Magnesium sulfate neither potentiates nor inhibits tissue plasminogen activator-induced thrombolysis


Victor J. Marder, Los Angeles Orthopaedic Medical Center, 2400 S. Flower St, Los Angeles, CA 90007, USA.
Tel.: +1 213 742 1519; fax: +1 213 742 6568; e-mail:


Summary. Background: Increasing circulating magnesium concentrations to 2-fold over normal baseline may afford a neuroprotective effect in patients with acute cerebral ischemia. Objectives: As patients receiving magnesium sulfate (MgSO4) in human clinical trials may also be candidates for subsequent thrombolytic therapy with tissue plasminogen activator (t-PA), preclinical assessment of possible inhibition or potentiation of fibrinolytic activity by MgSO4 has important clinical relevance. Methods: We utilized an in vitro system, in which D-dimer release served as a reflection of t-PA-induced clot lysis, to measure the effect of magnesium at the target concentration being tested in human stroke clinical trials, and at 2- and 3-fold higher levels. Clots from normal volunteers were exposed to t-PA at concentrations that correspond to therapeutic or endogenous plasma t-PA levels. Results: MgSO4 had no effect on t-PA-induced clot lysis at up to 3-fold target magnesium concentration (6× normal serum concentration). Conclusions: MgSO4 concentrations well above the targeted level in therapeutic stroke trials does not affect t-PA-induced fibrinolytic activity, and therefore is a suitable agent for trials of combined neuroprotective and thrombolytic therapy in patients with acute ischemic stroke.


Combined neuroprotective and thrombolytic therapy is a promising strategy for treatment of acute ischemic stroke [1–5]. Delivery of neuroprotective therapy by paramedics in the field prior to hospital arrival may stabilize the ischemic penumbra and prolong the time window for safe and beneficial reperfusion therapy. Moreover, as thrombolytic therapy with tissue plasminogen activator (t-PA) is an FDA-approved and national guideline-recommended therapy for acute cerebral ischemia within 3 h of onset, trials of novel neuroprotective therapies in broad acute ischemic stroke populations now perforce take place on a background of concomitant thrombolytic therapy in a substantial proportion of the study population.

Magnesium sulfate (MgSO4) is under active investigation as a neuroprotective therapy for acute ischemic stroke. Prior studies document neuroprotective mechanisms of MgSO4in vitro, including a diminution of excitotoxic neuronal injury by antagonizing presynaptic calcium entry, and protection of white matter in ischemia models by inhibition of sodium–calcium exchange [6]. Magnesium sulfate substantially reduces stroke volume in the great preponderance, although not all, preclinical stroke models [6–10]. In human clinical trials, MgSO4 reduces ischemic brain injury when given before the onset of brain ischemia in adults with aneurismal subarachnoid hemorrhage [11] and fetuses undergoing premature birth [12]. A phase III trial evaluating MgSO4 for focal ischemic stroke up to 12 h after onset found no overall benefit of therapy, but point estimates suggested potential benefit among the subgroup of patients treated within 3 h of symptom onset [13].

Given these signals of potential efficacy from animal models and human clinical trials, the FAST-MAG (Field Administration of Stroke Therapy–Magnesium) trial has been launched as a pivotal trial testing the efficacy of MgSO4 for acute cerebral ischemia when initiated within 2 h of defined symptom onset [14]. However, considering the hyperacute enrollment window, many patients in the FAST-MAG trial will receive concomitant intravenous t-PA therapy.

This study was instigated by an ongoing clinical trial of MgSO4 in ischemic stroke, in order to assess the possible interaction between magnesium and t-PA in vitro, and to determine whether prior exposure to MgSO4 may potentiate or inhibit the thrombolytic activity of t-PA.



Twenty healthy donors, aged 26–60 years, 11 males, underwent phlebotomy after appropriate review and approval by institutional review board and after informed consent was obtained. Blood was obtained from the same donors on two occasions, for a volume of 30 and 41 mL respectively. Blood was collected between 08.00 and 10.00 h by venipuncture into vacutainer tubes containing 3.2% buffered sodium citrate. Aliquots of 1 mL whole blood were used for clot preparation and the remainder was centrifuged (20 min at 3000 × g). Platelet poor plasma was collected from the centrifuged samples and stored at 8 °C.

Clot lysis

Whole blood (1 mL) was mixed with 50 U bovine thrombin (Enzyme Research Laboratories, South Bend, IN, USA) in glass culture tubes, then vortexed and incubated at 37 °C for 60 min. After incubation, the tubes were centrifuged (5 min at 2000 × g) to separate clot from loose red blood cells and serum. Clots were washed three times with sterile normal saline, by adding 3 mL saline, gently mixing, and siphoning washed solution from the intact clot. After washing, clots were removed from the culture tubes, blotted on 11 cm circular filter paper (Whatman International Ltd, Maidstone, UK) and the dimensions of the long and radial axes imprinted to verify approximate consistency in clot size. Clots were then re-suspended in 1 mL of platelet poor plasma from the same donor. Reagents were added in μL quantities to the clot/plasma suspension.


Reagents included t-PA (Alteplase, Genentech Inc, South San Francisco, CA, USA), magnesium sulfate (2.5 mg 5 mL−1, Abboject, Abbott Laboratories, Chicago, IL, USA), 0.9% normal saline, and aminocaproic acid (250 mg mL−1, Abbott Laboratories). Two concentrations of t-PA were tested, the first chosen to mimic the maximal plasma concentration in an 80 kg adult after a standard 1 h infusion of 0.9 mg kg−1 to an 80 kg adult (18 μg mL−1 plasma). The second t-PA concentration was chosen to typify the local concentration of t-PA after release by endothelium at the site of a vascular occlusion, calculated to be twice the concentration present in blood (2 × 5 ng mL−1, 2 × 70 pM) [15,16]. Experimental magnesium concentrations were calculated based upon the treatment regimen employed in both the IMAGES [13] and FAST-MAG [14] phase III clinical trials: a 4-g loading dose and 16-g maintenance dose to rapidly increase and sustain the serum magnesium level at twice normal levels. Experimental magnesium solutions were prepared at concentrations of 4 mEq L−1 (FAST-MAG target concentration), and at 8 mEq L−1 (2 × target) and 12 mEq L−1 (3 × target). After addition of reagents, duplicate suspended clots were incubated for 10 min at 37 °C; saline controls, MgSO4 only controls, and samples testing lysis of plasma alone were incubated for 60 min. After incubation, aminocaproic acid (Abbott Laboratories) was added to a final concentration of 0.19 m to terminate fibrinolysis.

Quantitation of thrombolysis

After addition of aminocaproic acid, the reaction suspension was centrifuged (30 min at 3500 × g) and supernatant was collected and assayed for D-dimer concentration by a latex agglutination method (STA Liatest D-Dimer; Diagnostica Stago, Asnières, France), as previously applied by others [17]. Our use of D-dimer release from clot as a reflection of clot dissolution is based on the report by Colucci et al. [18]. Tests for D-dimer were performed either immediately or after freezing (−80 °C) and thawing at 24 h. D-dimer values were expressed as fibrinogen equivalent units in μg mL−1, means of duplicate samples.

For study of magnesium-induced increase or decrease in thrombolysis by a low concentration of t-PA, incubations were continued for 60 min and a more sensitive, ELISA-based method was employed for measuring low concentrations of D-dimer (Asserachrom D-Di; Diagnostica Stago) [19].

Assay of plasma-derived soluble fibrin is based on the observations of Francis et al. [20], showing that t-PA-induced plasmin activity degraded fibrin polymers to D-dimer, detectable by an ELISA technique and verified by changes in immunoblot electrophoresis.

Statistical analysis

The null hypothesis was that increasing doses of MgSO4 do not affect t-PA-induced clot lysis, at either treatment or physiologic dose of t-PA. Means between treatment groups were compared using Student's t-test for paired data. Dose response was assessed by a straight-line estimate using the method of least squares. Analyses performed using JMP statistical software, version 5.1.2 (SAS Institute, Cary, NC, USA).


Serum magnesium levels attained following loading 4 g MgSO4 bolus in 20 participants in the FAST-MAG pilot trial are shown in Fig. 1. The mean and median concentrations after treatment were 3.7 and 3.5 mEq L−1 (range: 3.2–4.2), corresponding closely to the target concentration of twice normal magnesium level (2 × 1.8 mEq L−1 = 3.6 mEq L−1). No patient's serum magnesium level exceeded 1.6 × the target concentration. Control clots incubated with saline or MgSO4 alone for 10 or 60 min demonstrated miniscule (<5 μg mL−1) release of D-dimer (Fig. 2), indicating minimal spontaneous lysis. Incubation of clots with t-PA (18 μg mL−1) resulted in significant lysis at 10 min, increased further at 60 min for all 20 subjects (mean: 149.8 ± 65.3 μg mL−1 at 10 min vs. 214.1 ± 68.1 μg mL−1 at 60 min, P = 0.045). All subjects demonstrated increased lysis at 60 vs. 10 min.

Figure 1.

 Individual values of postbolus serum magnesium in 20 subjects who participated in the FAST-MAG Pilot trial [13]. Mean serum concentration is 3.72 mEq L−1 (95% confidence interval: 3.24–4.19).

Figure 2.

 Baseline clot lysis. Values in parentheses indicate duration of incubation. Control value reflects the results in 20 subjects; MgSO4 result shows duplicate values from one subject at concentrations of 16 mEq L−1 and 32 mEq L−1. Individual values for t-PA alone are means of duplicate determinations in 20 subjects.

Figure 3 shows the results of clot lysis in the presence of therapeutic t-PA and the target concentration of MgSO4, as well at 2-fold and 3-fold target MgSO4 concentrations. There were 20 observations for analysis at each treatment level. Doses of 4-fold target MgSO4 (16 mEq L−1) and higher were tested, but as these doses are not clinically applicable and far exceeded the postloading dose concentration of all FAST-MAG subjects (Fig. 1), the data is not presented. The best fit dose–response line has a slope approaching zero (R2 = 0.032, P = 0.10, root mean square error = 50.9 μg mL−1), consistent with no dose-dependent effect of MgSO4 on t-PA-induced clot lysis. To test the hypothesis that MgSO4 would decrease t-PA-induced clot lysis, clots were incubated for 60 min to produce maximal lysis with either t-PA (18 μg mL−1) alone or a combination of t-PA (18 μg mL−1) and MgSO4 at twice target concentration (8 mEq L−1). There was no difference in D-dimer release between the two groups (214.1 vs. 212.4, P = 0.92) indicating that MgSO4 did not inhibit t-PA-induced clot lysis (Fig. 4).

Figure 3.

 Influence of MgSO4 on tissue plasminogen activator (t-PA)-induced clot lysis in 20 subjects. Clot lysis quantified by D-dimer released into solution (see Methods). The target dose of MgSO4 is 4 mEq L−1 (twice the normal serum concentration), 8 mEq L−1 is 2-fold target and 12 mEq L−1 is 3-fold target. The best fit dose–response curve has a regression coefficient of R2 = 0.032, consistent with no effect of MgSO4 dose (P = 0.11).

Figure 4.

 Measurement of tissue plasminogen activator (t-PA)-induced clot lysis after 60 min incubation in duplicate samples from 10 subjects. Addition of MgSO4 at 2× target dose (8 mEq L−1) did not inhibit clot lysis.

To determine the degree to which a plasma source of D-dimer could have contributed to our estimation of clot lysis, samples of platelet-poor plasma were incubated for 10 and 60 min with t-PA (18 μg mL−1) (Table 1). After 10 min incubation, 42.6% of measured D-dimer could be accounted for by lysis of plasma soluble fibrin (52.4 μg mL−1 of 123.1 μg mL−1 from plasma plus clot). However, the presence of MgSO4 (2 × target concentration) in addition to t-PA did not increase the amount of D-dimer released from plasma (50.2 vs. 52.4 μg mL−1). As expected, D-dimer release from clot increased with longer incubation of 60 min, but plasma-derived D-dimer did not increase, representing only 27.1% of the total at 60 min (58.04 μg mL−1 of 214.14 μg mL−1).

Table 1.   Relative amounts of D-dimer derived from plasma or clot
 10 min60 min
t-PAt-PA + MgSO4t-PA
Plasma (μg mL−1)52.450.258.4
Plasma + clot (μg mL−1)123.1137.2214.4
D-dimer from plasma (%)42.636.627.1

To measure the effect of MgSO4 on lysis induced by physiologic concentrations of t-PA, as might be expected in an occluded vascular bed, a relatively low concentration of t-PA was combined with MgSO4 at target and at 2-fold target concentrations (Fig. 5). As in the observations shown in Fig. 3 (above), the best fit dose–response demonstrates a regression coefficient close to zero (R2 = 0.026, P = 0.65), indicating no dose-dependent effect of MgSO4 concentration on low dose t-PA-induced clot lysis.

Figure 5.

 Effect of MgSO4 on clot lysis induced with physiologic concentrations of t-PA (‘low dose’). The relatively low concentrations of released D-dimer were quantified by ELISA method. Line fit with regression coefficient = 0.026, P = 0.65, demonstrating no significant dose effect on degree of lysis.


We demonstrate here that magnesium concentrations at up to 6-fold normal serum values do not affect t-PA-induced in vitro clot lysis. As shown in Figs 3 and 4, MgSO4 caused neither an increase nor a decrease in clot lysis in the presence of t-PA at the typical concentration expected in acute cerebral ischemia patients following therapeutic infusions of 0.9 mg kg−1 t-PA [21]. The therapeutic index was wide, with no observed interaction of magnesium and t-PA at the target concentration of MgSO4 used in acute ischemic stroke trials (IMAGES [13], FAST-MAG Pilot [14], and FAST-MAG), or at two and three times the target MgSO4 concentration. This lack of potentiating or inhibitory effect on t-PA activity is reassuring. On the basis of our findings, patients treated with t-PA after prior exposure to MgSO4 would not be expected to have reduced fibrinolytic activity at the site of the thrombotic vascular occlusion, or an exaggerated risk of intracranial hemorrhage.

Our in vitro studies suggest that potentiation of endogenous t-PA activity is not likely to be a major mechanism of MgSO4 benefit in patients with acute thrombotic stroke. We estimate that t-PA is present locally in occluded cerebral vessels at up to twice the background concentration in the circulation [15], as the result of regulated endothelial secretion into the microcirculation in response to local thrombosis [16]. Our studies of MgSO4–t-PA interaction included conditions of low t-PA concentration, but as with high t-PA levels, there was neither an enhancement nor a depression of t-PA clot-lysis activity with up to 6-fold normal serum concentration of magnesium (Fig. 5).

We did not study in vivo conditions of thrombolysis or hemorrhage in animal models, so our conclusions are not definitive. However, it is unlikely that administered t-PA would have an increased fibrinolytic effect in patients who also receive MgSO4, as both the therapeutic and hemorrhagic effects of t-PA are mostly a reflection of blood t-PA concentration [22], and adding MgSO4 to a given concentration of t-PA in our in vitro systems did not increase t-PA-induced clot lysis. While the findings reported here appear to refute an effect of MgSO4 on t-PA-related thrombolysis, MgSO4 could still impact on clinical outcome in patients treated with t-PA by virtue of effects on the inflammatory response, matrix metalloproteinases, the blood–brain barrier, or vascular wall function or integrity. Our in vitro findings do not provide insight for proposed cardioprotective effects of MgSO4 in patients with acute myocardial infarction (MI) [23], but the lack of interaction is in accord with clinical trial observations. In trials of MgSO4 for acute MI, there was no evidence of adverse drug interactions among above 20 000 patients receiving both MgSO4 and fibrinolytic agents, including above 600 receiving MgSO4 and t-PA [24–26].


Our results show no effect of t-PA-induced clot lysis by MgSO4 at up to 6-fold normal serum magnesium concentration. This in vitro data complements human clinical trial experience in acute myocardial ischemia and suggests that patients who receive very early MgSO4 therapy as part of treatment for acute cerebral ischemia will experience neither increased nor decreased t-PA-related thrombolysis, nor increased susceptibility to intracranial hemorrhage as a complication of magnesium therapy.


This study was supported by Grant RO1 HL 074051 from the National Heart, Lung, and Blood Institute, Grant T32 HL66992 from the National Heart, Lung, and Blood Institute, and Grant P50 NS044378 from the National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD, USA.