Plasma exchange (PLEX) superficially resembles the historical practice of bloodletting for any disease. The removal of blood, separation from its formed elements, and their reinfusion with plasma replacement started as a treatment for hyperviscosity as early as 1914 and was accepted as standard therapy in 1960. (PLEX was first employed in 1952 in multiple myeloma to control hyperviscosity.)

PLEX machines are now standard equipment in the neurological intensive care unit, where they are used to treat severe Guillain-Barré syndrome (GBS), myasthenic crisis (PLEX remains a first-line therapy in one of the best understood antibody-mediated diseases with a close relationship between antibody levels and severity of symptoms), and a number of autoimmune encephalitides and demyelinating disorders. It has also recently been introduced in immune checkpoint inhibitor–related adverse events. PLEX can remove pathogenic antibodies (many of which have yet to be identified with these adverse effects), cytokines, and chemokines [1, 2].

Despite its prominence, PLEX has never been compared with sham treatment in acute neurologic disease, and therefore its acceptance as a definitive treatment for diseases could be debated. Where did its reputation come from? Its proponents quickly won the day, but only because there was little else to offer.

PLEX is based on the following assumption: if a plasma factor causes a disease, then removal of the factor should treat the disease. PLEX was originally described in 1914 by Vadim A. Yurevitch and Nicolay Rosenberg of the Imperial Medical and Surgical Academy of Saint Petersburg [3] and in the US by John Abel (Fig. 1) and colleagues from Johns Hopkins Hospital in 1914 [4]. They described the technique used in animal experiments. Paul Rubinstein was the first to use plasmapheresis to treat an immune-related disorder when he saved the life of an adolescent boy with thrombotic thrombocytopenic purpura at the Cedars of Lebanon Hospital in Los Angeles in 1959.

John Abel (public domain)

By its very nature, PLEX is a nonspecific therapy, and it depends on the delay with which the offending factor can regenerate. For example, antibodies, with their long half-lives (~ 21 days), may remain depleted for prolonged periods after a short course of PLEX. By this rationale, any autoimmune disease with demonstrable autoantibodies is a legitimate target for clinical trials. (Obviously, the practice of plasmapheresis differs from classical bloodletting because the two therapies spring from entirely different theoretical bases).

Therefore, it was not long before PLEX was considered in the treatment of GBS. Patients frequently received corticosteroids, although the effect was highly questionable and without confirmed benefit. Brettle and colleagues from Hammersmith Hospital offered one of the first studies to suggest the efficacy of PLEX when they described a solitary case of a patient with GBS who regained dramatic, sustained improvement of proximal limb strength after a second PLEX treatment. After four exchanges, the patient could sit and stand unaided [5]. They opened their article with the following:

Dr Hughes and his colleagues (Oct. 7, p. 750) have shown, in a controlled, randomised, multicentre trial, that prednisolone is of no benefit in acute polyneuropathy of undetermined aetiology (Guillain-Barré syndrome). This finding, unfortunately, seems to leave us with no effective therapy for our patients, other than “careful nursing, physiotherapy, and attention to o the risk of ventilatory failure and circulatory complications” [5].

Addison et al. from Massachusetts General Hospital also described improvement by PLEX in four patients with GBS [6]. These initially encouraging responses in a few cases prompted the development of a prospective study comparing PLEX with “conventional therapy.”

Guy McKhann from Johns Hopkins University School of Medicine organized a study in 245 patients. The trial was notable for specific inclusion criteria, strict definition of the disease, and outcome assessment using a previously tested grading scale [7]. Patients were randomized for 4 years into a PLEX or a conventional therapy group and analyzed for time spent on the respirator and outcome. McKhann’s group recruited the participation of 21 medical centers; Massachusetts General Hospital provided one of the largest numbers of patients. The study concluded that PLEX had an important effect on recovery. Medical benefit was most apparent when reviewing the time spent on the ventilator. The median time for patients to reach grade 2 or be able to walk unassisted declined from 85 to 53 days (Fig. 2) [7], and similar benefit was found in patients on the ventilator. Outcome at 6 months was similar for all patients. The observers knew which patients were treated, which could have influenced time on the ventilator and earlier weaning. Of note, early, rapid recovery after PLEX was not observed in this study; the differences became apparent after weeks of observation. Few investigators expected to find a positive response to PLEX during the course of GBS when asked before the results were released (AH Ropper, personal communication).

Results of US plasma exchange study [7], used with permission

IVIG became a consideration after a patient with chronic inflammatory demyelinating polyneuropathy (CIDP) fortuitously was treated with plasma infusion rather than PLEX and, to the surprise of all, improved [8]. This led to consideration of IVIG as the possible active component. (Immunoglobulin [pooled quantities of human immunoglobulin G] was rarely used in neurologic disorders and mostly in autoimmune thrombocytopenia.) The reasoning was simple: if IVIG worked in other autoimmune disorders, it could work in autoimmune CIDP. Patients with a successful response to plasma infusion had comparable results with IVIG. The first experience with IVIG in GBS was reported in 1988 in six patients with presumed severe deteriorating GBS. (Another two patients received IVIG combined with fresh frozen plasma infusion after they worsened following PLEX treatment.) A significant improvement in motor function was observed in four of six patients and was observed again after IVIG infusion was repeated in a patient with an early relapse. (CIDP was considered unlikely because of the absence of late relapses [9]).

Its ease of use was the major incentive to study IVIG prospectively. In the early days of GBS treatment, there were concerns about the potential of virus transmission and risk of anaphylaxis in immunoglobulin A–deficient patients. However, IVIG has been proven to be safe, with few side effects, and is comparable in expense with PLEX.

The IVIG study was limited to patients with severe GBS (i.e., unable to stand unassisted), and patients were randomized into 73 patients treated with PLEX and 74 patients treated with immunoglobulin. Again, outcome assessment was not blinded. Patients were treated with 0.4 g/kg/day of immunoglobulin for 5 days compared to PLEX of 200 to 250 mL/kg/body weight in five sessions within 7 to 14 days. The trial was discontinued after 150 patients and after a difference of 19% in favor of IVIG was found. Most noticeable in this trial was less improvement by one functional grade in the immunoglobulin at 4 weeks. In the North American trial, 52% of the patients improved one grade as opposed to 34% in the immunoglobulin study. The study also found a sizable proportion of patients (16%) who could continue because of hypotension or other side effects after one or more sessions of PLEX. Three years after this PLEX trial, IVIG was introduced as an alternative treatment for GBS [10]. After the PLEX study, another study compared PLEX in combination with IVIG. No difference was found between IVIG and PLEX; combining both therapies (PLEX followed by IVIG) also failed to yield a better outcome. Studies comparing IVIG with corticosteroids found no additional benefit.

PLEX has remained an important treatment modality in patients with GBS, although IVIG has become the preferred first approach [11]. There is still insufficient understanding of how IVIG works in inflammatory disorders. The potential pathways include inhibition of autoantibody-mediated injury. IVIG may also counteract the interaction of T cells with antigen-presenting cells. IVIG reduces interleukin-1 and tumor necrosis factor-α levels and increases transforming growth factor-β production. IVIG also prevents binding and complement activation of autoantibodies to their antigens by anti-idiotypic antibodies. Each of these mechanisms may be operative.

There is a comparable history regarding the use of PLEX in myasthenia gravis [12]. A course of daily PLEX was undertaken in three patients with severe myasthenia gravis who had failed to respond to anticholinesterases, thymectomy, and steroids. Contrary to GBS, their improvement was rapid and impressive. “This improvement began 2–4 days after initiation of daily exchanges and was progressive over the course of exchanges, continuing for some days after this was completed. The improvement was independent of whether the patient was taking anticholinesterase medication” [12]. There was a more logical reason for why it worked: this clinical response to PLEX strongly suggested that a plasma factor produced the neuromuscular transmission disorder in the acquired form of this disease, and they speculated that this factor was the circulating antibody to AChR. No similar antibody or other factor that was pathognomic in GBS and declined after PLEX has been found. So do these interventions work and help the patient or did we see a bunch of trials comparing “placebos”?. Not the past, but the future may tell.