Qualitative age interactions in breast cancer studies: A mini-review

Ismail Jatoi, William F. Anderson

Research output: Contribution to journalReview articlepeer-review

20 Scopus citations


A qualitative age interaction is defined as the reversal of relative risks or rates according to age at onset, and is often evident in studies that examine the etiology, prognosis and treatment of breast cancer. For example, incidence rates (or risks) are higher for aggressive when compared with indolent breast cancers prior to age 40-50 years, after which rates are higher for indolent tumors. Nulliparity and obesity decrease breast cancer risk in younger women, but increase risk in older women. Curves depicting the annual hazard of breast cancer death are shaped differently for the early- and late-onset tumors. Clinical trials for mammography screening, fenretinide chemoprevention and neo-adjuvant chemotherapy show opposite effects in younger and older women. Finally, high-risk/early onset breast cancers are more common among African-American women than Caucasian women, and this may partly account for the racial survival disparities. Taken together, these examples imply that aging may modify breast cancer risk, prognosis and treatment. These qualitative age interactions (or effect modifications) are important because they suggest that high-risk/early-onset and low-risk/late-onset breast cancers are different diseases, derived from different carcinogenic pathways. When age interactions are suspected, breast cancer studies should be stratified by early versus late age of onset or analyzed age specifically.

Original languageEnglish (US)
Pages (from-to)1781-1788
Number of pages8
JournalFuture Oncology
Issue number11
StatePublished - Nov 2010


  • age
  • breast cancer
  • early onset
  • interaction
  • late onset
  • qualitative

ASJC Scopus subject areas

  • Oncology
  • Cancer Research


Dive into the research topics of 'Qualitative age interactions in breast cancer studies: A mini-review'. Together they form a unique fingerprint.

Cite this