A useful design utilizing the information fraction in a group sequential clinical trial with censored survival data.

Lan and DeMets (1983) proposed the alpha spending function for group sequential trials to permit the use of unspecified frequencies and timings of interim analyses in the trial design. Regarding a trial with censored time to endpoint, Lan and DeMets (1989) later defined information time at an interim analysis in a maximum duration trial. To compare two survival curves utilizing such a design, information times for group sequential logrank and Wilcoxon-type statistics have been developed by assuming that the survival time follows an exponential distribution or a Weibull distribution without considering the censoring distribution. To better address the practical concerns inherent in clinical trials with survival endpoints, we present a new approach to adequately design a group sequential trial using the Harrington-Fleming (1982) test based on our proposed information fractions by assuming the censoring distribution depends on the patient's accrual time according to various entry distributions and by extending the underlying survival distribution to the generalized gamma distribution. We also determine associated sample sizes, expected number of events and expected stopping time. Two phase III trials of non-small-cell lung cancer originally designed using fixed-sample tests are utilized to illustrate the potential advantages of using a group sequential design with the proposed approach. This enhanced method facilitates the design and analysis of group sequential clinical trials studying survival endpoints by increasing implemental flexibility.

Ages at Onset of 5 Cardiometabolic Diseases Adjusting for Nonsusceptibility: Implications for the Pathogenesis of Metabolic Syndrome.

To shed light on the etiology of metabolic syndrome development, it is important to understand whether its 5 component disorders follow certain onset sequences. To explore disease progression of the syndrome, we studied the ages at onset of 5 cardiometabolic diseases: abdominal obesity, diabetes, hypertension, hypertriglyceridemia, and hypo-α-lipoproteinemia. In analyzing longitudinal data from the Cardiovascular Disease Risk Factors Two-Township Study (1989-2002) in Taiwan, we adjusted for nonsusceptibility, utilizing the logistic-accelerated failure time location-scale mixture regression models for left-truncated and interval-censored data to simultaneously estimate the associations of township and sex with the susceptibility probability and the age-at-onset distribution of susceptible individuals for each disease. We then validated the onset sequences of 5 cardiometabolic diseases by comparing the overall probability density curves across township-sex strata. Visualization of these curves indicates that women tended to have onsets of abdominal obesity and hypo-α-lipoproteinemia in young adulthood, hypertension and hypertriglyceridemia in middle age, and diabetes later; men tended to have onsets of abdominal obesity, hypo-α-lipoproteinemia, and hypertriglyceridemia in young adulthood, hypertension in middle age, and diabetes later. Different onset patterns of abdominal obesity, hypo-α-lipoproteinemia, and male hypertension were identified between townships. Our proposed method provides a novel strategy for investigating both pathogenesis and preventive measures of complex syndromes.

Heteroscedastic transformation cure regression models.

Cure models have been applied to analyze clinical trials with cures and age-at-onset studies with nonsusceptibility. Lu and Ying (On semiparametric transformation cure model. Biometrika 2004; 91:331?-343. DOI: 10.1093/biomet/91.2.331) developed a general class of semiparametric transformation cure models, which assumes that the failure times of uncured subjects, after an unknown monotone transformation, follow a regression model with homoscedastic residuals. However, it cannot deal with frequently encountered heteroscedasticity, which may result from dispersed ranges of failure time span among uncured subjects' strata. To tackle the phenomenon, this article presents semiparametric heteroscedastic transformation cure models. The cure status and the failure time of an uncured subject are fitted by a logistic regression model and a heteroscedastic transformation model, respectively. Unlike the approach of Lu and Ying, we derive score equations from the full likelihood for estimating the regression parameters in the proposed model. The similar martingale difference function to their proposal is used to estimate the infinite-dimensional transformation function. Our proposed estimating approach is intuitively applicable and can be conveniently extended to other complicated models when the maximization of the likelihood may be too tedious to be implemented. We conduct simulation studies to validate large-sample properties of the proposed estimators and to compare with the approach of Lu and Ying via the relative efficiency. The estimating method and the two relevant goodness-of-fit graphical procedures are illustrated by using breast cancer data and melanoma data. Copyright © 2016 John Wiley & Sons, Ltd.

Logistic-AFT location-scale mixture regression models with nonsusceptibility for left-truncated and general interval-censored data.

In conventional survival analysis there is an underlying assumption that all study subjects are susceptible to the event. In general, this assumption does not adequately hold when investigating the time to an event other than death. Owing to genetic and/or environmental etiology, study subjects may not be susceptible to the disease. Analyzing nonsusceptibility has become an important topic in biomedical, epidemiological, and sociological research, with recent statistical studies proposing several mixture models for right-censored data in regression analysis. In longitudinal studies, we often encounter left, interval, and right-censored data because of incomplete observations of the time endpoint, as well as possibly left-truncated data arising from the dissimilar entry ages of recruited healthy subjects. To analyze these kinds of incomplete data while accounting for nonsusceptibility and possible crossing hazards in the framework of mixture regression models, we utilize a logistic regression model to specify the probability of susceptibility, and a generalized gamma distribution, or a log-logistic distribution, in the accelerated failure time location-scale regression model to formulate the time to the event. Relative times of the conditional event time distribution for susceptible subjects are extended in the accelerated failure time location-scale submodel. We also construct graphical goodness-of-fit procedures on the basis of the Turnbull-Frydman estimator and newly proposed residuals. Simulation studies were conducted to demonstrate the validity of the proposed estimation procedure. The mixture regression models are illustrated with alcohol abuse data from the Taiwan Aboriginal Study Project and hypertriglyceridemia data from the Cardiovascular Disease Risk Factor Two-township Study in Taiwan.

Predictive effect of serial serum alanine aminotransferase levels on spontaneous HBeAg seroconversion in chronic genotype B and C HBV-infected children.

OBJECTIVE: The present study aimed to investigate the association between serial serum alanine aminotransferase (ALT) and spontaneous hepatitis B e antigen (HBeAg) seroconversion age in chronic hepatitis B virus (HBV)-infected children. PATIENTS AND METHODS: One hundred four HBeAg-positive chronic genotype B or C HBV-infected patients were included in this long-term prospective cohort study (mean initial age 7.20 years). Serial serum ALT levels and HBV serology markers were measured every 6 to 12 months. The 104 subjects made a total of 2525 visits during the study period, and the majority (93.6%) of visits were within a 1-year interval apart from previous visits. Cox proportional hazards model with time-dependent covariates was used in the survival analysis of HBeAg in these subjects. RESULTS: During the chronic course of HBV infection, the median remaining times to spontaneous HBeAg seroconversion were 8.35, 5.14, 4.25, 3.95, and 2.80 years after the ALT levels crossed 20, 30, 40, 60, and 150 IU/L, respectively. The incidence rate of spontaneous HBeAg seroconversion within 6 months when a subject entered the phase of ALT between 60 and 150 IU/L was 5.57 times that of the phase with ALT < 60 IU/L. The incidence rate of HBeAg seroconversion once ALT levels were above 150 IU/L was 9.87 times that of the phase of ALT < 60 IU/L. CONCLUSIONS: The ALT levels above 30 IU/L served as a cutoff of the inflammatory phase in chronic genotype B and C HBV-infected patients. Serial ALT levels in chronic HBV-infected subjects offer a predicted effect on the occurrence of spontaneous HBeAg seroconversion.

Medium-term course and outcome of schizophrenia depicted by the sixth-month subtype after an acute episode.

BACKGROUND/PURPOSE: The intermediate course of schizophrenia is a complex intertwined with the heterogeneity of the illness. This article attempts to simplify this complexity using a hypothetical tripartite based on the profile of symptoms at 6 months after acute treatment. METHODS: This is a prospective 5-year follow-up study including 163 schizophrenic inpatients in northern Taiwan comparing patients' demographic data at index admission, scores on the Positive and Negative Syndrome Scale (PANSS) for schizophrenia and social function scale measured at admission, 6-month follow-up, and annually, and scores on a neuropsychologic test battery measured approximately 5 years after recruitment. RESULTS: Patients were grouped into three subtypes based on their sixth-month symptomatology by Generalized Association Plots, designated as remitted (RM), persistent delusion/hallucination (PDH), and markedly blunting (MB) groups. These three subtypes presented with similar positive symptom profiles at recruitment, yet during follow-up, the PDH group tended to maintain the highest risk of having worse clinical symptomatology, social functioning, and neuropsychologic functioning, and the RM was the best outcome group. CONCLUSION: This three-subtype model provides a practical reference to predict medium-term outcomes by the subject's response to acute treatment and serves as a model to sort out part of the heterogeneous nature of schizophrenia that still should be examined by further psychopharmacological, neurobiological, and genetic studies.

High BRCA1 gene expression increases the risk of early distant metastasis in ER+ breast cancers.

Although the function of the BRCA1 gene has been extensively studied, the relationship between BRCA1 gene expression and tumor aggressiveness remains controversial in sporadic breast cancers. Because the BRCA1 protein is known to regulate estrogen signaling, we selected microarray data of ER+ breast cancers from the GEO public repository to resolve previous conflicting findings. The BRCA1 gene expression level in highly proliferative luminal B tumors was shown to be higher than that in luminal A tumors. Survival analysis using a cure model indicated that patients of early ER+ breast cancers with high BRCA1 expression developed rapid distant metastasis. In addition, the proliferation marker genes MKI67 and PCNA, which are characteristic of aggressive tumors, were also highly expressed in patients with high BRCA1 expression. The associations among high BRCA1 expression, high proliferation marker expression, and high risk of distant metastasis emerged in independent datasets, regardless of tamoxifen treatment. Tamoxifen therapy could improve the metastasis-free fraction of high BRCA1 expression patients. Our findings link BRCA1 expression with proliferation and possibly distant metastasis via the ER signaling pathway. We propose a testable hypothesis based on these consistent results and offer an interpretation for our reported associations.

Patient subgroups of schizophrenia based on the Positive and Negative Syndrome Scale: composition and transition between acute and subsided disease states.

The present study focuses on schizophrenia patient subgroups with specific symptom pattern using the Positive and Negative Syndrome Scale (PANSS). In this report, we intend to (1) provide a more appropriate analytic method for exploring the subgroups based on PANSS data, (2) validate identified subgroups with external variables, and (3) estimate probabilities of subgroup changes between 2 disease states. The analyzed data include 219 acute-state patients who had completed the PANSS within 1 week of index admission and 225 subsided-state patients who were living in the community and under family care. Regression extension of latent class analysis was performed. We found that acute schizophrenia can be classified into 4 subgroups--whole syndrome, whole syndrome without hostility, partial syndrome with negative symptoms, and partial syndrome with pure reality distortion--and that subsided schizophrenia can be classified into 3 subgroups--florid symptom, marked negative, and remitted. Patients of the whole syndrome, whole syndrome without hostility, partial syndrome with negative symptoms, and partial syndrome with pure reality distortion subgroups at the acute state were most likely to transit to the florid symptom (61%), florid symptom (48%), marked negative (42%), and remitted (56%) subgroups at the subsided state, respectively. Significant relationships of obtained subgroups with sociodemographic variables and neurocognitive variables were identified. These results of different subgroups will provide the background for facilitating current molecular, genetic, and neurobiological studies of schizophrenia.

Dimension reduction in survival regressions with censored data via an imputed spline approach.

Dimension reduction methods have been proposed for regression analysis with predictors of high dimension, but have not received much attention on the problems with censored data. In this article, we present an iterative imputed spline approach based on principal Hessian directions (PHD) for censored survival data in order to reduce the dimension of predictors without requiring a prespecified parametric model. Our proposal is to replace the right-censored survival time with its conditional expectation for adjusting the censoring effect by using the Kaplan-Meier estimator and an adaptive polynomial spline regression in the residual imputation. A sparse estimation strategy is incorporated in our approach to enhance the interpretation of variable selection. This approach can be implemented in not only PHD, but also other methods developed for estimating the central mean subspace. Simulation studies with right-censored data are conducted for the imputed spline approach to PHD (IS-PHD) in comparison with two methods of sliced inverse regression, minimum average variance estimation, and naive PHD in ignorance of censoring. The results demonstrate that the proposed IS-PHD method is particularly useful for survival time responses approximating symmetric or bending structures. Illustrative applications to two real data sets are also presented.

Statistical validation of endophenotypes using a surrogate endpoint analytic analogue.

Endophenotypes, which involve the same biological pathways as diseases but presumably are closer to the relevant gene actions than diagnostic phenotypes, have emerged as an important concept in the genetic studies of complex diseases. In this report, we develop a formal statistical methodology for validating endophenotypes. The proposed method was motivated by the conditioning strategy used for surrogate endpoints commonly seen in clinical research. We define an endophenotype to be "a trait for which a test of null hypothesis of no genetic heritability implies the corresponding null hypothesis based on the phenotype of interest". An index, the proportion of heritability explained, is used as an operational criterion of validation. Statistical inferences on this index are also developed. Usefulness of the proposed method is demonstrated through computer simulations and a study of assessing the Continuous Performance Test as an endophenotype of the schizophrenia spectrum.

Is Hyperuricemia, an Early-Onset Metabolic Disorder, Causally Associated with Cardiovascular Disease Events in Han Chinese?

BACKGROUND: Serum uric acid (SUA) has gradually been recognized as a potential risk factor for cardiovascular disease (CVD). However, whether the relationship is causal remains controversial. METHODS: We employed two methods to demonstrate the importance of SUA in CVD development. First, we examined the onset sequence of hyperuricemia in relation to five cardiometabolic (CM) diseases. Second, we conducted a Mendelian randomization (MR) study to causally infer the relationship between SUA and CVD. The information collected from the Cardiovascular Disease Risk Factors Two-Township Study (CVDFACTS) and Taiwan Biobank was used, respectively. RESULTS: The onset sequence study showed that hyperuricemia and hypo-alpha-lipoproteinemia (low HDL-C) have earlier ages of onset than other CM diseases. For the MR analysis, the high weighted genetic risk score (WGRS) group had a significantly increased cumulative lifetime risk of CVD compared with the low WGRS group (OR = 1.62, (1.17-2.23), P = 0.003). Sensitivity analysis using the WGRS derived from other populations' SUA-influential SNPs revealed similar results. CONCLUSIONS: We showed that hyperuricemia is an earlier-onset metabolic disorder than hypertension, hypertriglyceridemia, and diabetes mellitus, indicating that high SUA plays an upstream role in CM development. Moreover, our MR study results support the idea that hyperuricemia may play a causal role in CVD development. Further validation studies in more populations are needed.

Factors related to perceived needs of primary caregivers of patients with schizophrenia.

BACKGROUND/PURPOSE: Schizophrenia is a chronic mental illness, and sufferers are usually dependent on family, primary caregivers in particular. The present study was designed to assess the perceived needs of caregivers so that adequate services can be provided for them in the community. METHODS: A total of 177 primary caregivers were interviewed with the structured burden-and-need schedules to determine their perceived needs, and the related clinical and demographic factors. Fourteen perceived needs were identified and classified into different need clusters using the generalized association plots. A multiple regression of logistic model was adopted to explore the relationships between the related factors and perceived needs. RESULTS: Four clusters of perceived needs were identified, which included assistant patient care (77.6%), access to relevant information (66.1%), societal support (68.2%), and burden release (27.2%). These needs were significantly related to number of admissions, duration of illness, relationship between caregiver and patient, and education level of the caregiver. CONCLUSION: Four clusters of caregivers' perceived needs were identified and found to be related to psychopathologic and demographic factors. These data are of value in designing appropriate community psychiatric programs to improve the quality of care and enhance the capacity of primary caregivers to care for patients.

The Efficacy of 24-Month Metformin for Improving Menses, Hormones, and Metabolic Profiles in Polycystic Ovary Syndrome.

Context: The long-term effects of metformin in women with polycystic ovarian syndrome (PCOS) are inadequately studied. Objective: The effects of metformin on women with PCOS during 24 months with respect to menses, hormones, and metabolic profiles are assessed. Design: Prospective cohort. Setting: A reproductive endocrinology clinic in a university-affiliated medical center. Patients: One hundred nineteen women with PCOS, defined by the Rotterdam criteria, were enrolled. Intervention: Metformin was given daily for 24 months. Main Outcome Measures: The primary outcome was the proportion of patients with regular menstruation during treatment. Changes in anthropometric, hormonal, and metabolic parameters were also assessed. Analyses were performed using segmented regression analysis with a generalized estimating equation methodology. Outcomes are expressed as magnitude of change from the baseline. Results: Both overweight (OW) and normal-weight (NW) women with PCOS had increased menstrual frequency and decreased body mass index (BMI), testosterone, and luteinizing hormone levels in the first 6 months. Further stratification showed that NW women exhibiting elevated testosterone at baseline had the largest magnitude of improvement at 6 months [odds ratio (OR), 7.21; 95% confidence interval (CI), 2.35 to 22.17], whereas OW patients with normal testosterone were most likely to achieve normal menses at 12 months (OR, 0.63; 95% CI, 0.47 to 0.77). Conclusions: Metformin was associated with improvements in the menstrual cycle and most hormonal profiles in OW and NW women with PCOS during 24 months of treatment. Most parameters reached maximal response and steady-state after 6 months. Phenotypic differences in baseline BMI and testosterone level can be used as patient selection criteria or treatment prognostics.

Selection of DDX5 as a novel internal control for Q-RT-PCR from microarray data using a block bootstrap re-sampling scheme.

BACKGROUND: The development of microarrays permits us to monitor transcriptomes on a genome-wide scale. To validate microarray measurements, quantitative-real time-reverse transcription PCR (Q-RT-PCR) is one of the most robust and commonly used approaches. The new challenge in gene quantification analysis is how to explicitly incorporate statistical estimation in such studies. In the realm of statistical analysis, the various available methods of the probe level normalization for microarray analysis may result in distinctly different target selections and variation in the scores for the correlation between microarray and Q-RT-PCR. Moreover, it remains a major challenge to identify a proper internal control for Q-RT-PCR when confirming microarray measurements. RESULTS: Sixty-six Affymetrix microarray slides using lung adenocarcinoma tissue RNAs were analyzed by a statistical re-sampling method in order to detect genes with minimal variation in gene expression. By this approach, we identified DDX5 as a novel internal control for Q-RT-PCR. Twenty-three genes, which were differentially expressed between adjacent normal and tumor samples, were selected and analyzed using 24 paired lung adenocarcinoma samples by Q-RT-PCR using two internal controls, DDX5 and GAPDH. The percentage correlation between Q-RT-PCR and microarray were 70% and 48% by using DDX5 and GAPDH as internal controls, respectively. CONCLUSION: Together, these quantification strategies for Q-RT-PCR data processing procedure, which focused on minimal variation, ought to significantly facilitate internal control evaluation and selection for Q-RT-PCR when corroborating microarray data.

Using an Event-History with Risk-Free Model to Study the Genetics of Alcoholism.

Case-control genetic association studies typically ignore possible later disease onset in currently healthy subjects and assume that subjects with diseases equally contribute to the likelihood for inference, regardless of their onset age. Therefore, we used an event-history with risk-free model to simultaneously characterize alcoholism susceptibility and onset age in 65 independent non-Hispanic Caucasian males in the Collaborative Study on the Genetics of Alcoholism. Following data quality control, we analysed 22 single nucleotide polymorphisms (SNPs) on 12 candidate genes. The single-SNP analysis showed that the dominant minor allele of rs2134655 on DRD3 increases alcoholism susceptibility; the dominant minor allele of rs1439047 on NTRK2 delays the alcoholism onset age, but the additive minor allele of rs172677 on GRIN2B and the dominant minor allele of rs63319 on ALDH1A1 advance the alcoholism onset age; and the dominant minor allele of rs1079597 on DRD2 shortens the onset age range. Similarly, multiple-SNPs analysis revealed joint effects of rs2134655, rs172677 and rs1079597, with an adjustment for habitual smoking. This study provides a more comprehensive understanding of the genetics of alcoholism than previous case-control studies.

Symptom patterns and subgrouping of schizophrenic patients: significance of negative symptoms assessed on admission.

This study subgroups schizophrenic patients based on symptoms assessed on admission and examines the validity of the subgrouping using follow-up data and other clinical outcome variables. Schizophrenic patients (n=163) from consecutive admission received ratings on the positive and negative syndrome scale (PANSS) on admission and during a 1-year follow-up course. An exploratory graphic analysis on the admission PANSS derived four symptom dimensions: negative symptoms, disorganized thought, hostility/excitement and delusions/hallucinations. This yielded two subgroups of patients on admission, a group with marked negative (GWNEG) and a group without marked negative (GONEG) symptoms. Compared with the GONEG, the GWNEG had a poorer recovery rate, more impairment in attention, a slower response of the delusion/hallucination symptoms to neuroleptic treatment and a longer duration of index hospitalization. At a one-year follow-up, the GWNEG assessed on admission had persistently higher scores on the negative symptom and disorganized thought syndromes, less relapse rate, a shorter duration on job, as well as worse social functioning than the GONEG. Thus, the GONEG might comprise patients having a pure paranoid syndrome with quick and better treatment response, while the GWNEG comprises patients with the blunt-disorganization syndrome having a poorer outcome.

HLA typing associated with hepatitis B E antigen seroconversion in children with chronic hepatitis B virus infection: a long-term prospective sibling cohort study in Taiwan.

OBJECTIVE: To conduct a prospective cohort study to clarify the relationship between human leukocyte antigen (HLA) polymorphisms and the seroconversion of hepatitis B e antigen (HBeAg). STUDY DESIGN: In the prospective cohort study, 81 HBeAg-positive children with chronic hepatitis B virus (HBV) infection from 40 unrelated families were recruited and followed-up regularly for a mean period of 17.70 +/- 3.23 years. The association between HLA antigen and the age at HBeAg seroconversion was analyzed using Cox regression model with shared frailties under left truncation and right censorship. RESULTS: HLA-B61 and HLA-DQB1*0503 antigens predicted a higher HBeAg seroconversion rate (relative incidence = 6.17 and 3.22, P = .024 and .017, respectively). Within-family frailty in our sibling cohort study demonstrated a negligible or a low degree of within-family correlation with spontaneous HBeAg seroconversion in each HLA antigen. CONCLUSIONS: HLA class I antigen B61 and class II antigen DQB1*0503 are associated with earlier HBeAg seroconversion in Taiwanese children with chronic HBV infection.

Validation of a heteroscedastic hazards regression model.

A Cox-type regression model accommodating heteroscedasticity, with a power factor of the baseline cumulative hazard, is investigated for analyzing data with crossing hazards behavior. Since the approach of partial likelihood cannot eliminate the baseline hazard, an overidentified estimating equation (OEE) approach is introduced in the estimation procedure. It by-product, a model checking statistic, is presented to test for the overall adequacy of the heteroscedastic model. Further, under the heteroscedastic model setting, we propose two statistics to test the proportional hazards assumption. Implementation of this model is illustrated in a data analysis of a cancer clinical trial.