Inflammation is an enabling characteristic of the hallmarks of cancer. There has therefore been increasing interest in the clinical value of circulating inflammatory biomarkers in cancer. In this review, we summarize results on C-reactive protein (CRP), alone or as part of the Glasgow Prognostic Score (GPS, composed of CRP and serum albumin), as a biomarker of prognosis or prediction and monitoring of therapeutic response in patients with breast cancer. A systematic literature search was performed in Medline and Embase from 1990 to August 2021. The association of serum CRP and overall survival and disease/progression-free survival was summarized in meta-analyses using a random effects model. The results from a total of 35 included studies (20,936 patients) were divided according to three identified patient settings (metastatic, non-metastatic, and general setting). Most of the studies examined prognostic utility. Several larger studies observed associations between high serum CRP and poor survival, but the meta-analyses suggested a limited value in a non-metastatic and general breast cancer setting (populations with unknown or varied disease stage). In metastatic patients, however, more consistent findings supported an association between serum CRP and prognosis (hazard ratio for overall survival: 1.87 (95% CI 1.31-2.67). Only five studies examined a role in prediction or monitoring of therapeutic response. One study reported a significant association between serum CRP levels and response to chemotherapy. Findings regarding serum CRP as a biomarker in breast cancer appear inconsistent, particularly in non-metastatic and general breast cancer, where the prognostic value could not be confirmed. In patients with metastatic breast cancer we suggest that high serum CRP is an indicator of poor prognosis. Too few studies assessed the role of serum CRP in prediction or monitoring of treatment response to allow conclusions.