Eosinophilia is not a rare event in patients requiring hemodialysis, and has been described since the 70s [1, 2], either isolated or associated with minor symptoms such as abdominal pain or moderate hypotension during dialysis sessions. These clinical manifestations have been indiscriminately clustered as “reaction to dialyzer membrane materials” [3]. Some rare cases of major hypereosinophilia presenting as part of a so-called allergy to hemodialysis material have been reported in the last decades in the chronic dialysis setting [3, 4].

Multiple components of the extra corporeal circuit (dialysis membrane, catheters, dialysate, etc.) have been implicated in the occurrence of eosinophilia [5, 6]. Chapelet-Debout et al. recently reported 6 cases of major hypereosinophilia attributed to the tunneled central venous catheter [6], but in our case, the catheter was not tunneled and consisted in a classical percutaneous double lumen dialysis catheter used in intensive care units in emergency dialysis settings. The dialysate was also taken into consideration in our etiologic investigation, but in our ICU we have a central water treatment system integrated into the dialysis loop, and the purified osmosis water is tested monthly. During our patient’s ICU stay, bacterial cultures were < 100 UFC/ml, and testing for endotoxins was < 0.25 UI/ml, as required. Moreover, at the time of our diagnosis, no other patient undergoing dialysis from this loop developed any adverse effects, including eosinophilia (approximately 40 patients were treated during the time interval of our patient’s stay). Yet, the onset of eosinophilia 3 days after implementation of IHD and 20 days after the patient was started on CVVHD pointed to the dialyzer membrane as the culprit since eosinophilia typically develops within a couple of days following exposure to the dialysis membrane.

In the era of modified cellulose or synthetic membranes, as used in our ICU, severe symptomatic allergic reactions are infrequent. Thus, in recent years, hypereosinophilia (> 1.109 G/L) has been reported in an estimated 5% of patients on chronic hemodialysis [5]. It is therefore important, in the case of hypereosinophilia in a patient undergoing dialysis, to have an exhaustive, step-by-step, etiological diagnostic approach before blaming the dialysis equipment (Fig. 2).

Proposed step-by-step etiological approach for hypereosinophilia in a dialysis setting. AKI acute kidney injury, ANA antinuclear antibody, ANCA anti-neutrophil cytoplasmic antibody, CT computed tomography, DRESS drug rash with eosinophilia and systemic symptoms, IHD intermittent hemodialysis, SPE serum protein electrophoresis. *Osmosis water with bacterial cultures < 100 UFC/ml, and testing for endotoxins < 0.25 UI/ml regularly checked

Treatment with ACE-inhibitors is the major known factor associated with hypereosinophilia in this setting [5], potentially due to enhanced bradykinin release during allergic episodes, but our patient never received any.

The pathophysiology of hypereosinophilia and allergic reaction related to the hemodialysis circuit is unclear, although cytokine release following eosinophil activation by the membrane components has been hypothesized [7]. Direct and alternative complement pathways might also be involved, with C3a and C5a fractions potentially participating in eosinophil activation [8]. In the setting of long-term hemodialysis patients, resolution of eosinophilia has been achieved through treatment with systemic corticosteroids (0.5–1 mg/kg) and/or changing the membrane [9]. In our case, immunosuppression induced by the prolonged ICU stay, the absence of vital organ damage and the gradual but consistent resolution of eosinophilia following dialysis weaning discouraged us from using corticosteroids.

Hypereosinophilia linked to dialyzer membrane should be taken into consideration when dealing with critically ill patients after all competing causes have been carefully ruled out.