Hyper-estrogenism is more common in obese than in non-obese women. Con
sequently obesity has been shown to increase the risk of hormone depen
dent tumors. Some investigators have claimed that obesity at the time
of primary treatment may be an independent prognostic factor for breas
t cancer, but this issue is still controversial. Therefore, we conduct
ed a retrospective analysis to assess the influence of obesity at the
time of primary treatment on disease-free survival (DFS). Obesity was
defined as an excess of more than 25% of ideal weight according to Bro
ca's index ([Height (cm) -100]) 10%). The Cox-model was used for multi
variate analysis. Mean follow-up was 61 (range 6-126) months. 295 (62.
3%) patients were classified as of normal weight and 178 (37.6%) as ob
ese. Mean excess of ideal weight was 8.9 kilograms (kg) in premenopaus
al and 13.9 kg in postmenopausal patients (nonparametric t-test p<0.00
001). Patients with tumor size <20mm, 20-50mm and >50mm had a mean exc
ess of the real weight of 10.6kg, 12.5kg and 16.1kg, respectively (non
-parametric t-test p<0.0001). Percentual excess of real weight compare
d to ideal weight was 22.4 [+/-21.2] kg in patients without recurrence
and 21.5 [+/-21.9] in patients with recurrent disease (nonparametric
t-test p=0.7256). Univariate analysis revealed no significant associat
ion between obesity and the DFS. Multivariate analysis identified axil
lary lymph node involvement as the only statistically significant prog
nostic factor for disease-free survival (RR 1.55; 95%-confidence inter
val 1.02-2.36; p:0.0368). Because of the high correlations of node-sta
tus, tumor size and histological grading, the other factors failed to
be prognostically relevant in this analysis. Obesity was not found to
influence DFS of patients with primary breast cancer and is therefore
unlikely to constitute an independent prognostic factor. It may, howev
er, contribute to delayed diagnosis, since a significant proportion of
obese patients were diagnosed with local advanced disease.