Sinus Thrombosis Should Be Treated With Anticoagulation

Abstract

T HE TREATMENT OF CErebral venous sinus thrombosis (CVST) with heparin has been controversial for decades, but most experts now agree that most patients with CVST should receive rapid and full anticoagulation as soon as the diagnosis is made. Physicians who do not see many cases of CVST would be surprised to find out there is a dilemma at all. Treatment with anticoagulants for venous thrombosis in the legs and for pulmonary embolism is standard practice based on sound evidence. Adequately dosed heparin or, even better, full-dose low molecular weight preparations followed by some months of oral anticoagulants will prevent the extension of the thrombus and allow for natural recanalization. So why should CVST be an exception? The reason, of course, is the fear of cerebral hemorrhages. This concern is based on the fact that 30% to 40% of all patients with CVST have some degree of cerebral hemorrhage at the time of diagnosis. To explain this combination of thrombosis and hemorrhage, it is useful to consider briefly the pathophysiology of CVST. Cerebral venous sinus thrombosis often occurs in individuals who havean increasedgeneticoracquired thrombotic risk, sometimes in conjunction with an additional local factor (suchashead injury, infection, or surgery). Other individuals develop CVST without an apparent risk factor or cause. Thrombosis typically begins in the large sinuses and may extend into the cortical veins. The clinical evolution depends on the location and extension of the thrombus. When the thrombus is confined tooneofthemajorsuperficialsinuses, such as the superior sagittal or transverse sinuses, the symptoms are often limitedtothoseof intracranialhypertension: headache, papilledema, and sometimes a sixth nerve palsy. Thesesymptomscanbe treatedeffectively with the usual measures to reduce intracranial hypertension and withwaiting for spontaneousrecanalization of the sinuses. However, serious problems can develop in patients with thrombosis of multiple sinusesorwhenthecortical veinsbecome occluded. The possibilities for collateralvenousdrainagearelimited, resulting in increased capillary pressure, local cerebral edema, and petechialhemorrhages, leading in turn to a fully developed, often hemorrhagic venous infarct. Venous infarcts and hemorrhagescauseseizuresandneurologicaldeficitsandcanleadtodeath from cerebral herniation. The problemis thatwedonotknowwhichpatients with thrombosis are likely to develop a venous infarct or hemorrhage.Therefore, it seemsprudent to treat all patients with anticoagulants before they deteriorate. The risk of heparin-induced intracerebral hemorrhage has to be weighed against the risk of hemorrhage caused by additional thrombotic venous occlusions, and the only way to solve that dilemma is by clinical trials. In the 2 clinical trials that met the inclusion criteria for our Cochrane review, no new or enlarging hemorrhages occurred in the 40 patients treated with heparin. Of course, this may have been chance: 0 of 40 patients had a 95% confidence interval of 0% to 9%, so an incidence of up to 9% of hemorrhages cannot be excluded. Nevertheless, that no new cerebral hemorrhages developed in the trials after heparin treatment is reassuring. Many of these patients (18 of 40 [45%]) already had some intracerebral hemorrhage before the start of anticoagulation, which apparently did not worsen after heparin was initiated. The danger of not using heparin is illustrated by the fact that 2 individuals in the control group had new hemorrhagic infarcts and 2 others probably had a pulmonary embolism, one of which was fatal. Those who still doubt the benefit of heparin for CVST rightly point to the scarce evidence from the few and small trials. Indeed, the pooled effect, an absolute reduction in mortality of 13% in the patients treated with heparin, was not statistically significant (P=.08). Instead of simply discarding this reduction as nonsignificant and therefore unimportant, one can also assess the evidence considering what we know from other sources, such as pathophysiology and documented clinical experience. In the largest published prospective cohort (N=624) on this subject, 245 (39%) patients had hemorrhagic cerebral infarcts before treatment, and 83% of all patients were treated with heparin. The 30-day mortality was only 3.4%, and the final outcomes were also better than in many of the older published case series. Statistical significance was originally introduced to test hypotheses, and the level of significance, used to decide between a trueor false-positive result, was set by convention at =0.05. This does not imply that an effect that is not associated with the traditional P .05 (or with a 95% confidence interval that includes the zero effect point) has no meaning whatsoever. The best estimate we now have from the Author Affiliation: Department of Neurology, University of Amsterdam, Amsterdam, the Netherlands.