Despite the efficacy of gemcitabine-platinum regimen, the outcome of patients with recurrent and/or metastatic nasopharyngeal carcinoma (RM NPC) is poor. A phase II trial was conducted to determine the efficacy of erlotinib, given as maintenance therapy after gemcitabine-platinum chemotherapy in patients with RM NPC. Patients were treated with gemcitabine 1000 mg/m on days 1 and 8 as well as cisplatin 70 mg/m on day 1 (or carboplatin area under curve 5 on day 1, if contraindication to cisplatin) 3 weeks. After 6 chemotherapy cycles (or before in case of progression), patients were switched to erlotinib 150 mg/d 4 weeks. Primary end point was time to progression in patients without progressive disease after 6 chemotherapy cycles and treated with maintenance erlotinib. Epstein-Barr virus DNA plasma levels, measured using quantitative real-time polymerase chain reaction, were correlated with outcome. Of 20 enrolled patients, 19 patients received 96 chemotherapy cycles. Fifteen patients were switched to erlotinib and received 36 cycles (range: 1 to 6 cycles). Safety profiles observed with the chemotherapy combination and erlotinib were those expected. Of 12 patients evaluable for response to erlotinib, all progressed except 3 patients (25%) who had stable disease for 3, 4, and 7 months, respectively. Median time to progression was 6.9 months for 13 patients without progressive disease after 6 chemotherapy cycles and treated with erlotinib. No correlation was identified between Epstein-Barr virus DNA plasma levels and clinical outcome. Maintenance or second-line therapy with erlotinib after chemotherapy was not effective in RM NPC. Historical comparison with patients treated with the same chemotherapy alone until progression suggests that it may be detrimental to stop chemotherapy after 6 cycles if disease did not progress.