Article
Systematic review of treatment-related risk of secondary malignant neoplasms after Hodgkin lymphoma: issues in meta-analysis of time-to-rare-event data
Search Medline for
Authors
-
Jeremy Franklin
- Universität Köln, Köln, DE
-
Ingrid Kaul
- Universität Köln, Köln, DE
-
Dennis Eichenauer
- Universität Köln, Köln, DE
-
Andreas Engert
- Universität Köln, Köln, DE
| Published: | August 27, 2013 |
|---|
Outline
Text
Introduction and aim: Secondary malignant neoplasms (SMN) are a major late effect of treatment for Hodgkin lymphoma (HL). Single trials, even if large, have inadequate power to test for differences in SMN rates, while the many large-scale cohort-based studies of SMN after HL suffer from non-randomised, potentially biased comparisons of treatment strategies (see [1] for a discussion of issues in overviews of adverse events). Thus, the consequences of choice of first-line treatment for SMN risk remain unclear. A Cochrane systematic review addressing this question is being performed based on our previous such review in 2000-2004 [2].
Material and methods: Individual patient data (IPD) were collected from randomised controlled trials addressing 5 currently relevant study questions: additional radiotherapy (RT), RT field, RT dose, chemotherapy (CT) cycles, intensified CT. Incidence of SMN, overall survival (OS) and progression-free survival (PFS) are analysed meta-analytically. Due to small numbers of events, SMN is analysed using Peto’s (two-stage) method [3], [4], and also using a (one-stage) Cox proportional hazards regression model incorporating trial effects (in contrast, OS and PFS are analysed by two-stage inverse-variance synthesis of trial-specific hazard ratios from Cox regressions). Cumulative SMN incidences over time per treatment group are estimated, with confidence intervals constructed using Monte Carlo methods. Sensitivity analyses examine alternative methods for meta-analysis of rare events, such as the Mantel-Haenszel and exact methods.
Results: Data from 15 of the 21 eligible trials were obtained, including 9753 patients. For each study question, the number of trials (k), total number of patients (N), median length of follow-up and total number of SMN events (n) were: see Table 1 [Tab. 1].
The observed SMN frequencies in the eligible trials are mostly larger than the expected values based on published study results, which may be partly due to a longer median follow-up time in the updated IPD [5].
Discussion: The data collected for the present meta-analysis allow testing and estimation of treatment effects on SMN risk (as well as OS and PFS). First results will be presented and evaluated, including comparison of alternative meta-analytic methods.
References
- 1.
- Golder S, Loke YK, Bland M. Meta-analyses of adverse events data derived from randomised controlled trials as compared to observational studies: methodological overview. PLoS Med. 8(5): e1001026. DOI: 10.1371/journal.pmed.1001026
- 2.
- Franklin JG, Paus MD, Pluetschow A, Specht L. Chemotherapy, radiotherapy and combinedmodality for Hodgkin's disease, with emphasis on second cancer risk. Cochrane Database of Systematic Reviews. 2005;4. Art. No.: CD003187. DOI: 10.1002/14651858.CD003187.pub2.
- 3.
- CTSU. http://www.ctsu.ox.ac.uk/research/meta-trials/ebctcg/original-methods-for-ebctcg-meta-analyses
- 4.
- Bradburn MJ, Deeks JJ, Berlin JA, Russell Localio A. Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events. Stat Med. 2007; 26: 53-77
- 5.
- Franklin J, Kaul I, Eichenauer D, Engert E. Meta-analysis of treatment-related risk of second malignant neoplasms after Hodgkin lymphoma: updated design and progress. Ann Oncol. 2011;22(4):204.
