The number and timing of time -dependent covariate measurements for the Cox regression model

The problem of analyzing data with updated measurements in the time-dependent proportional hazards model arises frequently in practice. One available option is to reduce the number of intervals (or updated measurements) to be included in the Cox regression model. We empirically investigated the bias of the estimator of the time-dependent covariate while varying the effect of failure rate, sample size, true values of the parameters and the number of intervals. We also evaluated how often a time-dependent covariate needs to be collected and assessed the effect of sample size and failure rate on the power of testing a time-dependent effect.^ A time-dependent proportional hazards model with two binary covariates was considered. The time axis was partitioned into k intervals. The baseline hazard was assumed to be 1 so that the failure times were exponentially distributed in the ith interval. A type II censoring model was adopted to characterize the failure rate. The factors of interest were sample size (500, 1000), type II censoring with failure rates of 0.05, 0.10, and 0.20, and three values for each of the non-time-dependent and time-dependent covariates (1/4,1/2,3/4).^ The mean of the bias of the estimator of the coefficient of the time-dependent covariate decreased as sample size and number of intervals increased whereas the mean of the bias increased as failure rate and true values of the covariates increased. The mean of the bias of the estimator of the coefficient was smallest when all of the updated measurements were used in the model compared with two models that used selected measurements of the time-dependent covariate. For the model that included all the measurements, the coverage rates of the estimator of the coefficient of the time-dependent covariate was in most cases 90% or more except when the failure rate was high (0.20). The power associated with testing a time-dependent effect was highest when all of the measurements of the time-dependent covariate were used. An example from the Systolic Hypertension in the Elderly Program Cooperative Research Group is presented. ^

Factors determining disease duration in Alzheimer's disease:a postmortem study of 103 cases using the Kaplan-Meier estimator and Cox regression

Factors associated with duration of dementia in a consecutive series of 103 Alzheimer's disease (AD) cases were studied using the Kaplan-Meier estimator and Cox regression analysis (proportional hazard model). Mean disease duration was 7.1 years (range: 6 weeks-30 years, standard deviation = 5.18); 25% of cases died within four years, 50% within 6.9 years, and 75% within 10 years. Familial AD cases (FAD) had a longer duration than sporadic cases (SAD), especially cases linked to presenilin (PSEN) genes. No significant differences in duration were associated with age, sex, or apolipoprotein E (Apo E) genotype. Duration was reduced in cases with arterial hypertension. Cox regression analysis suggested longer duration was associated with an earlier disease onset and increased senile plaque (SP) and neurofibrillary tangle (NFT) pathology in the orbital gyrus (OrG), CA1 sector of the hippocampus, and nucleus basalis of Meynert (NBM). The data suggest shorter disease duration in SAD and in cases with hypertensive comorbidity. In addition, degree of neuropathology did not influence survival, but spread of SP/NFT pathology into the frontal lobe, hippocampus, and basal forebrain was associated with longer disease duration. © 2014 R. A. Armstrong.

A diszkrét kiválasztási modell becslése Cox-regresszióval (Parameter estimation for multinomial discrete choice model using Cox-regression)

Considering the so-called "multinomial discrete choice" model the focus of this paper is on the estimation problem of the parameters. Especially, the basic question arises how to carry out the point and interval estimation of the parameters when the model is mixed i.e. includes both individual and choice-specific explanatory variables while a standard MDC computer program is not available for use. The basic idea behind the solution is the use of the Cox-proportional hazards method of survival analysis which is available in any standard statistical package and provided a data structure satisfying certain special requirements it yields the MDC solutions desired. The paper describes the features of the data set to be analysed.

Sobrevida e fatores de risco para mortalidade neonatal em uma coorte de nascidos vivos de muito baixo peso ao nascer, na Região Sul do Município de São Paulo, Brasil

Estudos populacionais sobre mortalidade neonatal de nascimentos de muito baixo peso ao nascer contribuem para identificar sua complexa rede de fatores de risco. Foi estudada uma coorte de 213 recém-nascidos com peso inferior a 1.500g (112 óbitos neonatais e 101 sobreviventes) na Região Sul do Município de São Paulo, Brasil, em 2000/2001. Foram realizadas entrevistas domiciliares e obtidos dados de prontuários hospitalares. Foi realizada análise de sobrevida e empregada regressão múltipla de Cox. A elevada mortalidade na sala de parto, no primeiro dia de vida e ausência de sobreviventes < 700g dos nascimentos < 1.000g e com menos de 28 semanas sugere que condutas mais ativas destinam-se a nascituros de maior viabilidade. Mães residentes em favela, com história anterior de cesárea e aborto provocado, adolescentes, com sangramento vaginal e ausência de pré-natal aumentaram o risco de óbito neonatal. Partos cesarianos e internação em berçários mostraram efeito protetor. O peso ao nascer abaixo de 1.000g e Apgar menor que 7 foram risco. A elevada mortalidade está associada às condições de vida, características maternas e dos nascimentos e condições assistenciais. A melhoria da atenção pré-natal e ao recém-nascido pode atuar na redução da mortalidade.

Prognostic Factors in Patients with Malignant Salivary Gland Neoplasms in a Brazilian Population

Due to the difficulty of follow-up for long periods, information about the survival rates of malignant salivary gland tumors is deficient in the global scientific literature. This study was aimed at investigating the epidemiological profile and prognostic factors that might affect survival in patients with primary malignant salivary gland tumors in Brazil. Patients were investigated regarding histopathological subtypes, age, gender, anatomic localization, smoking and alcohol intake, tumor size, clinical stage, histological grade, recurrence, metastasis, and treatment on clinicopathological outcomes. Survival curves were generated using the Kaplan-Meier method, and both univariate and multivariate analyses were performed using the log rank test and Cox regression, respectively. A total of 63 cases were analyzed, females beingslightly predominant (50.8%), with ages ranging from 13 to 87 years. The most common diagnosis was adenoid cystic carcinoma and the most affected anatomical location was the parotid. Tumors were predominantly classified as stage I and high-grade at the diagnosis. The 5- and 10-year overall survival rates were 84.6% and 74.7%, respectively. Disease-free survival (DFS) rates were 71.6% (5 years) and 56.6% (10 years). Univariate analysis showed significant effects of tumor size and clinical stage on the DFS (P < 0.0001 for both), and Cox regression analysis confirmed clinical stage as an independent prognostic factor (P = 0.035). Our results highlight the relevance of clinical stage as an independent prognostic parameter for malignant salivary gland tumors.

Prognostic Significance of Myocardial Fibrosis Quantification by Histopathology and Magnetic Resonance Imaging in Patients With Severe Aortic Valve Disease

Objectives We sought to determine whether the quantitative assessment of myocardial fibrosis (MF), either by histopathology or by contrast-enhanced magnetic resonance imaging (ce-MRI), could help predict long-term survival after aortic valve replacement. Background Severe aortic valve disease is characterized by progressive accumulation of interstitial MF. Methods Fifty-four patients scheduled to undergo aortic valve replacement were examined by ce-MRI. Delayed-enhanced images were used for the quantitative assessment of MF. In addition, interstitial MF was quantified by histological analysis of myocardial samples obtained during open-heart surgery and stained with picrosirius red. The ce-MRI study was repeated 27 +/- 22 months after surgery to assess left ventricular functional improvement, and all patients were followed for 52 +/- 17 months to evaluate long-term survival. Results There was a good correlation between the amount of MF measured by histopathology and by ce-MRI (r = 0.69, p < 0.001). In addition, the amount of MF demonstrated a significant inverse correlation with the degree of left ventricular functional improvement after surgery (r = -0.42, p = 0.04 for histopathology; r = -0.47, p = 0.02 for ce-MRI). Kaplan-Meier analyses revealed that higher degrees of MF accumulation were associated with worse long-term survival (chi-square = 6.32, p = 0.01 for histopathology; chi-square = 5.85, p = 0.02 for ce-MRI). On multivariate Cox regression analyses, patient age and the amount of MF were found to be independent predictors of all-cause mortality. Conclusions The amount of MF, either by histopathology or by ce-MRI, is associated with the degree of left ventricular functional improvement and all-cause mortality late after aortic valve replacement in patients with severe aortic valve disease. (J Am Coll Cardiol 2010; 56: 278-87) (c) 2010 by the American College of Cardiology Foundation

Usefulness of Proteinuria as a Prognostic Marker of Mortality and Cardiovascular Events Among Patients Undergoing Percutaneous Coronary Intervention (Data from the Evaluation of Oral Xemilofiban in Controlling Thrombotic Events [EXCITE] Trial)

Proteinuria was associated with cardiovascular events and mortality in community-based cohorts. The association of proteinuria with mortality and cardiovascular events in patients undergoing percutaneous coronary intervention (PCI) was unknown. The association of urinary dipstick proteinuria with mortality and cardiovascular events (composite of death, myocardial infarction, or nonhemorrhagic stroke) in 5,835 subjects of the EXCITE trial was evaluated. Dipstick urinalysis was performed before PCI, and proteinuria was defined as trace or greater. Subjects were followed up for 210 days/7 months after enrollment for the occurrence of events. Multivariate Cox regression analysis evaluated the independent association of proteinuria with each outcome. Mean age was 59 years, 21% were women, 18% had diabetes mellitus, and mean estimated glomerular filtration rate was 90 ml/min/1.73 m(2). Proteinuria was present in 750 patients (13%). During follow-up, 22 subjects (2.9%) with proteinuria and 54 subjects (1.1%) without proteinuria died (adjusted hazard ratio 2.83, 95% confidence interval [CI] 1.65 to 4.84, p <0.001). The severity of proteinuria attenuated the strength of the association with mortality after PCI (low-grade proteinuria, hazard ratio 2.67, 95% CI 1.50 to 4.75; high-grade proteinuria, hazard ratio 3.76, 95% CI 1.24 to 11.37). No significant association was present for cardiovascular events during the relatively short follow-up, but high-grade proteinuria tended toward increased risk of cardiovascular events (hazard ratio 1.45, 95% CI 0.81 to 2.61). In conclusion, proteinuria was strongly and independently associated with mortality in patients undergoing PCI. These data suggest that such a relatively simple and clinically easy to use tool as urinary dipstick may be useful to identify and treat patients at high risk of mortality at the time of PCI. (C) 2008 Elsevier Inc. All rights reserved. (Am J Cardiol 2008;102:1151-1155)

Relation of Aortic Valve Calcium Detected by Cardiac Computed Tomography to All-Cause Mortality

Aortic valve calcium (AVC) can be quantified on the same computed tomographic scan as coronary artery calcium (CAC). Although CAC is an established predictor of cardiovascular events, limited evidence is available for an independent predictive value for AVC. We studied a cohort of 8,401 asymptomatic subjects (mean age 53 10 years, 69% men), who were free of known coronary heart disease and were undergoing electron beam computed tomography for assessment of subclinical atherosclerosis. The patients were followed for a median of 5 years (range 1 to 7) for the occurrence of mortality from any cause. Multivariate Cox regression models were developed to predict all-cause mortality according to the presence of AVC. A total of 517 patients (6%) had AVC on electron beam computed tomography. During follow-up, 124 patients died (1.5%), for an overall survival rate of 96.1% and 98.7% for those with and without AVC, respectively (hazard ratio 3.39, 95% confidence interval 2.09 to 5.49). After adjustment for age, gender, hypertension, dyslipidemia, diabetes mellitus, smoking, and a family history of premature coronary heart disease, AVC remained a significant predictor of mortality (hazard ratio 1.82, 95% confidence interval 1.11 to 2.98). Likelihood ratio chi-square statistics demonstrated that the addition of AVC contributed significantly to the prediction of mortality in a model adjusted for traditional risk factors (chi-square = 5.03, p = 0.03) as well as traditional risk factors plus the presence of CAC (chi-square = 3.58, p = 0.05). In conclusion, AVC was associated with increased all-cause mortality, independent of the traditional risk factors and the presence of CAC. (C) 2010 Published by Elsevier Inc. (Am J Cardiol 2010;106:1787-1791)

FGF-23 as a Predictor of Renal Outcome in Diabetic Nephropathy

Background and objectives Fibroblast growth factor 23 (FGF-23) has emerged as a new factor in mineral metabolism in chronic kidney disease (CKD). An important regulator of phosphorus homeostasis, FGF-23 has been shown to independently predict CKD progression in nondiabetic renal disease. We analyzed the relation between FGF-23 and renal outcome in diabetic nephropathy (DN). Design, setting, participants, & measurements DN patients participating in a clinical trial (enalapril+placebo versus enalapril+losartan) had baseline data collected and were followed until June 2009 or until the primary outcome was reached. Four patients were lost to follow-up. The composite primary outcome was defined as death, doubling of serum creatinine, and/or dialysis need. Results At baseline, serum FGF-23 showed a significant association with serum creatinine, intact parathyroid hormone, proteirturia, urinary fractional excretion of phosphate, male sex, and race. Interestingly, FGF-23 was not related to calcium, phosphorus, 25OH-vitamin D, or 24-hour urinary phosphorus. Mean follow-up time was 30.7 +/- 10 months. Cox regression showed that FGF-23 was an independent predictor of the primary outcome, even after adjustment for creatinine clearance and intact parathyroid hormone (10 pg/ml FGF-23 increase = hazard ratio, 1.09; 95% CI, 1.01 to 1.16, P = 0.02). Finally, Kaplan-Meier analysis showed a significantly higher risk of the primary outcome in patients with FGF-23 values of >70 pg/ml. Conclusions FGF-23 is a significant independent predictor of renal outcome in patients with macroalbuminuric DN. Further studies should clarify whether this relation is causal and whether FGF-23 should be a new therapeutic target for CKD prevention. Clin J Am Soc Nephrol 6: 241-247, 2011. doi: 10.2215/CJN.04250510

Antimalarial Treatment May Have a Time-Dependent Effect on Lupus Survival

Objective. To evaluate the beneficial effect of antimalarial treatment on lupus survival in a large, multiethnic, international longitudinal inception cohort. Methods. Socioeconomic and demographic characteristics, clinical manifestations, classification criteria, laboratory findings, and treatment variables were examined in patients with systemic lupus erythematosus (SLE) from the Grupo Latino Americano de Estudio del Lupus Eritematoso (GLADEL) cohort. The diagnosis of SLE, according to the American College of Rheumatology criteria, was assessed within 2 years of cohort entry. Cause of death was classified as active disease, infection, cardiovascular complications, thrombosis, malignancy, or other cause. Patients were subdivided by antimalarial use, grouped according to those who had received antimalarial drugs for at least 6 consecutive months (user) and those who had received antimalarial drugs for <6 consecutive months or who had never received antimalarial drugs (nonuser). Results. Of the 1,480 patients included in the GLADEL cohort, 1,141 (77%) were considered antimalarial users, with a mean duration of drug exposure of 48.5 months (range 6-98 months). Death occurred in 89 patients (6.0%). A lower mortality rate was observed in antimalarial users compared with nonusers (4.4% versus 11.5%; P < 0.001). Seventy patients (6.1%) had received antimalarial drugs for 6-11 months, 146 (12.8%) for 1-2 years, and 925 (81.1%) for >2 years. Mortality rates among users by duration of antimalarial treatment (per 1,000 person-months of followup) were 3.85 (95% confidence interval [95% CI] 1.41-8.37), 2.7 (95% CI 1.41-4.76), and 0.54 (95% CI 0.37-0.77), respectively, while for nonusers, the mortality rate was 3.07 (95% CI 2.18-4.20) (P for trend < 0.001). After adjustment for potential confounders in a Cox regression model, antimalarial use was associated with a 38% reduction in the mortality rate (hazard ratio 0.62, 95% CI 0.39-0.99). Conclusion. Antimalarial drugs were shown to have a protective effect, possibly in a time-dependent manner, on SLE survival. These results suggest that the use of antimalarial treatment should be recommended for patients with lupus.

Outcome of adults with asymptomatic severe aortic stenosis

Objectives: To evaluate clinical and echocardiographic variables that could be used to predict outcomes in patients with asymptomatic severe aortic valve stenosis. Management of asymptomatic severe aortic stenosis is controversial. Because prophylactic surgery may be protective, independent predictors of events that could justify early surgery have been sought. Methods: Outpatients (n= 133; mean [+/- SD] age, 66.2 +/- 13.6 years) with isolated severe asymptomatic aortic stenosis but normal left ventricular function and no previous myocardial infarction were followed up prospectively at a tertiary care hospital. Interventions: We use a ""wait-for-events"" strategy. Clinical and echocardiographic variables were analyzed. Results: Nineteen patients developed angina; 40, dyspnea; 5, syncope; and 7, sudden death during a mean follow-up period of 3.30 +/- 1.87 years. Event-free survival was 90.2 +/- 2.6% at 1 year, 73.4 +/-.9% at 2 years, 70.7 +/- 4.3% at 3 years, 57.8 +/- 4.7% at 4 years, 40.3 +/- 5.0% at 5 years, and 33.3 +/- 5.2% at 6 years. The mean follow-up period until sudden death (1.32 +/- 1.11 years) was shorter than that for dyspnea (2.44 +/- 1.84 years), syncope (2.87 +/- 1.26 years) and angina (3.03 +/- 1.68 years). Cox regression analysis disclosed only reduced but within normal limits ejection fraction as independent predictor of total events. Conclusions: Management on ""wait-for-events"" strategy is generally safe. Progressive left ventricular ejection fraction reduction even within normal limits identified patients at high risk for events in whom valve replacement surgery should be considered. (c) 2007 Elsevier Ireland Ltd. All rights reserved.

Ten-Year Follow-Up Survival of the Medicine, Angioplasty, or Surgery Study (MASS II) A Randomized Controlled Clinical Trial of 3 Therapeutic Strategies for Multivessel Coronary Artery Disease

Background-This study compared the 10-year follow-up of percutaneous coronary intervention (PCI), coronary artery surgery (CABG), and medical treatment (MT) in patients with multivessel coronary artery disease, stable angina, and preserved ventricular function. Methods and Results-The primary end points were overall mortality, Q-wave myocardial infarction, or refractory angina that required revascularization. All data were analyzed according to the intention-to-treat principle. At a single institution, 611 patients were randomly assigned to CABG (n = 203), PCI (n = 205), or MT (n = 203). The 10-year survival rates were 74.9% with CABG, 75.1% with PCI, and 69% with MT (P = 0.089). The 10-year rates of myocardial infarction were 10.3% with CABG, 13.3% with PCI, and 20.7% with MT (P < 0.010). The 10-year rates of additional revascularizations were 7.4% with CABG, 41.9% with PCI, and 39.4% with MT (P < 0.001). Relative to the composite end point, Cox regression analysis showed a higher incidence of primary events in MT than in CABG (hazard ratio 2.35, 95% confidence interval 1.78 to 3.11) and in PCI than in CABG (hazard ratio 1.85, 95% confidence interval 1.39 to 2.47). Furthermore, 10-year rates of freedom from angina were 64% with CABG, 59% with PCI, and 43% with MT (P < 0.001). Conclusions-Compared with CABG, MT was associated with a significantly higher incidence of subsequent myocardial infarction, a higher rate of additional revascularization, a higher incidence of cardiac death, and consequently a 2.29-fold increased risk of combined events. PCI was associated with an increased need for further revascularization, a higher incidence of myocardial infarction, and a 1.46-fold increased risk of combined events compared with CABG. Additionally, CABG was better than MT at eliminating anginal symptoms.

beta-Blocker Therapy and Mortality of Patients With Chagas Cardiomyopathy A Subanalysis of the REMADHE Prospective Trial

Background-Peculiar aspects of Chagas cardiomyopathy raise concerns about efficacy and safety of sympathetic blockade. We studied the influence of beta-blockers in patients with Chagas cardiomyopathy. Methods and Results-We examined REMADHE trial and grouped patients according to etiology (Chagas versus non-Chagas) and beta-blocker therapy. Primary end point was all-cause mortality or heart transplantation. Altogether 456 patients were studied; 27 (5.9%) were submitted to heart transplantation and 202 (44.3%) died. Chagas etiology was present in 68 (14.9%) patients; they had lower body mass index (24.1+/-4.1 versus 26.3+/-5.1, P=0.001), smaller end-diastolic left ventricle diameter (6.7+/-1.0 mm versus 7.0+/-0.9 mm, P=0.001), smaller proportion of beta-blocker therapy (35.8% versus 68%, P<0.001), and higher proportion of spironolactone therapy (74.6% versus 57.8%, P=0.003). Twenty-four (35.8%) patients with Chagas disease were under beta-blocker therapy and had lower serum sodium (136.6+/-3.1 versus 138.4+/-3.1 mEqs, P=0.05) and lower body mass index (22.5+/-3.3 versus 24.9+/-4.3, P=0.03) compared with those who received beta-blockers. Survival was lower in patients with Chagas heart disease as compared with other etiologies. When only patients under beta-blockers were considered, the survival of patients with Chagas disease was similar to that of other etiologies. The survival of patients with beta-blockers was higher than that of patients without beta-blockers. In Cox regression model, left ventricle end-diastolic diameter (hazard ratio, 1.78; CI, 1.15 to 2.76; P=0.009) and beta-blockers (hazard ratio, 0.37; CI, 0.14 to 0.97; P=0.044) were associated with better survival. Conclusions-Our study suggests that beta-blockers may have beneficial effects on survival of patients with heart failure and Chagas heart disease and warrants further investigation in a prospective, randomized trial.

Red Blood Cell Transfusions are Independently Associated with Intra-Hospital Mortality in Very Low Birth Weight Preterm Infants

Objective To test the hypothesis that red blood cell (RBC) transfusions in preterm infants are associated with increased intra-hospital mortality. Study design Variables associated with death were studied with Cox regression analysis in a prospective cohort of preterm infants with birth weight <1500 g in the Brazilian Network on Neonatal Research. Intra-hospital death and death after 28 days of life were analyzed as dependent variables. Independent variables were infant demographic and clinical characteristics and RBC transfusions. Results Of 1077 infants, 574 (53.3%) received at least one RBC transfusion during the hospital stay. The mean number of transfusions per infant was 3.3 +/- 3.4, with 2.1 +/- 2.1 in the first 28 days of life. Intra-hospital death occurred in 299 neonates (27.8%), and 60 infants (5.6%) died after 28 days of life. After adjusting for confounders, the relative risk of death during hospital stay was 1.49 in infants who received at least one RBC transfusion in the first 28 days of life, compared with infants who did not receive a transfusion. The risk of death after 28 days of life was 1.89 times higher in infants who received more than two RBC transfusions during their hospital stay, compared with infants who received one or two transfusions. Conclusion Transfusion was associated with increased death, and transfusion guidelines should consider risks and benefits of transfusion. (J Pediatr 2011; 159: 371-6).

Higher Serum sTNFR1 Level Predicts Conversion from Mild Cognitive Impairment to Alzheimer`s Disease

The activation of inflammatory cascades has been consistently demonstrated in the pathophysiology of Alzheimer`s disease (AD). Among several putative neuroinflammatory mechanisms, the tumor necrosis factor alpha (TNF-alpha) signaling system has a central role in this process. Recent evidence indicates that the abnormal production of inflammatory factors may accompany the progression from mild cognitive impairment (MCI) to dementia. We aimed to examine serum levels of TNF-alpha and its soluble receptors (sTNFR1 and sTNFR2) in patients with MCI and AD as compared to cognitively unimpaired elderly subjects. We further aimed to investigate whether abnormal levels of these cytokines predict the progression from MCI to AD upon follow-up. We utilized cross-sectional determination of serum levels of TNF-alpha, sTNFR1, and sTNFR2 (ELISA method) in a test group comprising 167 older adults (31 AD, 72 MCI, and 64 healthy controls), and longitudinal reassessment of clinical status after 18.9 +/- 10.0 months. At baseline, there were no statistically significant differences in serum TNF-alpha, sTNFR1, and sTNFR2 between patients with MCI and AD as compared to controls. Nevertheless, patients with MCI who progressed to AD had significantly higher serum sTNFR1 levels as opposed to patients who retained the diagnosis of MCI upon follow-up (p = 0.03). Cox regression analysis showed that high serum sTNFR1 levels predicted the conversion from MCI to AD (p = 0.003), whereas no significant differences were found with respect to serum levels of TNF-alpha and sTNFR2. Abnormal activation of TNF-alpha signaling system, represented by increased expression of sTNFR1, is associated with a higher risk of progression from MCI to AD.