Low hemoglobin concentration is associated with increased mortality, but there is disagreement with regard to the clinical definition of anemia. We aimed to evaluate the prevalence, clinical correlates and association with total and cause-specific long-term mortality across the hemoglobin distribution and for previously proposed definitions of anemia. Blood hemoglobin concentration and mean corpuscular volume was measured in participants of the Malmö diet and cancer study-a prospective cohort study, and related to baseline characteristics and outcomes during follow-up. Primary endpoints were all-cause mortality, cardiovascular mortality and cancer-related mortality. A U-shaped association of hemoglobin with total mortality was observed in spline regression analyses, with nadir at hemoglobin 150 g/L among men and 130 g/L among women. Mortality increased steeply with more strict definitions of anemia, hazard ratio: 1.36, 1.94 and 2.16 for hemoglobin <140/130 (men/women), 132/122 and 130/120 g/L, respectively. Similar trends were seen for both cancer- and cardiovascular mortality. The incidence of coronary disease and cancer did not differ across groups. Erythrocyte volume was an independent predictor of mortality, with the highest mortality observed for macrocytic anemia, which was less prevalent than microcytic and normocytic anemia. Dietary intake of iron and vitamin B12 were significantly lower and use of antithrombotic medications was significantly higher in subjects with anemia. The World Health Organisation definition of anemia was associated with increased mortality (hazard ratio 2.16) but excess mortality was also observed at higher hemoglobin levels. Of morphological subtypes, anemia with macrocytosis was rare but associated with the highest mortality.