Peri-operative anaemia is a common problem in patients undergoing cardiac surgery. Postoperative anaemia is not well understood relative to pre-operative anaemia; limited data exist on haemoglobin recovery and mortality after discharge, especially in the era of restrictive transfusion practice. We aimed to investigate the associations of pre-operative and persistent postoperative anaemia with 1-year mortality in patients undergoing valvular heart surgery. We identified patients who had undergone valvular heart surgery and allocated them to one of four groups based on their pre-operative (haemoglobin ≥ 130 g.l-1 and < 130 g.l-1 in men and ≥ 120 g.l-1 and < 120 g.l-1 in women) and postoperative (measured 2 months after surgery; haemoglobin ≥ 100 g.l-1 and < 100 g.l-1 in both men and women) anaemia status. The four groups were: pre- and postoperative non-anaemia (non-anaemia-non-anaemia); pre-operative anaemia-postoperative non-anaemia (anaemia-non-anaemia); pre- and postoperative anaemia (anaemia-anaemia); and pre-operative non-anaemia-postoperative anaemia (non-anaemia-anaemia). The primary outcome was 1-year mortality. Data from 2486 patients were included. Pre-operative anaemia was diagnosed in 1107 patients (44.5%) and 279 (11.9%) met the diagnostic criteria for persistent anaemia 2 months postoperatively. The overall 1-year mortality rate was 3.3%. The highest rate was observed in the anaemia-anaemia group (17.8%), followed by the non-anaemia-anaemia (13.1%), anaemia-non-anaemia (2.9%) and non-anaemia-non-anaemia (0.5%) groups. Multivariable Cox regression analysis showed that the non-anaemia-anaemia group had the highest risk of 1-year mortality (adjusted hazard ratio 14.44, 95%CI 4.88-42.69), followed by the anaemia-anaemia group (adjusted hazard ratio 10.94, 95%CI 4.41-27.16). Our study highlights the high prevalence of persistent anaemia following valvular heart surgery. Persistent anaemia 2 months postoperatively is associated with an increased risk of 1-year mortality.