Practical methodology of meta-analysis of individual patient data using a survival outcome

Effect of KRAS codon 12 or 13 mutations on survival with trifluridine/tipiracil in pretreated metastatic colorectal cancer: a meta-analysis

Background

KRAS gene mutations can predict prognosis and treatment response in patients with metastatic colorectal cancer (mCRC).

Methods

We undertook a meta-analysis of three randomized, placebo-controlled trials (RECOURSE, TERRA and J003) to investigate the impact of KRAS mutations in codons 12 or 13 on overall survival (OS) and progression-free survival in patients receiving trifluridine/tipiracil (FTD/TPI) for refractory mCRC.

Results

A total of 1375 patients were included, of whom 478 had a KRAS codon 12 mutation and 130 had a KRAS codon 13 mutation. In univariate analyses, the absence of a KRAS codon 12 mutation was found to significantly increase the OS benefit of FTD/TPI relative to placebo compared with the presence of the mutation {hazard ratio (HR), 0.62 [95% confidence interval (CI): 0.53-0.72] versus 0.86 (0.70-1.05), respectively; interaction P = 0.0206}. Multivariate analyses showed that taking confounding factors into account reduced the difference in treatment effect between the presence and the absence of KRAS codon 12 mutations, confirming that treatment benefit was maintained in patients with [HR, 0.73 (95% CI: 0.59-0.89)] and without [HR, 0.63 (95% CI: 0.54-0.74)] codon 12 mutations (interaction P = 0.2939). KRAS mutations in codon 13 did not reduce the OS benefit of FTD/TPI relative to placebo, and, furthermore, KRAS mutations at either codon 12 or codon 13 did not affect the progression-free survival benefit.

Conclusions

Treatment with FTD/TPI produced a survival benefit, relative to placebo, regardless of KRAS codon 12 or 13 mutation status in patients with previously treated mCRC.

●

KRAS mutations are associated with negative outcomes in patients with mCRC; codon 12 and 13 mutations are the most common.

●

FTD/TPI was associated with longer median overall survival vs placebo both in patients with wild-type KRAS and mutant KRAS.

●

FTD/TPI produced a survival benefit, relative to placebo, regardless of KRAS codon 12 or 13 mutation status in this patient group.

Association of RASSF1A, DCR2, and CASP8 Methylation with Survival in Neuroblastoma: A Pooled Analysis Using Reconstructed Individual Patient Data

Neuroblastoma (NB) is a heterogeneous tumor affecting children. It shows a wide spectrum of clinical outcomes; therefore, development of risk stratification is critical to provide optimum treatment. Since epigenetic alterations such as DNA methylation have emerged as an important feature of both development and progression in NB, in this study, we aimed to quantify the effect of methylation of three distinct genes (RASSF1A, DCR2, and CASP8) on overall survival in NB patients. We performed a systematic review using PubMed, Embase, and Cochrane libraries. Individual patient data was retrieved from extracted Kaplan–Meier curves. Data from studies was then merged, and analysis was done on the full data set. Seven studies met the inclusion criteria. Methylation of the three genes had worse overall survival than the unmethylated arms. Five-year survival for the methylated arm of RASSF1A, DCR2, and CASP8 was 63.19% (95% CI 56.55-70.60), 57.78% (95% CI 47.63-70.08), and 56.39% (95% CI 49.53-64.19), respectively, while for the unmethylated arm, it was 93.10% (95% CI 87.40–99.1), 84.84% (95% CI 80.04-89.92), and 83.68% (95% CI 80.28-87.22), respectively. In conclusion, our results indicate that in NB patients, RASSF1A, DCR2, and CASP8 methylation is associated with poor prognosis. Large prospective studies will be necessary to confirm definitive correlation between methylation of these genes and survival taking into account all other known risk factors. (PROSPERO registration number CRD42017082264).

Individual participant data meta-analysis of intervention studies with time-to-event outcomes: A review of the methodology and an applied example

Many randomized trials evaluate an intervention effect on time-to-event outcomes. Individual participant data (IPD) from such trials can be obtained and combined in a so-called IPD meta-analysis (IPD-MA), to summarize the overall intervention effect. We performed a narrative literature review to provide an overview of methods for conducting an IPD-MA of randomized intervention studies with a time-to-event outcome. We focused on identifying good methodological practice for modeling frailty of trial participants across trials, modeling heterogeneity of intervention effects, choosing appropriate association measures, dealing with (trial differences in) censoring and follow-up times, and addressing time-varying intervention effects and effect modification (interactions).We discuss how to achieve this using parametric and semi-parametric methods, and describe how to implement these in a one-stage or two-stage IPD-MA framework. We recommend exploring heterogeneity of the effect(s) through interaction and non-linear effects. Random effects should be applied to account for residual heterogeneity of the intervention effect. We provide further recommendations, many of which specific to IPD-MA of time-to-event data from randomized trials examining an intervention effect.We illustrate several key methods in a real IPD-MA, where IPD of 1225 participants from 5 randomized clinical trials were combined to compare the effects of Carbamazepine and Valproate on the incidence of epileptic seizures.

Does leadership support buffer the effect of workplace bullying on the risk of disability pensioning? An analysis of register-based outcomes using pooled survey data from 24,538 employees

PURPOSE

To investigate whether self-reported exposure to workplace bullying predicts the risk of disability pensioning among employees in two occupational groups-(1) employees working with clients or customers and (2) office workers and manual workers-and whether leadership support and occupational group moderates that association.

METHODS

Survey data from 24,538 employees (112,889 person years) were fitted to a national register containing information on disability-pension payments. Using multi-adjusted Cox-regression analysis, observations were followed in the register to assess the risk of disability pensioning. The average follow-up time was 4.6 years (standard deviation [SD] = 1.5).

RESULTS

Self-reported exposure to workplace bullying predicted an increased risk of disability pensioning (hazard ratio [HR] = 1.46; 95% confidence interval [CI]: 1.15-1.86). This association was moderated by leadership support: the association between workplace bullying and disability pensioning was significantly different for respondents who reported low leadership support (HR = 1.97; 95% CI: 1.38-2.80) compared to respondents who reported medium (HR = 1.03; 95% CI: 0.60-1.76) or high leadership support (HR = 1.08; 95% CI: 0.60-1.95). Further analyses showed similar associations between workplace bullying and the risk of disability pensioning among the two occupational groups.

CONCLUSIONS

Self-reported workplace bullying increases the risk of disability pensioning, and this association is buffered by leadership support. Workplace bullying should be considered an important workplace stressor. This study indicates that workplaces may enhance worker retention by actively promoting measures to eliminate the occurrence of workplace bullying and to enhance leadership support.

One-stage individual participant data meta-analysis models: estimation of treatment-covariate interactions must avoid ecological bias by separating out within-trial and across-trial information

Stratified medicine utilizes individual‐level covariates that are associated with a differential treatment effect, also known as treatment‐covariate interactions. When multiple trials are available, meta‐analysis is used to help detect true treatment‐covariate interactions by combining their data. Meta‐regression of trial‐level information is prone to low power and ecological bias, and therefore, individual participant data (IPD) meta‐analyses are preferable to examine interactions utilizing individual‐level information. However, one‐stage IPD models are often wrongly specified, such that interactions are based on amalgamating within‐ and across‐trial information. We compare, through simulations and an applied example, fixed‐effect and random‐effects models for a one‐stage IPD meta‐analysis of time‐to‐event data where the goal is to estimate a treatment‐covariate interaction. We show that it is crucial to centre patient‐level covariates by their mean value in each trial, in order to separate out within‐trial and across‐trial information. Otherwise, bias and coverage of interaction estimates may be adversely affected, leading to potentially erroneous conclusions driven by ecological bias. We revisit an IPD meta‐analysis of five epilepsy trials and examine age as a treatment effect modifier. The interaction is −0.011 (95% CI: −0.019 to −0.003; p = 0.004), and thus highly significant, when amalgamating within‐trial and across‐trial information. However, when separating within‐trial from across‐trial information, the interaction is −0.007 (95% CI: −0.019 to 0.005; p = 0.22), and thus its magnitude and statistical significance are greatly reduced. We recommend that meta‐analysts should only use within‐trial information to examine individual predictors of treatment effect and that one‐stage IPD models should separate within‐trial from across‐trial information to avoid ecological bias. © 2016 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.

Unwanted sexual attention at work and long-term sickness absence: a follow-up register-based study

Background

The current understanding of the relationship between unwanted sexual attention at work and long-term sickness absence (LTSA) is limited for three reasons: 1) the under-researched role of unwanted sexual attention perpetrated by individuals outside the work organization; 2) a widespread use of self-reported measures of sickness absence, with an unclear identification of sickness absence episodes of long duration; 3) the cross-sectional design of most existing studies. The aim of this study was therefore to investigate the relationship between self-reported unwanted sexual attention at work and subsequent LTSA (≥3 weeks), stratifying by gender and source of exposure (i.e., colleagues, managers and/or subordinates vs. clients/customers/patients).

Methods

This prospective study is based on a pooled sample of 14,605 employees from three Danish surveys conducted in 2000, 2004 and 2005, providing a total of 19,366 observations. A single questionnaire-based item was used to assess exposure to unwanted sexual attention. The pooled dataset was merged with Danish register data on LTSA. The risk of first-onset episode of LTSA (up to 18 months after baseline) in connection with unwanted sexual attention was examined using Cox proportional hazards models. We estimated Hazard ratios (HR) and 95 % confidence intervals (95 % CI) adjusted for age, influence at work, work pace, occupational group and mode of data collection. We also adjusted for repeated measures from individual respondents by stratifying the Cox models by wave of survey.

Results

Unwanted sexual attention from colleagues, managers and/or subordinates predicted LTSA among men (HR 2.66; 95 % CI 1.42-5.00). Among women, an elevated but non-statistically significant risk of LTSA (HR 1.18; 95 % CI 0.65-2.14) was found. Unwanted sexual attention from clients/customers/patients did not predict LTSA, neither among men nor among women.

Conclusions

The findings indicate a significantly elevated risk of LTSA, among men only, in relation to exposure to unwanted sexual attention from colleagues, managers and/or subordinates. This study therefore suggests both individual and organizational costs associated with unwanted sexual attention at work. Due to the low prevalence of unwanted sexual attention, larger studies with more statistical power are needed to confirm (or disconfirm) the present findings.

Get real in individual participant data (IPD) meta-analysis: a review of the methodology

Individual participant data (IPD) meta-analysis is an increasingly used approach for synthesizing and investigating treatment effect estimates. Over the past few years, numerous methods for conducting an IPD meta-analysis (IPD-MA) have been proposed, often making different assumptions and modeling choices while addressing a similar research question. We conducted a literature review to provide an overview of methods for performing an IPD-MA using evidence from clinical trials or non-randomized studies when investigating treatment efficacy. With this review, we aim to assist researchers in choosing the appropriate methods and provide recommendations on their implementation when planning and conducting an IPD-MA.

A Comparison of Seven Cox Regression-Based Models to Account for Heterogeneity Across Multiple HIV Treatment Cohorts in Latin America and the Caribbean

Many studies of HIV/AIDS aggregate data from multiple cohorts to improve power and generalizability. There are several analysis approaches to account for cross-cohort heterogeneity; we assessed how different approaches can impact results from an HIV/AIDS study investigating predictors of mortality. Using data from 13,658 HIV-infected patients starting antiretroviral therapy from seven Latin American and Caribbean cohorts, we illustrate the assumptions of seven readily implementable approaches to account for across cohort heterogeneity with Cox proportional hazards models, and we compare hazard ratio estimates across approaches. As a sensitivity analysis, we modify cohort membership to generate specific heterogeneity conditions. Hazard ratio estimates varied slightly between the seven analysis approaches, but differences were not clinically meaningful. Adjusted hazard ratio estimates for the association between AIDS at treatment initiation and death varied from 2.00 to 2.20 across approaches that accounted for heterogeneity; the adjusted hazard ratio was estimated as 1.73 in analyses that ignored across cohort heterogeneity. In sensitivity analyses with more extreme heterogeneity, we noted a slightly greater distinction between approaches. Despite substantial heterogeneity between cohorts, the impact of the specific approach to account for heterogeneity was minimal in our case study. Our results suggest that it is important to account for across cohort heterogeneity in analyses, but that the specific technique for addressing heterogeneity may be less important. Because of their flexibility in accounting for cohort heterogeneity, we prefer stratification or meta-analysis methods, but we encourage investigators to consider their specific study conditions and objectives.

Matching methods to create paired survival data based on an exposure occurring over time: a simulation study with application to breast cancer

Background

Paired survival data are often used in clinical research to assess the prognostic effect of an exposure. Matching generates correlated censored data expecting that the paired subjects just differ from the exposure. Creating pairs when the exposure is an event occurring over time could be tricky. We applied a commonly used method, Method 1, which creates pairs a posteriori and propose an alternative method, Method 2, which creates pairs in “real-time”. We used two semi-parametric models devoted to correlated censored data to estimate the average effect of the exposure HR¯(t): the Holt and Prentice (HP), and the Lee Wei and Amato (LWA) models. Contrary to the HP, the LWA allowed adjustment for the matching covariates (LWA_a) and for an interaction (LWA_i) between exposure and covariates (assimilated to prognostic profiles). The aim of our study was to compare the performances of each model according to the two matching methods.

Methods

Extensive simulations were conducted. We simulated cohort data sets on which we applied the two matching methods, the HP and the LWA. We used our conclusions to assess the prognostic effect of subsequent pregnancy after treatment for breast cancer in a female cohort treated and followed up in eight french hospitals.

Results

In terms of bias and RMSE, Method 2 performed better than Method 1 in designing the pairs, and LWA_a was the best model for all the situations except when there was an interaction between exposure and covariates, for which LWA_i was more appropriate. On our real data set, we found opposite effects of pregnancy according to the six prognostic profiles, but none were statistically significant. We probably lacked statistical power or reached the limits of our approach. The pairs’ censoring options chosen for combination Method 2 - LWA had to be compared with others.

Conclusions

Correlated censored data designing by Method 2 seemed to be the most pertinent method to create pairs, when the criterion, which characterized the pair, was an exposure occurring over time. In such a setting, the LWA was the most appropriate model.

Does affective organizational commitment and experience of meaning at work predict risk of disability pensioning? An analysis of register-based outcomes using pooled data on 40,554 observations in four occupational groups

BACKGROUND

The aim of this study is to investigate whether experience of meaning at work (MAW) and affective organizational commitment (AOC) predict risk of disability pensioning in four occupational groups.

METHODS

Survey data from 40,554 individuals were fitted to a national register (DREAM) containing information on payments of disability pension. Using multi-adjusted Cox-regression, observations were followed in the DREAM-register to assess risk of disability pensioning.

RESULTS

Low levels of MAW significantly increased risk of disability pensioning during follow-up referencing high levels of MAW. Respondents with medium levels of AOC had a significantly reduced risk of disability pensioning, when compared to respondents with high levels of AOC. Furthermore, results indicate an interaction effect between AOC and MAW in predicting risk of disability pension.

CONCLUSIONS

AOC and MAW are significantly associated with risk of disability pensioning. Promoting MAW and managing AOC in contemporary workplaces may contribute towards reducing risk of disability pensioning.

Permutation tests for centre effect on survival endpoints with application in an acute myeloid leukaemia multicentre study

When analysing multicentre data, it may be of interest to test whether the distribution of the endpoint varies among centres. In a mixed-effect model, testing for such a centre effect consists in testing to zero a random centre effect variance component. It has been shown that the usual asymptotic χ(2) distribution of the likelihood ratio and score statistics under the null does not necessarily hold. In the case of censored data, mixed-effects Cox models have been used to account for random effects, but few works have concentrated on testing to zero the variance component of the random effects. We propose a permutation test, using random permutation of the cluster indices, to test for a centre effect in multilevel censored data. Results from a simulation study indicate that the permutation tests have correct type I error rates, contrary to standard likelihood ratio tests, and are more powerful. The proposed tests are illustrated using data of a multicentre clinical trial of induction therapy in acute myeloid leukaemia patients.

Do psychosocial work conditions predict risk of disability pensioning? An analysis of register-based outcomes using pooled data on 40,554 observations

Aims: To investigate whether high psychosocial job demands (quantitative demands and work pace) and low psychosocial job resources (influence at work and quality of leadership) predicted risk of disability pensioning among employees in four occupational groups – employees working with customers, employees working with clients, office workers and manual workers – in line with the propositions of the Job Demands-Resources (JD-R) model. Methods: Survey data from 40,554 individuals were fitted to the DREAM register containing information on payments of disability pension. Using multi-adjusted Cox regression, observations were followed in the DREAM-register to assess risk of disability pensioning. Average follow-up time was 5.9 years (SD=3.0). Results: Low levels of influence at work predicted an increased risk of disability pensioning and medium levels of quantitative demands predicted a decreased risk of disability pensioning in the study population. We found significant interaction effects between job demands and job resources as combinations low quality of leadership and high job demands predicted the highest rate of disability pensioning. Further analyses showed some, but no statistically significant, differences between the four occupational groups in the associations between job demands, job resources and risk of disability pensioning. Conclusions: The study showed that psychosocial job demands and job resources predicted risk of disability pensioning. The direction of some of the observed associations countered the expectations of the JD-R model and the findings of the present study therefore imply that associations between job demands, job resources and adverse labour market outcomes are more complex than conceptualised in the JD-R model.

Do psychosocial job demands and job resources predict long-term sickness absence? An analysis of register-based outcomes using pooled data on 39,408 individuals in four occupational groups

PURPOSE

To investigate whether psychosocial job demands (work pace and quantitative demands) and job resources (influence at work and quality of leadership) predict long-term sickness absence (LTSA) for more than three consecutive weeks in four occupational groups.

METHODS

Survey data pooling 39,408 respondents were fitted to a national register containing information on payments of sickness absence compensation. Using multi-adjusted Cox regression, respondents were followed for an 18-month follow-up period to assess risk of LTSA.

RESULTS

In the entire study population, low and medium levels of influence at work and low quality of leadership predicted a significantly increased risk of LTSA, whereas medium levels of quantitative demands predicted a significantly reduced risk of LTSA. For employees working with clients and for office workers, low and medium influence at work associated with a significantly increased risk of LTSA. For employees working with clients, low quality of leadership predicted a significantly increased risk of LTSA. For manual workers, low influence at work predicted a significantly increased risk of LTSA and medium quantitative demands were associated with a significantly reduced risk of LTSA. For employees working with customers, medium quantitative demands predicted a significantly reduced risk of LTSA. Finally, in predicting LTSA, we found significant interaction effects between job demands and job resources.

CONCLUSIONS

The study indicates that a lack of job resources--particularly influence at work--are more important predictors of LTSA than high job demands.

CD49d is the strongest flow cytometry-based predictor of overall survival in chronic lymphocytic leukemia

PURPOSE

Although CD49d is an unfavorable prognostic marker in chronic lymphocytic leukemia (CLL), definitive validation evidence is lacking. A worldwide multicenter analysis was performed using published and unpublished CLL series to evaluate the impact of CD49d as an overall (OS) and treatment-free survival (TFS) predictor.

PATIENTS AND METHODS

A training/validation strategy was chosen to find the optimal CD49d cutoff. The hazard ratio (HR) for death and treatment imposed by CD49d was estimated by pooled analysis of 2,972 CLLs; Cox analysis stratified by center and stage was used to adjust for confounding variables. The importance of CD49d over other flow cytometry-based prognosticators (eg, CD38, ZAP-70) was ranked by recursive partitioning.

RESULTS

Patients with ≥ 30% of neoplastic cells expressing CD49d were considered CD49d+. Decrease in OS at 5 and 10 years among CD49d+ patients was 7% and 23% (decrease in TFS, 26% and 25%, respectively). Pooled HR of CD49d for OS was 2.5 (2.3 for TFS) in univariate analysis. This HR remained significant and of similar magnitude (HR, 2.0) in a Cox model adjusted for clinical and biologic prognosticators. Hierarchic trees including all patients or restricted to those with early-stage disease or those age ≤ 65 years always selected CD49d as the most important flow cytometry-based biomarker, with negligible additional prognostic information added by CD38 or ZAP-70. Consistently, by bivariate analysis, CD49d reliably identified patient subsets with poorer outcome independent of CD38 and ZAP-70.

CONCLUSION

In this analysis of approximately 3,000 patients, CD49d emerged as the strongest flow cytometry-based predictor of OS and TFS in CLL.

Multiple imputation for handling systematically missing confounders in meta-analysis of individual participant data

A variable is 'systematically missing' if it is missing for all individuals within particular studies in an individual participant data meta-analysis. When a systematically missing variable is a potential confounder in observational epidemiology, standard methods either fail to adjust the exposure-disease association for the potential confounder or exclude studies where it is missing. We propose a new approach to adjust for systematically missing confounders based on multiple imputation by chained equations. Systematically missing data are imputed via multilevel regression models that allow for heterogeneity between studies. A simulation study compares various choices of imputation model. An illustration is given using data from eight studies estimating the association between carotid intima media thickness and subsequent risk of cardiovascular events. Results are compared with standard methods and also with an extension of a published method that exploits the relationship between fully adjusted and partially adjusted estimated effects through a multivariate random effects meta-analysis model. We conclude that multiple imputation provides a practicable approach that can handle arbitrary patterns of systematic missingness. Bias is reduced by including sufficient between-study random effects in the imputation model.

A framework for developing, implementing, and evaluating clinical prediction models in an individual participant data meta-analysis

The use of individual participant data (IPD) from multiple studies is an increasingly popular approach when developing a multivariable risk prediction model. Corresponding datasets, however, typically differ in important aspects, such as baseline risk. This has driven the adoption of meta-analytical approaches for appropriately dealing with heterogeneity between study populations. Although these approaches provide an averaged prediction model across all studies, little guidance exists about how to apply or validate this model to new individuals or study populations outside the derivation data. We consider several approaches to develop a multivariable logistic regression model from an IPD meta-analysis (IPD-MA) with potential between-study heterogeneity. We also propose strategies for choosing a valid model intercept for when the model is to be validated or applied to new individuals or study populations. These strategies can be implemented by the IPD-MA developers or future model validators. Finally, we show how model generalizability can be evaluated when external validation data are lacking using internal-external cross-validation and extend our framework to count and time-to-event data. In an empirical evaluation, our results show how stratified estimation allows study-specific model intercepts, which can then inform the intercept to be used when applying the model in practice, even to a population not represented by included studies. In summary, our framework allows the development (through stratified estimation), implementation in new individuals (through focused intercept choice), and evaluation (through internal-external validation) of a single, integrated prediction model from an IPD-MA in order to achieve improved model performance and generalizability.

Meta-analysis of two-arm studies: Modeling the intervention effect from survival probabilities

Pooling the hazard ratios is not always feasible in meta-analyses of two-arm survival studies, because the measure of the intervention effect is not systematically reported. An alternative approach proposed by Moodie et al. is to use the survival probabilities of the included studies, all collected at a single point in time: the intervention effect is then summarised as the pooled ratio of the logarithm of survival probabilities (which is an estimator of the hazard ratios when hazards are proportional). In this article, we propose a generalization of this method. By using survival probabilities at several points in time, this generalization allows a flexible modeling of the intervention over time. The method is applicable to partially proportional hazards models, with the advantage of not requiring the specification of the baseline survival. As in Moodie et al.'s method, the study-level factors modifying the survival functions can be ignored as long as they do not modify the intervention effect. The procedures of estimation are presented for fixed and random effects models. Two illustrative examples are presented.

Work-related violence and incident use of psychotropics

Although the mental health consequences of domestic violence are well documented, empirical evidence is scarce regarding the mental health effects of violence in the workplace. Most studies have used data from small occupation-specific samples, limiting their generalizability. This article examines whether direct exposure to work-related violence is associated with clinically pertinent mental health problems, measured by purchases of psychotropics (antidepressants, anxiolytics, hypnotics), in a cross-occupational sample of 15,246 Danish employees free from using psychotropics at baseline. Self-reported data on work-related violence were merged with other data on purchases of medications through a national registry to estimate cause-specific hazard ratios during 3.6 years (1,325 days) of follow-up in the years 1996-2008. Outcomes were examined as competing risks, and analyses were adjusted for gender, age, cohabitation, education, income, social support from colleagues, social support from supervisor, and influence and quantitative demands at work. Work-related violence was associated with purchasing antidepressants alone (hazard ratio = 1.38, 95% confidence interval: 1.09, 1.75) or in combination with anxiolytics (hazard ratio = 1.74, 95% confidence interval: 1.13, 2.70) but not with purchasing anxiolytics or hypnotics only. The frequency of violent episodes and risk of caseness were unrelated. Work-related violence is associated with increased risk of clinically pertinent mental health problems. Reducing levels of work-related violence may help to prevent mental disorders in the working population.

Thalidomide for previously untreated elderly patients with multiple myeloma: meta-analysis of 1685 individual patient data from 6 randomized clinical trials

The role of thalidomide for previously untreated elderly patients with multiple myeloma remains unclear. Six randomized controlled trials, launched in or after 2000, compared melphalan and prednisone alone (MP) and with thalidomide (MPT). The effect on overall survival (OS) varied across trials. We carried out a meta-analysis of the 1685 individual patients in these trials. The primary endpoint was OS, and progression-free survival (PFS) and 1-year response rates were secondary endpoints. There was a highly significant benefit to OS from adding thalidomide to MP (hazard ratio = 0.83; 95% confidence interval 0.73-0.94, P = .004), representing increased median OS time of 6.6 months, from 32.7 months (MP) to 39.3 months (MPT). The thalidomide regimen was also associated with superior PFS (hazard ratio = 0.68, 95% confidence interval 0.61-0.76, P < .0001) and better 1-year response rates (partial response or better was 59% on MPT and 37% on MP). Although the trials differed in terms of patient baseline characteristics and thalidomide regimens, there was no evidence that treatment affected OS differently according to levels of the prognostic factors. We conclude that thalidomide added to MP improves OS and PFS in previously untreated elderly patients with multiple myeloma, extending the median survival time by on average 20%.

Gait speed and survival in older adults

CONTEXT

Survival estimates help individualize goals of care for geriatric patients, but life tables fail to account for the great variability in survival. Physical performance measures, such as gait speed, might help account for variability, allowing clinicians to make more individualized estimates.

OBJECTIVE

To evaluate the relationship between gait speed and survival.

DESIGN, SETTING, AND PARTICIPANTS

Pooled analysis of 9 cohort studies (collected between 1986 and 2000), using individual data from 34,485 community-dwelling older adults aged 65 years or older with baseline gait speed data, followed up for 6 to 21 years. Participants were a mean (SD) age of 73.5 (5.9) years; 59.6%, women; and 79.8%, white; and had a mean (SD) gait speed of 0.92 (0.27) m/s.

MAIN OUTCOME MEASURES

Survival rates and life expectancy.

RESULTS

There were 17,528 deaths; the overall 5-year survival rate was 84.8% (confidence interval [CI], 79.6%-88.8%) and 10-year survival rate was 59.7% (95% CI, 46.5%-70.6%). Gait speed was associated with survival in all studies (pooled hazard ratio per 0.1 m/s, 0.88; 95% CI, 0.87-0.90; P < .001). Survival increased across the full range of gait speeds, with significant increments per 0.1 m/s. At age 75, predicted 10-year survival across the range of gait speeds ranged from 19% to 87% in men and from 35% to 91% in women. Predicted survival based on age, sex, and gait speed was as accurate as predicted based on age, sex, use of mobility aids, and self-reported function or as age, sex, chronic conditions, smoking history, blood pressure, body mass index, and hospitalization.

CONCLUSION

In this pooled analysis of individual data from 9 selected cohorts, gait speed was associated with survival in older adults.