The objective of this study was to define prognostic serum biomarkers that could serve as surrogate survival endpoints during second-line treatment for advanced pancreatic cancer. This retrospective single-center study included patients treated with second-line therapy for advanced exocrine pancreatic cancer. A pretreatment value and at least one serial measurement during the first two cycles of second-line chemotherapy for CA 19-9, CEA, CRP, and LDH had to be available in order to evaluate the prognostic role of kinetics on overall survival. A cutoff of a >20% increase from baseline during treatment was defined in order to form groups with suspected different outcomes. The effect of serial biomarker changes on survival was modeled by Cox proportional hazards regression in univariate and multivariate analyses. Overall, 70 patients treated with second-line therapy for advanced disease were included; 94% had distant metastases at treatment initiation. Median time to progression was 2.7 months and median survival 5.4 months. Univariate analysis found that an increase of >20% during treatment was significantly associated with a worse overall survival for CA 19-9 (HR 2.00, p = 0.018), CEA (HR 2.38, p = 0.004), and CRP (HR 3.06, p < 0.001). These associations remained significant within multivariate analysis for CEA (HR 2.86, p = 0.001) and CRP (HR 3.20, p = 0.001). Serum biomarker kinetics might serve as useful prognostic tools during second-line chemotherapy in advanced pancreatic cancer.