Hyperkalaemia is a serious adverse effect of unfractionated heparin, but the effect of low-molecular-weight heparins (LMWHs) on potassium levels is not clear. Previous studies have shown a disparity of results depending on the dose and type of LMWH used. To analyse potassium level variations in medical and surgical inpatients receiving the LMWH bemiparin sodium at prophylactic doses and assess the consequent effective risk of hyperkalaemia in a real-life setting. This was a prospective observational study conducted over a 9-month period in a university teaching hospital. Patients consecutively admitted to internal medicine wards for general medical conditions (n = 145) or to traumatology wards for hip fractures (n = 98) and who received prophylactic bemiparin sodium were enrolled in the study. The intervention consisted of daily dosages of bemiparin sodium (Hibor) 3500 IU (56%) or 2500 IU (44%) for a minimum of 5 days. The mean age of participants was 80.5 years, with 91.8% being aged > or =65 years. Eighty-six percent of patients had co-morbidities and 79.4% were taking medication affecting potassium homeostasis. The main outcome measures were variations in serum potassium levels observed within 4-8 days of starting bemiparin sodium and the presence of hyperkalaemia (serum potassium >5.1 mmol/L) while on bemiparin sodium treatment. After patients had received bemiparin sodium for a median 6-day period, the mean (+/-SD) serum potassium level increased from 4.1 +/- 0.5 to 4.3 +/- 0.5 mmol/L (p < 0.001). Hyperkalaemia >5.1 mmol/L developed in ten patients (4.1%), but serum potassium levels >5.5 mmol/L related to bemiparin sodium were present in only two (0.8%). Laboratory tests between the fourth and eighth days identified all but one case of hyperkalaemia. Patients were not symptomatic and discontinuation of bemiparin sodium treatment was not required. There were no statistically significant differences in potassium disturbances between older (aged > or =65 years) and younger (aged <65 years) patients. The maximum serum potassium level showed a significant inverse correlation with bodyweight (R = -0.731; p = 0.016) and creatinine clearance (R = -0.640; p = 0.046), and a positive correlation with the individual variation in serum potassium levels (R = 0.692; p = 0.027) and with serum potassium levels after 4-8 days on bemiparin sodium treatment (R = 0.741; p = 0.014). Baseline potassium level (odds ratio [OR] 26.5, 95% CI 4.7, 150.3; p < 0.001) and treatment with ACE inhibitors (OR 10.5, 95% CI 1.9, 57.8; p = 0.007) were the only predictors of hyperkalaemia at admission (c-statistic 0.88, 95% CI 0.78, 0.99). For patients not receiving ACE inhibitors, a baseline serum potassium >4.6 mmol/L was considered the cut-off value for predicting hyperkalaemia (sensitivity 90% and specificity 70%). Serum potassium levels in in-hospital traumatology and medical patients increased significantly with bemiparin sodium prophylaxis but the incidence of relevant hyperkalaemia was low. Patients taking bemiparin sodium who are treated with ACE inhibitors or who have a baseline potassium level >4.6 mmol/L should be monitored for serum potassium levels between days 4 and 8 of hospital admission.