Sample size for positive and negative predictive value in diagnostic research using case-control designs

Important properties of diagnostic methods are their sensitivity,specificity, and positive and negative predictive values (PPVand NPV). These methods are typically assessed via case–controlsamples, which include one cohort of cases known to have thedisease and a second control cohort of disease-free subjects.Such studies give direct estimates of sensitivity and specificitybut only indirect estimates of PPV and NPV, which also dependon the disease prevalence in the tested population. The motivatingexample arises in assay testing, where usage is contemplatedin populations with known prevalences. Further instances includebiomarker development, where subjects are selected from a populationwith known prevalence and assessment of PPV and NPV is crucial,and the assessment of diagnostic imaging procedures for rarediseases, where case–control studies may be the only feasibledesigns. We develop formulas for optimal allocation of the samplebetween the case and control cohorts and for computing samplesize when the goal of the study is to prove that the test procedureexceeds pre-stated bounds for PPV and/or NPV. Surprisingly,the optimal sampling schemes for many purposes are highly unbalanced,even when information is desired on both PPV and NPV.     

Cytomegalovirus infection and coronary heart disease risk: a meta-analysis

The chronic inflammatory process including cytomegalovirus (CMV) infection has been hypothesized to induce the progression of atherosclerosis in coronary heart disease (CHD). Numbers studies were conducted to analyze the association between CMV infection and risk of CHD, but no clear consensus had been reached. To assess this relationship more precisely, a meta-analysis was performed. The electronic databases PubMed, Embase, and CNKI were searched; data were extracted and analyzed independently by two investigators. Ultimately, 55 studies, involving 9,000 cases and 8,608 controls from six prospective studies (all with a nested case–control design) and 49 retrospective case–control studies were included. Overall, people exposed to CMV infection had an odds ratio (OR) of 1.67 (95% CI, 1.56–1.79) for CHD risk, relative to those not exposed. CMV infection was clearly identified as a risk factor for CHD in both prospective studies (OR, 1.31; 95% CI, 1.132–1.517) and retrospective studies (OR, 1.79; 95% CI, 1.659–1.939), and in both Asian group (OR, 2.69; 95% CI, 2.304–3.144) and non-Asian group (OR, 1.48; 95% CI, 1.371–1.600). Interestingly, in the subgroup analyses by detection methods of CMV, the increased risk (OR, 8.121) was greater among studies using polymerase chain reaction than the risk (OR, 1.561) among studies using enzyme-linked immunosorbent assay. In conclusion, this meta-analysis suggested that CMV infection is associated with an increased risk for CHD, especially among Asian populations.     

Power analysis for case–control association studies of samples with known family histories

Genome-wide case–control studies have been widely used to identify genetic variants that predispose to human diseases. Such studies are powerful in detecting common genetic variants with moderate effects, but quickly lose power as allele frequency and genotype relative risk decrease. Because patients with one or more affected relatives are more likely to inherit disease-predisposing alleles of a genetic disease than patients without family histories of the disease, sampling patients with affected relatives almost always increases the frequency of disease predisposing alleles in cases and improves the power of case–control association studies. This paper evaluates the power of case–control studies that select cases and/or controls according to their family histories of disease. Our results showed that this study design can dramatically increase the power of a case–control association study for a wide range of disease types. Because each additional affected relative of a patient reduces the required sample size roughly by a pair of case and control, inclusion of cases with affected relatives can dramatically decrease the required sample size and thus the cost of such studies.     

Self-reported ethnicity, genetic structure and the impact of population stratification in a multiethnic study

It is well-known that population substructure may lead to confounding in case–control association studies. Here, we examined genetic structure in a large racially and ethnically diverse sample consisting of five ethnic groups of the Multiethnic Cohort study (African Americans, Japanese Americans, Latinos, European Americans and Native Hawaiians) using 2,509 SNPs distributed across the genome. Principal component analysis on 6,213 study participants, 18 Native Americans and 11 HapMap III populations revealed four important principal components (PCs): the first two separated Asians, Europeans and Africans, and the third and fourth corresponded to Native American and Native Hawaiian (Polynesian) ancestry, respectively. Individual ethnic composition derived from self-reported parental information matched well to genetic ancestry for Japanese and European Americans. STRUCTURE-estimated individual ancestral proportions for African Americans and Latinos are consistent with previous reports. We quantified the East Asian (mean 27%), European (mean 27%) and Polynesian (mean 46%) ancestral proportions for the first time, to our knowledge, for Native Hawaiians. Simulations based on realistic settings of case–control studies nested in the Multiethnic Cohort found that the effect of population stratification was modest and readily corrected by adjusting for race/ethnicity or by adjusting for top PCs derived from all SNPs or from ancestry informative markers; the power of these approaches was similar when averaged across causal variants simulated based on allele frequencies of the 2,509 genotyped markers. The bias may be large in case-only analysis of gene by gene interactions but it can be corrected by top PCs derived from all SNPs.     

Peptidoglycan recognition proteins Pglyrp3 and Pglyrp4 are encoded from the epidermal differentiation complex and are candidate genes for the Psors4 locus on chromosome 1q21

Psoriasis is a common inflammatory skin disease caused by genetic and environmental factors, including bacterial and viral infections. Since the skin is in constant contact with commensal and pathogenic microorganisms, we examined well-supported psoriasis genetic linkage intervals to identify genes encoding innate immune pattern recognition proteins that may play a role in pathogenesis. Two peptidoglycan recognition proteins, Pglyrp3 and Pglyrp4, are localized to the Psors4 locus on chromosome 1q21 in a gene cluster known as the epidermal differentiation complex (EDC). We show that these genes are expressed in the skin as well as in germinal centers in the tonsil. We tested 13 SNPs in or near these genes for association with psoriasis in two independent patient collections: a family-based patient set comprised of 375 individuals from 101 families, and a case–control patient collection of 282 patients with moderate to severe psoriasis and 192 healthy controls. In the family-based analysis, several SNPs in the Pglyrp3–Pglyrp4 locus show association with psoriasis (0.01<P<0.05). Multiple-SNP haplotypes incorporating Pglyrp3 and Pglyrp4 SNPs also show significant association in the transmission disequilibrium test (TDT; P<0.01). In the case–control test, none of the SNPs that we tested show association with psoriasis when analyzed in single-SNP or haplotype-based tests. The discordance between the TDT and case–control results suggests that the two populations are significantly different in disease etiology, that the polymorphism responsible for the Psors4 linkage is elsewhere in the Pglyrp locus, or that the causative Psors4 polymorphism is in a location near but not in the Pglyrp locus. These data are consistent with previous reports of association of psoriasis with genes on 1q21, and suggest a role for Pglyrps in skin biology. Electronic Supplementary Material Supplementary material is available for this article at and is accessible for authorized users.     

Adjusting for selection bias in retrospective, case-control studies

Retrospective case–control studies are more susceptibleto selection bias than other epidemiologic studies as by designthey require that both cases and controls are representativeof the same population. However, as cases and control recruitmentprocesses are often different, it is not always obvious thatthe necessary exchangeability conditions hold. Selection biastypically arises when the selection criteria are associatedwith the risk factor under investigation. We develop a methodwhich produces bias-adjusted estimates for the odds ratio. Ourmethod hinges on 2 conditions. The first is that a variablethat separates the risk factor from the selection criteria canbe identified. This is termed the "bias breaking" variable.The second condition is that data can be found such that a bias-correctedestimate of the distribution of the bias breaking variable canbe obtained. We show by means of a set of examples that suchbias breaking variables are not uncommon in epidemiologic settings.We demonstrate using simulations that the estimates of the oddsratios produced by our method are consistently closer to thetrue odds ratio than standard odds ratio estimates using logisticregression. Further, by applying it to a case–controlstudy, we show that our method can help to determine whetherselection bias is present and thus confirm the validity of studyconclusions when no evidence of selection bias can be found.     

Lipoprotein-associated phospholipase A2 gene V279F polymorphisms and coronary heart disease: a meta-analysis

Lipoprotein-associated phospholipase A2 (LP-PLA2) may play an important role in the pathophysiology of coronary heart disease (CHD). The polymorphism of LP-PLA2 gene caused LP-PLA2 enzyme activity depressing or lost. But there is not a definite conclusion for the association of between the LP-PLA2 gene polymorphism and CHD risk. To assess the relationship between LP-PLA2 gene V279F polymorphism and CHD, a comprehensive Meta-analysis was performed. All the case–control studies evaluating the association of between the LP-PLA2 gene V279F polymorphism and CHD risk were identified. Seven case–control studies involving 3,614 patients with CHD and 4,334 controls were included. The crude odds ratios (ORs) of meta-analysis under the different gene model were not significant. But in the stratified analysis by study size, ethnicity, cases definition, and source of controls under the additive model, the association was evident in ethnicity for Japanese group (OR = 1.38, 95%CI = 1.22–1.56), cases definition for MI (OR = 1.22, 95%CI = 1.01–1.49), source of controls for the based-hospital (OR = 1.42, 95%CI = 1.24–1.59). These data suggested that the V279F polymorphism in LP-PLA2 gene may contribute to CHD development. But there is necessary that more well-designed large studies are required for the validation of this association.     

Association of C47T polymorphism in <Emphasis Type="Italic">SOD2</Emphasis> gene with coronary artery disease: a case–control study and a meta-analysis

Oxidative damage promotes atherosclerosis. SOD2 is an important antioxidant enzyme. A case–control study and a meta-analysis were performed to assess the association of C47T polymorphism in SOD2 gene with premature, late-onset and overall coronary artery disease (CAD) risk. A hospital-based case–control study was conducted with 269 premature CAD cases, 278 late-onset CAD cases and 299 healthy controls. Polymerase chain reaction (PCR) and Pyrosequencing were used to detect the polymorphism. Multinomial logistic regression model was performed to estimate odds ratio (OR) with 95% confidence intervals (CIs) and adjust potential confounders. A meta-analysis was performed using eight outcomes including our result. Fixed or random effect pooled measure was selected on the basis of homogeneity test among studies. Heterogeneity among studies was evaluated using I ². Meta-regression was used to explore potential sources of between-study heterogeneity. Publication bias was estimated using Peters’s linear regression test. In our case–control study, compared with the TT as the reference, the mutant genotype of CC + TC was significantly associated with a reduced premature CAD risk both in univariate (OR = 0.60, 95% CI = 0.41–0.87) and multivariate (OR = 0.59, 95% CI = 0.40–0.87) logistic regressions, but not with late-onset CAD risk. After excluding one article that deviated from Hardy–Weinberg equilibrium in controls, this meta-analysis showed a significant association of the C allele with reduced risk of CAD in dominant (FEM: OR = 0.69, 95% CI = 0.61–0.78), recessive (OR = 0.64, 95% CI = 0.50–0.82), and codominant (FEM: OR = 0.73, 95% CI = 0.65–0.80) models. Our study suggested that the mutant genotype of CC + TC was significantly associated with a reduced CAD risk.     

Varied association of prothrombin G20210A polymorphism with coronary artery disease susceptibility in different ethnic groups: evidence from 15,041 cases and 21,507 controls

Published data on the association between prothrombin G20210A polymorphism and coronary artery disease (CAD) risk are inconclusive. To derive a more precise estimation of the relationship, a meta-analysis was performed. A total of 42 case–control studies including 15,041 cases and 21,507 controls were included in this meta-analysis. Overall, significantly elevated CAD risk was associated with prothrombin G20210A polymorphism (OR, 1.22; 95% CI 1.07–1.40; P = 0.003) when 39 eligible studies were pooled into the meta-analysis. In the subgroup analysis, borderline statistically increased risk was found for myocardial infarction in 22 case–control studies (OR, 1.27; 95% CI 1.00–1.61; P = 0.05). When stratified by ethnicity, significantly elevated risk was found in Europeans (OR, 1.19; 95% CI, 1.02–1.38; P = 0.02). However, no statistical differences were found among Americans and Asians. In summary, this meta-analysis indicated that prothrombin G20210A allele is a low-penetrant risk factor for developing CAD in Europeans.     

A General and Robust Framework for Secondary Traits Analysis

Case–control designs are commonly employed in genetic association studies. In addition to the case–control status, data on secondary traits are often collected. Directly regressing secondary traits on genetic variants from a case–control sample often leads to biased estimation. Several statistical methods have been proposed to address this issue. The inverse probability weighting (IPW) approach and the semiparametric maximum-likelihood (SPML) approach are the most commonly used. A new weighted estimating equation (WEE) approach is proposed to provide unbiased estimation of genetic associations with secondary traits, by combining observed and counterfactual outcomes. Compared to the existing approaches, WEE is more robust against biased sampling and disease model misspecification. We conducted simulations to evaluate the performance of the WEE under various models and sampling schemes. The WEE demonstrated robustness in all scenarios investigated, had appropriate type I error, and was as powerful or more powerful than the IPW and SPML approaches. We applied the WEE to an asthma case–control study to estimate the associations between the thymic stromal lymphopoietin gene and two secondary traits: overweight status and serum IgE level. The WEE identified two SNPs associated with overweight in logistic regression, three SNPs associated with serum IgE levels in linear regression, and an additional four SNPs that were missed in linear regression to be associated with the 75th quantile of IgE in quantile regression. The WEE approach provides a general and robust secondary analysis framework, which complements the existing approaches and should serve as a valuable tool for identifying new associations with secondary traits.     

Causal mediation analysis in nested case-control studies using conditional logistic regression

The paper proposes an approach to causal mediation analysis in nested case-control study designs, often incorporated with countermatching schemes using conditional likelihood, and we compare the method's performance to that of mediation analysis using the Cox model for the full cohort with a continuous or dichotomous mediator. Simulation studies are conducted to assess our proposed method and investigate the efficiency relative to the cohort. We illustrate the method using actual data from two studies of potential mediation of radiation risk conducted within the Adult Health Study cohort of atomic-bomb survivors. The performance becomes comparable to that based on the full cohort, illustrating the potential for valid mediation analysis based on the reduced data obtained through the nested case-control design.     

Association of the GNAS locus with severe malaria

Functional studies have demonstrated an interaction between the stimulatory G protein alpha subunit (G-alpha-s) and the malaria parasite at a cellular level. Obstruction of signal transduction via the erythrocyte G-alpha-s subunit reduced invasion by Plasmodium falciparum parasites. We sought to determine whether this signal pathway had an impact at the disease level by testing polymorphisms in the gene encoding G-alpha-s (GNAS) for association with severe malaria in a large multi-centre study encompassing family and case–control studies from The Gambia, Kenya and Malawi, and a case–control study from Ghana. We gained power to detect association using meta-analysis across the seven studies, with an overall sample size approximating 4,000 cases and 4,000 controls. Out of 12 SNPs investigated in the 19 kb GNAS region, four presented signals of association (P < 0.05) with severe malaria. The strongest single-locus association demonstrated an odds ratio of 1.13 (1.05–1.21), P = 0.001. Three of the loci presenting significant associations were clustered at the 5-prime end of the GNAS gene. Accordingly, haplotypes constructed from these loci demonstrated significant associations with severe malaria [OR = 0.88 (0.81–0.96), P = 0.005 and OR = 1.12 (1.03–1.20), P = 0.005]. The evidence presented here indicates that the influence of G-alpha-s on erythrocyte invasion efficacy may, indeed, alter individual susceptibility to disease.     

Genetic variation in <Emphasis Type="Italic">IGFBP2</Emphasis> and <Emphasis Type="Italic">IGFBP5</Emphasis> is associated with breast cancer in populations of African descent

The insulin-like growth factor (IGF) signaling pathway is thought to play a major role in the etiology of breast cancer. Although incidence rates of breast cancer overall are lower in African Americans than in Caucasians, African-American women have a higher incidence under age 40 years, are diagnosed with more advanced disease, and have poorer prognosis. We investigated the association of breast cancer and genetic variants in genes in the IGF signaling pathway in a population-based case–control study of African-American women. We found significant associations at a locus encompassing parts of the IGFBP2 and IGFBP5 genes on chromosome 2q35, which we then replicated in a case–control study of Nigerian women. Based on those initial findings, we genotyped a total of 34 single nucleotide polymorphisms (SNPs) across the region in both study populations. Statistically significant associations with breast cancer were observed across approximately 50 kb of DNA sequence encompassing three exons in the 3′ end of IGFBP2 and three exons in the 3′ end of IGFBP5. SNPs were associated with breast cancer risk with P values as low as P = 0.0038 and P = 0.01 in African-Americans and Nigerians, respectively. This study is the first to report associations between genetic variants in IGFBP2 and IGFBP5 and breast cancer risk.     

Polymorphism of vascular endothelial growth factor −1154G>A (rs1570360) with cancer risk: a meta-analysis of 16 case–control studies

Published data on the association of vascular endothelial growth factor (VEGF) −1154G>A polymorphism with cancer risk is inconclusive. To derive a more precise estimation of association between VEGF −1154G>A polymorphism and the risk of cancer, we performed a meta-analysis of 7,071 cancer cases and 7,693 controls from 16 published case–control studies. Our meta-analysis didn’t reveal an association between VEGF −1154G>A polymorphism and overall cancer risk (GG vs. AA: OR: 1.08, 95% CI: 0.96–1.20; GA vs. AA: OR: 1.04, 95% CI: 0.93–1.17; recessive model: GG+GA vs. AA: OR: 1.06, 95% CI: 0.95–1.18; dominant model: GG vs. GA+AA, OR: 1.11, 95% CI: 1.00–1.24). Nevertheless, for non-Caucasians, GG homozygote may have higher cancer risk compared with either A carriers (OR: 1.58, 95% CI: 1.12–2.23) or AA homozygote (OR: 1.43, 95% CI: 1.17–1.76). No significant heterogeneity was detected except in the dominant model and “prostate cancer” subgroup analysis. More studies with larger samples are warranted to confirm these findings.     

Design and Analysis of Multiple Events Case–Control Studies

Summary In case–control research where there are multiple case groups, standard analyses fail to make use of all available information. Multiple events case–control (MECC) studies provide a new approach to sampling from a cohort and are useful when it is desired to study multiple types of events in the cohort. In this design, subjects in the cohort who develop any event of interest are sampled, as well as a fraction of the remaining subjects. We show that a simple case–control analysis of data arising from MECC studies is biased and develop three general estimating‐equation‐based approaches to analyzing data from these studies. We conduct simulation studies to compare the efficiency of the various MECC analyses with each other and with the corresponding conventional analyses. It is shown that the gain in efficiency by using the new design is substantial in many situations. We demonstrate the application of our approach to a nested case–control study of the effect of oral sodium phosphate use on chronic kidney injury with multiple case definitions.     

Stopover of migrating birds: simultaneous analysis of different marking methods enhances the power of capture–recapture analyses

Analyses of stopover parameters of migrating birds with Cormack–Jolly–Seber (CJS) capture–recapture models often suffer from low precision due to sparse data sets. Low recapture rates result in low power to detect violations of the underlying assumptions and factors influencing stopover behaviour. We studied stopover behaviour of Palearctic migrant passerines in an oasis in Mauritania, West Africa. Using capture–recapture data and systematic observations of colour-ringed birds, we analysed the effect of increased sample size on probability of stay and recapture probability and the influence of a possible trap response on these parameters. We analysed capture–recapture data with the conventional CJS model and compared the results with those from a multistate model using in addition resighting data. The analyses including resighting data resulted in a higher precision of the estimates of the probabilities of stay compared to analyses using only capture–recapture data of the same individuals. Moreover, the power to detect transients was substantially enhanced. Capture had no effect on the estimates of probability to stay and recapture probability; birds did not leave the stopover site or avoid nets as a reaction to capture. The estimates of probability of stay were up to 15.7% higher when resighting data were included, probably due to the higher power to detect transients and the elimination of the bias induced by non-random temporary emigration when both data types are considered. As a consequence, stopover duration would have been underestimated when only the capture–recapture data were available. We conclude that additional resightings of colour ringed birds can produce more accurate results needed for enhancing our understanding of stopover ecology of migrants. Electronic supplementary material Supplementary material is available in the online version of this article at and is accessible for authorized users. From the Swiss Ornithological Institute project on bird migration across the Sahara.     

An Association Analysis of Reelin Gene (<Emphasis Type="Italic">RELN</Emphasis>) Polymorphisms with Childhood Epilepsy in Eastern Indian Population from West Bengal

Epilepsy is a common neurological condition characterized by unprovoked seizure attacks. Early brain developmental abnormalities involving neuronal migration and lamination are implicated in childhood epilepsy. Reelin, a neuronal-signaling molecule plays a crucial role in these migratory processes. Therefore, reelin gene (RELN), which is located on human chromosome 7q22 is considered as a potential candidate gene for childhood epilepsy. In this study, we recruited 63 patients with childhood-onset epilepsy and 103 healthy controls from West Bengal in India. Genomic DNA isolated from leukocytes of cases and control individuals were used for genotyping analysis of 16 markers of RELN. Case–control analysis revealed significant over-representation of G/C and (G/C+C/C) genotypes, and C allele of exon 22 G/C marker (rs362691) in cases as compared to controls. Pair-wise linkage disequilibrium analysis demonstrated two separate LD blocks with moderately high D′ values in epileptic cases. Based on these data, we have carried out haplotype case–control analysis. Even though we found over-representation of A-C haplotype of intron 12 A/C/exon 22 G/C markers and haplotype combination involving G-allele of exon 22 marker in cases and controls, respectively, the overall test was not significant. LD in this region involving this marker was also more robust in epileptic cases. Taken together, the results provide possible evidences for association of exon 22 G/C marker or any marker in the vicinity, which is in LD with this marker with epilepsy in the West Bengal population. Further investigations involving higher sample sizes are warranted to validate the present finding.     

A novel survival multifactor dimensionality reduction method for detecting gene–gene interactions with application to bladder cancer prognosis

The widespread use of high-throughput methods of single nucleotide polymorphism (SNP) genotyping has created a number of computational and statistical challenges. The problem of identifying SNP–SNP interactions in case–control studies has been studied extensively, and a number of new techniques have been developed. Little progress has been made, however, in the analysis of SNP–SNP interactions in relation to time-to-event data, such as patient survival time or time to cancer relapse. We present an extension of the two class multifactor dimensionality reduction (MDR) algorithm that enables detection and characterization of epistatic SNP–SNP interactions in the context of survival analysis. The proposed Survival MDR (Surv-MDR) method handles survival data by modifying MDR’s constructive induction algorithm to use the log-rank test. Surv-MDR replaces balanced accuracy with log-rank test statistics as the score to determine the best models. We simulated datasets with a survival outcome related to two loci in the absence of any marginal effects. We compared Surv-MDR with Cox-regression for their ability to identify the true predictive loci in these simulated data. We also used this simulation to construct the empirical distribution of Surv-MDR’s testing score. We then applied Surv-MDR to genetic data from a population-based epidemiologic study to find prognostic markers of survival time following a bladder cancer diagnosis. We identified several two-loci SNP combinations that have strong associations with patients’ survival outcome. Surv-MDR is capable of detecting interaction models with weak main effects. These epistatic models tend to be dropped by traditional Cox regression approaches to evaluating interactions. With improved efficiency to handle genome wide datasets, Surv-MDR will play an important role in a research strategy that embraces the complexity of the genotype–phenotype mapping relationship since epistatic interactions are an important component of the genetic basis of disease.     

Interferon gamma +874 T/A polymorphism contributes to cancer susceptibility: a meta-analysis based on 17 case–control studies

Interferon gamma (IFN-γ) plays a pivotal role in antiproliferative, antitumor and antiviral activities. The +874 polymorphism in IFN gene region reportedly affects cancer risk. However, pertinent studies offer conflicting results. To derive a more precise estimation, we performed a meta-analysis based on 1,929 cases and 2,830 controls from 17 published case–control studies, assessing the strength of the association using odds ratios with 95% confidence intervals. Our meta-analysis showed the evidence that IFN-γ +874 T/A was not associated with increased cancer risk in ethnicity and source of controls. However, stratified analysis by cancer type indicated a significantly increased risk of cervical cancer (AT vs. TT: OR = 1.10, 95% CI = 1.02–1.19, P = 0.961 for heterogeneity). Further prospective researches with a larger single study are required to evaluate any association with other types of cancer or in other populations.