TAM has proven to benefit all groups of patients with breast cancer and hormone sensitive tumours in the adjuvant setting and has a role in the prevention of breast cancer in several situations. The present study was aimed to investigate the role of TAM in patients operated for breast cancer who had never received TAM and it was powered to detect an effect in the order of that ascertained in the adjuvant setting.
The eligible population consisted partly of premenopausal patients who, several years before, did not receive TAM either because of a low risk of relapse or because they had received oophorectomy, chemotherapy or both. It should be noted that the acceptance of TAM as a standard therapy for premenopausal patients with hormone sensitive tumour was slower in Italy than in other countries and that, as a consequence, a certain population of patients diagnosed and treated in the Eighties and early Nineties who never received TAM according to the present indications exists. In addition, both pre- and postmenopausal patients who did not receive TAM either because their tumour was hormone insensitive or for other reasons, were eligible for the study. One hundred and sixty-four patients had an unknown ER status. Efforts were made to retrieve the information whenever possible, but some data of patients who had been operated elsewhere a long time before are missing. This may reflect the situation in contexts where it is less likely that the patients received tamoxifen as appropriate after primary treatment. Obviously, the lack of ER evaluation in a relevant proportion of the study population represents a major limitation of the study.
At this 5-year analysis, few events (n = 83) occurred and only a minority of these (n = 30) were of metastatic nature. This was predictable, taking into account the nature of the study in which the selection mechanism tended to exclude from randomization biologically aggressive tumours with a high propensity to early relapse. In addition, the majority of patients had received active adjuvant treatment, which may have contributed to the low number of events making difficult to ascertain the role of tamoxifen.
Some differences in outcome were detected in the two groups.
A statistically significant smaller number (1 vs 10) of ER-positive contralateral breast cancers, a trend towards fewer contralateral breast cancers (4 vs 10) and to a longer event-free survival in ER-positive cases were noted in the TAM-treated group. Contralateral ER-negative breast cancer occurred more frequently in the TAM group, supporting the inability of TAM to prevent ER-negative secondary tumours as previously described .
TAM was basically well tolerated and no serious adverse events occurred. The toxicity encountered was mostly related to the hormonal effects of TAM. Only one case of endometrial cancer occurred in a TAM-treated patient.
In the only published study with a design and size comparable to that of the present study, Delozier et al  noted 109 events in a population of 494 randomized patients followed-up for 10 years. An 83% 10-year disease-free survival in TAM- treated patients was reported, as compared to 75% in controls (p = 0.01). No difference in overall survival in the whole population was noted, but subgroups with node-positive or ER-positive disease had a better survival with TAM. Different from our study, TAM was planned to be administered continuously lifelong at the dose of 30 mg/day.
In a smaller study , 2 years of TAM had no influence upon disease-free survival, although there were more deaths (mostly unrelated to breast cancer) in the placebo group.
Delozier's data bring into question the adequacy of a 2-year treatment with TAM as compared to a standard 5-year treatment or longer. When designing our study, we were particularly worried by the carcinogenic effects of TAM, especially in this population of patients with a high likelihood of permanent cure, and decided to limit the treatment duration to 2 years, which had shown an effect in previous randomized studies as demonstrated both in the 1992 and the 1998 Overviews [15, 16], with a respective 27% and 24% reduction of recurrences. Subsequent 2005 Overview data  continued to indicate a 21% reduction in the risk of recurrence (26% in ER-positive, 11% in ER-poor cases) following 1-2 years of adjuvant TAM. In a recent study , 2 years of TAM was able to halve the risk of contralateral breast cancer in premenopausal women of all ages, although the effect was more evident in women younger than 40 years. Interestingly, the protective effect of 2 years of TAM was persistent during the whole follow-up period (median follow-up, 14 years).
Regarding the use of TAM in ER-negative tumours, although the 1998 Overview  showed a beneficial effect of TAM, subsequent reports [5, 18] indicated a potential deleterious effect of 5 years of TAM. This has not emerged in this study, where TAM was used in a different setting.
At the time this study was designed, the effect of TAM on contralateral breast cancer appeared to be independent of ER status [15, 16]. Subsequent studies indicated that its effect is limited to, or prevalent in women who originally had ER-positive breast cancer . In our study, the results thus far are inconclusive on this issue.