Plasma exchange (PE) and immunoadsorption (IA) are frequently used for treatment of various autoimmune-mediated neurological diseases, including multiple sclerosis (MS), chronic inflammatory demyelinating polyneuropathy (CIDP), and Guillain–Barré syndrome (GBS). Although both methods are generally regarded as well-tolerated treatment options, evidence for safety and tolerability is low for most indications and largely relies on small case series. In this study, we retrospectively analysed adverse events (AEs) and laboratory changes in 284 patients with various neurological indications who received either PE (n = 65, 113 cycles) or IA (n = 219, 435 cycles) between 2013 and 2020 in our Neurology department. One standard treatment cycle for PE as well as IA consisted of five treatments on five consecutive days. During every treatment, the 2.0–2.5-fold individual plasma volume (PV) was treated in IA, while in PE, the 0.7-fold individual PV was replaced by human albumin solution. Overall, both methods showed an excellent safety profile; no deaths of life-threatening adverse events were recorded. Severe AEs (corresponding to grade 3 on the Common Terminology Criteria for Adverse Events grading scale v5.0) including three patients with sepsis, one pneumonia, and one pneumothorax were present in 5/435 IA cycles (1.1%); in the PE group, no severe AEs were recorded. Furthermore, although advantageous tolerability is generally considered the main advantage of IA over PE, we found that overall frequency of AEs (including grades 1 and 2) was higher in IA (67.1% of all cycles) compared to PE (35.4%; p < 0.001). The low incidence of AEs in PE might be caused by the lower PV exchanged during each treatment (0.7-fold) compared to previous studies which predominantly exchanged the 1.0–1.5-fold PV. In order to verify this hypothesis as well as confirming the efficacy of this lower-dosed scheme, prospective studies comparing different treatment regimens are needed.