Background: As a risk classification system of metastatic germ cell tumors, the International Germ Cell Consensus (IGCC) classification was proposed in 1997 and has received broad approval. We aimed to clarify the significance of the IGCC classification in the 2000s especially in intermediate and poor-prognosis patients in Japan. Methods: We analyzed 118 patients with intermediate- and poor-prognosis metastatic non-seminomatous germ cell testicular cancer treated at five university hospitals in Japan between 2000 and 2010. Data were collected on age, levels of serum AFP, HCG and LDH, lung metastases, spread to non-pulmonary visceral metastases (NPVM) and on treatment details (first, second and third-line chemotherapy and post-chemotherapy surgery) and survival data. Results: The median age at diagnosis was 31 years (range, 2-54 years). The median follow-up period of all patients was 57 months. Sixty-eight patients were classified as poor prognosis, having LDH elevation in 14 (21%), AFP in 13 (19%), HCG in 27 (40%), presence of NPVM in 44 (65%). As first line chemotherapy, 93 (79%) were treated by BEP (bleomycin, etoposide and cisplatin)-containing regimen. Of 118 patients, 75 (64%) received second-line chemotherapy, in which the most frequently used regimen was taxane-containing regimen, including TIP (paclitaxel, ifosfamide and cisplatin), TIN (paxlitaxel, ifosfamide and nedaplatin) and DIN (docetaxel, ifosfamide and nedaplatin). Third-line chemotherapy was carried out in 33 patients (28%). Eighty-nine patients (75%) underwent post-chemotherapy surgery. The 5-year overall survivals for intermediate (n=50) and poor (n=68) prognosis was 89% and 83% (P=0.23), respectively. In poor prognosis patients, patients with more than 2 poor-prognostic factors had significantly worse survival than those with only one prognostic factor (72% vs 91%, P=0.01). Conclusions: There was a trend of increase in survival for any risk groups and, in particular, large increase in survival for patients with a poor prognosis. Further classification of poor-prognosis patients into two subgroups has a potential to identify a patient group with very poor-prognosis.