Clinical Trial Service Unit and Epidemiological Studies Unit, Radcliffe Infirmary, Oxford OX2 6HE.

• Sarah Darby professor
• Paul McGale statistician
• Richard Peto codirector
• Department of Medical Epidemiology, Karolinska Institute, 171 77 Stockholm, Sweden
• Fredrik Granath statistician
• Per Hall professor
• Department of Medicine, Karolinska Institute
• Anders Ekbom professor

Correspondence to: S Darby sarah.darby@ctsu.ox.ac.uk

During radiotherapy for breast cancer, there is often some irradiation of the heart and major blood vessels, which could increase cardiovascular mortality many years later.³ The dose of radiation to the heart is generally higher when the left rather than the right breast is affected. Therefore, indirect evidence on the magnitude of any risk is available where the tumour laterality (left or right breast) can be linked to subsequent cardiovascular mortality.^{1 2} Studies of the survivors of the atomic bombing of Japan who received single doses to the whole body of 0-4 Gy show that the cardiovascular disease risk is dose related and increases by about 14% per gray.⁴

Since 1970, the nationwide Swedish cancer registry has recorded the laterality of breast cancers but not the use of radiotherapy. Unpublished data from regional Swedish registries suggest that about 30% of women with early breast cancer during the 1970s and early 80s received radiotherapy. We linked registry records (1970-96) with national mortality records. The study was approved by the ethics committee of the Karolinska Institute.

After we excluded women whose cancer was diagnosed at autopsy or outside Sweden and those with previously registered cancers (except squamous cell skin cancer), 89 407 women aged 18-79 with unilateral breast cancer remained. We stratified analyses of subsequent mortality in groups of five years by calendar year of diagnosis, time since diagnosis, and age at diagnosis. Stratification by age was necessary because the proportion of left sided tumours increases with age.⁵ Each womans contribution to the person years at risk ran from the date of diagnosis until her date of death, date of emigration, 100th birthday, or 1 January 1997, whichever was earliest. We used Poisson regression to calculate mortality ratios, left versus right, from the numbers of deaths and person years. Ratios greater than one indicate greater mortality in women with left sided tumours than in women with right sided tumours.

Mortality from breast cancer was identical in women with left sided or right sided tumours (table). Mortality from cardiovascular diseases was higher in women with left sided tumours. Little excess occurred in the first 10 years after diagnosis (mortality ratio 1.01; 95% confidence interval 0.96 to 1.07), but later the ratio was 1.10 (1.03 to 1.18; P=0.004), 1.13 (1.03 to 1.25; P=0.01) for ischaemic heart disease (half of all cardiovascular mortality), and 1.08 (0.98 to 1.18) for other cardiovascular deaths (about 30% of which probably involved heart disease). For the remaining causes, mortality in women with left sided tumours did not differ significantly from that in women with right sided tumours.

Mortality ratio for women with left sided breast cancer versus women with right sided breast cancer during and after the first 10 years from diagnosis of breast cancer among 89 407 women registered during 1970-96 at the Swedish cancer registry

ICD-9=International classification of diseases, ninth revision.

*P=0.04. †P=0.004. †P=0.05. §P=0.01.

Most of the late cardiovascular deaths involved women treated for breast cancer in the 1970s, and improvements in radiotherapy techniques since then have tended to reduce radiation dose to the heart. For women treated in the 1980s, however, the cardiovascular ratio, left versus right, was still 1.11 but with a wide 95% confidence interval (0.95 to 1.29).

Comment

A mortality ratio, left versus right, of 1.10 for cardiovascular disease more than 10 years after diagnosis of breast cancer is compatible with a substantial hazard among some of those actually irradiated. For example, if about 30% of women surviving 10 years after breast cancer had been irradiated then a cardiovascular mortality ratio of 1.10 in all women and 1.00 in unirradiated women would suggest a ratio of 1.33 in those irradiated. This could be produced by a 60% increase in late cardiovascular mortality after irradiation for a left sided tumour and a 20% increase after irradiation for a right sided tumour. The confidence interval for the observed ratio of 1.10 is, however, wide, so the true cardiovascular hazard from radiotherapy in the 1970s and 80s remains uncertain.

Competing interests: None declared.

Sarah Darby, Paul McGale, Richard Peto, Fredrik Granath, Per Hall and Anders Ekbom

1. Paszat L, Mackillop WJ, Groome PA, Boyd C, Schulze K, Holowaty E. Mortality from myocardial infarction after adjuvant radiotherapy for breast cancer in the surveillance, epidemiology and end-results cancer registries. J Clin Oncol1998;16:2625-31.
2. Rutqvist LE, Johansson H. Mortality by laterality of the primary tumour among 55 000 breast cancer patients from the Swedish Cancer Registry. Br J Cancer 1990;61:866-8.
3. Early breast cancer trialists collaborative group (EBCTCG). Favourable and unfavourable effects on long-term survival of radiotherapy for early breast cancer: an overview of the randomised trials. Lancet 2000;355:1757-70.
4. Shimizu Y, Pierce DA, Preston DL, Mabuchi K. Studies of the mortality of atomic bomb survivors: non-cancer mortality 1950-1990. Radiat Res 1999;152:374-89. (Report 12, part 11.)
5. Weiss HA, Devesa SS, Brinton LA. Laterality of breast cancer in the United States. Cancer Causes Control 1996;7:539-43.