Heparin in Acute Ischemic Stroke: The Case for a New Clinical Trial

More than 80 years after the discovery of heparin it is still uncertain whether its administration to acute ischemic stroke patients results in harmful or beneficial effects. During the second half of this decade several clinical trials showed that unfractionated heparin (UH) or fractionated heparins (FH) either provided no clear clinical benefit to the patients, or resulted in an excessive risk of untoward complications. Despite these discouraging results some clinicians are still reluctant to put the last nail to the coffin of heparin in the setting of acute ischemic stroke. In the following lines, the main results of these trials will be summarized and arguments will be given to judge whether the use of heparin in acute ischemic stroke still merits further clinical research or, rather, it deserves a place among the long list of outdated therapies. Recent experimental data gathered in murine models of ischemic stroke models will be discussed to emphasize other than the antithrombotic effects of heparin. Our more recent experience with the use of UH in stroke patients will also be reviewed to provide some indirect comparisons between the results obtained in clinical trials with those procured in daily clinical practice. These data will serve to address the crucial question whether another trial of heparin is still necessary in acute ischemic stroke.During the last 3 years four trials of anticoagulation in acute ischemic stroke have been reported in final [1, 2, 3]or preliminary [4]form. Only one of these studies [1]was able to demonstrate that active treatment was better than placebo in improving clinical outcome after ischemic stroke. Nevertheless, the beneficial effect of nadroparin was observed at 6 months after stroke onset but not before. While the IST trial found that patients allocated to UH had significantly fewer recurrent strokes within 14 days (2.9 vs. 3.8%) than those patients not treated with UH, this benefit was offset by a similar-sized increase in hemorrhagic strokes (1.2 vs. 0.4%), so the difference in death or nonfatal recurrent stroke was not significant [2]. As summarized in table 1, the other trials were unable to reject the null hypotheses that was previously defined, although the death and dependency rates observed across studies varied substantially. As shown in table 2, the methodology and treatment regimens given to studied patients also differed among the trials. With the exception of the TOAST study, which adjusted the rate of treatment infusion to achieve the intended biological effect, the remaining trials evaluated the effects of different doses of low-molecular-weight heparins (LMWH) or UH without dose adjustments or biological monitoring. In all the trials, patients received active treatment or placebo within no less than 24 h from stroke onset, although in many instances treatment delay was even longer. In the two nadroparin trials there were significant dose differences implying that a patient weighing 80 kg received 6,880 IU every 12 or 24 h in the FISS bis study but only 4,100 IU in the FISS study. Although all the trials recommended to perform a CT scan at the time of randomization to rule out a hemorrhagic infarct, 5,600 IST patients were scanned after randomization and treatment initiation. Subsequently, a number of hemorrhagic strokes were wrongly included in the largest trial. The type of stroke randomized to receive active treatment or placebo varied across the trials. As shown in table 3, 1 out of 3–4 patients were admitted in the trials with symptoms related to a lacunar stroke. Table 4 highlights the main demographic and epidemiological traits of studied patients and shows the high proportion of patients over 80 included in IST, and the lower proportion of patients with past history of stroke admitted into the FISS study. To what extent these differences accounted for the great variation of death and dependency observed at the end of individual trials deserves careful thought [5]. Table 5 highlights the direct relationship between increasing doses of the anticoagulant regimen and the rate of bleeding complications. Somehow unexpectedly, the highest bleeding rates were observed in trials involving LMWH and heparinoids, although it is uncertain whether these findings were related to the properties and dose of allocated treatments or they resulted from the characteristics of the studied population.Heparins represent a large group of agents which includes sodium and calcium UH, a large number of LMWH [6]and heparinoids. These agents are administrated to stroke patients to prevent thrombus progression, hinder the occurrence of early stroke recurrence, and to prevent deep venous thrombosis and pulmonary embolism [7]. As most clinicians who treat stroke patients are familiarized with the use of UH but much less with the use of LMWH and heparinoids, this review emphasizes the mechanisms of the latter two types of drugs.LMWH are glycosaminoglycans consisting of chains of alternating residues of D-glucosamine and uronic acid, either glucuronic acid or iduronic acid, which are produced by controlled enzymatic or chemical depolymerization processes of UH that yield chains with a mean molecular weight of about 5,000. Irrespective of their molecular weight, heparins exert their anticoagulant activity by producing a conformational change of antithrombin that accelerates its interaction with thrombin and activated factor X (factor Xa). The conformational change of antithrombin is mediated by a pentasaccharide sequence of the heparin chain. Whereas any pentasaccharide-containing heparin chain can inhibit the action of factor Xa, only sequences of at least 18 saccharide units are able to inactivate thrombin if they bind to antithrombin. As most of the chains of UH are at least 18 saccharide units long, but fewer than half of those of LMWH have this length, UH has equivalent activity against factor Xa and thrombin but LMWH have greater activity against factor Xa. Nevertheless, as direct comparisons between UH and LMWH are lacking in the stroke population, the relationship between ex vivo measures of IIa and Xa inhibition and clinical antithrombotic and antihemostatic effects remains unsettled. Further, it is now clear that the adjustment of the potency of LMWH simply on the basis of the anti-Xa activities is not adequate as the antithrombin III affinity factors responsible for the anti-Xa activity only constitutes 20–30% of the material [8]. The remaining 70–80% include other actions, such as the release of tissue factor pathway inhibitor, the interaction with heparin cofactor II, the inhibition of leukocyte procoagulant actions, the promotion of fibrinolysis or the modulation of vascular endothelium. LMWH offer a few advantages over UH for they have a more predictable anticoagulant response, better bioavailability at low doses, dose-independent clearance mechanisms, and longer half-life. These differences are explained because LMHW have less binding to plasma proteins released from activated platelets or endothelial cells, less binding to the endothelium and macrophages, and they are eliminated by a nonsaturable renal mechanism [9, 10]. As a result, laboratory monitoring of LMHW is unnecessary except in special circumstances, such as in patients with renal insufficiency or extreme weights. LMHW are also said to cause less bleeding than UH [11], although the incidence of bleeding events in reported trials ranges from 1.5 to 13.9% [12]. Both UH and LMHW are limited by their inability to inactivate thrombin bound to fibrin, which represents an important trigger for clot extension at sites of vascular injury [13]. Additional drawbacks of LMWH concern their lack of dose adjustment guidelines and specific antagonist in case of bleeding. Moreover, as LMWH differ from each other in manufacturing, physicochemical, biological, pharmacokinetic, and biotransformation patterns, the clinical findings associated with a given LMWH preparation cannot be extrapolated to another one or generalized to the whole LMWH family without adequate clinical testing [14]. It must also be remembered that most of the studies on the comparison between different LMWH were primarily based on clinical data obtained in general surgery patients. As a result, these observations cannot be directly extrapolated to the stroke population.Danaparoid sodium is a heparinoid composed by a mixture of glysoaminoglycans with a mean molecular weight of 5,500 daltons that is isolated from porcine intestinal mucosa. Its anti-factor Xa activity is attributed to its heparan sulfate component. Alike other LMWH, danaparoid sodium is not inactivated by endogenous heparin-neutralizing factors and it has virtually no effect on platelet function. In light of these properties, heparinoids are theoretically better suited than UH for safe administration in acute ischemic stroke.Recent investigations have demonstrated that inflammation plays a pivotal role in the pathophysiology of brain ischemia [15, 16]. Despite the great advances achieved in this exciting field, little information exists on how heparin could modify these processes in human ischemia. Histologically, the inflammatory reaction in brain ischemia is characterized by the infiltration of leukocytes, mainly polymorphonuclear leukocytes and monocytes/macrophages. Intense leukocyte infiltration of the brain parenchyma is described 2–3 days after the stroke, although experimental models of MCA occlusion recognize the presence of leukocytes in the microvessels as early as 30 min after occlusion. This inflammatory response is preceded by the expression of several cytokines such as tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), resulting in chemotactic cytokine release, leukocyte adhesion molecule upregulation, and conversion of the local endothelium to a prothrombotic state. Leukocyte-endothelial cell interaction is essentially regulated by cell adhesion molecules which include the selectins, the Ig superfamily and the integrin superfamily. There are a number of mechanisms by which leukocytes could produce deleterious effects on ischemic brain parenchyma [17]. The release of vasoconstrictive mediators, such as superoxide anions, thromboxane A2, endothelin-1, and prostaglandin H2, the alteration in cerebral artery vasoreactivity, and the release of cytotoxic enzymes, free oxygen radicals, NO, IL-1β, TNF-α and products of the phospholipid cascade have been invoked. Many of the molecules responsible for leukocyte adhesion and extravasation have been elucidated [18]. There is a first stage termed rolling, in which trafficking leukocytes contact with the vasculature at the site of ischemia where selectins mediate [19]. Selectins expressed on leukocytes, platelets, and endothelial cells have been described [20]. In a second stage, the surface of contact between flattened leukocytes and the endothelium increases through the action of integrins, which includes the CD18 family expressed on leukocytes. Intercellular adhesion molecules expressed on the endothelium, such as ICAM-1 and ICAM-2, are also at play. Cytokines may be the key components in the activation and recruitment of leukocytes into the ischemic tissue, although the signals mediating these actions are not clear. Most cytokines are upregulated at 1 h postischemia and have a peak response at 6–12 h [21, 22]. Therefore, it is likely that inhibition of the inflammatory response needs to be accomplished as early as possible after symptoms onset to be clinically meaningful.The mechanism of action of heparin for brain protection has been thought to be dependent on its antithrombotic property. Other effects, including the regulation of angiogenesis, lipoprotein lipase modulation, maintenance of endothelial wall competence, and inhibition of vascular smooth muscle proliferation after endothelial injury, might as well be relevant [23, 24, 25, 26, 27]. The potential role of heparin as anti-inflammatory agent is being increasingly recognized and this property is currently evaluated in inflammatory conditions such as Crohn’s disease and ulcerative colitis [28, 29, 30]. The anti-inflammatory effects of heparin in animal stroke models are also under current investigation [31].It is of great clinical importance to establish whether UH and FH are equally effective as anti-inflammatory agents. Both types of drug inhibit leukocyte rolling on the vessel wall [32, 33]and evidence has accumulated that this activity depends in part on their ability to block selectins on leukocyte and platelets [34]. Although the expression of adhesion molecules seems to be unaffected by UH, this drug binds leukocyte integrin Mac-1 (CD11b/CD18), therefore interfering the cell adhesive interactions between leukocytes and endothelial cells [25, 35]. Moreover, LMWH seem to reduce leukocyte adherence to endothelial cells to a slightly lesser extent than UH [33]. Recently, it has been shown in vitro that heparin attenuates in a dose-dependent manner the increase in inducible nitric oxide synthase (iNOS) and NO release after cytokine activation [36]. This finding suggests that heparin influences the cytokine/membrane interaction or the early steps of signal transduction in inflamed endothelial cells.The clinical and morphological consequences of the modification of inflammatory processes with the use of heparin have been evaluated in murine models of transient brain ischemia. Yanaka et al. [31]showed in a rat ischemic model that animals receiving UH or dextran sulfate showed a significant reduction in leukocyte accumulation, infarct size, and neurological dysfunction 48 h after reperfusion when compared to untreated animals. Further, animals that received UH showed significantly better results than animals that received an equivalent anticoagulant dose of LMWH. Higher doses of UH were more effective than lower doses. According to these results, the investigators concluded that the relative potency of sulfated polysaccharides in the inhibition of leukocyte accumulation and the reduction of infarct size can be attributed to their degree of sulfation [31]. The same research group assessed the optimum timing of the administration of heparin. In this study, Yanaka et al. [37]compared the extent of leukocyte activation, infarction size and neurological function of rats subjected to transient MCA occlusion and allocated to heparin at either 3, 6 or 24 h after reperfusion. In these experiments, animals treated within 3 h after reperfusion did significantly better than animals treated at 6 h after reperfusion. Leukocyte activation, as reflected by the myeloperoxidase activity, and the size of the infarction were also significantly reduced in the group of animals treated at an earlier phase. From these experimental data emerges the view that the anti-inflammatory efficiency of heparin depends on the type and dose of heparin and on the delay to its administration.It has been claimed that the clinical characteristics of patients included in clinical trials might differ from those treated in conventional neurological wards or in stroke units. In particular, the inclusion of heterogeneous stroke subtypes in these trials is worrying. During the last 5 years, 725 acute ischemic stroke patients admitted to our stroke unit received full doses of UH within 48 h from stroke onset, as soon as CT had ruled out the presence of blood within the brain. Overall, the stroke population that is treated in the stroke unit represents 78% of all ischemic strokes admitted to our institution, a referral teaching hospital with approximately 600 stroke admissions per year. As admission into the stroke unit only depends upon its bed availability, and not on the cause, duration and severity of symptoms, we believe that our data reflect adequately the type of patients admitted into referral hospitals of large urban communities. In all instances the biological effect of UH was first assessed 6 h after treatment onset by measuring the activated thromboplastin time (APTT). Afterwards, APTT measurements were repeated as necessary to adjust the dose of heparin at 1.5–2 times control values. According to our clinical protocol, UH was administered to most acute patients with nonlacunar strokes, although individuals with progressing lacunar stroke were also anticoagulated if other therapeutic strategies failed. In agreement with previous observations [38], patients were anticoagulated early irrespective of their age, infarction size or neurological condition on admission. Stuporous patients and those with active seizures or recent history of bleeding were not anticoagulated. In addition to the pretreatment CT scan, most patients had prior to death or hospital discharge a second brain CT scan or a brain MRI (60%) to better document the characteristics of the infarction. From a clinical standpoint, hemorrhagic worsening referred to any clinical deterioration which was associated to the presence of blood on CT scan or MRI, or that required transfusions. Aware of the potential risks of early anticoagulation, we closely monitored the patients’ neurological condition with standard neurological scales (Mathew scale) and neuroimaging. Daily platelet counts were also performed to detect the incidence of thrombocytopenia. Many patients had videofluoroscopic studies of the upper digestive tract to determine whether food could be safely given by oral route. Blood pressure, serum glucose levels, and body temperature were monitored and treated as required [39]. If needed, rehabilitation was initiated in most cases the first day of hospital admission. As indicated in table 6, whereas the prevalence of risk factors and baseline hemodynamic characteristics of patients included in our series were equivalent to the figures provided in reported trials, our series included a much lower proportion of patients with lacunar stroke.Overall, the infusion of UH had to be permanently stopped in 5% of the patients because a CT/MRI disclosed at follow-up moderate to large of blood within the brain hemorrhagic worsening in only 12 patients of the as indicated in table The incidence of was less than The death rate and outcome at discharge are also in table further whether the time delay to treatment from clinical onset neurological outcome at hospital To the effect of stroke on we this to the group of patients with first stroke associated with of this population have been reported will soon be in As reflected in the model of table the of included age, admission neurological CT findings and delay to heparin initiation. delay less than 6 h to was also associated with and early we had demonstrated that the baseline clinical condition and prevalence of risk factors were in the two that were in to treatment In this population, hemorrhagic worsening and early stroke in APTT were achieved at 24 h in less than half the patients. This laboratory finding is of great clinical as we not only found that stroke was associated with lower mean APTT but significantly mean APTT were also in patients with hemorrhagic In agreement with our previous age, admission stroke blood pressure, and baseline CT findings did not a risk of hemorrhagic worsening reported heparin trials in acute ischemic stroke have so been although drawbacks in their could in part be of adequate biological monitoring of the anticoagulant effect achieved in treated patients, inclusion of heterogeneous of stroke subtypes more an excessive delay to treatment could among the factors associated with treatment these studies have provided data for the largest IST trial showed that UH more events than at early stroke not be the only it is likely that the clinical effect of heparin depends on the delay to its administration as experimental data a modulation of the inflammatory cascade the the drug is This is further by clinical data a direct relationship between the of the inflammatory response, outcome the risk of early stroke As heparins a heterogeneous group of these effects clinical testing in human stroke. UH has provided the evidence for an anti-inflammatory effect and therefore be the first for testing in a large clinical trial. previous treatment be initiated within a much more time is an that will more than patients with nonlacunar ischemic stroke and symptoms less than 12 h to the relative of full doses of UH of CT prior to treatment onset and monitoring of the APTT active treatment will be to the risk of bleeding. is to in the in to of