A 52 year-old male unemployed labourer was referred because of recurrent thromboembolism. An episode of thrombophlebitis migrans 3 months earlier had been followed by a pulmonary embolus a week later, and then a deep vein thrombosis despite apparently adequate anticoagulant therapy with warfarin 10 mg daily (INRs between 2 and 4). For 9 years he had suffered from hypertension for which he took lisinopril 2.5 mg daily, had smoked 20 cigarettes daily since his teens and drank 12–15 pints of beer each week. On presentation, his right leg remained hot, swollen and tender to the thigh but there were no other abnormal findings. Investigations revealed a normochromic, normocytic anaemia (Hb 11.6 g dl −1 ) and his erythrocyte sedimentation rate, which had been 11 mm h−1, had risen to 35 mm h−1. Renal and liver function tests, autoantibody screen, and prostate-specific antigen were normal. Antithrombin III, protein C and S concentrations were all within normal limits and ultrasound of the abdomen and pelvis, barium meal and enema, and CT of the thorax, abdomen and pelvis, bronchoscopy and brushings also failed to reveal any abnormality. He was commenced on full-dose (porcine mucosal) heparin for 10 days and the warfarin dose was increased to 15 mg daily until his INR had risen to 4.0, when his plasma warfarin concentration was measured at 3.1 mg l−1. Patients vary markedly in their requirement for warfarin, but 95% of subjects need more than 1 and less than 9 mg day−1. Subjects requiring 9 mg or more are therefore classed as warfarin-resistant. Warfarin resistance may be due to several factors (either pharmacokinetic, pharmacodynamic) or due to poor concordance (compliance). The important causes of warfarin resistance are listed in Table 1. Pharmacokinetic causes include enhanced warfarin metabolism. This may be due to enzyme induction, particularly by certain drugs (e.g. barbiturates, carbamazepine and rifampicin). Impairment of absorption of warfarin (it is normally completely absorbed from the gut) can occur due to binding by cholestyramine, colestipol or charcoal. Poor concordance may simulate a pharmacokinetic cause, since the ingested dose is lower than that intended to be taken. Pharmacodynamic causes of resistance include ingestion of excess vitamin K in certain health foods or weight-reducing diets or in enteral feeds. Oestrogens, griseofulvin, 6-mercaptopurine and haloperidol may cause diminution of anticoagulant effect in some individuals, although the mechanisms are poorly understood [1]. Patients with hyperlipidaemia (particularly hypercholesterolaemia may be relatively resistant to warfarin [2, 3] and although the effect appears to have a pharmacodynamic basis [3], the cause is unknown. Finally, an extremely rare form of hereditary warfarin resistance has been described in two human kindreds [4, 5]. The trait is inherited in an autosomal dominant fashion and a mutation of the receptor site shared by warfarin and vitamin K is postulated, although the locus of the mutation is unknown. Measurement of the plasma warfarin concentration may be helpful in all these circumstances, since at steady state, the reference range is 0.8 to 2.4 mg l−1 [6]. The reference range for warfarin clearance (calculated by dividing the daily warfarin dose by the steady-state plasma concentration) is 2.5 to 8.7 l day−1. Our patient’s clearance was within the reference range (4.8 l day−1 ), confirming that his resistance was not due to enhanced elimination but appeared to be of a pharmacodynamic nature (Figure 1). The reason for this resistance was not immediately apparent. Plasma concentrations below the reference range at this dose would suggest either rapid metabolism or poor concordance. A similar algorithm is available to investigate causes of warfarin sensitivity [6]. Algorithm to determine cause of warfarin resistance. Recurrent thromboemblism, despite an apparently adequate dose of warfarin and an INR within the recommended range for prevention of recurrence is not warfarin resistance. Since the drug has produced its required hypoprothrombinaemic effect, warfarin inefficacy may perhaps be a better term for this rare phenomenon. It is sometimes seen in patients with underlying cancer [7]. The heparin infusion was discontinued. Four hours later he developed painful and persistent penile erection (priapism) which he only reported after a further 4 h when it failed to settle. An urgent urological referral was sought and intracavernosal metaraminol was injected and surgical decompression of the penis performed by aspiration of blood. The platelet count was normal and heparin was recommenced at full dose by intravenous infusion. Subsequent thrombophilia screen was also essentially normal except for the presence of spontaneous platelet aggregation and secondary platelet aggregation after addition of ADP at concentrations of less than 0.438 μmol. While receiving heparin an inferior vena caval umbrella (Greenfield filter) was inserted. The heparin was then given at full doses by subcutaneous injection and the patient was eventually discharged on this treatment. Priapism is defined as a persistent erection that cannot be relieved by sexual intercourse or masturbation [8]. It is a urological emergency, since persistence may be associated with a high rate of subsequent impotence. Although up to 50% of cases may be idiopathic, it is also associated with a variety of haematological disorders (e.g. multiple myeloma, leukaemia and sickle cell disease) or drug therapy. Drug-induced causes include phenothiazines, butyrophenones, antidepressants, antihypertensives and anticoagulants (Table 2). Our patient was receiving lisinopril which is not one of the antihypertensives reported to cause priapism. The relationship between anticoagulants and priapism was first recognised by Duggan & Morgan [9] in four patients receiving heparin after myocardial infarction. Because the clotting times of these patients were low (in two cases before the onset and in 2 some period after the onset of priapism), the authors speculated that a rebound hypercoagulability may occur. This has certainly been demonstrated after abrupt heparin discontinuation in the context of acute coronary syndromes and post-angioplasty [10, 11]. Priapism has been seen when heparin was discontinued in patients undergoing haemodialysis [12]. Burke and co-workers [13] described priapism in a patient with deep vein thrombosis, 7 h after intravenous infusion of heparin was discontinued. It is of interest that in the case report of Burke et al. [13], the subject’s platelets were shown to spontaneously aggregate and to be sensitive to ADP induced aggregation whilst on heparin. The same abnormalities were also seen in our subject. Thompson and colleagues [14] showed that heparin can increase platelet aggregation by ADP and adrenaline, both in vitro and in vivo. Heparin alone can induce platelet aggregation which is more pronounced with heparin fractions of high molecular weight (average 20 000) than low molecular weight (average 7000) [15] but the clinical significance of this is unclear. Priapism has been described in association with warfarin [16, 17] but it appears to be much less frequent. Again, the mechanism is unknown but withdrawal has been implicated as a precipitating factor [18]. The treatment of priapism remains largely surgical but α-adrenoceptor agonists (e.g. intracavernosal metaraminol) have been used, perhaps because several drugs causing priapism have α-adrenoceptor antagonist properties. Diphenhydramine has also been used successfully in priapism secondary to chlorpromazine, α-adrenoceptor blockers and trazodone, and cyproheptadine and benztropine have been employed successfully in trazodone-induced priapism. None of these studies was controlled. Further information is available in the review by Banos and colleagues [8]. The risk of impotence after priapism is around 40% overall but is related to the duration of symptoms prior to treatment, inversely related to the age of the patient, and poorest when heparin or a combination of alcohol and psychopharmaceuticals was the aetiological factor [19]. One month later he developed dysphagia for solids, low back pain of sudden onset with radiation down his left thigh. Bone-scan revealed increased uptake in his lumbar vertebra and CT thorax now showed a lesion pressing against the oesophagus. Local radiotherapy improved his dysphagia and back pain but he slowly deteriorated and died 8 weeks later. At post-mortem, extensive undifferentiated carcinoma was found at many sites. A primary site could not be identified. Thrombosis is the most frequent complication of cancer and the second most frequent cause of death. It is also clear that the thrombotic episodes may precede the diagnosis of underlying cancer by months (as in this man) or sometimes even years. Tumour cells and their products can interact with platelets and recently, a promoter of coagulation, cancer procoagulant, has been discovered in patients with solid tumors as well as in acute promyelocytic leukaemia [20]. It is thought to be a vitamin K dependent enzyme, but there may be several other procoagulants with different structures. There is a clinically important relationship between idiopathic deep vein thrombosis and the subsequent development of overt cancer, particularly in patients with recurrent venous thromboembolism [21] and even patients with a first episode in the absence of conventional risk factors have a 10% likelihood of having cancer [22]. Monreal and co-workers [23] suggested that measurement of carcinoembryonic antigen, lactic dehydrogenase, chest radiography and abdominal ultrasound (or CT scan) in such patients sometimes detected malignancy that otherwise might have been overlooked. Unfortunately, extensive investigations were not helpful in our patient on first presentation. Not only does cancer increase thromboembolic risk, but surgery, chemotherapy and hormonal therapy may add to that risk. Full dose anticoagulant therapy may be associated with increased risk of bleeding, sometimes from the site of the tumour [24]. In such individuals, placement of a filter in the inferior vena cava may be as effective as, and yet safer than, anticoagulant therapy [24, 25]. In patients with stage IV breast cancer, low-dose (mean 2.6, s.d. 1.2 mg) warfarin therapy was safe and effective in reducing the excessive risk of thromboembolic complications in patients receiving chemotherapy [26]. In other subjects like our patient, warfarin may fail to protect against recurrent thromboembolism and long-term adjusted high-dose heparin may be required [7]. Recent trials of low molecular weight heparins have indicated that they may be more effective in preventing recurrent venous thromboembolism in subjects with cancer, to be less likely to produce major bleeding and to perhaps even reduce mortality, although prospective studies are need to confirm the latter observation [27]. • Warfarin resistance may be caused by several pharmacokinetic and pharmacodynamic factors Plasma warfarin concentrations may be useful in determining the possible underlying mechanism. • Recurrent idiopathic venous thromboembolism may be associated with occult malignancy and such individuals should be carefully investigated. • Warfarin inefficacy (the inability of adequate anticoagulation to prevent recurrent thrombosis) is rare, but should always prompt thorough investigation for underlying risk factors for thrombosis, including cancer. • Priapism may be caused by a variety of drugs, including heparin and perhaps (much more rarely) warfarin. The risk appears to be greatest just after heparin discontinuation. The prognosis for subsequent potency is poorest after heparin-induced priapism. • Anticoagulants, including low-dose warfarin and low molecular weight heparins may be effective in protecting against thrombotic disorders in cancer patients, and insertion of an inferior vena caval umbrella catheter should be considered in subjects in whom anticoagulant therapy is ineffective or contra-indicated.
No takes yet. Share an insight, caveat, or question.
Routledge et al. (1998) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: