The Haemostasis and Thrombosis Task Force of the British Committee for Standards in Haematology has produced a concise practical guideline to highlight the key issues in the management of heparin-induced thrombocytopenia (HIT) for the practicing physician in the UK. The guideline is evidence-based and levels of evidence are included in the body of the article. All patients who are to receive heparin of any sort should have a platelet count on the day of starting treatment. For patients who have been exposed to heparin in the last 100 d, a baseline platelet count and a platelet count 24 h after starting heparin should be obtained. For all patients receiving unfractionated heparin (UFH), alternate day platelet counts should be performed from days 4 to 14. For surgical and medical patients receiving low-molecular-weight heparin (LMWH) platelet counts should be performed every 2–4 d from days 4 to 14. Obstetric patients receiving treatment doses of LMWH should have platelet counts performed every 2–4 d from days 4 to 14. Obstetric patients receiving prophylactic LMWH are at low risk and do not need routine platelet monitoring. If the platelet count falls by 50% or more, or falls below the laboratory normal range and/or the patient develops new thrombosis or skin allergy between days 4 and 14 of heparin administration HIT should be considered and a clinical assessment made. If the pretest probability of HIT is high, heparin should be stopped and an alternative anticoagulant started at full dosage unless there are significant contraindications while laboratory tests are performed. Platelet activation assays using washed platelets have a higher sensitivity than platelet aggregation assays but are technically demanding and their use should be restricted to laboratories experienced in the technique. Non-expert laboratories should use an antigen-based assay of high sensitivity. Only IgG class antibodies need to be measured. Useful information is gained by reporting the actual optical density, inhibition by high concentrations of heparin, and the cut-off value for a positive test rather than simply reporting the test as positive or negative. In making a diagnosis of HIT the clinician's estimate of the pretest probability of HIT together with the type of assay used and its quantitative result (enzyme-linked immunosorbent assay, ELISA, only) should be used to determine the overall probability of HIT. Clinical decisions should be made following consideration of the risks and benefits of treatment with an alternative anticoagulant. For patients with strongly suspected or confirmed HIT, heparin should be stopped and full-dose anticoagulation with an alternative, such as lepirudin or danaparoid, commenced (in the absence of a significant contraindication). Warfarin should not be used until the platelet count has recovered. When introduced in combination with warfarin, an alternative anticoagulant must be continued until the International Normalised Ratio (INR) is therapeutic for two consecutive days. Platelets should not be given for prophylaxis. Lepirudin, at doses adjusted to achieve an activated partial thromboplastin time (APTT) ratio of 1·5–2·5, reduces the risk of reaching the composite endpoint of limb amputation, death or new thrombosis in patients with HIT and HIT with thrombosis (HITT). The risk of major haemorrhage is directly related to the APTT ratio, lepirudin levels and serum creatinine levels. The patient's renal function needs to be taken into careful consideration before treatment with lepirudin is commenced. Severe anaphylaxis occurs rarely in recipients of lepirudin and is more common in previously exposed patients. Danaparoid in a high-dose regimen is equivalent to lepirudin in the treatment of HIT and HITT. Danaparoid at prophylactic doses is not recommended for the treatment of HIT or HITT. Patients with previous HIT who are antibody negative (usually so after >100 d) who require cardiac surgery should receive intraoperative UFH in preference to other anticoagulants that are less validated for this purpose. Pre- and postoperative anticoagulation should be with an anticoagulant other than UFH or LMWH. Patients with recent or active HIT should have the need for surgery reviewed and delayed until the patient is antibody negative if possible. They should then proceed as above. If deemed appropriate early surgery should be carried out with an alternative anticoagulant. We recommend discussion of these complex cases requiring surgery with an experienced centre. The diagnosis must be clearly recorded in the patient's medical record.