HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH REGISTER		Author Keyword(s) Vol Page [Advanced]

Electronic Letters to:

Original articles: M K Schubauer-Berigan, J A Deddens, K Steenland, W T Sanderson, and M R Petersen Adjustment for temporal confounders in a reanalysis of a case–control study of beryllium and lung cancer Occup Environ Med 2008; 65: 379-383 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Beryllium and Lung cancer: Study design artifacts

David C. Deubner, Paul S. Levy and H. Daniel Roth   (16 November 2007)

Beryllium and Lung cancer: Study design artifacts 16 November 2007
David C. Deubner, Occupational physician Brush Wellman Inc, Paul S. Levy and H. Daniel Roth Send letter to journal: Re: Beryllium and Lung cancer: Study design artifacts david_deubner{at}brushwellman.com David C. Deubner, et al.	To the editor In their recent article (1), Mary Schubauer-Berigan and colleagues elaborate on a prior study (2) in order to examine potential confounding by birth cohort and hire age of the relationship between beryllium exposure and lung cancer, a worthwhile objective. The prior study is a matched nested case-control study that employs incidence density sampling to match controls to cases and uses exposure lagging to take disease latency into consideration. However, its study design is unable to distinguish between null and alternate hypotheses owing to artifacts caused by imbalances in the matching of controls to cases on age of censor. These artifacts result in case-control differences in hire age and date of birth as well as in controls having an artificially high rate of truncated lagged exposure when compared to cases. This sequence of measurement error, in turn, produces artificially high lagged exposure odds ratios (3,4,5). Our simulations (unpublished) demonstrate that these study design artifacts are not completely reversed through analysis, for instance, by simply adding hire age to a conditional logistic regression model. An alternative approach is required, such as closer matching of cases and controls on age at censor (4). The current study (1) inherits the artifacts present in the prior study (2), as it uses the same cases, differentially matched controls, and calculation of artifacts in lagged exposure, with minor changes in the method for substituting small values for zero. The authors have not shown that the current study design eliminates or controls the inherent artifacts or that it corrects the inability of the prior study design to distinguish null and alternative hypotheses. 1) Schubauer-Berigan MK, Deddens JA, Steenland K et al. Adjustment for temporal confounders in a reanalysis of a case-control study of beryllium and lung cancer. Occup Environ Med doi:10.1136/oem.2007.033654. 2) Sanderson WT, Ward EM, Steenland L et al. Lung cancer case-control study of beryllium workers. Am J Ind Med 2001; 39:133-144. 3) Deubner DC, Roth HD, Levy PS. Empirical evaluation of complex study designs: workplace exposure and cancer. J Occup Environ Med 2007, 49:953-959. 4) Levy PS, Roth HD, Deubner DC. Exposure to beryllium and occurrence of lung cancer: a reexamination of findings from a nested case-control study. J Occup Environ Med 2007, 49:96-101 5) Levy P.S., Deubner D.C. and Levy P.S., Authors’ response to letter to the editor. J Occup Environ Med 2007, 49:707-711 Sincerely, David C. Deubner H. Daniel Roth Paul S. Levy David C. Deubner is an employee of Brush Wellman Inc, a manufacturer of beryllium materials H. Daniel Roth (Roth Associates, Rockville, Maryland, USA) and Paul S. Levy (Research Triangle Institute International, Research Triangle Park, North Carolina, USA) are supported in part through contracts, direct and indirect, with Brush Wellman Inc.