Consistent Lack of Association Between Breast Cancer and Oral Contraceptives Using Either Hospital or Neighborhood Controls

Alfredo Morabia, Moyses Szklo, W. Stewart, L. Schuman, D. B. Thomas

Research output: Contribution to journalArticle

Abstract

Background. Case-control studies of oral contraceptive use and breast cancer have used neighborhood, population, or hospital controls. Methods. To determine whether this association differs according to type of controls, interview data from 131 incident cases of breast cancer were compared with those of 189 hospital controls and 182 neighborhood controls Study subjects were white females recruited between 1973 and 1975 from among residents of Baltimore City and County ages 18 to 59 years. Results. Adjusted relative odds of breast cancer related to ever versus never use of oral contraceptives were 1.0 and 0.8, using hospital and neighborhood controls, respectively. Relative odds did not increase proportionally to duration of oral contraceptive use or to progestogen potency. Although few subjects had used oral contraceptives for more than 2 years, almost all pill brands contained more than 49 μg of ethinyl estradiol or of mestranol. Conclusions. Results from the present study do not support the hypothesis that early preparations of oral contraceptives increased breast cancer risk among white women, regardless of whether controls are neighbors of the cases or hospital patients. However, its conclusions cannot be generalized to women who began taking the pill before their first full-term birth or took it for more than 2 years.

Original languageEnglish (US)
Pages (from-to)178-186
Number of pages9
JournalPreventive Medicine
Volume22
Issue number2
DOIs
StatePublished - Mar 1993

ASJC Scopus subject areas

  • Epidemiology
  • Public Health, Environmental and Occupational Health

Fingerprint Dive into the research topics of 'Consistent Lack of Association Between Breast Cancer and Oral Contraceptives Using Either Hospital or Neighborhood Controls'. Together they form a unique fingerprint.

Cite this