Journal of the National Cancer Institute Advance Access originally published online on July 10, 2007
JNCI Journal of the National Cancer Institute 2007 99(14):1131; doi:10.1093/jnci/djm041
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author 2007. Published by Oxford University Press.
CORRESPONDENCE |
Response: Re: Pregnancies, Breastfeeding, and Breast Cancer Risk in the International BRCA1/2 Carrier Cohort Study (IBCCS)
Affiliations of authors: INSERM U794/Service de Biostatistiques de l’Institut Curie, Paris, France (NA); Department of Dermatology, University of Utah School of Medicine, Salt Lake City, UT (DEG); Cancer Research UK, Genetic Epidemiology Unit, Department of Public Health and Primary Care, University of Cambridge, UK (DFE, ACA); Department of Epidemiology, The Netherlands Cancer Institute, Amsterdam, Netherlands (MR, RB); Division of Epidemiology, German Cancer Research Center, Heidelberg, Germany (JCC)
Correspondence to: Nadine Andrieu, PhD, Inserm U794/Service de Biostatistiques, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (e-mail: nadine.andrieu{at}curie.net) or Jenny Chang-Claude, PhD, Division of Cancer Epidemiology, German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany (e-mail: j.chang-claude{at}dkfz-heidelberg.de).
Huiart and Sylvester were interested in one of the findings reported in our recent article, that a late age at first full-term pregnancy was associated with a reduction in the risk of breast cancer in the women with a mutation in the BRCA1 gene. Huiart and Sylvester cautioned that the observed reduction in risk may only be due to immortal time bias and noted that immortal time bias could be avoided by using "models in which the categories describing age at first pregnancy are modeled as time-dependent variables." They demonstrated their point using a simulated dataset. We appreciate having the opportunity to clarify the methodology of our study, which was perhaps not as transparent to some readers as we had assumed. We described in the statistical methods section of our paper that time-dependent analyses were, in fact, performed: "Parity and menopausal status changed over time; therefore, they were analyzed as time-dependent covariates to account for any potential interaction with time." We used parity as a generic term for all variables related to "having given birth" as defined in the Collins dictionary (1). It would be illogical to consider only the number of full-term pregnancies as time dependent but not the other related variables. In addition, we mentioned that "eighteen women who had missing values for age at pregnancy were excluded from the analysis," which we assumed would be an indication that the parity variables used were defined on the time scale. In particular, we modeled age at first pregnancy in the exact manner as Huiart and Sylvester in the time-dependent analysis of their simulated data.
The exercise performed by Huiart and Sylvester is illustrative but not relevant to our paper. There are several limitations in our study that may be reasons for caution when interpreting the findings, which we discussed extensively, but our results were not due to a statistical artifact.
REFERENCES
(1) Makins M, Adams D, eds. Collins English dictionary. (1992) 3rd ed. Glasgow (U.K.): HarperCollins.
Related Correspondence
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 1130-1131.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||