Hazard Ratios For Association Research Articles

Medicine utilization studies in Australian individual-level dispensing data: A blinded, multi-center replicated analysis.

Medicine dispensing data require extensive preparation when used for research and decisions during this process may lead to results that do not replicate between independent studies. We conducted an experiment to examine the impact of these decisions on results of a study measuring discontinuation, intensification, and switching in a cohort of patients initiating metformin. Four Australian sites independently developed a HARmonized Protocol template to Enhance Reproducibility (HARPER) protocol and executed their analyses using the Australian Pharmaceutical Benefits Scheme 10% sample dataset. Each site calculated cohort size and demographics and measured treatment events including discontinuation, switch to another diabetes medicine, and intensification (addition of another diabetes medicine). Time to event and hazard ratios for associations between cohort characteristics and each event were also calculated. Concordance was assessed by measuring deviations from the calculated median of each value across the sites. Good agreement was found across sites for the number of initiators (median: 53 127, range: 51 848-55 273), gender (56.9% female, range: 56.8%-57.1%) and age group. Each site employed different methods for estimating days supply and used different operational definitions for the treatment events. Consequently, poor agreement was found for incidence of discontinuation (median 55%, range: 34%-67%), switching (median 3.5%, range: 1%-7%), intensification (median 8%, range: 5%-12%), time to event estimates and hazard ratios. Differences in analytical decisions when deriving exposure from dispensing data affect replicability. Detailed analytical protocols, such as HARPER, are critical for transparency of operational definitions and interpretations of key study parameters.

Abstract 785: Is genetic ancestry associated with mortality risk after accounting for neighborhood socioeconomic status? A study of African American men

Abstract Background Genetic ancestry, a correlate of self-identified race/ethnicity, could be associated with mortality risk, allowing for more targeted treatments for cancer and other chronic diseases. However, associations may be confounded by neighborhood socioeconomic status (nSES). We investigated whether African ancestry and nSES were independently associated with all-cause mortality in a cohort of African American trial participants. Methods We constructed a cohort of American-born participants of African descent from the Prostate, Lung, Colorectal and Ovarian screening trial. Participants were enrolled from 1993-1998 and followed through 2019 or until death, whichever came first. Genetic ancestry (% African) was estimated using a principal components analysis of genome-wide SNP data and validated against European, African, and Asian panels. A nSES index was estimated using principal component analysis of census tract measures of income, housing, and employment. Census tract was assessed in 2012 for residences of living participants. Last known address was assigned to participants who died prior to 2012. We assigned 2000 census data to those who died prior to 2010, and assigned 2006-2010 American Community Survey data to those who were alive after 2010. Hazard ratios for associations between these variables and all-cause mortality were estimated using sequentially adjusted Cox proportional hazards models with sociodemographic, behavioral, and neighborhood contextual covariates. We tested whether mortality associations with ancestry varied by nSES using multiplicative interaction terms. All tests were two-sided with alpha=0.05. Results Among 2239 African American participants, there were 901 deaths over 40,767 person-years of follow-up. Mean baseline age was 61.7, 45% were men, and the mean percentage of African ancestry was 74.3. In unadjusted models, a 10% increase in African ancestry was associated with a 7% higher rate of all-cause mortality (HR: 1.07, 95% CI: 1.02, 1.12), but this attenuated to the null after adjusting for covariates (aHR: 1.01, 95% CI: 0.96, 1.06). Participants in nSES Q5 had 45% lower rate of mortality compared to those in Q1 in unadjusted models (HR: 0.55, 95% CI: 0.44, 0.68, Ptrend<.0001); similarly, this was attenuated after adjusting for covariates (aHR: 0.74, 95% CI: 0.57, 0.98, Ptrend=0.017). There was no evidence of statistical interaction between ancestry and nSES. Discussion Our findings suggest that nSES is a stronger independent correlate of mortality than genetic ancestry, which was only weakly associated with mortality after accounting for other factors. Larger studies with wider variability in African ancestry and geographic context are needed to confirm these findings. Citation Format: Hari S. Iyer, Scarlett L. Gomez, Iona Cheng, Timothy R. Rebbeck. Is genetic ancestry associated with mortality risk after accounting for neighborhood socioeconomic status? A study of African American men [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 785.

Efficacy of rapid decongestion strategy in patients hospitalized for acute heart failure

Abstract Background Clinical congestion is the most dominant feature in patients with acute decompensated heart failure (HF). However, uncertainty exists due to the permutations and combinations of congestion status and decongestion strategies. We aimed to investigate the impact of congestion status and its improvement on 1-year mortality. Methods We prospectively included 453 consecutive patients hospitalized for acute decompensated HF between July 2015 and March 2017. Congestion was evaluated using the congestion score which consists of signs and symptoms including dyspnea, fatigue, orthopnea, jugular vein distension, rales, and edema. This score ranges from 0 to 3 for each value, and calculated by summing each score. We assessed this score at admission, day 3, and discharge. Results The median age was 81 (interquartile range 75–87) years, and 54.1% of the subjects were male. The 1-year mortality rate was 22.7%. The congestion scores at admission, day 3, and discharge were 10.7±3.9, 3.4±3.5, and 0.3±0.8, respectively. The rate of improvement during the first 3 days was 78% and 46.6% had residual congestion, defined as scores at day 3 ≥3. The score at day 3 and improvement rate during the first 3 days were related to 1-year all-cause mortality (P<0.001). We examined combined predictive values by calculating multivariable-adjusted hazard ratios for associations of residual congestion and improvement rate during the first 3 days (cut-off value 64%); and prognostic variables identified by univariate Cox regression model (age, body mass index, systolic blood pressure, potassium level, albumin level, the prevalence of anemia and hypertension, left ventricular ejection fraction, ischemic etiology, previous HF hospitalization). Residual congestion and lesser improvement (% improvement <64%) were associated with higher relative risk of 1-year all-cause mortality than residual congestion and higher improvement (% improvement ≥64%) [adjusted hazard ratio (aHR) 2.33, 95% CI 1.11–4.91, P=0.025], or resolved congestion (aHR 2.17, 95% CI 1.30–3.63, P=0.003). Similar analysis revealed significant relationship of the score at day 3 and the rate of improvement from admission to day 3 to cardiovascular mortality. Combined predictive values of residual congestion and lesser improvement with adjustment for prognostic variables identified by univariate Cox regression model (age, body mass index, systolic blood pressure, potassium level, albumin level, the prevalence of anemia, reduced eGFR, and hypertension, left ventricular ejection fraction, ischemic etiology, previous HF hospitalization) were higher than those with residual congestion and higher improvement (aHR 3.04, 95% CI 1.15–8.03, P=0.025), or resolved congestion (aHR 3.17, 95% CI 1.65–6.11, P<0.001). Conclusions This study suggested that rapid decongestion therapy after hospital admission could be prerequisite to improve 1-year mortality in acute decompensated HF. Funding Acknowledgement Type of funding source: None

Use of Flutemetamol F 18–Labeled Positron Emission Tomography and Other Biomarkers to Assess Risk of Clinical Progression in Patients With Amnestic Mild Cognitive Impairment

ImportancePatients with amnestic mild cognitive impairment (aMCI) may progress to clinical Alzheimer disease (AD), remain stable, or revert to normal. Earlier progression to AD among patients who were β-amyloid positive vs those who were β-amyloid negative has been previously observed. Current research now accepts that a combination of biomarkers could provide greater refinement in the assessment of risk for clinical progression.ObjectiveTo evaluate the ability of flutemetamol F 18 and other biomarkers to assess the risk of progression from aMCI to probable AD.Design, Setting, and ParticipantsIn this multicenter cohort study, from November 11, 2009, to January 16, 2014, patients with aMCI underwent positron emission tomography (PET) at baseline followed by local clinical assessments every 6 months for up to 3 years. Patients with aMCI (365 screened; 232 were eligible) were recruited from 28 clinical centers in Europe and the United States. Physicians remained strictly blinded to the results of PET, and the standard of truth was an independent clinical adjudication committee that confirmed or refuted local assessments. Flutemetamol F 18–labeled PET scans were read centrally as either negative or positive by 5 blinded readers with no knowledge of clinical status. Statistical analysis was conducted from February 19, 2014, to January 26, 2018.InterventionsFlutemetamol F 18–labeled PET at baseline followed by up to 6 clinical visits every 6 months, as well as magnetic resonance imaging and multiple cognitive measures.Main Outcomes and MeasuresTime from PET to probable AD or last follow-up was plotted as a Kaplan-Meier survival curve; PET scan results, age, hippocampal volume, and aMCI stage were entered into Cox proportional hazards logistic regression analyses to identify variables associated with progression to probable AD.ResultsOf 232 patients with aMCI (118 women and 114 men; mean [SD] age, 71.1 [8.6] years), 98 (42.2%) had positive results detected on PET scan. By 36 months, the rates of progression to probable AD were 36.2% overall (81 of 224 patients), 53.6% (52 of 97) for patients with positive results detected on PET scan, and 22.8% (29 of 127) for patients with negative results detected on PET scan. Hazard ratios for association with progression were 2.51 (95% CI, 1.57-3.99; P < .001) for a positive β-amyloid scan alone (primary outcome measure), 5.60 (95% CI, 3.14-9.98; P < .001) with additional low hippocampal volume, and 8.45 (95% CI, 4.40-16.24; P < .001) when poorer cognitive status was added to the model.Conclusions and RelevanceA combination of positive results of flutemetamol F 18–labeled PET, low hippocampal volume, and cognitive status corresponded with a high probability of risk of progression from aMCI to probable AD within 36 months.

Abstract A60: Use of statin medications and reduced risk of esophageal adenocarcinoma in persons with Barrett's esophagus

Abstract Background: The incidence of esophageal adenocarcinoma (EA) has increased dramatically in many developed countries over the past 30 years, with little improvement in its rapid mortality. Persons with the metaplastic condition, Barrett's esophagus (BE), experience increased risk of EA; they are typically followed in surveillance programs to facilitate early diagnosis of cancer, and may therefore benefit from safe and effective chemopreventive agents. Non-steroidal anti-inflammatory drugs (NSAID) have received much attention due to the fact that NSAID use seems to be negatively associated with the development of various cancers, including EA. However, concern over the potential toxicities associated with NSAID use highlights the importance of identifying alternative agents. The general class of statin medications (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) have received recent attention for their potential chemopreventive properties in human and animal studies, but little is known on the possible effects of statin use on the development of EA. Methods: We assessed the association between statin use and risk of progression to EA in a prospective cohort study of 411 persons with BE followed for a total of 33,656 person-months between 1995 and 2009. Data on use of medications, cigarette smoking, anthropometry and other potential risk factors were collected at baseline through personal interviews and updated over the course of follow-up. Proportional hazards regression was used to calculate the hazard ratios (HR) for NSAID and statin use, while adjusting as necessary for the effects of age, sex, cigarette use, and waist-to-hip ratio. Results: Overall, 14% of participants reported statin use at baseline, with 40% reporting use on at least one follow-up visit. NSAID use was more common, with 41% of the cohort reporting regular NSAID use at baseline and 70% reporting regular use on at least one follow-up visit. The main analyses focused on the 396 persons with at least 5 months of follow up (median 79.5 months; range 5.7 −170.7 months). The HR for statin use, entered into models as a time-varying covariate, was 0.59 (95% CI=0.26-1.33) when adjusted for age, sex, and smoking history; when NSAID use was added to the model, the HR weakened slightly to 0.68 (95% CI=0.30-1.55). Among persons with high-grade dysplasia, the HR for association between statin use and EA was 0.45 (95%CI=0.15, 1.29). Among persons with high-grade dysplasia, analyses of the joint associations between statins and NSAID was consistent with a multiplicative relationship, with an estimated HR of 0.19 (95%CI=0.06-0.65) comparing persons using both medications to those using neither. Conclusions: While reduced risks of EA that we observed among statin users may be explained by chance, the point estimates are similar in magnitude to those previously reported for NSAID use in this cohort, and are unlikely to be confounded by known risk factors. Among those with highgrade dysplasia at baseline, there is a suggestion that concurrent use of NSAID and statins is more protective than taking NSAID or statins individually. Continued study of this issue in this and other cohorts is warranted. Citation Information: Cancer Prev Res 2010;3(12 Suppl):A60.

Are more genes better in clinical-genomic studies? A mathematical model to define identification power for clinically relevant genes

e22136 Background: Modern molecular technologies that drive personalized medicine can generate expression data for thousands of candidate genes, or indeed, data for the “whole genome”. Clinical-genomic studies aim to identify genes that are truly associated with clinical outcome. We investigated the impact of large numbers of genes with little or no association with clinical outcome on the statistical power of studies to identify individual genes with strong association. Methods: We adapted Efron's (Ann Stat 2007) empirical Bayes approach to develop a method to calculate the identification power for individual genes in analyses that control the false discovery rate (FDR), the expected proportion of false positives. The identification power is the probability that a gene with a given true magnitude of association with clinical outcome will be identified when we control the FDR at a specified level. The identification power also depends on the proportion of genes studied that are not associated with outcome and the distribution of the degree of association among genes that are. Results: The identification power for clinically relevant genes decreases dramatically as the proportion of genes having no association with clinical outcome increases. For example, in a scenario in which 100 genes have some association with clinical outcome [median hazard ratio (HR) 1.125], increasing the number of genes having no association from 400 to 4000 decreases the identification power an individual gene having strong association (HR 1.42) from 80% to 36%. Similarly, when the number of genes having no association is 400 and the median HR among the 100 genes that have an association is decreased from 1.125 to 1.06, identification power for a gene with an association HR of 1.42 decreases from 80% to 62%. Conclusions: Identification power can be used to optimize strategies for gene finding and enable personalized medicine. Although technology allows the assay of ever-increasing numbers of genes, inclusion in a single study of many genes unrelated to clinical outcome is detrimental to the identification power for clinically relevant genes. Even with increased power, replication of results in independent studies will continue to be critically important. [Table: see text]