Variability in the efficacy of the IL1 receptor antagonist anakinra for treating Erdheim-Chester disease

To the editor:

Erdheim-Chester disease (ECD) is a rare non-Langerhans cell histiocytosis characterized by foamy histiocytes, sometimes associated with central nervous system (CNS) and cardiac infiltration.¹  Several reports have suggested that an interleukin (IL)1 receptor antagonist, anakinra, is effective in treating ECD.^2-8 

Twelve patients (7 men and 5 women; median age, 70 years; range, 22-80 years) with biopsy-proven ECD⁹  and previous failure (cases 2, 3, 4, 9, and 11), poor tolerance, or contraindication (cases 1, 2, 5, 6, 7, 8, 10, and 12) to interferon-α (IFN-α) therapy received alternative treatment with daily subcutaneous injection of 100 mg anakinra. Data collection and analysis were performed retrospectively. Three patients rapidly stopped the treatment (2 for failure and 1 for pain at the injection site). Table 1 shows the characteristics of the 12 initial and the 9 remaining patients at baseline. Eight patients had long bone involvement, 8 had cardiac disease (4 with a pericardial effusion or thickening and 8 with a pseudomass in the right atrium), 3 had retro-orbital infiltration, 1 had a pseudo-meningioma of the spine, and 1 had pachymeningitis. The median duration of ECD disease when anakinra was started was 5 years (range, 1-12 years). Ten of the 12 patients were treated prior to inclusion with IFN-α (pegylated or not). For 5 of them, anakinra was given just after IFN-α treatment.

The median duration of anakinra treatment (always given as a monotherapy) among the 9 patients was 22 months (range, 3-34 months). Clinical symptoms improved in 50% of the patients (3 of 7 for fatigue, among them 2 stopped IFN-α just before, 1 of 1 for fever, 2 of 6 for bone pain, 1 of 2 for sinus disturbance). The median C-reactive protein (CRP) value at baseline was 10 mg/L (range, 4-30 mg/L) and at the end of the treatment was 10 mg/L (range, 4-43 mg/L), which was not significantly different. All patients had a positron emission tomography (PET) evaluation at baseline and at the end of follow-up. Five had progression of the disease, 2 had stable disease, 1 had a partial response, and 1 had a complete metabolic response. Among the 6 patients who had cardiac involvement, 5 underwent baseline and final cardiac magnetic resonance imaging (MRI), which displayed stability in 3 and progression in 2 (among them, 1 had tamponade). Retro-orbital infiltration evaluated by cerebral MRI was stable in 1 of 1 patients and CNS infiltration appeared in 1 patient free of CNS involvement at the beginning of the treatment.

Tolerance was variable. Four of the 9 patients included in the long-term analysis had side effects, including pain at the injection site (n = 4), edema (n = 1), severe sepsis (n = 1), and headache (n = 1). Overall, anakinra was stopped in 11 of 12 patients because of poor tolerance (n = 4) or progression of disease (n = 7).

Proinflammatory cytokines, such as IL-1, IL-6, and tumor necrosis factor-α, are strongly increased in ECD lesions. Whether IL-1 blood levels are increased is still controversial; this finding has been reported in isolated cases but not in a larger series of 37 patients.¹⁰  Altogether, these findings suggested that inhibition of the IL-1 pathway could be a promising therapeutic area for ECD treatment.

Here, we report a series of 12 patients with ECD exhibiting multisystem histiocyte infiltration treated with anakinra. Tolerance was variable and led to discontinuation of the treatment in 33% of patients. In contrast with 6 of 8 previous case reports, we did not observe an improvement in CRP levels. This may be due to IFN-α therapy, which may have lowered basal CRP levels in 4 patients who received this treatment just before receiving anakinra. The efficacy of anakinra assessed by fluorodeoxyglucose (FDG)-PET was variable. The extent of the disease evaluated by computed tomography or MRI imaging was not improved in our patients receiving anakinra, regardless of the site (retro-orbital, cardiac, spine, brain, pleura, or retroperitoneal). Conversely, CNS infiltration and tamponade occurred during treatment in 1 patient each.

Aouba et al²  first reported the dramatic efficacy of anakinra in 2 ECD patients with a dramatic improvement in retroperitoneal infiltration and regression of hydronephrosis. Since then, the use of IL-1 blockade in ECD has been reported in 6 additional reports^3-8  (Table 2), which did not mention any changes in bone lesions or retroperitoneal infiltration. Killu et al⁵  described a patient with a right atrial mass in which FDG uptake decreased under anakinra treatment, although there was no MRI evaluation of the cardiac infiltration. Some of our patients, conversely to several previous reports including the one by Aouba et al, were refractory to IFN-α therapy, which could be the hallmark of more severe disease and can explain the different responses to treatment.

Altogether, these results show that anakinra displays a variable efficacy in ECD sometimes, with progression of cardiac or CNS involvement.

In conclusion, anakinra is possibly effective for treating constitutional symptoms, has variable efficacy in organs as measured by FDG-PET scan, and did not lead to measurable regression of tumors. For ECD patients with cardiac or CNS infiltration, other treatments, such as IFN-α or B-Raf inhibitors, should be considered first. Increasing the anakinra dose should also be considered in the future, as has been proposed in other inflammatory conditions.