Safety, Tolerability, and Feasibility of Young Plasma Infusion in the Plasma for Alzheimer Symptom Amelioration Study

Key Points

Question

Is young plasma safe, feasible, and tolerable in patients with mild to moderate Alzheimer disease dementia?

Findings

This randomized clinical trial consisted of a double-blind crossover group of 9 patients and an open-label group of 9 patients. Patients with mild to moderate Alzheimer disease dementia were able to tolerate 4 weekly infusions of young plasma.

Meaning

Safe and well tolerated in a small sample of patients with mild to moderate Alzheimer disease dementia, young plasma infusions warrant further analysis in a larger study using a double-blinded design with a placebo control.

Abstract

Importance

Young mouse plasma restores memory in aged mice, but, to our knowledge, the effects are unknown in patients with Alzheimer disease (AD).

Objective

To assess the safety, tolerability, and feasibility of infusions of young fresh frozen plasma (yFFP) from donors age 18 to 30 years in patients with AD.

Design, Setting, and Participants

The Plasma for Alzheimer Symptom Amelioration (PLASMA) study randomized 9 patients under a double-blind crossover protocol to receive 4 once-weekly infusions of either 1 unit (approximately 250 mL) of yFFP from male donors or 250 mL of saline, followed by a 6-week washout and crossover to 4 once-weekly infusions of an alternate treatment. Patients and informants were masked to treatment and subjective measurements. After an open-label amendment, 9 patients received 4 weekly yFFP infusions only and their subjective measurements were unmasked. Patients were enrolled solely at Stanford University, a tertiary academic medical center, from September 2014 to December 2016, when enrollment reached its target. Eighteen consecutive patients with probable mild to moderate AD¹ dementia, a Mini-Mental State Examination (score of 12 to 24 inclusive), and an age of 50 to 90 years were enrolled. Thirty-one patients were screened and 13 were excluded: 11 failed the inclusion criteria and 2 declined to participate.

Interventions

One unit of yFFP from male donors/placebo infused once weekly for 4 weeks.

Main Outcome and Measures

The primary outcomes were the safety, tolerability, and feasibility of 4 weekly yFFP infusions. Safety end point analyses included all patients who received the study drug/placebo.

Results

There was no difference in the age (mean [SD], 74.17 [7.96] years), sex (12 women [67%]), or baseline Mini-Mental State Examination score (mean [SD], 19.39 [3.24]) between the crossover (n = 9) and open-label groups (n = 9). There were no related serious adverse events. One patient discontinued participation because of urticaria and another because of an unrelated stroke. There was no statistically significant difference between the plasma (17 [94.4%]) and placebo (9 [100.0%]) cohorts for other adverse events, which were mild to moderate in severity. The most common adverse events in the plasma group included hypertension (3 [16.7%]), dizziness (2 [11.1%]), sinus bradycardia (3 [16.7%]), headache (3 [16.7%]), and sinus tachycardia (3 [16.7%]). The mean visit adherence (n = 18) was 86% (interquartile range, 87%-100%) and adherence, accounting for a reduction in the total visit requirement due to early patient discontinuation, was 96% (interquartile range, 89%-100%).

Conclusions and Relevance

The yFFP treatment was safe, well tolerated, and feasible. The study’s limitations were the small sample size, short duration, and change in study design. The results warrant further exploration in larger, double-blinded placebo-controlled clinical trials.

Trial Registration

ClinicalTrials.gov Identifier: NCT02256306

This randomized clinical trial examines the safety, tolerability, and feasibility of young plasma infusion in patients with mild to moderate Alzheimer disease dementia in the Plasma for Alzheimer Symptom Amelioration study.

Introduction

Alzheimer disease (AD) is a progressive neurodegenerative disease that affects cognition and function. There is no known disease-modifying therapy, making drug discovery an area of unmet medical need. The increasing number of patients with AD warrants a pursuit of treatments that can improve or maintain cognitive function.

Plasma from young mice has been demonstrated to restore memory and stimulate synaptic plasticity in the hippocampi of aged mice in parabiotic models and via direct injection.^1,2 In one publication by our group,¹ aged mice given young mouse plasma demonstrated enhanced learning and memory on a radial arm water maze test and increased freezing in both contextual memory and fear-conditioning memory testing results compared with aged mice that were given aged plasma. Another study by our group² demonstrated that aged immunodeficient mice that were treated with human cord plasma had decreased escape latency results on the Barnes maze test and increased freezing levels during a hippocampal-dependent contextual memory test compared with vehicle-treated mice of the same age. Independent studies have reported that treatment with young plasma increased cerebral blood flow and olfactory memory,³ increased hippocampal-dependent memory as assessed by a novel object recognition test, and reduced anxiety⁴ in aged mice. Young plasma also improved cognitive function as assessed by spatial working memory and contextual fear condition tests and reduced inflammatory processes in a transgenic mouse model for AD.⁵ These promising preclinical studies suggested that circulating factors in young plasma may have beneficial effects on human cognition. Plasma infusions are known to be safe and are typically administered to promote coagulation in people with bleeding disorders due to factor deficiencies. Whether the benefits seen in mouse studies apply to humans is, to our knowledge, unknown, as the cognitive effects of plasma have not yet been studied in aged humans or in patients with AD.

The primary objective of this study was to assess the safety, tolerability, and feasibility of infusions of young fresh frozen plasma (yFFP) from young donors in patients with mild to moderate AD dementia. The exploratory objectives were cognition, functional ability, mood, and functional connectivity in the brain’s default mode network. To our knowledge, there are no publications describing the weekly administration of young plasma in this patient population.

Methods

Study Oversight

This study was approved by the Stanford University institutional review board and registered at ClinicalTrials.gov (NCT02256306). Written informed consent was obtained from patients and caregivers.

Study Design and Enrollment

The target enrollment was 18 patients with mild to moderate AD dementia. Patients were enrolled at Stanford University from September 2014 until December 2016 when the enrollment reached its target. The inclusion criteria were an age of 50 to 90 years, probable Alzheimer disease (National Institute on Aging–Alzheimer Association criteria⁶), a Mini-Mental State Examination score of 12 to 24 (inclusive), and the availability of a study partner who knew the patient well to attend all visits with patients. See eAppendix 1 in Supplement 1 for the exclusion criteria. Nine patients were initially enrolled and randomized to receive treatment under a double-blind crossover protocol with 4 once-weekly infusions of either 1 unit (approximately 250 mL) of plasma from male donors aged 18 to 30 years (yFFP) or 250 mL of saline, followed by a 6-week washout and 4 once-weekly infusions of the alternate treatment, followed by another 6-week washout period (eTable 1 in Supplement 1; Figure 1 and Figure 2). Randomization was performed by a coordinator who was uninvolved in the study. Raters, patients, and caregivers were masked to the intervention until study completion and database lock. The randomization coordinator selected treatment assignments from a randomization envelope and provided assignments to the study coordinator, who requested the plasma from the Stanford Blood Bank or saline on the day of dosing. To ensure that the patient and caregiver remained masked, intravenous tubing and bags were covered. In October 2015, after 9 of the 18 patients were enrolled, the protocol was amended and the remaining 9 patients were enrolled under an open-label design in which patients received 4 once-weekly infusions of only yFFP (Figure 2; eTable 2 in Supplement 1). These patients and caregivers knew they received treatment with yFFP and no randomization was necessary because no placebo was provided after the amendment occurred. This protocol change condensed the total time of participation from 5 months to less than 3 months, which was intended to improve enrollment efficiency, reduce the burden on patients and caregivers, and decrease costs. Male donor plasma was used to minimize the known increased risk of transfusion-related acute lung injury that is associated with multiparous female donor blood.

Figure 1. Consolidated Standards of Reporting Trials Diagram of Patient Enrollment.

^aOne patient received placebo only and withdrew prior to receiving plasma because of an unrelated stroke.

Figure 2. Plasma for Alzheimer Symptom Amelioration Study Experimental Paradigm.

^aBaseline 1, n = 4; post 1, n = 4; baseline 2, n = 4; post 2, n = 4.

^bBaseline 1, n = 5; post 1, n = 5; baseline 2, n = 3; post 2, n = 2.

^cBaseline 1, n = 9; post 1, n = 7.

Outcomes

The primary outcomes were safety, tolerability, and feasibility of once-weekly plasma infusions in patients with mild to moderate AD dementia. Safety measures were summarized by the number of adverse events (listed by the Common Criteria for Adverse Events, version 4.0), compliance rate, and routine clinical and laboratory assessments. The primary end points for safety included all patients who received any study drug (including the control) and excluded patients who withdrew consent before receiving any treatment. Feasibility was defined as adherence to visits (the number of visits completed divided by the total scheduled visits). The exploratory outcomes were a change in the baseline for the following measures: Alzheimer Disease Assessment Scale-Cog (ADAS-Cog) 13-item version, Trail Making Test (TMT) parts A and B, Geriatric Depression Scale, Neuropsychiatric Inventory, Clinical Dementia Rating Scale Sum of Boxes, Functional Activities Questionnaire, and the Alzheimer Disease Cooperative Study Activities of Daily Living Inventory in mild cognitive impairment. These tests were administered before each block of infusions (to serve as a baseline before each treatment period), after the fourth infusion, after the eighth infusion (when applicable), and after the final 6-week washout period, which was not used for analysis. The exploratory outcome analyses were performed on all randomized or enrolled patients who received the study drug or placebo and who participated in at least 1 postbaseline assessment (modified intention to treat). Additional imaging exploratory outcomes were the effects of yFFP on functional connectivity and are described later in this article.

Statistical Analyses

Statistical analyses were conducted using R, version 3.2.2 (R Foundation) and SPSS, version 22 (IBM). The Fisher exact test was used to compare the number of related adverse events by treatment group. The treatment effect on each cognitive, mood, and functional measure was evaluated by fitting a mixed-effects linear regression model with the outcome at the posttreatment visit as the response, the treatment indicator as the independent variable of the primary interest, and a subject-specific random intercept to account for within-person correlations between 2 outcomes for patients. Specifically, the regression model is

Y_ij = α + βTr_ij + γ₁j +γ₂Y_i0 + α_i + ε_ij

in which Y_ij is the outcome of the ith patient at the jth session, j = 1,2, Trt_ij = 0/1 is the corresponding treatment indicator, Y_i0 is the baseline outcome, and α_i is the subject-specific random intercept. Therefore, the regression analysis also adjusted for the session of the treatment and baseline outcome. The regression coefficient β represents the treatment effect, summarizing both the within-person comparisons among patients who crossed over from placebo to plasma infusion or vice versa and between-group comparisons between patients receiving placebo and patients receiving plasma infusions. The mixed-effect regression model adjusted for data missing at random. The treatment effect on each measure was summarized by point estimator, 95% confidence intervals, and P values. We did not correct for multiple comparisons because of the small sample size and exploratory nature of analyses.

Apolipoprotein E Genotyping

Apolipoprotein E (APOE) genotyping was performed by Progenika (San Marcos, Texas). DNA was isolated from patients’ whole blood cell counts and DNA amplification was performed on a region of APOE exon 4 that contained the polymorphisms determining APOE ε2, ε3, and ε4 haplotypes. Sanger sequencing was performed to ascertain patients’ haplotypes. The effect of ApoE4 status was examined by conducting a subgroup analysis and comparing the average change between subgroups of patients with different ApoE4 statuses.

Magnetic Resonance Imaging Acquisition and Preprocessing

Magnetic resonance imaging (MRI) scanning was performed at screening and after each block of infusions using a 3-T scanner (General Electric). Patients completed one T1-weighted anatomical scan (3-dimensional fast spoiled gradient echo; flip angle = 15°; matrix size = 256 × 256; 158 axial slices; in-plane resolution = 0.85 mm; slice thickness = 1 mm) and 2 resting-state functional MRI (fMRI) scans (echo-planar imaging [General Electric]; echo time = 30 milliseconds; flip angle = 80°; in-plane resolution = 3.4 mm; slice thickness = 4.0 mm; gap = 0.5 mm; 31 axial slices; repetition time = 2 seconds; 240 volumes). All resting-state fMRI scans were preprocessed using the FSL software package⁷ and integrated in-house processing pipelines. Preprocessing included correction for head motion through volume realignment, correction for magnetic field inhomogeneities (B0 correction), spatial smoothing (6 mm full width at half maximum), advanced head motion de-noising using ICA-based Automatic Removal of Motion Artifacts,⁸ and a 0.01-Hz high-pass filter. All further analyses were performed in Montreal Neurological Institute-152 standard space to allow between-patient comparisons (eAppendix 1 in Supplement 1). Where available, we combined results for the 2 within-session resting-state functional scans by means of within-patient fixed-effects analyses.

Resting-State fMRI Analyses

Resting-state fMRI analyses allows for the investigation of characteristics of spontaneous brain activity, which has been found to cluster into meaningful large-scale networks.⁹ We assessed the network integrity and identifiability of 10 well-described and replicated resting-state networks.⁹ To assess network integrity, we conducted group-level analyses for 3 differential within-patient effects: (1) baseline + plasma (mean effect), (2) baseline > plasma, and (3) plasma > baseline. Group-level statistics were obtained using nonparametric testing (FSL randomize). Results were considered significant at P < .005 (ie, P < .05 per 10 investigated networks). To assess network identifiability, we extracted the mean z statistic within the 10 resting-state networks and compared those with the mean z statistic outside each respective network mask. A higher inside-outside ratio indicates that the network is easier to identify against the background noise. We assessed the effect of sessions on these scores using unpaired t tests across all networks, including all available patients. Effects were considered significant at P < .05. Finally, we obtained seed-based functional connectivity maps for left and right hippocampus regions of interest. These connectivity maps were entered into group-level comparisons, as described previously. In addition, for each patient and each session we extracted the connectivity strength between each of the hippocampus regions of interest and each of the 10 large-scale networks. We used these mean z values for correlation analyses against behavioral metrics that were obtained for each patient (eAppendix 1 in Supplement 1).

Results

Demographics

There were no statistically significant differences in age, sex, baseline Mini-Mental State Examination score, or ApoE4 genotype between the plasma crossover and plasma-only groups (Table 1).

Table 1. Characteristics of Study Patients.

Characteristic	Crossover (n = 9)	Plasma Only (n = 9)	P Value	All Patients (n = 18)
Age, mean (SD), y	72.33 (10.61)	76.00 (3.84)	.34	74.17 (7.96)
Female, No. (%)	6 (67)	6 (67)	>.99	12 (67)
MMSE score, mean (SD)	19.89 (3.72)	18.89 (2.80)	.53	19.39 (3.24)
ApoE4 carrier, No./total No. (%)	7/8 (87.5%)	7/9 (77.8%)	.46	14/17 (82.4%)
Homozygous, No. (%)	3 (33.3%)	2 (22.2%)	.61	5 (27.8%)

Abbreviations: ApoE, apolipoprotein E; MMSE, Mini-Mental State Examination.

Safety, Tolerability, and Feasibility

There were 71 adverse events reported and 21 (30%) were likely related. One patient in the open-label cohort discontinued treatment owing to urticaria during plasma infusion (Figure 1). Another patient in the crossover cohort discontinued participation because of an unrelated stroke that occurred during the end of the washout period following the block of placebo infusions. This patient did not receive plasma treatment because discontinuation occurred before the block of plasma infusions. There were no related serious adverse events. All other adverse events were mild or moderate in severity and the most common in the plasma treatment group included hypertension (5 events in 3 patients), dizziness (2 [12%]), sinus bradycardia (3 [18%]), headache (3 [18%]), and sinus tachycardia (3 [18%]) (Table 2). There was no statistically significant difference between the plasma (n = 17) and placebo (n = 9) cohorts for these adverse events. For all 18 patients, mean visit adherence was 86% (interquartile range, 87%-100%) and adherence, accounting for the reduction in total visit requirements due to patient discontinuation early from the study, was 96% (interquartile range, 89%-100%), suggesting good overall feasibility.

Table 2. Related Adverse Events^a.

Related Adverse Event	No. (%)		P Value
	Plasma (n = 17)	Placebo (n = 9)
Dizziness	2 (12)	0	.53
Fatigue	1 (6)	0	>.99
Headache	3 (18)	0	.53
Hypertensionb	3 (18)	2 (22)	>.99
Presyncope	1 (6)	0	>.99
Respiratory-other (tachypnea)	1 (6)	1 (11)	>.99
Sinus
Bradycardia	3 (18)	1 (11)	>.99
Tachycardia	3 (18)	1 (11)	>.99
Somnolence	1 (6)	0	>.99
Urticaria	1 (6)	0	>.99
Vertigo	1 (6)	0	>.99

^aMild and moderate related adverse events. Each event occurred once per patient except for hypertension.

^bHypertension events: 5 events in 3 patients in the plasma group; 3 events in 2 patients in placebo group. Percentages listed as number of patients with event/number of patients in the treatment group. A statistical analysis using the Fisher exact test was used to compare the number of related adverse events by treatment group.

Exploratory End Points

Baseline assessments were conducted for all 18 patients in the originally assigned groups. Of the 9 patients in the open-label plasma cohort, 7 (77.8%) had post-yFFP treatment assessments, and all 9 patients in the crossover cohort had posttreatment assessment 1 (Figure 2). Seven patients completed baseline assessment 2, of whom 6 (85.7%) completed post-treatment assessment 2 (Figure 2). There were no changes on the total score of the ADAS-Cog 13 (−1.04 points; 95% CI, −3.49 to 1.41; P = .27) and TMT part A (−4.64 points; 95% CI, −12.28 to 3.01; P = .18) (eAppendix 2 in Supplement 1). Less than half the patients validly performed the TMT part B, which was excluded from the analysis. On functional measures, there was improvement on the Functional Activities Questionnaire (−4.56 points; 95% CI, −6.11 to −3.01; P = .001) and the Alzheimer’s Disease Cooperative Study Activities of Daily Living Inventory in mild cognitive impairment (3.17 points; 95% CI, 0.70-5.65; P = .03) after plasma treatment (eFigure 1 in Supplement 1), but no statistical adjustments were made for multiple testing given the small sample size (eFigure 2 and eTable 3 in Supplement 1). There were no changes on the Clinical Dementia Rating Scale Sum of Boxes (0.26 points; 95% CI, −0.12 to 0.64; P = .13) nor on measures of psychiatric symptoms and mood (Neuropsychiatric Inventory, −73.57 points; 95% CI, −11.57 to 4.43; P = .33; or Geriatric Depression Scale, 1.45 points; 95% CI, −0.64 to 3.54; P = .14). Because of the large number of ApoE 4 carriers, a similar prevalence in both cohort groups, and the small sample size, it was not possible to assess whether the exploratory end points were differentially affected by ApoE 4 status.

Baseline imaging data were available for 15 patients (83.3%), of whom 9 (60%) also had good-quality imaging data available after plasma administration. There were no effects of yFFP administration on the overall structure of the 10 resting-state networks and both hippocampal seed-based connectivity maps for whole-brain functional connectivity comparisons. Across networks, we observed lower z scores inside the networks after plasma administration compared with the baseline (t = −2.3; P < .02; eFigure 3 in Supplement 1). For the z statistics outside the network (t = −1.68; P < .09) and the inside/outside ratio (t = 1.64; P < .10), the effects did not reach statistical significance. The change in composite memory performance on the ADAS-Cog 13 was associated with a change in the default mode network seed-based correlations from the right hippocampus (r = −0.756; n = 12 [80%]; P = .004; eFigure 4 in Supplement 1).

Discussion

Studies using young plasma have demonstrated improved memory and synaptic plasticity in aged mice, suggesting a possible therapeutic role of young plasma for treating memory impairment in patients with AD. In what, to our knowledge, is the first study to assess the safety and feasibility of yFFP in people with mild to moderate AD dementia, we found that yFFP was safe, well tolerated, and feasible. There were no related serious adverse events and all other adverse events were mild or moderate in severity. There were no statistically significant differences in adverse events between the plasma and placebo cohorts.

The efficacy of yFFP in patients with mild to moderate AD could not be determined because of the small sample size, a change in the design of the study, and the short duration of treatment. Therefore, assessments of cognition, mood, functional ability, and default mode network changes were exploratory. Analyses of these measures did not find that infusions of yFFP altered mood, global cognition, or functional connectivity. However, improvements in functional abilities were reported by caregivers. These findings could be further explored with a larger study that is powered to determine clinical and statistical significance.

Limitations

There were several limitations to this study. The change in design from a double-blind crossover study to an open-label study facilitated a more efficient enrollment and reduced the cost of the study, but statistical analyses of outcome measures needed to be modified, efficacy could not be determined, and the analysis for the safety of drug-placebo differences became less robust. Additionally, the small sample size and short study duration, as mentioned previously, limited the analyses for efficacy. The high prevalence of ApoE4 carriers in this study prohibited the generalizability to other individuals with mild to moderate AD dementia without further investigation in a larger trial. The improvements observed on informant rating scales were subjective. The crossover nature of the original design of the study did not account for potentially sustained effects of yFFP and may not be best suited for neurodegenerative disease because of a potentially changing baseline. Further studies should include an analysis for dosing, pharmacokinetics, and pharmacodynamics.

Conclusions

The results from this study using yFFP in patients with AD demonstrate that treatment with yFFP is safe and warrant further exploration in larger double-blinded clinical trials that use measures that are designed to detect change within the treatment time frame and are powered to determine efficacy.

Supplement 1.

eAppendix 1. Methods: Exclusion criteria, MRI processing, Resting state fMRI analyses, Dual regression analyses, Group-level network analyses, Network identifiability analyses, Seed-based functional connectivity analyses.

eAppendix 2. Results: Cognitive sub-analyses, Functional sub-analyses, Correlation of cognitive and MRI data.

eTable 1. Schedule of Events (Double-blind Crossover Design).

eTable 2. Schedule of Events (Open-label Amendment).

eFigure 1. Cognitive, Functional, and Mood Results.

eFigure 2. Individual patient plots of functional measures.

eTable 3. FAQ and ADCS-MCI-ADL Change Scores by Group.

eFigure 3. Across network Z-score analyses.

eFigure 4. Correlation between change in ADAS-Cog 13 memory composite scores and seed-based connectivity measures.

eReferences.

Supplement 2.

Protocol amended.

Supplement 3.

Protocol original.

Supplement 4.

Statistical analysis plan.

Supplement 5.

Data sharing statement.