Besides infectious complications, severe acute flares of ANCA-associated vasculitis are common reasons that warrant ICU admission [1]. Plasma exchange (PLEX) has been proposed as an urgent adjuvant treatment in patients with life-threatening organ dysfunctions [2, 3]. In order to explore this question, we conducted a retrospective monocenter study in our tertiary ICU. We included patients admitted to the ICU for acute respiratory failure related to DAH, diagnosed as ANCA-associated vasculitis, and who received urgent initiation of PLEX. DAH was defined by bilateral infiltrates on chest X-ray and macroscopically bloody bronchoalveolar lavage with hemorrhagic and siderophagic alveoliitis. PLEX was performed daily with 1.2 plasma volume plasmapheresis primarily substituted with fresh frozen plasma and then albumin 5% and fresh frozen plasma when needed to maintain a prothrombin time > 50% and a fibrinogen level > 1.5 g/L. The main outcome was the evolution of oxygenation over the first seven days, using the SpO2/FiO2 ratio. We present data as median [interquartile range] or number (percentage) as appropriate. P for trend for continuous variables was calculated using a Cuzick test.

Between 2006 and 2014, 12 patients were treated by PLEX in the ICU for ANCA-vasculitis with respiratory symptoms (Table 1). All patients received high-dose corticosteroids (≥ 1 mg/kg prednisone-equivalent) and additional immunosuppressive drugs, either cyclophosphamide (nine within 24 h before or after ICU admission and one after ICU discharge) or rituximab administrated in the ICU (n = 2). One patient died from refractory multiple organ failure related to septic shock. Invasive mechanical ventilation was required in five patients (two received high-frequency oscillation ventilation). One patient received adjuvant nitric oxide. Duration of invasive mechanical ventilation ranged from 6 to 20 days. Three patients successfully received non-invasive ventilation. Oxygenation improved over the first week, as shown by the increase in the SpO2/FiO2 ratio from 183 [137–321] to 353 [239–432] (p value for trend 0.003), along with a decrease in the level of ventilatory support (Fig. 1). In contrast, only one out of five patients could be weaned off dialysis.

Table 1 Characteristics of patients
Full size table
Fig. 1
figure 1

Evolution of respiratory dysfunction as assessed every 12 h over the first 7 days (d1 to d7) from initiation of plasma exchange. Black dots and lines represent the SpO2/FiO2 ratio (median and interquartile range). Background histograms show the distribution of ventilatory support. NIV non-invasive ventilation, IMV invasive mechanical ventilation

Full size image

In conclusion, this suggests the addition of PLEX results in fast respiratory recovery in most patients. This contrasts with the limited impact on renal function. The effects of PLEX are presumably related to fast removal of auto-antibodies as well as pro-inflammatory mediators likely to induce and/or sustain the increased permeability of the alveolo-capillar barrier.