Extension of Kaplan-Meier methods in observational studies with time-varying treatment

Treatment Persistence and Medication Switch Associated With Subsequent Fractures After Osteoporotic Fractures

CONTEXT

Despite prevalent anti-osteoporosis medication (AOM) switching in real-world osteoporosis management, few studies have evaluated the impact of persistent AOM treatment, allowing for AOM switching, on the risk of subsequent fracture.

OBJECTIVE

We examined the association between persistence in AOM and subsequent fractures, allowing for medication switching among patients with osteoporotic fractures.

METHODS

This retrospective cohort study used Taiwan National Health Insurance claims data to select patients who initiated AOM between 2013 and 2016. Treatment persistence was defined as use of any AOM on a given day of interest with a 45-day grace period. Medication switch was allowed for persistence if remaining on treatment. AOMs with long-lasting inhibition of bone resorption (zoledronate and denosumab) were categorized as high-potency; others as low-potency. Multivariate Cox models were used to evaluate risk of subsequent fractures ≥3 months after initiating AOM.

RESULTS

A total of 119 473 patients were included (mean [SD] follow-up 46.4 [15.6] months), and 26.8% switched from the index AOM. Within 1 year, 52% remained persistent with AOM. Compared to patients with persistent AOM, those not persistent had higher risk of subsequent hip (adjusted hazard ratio [aHR] = 1.31; 95% CI, 1.21-1.42), vertebral (aHR = 1.17; 95% CI, 1.13-1.22), and radius fractures (aHR = 1.16; 95% CI, 1.08-1.25). Patients with persistent AOM who switched from high- to low-potency AOM had higher risk of subsequent vertebral fractures than those with persistent AOM and no potency switch (aHR = 1.28; 95% CI, 1.02-1.60).

CONCLUSION

Patients with non-persistent AOM had higher risk of subsequent fractures than persistent users when allowing AOM switch. Switching AOM potency may influence the risk of subsequent vertebral fractures and warrants further investigation.

Considering Questions Before Methods in Dementia Research With Competing Events and Causal Goals

Studying causal exposure effects on dementia is challenging when death is a competing event. Researchers often interpret death as a potential source of bias, although bias cannot be defined or assessed if the causal question is not explicitly specified. Here we discuss 2 possible notions of a causal effect on dementia risk: the "controlled direct effect" and the "total effect." We provide definitions and discuss the "censoring" assumptions needed for identification in either case and their link to familiar statistical methods. We illustrate concepts in a hypothetical randomized trial on smoking cessation in late midlife, and emulate such a trial using observational data from the Rotterdam Study, the Netherlands, 1990-2015. We estimated a total effect of smoking cessation (compared with continued smoking) on 20-year dementia risk of 2.1 (95% confidence interval: -0.1, 4.2) percentage points and a controlled direct effect of smoking cessation on 20-year dementia risk had death been prevented of -2.7 (95% confidence interval: -6.1, 0.8) percentage points. Our study highlights how analyses corresponding to different causal questions can have different results, here with point estimates on opposite sides of the null. Having a clear causal question in view of the competing event and transparent and explicit assumptions are essential to interpreting results and potential bias.

Utilizing Electronic Health Records (EHR) and Tumor Panel Sequencing to Demystify Prognosis of Cancer of Unknown Primary (CUP) patients

Cancer of unknown primary (CUP) is a type of cancer that cannot be traced back to its original site and accounts for 3-5% of all cancers. It does not have established targeted therapies, leading to poor outcomes. We developed OncoNPC, a machine learning classifier trained on targeted next-generation sequencing data from 34,567 tumors from three institutions. OncoNPC achieved a weighted F1 score of 0.94 for high confidence predictions on known cancer types (65% of held-out samples). When applied to 971 CUP tumors from patients treated at the Dana-Farber Cancer Institute, OncoNPC identified actionable molecular alterations in 23% of the tumors. Furthermore, OncoNPC identified CUP subtypes with significantly higher polygenic germline risk for the predicted cancer type and significantly different survival outcomes, supporting its validity. Importantly, CUP patients who received first palliative intent treatments concordant with their OncoNPC-predicted cancer sites had significantly better outcomes (H.R. 0.348, 95% C.I. 0.210 - 0.570, p-value 2.32 × 10-5). OncoNPC thus provides evidence of distinct CUP subtypes and offers the potential for clinical decision support for managing patients with CUP.

Effect of Quality of Caregiver-Adolescent Relationship on Sexual Debut, Transactional Sex, and on Age-Disparate Relationships Among Young Women in Rural South Africa Enrolled in HPTN 068

BACKGROUND

Adolescent girls and young women (AGYW) remain burdened by the HIV epidemic. Positive caregiver-child relationships are associated with safe sexual behaviors in young people; yet, this literature often highlights the role of parent-child communication and parental monitoring, neglecting the importance of emotional relationships between a caregiver and adolescent.

SETTING

We used longitudinal data from HIV Prevention Trial Network 068-conducted among 2533 AGYW (13-20 years) over a period of 5 years in Agincourt, South Africa.

METHOD

Kaplan-Meier and Cox models were used to estimate the effect of quality of caregiver-adolescent relationships (caring and closeness) on sexual debut, and log-binomial models with generalized estimating equations were used to examine the relationship between our exposures and transactional sex and age-disparate relationships.

RESULTS

Sexual debut was delayed among those who reported high levels of caregiver caring [hazard ratio: 0.80, 95% confidence interval (CI): 0.69 to 0.93] and caregiver closeness (hazard ratio: 0.80, 95% CI: 0.68 to 0.95). AGYW who reported high quality caregiver-adolescent relationships had a lower risk of transactional sex [caring: risk ratio (RR): 0.67, 95% CI: 0.58 to 0.78; closeness: RR: 0.58; 95% CI: 0.50 to 0.67]. Similarly, those with high-quality caregiver-adolescent relationships were less likely to be in an age-disparate relationship (caring: RR: 0.68, 95% CI: 0.58 to 0.79; closeness: RR: 0.77, 95% CI: 0.66 to 0.90).

CONCLUSIONS

Findings indicate high-quality caregiver-adolescent relationships are associated with delayed sexual debut, a lower risk of transactional sex, and having an older partner. Family-centered interventions are needed to improve relationships between AGYW and caregivers.

Dexamethasone in hospitalised coronavirus-19 patients not on intensive respiratory support

Background

Dexamethasone decreases mortality in coronavirus disease 2019 (COVID-19) patients on intensive respiratory support (IRS) but is of uncertain benefit if less severely ill. We determined whether early (within 48 h) dexamethasone was associated with mortality in patients hospitalised with COVID-19 not on IRS.

Methods

We included patients admitted to US Veterans Affairs hospitals between 7 June 2020 and 31 May 2021 within 14 days after a positive test for severe acute respiratory syndrome coronavirus 2. Exclusions included recent prior corticosteroids and IRS within 48 h. We used inverse probability of treatment weighting (IPTW) to balance exposed and unexposed groups, and Cox proportional hazards models to determine 90-day all-cause mortality.

Results

Of 19 973 total patients (95% men, median age 71 years, 27% black), 15 404 (77%) were without IRS within 48 h. Of these, 3514 out of 9450 (34%) patients on no oxygen received dexamethasone and 1042 (11%) died; 4472 out of 5954 (75%) patients on low-flow nasal cannula (NC) only received dexamethasone and 857 (14%) died. In IPTW stratified models, patients on no oxygen who received dexamethasone experienced 76% increased risk for 90-day mortality (hazard ratio (HR) 1.76, 95% CI 1.47–2.12); there was no association with mortality among patients on NC only (HR 1.08, 95% CI 0.86–1.36).

Conclusions

In patients hospitalised with COVID-19, early initiation of dexamethasone was common and was associated with no mortality benefit among those on no oxygen or NC only in the first 48 h; instead, we found evidence of potential harm. These real-world findings do not support the use of early dexamethasone in hospitalised COVID-19 patients without IRS.

Although commonly used, dexamethasone within 48 h of admission was associated with increased 90-day mortality in patients hospitalised with COVID-19 not on oxygen and with no mortality benefit in patients on low-flow nasal cannulahttps://bit.ly/3l2aqjb

Using marginal structural models to analyze the impact of subsequent therapy on the treatment effect in survival data: Simulations and clinical trial examples

We explore the impact of time-varying subsequent therapy on the statistical power and treatment effects in survival analysis. The marginal structural model (MSM) with stabilized inverse probability treatment weights (sIPTW) was used to account for the effects due to the subsequent therapy. Simulations were performed to compare the MSM-sIPTW method with the conventional method without accounting for the time-varying covariate such as subsequent therapy that is dependent on the initial response of the treatment effect. The results of the simulations indicated that the statistical power, thereby the Type I error, of the trials to detect the frontline treatment effect could be inflated if no appropriate adjustment was made for the impact due to the add-on effects of the subsequent therapy. Correspondingly, the hazard ratio between the treatment groups may be overestimated by the conventional analysis methods. In contrast, MSM-sIPTW can maintain the Type I error rate and gave unbiased estimates of the hazard ratio for the treatment. Two real examples were used to discuss the potential clinical implications. The study demonstrated the importance of accounting for time-varying subsequent therapy for obtaining unbiased interpretation of data.

Burden of Antimicrobial Resistance: Compared to What?

The increased focus on the public health burden of antimicrobial resistance (AMR) raises conceptual challenges, such as determining how much harm multidrug-resistant organisms do compared to what, or how to establish the burden. Here, we present a counterfactual framework and provide guidance to harmonize methodologies and optimize study quality. In AMR-burden studies, 2 counterfactual approaches have been applied: the harm of drug-resistant infections relative to the harm of the same drug-susceptible infections (the susceptible-infection counterfactual); and the total harm of drug-resistant infections relative to a situation where such infections were prevented (the no-infection counterfactual). We propose to use an intervention-based causal approach to determine the most appropriate counterfactual. We show that intervention scenarios, species of interest, and types of infections influence the choice of counterfactual. We recommend using purpose-designed cohort studies to apply this counterfactual framework, whereby the selection of cohorts (patients with drug-resistant, drug-susceptible infections, and those with no infection) should be based on matching on time to infection through exposure density sampling to avoid biased estimates. Application of survival methods is preferred, considering competing events. We conclude by advocating estimation of the burden of AMR by using the no-infection and susceptible-infection counterfactuals. The resulting numbers will provide policy-relevant information about the upper and lower bound of future interventions designed to control AMR. The counterfactuals should be applied in cohort studies, whereby selection of the unexposed cohorts should be based on exposure density sampling, applying methods avoiding time-dependent bias and confounding.

Displaying survival of patient groups defined by covariate paths: Extensions of the Kaplan-Meier estimator

Extensions of the Kaplan-Meier estimator have been developed to illustrate the relationship between a time-varying covariate of interest and survival. In particular, Snapinn et al and Xu et al developed estimators to display survival for patients who always have a certain value of a time-varying covariate. These estimators properly handle time-varying covariates, but their clinical interpretation is limited. It is of greater clinical interest to display survival for patients whose covariates lie along certain defined paths. In this article, we propose extensions of Snapinn et al and Xu et al's estimators, providing crude and covariate-adjusted estimates of the survival function for patients defined by covariate paths. We also derive analytical variance estimators. We demonstrate the utility of these estimators with medical examples and a simulation study.

Estimating the Effect of Healthcare-Associated Infections on Excess Length of Hospital Stay Using Inverse Probability-Weighted Survival Curves

BACKGROUND

Studies estimating excess length of stay (LOS) attributable to nosocomial infections have failed to address time-varying confounding, likely leading to overestimation of their impact. We present a methodology based on inverse probability-weighted survival curves to address this limitation.

METHODS

A case study focusing on intensive care unit-acquired bacteremia using data from 2 general intensive care units (ICUs) from 2 London teaching hospitals were used to illustrate the methodology. The area under the curve of a conventional Kaplan-Meier curve applied to the observed data was compared with that of an inverse probability-weighted Kaplan-Meier curve applied after treating bacteremia as censoring events. Weights were based on the daily probability of acquiring bacteremia. The difference between the observed average LOS and the average LOS that would be observed if all bacteremia cases could be prevented was multiplied by the number of admitted patients to obtain the total excess LOS.

RESULTS

The estimated total number of extra ICU days caused by 666 bacteremia cases was estimated at 2453 (95% confidence interval [CI], 1803-3103) days. The excess number of days was overestimated when ignoring time-varying confounding (2845 [95% CI, 2276-3415]) or when completely ignoring confounding (2838 [95% CI, 2101-3575]).

CONCLUSIONS

ICU-acquired bacteremia was associated with a substantial excess LOS. Wider adoption of inverse probability-weighted survival curves or alternative techniques that address time-varying confounding could lead to better informed decision making around nosocomial infections and other time-dependent exposures.

Digoxin, mortality, and cardiac hospitalizations in patients with atrial fibrillation and heart failure with reduced ejection fraction and atrial fibrillation: An AF-CHF analysis

BACKGROUND

Recent publications have raised serious concerns regarding the safety of digoxin for atrial fibrillation (AF). However, the subgroup of patients with reduced ejection fraction and AF have been speculated to derive clinical benefit from digoxin. We aimed to assess the impact of digoxin on mortality and cardiovascular hospitalizations in the Atrial Fibrillation and Congestive Heart Failure (AF-CHF) trial since all AF-CHF patients had an ejection fraction ≤35% and AF.

METHODS AND RESULTS

Using marginal structural modeling, a contemporary statistical method that overcomes limitations of traditional modeling techniques and reduces bias, we assessed the impact of digoxin on the pre-specified primary and secondary outcomes of the AF-CHF trial, i.e., all-cause, cardiac and arrhythmic death as well as cardiovascular hospitalization. Among 1376 patients, 869 (65%) were on digoxin at one-year follow-up. Over a mean (SD) follow-up of 37 (19) months (maximum 74 months), 445 (32%) patients died, 357 (26%) from cardiovascular causes and 159 (12%) from arrhythmic death. Digoxin was significantly associated with all-cause, cardiac, and arrhythmic death, with estimated hazard ratios (HR) of 1.39 (95% confidence interval [CI] 1.11-1.73, P = 0.004), 1.44 (95% CI 1.13-1.82, P = 0.003), and 2.03 (95% CI 1.63-2.54, P < 0.0001), respectively. Digoxin was not associated with cardiovascular hospitalizations [HR 1.12 (95% CI 0.91-1.37), P = 0.29].

CONCLUSION

Digoxin is associated with increased all-cause mortality among patients with combined heart failure with reduced ejection fraction and AF, which is predominantly driven by arrhythmic deaths. In contrast, cardiovascular hospitalizations were not impacted by digoxin.

Age-disparate partnerships and incident HIV infection in adolescent girls and young women in rural South Africa

OBJECTIVE

Adolescent girls and young women (AGYW) have a much higher risk of HIV infection than young men of the same age. One hypothesis for this disparity is AGYW are more likely to be in sexual partnerships with older men with HIV; however, evidence has been inconclusive.

DESIGN

We used longitudinal data from a randomized trial in South Africa (HPTN 068) to determined whether partner age difference is associated with incident HIV infection in AGYW.

METHODS

Age difference was examined continuously and dichotomously (≥5 years). We examined inverse probability of exposure weighted survival curves and calculated time-specific risk differences and risk ratios over 5.5 years of follow-up. We also used a marginal structural Cox model to estimate hazard ratios over the entire study period.

RESULTS

Risk of HIV was higher in AGYW with an age-disparate partnership versus not and the risk difference was largest at later time points. At 5.5 years, AGYW with an age-disparate partnership had a 12.6% (95% confidence interval 1.9-23.3) higher risk than AGYW with no age-disparate partnerships. The weighted hazard ratio was 1.91 (95% confidence interval 1.33-2.74), an association that remained after weighting for either transactional or condomless sex, and after examining continuous age-differences.

CONCLUSION

Age-disparate partnerships increased risk of HIV infection, even after accounting for transactional sex and condomless sex. The relationship between age-disparate partnerships and HIV infection may be explained by increased exposure to infection from men in a higher HIV prevalence pool rather than differences in sexual behaviour within these partnerships.

Transactional sex and incident HIV infection in a cohort of young women from rural South Africa

Objective:

In sub-Saharan Africa, young women who engage in transactional sex (the exchange of sex for money or gifts) with a male partner show an elevated risk of prevalent HIV infection. We analyse longitudinal data to estimate the association between transactional sex and HIV incidence.

Design:

We used longitudinal data from a cohort of 2362 HIV-negative young women (aged 13–20 years) enrolled in a randomized controlled trial in rural, South Africa who were followed for up to four visits over 6 years.

Methods:

The effect of transactional sex on incident HIV was analysed using stratified Cox proportional hazards models and cumulative incidence curves. Risk ratios were estimated using log-binomial models to compare the effects across visits.

Results:

HIV incidence was higher for young women that reported transactional sex (hazard ratio 1.59, 95% confidence interval 1.02–2.19), particularly when money and/or gifts were received frequently (at least weekly) (hazard ratio 2.71, 95% confidence interval 1.44–5.12). We also find that effects were much stronger during the main trial and dissipate at the postintervention visit, despite an increase in both transactional sex and HIV.

Conclusion:

Transactional sex elevates the risk of HIV acquisition among young women, especially when it involves frequent exchanges of money and/or gifts. However, the effect was attenuated after the main trial, possibly due to the changing nature of transactional sex and sexual partners as women age. These findings suggest that reducing transactional sex among young women, especially during adolescence, is important for HIV prevention.

Efficacy of thoracic radiotherapy in patients with stage IIIB-IV epidermal growth factor receptor-mutant lung adenocarcinomas who received and responded to tyrosine kinase inhibitor treatment

PURPOSE

Large-scale, prospective, randomized studies of the efficacy of thoracic radiotherapy (RT) in patients with unresectable stage IIIB-IV epidermal growth factor receptor (EGFR)-mutant lung adenocarcinomas who received and responded to EGFR tyrosine kinase inhibitor (TKI) treatment are not currently available. Therefore, we designed a propensity score-matched, nationwide, population-based, cohort study for estimating the effects of thoracic RT on patients with EGFR-mutant lung adenocarcinomas.

PATIENTS AND METHODS

We analyzed patients with unresectable stage IIIB-IV EGFR mutant lung adenocarcinomas and categorized them into two groups according to treatment modality and compared their outcomes; groups 1 and 2 consisted of patients who received EGFR TKI treatment alone until tumor progression and those who received and responded to EGFR TKI treatment and subsequently received thoracic RT for lung tumors, respectively. The patients in groups 2 and 1 were matched at a ratio of 1:4.

RESULTS

The matching process yielded a final cohort of 1475 patients (1180 and 295 patients in groups 1 and 2, respectively) who were eligible for further analysis. According to both univariate and multivariate Cox regression analyses, the adjusted hazard ratios (aHRs) (95% confidence interval [CI]) derived for thoracic RT for lung tumor after EGFR TKI use and tumor response (group 2) compared with EGFR TKI treatment alone (group 1) was 0.72 (0.60-0.85).

CONCLUSIONS

Thoracic RT might be associated with overall survival in patients with unresectable stage IIIB-IV EGFR-mutant lung adenocarcinomas who received and responded to EGFR TKI treatment.

Potential utility of a longitudinal relative dose intensity of molecularly targeted agents in phase 1 dose-finding trials

Phase 1 trials of molecularly targeted agents (MTA) often do not use toxicity data beyond the first cycle of treatment to determine a recommended phase 2 dose (RP2D). We investigated the potential utility of longitudinal relative dose intensity (RDI) that may be a better new way of determining a more accurate RP2D as a lower dose that is presumably more tolerable over the long term without compromising efficacy. All consecutive patients who were initially treated using a single MTA at the conventional RP2D or at one level lower dose (OLLD) of that RP2D in 9 phase 1 trials sponsored by the National Cancer Institute were included. The associations between longitudinal RDI, time to first progression, and response rate were analyzed. The RDI of the conventional RP2D group were maintained a rate of ≥70% throughout 10 cycles, and were higher than those of the OLLD group, although in both groups the RDI gradually decreased with additional treatment cycles. The RP2D group was similar to the OLLD group with respect to time to first progression and response rate. In both groups, however, the decreasing RDI over time was significantly associated with shorter time to first disease progression; therefore, the longitudinal RDI, which takes into account lower grade toxicity occurrences, may be useful in determining a more desirable dose to use in phase 2 and 3 studies.

The effect of school attendance and school dropout on incident HIV and HSV-2 among young women in rural South Africa enrolled in HPTN 068

OBJECTIVE

To estimate the association between school attendance, school dropout, and risk of incident HIV and herpes simplex virus type 2 (HSV-2) infection among young women.

DESIGN

We used longitudinal data from a randomized controlled trial in rural Mpumalanga province, South Africa, to assess the association between school days attended, school dropout, and incident HIV and HSV-2 in young women aged 13-23 years.

METHODS

We examined inverse probability of exposure weighted survival curves and used them to calculate 1.5, 2.5, and 3.5-year risk differences and risk ratios for the effect of school attendance on incident HIV and HSV-2. A marginal structural Cox model was used to estimate hazard ratios for the effect of school attendance and school dropout on incident infection.

RESULTS

Risk of infection increased over time as young women aged, and was higher in young women with low school attendance (<80% school days) compared with high (≥80% school days). Young women with low attendance were more likely to acquire HIV [hazard ratio (HR): 2.97; 95% confidence interval (CI): 1.62, 5.45] and HSV-2 (HR: 2.47; 95% CI: 1.46, 4.17) over the follow-up period than young women with high attendance. Similarly, young women who dropped out of school had a higher weighted hazard of both HIV (HR 3.25 95% CI: 1.67, 6.32) and HSV-2 (HR 2.70; 95% CI 1.59, 4.59).

CONCLUSION

Young women who attend more school days and stay in school have a lower risk of incident HIV and HSV-2 infection. Interventions to increase frequency of school attendance and prevent dropout should be promoted to reduce risk of infection.

Inverse probability weighting to estimate causal effect of a singular phase in a multiphase randomized clinical trial for multiple myeloma

Background

Randomization procedure in randomized controlled trials (RCTs) permits an unbiased estimation of causal effects. However, in clinical practice, differential compliance between arms may cause a strong violation of randomization balance and biased treatment effect among those who comply. We evaluated the effect of the consolidation phase on disease-free survival of patients with multiple myeloma in an RCT designed for another purpose, adjusting for potential selection bias due to different compliance to previous treatment phases.

Methods

We computed two propensity scores (PS) to model two different selection processes: the first to undergo autologous stem cell transplantation, the second to begin consolidation therapy. Combined stabilized inverse probability treatment weights were then introduced in the Cox model to estimate the causal effect of consolidation therapy miming an ad hoc RCT protocol.

Results

We found that the effect of consolidation therapy was restricted to the first 18 months of the phase (HR: 0.40, robust 95 % CI: 0.17-0.96), after which it disappeared.

Conclusions

PS-based methods could be a complementary approach within an RCT context to evaluate the effect of the last phase of a complex therapeutic strategy, adjusting for potential selection bias caused by different compliance to the previous phases of the therapeutic scheme, in order to simulate an ad hoc randomization procedure.

Trial registration

ClinicalTrials.gov: NCT01134484 May 28, 2010 (retrospectively registered)

EudraCT: 2005-003723-39 December 17, 2008 (retrospectively registered)

Testosterone therapy in men with testosterone deficiency: are the benefits and cardiovascular risks real or imagined?

In the adult male, testosterone (T) deficiency (TD) also known as male hypogonadism, is a well-established medical condition, which has been recognized for more than a century. T therapy in men with TD was introduced as early as 1940s and was reported to improve overall health with no concomitant serious adverse effects. A wealth of recent studies demonstrated that T therapy in men with TD is associated with increased lean body mass, reduced fat mass and waist circumference, improvement in glycemic control, and reduced obesity. T therapy is also associated with improvements in lipid profiles, amelioration of metabolic syndrome (Met S) components, reduced inflammatory biomarkers, reduced systolic and diastolic blood pressure, and improvements in sexual function. More importantly, T therapy is associated with amelioration of diabetes and reduced mortality. However, few studies, marred with serious methodological and analytical flaws reported between 2010 and 2014, suggested that T therapy is associated with increased cardiovascular (CV) risk. As summarized in this review, a thorough and critical analysis of these studies showed that the risks purported are unsubstantiated and such studies lacked credible scientific and clinical evidence. Moreover, recent observational, registry studies, clinical trials, and meta-analyses, all revealed no increase in CV risks in men receiving T therapy. In this review, the benefits of T therapy in adult men with TD and the lack of credible evidence suggesting that T therapy is linked to increased CV risks are discussed. It should be noted that the literature is replete with studies demonstrating beneficial effects of T therapy on CV and overall health.

Testosterone deficiency and cardiovascular mortality

New concerns have been raised regarding cardiovascular (CV) risks with testosterone (T) therapy (TTh). These concerns are based primarily on two widely reported retrospective studies. However, methodological flaws and data errors invalidate both studies as credible evidence of risk. One showed reduced adverse events by half in T-treated men but reversed this result using an unproven statistical approach. The authors subsequently acknowledged serious data errors including nearly 10% contamination of the dataset by women. The second study mistakenly used the rate of T prescriptions written by healthcare providers to men with recent myocardial infarction (MI) as a proxy for the naturally occurring rate of MI. Numerous studies suggest T is beneficial, including decreased mortality in association with TTh, reduced MI rate with TTh in men with the greatest MI risk prognosis, and reduced CV and overall mortality with higher serum levels of endogenous T. Randomized controlled trials have demonstrated benefits of TTh in men with coronary artery disease and congestive heart failure. Improvement in CV risk factors such as fat mass and glycemic control have been repeatedly demonstrated in T-deficient men treated with T. The current evidence does not support the belief that TTh is associated with increased CV risk or CV mortality. On the contrary, a wealth of evidence accumulated over several decades suggests that low serum T levels are associated with increased risk and that higher endogenous T, as well as TTh itself, appear to be beneficial for CV mortality and risk.

Severe COPD Exacerbation Risk and Long-Acting Bronchodilator Treatments: Comparison of Three Observational Data Analysis Methods

OBJECTIVE

Results from three observational methods for assessing effectiveness of long-acting bronchodilator therapies for reducing severe exacerbations of chronic obstructive pulmonary disease (COPD) were compared: intent-to-treat (ITT), as protocol (AP), and an as-treated analysis that utilized a marginal structural model (MSM) incorporating time-varying covariates related to treatment adherence and moderate exacerbations.

STUDY DESIGN AND SETTING

Severe exacerbation risk was assessed over a 2-year period using claims data for patients aged ≥40 years who initiated long-acting muscarinic antagonist (LAMA), inhaled corticosteroid/long-acting beta-agonist (ICS/LABA), or triple therapy (LAMA + ICS/LABA).

RESULTS

A total of 5475 COPD patients met inclusion criteria. Six months post-initiation, 53.5 % of patients discontinued using any therapy. The ITT analysis found an increased severe exacerbation risk for triple therapy treatment (hazard ratio [HR] 1.24; 95 % confidence interval [CI] 1.00-1.53). No increased risk was found in the AP (HR 1.00; 95 % CI 0.73-1.36), or MSM analyses (HR 1.11; 95 % CI 0.68-1.81). The MSM highlighted important associations among post-index events.

CONCLUSION

Neglecting to adjust for treatment discontinuation may produce biased risk estimates. The MSM approach is a promising tool to compare chronic disease management by illuminating relationships between treatment decisions, adherence, patient choices, and outcomes.

Testosterone therapy and cardiovascular risk: advances and controversies

Two recent studies raised new concerns regarding cardiovascular (CV) risks with testosterone (T) therapy. This article reviews those studies as well as the extensive literature on T and CV risks. A MEDLINE search was performed for the years 1940 to August 2014 using the following key words: testosterone, androgens, human, male, cardiovascular, stroke, cerebrovascular accident, myocardial infarction, heart attack, death, and mortality. The weight and direction of evidence was evaluated and level of evidence (LOE) assigned. Only 4 articles were identified that suggested increased CV risks with T prescriptions: 2 retrospective analyses with serious methodological limitations, 1 placebo-controlled trial with few major adverse cardiac events, and 1 meta-analysis that included questionable studies and events. In contrast, several dozen studies have reported a beneficial effect of normal T levels on CV risks and mortality. Mortality and incident coronary artery disease are inversely associated with serum T concentrations (LOE IIa), as is severity of coronary artery disease (LOE IIa). Testosterone therapy is associated with reduced obesity, fat mass, and waist circumference (LOE Ib) and also improves glycemic control (LOE IIa). Mortality was reduced with T therapy in 2 retrospective studies. Several RCTs in men with coronary artery disease or heart failure reported improved function in men who received T compared with placebo. The largest meta-analysis to date revealed no increase in CV risks in men who received T and reduced CV risk among those with metabolic disease. In summary, there is no convincing evidence of increased CV risks with T therapy. On the contrary, there appears to be a strong beneficial relationship between normal T and CV health that has not yet been widely appreciated.

Estimating the effects of time-varying exposures in observational studies using Cox models with stabilized weights adjustment

PURPOSE

Assessing the safety and effectiveness of medical products with observational electronic medical record data is challenging when the treatment is time-varying. The objective of this paper is to develop a Cox model stratified by event times with stabilized weights (SWs) adjustment to examine the effect of time-varying treatment in observational studies.

METHODS

Time-varying SWs are calculated at unique event times and are used in a Cox model stratified by event times to estimate the effect of time-varying treatment. We applied this method in examining the effect of an antiplatelet agent, clopidogrel, on events, including bleeding, myocardial infarction, and death after a drug-eluting stent was implanted in coronary artery. Clopidogrel use may change over time on the basis of patients' behavior (e.g., non-adherence) and physicians' recommendations (e.g., end of duration of therapy). We also compared the results with those from a Cox model for counting processes adjusting for all covariates used in creating SWs.

RESULTS

We demonstrate that the (i) results from the stratified Cox model without SWs adjustment and the Cox model for counting processes without covariate adjustment are identical in analyzing the clopidogrel data; and (ii) the effects of clopidogrel on bleeding, myocardial infarction, and death are larger in the stratified Cox model with SWs adjustment compared with those from the Cox model for counting processes with covariate adjustment.

CONCLUSIONS

The Cox model stratified by event times with time-varying SWs adjustment is useful in estimating the effect of time-varying treatments in observational studies while balancing for known confounders.