Natural history of severe eosinophilia with uncertain aetiology and proposals on a practical approach to its management.

BACKGROUND: Eosinophilia is commonly encountered during clinical practice. Some can be attributed to well-defined causes while others cannot. Optimal management of hypereosinophilia with unknown aetiology is uncertain as the natural history is not well described. METHODS: We retrospectively studied patients with hypereosinophilia (>5 × 10(9)/L) and described the characteristics, natural history and treatment of those with eosinophilia of uncertain aetiology. RESULTS: There were 141 patients with hypereosinophilia: 87 with well-defined causes, 54 with uncertain aetiology. The latter was managed as hypereosinophilic syndrome (HES) (n = 5), idiopathic hypereosinophilia (IH) (n = 11), presumptive helminthic infection (n = 11) and reactive eosinophilia (n = 5), while 22 were insufficiently investigated and did not have definite working diagnoses. Their median age and peak eosinophil count were 64 (22 to 94) years and 10.0 (5.2-33.9) × 10(9)/L respectively. Forty-six per cent had symptoms attributable to eosinophilia, with the HES and insufficiently investigated groups having the highest (100%) and lowest (27%) percentages respectively. HES and IH patients were most extensively investigated. All 14 HES or IH patients who received steroids responded. All presumptive helminthic infection patients received mebendazole: nine responded, and two had unassessable responses. For the remaining patients, seven received steroids and all responded; one received mebendazole but defaulted; 19 were not treated: 11 resolved spontaneously. No non-HES patients developed eosinophilia-related organ dysfunction. No mortality was caused by hypereosinophilia. CONCLUSIONS: Patients with hypereosinophilia of uncertain aetiology can be empirically managed according to working diagnoses derived from history taking, examination and selective investigations. Most patients have benign short-term outcomes, but longer monitoring is required to assess long-term outcomes from untreated hypereosinophilia.