PEXIVAS

• ANCA
• Plasmapheresis
• PEXIVAS
• renal
• pulmonary hemorrhage
• severe

Kidney involvement in ANCA-associated vasculitis is common (64%–85%) and associated with high morbidity and mortality. Despite recent advances in therapy, a significant number of patients still progress to kidney failure or death. As such, therapies that improve kidney outcomes have been pursued.

The rationale for using plasma exchange in ANCA-associated vasculitis rests heavily on the assumptions that ANCAs are pathogenetic and can be efficiently removed by plasma exchange. Plasma exchange could also exert a beneficial effect by eliminating mediators of inflammation and tissue injury. Small historical studies of plasma exchange added to standard remission induction therapy yielded conflicting results (1). The strongest support for plasma exchange comes from the MEPEX trial, which randomized 137 patients with ANCA-associated vasculitis (52% MPO-ANCA and 43% PR3-ANCA) and baseline serum creatinine >5.8 mg/dl to either seven sessions of plasma exchange or three intravenous pulses of methylprednisolone 1 g in addition to remission induction therapy (2). Plasma exchange significantly reduced the risk of progression to kidney failure at 12 months, but patient survival was not affected (2). However, a post hoc analysis found that the certainty of the effect was lost over longer-term follow-up (3). A meta-analysis of randomized controlled trials that compared standard care with standard care plus plasma exchange in adults with kidney vasculitis or idiopathic rapidly progressive GN calculated that plasma exchange reduced the risk of kidney failure by 20%. However, the number of patients was small, and sensitivity analyses were insufficiently robust to draw reliable conclusions (1). A trial of 500 patients showing at least a 25% relative risk reduction would be needed to confirm a treatment effect of at least moderate size (1).

As such, the PEXIVAS trial was conducted in patients with ANCA-associated vasculitis and kidney involvement (with an eGFR of <50 ml/min per 1.73 m²) and/or pulmonary involvement (with diffuse alveolar hemorrhage) (4). The trial had a two-by-two factorial design: initial treatment with plasma exchange was compared with no plasma exchange, and a standard dose oral glucocorticoid regimen was compared with a reduced dose regimen. Patients were treated with either oral or intravenous cyclophosphamide (85%) or rituximab (15%) according to the investigator’s preference. In contrast to MEPEX, all patients were treated with daily intravenous methylprednisolone for 1–3 days for a maximum cumulative dose of 1–3 g. A total of 704 patients (59% MPO-ANCA and 41% PR3-ANCA) were enrolled from June 2010 to September 2016, 98% of whom had kidney involvement. Kidney biopsy was not required, and concomitant antiglomerular basement membrane antibody was an exclusion criterion. At entry, the median serum creatinine was 3.7 mg/dl, and 20% required hemodialysis. After a median follow-up of 2.9 years (range, 2–7 years), the primary composite end point (death from any cause or kidney failure) occurred in 28.4% of the plasma exchange group and in 31.0% of the control group (P=0.27). The effects of plasma exchange on kidney failure and death were concordant.

The neutral findings of PEXIVAS were met with disappointment by some in the nephrologic community and elicited a number of critical comments (5,6). One criticism pertains to the low intensity of the plasma exchange regimen used in PEXIVAS (up to seven sessions over a 14-day period) compared with the regimen advocated in antiglomerular basement membrane disease (daily sessions for 2–3 weeks) (5). ANCA titers were not measured in PEXIVAS, but the plasma exchange regimen was identical to that used in MEPEX and in a small study revealing that five to seven sessions of plasma exchange decreased ANCA by 88.3% (7). Because the vast majority of patients (96.3%) received plasma exchange as per protocol, the absence of benefit cannot be attributed to a lack of adherence to the plasma exchange regimen. It has also been argued that fresh frozen plasma would have been a better replacement fluid than the 3%–5% albumin that was used because restitution of consumed plasma components, such as factor H and DNase, may contribute to control of ANCA-associated vasculitis (5). Such criticism is purely speculative and lacks clinical data to support it. The American Society for Apheresis guidelines (https://www.apheresis.org) recommend the use of albumin on the basis of its superior side effect profile compared with fresh frozen plasma. The MEPEX trial also used albumin as plasma substitute.

A more pertinent issue is that of a possible type 2 error. Was the power of PEXIVAS too low to demonstrate an effect in the two groups that are expected to derive the most benefit from a rapid reduction in ANCA titers: those presenting with severe kidney involvement or with severe pulmonary hemorrhage? In the subgroup of patients with baseline serum creatinine >5.6 mg/dl or undergoing dialysis (29% of the enrolled patients), PEXIVAS did not reveal a significant beneficial effect of plasma exchange. Although the combined outcome of kidney failure or death was not statistically different at 12 months, a visual appreciation of the survival curves suggests that a potential benefit of plasma exchange during the first 12 months was subsequently lost, which would be consistent with the findings in MEPEX. It could be argued that, even in the absence of a long-term benefit, a dialysis-free holiday of 12 months has a substantial effect on quality of life and is worth the trouble of seven sessions of plasma exchange. PEXIVAS may have been insufficiently powered to detect small or temporary effects of plasma exchange on impactful outcomes, such as death and kidney failure requiring dialysis.

The neutral results of PEXIVAS in the subgroup of patients with severe kidney involvement are corroborated by a recent retrospective cohort study conducted at the Mayo Clinic (8). A total of 252 patients with ANCA-associated vasculitis and severe active kidney disease (eGFR<30 ml/min per 1.73 m²) received cyclophosphamide (n=161) or rituximab (n=64) as remission induction therapy, of which 51 additionally were treated with plasma exchange. The PLEX regimen was more aggressive than that of PEXIVAS and consisted of seven to 14 daily sessions. Plasma exchange had no effect on remission induction at 6 months and on progression to kidney failure or survival at 18 and 24 months (8).

Even though patients with preexisting kidney disease were excluded, PEXIVAS was also criticized because kidney biopsy data were not reported (4). Although an uneven distribution of pathology scores among the groups seems unlikely in a randomized trial of this size, it cannot be excluded that enrollment of patients with severe chronic damage obscured a beneficial effect in patients with acute necrotizing GN and minimal scarring. In the Mayo Clinic cohort, diagnosis was biopsy proven in 222 patients, and of these, 199 kidney biopsies were reviewed. The histologic pattern was focal in 39 (19.6%), crescentic in 49 (24.6%), mixed in 73 (36.7%), and sclerotic in 38 (19.1%) patients. There were no differences in the distribution of the histologic categories in patients who received plasma exchange in comparison with the patients who did not receive plasma exchange (8). Only a minority of patients had kidney involvement not amenable to aggressive treatment. Although suffering from the limitations intrinsic to observational studies, the results from the Mayo Clinic cohort, obtained with a more aggressive plasma exchange regimen in a large group of patients with severe kidney involvement with availability of kidney biopsy results, support and add to those of PEXIVAS.

Finally, although 191 (27%) of the patients in PEXIVAS had pulmonary hemorrhage at presentation, only 61 (9%) patients with severe pulmonary hemorrhage, defined as oxygen saturation ≤85% on room air or requirement of mechanical ventilation, were included (4). In neither subgroup, plasma exchange had a statistically significant effect on the primary outcome of kidney failure or death from any cause, which has been attributed to a lack of power (6). Outcome for death alone was not analyzed. Because physicians may perceive little clinical equipoise regarding plasma exchange in patients with severe pulmonary hemorrhage, they may be reluctant to include these patients in a randomized clinical trial. In this setting, cohort studies may provide additional information. In a Mayo Clinic cohort of 73 patients with pulmonary hemorrhage, 41 were admitted to the intensive care unit, and 34 had respiratory failure requiring mechanical ventilation (9). Plasma exchange (a median of seven sessions) was added to standard remission induction therapy in 32 (44%) patients, of which 21 (62%) required mechanical ventilation. Conversely, 41 (56%) patients, of which 13 (38%) required mechanical ventilation, were not treated with plasma exchange. There was no difference in hospital mortality, length of hospital or intensive care unit stay, or duration of mechanical ventilation between the groups, and after adjustment for the imbalance of more critically ill patients receiving plasma exchange, the use of plasma exchange was not associated with achieving complete remission at 6 months. An aggregate analysis with ten other studies of patients with pulmonary hemorrhage revealed that resolution of pulmonary hemorrhage and being alive at hospital discharge occurred in 69 of 104 (66%) patients treated with plasma exchange, compared with 51 of 68 (75%) patients who did not undergo plasma exchange (P=0.22) (9).

In conclusion, the burden of proof does not justify the use of plasma exchange in patients with ANCA-associated vasculitis and kidney or pulmonary involvement. Doubts have been raised for the subgroups of patients with severe active kidney disease and severe pulmonary hemorrhage in view of the wide confidence intervals. Because the numbers of patients in these subgroups were already small, it is highly unlikely that further analysis of the PEXIVAS data will shed more light, although “if you torture the data long enough, it will confess to anything” (10). Understandably, little enthusiasm (and funding) will be garnered to conduct another trial of this magnitude. In the subgroups of patients for whom immediate downregulation of the inflammatory cascades can be lifesaving, perhaps attention should shift to pharmacologic intervention in the complement pathway. The oral complement C5a receptor inhibitor avacopan can effectively replace high-dose glucocorticoids in patients with ANCA-associated vasculitis (11). Further studies will help to define the potential role of complement inhibition in ANCA-associated vasculitis.

Regardless of the strength of the evidence against the routine use of plasma exchange in severe ANCA-associated vasculitis, with or without pulmonary hemorrhage, and the potential benefits of new therapies, some will continue to believe in its value. Old habits tend to die hard. As Claude Bernard said, “It’s what we think we know that keeps us from learning.”

• Copyright © 2021 by the American Society of Nephrology