We evaluated the correlation between histologic and cytologic specimens in the determination of estrogen receptor (ER) and progesterone receptor (PR) status in breast carcinoma and investigated the causes of clinically significant discrepancies. We analyzed 70 immunoassays for ER and 60 for PR from 71 patients with breast carcinoma. Concordance between cytology and histology was 89% for ER and 63% for PR using scores from pathology reports. Concordance between cytology and histology was 98% for ER and 91% for PR using consensus scores (obtained after reevaluation by the team of pathologists). Thirty of 130 (23%) tests had clinically relevant discrepancies, 53% of which were caused by wrong interpretation of cytologic findings, 10% by wrong interpretation of histologic findings, 17% by sampling error and 20% were not available for reevaluation. Wrong interpretation of the results for ER and PR status in cytology was a far more frequent cause of clinically relevant discrepancies than sampling errors. The use of strict criteria is recommended. Diagn. Cytopathol. 2002;26:137–141; DOI 10.1002/dc.10079 © 2002 Wiley-Liss, Inc.