it is important to estimate additive interaction. A confidence interval procedure proposed by Hosmer and Lemeshow has become widespread. This article first reveals that the Hosmer-Lemeshow procedure makes an assumption that confidence intervals for risk ratios are symmetric and then presents an alternative that uses the conventional asymmetric intervals for risk ratios to set confidence limits for measures of additive interaction. For the four-by-two table, the calculation involved requires no statistical programs but only elementary calculations. Simulation results demonstrate that this new approach can perform almost as well as the bootstrap. Corresponding calculations in more complicated situations can be simplified by use of routine output from multiple regression programs. The approach is illustrated with three examples. A Microsoft Excel spreadsheet and SAS codes for the calculations are available from the author and the Journals website, respectively.

Extension of the modified Poisson regression model to prospective studies with correlated binary data

The Poisson regression model using a sandwich variance estimator has become a viable alternative to the logistic regression model for the analysis of prospective studies with independent binary outcomes. The primary advantage of this approach is that it readily provides covariate-adjusted risk ratios and associated standard errors. In this article, the model is extended to studies with correlated binary outcomes as arise in longitudinal or cluster randomization studies. The key step involves a cluster-level grouping strategy for the computation of the middle term in the sandwich estimator. For a single binary exposure variable without covariate adjustment, this approach results in risk ratio estimates and standard errors that are identical to those found in the survey sampling literature. Simulation results suggest that it is reliable for studies with correlated binary data, provided the total number of clusters is at least 50. Data from observational and cluster randomized studies are used to illustrate the methods.

The relationship between infliximab concentrations, antibodies to infliximab and disease activity in Crohn's disease

Objective Although low infliximab trough concentrations and antibodies to infliximab (ATI) are associated with poor outcomes in patients with Crohns disease (CD), the clinical relevance of ATI in patients with adequate infliximab concentrations is uncertain. We evaluated this question using an assay sensitive for identification of ATI in the presence of infliximab. Design In an observational study, 1487 trough serum samples from 483 patients with CD who participated in four clinical studies of maintenance infliximab therapy were analysed using a fluid phase mobility shift assay. Infliximab and ATI concentrations most discriminant for remission, defined as a C-reactive protein concentration of ≤5 mg/L, were determined by receiver operating characteristic curves. A multivariable regression model evaluated these factors as independent predictors of remission. Results Based upon analysis of 1487 samples, 77.1% of patients had detectable and 22.9% had undetectable infliximab concentrations, of which 9.5% and 71.8%, respectively, were positive for ATI. An infliximab concentration of >2.79 μg/mL (area under the curve (AUC)=0.681; 95% CI 0.632 to 0.731) and ATI concentration of <3.15 U/mL (AUC=0.632; 95% CI 0.589 to 0.676) were associated with remission. Multivariable analysis showed that concentrations of both infliximab trough (OR 1.8; 95% CI 1.3 to 2.5; p<0.001) and ATI (OR 0.57; 95% CI 0.39 to 0.81; p=0.002) were independent predictors of remission. Conclusions The development of ATI increases the probability of active disease even at low concentrations and in the presence of a therapeutic concentration of drug during infliximab maintenance therapy. Evaluation of strategies to prevent ATI formation, including therapeutic drug monitoring with selective infliximab dose intensification, is needed.

Early combined immunosuppression for the management of Crohn's disease (REACT): a cluster randomised controlled trial

BACKGROUND Conventional management of Crohns disease features incremental use of therapies. However, early combined immunosuppression (ECI), with a TNF antagonist and antimetabolite might be a more effective strategy. We compared the efficacy of ECI with that of conventional management for treatment of Crohns disease. METHODS In this open-label cluster randomised controlled trial (Randomised Evaluation of an Algorithm for Crohns Treatment, REACT), we included community gastroenterology practices from Belgium and Canada that were willing to be assigned to either of the study groups, participate in all aspects of the study, and provide data on up to 60 patients with Crohns disease. These practices were randomly assigned (1:1) to either ECI or conventional management. The computer-generated randomisation was minimised by country and practice size. Up to 60 consecutive adult patients were assessed within practices. Patients who were aged 18 years or older; documented to have Crohns disease; able to speak or understand English, French, or Dutch; able to access a telephone; and able to provide written informed consent were followed up for 2 years. The primary outcome was the proportion of patients in corticosteroid-free remission (Harvey-Bradshaw Index score ≤ 4) at 12 months at the practice level. This trial is registered with ClinicalTrials.gov, number NCT01030809. FINDINGS This study took place between March 15, 2010, and Oct 1, 2013. Of the 60 practices screened, 41 were randomly assigned to either ECI (n=22) or conventional management (n=19). Two practices (one in each group) discontinued because of insufficient resources. 921 (85%) of the 1084 patients at ECI practices and 806 (90%) of 898 patients at conventional management practices completed 12 months follow-up and were included in an intention-to-treat analysis. The 12 month practice-level remission rates were similar at ECI and conventional management practices (66·0% [SD 14·0] and 61·9% [16·9]; adjusted difference 2·5%, 95% CI -5·2% to 10·2%, p=0·5169). The 24 month patient-level composite rate of major adverse outcomes defined as occurrence of surgery, hospital admission, or serious disease-related complications was lower at ECI practices than at conventional management practices (27·7% and 35·1%, absolute difference [AD] 7·3%, hazard ratio [HR]: 0·73, 95% CI 0·62 to 0·86, p=0·0003). There were no differences in serious drug-related adverse events. INTERPRETATION Although ECI was not more effective than conventional management for controlling Crohns disease symptoms, the risk of major adverse outcomes was lower. The latter finding should be considered hypothesis-generating for future trials. ECI was not associated with an increased risk of serious drug-related adverse events or mortality. FUNDING AbbVie Pharmaceuticals.

C-Reactive Protein, Fecal Calprotectin, and Stool Lactoferrin for Detection of Endoscopic Activity in Symptomatic Inflammatory Bowel Disease Patients: A Systematic Review and Meta-Analysis.

Objectives:Persistent disease activity is associated with a poor prognosis in inflammatory bowel disease (IBD). Therefore, monitoring of patients with intent to suppress subclinical inflammation has emerged as a treatment concept. As endoscopic monitoring is invasive and resource intensive, identification of valid markers of disease activity is a priority. The objective was to evaluate the diagnostic accuracy of C-reactive protein (CRP), fecal calprotectin (FC), and stool lactoferrin (SL) for assessment of endoscopically defined disease activity in IBD.Methods:Databases were searched from inception to November 6, 2014 for relevant cohort and case-control studies that evaluated the diagnostic accuracy of CRP, FC, or SL and used endoscopy as a gold standard in patients with symptoms consistent with active IBD. Sensitivities and specificities were pooled to generate operating property estimates for each test using a bivariate diagnostic meta-analysis.Results:Nineteen studies (n=2499 patients) were eligible. The pooled sensitivity and specificity estimates for CRP, FC, and SL were 0.49 (95% confidence interval (CI) 0.34–0.64) and 0.92 (95% CI 0.72–0.96), 0.88 (95% CI 0.84–0.90) and 0.73 (95% CI 0.66–0.79), and 0.82 (95% CI 0.73–0.88) and 0.79 (95% CI 0.62–0.89), respectively. FC was more sensitive than CRP in both diseases and was more sensitive in ulcerative colitis than Crohn’s disease.Conclusions:Although CRP, FC, and SL are useful biomarkers, their value in managing individual patients must be considered in specific clinical contexts.

Polygenic determinants of severe hypertriglyceridemia

Recent genome-wide association (GWA) studies have identified new genetic determinants of complex quantitative traits, including plasma triglyceride (TG). We hypothesized that common variants associated with mild TG variation identified in GWA studies would also be associated with severe hypertriglyceridemia (HTG). We studied 132 patients of European ancestry with severe HTG (fasting plasma TG > 10 mmol/l), who had no mutations found by resequencing of candidate genes, and 351 matched normolipidemic controls. We determined genotypes for: GALNT2 rs4846914, TBL2/MLXIPL rs17145738, TRIB1 rs17321515, ANGPTL3 rs12130333, GCKR rs780094, APOA5 rs3135506 (S19W), APOA5 rs662799 (-1131T > C), APOE (isoforms) and LPL rs328 (S447X). We found that: (i) genotypes, including those of APOA5 S19W, APOA5 -1131T > C, APOE, GCKR, TRIB1 and TBL2/MLXIPL, were significantly associated with severe HTG; (ii) odds ratios for these genetic variables were significant in both univariate and multivariate regression analyses, irrespective of the presence or absence of diabetes or obesity; (iii) a significant fraction-about one-quarter-of the explained variation in disease status was associated with these genotypes. Therefore, common SNPs (single nucleotide polymorphisms) that are associated with mild TG variation in GWA studies of normolipidemic subjects are also associated with severe HTG. Our findings are consistent with the emerging model of a complex genetic trait. At the extremes of a quantitative trait, such as severe HTG, are found the cumulative contributions of both multiple rare alleles with large genetic effects and common alleles with small effects.

Sample size formulas for estimating intraclass correlation coefficients with precision and assurance

The number of subjects required to estimate the intraclass correlation coefficient in a reliability study has usually been determined on the basis of the expected width of a confidence interval. However, this approach fails to explicitly consider the probability of achieving the desired interval width and may thus provide sample sizes that are too small to have adequate chance of achieving the desired precision. In this paper, we present a method that explicitly incorporates a prespecified probability of achieving the prespecified width or lower limit of a confidence interval. The resultant closed-form formulas are shown to be very accurate.

The merits of testing Hardy-Weinberg equilibrium in the analysis of unmatched case-control data: a cautionary note.

Testing for departures from the assumption of Hardy‐Weinberg equilibrium (HWE) has been widely recommended as a preliminary step in the analysis of genetic case‐control studies. Some authors suggest using a two‐stage procedure in which gene/disease associations are ultimately evaluated using either the Pearson chi‐square procedure or the Cochran‐Armitage test for trend. Other authors go further and encourage investigators to discard data that are in violation of HWE, essentially using the test as a tool for identifying genotyping errors. In this paper we show that 1) testing for HWE should not be used as a tool to identify genotyping errors; and 2) it is not necessary, and possibly even harmful, to test the HWE assumption before testing for association between alleles and disease. Instead one should inherently account for deviations from HWE with an adjusted chi‐square test statistic, a procedure which in the present context is identical to the trend test. Examples from previous reports are used to illustrate the methodology.