Prevalence of diagnosed chronic immune thrombocytopenic purpura in the US: analysis of a large US claim database: a rebuttal.

Abstract

We read with interest the paper by Segal and Powe [1] on the prevalence of immune thrombocytopenic purpura (ITP). ITP is a disease caused by inadequate platelet production, as well as increased platelet destruction, and is traditionally categorized into acute and chronic forms. For the latter, thrombocytopenia needs to be present for at least 6 months [2]. In this paper, Segal and Powe refer to ITP as immune (rather than idiopathic) thrombocytopenia purpura. There is relatively little epidemiological evidence on chronic ITP, and its prevalence in the US is poorly documented. The paper by Segal and Powe is one of the first population-based studies of prevalence of ITP in the US that provides an estimate of chronic ITP (albeit informal). However, as the authors noted, there were some limitations in their study. Most of these were related to the limitations in the data source. In the analysis presented here, we aimed to estimate the diagnosed prevalence of chronic ITP in the US in 2005, using a large US claims data base. In particular, our analysis overcomes some of the main limitations in Segal and Powes paper. (i) Segal and Powe used only one calendar year (2002) of data and identified potentially chronic ITP patients as those with at least two ITP ICD-9 codes (287.3), which were separated by at least 6 months. Therefore, a patient whose first ITP ICD-9 code occurred in the second half of 2002 had no chance of presenting a second code within the 2002 calendar year. Hence, as the authors pointed out, their estimation of the prevalence of chronic ITP in all people under 65 years (4.5 per 100 000) may be a considerable underestimate. (ii) Segal and Powe se stimation only includes patients under 65 years old. The increased risk of ITP with age is well documented [3–5] and excluding patients over 64 years of age would lead to an underestimate. We analyzed the Integrated Healthcare Information System (IHCIS) database, one of the largest US health care managed databases, fully de-identified, with over 70 million patients from more than 45 health plans. It covers 7 out of 10 census regions and contains patient demographics, age, gender and morbidity. Our patient population consisted of patients enrolled to one of the health plans before or any time between 2002 and 2006 with a continuous enrollment throughout the year 2004. Chronic ITP cases were defined as patients with at least two diagnoses (ICD-9 code 287.3) for primary thrombocytopenia separated by at least 6 months between 2002 and 2006. A patient sfi rst identified 287.3 ICD-9 code over the period 2002–2006 had to be in 2004 or earlier and the last one in 2004 or later. Prevalence rates of chronic ITP were calculated as the total number of identified chronic ITP cases divided by the total population with continuous enrollment in the data base in 2004 stratified by age group and gender. The total age- and gender-adjusted rate was estimated as the total number of expected ITP cases divided by the total 2005 US population (given the observed age- and gender-specific rates, and the structure of the 2005 US population). To assess chronic idiopathic thrombocytopenia rather than chronic immune thrombocytopenia, and following the Segal and Powe approach, we repeated the above estimates, excluding patients with a diagnosis of human immunodeficiency virus (HIV) infection, hematological malignancies or aplastic anemia, any time between 2002 and 2004.