Have we been underestimating the relative risk of breast cancer after HRT use?


Year:

Session type:

Ailie Grzybek, Colin McCowan, Frank Sullivan, Alastair Thompson, John Dewar

University of Dundee, Dundee, UK

Abstract

The Million Women Study (MWS) reported an 18-20/1000 increase in breast cancer risk per annum with oestrogen/progestagen hormone replacement therapy (HRT) use. Subsequent evidence has demonstrated a reduction in breast cancers following decreased HRT use. This study calculated the expected (MWS) and observed incidence of breast cancer using encashed HRT prescription records for a population served by a single regional cancer centre.

A retrospective population cohort study of encashed prescription data for all women aged 50-64 years who commenced HRT and received more than one months prescription between January 1992 and December 2003 was linked to breast cancer diagnosis. Women were classed as HRT users if they had ever used HRT.

387 cancers (mean age 54.4 years) were diagnosed in 18,308 women (median follow up 7.7 years; mean 4.99 years of HRT use). Almost 70% of patients used an oestrogen/progestagen HRT preparation, 26% oestrogen only and 5% tibolone. There were 282 breast cancers expected in the study population, calculated using the attributable fraction based on MWS relative risks for each HRT preparation. This was significantly lower than the 387 cancers observed in this cohort (Chi square 89.73; df 5; P<0.001), although the MWS predictions were more accurate in women aged 55-69.

The MWS relative risk underestimated the risk of developing breast cancer. While the difference may be due to our classification of women as HRT users irrespective of when they stopped or the longer follow up in the present study, this data sounds a further, cautionary, note.

The Million Women Study (MWS) reported an 18-20/1000 increase in breast cancer risk per annum with oestrogen/progestagen hormone replacement therapy (HRT) use. Subsequent evidence has demonstrated a reduction in breast cancers following decreased HRT use. This study calculated the expected (MWS) and observed incidence of breast cancer using encashed HRT prescription records for a population served by a single regional cancer centre.

A retrospective population cohort study of encashed prescription data for all women aged 50-64 years who commenced HRT and received more than one months prescription between January 1992 and December 2003 was linked to breast cancer diagnosis. Women were classed as HRT users if they had ever used HRT.

387 cancers (mean age 54.4 years) were diagnosed in 18,308 women (median follow up 7.7 years; mean 4.99 years of HRT use). Almost 70% of patients used an oestrogen/progestagen HRT preparation, 26% oestrogen only and 5% tibolone. There were 282 breast cancers expected in the study population, calculated using the attributable fraction based on MWS relative risks for each HRT preparation. This was significantly lower than the 387 cancers observed in this cohort (Chi square 89.73; df 5; P<0.001), although the MWS predictions were more accurate in women aged 55-69.

The MWS relative risk underestimated the risk of developing breast cancer. While the difference may be due to our classification of women as HRT users irrespective of when they stopped or the longer follow up in the present study, this data sounds a further, cautionary, note.