OlympiA
In high-risk HER2-negative early breast cancer with germline BRCA1/2 pathogenic variants, does adjuvant olaparib improve outcomes after local treatment and chemotherapy?
Trial design and population
1,836 patients with germline BRCA1/2-mutated HER2-negative early breast cancer at high risk after chemotherapy and definitive local treatment.
Olaparib 300 mg orally twice daily for 1 year vs placebo.
1,836
Invasive DFS; OS key secondary endpoint
Key results
At 4 years, invasive DFS was 82.7% vs 75.4% (HR 0.63, 95% CI 0.51–0.78); distant DFS was 86.5% vs 79.1%. Four-year OS was 89.8% vs 86.4% (HR 0.68, 98.5% CI 0.47–0.97).
Source-reported outcome visualization
4-year overall survival · 4 years
Source-reported endpoint and unit; median survival is not converted into an NNT.
Absolute Effect
Absolute 4-year OS gain 3.4 points; NNT 30 based on 4-year OS. Four-year invasive DFS gain 7.3 points; NNT 14 based on 4-year invasive DFS.
NNT: 30 (4-year OS); 14 (4-year invasive DFS). Endpoint/time horizon are stated above; do not convert median survival differences into NNT.
NNH: Not quantifiable. Anaemia, fatigue, nausea and dose interruptions are important; rare MDS/AML is a serious labelled risk. NNH not calculated without a matched fixed-time excess-harm endpoint.
Safety and treatment burden
Anaemia, fatigue, nausea and dose interruptions are important; rare MDS/AML is a serious labelled risk. NNH not calculated without a matched fixed-time excess-harm endpoint.
Follow-up, maturity and limitations
Updated analysis with approximately 3.5 years median follow-up; 4-year estimates reported.
Requires germline BRCA1/2 status, HER2-negative disease and high-risk clinicopathologic eligibility; do not extrapolate to unselected disease.
Clinical interpretation
Current-practice context · reviewed 20 October 2026
Historical evidence is separated from current sequencing. Confirm local regulatory status, current guidelines, biomarkers, prior therapies and patient-specific contraindications before applying these results. Newer treatments may have changed the preferred sequence; where NNT/NNH cannot be defensibly quantified, no surrogate calculation is made.