The ALIAS Pilot Trial
A Dose-Escalation and Safety Study of Albumin Therapy for Acute Ischemic Stroke—II: Neurologic Outcome and Efficacy Analysis
Background and Purpose— High-dose human albumin (ALB) is robustly neuroprotective in rodent stroke models. A phase I dose-escalation study was conducted to assess the safety of ALB therapy in ischemic stroke. We analyzed the data for preliminary evidence of treatment efficacy.
Methods— Eighty-two subjects with acute ischemic stroke (NIH Stroke Scale [NIHSS] of 6 or above) received 25% ALB beginning within 16 hours of stroke onset. Six successive ALB dose tiers were assessed (range, 0.34 to 2.05 g/kg). Forty-two patients also received standard-of-care intravenous tissue plasminogen activator (tPA). Efficacy outcomes were determined at 3 months. We compared the highest three, putatively therapeutic ALB dose tiers (1.37 to 2.05 g/kg) with the lowest three, presumed subtherapeutic doses (0.34 to 1.03 g/kg) and with historical cohort data derived from the NINDS rt-PA Stroke Study.
Results— After adjusting for the tPA effect, the probability of good outcome (defined as modified Rankin Scale 0 to 1 or NIH Stroke Scale 0 to 1 at 3 months) at the highest three ALB doses was 81% greater than in the lower dose-tiers (relative risk [RR], 1.81; 95% confidence interval [CI], 1.11 to 2.94) and was 95% greater than in the comparable NINDS rt-PA Stroke Study cohort (RR, 1.95; 95% CI, 1.47 to 2.57). The tPA-treated subjects who received higher-dose ALB were three times more likely to achieve a good outcome than subjects receiving lower-dose ALB, suggesting a positive synergistic effect between ALB and tPA.
Conclusions— Our data suggest that high-dose ALB therapy may be neuroprotective after ischemic stroke. These results have led to a multicenter, randomized, placebo-controlled efficacy trial of ALB in acute ischemic stroke—the ALIAS Phase III Trial.
Preclinical studies in rodent models of ischemic stroke have established that the administration of high doses of human albumin confers robust neuroprotection as assessed both neurobehaviorally and histopathologically.2–5 These observations led us to conduct a National Institutes of Health (NIH)-funded pilot clinical trial, the ALIAS (Albumin in Acute Stroke) Pilot Trial—an open-label, dose-escalation study designed primarily to evaluate the safety of administering moderate to high doses of 25% human albumin (ALB) to patients with acute ischemic stroke. In the companion article to this report,6 we present physiological data, laboratory variables, and safety outcomes in the 82 subjects studied. These results indicate that ALB doses of up to 2.05 g/kg body weight—a dose lying within the highly therapeutic dose range established preclinically2–4—could be safely administered without significant adverse experiences.6 The only ALB-related adverse event encountered was mild-to-moderate pulmonary edema, which occurred in approximately 13% of subjects but was readily manageable with diuretics.
In this report, we present the clinical outcome results, and we probe these data for suggestions of possible therapeutic efficacy.
Materials and Methods
The companion article6 presents the full details of the ALIAS Pilot Trial design and study subjects. In brief, the ALIAS Pilot Trial was an NIH-funded open-label, dose-escalation trial designed to establish the safety of administering substantial volumes of 25% ALB to subjects 18 years or older with moderate-to-severe (NIH Stroke Scale [NIHSS] score ≥6) acute ischemic stroke who were able to receive the treatment within 16 hours from symptom onset. The subjects were recruited at the University of Miami/Jackson Memorial Hospital in Miami, Fla, and the University of Calgary/Foothills Medical Centre in Calgary, Alberta, Canada. The Medical University of South Carolina provided data management and statistical analyses. Two patient cohorts were entered into the trial: (1) those who received standard-of-care intravenous (IV) tissue plasminogen activator (tPA); and (2) those who did not receive tPA. All subjects were treated with ALB within 16 hours of stroke onset and were followed for 3 months. In successive cohorts, six ALB dose tiers were studied, ranging from 0.34 to 2.05 g/kg body weight.
The NIHSS was evaluated at baseline; at 4, 12, 24, 48, and 72 hours posttreatment; at discharge; and 1 month posttreatment. At 3 months, the NIHSS, modified Rankin Scale (mRS), and Barthel Index were measured. For subjects who died before the 3-month assessment, we imputed their outcome scores with worst possible values, that is, 42, 6, and 0 for NIHSS, mRS, and Barthel Index, respectively. For subjects who were lost to follow up before the 3-month assessment, we carried forward the NIHSS scores from the 1-month assessment or, if that was also missing, from discharge, that is, we applied the last-observation-carried-forward (LOCF) imputation method. The raw mRS and Barthel Index scores at 3 months are missing for the latter subjects; however, their dichotomized mRS score (0 to 1 versus 2 to 6) and the dichotomized Barthel Index score (<95 versus ≥95) were imputed with 0 to 1 or ≥95, respectively, for those with a LOCF NIHSS score of 0 to 1, and otherwise for a LOCF NIHSS score >1.
Because the ALIAS Pilot Trial was an open-label study without concurrent controls, we used two analytical approaches in evaluating for signs of efficacy. First, we made statistical comparisons between subjects receiving the lowest three dose tiers (0.34, 0.68, and 1.03 g/kg) and those receiving the highest three dose tiers (1.37, 1.71, and 2.05 g/kg). ALB dose tiers I to III are presumed to be subtherapeutic based on the preclinically established effective ALB dose range of 1.25 to 2.50 g/kg.4 By contrast, dose tiers IV to VI are squarely within the preclinically established therapeutic range. Second, data from the higher dose-tier group were compared with outcome data obtained from a subset of the NINDS rt-PA Stroke Study.1 The subset consisted of subjects whose baseline NIHSS score was ≥6. The analyses conducted here were not prespecified but rather were conducted in a purely exploratory, post hoc fashion.
In the exploratory analysis that follows, we defined favorable outcome as an NIHSS score of 0 to 1, an mRS of 0 to 1, or both, at 3 months. We chose this composite outcome for the following reasons. In experimental studies, the “gold standard” of neuroprotective efficacy is the volume of brain tissue salvage. In patients with middle cerebral artery strokes, ischemic lesion volume correlates strongly with NIHSS score.7 On the other hand, outcome-preference surveys indicate that a majority of stroke victims or potential stroke victims most value a complete functional recovery.8,9 The latter is best reflected by an mRS score of 0 or 1.
Univariate analyses of the outcome data were conducted by using the nonparametric Wilcoxon rank sum test. Relative risk or risk ratios (RR) were obtained by applying the generalized linear model with log link.10 Because the Pilot Trial was a cohort study, the RR estimates are appropriate summary measures of the ALB effect.
The study population descriptors are presented in Table 3 of the companion article.6 Eighty-two subjects participated; 42 received tPA and 40 did not. The mean age was 65.2 years (standard deviation [SD] 1.7), and the proportions of males and females were approximately equal. In the tPA cohort, median time to standard-of-care IV tPA initiation was 2.5 hours after stroke onset, and 6.1 hours to ALB initiation (mean [SD], 6.5 [3.0] hours). In the non-tPA cohort, the median time to ALB initiation was 8.1 hours (mean [SD], 9.1 [3.3] hours). The baseline NIHSS score ranged from 6 to 37 with mean of 13.6 (SD 6.9). Six subjects (four in the tPA and two in the non-tPA cohort) died before the 3-month follow-up assessment, and two subjects (one in each cohort) were lost to follow up.
The distribution of the outcome measures by each dose tier and cohort is provided in Table 1. Figure 1A and 1B show the mean change from baseline NIHSS score over time for ALB dose tiers I to III and IV to VI, respectively, in the tPA and non-tPA cohorts.
The distributions of mRS and Barthel Index scores at 3 months by tier groups are illustrated in Figures 2 and 3⇓, respectively. The change from baseline NIHSS scores at 3 months by low-dose (tiers I to III) and high-dose (tiers IV to VI) ALB groups is displayed for the tPA and non-tPA cohorts in Figure 4. In the tPA cohort, the mean degree of improvement in the NIHSS score at 3 months was significantly greater (P=0.015) in subjects of dose tiers IV to VI (N=22) than in subjects of dose tiers I to III (N=20) (Figure 4). For the non-tPA cohort, the difference in respective values was not statistically significant.
Overall, the mean time from symptom onset to initiation of ALB therapy was 6.8 (SD 3.0) hours among subjects with a good outcome and 8.7 (SD 3.6) hours in those who did not have a good outcome. For every hour later that a subject was treated with ALB, his or her likelihood of having a good outcome was reduced by 9% (ie, RR, 0.91; 95% CI, 0.84 to 0.99). This was the case regardless of the ALB dose-tier group. In the tPA cohort, the mean time was 5.7 (SD 1.9) hours in the good-outcome group and 7.4 (SD 3.6) hours in the not-good outcome group. In the non-tPA cohort, the corresponding values were 8.1 (SD 2.6) and 10.0 (SD 3.0) hours, respectively.
We compared subjects of the present trial who received ALB dose tiers IV to VI (in the putatively therapeutic range) with outcome data from a subset of subjects in the NINDS tPA Stroke Study, Part 2,1 (N=146 in the tPA cohort; N=156 in the non-tPA cohort). Overall, the unadjusted probability of good outcome in the high ALB dose tiers was 97% greater than that for the comparable NINDS rt-PA Stroke Study cohort (RR, 1.97; 95% CI, 1.47 to 2.63). The RRs, adjusted for tPA, as well as adjusted for tPA, ALB×tPA interaction, baseline NIHSS score, and age, are provided in Table 2. These comparisons against historical controls (1) suggest that ALB therapy at dose tiers IV to VI is highly effective in improving outcome; and (2) confirm that tPA is also effective.
We also compared the data of the lower and higher ALB dose tiers in a similar manner (Table 3) under the assumption that the lower dose-tier subjects approximate a control group in that lower ALB doses are likely to be nontherapeutic.4 Again, overall, the unadjusted probability of good outcome was 77% greater at the higher ALB doses than in the lower dose tiers (RR, 1.77; 95% CI, 1.09 to 2.89). The RRs in Table 3 are adjusted for the tPA effect as well as for tPA, ALB×tPA interaction, baseline NIHSS, age, and time from symptom onset to ALB treatment initiation. Although the ALB effect remains highly significant in the tPA cohort, caution is advised in its interpretation as a result of the wide confidence intervals on the RRs arising from the much smaller sample sizes for this analysis than in the previous comparison.
In these analyses, the RRs and the absolute risk differences between the tPA and non-tPA cohorts are suggestive of an interaction effect between ALB and tPA.
The primary aim of the ALIAS Pilot Trial was to establish the safety of high-dose ALB administration in patients with acute ischemic stroke.6 As such, the trial was a nonrandomized, open-label study in which all subjects received ALB in escalating dosages; there was no untreated concurrent control group.
An additional constraint on the efficacy analysis was that we permitted a 16-hour time to ALB treatment to maximize recruitment under the assumption that the adverse effects of ALB would not be influenced by time to treatment. This resulted in the fact that the tPA cohort (having arrived within 3 hours of stroke onset) was available to receive ALB, on average, 2.6 hours earlier than the non-tPA cohort—in many instances, at time points that were putatively within ALB’s window of therapeutic efficacy.4 This difference is critically important in that preclinical results suggested that any neuroprotective effect of ALB would diminish significantly if ALB therapy were delayed by more than 4 to 5 hours from stroke onset.4 Thus, caution must be exercised when comparing neurologic outcomes in the tPA and non-tPA cohorts of this trial and, particularly, when assessing possible synergism between ALB and tPA.
Nevertheless, exploratory efficacy analyses of the tPA cohort, when conducted against historical controls from the NINDS rt-PA Stroke Study Part 2 Trial (Table 2), as well as when carried out by comparing low- versus high-dose ALB subjects of the present trial (Table 3), strongly support the view that the therapeutic effect of tPA is substantially enhanced when higher doses ALB are concurrently administered. The estimated RRs support an effect of ALB for the tPA cohort (Tables 2 and 3⇑). In the comparison of high ALB dose tiers with the NINDS rt-PA Study cohort (Table 2), we also observed a strong ALB effect in the non-tPA cohort. However, possibly because of the much longer time to ALB treatment initiation in the non-tPA cohort, which we believe to lie well outside of the 4- to 5-hour therapeutic window of ALB established in preclinical studies,4 it was not possible to observe an ALB effect in the non-tPA cohort (Table 2).
There are mechanistic data to support a pathophysiological basis for a possible synergistic effect of tPA and ALB. Namely, an important component of ALB’s salutary effects in cerebral ischemia is mediated within the microvasculature. Microvascular endothelial cells express several specific albumin-binding sites on their surface.11–13 By binding to the endothelial glycocalyx, albumin maintains the normal permeability of microvessel walls; and, by its transcytosis across endothelium, it serves as a carrier for various small molecules.12,14,15 Recent work suggests that albumin may be a factor mediating the effect of blood coagulation on vascular tone and capillary permeability. Albumin exerts complex influences on erythrocyte aggregation, increasing low-shear viscosity but decreasing erythrocyte sedimentation under low-flow conditions.16 In addition, serum albumin reacts with nitric oxide to form a stable S-nitrosothiol that has endothelium-derived relaxing factor-like properties.17
Albumin is also an important inhibitor of platelet aggregation.18–20 Albumin increases the production of the antiaggregatory prostaglandin, PGD2, from cyclic endoperoxides.19 It also binds platelet-activating factor (PAF) with high affinity21,22 and decreases PAF-induced responses in platelets.22 Coating of thrombogenic surfaces with S-nitrosylated albumin reduces platelet adhesion and aggregation—an effect attributable both to the direct antiplatelet actions of nitric oxide and the antiadhesive properties of albumin itself.18 In the Atherosclerosis Risk in Communities study of over 14 000 subjects followed for approximately 5 years, lower serum albumin levels were associated with an increased incidence of coronary heart disease at the baseline examination in current smokers.23 Hypoalbuminemia was reported to reduce platelet aggregability in patients undergoing ambulatory peritoneal dialysis.24
In our preclinical studies of focal ischemia, ALB therapy improved local cerebral perfusion in ischemic regions having critically reduced cerebral blood flow.25,26 In studies using laser-scanning confocal microscopy of the cortical vasculature in focal ischemia, we found that during the first 15 to 30 minutes of postischemic recirculation, prominent foci of vascular stasis developed within cortical venules, associated with thrombus-like stagnant foci and adherent intravenular corpuscular structures (thought to be activated neutrophils adhering to the venular endothelium).27 Administration of intravenous ALB (2.5 g/kg) was followed by a prompt improvement of venular flow and disappearance of adherent corpuscles and thrombotic material, whereas saline administration did not affect these phenomena.27
Taken together, these data permit the speculation that ALB therapy might act to maintain microvascular patency, enhance local perfusion, and discourage rethrombosis after tPA-induced reperfusion. This hypothesis is amenable to mechanistic exploration in animal models of focal ischemia.
Conclusions and Future Directions
The ALIAS Pilot Trial has demonstrated the safety and feasibility of administering high-dose 25% ALB solution to patients with acute ischemic stroke. In addition, despite our relatively small sample size, strong preliminary suggestions of therapeutic efficacy emerged.
Based on these results and supported by an NIH Planning Grant, we have designed, and received NIH funding for, the ALIAS Phase III Trial—a large (N=1800), randomized, double-blind, placebo-controlled multicenter clinical trial designed to ascertain definitively the therapeutic efficacy of ALB in acute ischemic stroke. Funding began in September 2005. The inclusion and exclusion criteria closely resemble those of the present ALIAS Pilot Trial except for the requirement that infusion of the study drug (2.0 g/kg of 25% ALB or a comparable volume of isotonic saline) be begun within 5 hours of stroke onset. A standard-of-care tPA cohort and a non-tPA cohort will again be studied separately as a result of the potential differential magnitude of the ALB effect in these two cohorts. The primary outcome measure is a favorable neurologic outcome, defined as NIHSS score of 0 to 1, and/or mRS score of 0 to 1, at 3 months postrandomization. Subjects will be followed out to 1 year for quality-of-life assessment. The trial is estimated to conclude in 2010.
As described in the companion article to this report,6 subjects of the ALIAS Pilot Trial who received tPA had a lower incidence of symptomatic intracerebral hemorrhage than occurred in the NINDS rt-PA Stroke Study.1 The ALIAS Phase III Trial will allow us to examine whether ALB therapy, in fact, reduces symptomatic intracerebral hemorrhage after tPA.
University of Miami: Ludmila Belayev, MD; Eduardo DeMarchena, MD; Alexandre C. Ferreira, MD; Alejandro M. Forteza, MD; Myron D. Ginsberg, MD; Kathy Hesse, RN; J. Andres Hidalgo, MD; Sebastian Koch, MD; Alejandro Rabinstein, MD; Jose G. Romano, MD; and Diego Tamariz, MD.
University of Calgary: Philip A. Barber, MB, ChB; Alastair M. Buchan, MB, BCh; Andrew M. Demchuk, MD; Michael D. Hill, MD, MSc; Albert J. Kryski, Jr., MD, PhD, FRCPC; Karla J. Ryckborst, BA, RN; Tim Watson, MD; and the Calgary Stroke Team.
Medical University of South Carolina: Catherine Dillon, BS; Elizabeth Gieseke; Renee H. Martin, PhD; Yuko Y. Palesch, PhD; Julia Pauls, MES; Jon Taie; Barbara C. Tilley, PhD; and James A. Vaughan, MES.
Safety Evaluation Committee
Alastair M. Buchan, MB, BCh (neurology, Calgary); Alexandre C. Ferreira, MD (cardiology, Miami); Alejandro M. Forteza, MD (neurology, Miami); Myron D. Ginsberg, MD (neurology, Miami); Neal S. Kleiman, MD (cardiology, Baylor, Houston); and Stephan A. Mayer, MD (neurology, Columbia University).
Data Safety and Monitoring Board
Harold Adams, MD, University of Iowa (Chair); Shunichi Homma, MD, Columbia University; George Howard, PhD, University of Alabama at Birmingham; Amin Kassam, MD, University of Pittsburgh; Walter J. Koroshetz, MD, Massachusetts General Hospital; Derk W. Krieger, MD, Cleveland Clinic Foundation; and Claudia S. Moy, PhD, NINDS/NIH.
Sources of Funding
This study was supported by NIH Pilot Clinical Trial Grant NS 40406 (M.D.G.) and NIH Planning Grant NS 48784 (M.D.G.).
- Received March 24, 2006.
- Accepted April 13, 2006.
Belayev L, Zhao W, Pattany PM, Weaver RG, Huh PW, Lin B, Busto R, Ginsberg MD. Diffusion-weighted magnetic resonance imaging confirms marked neuroprotective efficacy of albumin therapy in focal cerebral ischemia. Stroke. 1998; 29: 2587–2599.
Belayev L, Liu Y, Zhao W, Busto R, Ginsberg MD. Human albumin therapy of acute ischemic stroke: marked neuroprotective efficacy at moderate doses and with a broad therapeutic window. Stroke. 2001; 32: 553–560.
Ginsberg MD, Hill MD, Palesch YY, Ryckborst KJ, Tamariz D. The ALIAS Pilot Trial: a dose-escalation and safety study of albumin therapy for acute ischemic stroke. I. Physiological responses and safety results. Stroke. 2006; 37: 2100–2106.
Schiemanck SK, Post MW, Witkamp TD, Kappelle LJ, Prevo AJ. Relationship between ischemic lesion volume and functional status in the 2nd week after middle cerebral artery stroke. Neurorehabil Neural Repair. 2005; 19: 133–138.
Solomon NA, Glick HA, Russo CJ, Lee J, Schulman KA. Patient preferences for stroke outcomes. Stroke. 1994; 25: 1721–1725.
Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol. 2005; 162: 199–200.
Ghitescu L, Fixman A, Simionescu M, Simionescu N. Specific binding sites for albumin restricted to plasmalemmal vesicles of continuous capillary endothelium: receptor-mediated transcytosis. J Cell Biol. 1986; 102: 1304–1311.
Schnitzer JE, Sung A, Horvat R, Bravo J. Preferential interaction of albumin-binding proteins, gp30 and gp18, with conformationally modified albumins. Presence in many cells and tissues with a possible role in catabolism. J Biol Chem. 1992; 267: 24544–24553.
Ammit AJ, O’Neill C. Studies of the nature of the binding by albumin of platelet-activating factor released from cells. J Biol Chem. 1997; 272: 18772–18778.
Nelson JJ, Liao D, Sharrett AR, Folsom AR, Chambless LE, Shahar E, Szklo M, Eckfeldt J, Heiss G. Serum albumin level as a predictor of incident coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) study. Am J Epidemiol. 2000; 151: 468–477.
Belayev L, Liu Y, Busto R, Zhao W, and Ginsberg MD. Sequential laser-Doppler perfusion imaging after cerebral ischemia in rats and mice; effect of albumin. Soc Neurosci Abstr. 2000; 26: 1032(No. 383.2).
Belayev L, Pinard E, Nallet H, Seylaz J, Liu Y, Riyamongkol P, Zhao W, Busto R, Ginsberg MD. Albumin therapy of transient focal cerebral ischemia: in vivo analysis of dynamic microvascular responses. Stroke. 2002; 33: 1077–1084.