The Albumin in Subarachnoid Hemorrhage (ALISAH) Multicenter Pilot Clinical Trial
Safety and Neurologic Outcomes
Background and Purpose—Human albumin has been shown to exert neuroprotective effects in animal models of cerebral ischemia and humans with various intracranial pathologies. We investigated the safety and tolerability of 25% human albumin in patients with subarachnoid hemorrhage.
Methods—The Albumin in Subarachnoid Hemorrhage (ALISAH) Pilot Clinical Trial was an open-label, dose-escalation study. We intended to study 4 different dosages of albumin of increasing magnitude (0.625 g/kg: Tier 1; 1.25 g/kg: Tier 2; 1.875 g/kg: Tier 3; and 2.5 g/kg: Tier 4). Each dosage was to be given to 20 adult patients. Treatment was administered daily for 7 days. We investigated the maximum tolerated dose of albumin based on the rate of severe-to-life-threatening heart failure and anaphylactic reaction and functional outcome at 3 months.
Results—We treated 47 adult subjects: 20 in Tier 1; 20 in Tier 2; and 7 in Tier 3. We found that doses ranging up to 1.25 g/kg/day×7 days were tolerated by patients without major dose-limiting complications. We also found that outcomes trended toward better responses in those subjects enrolled in Tier 2 compared with Tier 1 (OR, 3.0513; CI, 0.6586–14.1367) and with the International Intraoperative Hypothermia for Aneurysm Surgery Trial cohort (OR, 3.1462; CI, 0.9158–10.8089).
Conclusions—Albumin in doses ranging up to 1.25 g/kg/day×7 days was tolerated by patients with subarachnoid hemorrhage without major complications and may be neuroprotective. Based on these results, planning of the ALISAH II, a Phase III, randomized, placebo-controlled trial to test the efficacy of albumin, is underway.
Subarachnoid hemorrhage (SAH) is a neurological emergency that carries high morbidity and mortality. Half of the approximately 30 000 Americans who have SAH every year will die and one third of survivors will be left dependent.1 Significant advances in the diagnosis, management, and prevention of complications in patients with SAH have occurred in the past few decades. However, improvement of clinical outcome has been modest. Newer treatments with proven neuroprotective benefits on clinical outcomes are needed. Nimodipine is the only available treatment with proven marginal efficacy.2
Administration of high doses of 25% human albumin (ALB) has been associated with improved outcomes and reduced costs in patients with SAH.3 Furthermore, several preclinical and clinical studies demonstrated the role of ALB as a neuroprotective treatment in cerebral ischemia.4–9 These findings formed the rationale for the design of a Phase I pilot study investigating the safety and tolerability of ALB in patients with SAH.10
Materials and Methods
The National Institutes of Health funded the Albumin in Subarachnoid Hemorrhage (ALISAH) pilot study, initiated in May 2006 and terminated in May 2010. The study was originally planned for 3 years but mostly due to the principal investigators transferring institutions and initiation of 2 non-US sites, 1 extra year was needed. ALISAH investigated the intravenous administration of ALB in subjects with SAH. The primary objective of this prospective dose-finding pilot study was to demonstrate the tolerability and safety of 4 dosages of ALB in patients with SAH. For each dosage group, 20 patients who met the eligibility criteria were to be enrolled. Maximum tolerated dosage of ALB determination was based on the rate of treatment-related severe or life-threatening heart failure.10 Secondary objectives were to obtain preliminary estimates of the ALB treatment effect using incidence of neurological deterioration within 15 days of symptom onset as well as incidence of rebleeding, hydrocephalus, seizures, delayed cerebral ischemia (DCI), and vasospasm (both symptomatic and by transcranial Doppler ultrasound criteria) within 15 days after symptom onset. In addition, the Glasgow Outcome Scale, Barthel Index, modified Rankin Scale, National Institutes of Health Stroke Scale, and Stroke Impact Scale at 3 months after onset of symptoms were collected to assess residual neurological deficits.
ALISAH was conducted at 6 North American sites (Supplemental Table I; http://stroke.ahajournals.org). The Central Coordinating Center was at the Baylor College of Medicine. The Statistical and Data Management Center was at the Medical University of South Carolina in Charleston, SC. Each site obtained Institutional Review Board/Ethics Committee approval, and the study was registered at www.clinicaltrials.gov before patient recruitment.
All patients presenting with SAH to the participating clinical sites were screened for study eligibility10 (Supplemental Table II). A cardiologist or cardiology fellow performed cardiological evaluations and a qualified neurological surgeon or intensive care unit physicians performed a neurological assessment on study entry. Once informed written consent was obtained, patients were enrolled into the study.
Subjects were allocated in a dose-escalation design into 1 of 4 dosage groups of 25% ALB: 0.625 g/kg/day×7 days; 1.25 g/kg/day×7 days; 1.875 g/kg/day×7 days; and 2.5 g/kg/day×7 days. The tolerability and safety of the previous dosage tier of ALB was evaluated by the National Institutes of Health-appointed Data and Safety Monitoring Board before advancing to the next dosage. The estimated volume of infusion for a 70-kg subject ranged from 175 to 700 mL, and the delivery time was 3 hours for all dosage groups. Dosage ranges of ALB were chosen based on the range administered to patients for induction of hypervolemia and hyperdynamia after SAH including our preliminary data3,10 and dosage ranges studied in animal models of focal cerebral ischemia.7 The duration of treatment (7 days) was chosen on the premise that it would cover the highest risk period for DCI (peak occurrence at Days 8–10). Because the half-life of human ALB is approximately 21 days with a degradation rate of 3.7% per day and a transcapillary escape rate of 4% to 5% per hour,10,11 we expected that the volume repletion and other neuroprotective effects of 25% ALB would persist for ≥24 hours beyond the 7 days. In addition, animal models of focal cerebral ischemia have shown significantly higher plasma oncotic pressures and serum ALB concentrations up to 3 days after treatment.5
Patient Management During the Acute Period
All subjects had vital signs monitored hourly and were assessed by intensive care unit nursing staff and/or site investigators for any new episode of neurological or cardiovascular deterioration. In addition, subjects had daily laboratory evaluations, assessment of neurological function (Glasgow Coma Scale, National Institutes of Health Stroke Scale), transcranial Doppler ultrasound, complete cardiac examinations, evaluation of intravascular volume status, and monitoring of vital signs (including weight) for 15 days or until hospital discharge if before that.
All subjects were treated with maintenance fluids of 0.9% normal saline at 80 to 125 mL/hr (total 2–3 L per day) with the goal central venous pressure between 5 and 8 mm Hg. In the event that subjects required further fluid administration beyond their maintenance to maintain the desired central venous pressure, they received 250 to 500 mL intermittent intravenous boluses of normal saline as needed. Extra ALB administration was not allowed, and the initial fluid balance goal was set at <+2000 mL/day.
Neurological deterioration after treatment was defined as a decline in >2 points in the Glasgow Coma Scale.10,12 Investigators determined causes for such deterioration. Acute left heart failure was defined as pulmonary edema occurring during the 7 days and up to 48 hours of treatment administration.10,13–15 All neurological and cardiovascular complications were managed according to a prespecified management protocol.
Patients returned to the clinic 90 days after enrollment. During this follow-up visit, subjects underwent a head CT and were assessed using the Glasgow Outcome Scale, modified Rankin Scale, National Institutes of Health Stroke Scale, Barthel Index, and the Stroke Impact Scale. Information on adverse events and concomitant medications since the time of discharge was also collected. After the 90-day evaluation, subjects were discharged from the study.
Data management was handled by the Statistical and Data Management Center at the Medical University of South Carolina. The Statistical and Data Management Center developed the Case Report Forms with input from the Central Coordinating Center. The clinical site staff entered data electronically into the database through the WebDCU System, a user-friendly menu-driven system with built-in warnings and rules to facilitate the data collection process and ensure sufficient quality control.
Statistical Considerations and Safety Monitoring
Sample size consideration for this Phase I dose-escalation study was based on the feasibility of recruiting patients in a 3-year study period at 5 sites.10 The recruitment yield would be a maximum of 80 patients or 20 patients per dosage group. Statistical analyses were mainly descriptive.
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 1 of the previously mentioned outcomes. Safety guidelines for escalation to the next ALB dosage level were based on the rate of cardiovascular SAEs defined as severe or life-threatening heart failure considered to be related (probably, possibly, and definitely) to ALB treatment.10,16 On completion of each ALB dosage level, if ≥2 subjects out of 20 experienced 1 of these events, then the independent Medical Safety Monitor and Data and Safety Monitoring Board could suggest termination of escalation to the next dosage level with the maximum tolerated dosage designated as the 1 dosage level below the current level.10
In addition, we analyzed the functional clinical outcome and embarked on comparisons between the safe dosage tiers (Tiers 1 and 2) and data from the International Intraoperative Hypothermia for Aneurysm Surgery Trial (IHAST).17 Historical data from the latter was chosen because the inclusion criteria were very similar to ALISAH allowing for more comparable populations than seen in other studies. Because ALISAH was an open-label study without concurrent controls, 2 analytic approaches to evaluate for potential treatment effects were adopted. First, statistical comparisons were made between subjects receiving the 2 dosage tiers deemed to be safe based on the safety analyses (Tier 1: 0.625 and Tier 2: 1.25 g/kg/day×7 days). Second, data from ALB Tiers 1 and 2 were compared with the entire IHAST cohort, IHAST normothermia, and IHAST hypothermia groups. The analyses conducted here were not prespecified but rather were conducted in a purely exploratory, post hoc fashion. Favorable outcome was defined as a score of 0 to 1 on the Glasgow Outcome Scale, reflecting good recovery at 3 months. The Glasgow Outcome Scale is the most commonly used outcome measure in SAH clinical trials. For these exploratory comparisons, relative risks were generated using the generalized linear model with log link.
A total of 383 subjects were screened and 47 (12.3%) enrolled in the ALISAH study. The most common reasons for exclusion were unavailable written informed consent (17%); no aneurysm found on cerebral angiography (14%); aneurysm treatment >72 hours from symptom onset (12%); stupor or coma (11%); age >80 years (7%); ALB administration before screening (7%); symptoms not related to SAH (6%); and unreliable time of onset (5%). The majority of subjects were female (72.3%), white (87%), and median age was 51 years (Table 1). Most subjects were current smokers and all were fully independent before symptom onset. Ruptured aneurysms were mostly located in the anterior circulation and were treated within 24 hours of symptom onset.
We enrolled 20 subjects in dosage Tier 1, 20 subjects in dosage Tier 2, and 7 subjects in dosage Tier 3. Two subjects died at the time of being in the study: 1 in Tier 1 and 1 in Tier 3. The former died of septic shock secondary to Gram-negative ventriculitis 2 weeks after symptom onset and was adjudicated as not related to ALB. The latter developed pulmonary edema possibly related to ALB and aspiration pneumonia with acute respiratory distress syndrome, which was the presumed cause of death.
Most subjects completed ALB treatment. In the first dosage Tier 3 subjects did not complete ALB infusion: 1 withdrew consent after developing DCI on Day 6; 1 skipped 1 day of treatment due to mild pulmonary edema on Day 3; and 1 developed mild pulmonary edema on Day 3 of treatment and medication was held. The latter was reported as a protocol violation because the clinical site did not consult with the primary coordinating center. In the second dosage Tier 4 subjects did not complete ALB infusion: 1 withdrew consent after 5 days of treatment; 1 skipped 1 day of treatment due to neurological deterioration on Day 2; 1 developed serious respiratory insufficiency possibly related to ALB after the first day of treatment; and 1 patient developed SAE with respiratory failure unlikely related to ALB on Day 3 of treatment.
Physiological and Laboratory Data
Daily serum ALB concentrations increased from baseline in all dosage tiers during the treatment period (Figure 1A). The changes were more prominent for Tier 2 and were sustained up to 15 days after enrollment (Supplemental Table III).
Daily maximum mean arterial blood pressures also increased from baseline for all treatment tiers during the treatment period (Figure 1B). These values tended to decrease to baseline after treatment was completed (Supplemental Table IV). Baseline blood pressures were higher for subjects in Tier 2 compared with the other tiers. This was not related to a prior known history of hypertension. We observed no differences in serum osmolality, serum creatinine, daily fluid balance, and daily weight from baseline in all the dosage tiers (Supplemental Tables V–VIII). In addition, there was no difference in central venous pressure values. However, due to concerns with higher values of fluid intake potentially leading to higher incidence of cardiac complications, the Data and Safety Monitoring Board recommended reducing the fluid balance value to <+1000 mL/day after 22 subjects had been enrolled.
There were 171 adverse events in 32 subjects. The only expected and observed adverse events related to ALB were those related to volume overload leading to acute heart failure or pulmonary edema. There were 17 SAEs reported in 7 subjects (15%) but only 3 adjudicated as related to ALB (Table 2). In the second dosage Tier 1 subject experienced pulmonary edema, which was possibly related to ALB. In the third dosage Tier 2 subjects experienced pulmonary edema with 1 possibly and 1 definitely related to ALB. Because of the 2 SAEs related to ALB in dosage Tier 3, the study was terminated after 47 patients had been enrolled and completed treatment.
The summary functional outcome scores for all the tiers and historical IHAST treatment groups are presented in Table 3. We also present the distribution of the Glasgow Outcome Scale, modified Rankin Scale, and Barthel Index scores at 3 months in Figure 2. There was a consistent dose–response relationship in that those subjects in Tier 2 did better overall than those in Tier 1 on all functional outcome measures, and those in Tier 3 did worse overall. In addition, the overall incidence of DCI secondary to symptomatic vasospasm was low (17%). However, the proportion was lower for subjects in Tier 2 (15%) compared with those in Tier 3 (20%). We also reviewed head CT scans at 90 days to investigate new cerebral infarctions that were not present at the baseline study. We found a total of 7 new cerebral infarctions (3 in dosage Tier 1, 3 in dosage Tier 2, and 1 in dosage Tier 3).
Subjects in ALB dosage Tiers 1 and 2 were compared with subjects in the IHAST study (total N=1000; hypothermia group=499; normothermia group=501).17 Subjects in Tier 2 had better outcomes compared with those of Tier 1 suggesting a dose–response (Table 4). In addition, when compared with IHAST subjects, outcomes for Tier 2 subjects were better.
We have shown that large doses of ALB up to 1.25 g/kg/day×7 days are safe in patients with aneurysmal SAH. The safety stopping rule of at least 2 events of severe-to-life-threatening heart failure in dosage Tiers 1 and 2 was not met. Dosages higher than 1.25 g/kg/day were associated with significant cardiovascular complications including 2 SAEs related to ALB. The latter resulted in early termination of the study. It is important to note that in all instances, pulmonary edema was easily treated and resolved. We have also demonstrated that our treatment protocol was feasible and successfully implemented in several international centers. Our data are supported by findings from ALB studies in patients with acute ischemic stroke.8,9
The main physiological effects of ALB treatment were elevation in the serum ALB concentration and mean arterial blood pressure. The latter improved to baseline values after treatment completion. Serum ALB values remained elevated 7 days after treatment suggesting that any potential beneficial effects of ALB may remain throughout the critical period of DCI risk. We observed no changes in serum osmolality or renal function. In addition, we found that fluid intake increased in direct relationship to higher dosage tiers. However, there was no correlation between higher fluid intake or fluid balances and cardiovascular complications. This suggests that factors other than pure intravascular volume augmentation may play a role. We speculate that diastolic cardiac dysfunction may be a contributing factor rather than systolic abnormalities due to the fact that left ventricular ejection fractions were within normal limits in all our subjects. Note, however, that the sample size was small and therefore our findings will have to be validated in a larger cohort of SAH subjects.
The data also suggest that high dosage of ALB up to 1.25 g/kg/day×7 days is not only safe but also may be neuroprotective. This is supported by data from ischemic stroke subjects treated with ALB8,9,18 and retrospective data from patients with SAH.3 ALB has several potential mechanisms that could explain its neuroprotective effects. The increase in serum albumin concentrations is expected to increase the serum oncotic pressure, which in turn will mobilize interstitial fluid and improve organ function, including the brain.19 ALB also possesses antioxidant and scavenger properties in part by its potential to replete thiol stores,20–22 and it can modulate apoptosis.23 In addition, ALB administration may also improve microcirculatory blood flow, increase organ blood flow, decrease leukocyte rolling and adherence, and reduce the inflammatory response.24
The potential reduction DCI secondary to symptomatic vasospasm may be explained by the interaction of ALB with the ALB-specific binding sites of the microvasculature endothelium.25–27 By binding to the endothelial glycocalyx, ALB maintains the normal permeability of microvessel walls.26,28,29 It has also been suggested that ALB may be a factor in mediating the effect of blood coagulation on vascular tone and capillary permeability.30 Moreover, serum ALB reacts with nitric oxide to form a stable S-nitrosothiol that has endothelium-derived relaxing factor-like properties.31
Our study has several limitations. ALISAH is an early-phase design and we do not have concurrent controls. In addition, the study was neither randomized nor powered to test for efficacy effects. Therefore, caution is advised in the interpretation of our data and comparison with the IHAST study. Moreover, we did not obtain head CT scans after aneurysmal treatment. This limits our ability to determine whether the cerebral infarctions found at 90 days were due to treatment modalities or ischemic complications from SAH. Lastly, the severity of radiological infarctions cannot be ascertained because we did not measure infarction volume.
The ALISAH Pilot Study has demonstrated that large ALB dosages up to 1.25 g/kg/day×7 days are safe and the treatment protocol is feasible. These dosages have been found to be neuroprotective in animal models of cerebral ischemia. Pulmonary edema, the main systemic complication related to ALB, was easily managed in the intensive care unit setting. Despite the limited sample size, the ALISAH data also provide preliminary evidence that high-dose ALB may be neuroprotective after aneurysmal SAH. Based on these encouraging results, initial planning of ALISAH II, a large Phase III multicenter, randomized, placebo-controlled clinical trial to evaluate the efficacy of ALB in subjects with SAH, is underway.
Sources of Funding
This study was supported by National Institutes of Health Pilot Clinical Trial Grant NS049135 (Principal Investigator, J. I. Suarez). ALISAH is operated under Investigational New Drug approval from the Food and Drug Administration: BB-IND #13022.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Neurological Disorders and Stroke of the National Institutes of Health. Study intervention was provided at no cost by Grifols International (Barcelona, Spain).
↵* The complete list of investigators is provided in the Appendix.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.633958/-/DC1.
- Received August 26, 2011.
- Revision received October 17, 2011.
- Accepted November 8, 2011.
- © 2012 American Heart Association, Inc.
- Bederson JB,
- Connolly ES,
- Batjer HH,
- Dacey RG,
- Dion JE,
- Diringer MN,
- et al
- Rinkel GJ,
- Feigin VL,
- Algra A,
- Vermeulen M,
- van Gijn J
- Belayev L,
- Alonso OF,
- Busto R,
- Zhao W,
- Ginsberg MD
- Belayev L,
- Zhao W,
- Pattany PM,
- Weaver RG,
- Huh PW,
- Lin B
- Belayev L,
- Liu Y,
- Zhao W,
- Busto R,
- Ginsberg MD
- Ginsberg MD,
- Hill MD,
- Palesch YY,
- Ryckborst KJ,
- Tamariz D
- Palesch YY,
- Hill MD,
- Ryckborst KJ,
- Tamariz D,
- Ginsberg MD
- Peters T
- Peters T
- Lennihan L,
- Mayer SA,
- Fink ME,
- Beckford A,
- Paik M,
- Zhang H,
- et al
- Hill MD,
- Martin RH,
- Palesch YY,
- Tamariz D,
- Waldman BD,
- Ryckborst KJ,
- et al
- Zoellner H,
- Hofler M,
- Beckmann R,
- Hufnagl P,
- Vanyek E,
- Bielek E,
- et al
- Ghitescu L,
- Fixman A,
- Simionercu N
- He P,
- Curry FE
- Schnitzer JE,
- Sung A,
- Hovart R,
- Bravo J