The ALIAS Pilot Trial
A Dose-Escalation and Safety Study of Albumin Therapy for Acute Ischemic Stroke—I: Physiological Responses and Safety Results
Background and Purpose— In preclinical stroke models, high-dose human albumin confers robust neuroprotection. We investigated the safety and tolerability of this therapy in patients with acute ischemic stroke.
Methods— The ALIAS (Albumin in Acute Stroke) Pilot Clinical Trial used a multiple-tier, open-label, dose-escalation design. Subjects with acute ischemic stroke (NIH Stroke Scale [NIHSS] of 6 or above) received a 2-hour infusion of 25% human albumin (ALB) beginning within 16 hours of stroke onset. Six successive ALB dose tiers were assessed ranging from 0.34 to 2.05 g/kg. Neurologic and cardiac function was sequentially monitored. At 3 months, the NIHSS, modified Rankin Scale, and Barthel Index were measured.
Results— Eighty-two subjects (mean age, 65 years) received ALB at 7.8±3.4 hours after stroke onset (mean±standard deviation). Forty-two patients also received standard-of-care intravenous tissue plasminogen activator (tPA). Vital signs were unaltered by ALB treatment. Dose-related increases in plasma albumin and mild hemodilution were maximal at 4 to 12 hours. Age-related plasma brain natriuretic peptide levels increased at 24 hours after ALB but did not predict cardiac adverse events. The sole ALB-related adverse event was mild or moderate pulmonary edema in 13.4% of subjects, which was readily managed with diuretics. In the tPA-treated subgroup, symptomatic intracranial hemorrhage occurred in only one of 42 subjects.
Conclusions— Twenty-five percent human albumin in doses ranging up to 2.05 g/kg was tolerated by patients with acute ischemic stroke without major dose-limiting complications. tPA therapy did not affect the safety profile of ALB. The companion article presents neurologic outcome data and efficacy analysis in these subjects.
Multiple biochemical and molecular pathways mediate brain injury in ischemic stroke. Experimental studies using a great variety of neuroprotective treatment strategies have established the principle that ischemic tissue can indeed be protected from injury by prompt intervention.1,2 Unfortunately, however, the successful translation of these potential strategies to the clinic has not yet occurred.3
In extensively replicated laboratory studies conducted in a standardized rodent model of focal ischemic stroke (2-hour middle cerebral artery occlusion by the modified intraluminal suture method),4 we have shown that the prompt administration of high-dose 25% human serum albumin (ALB) is markedly neuroprotective as reflected in improved neurologic score, decreased volume of brain infarction by histopathology, and reduced brain swelling.5–8 In addition, we found that ALB therapy improves cerebral perfusion,8,9 normalizes changes on diffusion-weighted magnetic resonance imaging,6 reverses postischemic microvascular stasis,10 and contributes to the systemic mobilization and supply of free fatty acids to the postischemic brain.11 In these studies, ALB doses of 1.25 g/kg to 2.5 g/kg were markedly neuroprotective, with a therapeutic window of 4 to 5 hours.7 In a pooled analysis of these series, cortical infarcts lying at or below the median size in controls were reduced in size by over 99% by ALB therapy.12 Even when ALB therapy was markedly delayed, it was not detrimental.13 In all of these studies, ALB treatment was completely benign, that is, devoid of adverse side effects.
These preclinical findings formed the rationale for a pilot-phase safety and feasibility study in humans, the results of which we report here.
Materials and Methods
A National Institutes of Health-funded pilot clinical trial, the ALIAS (Albumin in Acute Stroke) Pilot Trial, was initiated in August 2001 to investigate the intravenous use of 25% ALB in subjects with acute ischemic cerebral infarction. The primary objective was to use a multiple-tier, open-label, dose-escalation design to establish the safety of this therapy in acute ischemic stroke. The trial’s aim was to ascertain whether subjects with acute ischemic stroke would tolerate per-kilogram doses of ALB shown to be neuroprotective in preclinical studies without experiencing cardiovascular or neurologic complications or other adverse events. The trial’s secondary objective was to gain experience in implementing standardized measures for assessing neurologic deficit, cardiovascular status, and neurologic outcome as a prelude to future phase II–III trials. The trial was conducted at two North American clinical sites: the University of Miami/Jackson Memorial Hospital and the University of Calgary/Foothills Medical Centre, in Calgary, Alberta, Canada. The biostatistical and data management site was at the Medical University of South Carolina.
Two patient subgroups were entered into the trial: (1) subjects who received standard-of-care intravenous tissue plasminogen activator (tPA) as well as ALB; and (2) subjects who did not receive tPA. All subjects were treated with ALB within 16 hours of stroke onset. They were followed for 3 months for the assessment of their NIH Stroke Scale (NIHSS) score, modified Rankin Scale (mRS), and Barthel Index.
The inclusion criteria were: (1) history of a focal neurologic syndrome of acute onset and at least a moderate initial neurologic deficit, defined as an NIHSS score of 6 or greater; (2) computed tomography (CT) or magnetic resonance scan excluding hemorrhage; (3) ability to commence the intravenous infusion of 25% ALB within 16 hours after the onset of stroke symptoms (defined as the time at which the onset of a neurologic abnormality was observed or, with symptoms first noticed on awakening from sleep, the last time the patient was observed to be without stroke symptoms); (4) age 18 years or above; and (5) informed consent executed by the patient, responsible family member, or legally authorized representative. The exclusion criteria are listed in Table 1.
The time of stroke onset was established by a stroke service neurologist, who conducted a complete medical history, physical examination, and baseline neurologic evaluation (including the NIHSS and a historical determination of the mRS). A complete cardiologic evaluation, including clinical examination and electrocardiogram, was conducted by a cardiologist or internist. (In the initial portion of the trial, an echocardiogram was also conducted at baseline.)
Baseline CT scans of the brain were performed on admission to exclude intracranial hemorrhage and to provide baseline information as to early hypodensities and/or brain swelling compatible with ischemic stroke. CT scans were scored using the Alberta Stroke Programme Early CT Score (ASPECTS)—a well-validated scale that exhibits a strong inverse correlation with stroke severity on the NIHSS.14 CT scans and ASPECTS scoring were repeated at 24 hours.
Pretreatment laboratory tests included hemoglobin, hematocrit, white blood cell count with differential, and platelet count; plasma blood urea nitrogen, creatinine, potassium, sodium, chloride, bicarbonate, and glucose; activated partial thromboplastin time; prothrombin time; serum albumin, globulin and fibrinogen levels; creatine kinase (CK) and isoenzymes if abnormal; serum osmolality; complete urinalysis; and chest x-ray.
Our preclinical studies established that ALB was robustly neuroprotective at doses of 2.0 to 2.5 g/kg body weight5,6 and that an equivalent degree of neuroprotection could also be achieved at one- half of that dose, or 1.25 g/kg.7 Translated to humans, the dose range of 1.25 to 2.5 g/kg corresponds to volumes of 25% human albumin solution of approximately 350 to 700 mL. Accordingly, our dose-escalation design began at 100 mL and ranged up to 600 mL for a typical (73-kg) subject. Six ALB dose tiers were used: tier I, 0.34 g ALB per kg body weight; II, 0.68; III, 1.03; IV, 1.37; V, 1.71; and VI, 2.05. The ALB (25% human albumin solution; Baxter BioScience) was infused intravenously over 2 hours.
In January 2004 (after the first 52 subjects had been enrolled), the study protocol was amended to include the on-protocol prophylactic administration of the diuretic furosemide, 10 to 20 mg intravenously, at 24 hours after ALB infusion. The rationale for this measure was to treat prophylactically those patients at higher risk of symptomatic congestive heart failure. The 24-hour time point was chosen to avoid early volume depletion, which might exacerbate brain ischemia. In practice, the investigator was permitted to withhold furosemide at his or her discretion based on the physical examination and chest x-ray findings.
Monitoring and Follow-Up Evaluations
Vital signs were monitored at 15-minute intervals throughout the ALB infusion period, at 30-minute intervals over the next 24 hours, and at 4-hour intervals out to 72 hours. Subjects were reevaluated cardiologically at 4 hours, 12 hours, 24 hours, and 48 hours. Brain CT scans and chest x-ray films were repeated at 24 to 48 hours and at additional time points if clinically indicated. NIHSS was evaluated at 4, 12, 24, 48, and 72 hours posttreatment and at discharge. Other laboratory measures were periodically reassessed. At 1 month and 3 months, NIHSS, mRS, and the Barthel Index were measured.
All study data, including all adverse events and concomitant medications, were collected on case report forms and independently double-key-entered into a database at the Medical University of South Carolina, where range verification, consistency checks, quality assurance, and backup were performed.
The ALIAS Pilot Trial was designed to determine whether per-kilogram doses of ALB shown to be neuroprotective in animal studies could be tolerated by subjects with acute ischemic stroke without the occurrence of unacceptable cardiac or other adverse events. The study design was based on the method proposed by Storer,15 in which a cohort of subjects is treated at a given level and their toxicity experience is assessed before escalating to the next level if deemed appropriate. Our approach did not incorporate explicit a priori stopping rules.
At each ALB dose tier, our intent was to enroll approximately six subjects who also received tPA and approximately six subjects who did not, and the study was initially structured to permit separate analysis of ALB-related adverse events in the tPA and non-tPA cohorts at each ALB dose tier. The safety adjudications were performed by a Safety Evaluation Committee (SEC) composed of four neurologists and two cardiologists (Appendix) once each group of approximately six subjects reached the 72-hour time point. As the trial progressed, it became evident that the adverse-event profile did not differ in the tPA and non-tPA cohorts, so these groups were merged for adjudication. A Data Safety and Monitoring Board (DSMB) constituted by the NINDS Clinical Trials Group considered the SEC’s adjudications at each dose tier and decided on escalation to the next dose tier.
The study population descriptors are shown in Table 2. Among the 82 subjects, the mean age was 65 years; males and females were equally represented. Sixty subjects were non-Hispanic white; 9 were Hispanic, 11 black, one Asian, and one Native American. The mean initial NIHSS score was 13.1. The subjects’ stroke subtype classification according to the TOAST definitions16 were: 24 (29%) undetermined, 22 (27%) cardioembolic, 16 (20%) large artery, 15 (18%) small vessel/lacunar, 4 (5%) other determined etiology, and one (1%) two or more etiologies. Forty-two subjects received standard-of-care intravenous tPA, which was begun at 2.4±0.8 hours after stroke onset; in these subjects, ALB infusion was begun at 6.5±3.0 hours (mean±standard deviation [SD]). Forty subjects did not receive tPA; in these subjects, ALB infusion was begun at 9.1±3.3 hours after stroke onset. Within each dose tier, there were no differences in the ALB volumes administered in the tPA and non-tPA cohorts.
Physiological and Laboratory Variables
Mean blood pressure after ALB administration in the six dose tiers is shown in Figure 1. A tendency for mean arterial pressure (MAP) to increase slightly over the first 5 days (by approximately 2 mm Hg) was noted in the higher dose tiers; this was clinically inconsequential. Pulse rate, respirations, and body temperature were unaltered by ALB infusion at any dose tier.
Baseline plasma albumin levels averaged 3.8±0.4 g/dL (mean±SD). ALB infusion led to a dose-related increase in plasma albumin, most marked at 4 hours postinfusion and declining toward preinfusion levels at 48 to 72 hours (Figure 2). In dose tier VI, the increase at 4 hours averaged 2.0 g/dL above baseline. Correspondingly, ALB administration produced a dose-dependent hemodilution, maximal at 4 to 12 hours but still evident at 48 hours postinfusion (Figure 2). In tier VI, the decrease in hematocrit at 4 hours averaged 9.8 points (23%) below preinfusion values.
Supplemental Table I (available online at http://stroke.ahajournals.org) summarizes laboratory values at baseline and 24 hours after ALB treatment. Apart from the alterations described here, ALB induced no important changes. At echocardiography, six patients had an estimated ejection fraction <60% (ranging from 40 to 55%). Mild (n=18) or moderate (n=2) aortic insufficiency was observed in 20 patients. Mild (n=28) or moderate (n=3) mitral insufficiency was observed in 31 patients. One patient had a compensated flail mitral valve posterior leaflet with moderate mitral insufficiency. Two patients had left ventricular hypertrophy. There was no apparent relationship between echocardiographic findings and subsequent evolution of congestive heart failure/pulmonary edema.
Plasma brain natriuretic peptide (BNP) was measured in subjects of tiers IV, V, and VI before ALB infusion (n=40) and 24 hours later (n=38). Elevated BNP levels are thought to reflect increased cardiac filling pressure.17 In all subjects, BNP levels increased after ALB administration, and subjects’ age was highly correlated with BNP, both at baseline and post-ALB infusion (Figure 3). However, there was no correlation between BNP (either pre- or post-ALB infusion) and initial stroke severity (NIHSS score). Interestingly, there was also no correlation between the extent of BNP increase and the presence or absence of cardiac adverse events (described subsequently). The extent of increase in plasma BNP levels post-ALB did not differ in subjects with versus without cardiac adverse events.
The only observed adverse events bearing a relationship to ALB therapy in this trial (comprising “unlikely,” “possible,” “probable,” or “definite”) were those associated with intravascular volume expansion: shortness of breath and frank pulmonary edema/congestive heart failure in the hours or days after ALB administration. Eleven of the 82 subjects (13.4%) had clinical signs of pulmonary edema, which was of mild-to-moderate severity. These subjects were evenly distributed between the tPA and non-tPA cohorts (Table 3). At dose tiers I and II, these adverse events were only rarely observed (one of 28 subjects). At dose tier III (1.03 g/kg ALB), two of 12 subjects developed mild-to-moderate pulmonary edema, and the SEC cardiologists suspected that we were beginning to enter a dose range that produced evidence of mild pulmonary congestion, at least in elderly subjects with altered ventricular compliance or pulmonary systolic hypertension. At this point, we amended the protocol to require chest x-rays at both 24 and 48 hours after ALB administration; and we introduced plasma BNP measurements at baseline and 24 hours. At the request of the DSMB, we expanded the number of subjects in dose tier IV to 18. Of these 18 subjects, clinical evidence of pulmonary edema was observed in five cases with chest x-ray evidence in four; three of these subjects had received diuretics. Beginning with dose tier V (1.37 g/kg ALB), the prophylactic administration of furosemide at 24 hours was recommended (Table 3). In the 12 subjects of dose tier V, very mild clinical signs of pulmonary congestion/edema occurred in two subjects, one of whom also had chest x-ray changes. Prophylactic furosemide was administered to four of these 12 subjects. Twelve subjects were enrolled in the final dose tier, tier VI (2.05 g/kg). One tier VI subject developed clinical and chest x-ray signs of mild pulmonary edema, which was easily managed with two doses of furosemide. None of the other 11 tier VI subjects received furosemide. Four subjects of tier VI showed mild interstitial edema on chest x-ray without clinical findings. Overall, nine of 82 subjects were treated with furosemide; two subjects were not diuresed because symptoms were very mild and resolved spontaneously. Three of the nine furosemide-treated subjects required two or more doses for diuresis.
One patient with known coronary artery disease developed evidence of cardiac ischemia several hours after albumin administration. On electrocardiogram, lateral ST segment depression was noted. This reversed with administration of nitroglycerin. There was no elevation in serum cardiac troponin levels over the next 24 hours. A subsequent stress thallium cardiac perfusion study was unremarkable for any reversible ischemic deficit. It was presumed that left ventricular wall stress was increased by albumin administration, resulting in angina pectoris.
Serious adverse events (SAEs) were defined as those resulting in: death from any cause, a life-threatening adverse experience, prolongation of hospitalization, persistent or significant disability, or an important medical event requiring medical or surgical intervention to prevent one of the previously mentioned outcomes. In this trial, SAEs occurred in 17 subjects (21%) (Table 4), but only one SAE was adjudicated to be possibly related to ALB administration—in a subject in ALB dose tier III with paroxysmal atrial fibrillation (presumed resulting from left atrial stretching from volume expansion) and pulmonary edema resulting in prolonged hospitalization.
In the opinion of the trial’s safety monitors, no other adverse events were observed that bore a conceivable relationship to ALB. For example, no allergic reactions to ALB were noted.
Only one subject of the series developed symptomatic intracranial hemorrhage: an 80-year-old patient with striatocapsular infarction who received tPA and ALB (dose tier I); at 24 hours, he developed hemorrhagic conversion (PH 1 in the classification of Pessin et al18) and intraventricular hemorrhage. Thus, the overall incidence of symptomatic ICH in the tPA cohort of this trial was one of 42 subjects, or 2.4% (95% confidence interval, 0.06 to 12.6)—an incidence below the 6.4% incidence reported in the NINDS tPA trial.19
The primary intent of ALIAS Pilot Trial was to establish the safety of administering substantial volumes of 25% ALB to patients with acute ischemic stroke. This objective was achieved; we succeeded in administering ALB doses of up to 2.05 g/kg to acute stroke subjects without encountering significant dose-limiting adverse events. The sole ALB-related complication in these subjects was the occurrence of clinical signs of mild-to-moderate pulmonary edema in approximately 13% of subjects. Its onset was not immediate; rather, it tended to commence typically at approximately 24 hours post-ALB infusion, a probable consequence of delayed intravascular volume expansion triggered by the oncotic load. In all cases, this complication was readily manageable. Prophylactic administration of furosemide at 24 hours was recommended for consideration at the higher dose tiers (Table 3), however, in practice, was needed in only a minority of subjects. The benign ALB safety profile observed in this pilot clinical trial agrees with our preclinical experience in rodent ischemia models. As the trial progressed, it became evident that the more stringent cardiac precautions taken at the trial’s inception (eg, echocardiographic exclusion criteria) could be relaxed. Adherence to the clinical exclusion criteria was sufficient to identify and exclude patients at high risk of congestive heart failure.
As noted earlier, our preclinical studies of ALB in focal ischemic stroke showed that ALB doses of 1.25 to 2.50 g/kg body weight were highly neuroprotective.5–8 The ALB dose tiers IV, V, and VI (1.37 to 2.05 g/kg) used in the present dose-escalation clinical trial fall squarely within the therapeutically efficacious dose range demonstrated preclinically. The ability for ischemic stroke subjects to tolerate ALB doses equivalent to those shown to be highly therapeutic in animals is a result of great importance. In several previous clinical trials of preclinically promising neuroprotectants, a major impediment to success was the inability for patients to tolerate drug doses shown to be neuroprotective in animal models.20
An important strength of the ALIAS Pilot Trial is that it demonstrated safety of ALB therapy in a diverse cohort of ischemic stroke subjects that was highly representative of the stroke population at large with respect to gender, ethnicity, age range (25 to 88 years), initial stroke severity (NIHSS range, 6 to 31), and stroke etiologies. Numbers of subjects treated versus not treated with standard-of-care intravenous tPA were approximately equal.
The main physiological consequences of ALB administration were a transient dose-dependent hemodilution and elevation of plasma albumin levels (Figure 2) and an age-related post-ALB elevation of plasma BNP levels (Figure 3), which, however, was not predictive of cardiac adverse events. There were no consequential changes in blood pressure or heart rate resulting from ALB administration, and no alterations of plasma electrolytes, renal function, or coagulation variables. Patients with significant cardiac dysfunction were successfully screened out by history, and this was confirmed by the mild age-related echocardiographic changes in the majority of patients.
In summary, the ALIAS Pilot Trial has demonstrated that ALB doses shown to be neuroprotective in preclinical studies may be safely administered to acute stroke patients without major dose-limiting complications. Concomitant tPA therapy did not affect the safety profile of ALB. The main adverse event, mild-to-moderate pulmonary edema, was infrequent and readily manageable. In the companion article, we present neurologic outcomes and suggestions of neuroprotective efficacy.
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 NS48784 (M.D.G.).
- Received March 24, 2006.
- Accepted April 13, 2006.
Ginsberg MD. Neuroprotection in brain ischemia—an update—parts I and II. Neuroscientist. 1995; 1: 95–103, 164–175.
Cheng YD, Al Khoury L, Zivin JA. Neuroprotection for ischemic stroke: two decades of success and failure. NeuroRx. 2004; 1: 36–45.
Belayev L, Alonso OF, Busto R, Zhao W, Ginsberg MD. Middle cerebral artery occlusion in the rat by intraluminal suture. Neurological and pathological evaluation of an improved model. Stroke. 1996; 27: 1616–1622.
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.
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.
Ginsberg MD, Belayev L, Bazan NG, Marcheselli VL, Hill MD, Palesch YY, Khoutorova L, Rodriguez de Turco EB, Ryckborst K, Tamariz D, Busto R. Albumin-based neurotherapeutics for acute ischemic stroke: from bench to bedside. In: Krieglstein J, Klumpp S, eds. Pharmacology of Cerebral Ischemia 2004. Stuttgart: Medpharm Scientific Publishers; 2004.
Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE III. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993; 24: 35–41.
Pessin MS, Teal PA, Caplan LR. Hemorrhagic infarction: guilt by association? AJNR Am J Neuroradiol. 1991; 12: 1123–1126.
Recommendations for clinical trial evaluation of acute stroke therapies. Stroke. 2001; 32: 1598–1606.