Objective: Recent studies suggest that subtle vitamin K depletion has far-reaching consequences. As this entity is not associated with prothrombin time elevation, it is important to determine whether alternate methods can help identify it. We investigated subtle vitamin K depletion in a hospital setting and determined whether protein–calorie malnutrition predicts its presence.Design, setting, subjects: Using a high-pressure liquid chromatography (HPLC) assay of plasma phylloquinone and a food frequency questionnaire for phylloquinone intake, we examined the phylloquinone status of 27 hospitalized patients with normal coagulation parameters, no liver disease, and no recent warfarin use. We assessed protein–calorie nutritional status with Reilly’s criteria and anthropometrics.Results: 51% of patients (95% Cl=31% to 70%) had evidence of subtle vitamin K depletion as defined by a subnormal plasma phylloquinone concentration. Patients whose phylloquinone intake was less than the Recommended Daily Allowance (RDA) over the preceding year had lower plasma phylloquinone concentrations when compared to other patients: median (range) 0.106 nmol/l (0.022–0.461) versus 0.301 nmol/l (0.067–3.928), respectively (P=0.023). Plasma phylloquinone concentrations were no different, however, between well-nourished and malnourished patients: median (range) 0.245 nmol/l (0.022–0.522) versus 0.188 nmol/l (0.067–3.928), respectively (P=0.782).Conclusions: Subtle vitamin K depletion is common among hospitalized patients and protein–calorie malnutrition does not predict its presence.