Background: The 1998 St Gallen classification was devised to guide clinicians in the use of adjuvant systemic therapy for women with early breast cancer. In this study, the classification was applied to a historical group of patients with node-negative breast cancer who were treated without adjuvant therapy. Methods: The St Gallen classification was applied to 421 women with breast cancer treated with conservative surgery and radiotherapy alone between 1979 and 1994. Primary tumour characteristics were reviewed centrally. Results: When the most stringent version of the St Gallen classification was applied (grade 2 or 3 tumours classified as 'high risk'), only 10 per cent of patients were 'low risk', with a 10-year distant relapse-free survival (DRFS) rate of 100 per cent, and 15 per cent were at 'intermediate risk' (10-year DRFS rate of 94 per cent). The high-risk group (75 per cent of women) had a 10-year DRFS rate of 77 per cent (P < 0.01). If the St Gallen classification had been applied to all patients in this series who were aged less than 70 years, up to 91 per cent would have been recommended to have chemotherapy. Conclusion: The St Gallen classification is an inaccurate measure of prognosis for patients with node-negative breast cancer and should be used with caution.