INTRODUCTION At least 15 lymph nodes should be retrieved for proper TNM-staging in gastric cancer. We evaluated nodal harvest and examined its relation to stage distribution and survival at a population-based level, including the value of N-ratio (metastatic/evaluated) as a staging modality. METHODS All patients resected for primary M0 gastric cancer diagnosed in 1999-2007 in the Dutch Eindhoven Cancer Registry area were included (N = 880). Determinants of lymph node evaluation and their relationship with stage and survival were assessed in multivariable regression analyses. N-ratio categories were determined (N-ratio 0, 0%; N-ratio 1, 0.1%-19%; N-ratio 2, 20%-29%; N-ratio 3, ≥30%) RESULTS The median number of lymph nodes examined was 7, dependent on N-category (N0: 7; N+: 8). It varied between departments of pathology from 5 to 9. This variation remained after adjustment for relevant patient- and tumour factors. Stage distribution differed between pathology departments (proportion N0 ranging from 14% to 21%, p = 0.003). Among resected patients with N0M0 disease and <7 nodes examined, 5-year survival was 56%, compared to 69% among patients with ≥7 nodes examined (p = 0.012). Five-year survival for N-ratio 0 was 58%, N-ratio 1 50%, N-ratio 2 18% and N-ratio 3 11% (p < 0.0001), while 5-year survival ranged from 58% for N0, 17% for N1, and 11% for N2/3 (p < 0.0001). CONCLUSION In this series of patients with a relatively low number of evaluated lymph nodes, a high prognostic accuracy of N-ratio was found. However, improvement in nodal assessment is mandatory.