Plasma exchange with albumin replacement for Alzheimer's disease treatment induced changes in serum and cerebrospinal fluid inflammatory mediator levels

Abstract

Objective: There is extensive literature indicating that inflammatory pathways are affected in Alzheimer's disease (AD). We examined whether plasma exchange with albumin replacement (PE-Alb) can impact the inflammatory status of AD patients and alter the relationship between inflammatory mediators and cognitive measures.

Methods: Serum and cerebrospinal fluid (CSF) samples from 142 AD patients participating in the AMBAR trial (14-month schedule of PE-Alb treatment vs. placebo [sham PE-Alb]) were analyzed for changes from baseline for 19 inflammatory mediators (6 inflammatory cytokines, 9 chemokines, and 4 vascular injury indicators) at representative time points across the AMBAR study (lasting effects) as well as in pre- versus post-PE-Alb procedure (acute effects). Association between mediator changes and clinical outcomes reported in the AMBAR study (cognitive, functional, behavioral function, and global change tests) was assessed.

Results: PE-Alb significantly reduced IFN-γ, eotaxin, MIP-1α and ICAM-1 levels in serum, and eotaxin-3 and MIP-1β levels in CSF, at various time points during treatment (p < 0.05; false discovery rate-corrected). Vascular injury indicators were the mediators mostly affected by post- versus pre-PE-Alb level reduction. Increased serum MIP-1α levels were associated with worsening in ADAS-Cog, CDR-sb, and ADCS-CGIC scores in the placebo group, but not in the PE-Alb-treated group.

Interpretation: Peripheral intervention could affect AD by reducing inflammatory mediators in both peripheral and central compartments. Changes in MIP-1α due to PE-Alb were associated with changes in clinical outcomes.

Figure 1

Outline of treatment periods, plasma exchange sessions, visits, and distribution of treatment groups in the AMBAR trial. Visits with serum (shaded arrows) and cerebrospinal fluid (blank arrows) samples available for the present study are indicated. IVIG, intravenous immunoglobulin; LVPE, low‐volume plasma exchange; TPE, therapeutic plasma exchange.

Figure 2

Change from baseline in clinical outcome tests score (ADAS‐Cog, ADCS‐ADL, CDR‐sb, and ADCS‐CGIC) in the patient cohort of this study. FV, final visit (end of LVPE period); IV, intermediate visit (end of therapeutic plasma exchange period); LVPE 7, low‐volume plasma exchange 7; PE‐Alb, plasma exchange with albumin replacement.

Figure 3

Heatmap representation of least squares mean ratio between the group of patients treated with plasma exchange with albumin replacement (PE‐Alb; n = 54–102) and placebo group (n = 16–36). Gradient in red means decreasing ratio <1 (favors placebo) whereas gradient in blue means increasing ratio >1 (favors PE‐Alb treatment). See Figure 1 for details of treatment periods, visits, and treatment patient groups in the AMBAR trial. CSF, cerebrospinal fluid; IP, intensive period (therapeutic plasma exchange); MP, maintenance period (low‐volume plasma exchange).

Figure 4

Plots of the changes in inflammatory mediator levels (effect size in PE‐Alb‐treated with respect to placebo) before and after PE‐Alb (acute effects). (A) TPE 1 (therapeutic plasma exchange 1; Month 0.2); (B) LVPE 7 (low‐volume plasma exchange 7; Month 9).

Figure 5

Predicted trajectories of clinical outcome (change from baseline [CFB] at the end of the study of clinical outcome tests score (ADAS‐Cog. CDR‐sb, and ADCS‐CGIG) according to levels of MIP‐1α in placebo and plasma exchange with albumin replacement (PE‐Alb)‐treated patients. Shaded areas represent the 95% confidence interval.