Cyclic thrombocytopenia with statistically significant neutrophil oscillations

Abstract

Cyclic thrombocytopenia is often misdiagnosed as immune thrombocytopenia due to similar clinical features, a fact of significance because cyclic thrombocytopenia generally responds poorly to treatments used successfully in immune thrombocytopenia. A precise diagnosis must establish the statistical significance of periodicity of the platelet counts using statistical methods (eg, Lomb-Scargle periodogram).

Keywords: bleeding; cyclic neutropenia; cyclic thrombocytopenia; immune thrombocytopenia; thrombopoietin.

Figure 1

Patient 1 data, with time measured in days since diagnosis. A, Platelet counts (data shown is from April 2003 to April 2006). B, Platelet power spectrum. C, Neutrophil counts (data shown is from April 2003 to April 2006). D, Neutrophil power spectrum. E, Platelet counts and TPO levels over one cycle (data shown is from 23 June 2006 to 30 July 2006, with TPO data points denoted by squares). F, Platelet count for three cycles during three periods of daily treatment with 50 mg of eltrombopag (data shown is from 5 November 2010 to 14 March 2011, with periods of treatment demarcated by dashed lines and double arrows)

Figure 2

Patient 2 data, with time measured in days since diagnosis. A, Platelet counts (data shown is from April 2003 to April 2006). B, Platelet power spectrum. C, Neutrophil counts (data shown is from April 2003 to April 2006). D, Neutrophil power spectrum. E, Platelet count before and after daily treatment with 50 mg of eltrombopag (data shown is from 5 January 2009 to 2 March 2009, with period of treatment demarcated by dashed lines and double arrows)