We evaluated the efficacy and tolerability of a dual therapy with rabeprazole and amoxicillin (AMX) as an empiric third-line rescue therapy. In patients with failure of first-line treatment with a proton pump inhibitor (PPI)- AMX-clarithromycin regimen and second-line treatment with the PPI-AMX-metronidazole regimen, a third-line eradication regimen with rabeprazole (10 mg q.i.d.) and AMX (500 mg q.i.d.) was prescribed for 2 wk. Eradication was confirmed by the results of the 13C-urea breath test (UBT) at 12 wk after the therapy. A total of 46 patients were included; however, two were lost to followup. The eradication rates as determined by per-protocol and intention-to-treat analyses were 65.9% and 63.0%,respectively. The pretreatment UBT results in the subjects showing eradication failure; those patients showing successful eradication comprised 32.9 ± 28.8 permil and 14.8 ± 12.8 permil, respectively. The pretreatment UBT results in the subjects with eradication failure were significantly higher than those in the patients with successful eradication (P = 0.019). A low pretreatment UBT result (≤ 28.5 permil) predicted the success of the eradication therapy with a positive predictive value of 81.3% and a sensitivity of 89.7%. Adverse effects were reported in 18.2% of the patients, mainly diarrhea and stomatitis. Dual therapy with rabeprazole and AMX appears to serve as a potential empirical third-line strategy for patients with low values on pretreatment UBT. In patients with failure of first-line treatment with a proton pump inhibitor (PPI) - AMX-clarithromycin regimen and second -line treatment with the PPI-AMX-metronidazole regimen, a third-line eradication regimen with rabeprazole (10 mg qid) and AMX (500 mg qid) was prescribed for 2 weeks. Eradication was confirmed by the results of the 13C-urea breath A total of 46 patients were included; however, two were lost to follow. The eradication rates as determined by per-protocol and intention-to-treat analyzes were 65.9% and 63.0%, respectively respectively. The pretreatment UBT results in the subjects showing eradication failure; those patients showing successful eradication failure 32.9 ± 28.8 permil and 14.8 ± 12.8 permil, respectively. The pretreatment UBT results in the subjects with eradication failure were signifi cantly higher than those in the patients with successful eradication (P = 0.019). A low pretreatment UBT result (≤ 28.5 permil) predicted the success of the eradication therapy with a positive predictive value of 81.3% and a sensitivity of 89.7%. were reported in 18.2% of the patients, mainly diarrhea and stomatitis. Dual therapy with rabeprazole and AMX appears to serve as a potential empirical third-line strategy for patients with low values ​​on pre treatment UBT.