Dichotomous Covariate Research Articles

Multiple imputation of missing values was not necessary before performing a longitudinal mixed-model analysis

Background and ObjectivesAs a result of the development of sophisticated techniques, such as multiple imputation, the interest in handling missing data in longitudinal studies has increased enormously in past years. Within the field of longitudinal data analysis, there is a current debate on whether it is necessary to use multiple imputations before performing a mixed-model analysis to analyze the longitudinal data. In the current study this necessity is evaluated. Study Design and SettingThe results of mixed-model analyses with and without multiple imputation were compared with each other. Four data sets with missing values were created—one data set with missing completely at random, two data sets with missing at random, and one data set with missing not at random). In all data sets, the relationship between a continuous outcome variable and two different covariates were analyzed: a time-independent dichotomous covariate and a time-dependent continuous covariate. ResultsAlthough for all types of missing data, the results of the mixed-model analysis with or without multiple imputations were slightly different, they were not in favor of one of the two approaches. In addition, repeating the multiple imputations 100 times showed that the results of the mixed-model analysis with multiple imputation were quite unstable. ConclusionIt is not necessary to handle missing data using multiple imputations before performing a mixed-model analysis on longitudinal data.

Estimation using all available covariate information versus a fixed look‐back window for dichotomous covariates

When using claims data, dichotomous covariates (C) are often assumed to be absent unless a claim for the condition is observed. When available historical data differs among subjects, investigators must choose between using all available historical data versus data from a fixed window to assess C. Our purpose was to compare estimation under these two approaches. We simulated cohorts of 20,000 subjects with dichotomous variables representing exposure (E), outcome (D), and a single time-invariant C, as well as varying availability of historical data. C was operationally defined under each paradigm and used to estimate the adjusted risk ratio of E on D via Mantel-Haenszel methods. In the base case scenario, less bias and lower mean square error were observed using all available information compared with a fixed window; differences were magnified at higher modeled confounder strength. Upon introduction of an unmeasured covariate (F), the all-available approach remained less biased in most circumstances and rendered estimates that better approximated those that were adjusted for the true (modeled) value of C in all instances. In most instances considered, operationally defining time-invariant dichotomous C based on all available historical data, rather than on data observed over a commonly shared fixed historical window, results in less biased estimates.

Correction: ‘Generating survival times to simulate Cox proportional hazards models with time‐varying covariates’

In a recent paper, we derived closed-form expressions for simulating survival times from a Cox proportional hazards model with time-dependent covariates 1. We considered three different distributions for the distribution of event times: exponential, Weibull and Gompertz. We considered three different types of time-dependent covariates: (i) a dichotomous time-varying covariate that can change at most once from untreated to treated; (ii) a dichotomous time-varying covariate with a subject being able to move repeatedly between treatment states; (iii) a continuous time-varying covariate of the form kt, where t denotes time. For each of the nine different scenarios considered, we derived a closed-form expression that allows one to simulate survival times from the specified distribution of event times, given the specified type of time-varying covariate. Each derivation involved determining the cumulative hazard function and the inverse of the cumulative hazard function. Unfortunately, we made two errors in the derivations 2. We made a minor error in deriving the expression for the cumulative hazard function for a Weibull distribution of event times under the first type of time-varying covariate (Section 3.1.2). In the final expression for the cumulative hazard function, H(t,x,z(t)), for the case when t ⩾ t0, we inadvertently did not cancel the υ from the denominator when it was canceled from the numerator. However, we correctly cancelled this term from the denominator in the subsequent derivations in this section when the inverse of the cumulative hazard function was determined. Thus, the final closed-form expression for generating survival times in this scenario is correct. A second error occurred in deriving a closed-form expression for the cumulative hazard function for a Weibull distribution of event times with a continuous time-varying covariate (Section 3.2.2). When determining the cumulative hazard function, we incorrectly evaluated the following integral: . The correct integration is , where Γ(x) denotes the gamma function and Γ(x,b) the upper incomplete gamma function. Accordingly, the cumulative hazard function involves the lower incomplete gamma function. As such, a closed-form expression for the cumulative hazard function does not exist in this particular scenario. Thus, a closed-form expression for generating survival times from a Weibull distribution with a continuous time-varying covariate does not exist. As noted elsewhere, one can evaluate this integral numerically 2. Alternatively, researchers wanting to simulate survival times with continuous time-dependent covariates are encouraged to consider either an exponential distribution of event times (Section 3.2.1, formula (4)) or a Gompertz distribution of event times (Section 3.2.3, formula (6)).

Binary response models with [formula omitted]-phase case-control data

In this study, a more general single-index regression model was presented to characterize the relationship between a dichotomous response and covariates of interest. With M-phase (M≥2) case-control data supplemented by information on a response and certain covariates, we propose a pseudo maximum likelihood estimation for the index coefficients. In the receiver operating characteristic curve analysis, an estimation for the accuracy measure is further provided and is borrowed to seek an optimal linear predictor. As for the hypothesis of model correctness, a pseudo least squares approach is employed as an aid to devising suitable testing procedures. Moreover, the general theoretical frameworks of these estimators are well developed. Finally, extensive simulations and two empirical applications are used to illustrate the applicability of our methodology.

Log-Link Regression Models for Ordinal Responses

The adjacent-categories, continuation-ratio and proportional odds logit-link regression models provide useful extensions of the multinomial logistic model to ordinal response data. We propose fitting these models with a logarithmic link to allow estimation of different forms of the risk ratio. Each of the resulting ordinal response log-link models is a constrained version of the log multinomial model, the log-link counterpart of the multinomial logistic model. These models can be estimated using software that allows the user to specify the log likelihood as the objective function to be maximized and to impose constraints on the parameter estimates. In example data with a dichotomous covariate, the unconstrained models produced valid coefficient estimates and standard errors, and the constrained models produced plausible results. Models with a single continuous covariate performed well in data simulations, with low bias and mean squared error on average and appropriate confidence interval coverage in admissible solutions. In an application to real data, practical aspects of the fitting of the models are investigated. We conclude that it is feasible to obtain adjusted estimates of the risk ratio for ordinal outcome data.

A Simulation-Based Evaluation of the Asymptotic Power Formulas for Cox Models in Small Sample Cases

Cox proportional hazards (PH) models are commonly used in medical research to investigate the associations between covariates and time-to-event outcomes. It is frequently noted that with less than 10 events per covariate, these models produce spurious results and therefore should not be used. Statistical literature contains asymptotic power formulas for the Cox model which can be used to determine the number of events needed to detect an association. Here, we investigate via simulations the performance of these formulas in small sample settings for Cox models with one or two covariates. Our simulations indicate that when the number of events is small, the power estimate based on the asymptotic formula is often inflated. The discrepancy between the asymptotic and empirical power is larger for the dichotomous covariate especially in cases where allocation of sample size to its levels is unequal. When more than one covariate is included in the same model, the discrepancy between the asymptotic power and the empirical power is even larger, especially when a high positive correlation exists between the two covariates.

Generating survival times to simulate Cox proportional hazards models with time‐varying covariates

Simulations and Monte Carlo methods serve an important role in modern statistical research. They allow for an examination of the performance of statistical procedures in settings in which analytic and mathematical derivations may not be feasible. A key element in any statistical simulation is the existence of an appropriate data-generating process: one must be able to simulate data from a specified statistical model. We describe data-generating processes for the Cox proportional hazards model with time-varying covariates when event times follow an exponential, Weibull, or Gompertz distribution. We consider three types of time-varying covariates: first, a dichotomous time-varying covariate that can change at most once from untreated to treated (e.g., organ transplant); second, a continuous time-varying covariate such as cumulative exposure at a constant dose to radiation or to a pharmaceutical agent used for a chronic condition; third, a dichotomous time-varying covariate with a subject being able to move repeatedly between treatment states (e.g., current compliance or use of a medication). In each setting, we derive closed-form expressions that allow one to simulate survival times so that survival times are related to a vector of fixed or time-invariant covariates and to a single time-varying covariate. We illustrate the utility of our closed-form expressions for simulating event times by using Monte Carlo simulations to estimate the statistical power to detect as statistically significant the effect of different types of binary time-varying covariates. This is compared with the statistical power to detect as statistically significant a binary time-invariant covariate. Copyright © 2012 John Wiley & Sons, Ltd.

Applications of the Bootstrap in ROC Analysis

The problem of estimating standard errors for diagnostic accuracy measures might be challenging for many complicated models. We can address such a problem by using the Bootstrap methods to blunt its technical edge with resampled empirical distributions. We consider two cases where bootstrap methods can successfully improve our knowledge of the sampling variability of the diagnostic accuracy estimators. The first application is to make inference for the area under the ROC curve resulted from a functional logistic regression model which is a sophisticated modelling device to describe the relationship between a dichotomous response and multiple covariates. We consider using this regression method to model the predictive effects of multiple independent variables on the occurrence of a disease. The accuracy measures, such as the area under the ROC curve (AUC) are developed from the functional regression. Asymptotical results for the empirical estimators are provided to facilitate inferences. The second application is to test the difference of two weighted areas under the ROC curve (WAUC) from a paired two sample study. The correlation between the two WAUC complicates the asymptotic distribution of the test statistic. We then employ the bootstrap methods to gain satisfactory inference results. Simulations and examples are supplied in this article to confirm the merits of the bootstrap methods.

Incorporating a Patient Dichotomous Characteristic in Cancer Phase I Clinical Trials Using Escalation with Overdose Control

We describe a design for cancer phase I clinical trials that takes into account patients heterogeneity thought to be related to treatment susceptibility. The goal is to estimate the maximum tolerated dose (MTD) given patient’s specific dichotomous covariate value. The design is Bayesian adaptive and is an extension of escalation with overdose control (EWOC). We will assess the performance of this method by comparing the following designs via extensive simulations: (1) design using a covariate; patients are accrued to the trial sequentially and the dose given to a patient depends on his/her baseline covariate value, (2) design ignoring the covariate; patients are accrued to the trial sequentially and the dose given to a patient does not depend on his/her baseline covariate value, and (3) design using separate trials; in each group, patients are accrued to the trial sequentially and EWOC is implemented in each group. These designs are compared with respect to safety of the trial and efficiency of the estimates of the MTDs via extensive simulations. We found that ignoring a significant baseline binary covariate in the model results in a substantial number of patients being overdosed. On the other hand, accounting for a nonsignificant covariate in the model has practically no effect on the safety of the trial and efficiency of the estimates of the MTDs.

Factors affecting variability in the urinary biomarker 1,6-hexamethylene diamine in workers exposed to 1,6-hexamethylene diisocyanate

Although urinary 1,6-hexamethylene diamine (HDA) is a useful biomarker of exposure to 1,6-hexamethylene diisocyanate (HDI), a large degree of unexplained intra- and inter-individual variability exists between estimated HDI exposure and urine HDA levels. We investigated the effect of individual and workplace factors on urine HDA levels using quantitative dermal and inhalation exposure data derived from a survey of automotive spray painters exposed to HDI. Painters' dermal and breathing-zone HDI-exposures were monitored over an entire workday for up to three separate workdays, spaced approximately one month apart. One urine sample was collected before the start of work with HDI-containing paints, and multiple samples were collected throughout the workday. Using mixed effects multiple linear regression modeling, coverall use resulted in significantly lower HDA levels (p = 0.12), and weekday contributed to significant variability in HDA levels (p = 0.056). We also investigated differences in urine HDA levels stratified by dichotomous and classification covariates using analysis of variance. Use of coveralls (p = 0.05), respirator type worn (p = 0.06), smoker status (p = 0.12), paint-booth type (p = 0.02), and more than one painter at the shop (p = 0.10) were all found to significantly affect urine HDA levels adjusted for creatinine concentration. Coverall use remained significant (p = 0.10), even after adjusting for respirator type. These results indicate that the variation in urine HDA level is mainly due to workplace factors and that appropriate dermal and inhalation protection is required to prevent HDI exposure.

Power Calculation for Likelihood Ratio Tests in the Nested Case-control Designs with a Randomly Sampled Control Per Case

This paper presents an asymptotic method of power calculations for likelihood ratio tests in the nested case-control designs with a randomly sampled control per case. It is an extension of the approach described in Self et al. (1992) for proportional hazards models. Our approach here focuses on a simple scenario with 1 : 1 case-control ratio, simple random sampling design, and two independent dichotomous covariates with no interaction effects. The approximation of the noncentrality of the noncentral chi-square distribution for the likelihood ratio statistic is provided. Simulation studies are conducted to examine the accuracy for several parameter values and data configurations. Overall the results suggest that estimates of power using our proposed method are consistent with those of actual power from Monte Carlo simulation. Therefore, the proposed approach can be practically useful in assessing the statistical power for the simple nested case-control design.

Anaesthetic effects of propofol polymeric micelle: a novel water soluble propofol formulation

BackgroundAs a result of its very low water solubility, propofol is generally presented as a lipid-based formulation with well-characterized limitations. MethodsPropofol (99.7%) was added directly to an aqueous solution of poly(N-vinyl-2-pyrrolidone)-block-poly(d,l-lactide)copolymers (PVP-PLA) block copolymers and stirred in order to obtain a clear solution. This formulation was filtered sterile and then lyophilized to its solid form Propofol-PM (propofol polymeric micelle) which reconstitutes to a propofol 1%w/v (10 mg ml−1) clear aqueous solution of 30–60 nm propofol-containing micelles. Population pharmacokinetic data from whole blood and plasma were obtained by administering reconstituted Propofol-PM formulations and a 1% oil in water formulation, Diprivan® to male Sprague–Dawley rats (n = 40) at a dose of 10 mg kg−1. Preliminary recovery data were obtained from a further small study. ResultsThe pharmacokinetics were best described using a two-compartment mamillary population model, which incorporated sample matrix (blood or plasma) and propofol formulation (Diprivan® or Propofol-PM) as covariates. Sample matrix was applied to all structural model parameters as a dichotomous covariate. An influence of propofol formulation was observed for all parameters (excluding distributional clearance) but only when plasma was used for propofol quantification. In this preliminary pharmacodynamic study, there was no statistically significant difference in the timing of the recovery endpoints between the Propofol-PM formulation and Diprivan® groups. ConclusionsPropofol-PM formulations produce anaesthesia in rats. Whole blood pharmacokinetics of Propofol-PM did not differ from those observed with Diprivan®.

The relationship between the development of rash and clinical and quality of life outcomes in colorectal cancer patients treated with cetuximab in NCIC CTG CO.17

4130 Background: The National Cancer Institute of Canada Clinical Trials Group CO.17 trial, conducted in collaboration with the Australasian Gastrointestinal Trials Group, demonstrated cetuximab monotherapy (vs best supportive care - BSC) improved overall (OS) and progression-free survival (PFS) and maintained quality of life (QoL) in patients previously treated for advanced colorectal cancer that expressed the Epidermal Growth Factor Receptor (EGFR). We analyzed the correlation between the occurrence of rash and clinical and QoL outcomes in CO.17 patients who received cetuximab therapy. Methods: Rash was graded weekly according to NCI CTC version 2.0 criteria. In the principle analysis, patients were classified into the “no rash” group until rash of any grade developed (= entered “rash” group). A Cox model with rash as a dichotomous time-dependent covariate (CMTDC) was used to compare clinical and QoL outcomes between these two groups. Two additional exploratory landmark-type analyses (LTA) were also performed by first excluding all patients who died within 28 days of randomization, and then classifying patients according to severity of rash (grade 0, 1, 2+) based on 1) the worst grade of rash ever developed (LTA1) and 2) the worst grade of rash on or before day 28 of the study (LTA2). These three groups were compared in Cox models as fixed covariates. Results: In patients who received cetuximab and experienced rash, the median time to onset of rash was 10 days and 90% experienced rash within 29 days. The hazard ratio (HR) of OS between the “rash” and “no rash” groups estimated from CMTDC was 0.58 (95% CI: 0.38–0.87; p=0.009). The median OS and corresponding HR from two landmark- type analyses are presented in the table below. Results for other clinical outcomes and QoL will be presented. Conclusions: All analyses demonstrated a strong correlation between rash development and clinical benefit in colorectal patients treated with cetuximab. Median Overall Survival Worst ever grade of rash (LTA1) Worst grade of rash by day 28 (LTA2) Grade 0 2.6 months 4.3 months Grade 1 4.8 months 4.6 months Grade 2+ 8.4 months 9.1 months HR: Grade 2+ vs Grade 0 0.33 (95% CI 0.22–0.50; p<0.001) 0.52 (95% CI: 0.37, 0.74; p=0.0002) HR: Grade 2+ vs Grade 1 0.54 (95% CI: 0.41–0.72; p<0.0001) 0.53 (95% CI: 0.39–0.72; p<0.0001) No significant financial relationships to disclose.

COX-2 expression predicts prostate-cancer outcome: analysis of data from the RTOG 92-02 trial

COX-2 is overexpressed in some cancers, including prostate cancer; however, little is known about the effect of COX-2 overexpression on outcome in radiation-treated patients with prostate cancer. We aimed to study COX-2 overexpression and outcome in a well-defined cohort of men who received treatment with short-term androgen deprivation (STAD) plus radiotherapy or long-term androgen deprivation (LTAD) plus radiotherapy. Men with prostate cancer who had participated in the Radiation Therapy Oncology Group (RTOG) 92-02 trial and for whom sufficient diagnostic tissue was available for immunohistochemical staining and image analysis of COX-2 expression were enrolled in this study. Patients in the 92-02 trial had been randomly assigned to treatment with STAD plus radiotherapy or LTAD plus radiotherapy. Multivariate analyses by Cox proportional hazards models were done to assess whether associations existed between COX-2 staining intensity and the RTOG 92-02 primary endpoints of biochemical failure (assessed by the American Society for Therapeutic Radiology and Oncology [ASTRO] and Phoenix criteria), local failure, distant metastasis, cause-specific mortality, overall mortality, and any failure. 586 patients with sufficient diagnostic tissue for immunohistochemical staining and image analysis of COX-2 expression were included in this study. In the multivariate analyses, the intensity of COX-2 staining as a continuous covariate was an independent predictor of distant metastasis (hazard ratio [HR] 1.181 [95% CI 1.077-1.295], p=0.0004); biochemical failure by two definitions (ASTRO HR 1.073 [1.018-1.131], p=0.008; Phoenix HR 1.073 [1.014-1.134], p=0.014); and any failure (HR 1.068 [1.015-1.124], p=0.011). The higher the expression of COX-2, the greater the chance of failure. As a dichotomous covariate, COX-2 overexpression seemed to be most discriminating of outcome for those who received STAD compared with those who received LTAD. To our knowledge, this is the first study to establish an association of COX-2 expression with outcome in patients with prostate cancer who have had radiotherapy. Increasing COX-2 expression was significantly associated with biochemical failure, distant metastasis, and any failure. COX-2 inhibitors might improve patient response to radiotherapy in those treated with or without androgen deprivation. Our findings suggest that LTAD might overcome the effects of COX-2 overexpression. Therefore, COX-2 expression might be useful in selecting patients who need LTAD.

A comparative investigation of methods for logistic regression with separated or nearly separated data

In logistic regression analysis of small or sparse data sets, results obtained by classical maximum likelihood methods cannot be generally trusted. In such analyses it may even happen that the likelihood meets the convergence criteria while at least one parameter estimate diverges to +/-infinity. This situation has been termed 'separation', and it typically occurs whenever no events are observed in one of the two groups defined by a dichotomous covariate. More generally, separation is caused by a linear combination of continuous or dichotomous covariates that perfectly separates events from non-events. Separation implies infinite or zero maximum likelihood estimates of odds ratios, which are usually considered unrealistic. I provide some examples of separation and near-separation in clinical data sets and discuss some options to analyse such data, including exact logistic regression analysis and a penalized likelihood approach. Both methods supply finite point estimates in case of separation. Profile penalized likelihood confidence intervals for parameters show excellent behaviour in terms of coverage probability and provide higher power than exact confidence intervals. General advantages of the penalized likelihood approach are discussed.

Noninvasive assessment of energy expenditure in children

This study establishes an affordable, simple, and noninvasive method to assess energy expenditure (EE) in children, an underrepresented group. The method is based on regression modeling, where prediction of oxygen consumption (VO(2)), a proxy of EE, was deduced from heart rate (HR) and several variables that adjusted for interindividual variability. Limb activities (arms vs. legs) and posture (sitting vs. standing) were represented in the regression as dichotomous covariates. The order of activities and intensities was randomized. Seventy-four children (aged 7-10 years), raised at sea-level (Seattle, WA), comprised the sample. Anthropometric measures were taken, and VO(2) and HR were measured for activities using the arms in sitting and standing positions (mixing and punching), as well as walking at different velocities on a treadmill. Repeated measures and least square regression estimation were used. HR, body mass, number of hours of physical activity per week (HPA), an interaction term between sitting and standing resting HR, and the two dichotomous variables, sex and limbs, were significant covariates; posture was not. Several equations were developed for various field uses. The equations were built from sea-level data, but ultimately this method could serve as a baseline for developing a similar approach in other populations, where noninvasive estimation of EE is imperative in order to gain a better understanding of children's energetic issues.

Insights into different results from different causal contrasts in the presence of effect-measure modification.

Both propensity score (PS) matching and inverse probability of treatment weighting (IPTW) allow causal contrasts, albeit different ones. In the presence of effect-measure modification, different analytic approaches produce different summary estimates. We present a spreadsheet example that assumes a dichotomous exposure, covariate, and outcome. The covariate can be a confounder or not and a modifier of the relative risk (RR) or not. Based on expected cell counts, we calculate RR estimates using five summary estimators: Mantel-Haenszel (MH), maximum likelihood (ML), the standardized mortality ratio (SMR), PS matching, and a common implementation of IPTW. Without effect-measure modification, all approaches produce identical results. In the presence of effect-measure modification and regardless of the presence of confounding, results from the SMR and PS are identical, but IPTW can produce strikingly different results (e.g., RR = 0.83 vs. RR = 1.50). In such settings, MH and ML do not estimate a population parameter and results for those measures fall between PS and IPTW. Discrepancies between PS and IPTW reflect different weighting of stratum-specific effect estimates. SMR and PS matching assign weights according to the distribution of the effect-measure modifier in the exposed subpopulation, whereas IPTW assigns weights according to the distribution of the entire study population. In pharmacoepidemiology, contraindications to treatment that also modify the effect might be prevalent in the population, but would be rare among the exposed. In such settings, estimating the effect of exposure in the exposed rather than the whole population is preferable.

A method to automate probabilistic sensitivity analyses of misclassified binary variables

Misclassification bias is present in most studies, yet uncertainty about its magnitude or direction is rarely quantified. The authors present a method for probabilistic sensitivity analysis to quantify likely effects of misclassification of a dichotomous outcome, exposure or covariate. This method involves reconstructing the data that would have been observed had the misclassified variable been correctly classified, given the sensitivity and specificity of classification. The accompanying SAS macro implements the method and allows users to specify ranges of sensitivity and specificity of misclassification parameters to yield simulation intervals that incorporate both systematic and random error. The authors illustrate the method and the accompanying SAS macro code by applying it to a study of the relation between occupational resin exposure and lung-cancer deaths. The authors compare the results using this method with the conventional result, which accounts for random error only, and with the original sensitivity analysis results. By accounting for plausible degrees of misclassification, investigators can present study results in a way that incorporates uncertainty about the bias due to misclassification, and so avoid misleadingly precise-looking results.

A direct comparison of the natriuretic peptides and their relationship to survival in chronic heart failure of a presumed non‐ischaemic origin

The natriuretic peptides have been validated as sensitive and specific markers of left ventricular dysfunction; brain natriuretic peptide (BNP), N-terminal atrial natriuretic peptide (NT-proANP) and N-terminal brain natriuretic peptide (NT-proBNP) elevations have been associated with New York Heart Association (NYHA) Class I-IV heart failure. We directly compared the association of each of these markers with 1-year survival in 173 patients with chronic heart failure of a presumed nonischaemic origin entering the PRAISE-2 Trial, a clinical study which assessed the therapeutic effect of Amlodipine in patients with NYHA Class III and IV heart failure and a left ventricular ejection fraction (LVEF) <30%. BNP, NT-proBNP, and NT-proANP levels were all correlated with 1-year mortality by univariate Cox proportional hazards analyses. With respect to multivariate Cox proportional hazards regression models containing variables deemed significant in univariate analyses, NT-proANP alone was identified as an independent predictor of 1-year mortality when log-transformed continuous covariates were utilized in the analysis. When the analysis was repeated using dichotomous covariates, NT-proANP remained the most significant predictor of 1-year mortality, followed by NT-proBNP, NYHA classification and BNP. We conclude that all three natriuretic peptides are significant predictors of short-term mortality in subjects with chronic congestive heart failure (CHF) of a presumed nonischaemic origin. Larger prospective studies are required to validate the clinical utility of NT-proANP as a discriminating marker of short-term survival, and to validate proposed cutoffs of approximately 2300 pmol/l for NT-proANP, 1500 pg/ml for NT-proBNP, and 50 pmol/l for BNP as prognostic indicators of adverse short-term outcome.

The androgen receptor CAG repeat polymorphism and modification of breast cancer risk in BRCA1 and BRCA2 mutation carriers

IntroductionThe androgen receptor (AR) gene exon 1 CAG repeat polymorphism encodes a string of 9–32 glutamines. Women with germline BRCA1 mutations who carry at least one AR allele with 28 or more repeats have been reported to have an earlier age at onset of breast cancer.MethodsA total of 604 living female Australian and British BRCA1 and/or BRCA2 mutation carriers from 376 families were genotyped for the AR CAG repeat polymorphism. The association between AR genotype and disease risk was assessed using Cox regression. AR genotype was analyzed as a dichotomous covariate using cut-points previously reported to be associated with increased risk among BRCA1 mutation carriers, and as a continuous variable considering smaller allele, larger allele and average allele size.ResultsThere was no evidence that the AR CAG repeat polymorphism modified disease risk in the 376 BRCA1 or 219 BRCA2 mutation carriers screened successfully. The rate ratio associated with possession of at least one allele with 28 or more CAG repeats was 0.74 (95% confidence interval 0.42–1.29; P = 0.3) for BRCA1 carriers, and 1.12 (95% confidence interval 0.55–2.25; P = 0.8) for BRCA2 carriers.ConclusionThe AR exon 1 CAG repeat polymorphism does not appear to have an effect on breast cancer risk in BRCA1 or BRCA2 mutation carriers.