The Incidence and Characterization of ITP in Prostate Cancer

INTRODUCTION:

Idiopathic thrombocytopenic purpura (ITP) is a common autoimmune disorder. The association between ITP and lymphoid neoplasms is well known, but the association of ITP with solid tumors is poorly documented. To date only one case of prostate cancer associated with ITP has been reported in the literature and the association was unclear. The purpose of this study is to estimate the incidence of ITP in men with prostate cancer and to characterize its clinical course.

METHODS:

We performed a retrospective cohort study using a large multi-institutional database of 3,258,677 US veterans to identify those with and without prostate cancer (ICD-9 code 185). We evaluated them during the years 2006-2011 for the development of ITP using ICD-9 code 287.31. We examined the association of the presence of prostate cancer with incident ITP using Cox proportional hazard analyses, with adjustments for age, race, and presence of systemic lupus erythematosus, lymphomas and rheumatoid arthritis.

In order to increase the specificity and sensitivity of our evaluation and to analyze the characteristics associated with ITP in prostate cancer, we then evaluated patients visiting the Memphis Veterans Administration Medical Center (VAMC) between January 2006 and December 2011. We identified 3973 patients with prostate cancer and randomly selected 4801 patients without prostate cancer during the same time period. In both of these cohorts, we performed an in depth chart review of all the patients with thrombocytopenic disorders, expanding the ICD.9 codes to 287.31 (ITP) and 287.5 (Thrombocytopenia NOS). We identified patients with ITP based on the American Society of Hematology criteria for the diagnosis. We then reviewed the ITP identified charts for characteristics of disease.

RESULTS:

Among 3,258,677 US veterans without a baseline diagnosis of thrombocytopenia, we identified 265,446 patients with prostate cancer and 2,993,231 patients without prostate cancer. In both these cohorts 1,847 patients developed ITP according to ICD.9 code 287.31. The incidence of ITP in patients with prostate cancer was 12.3/100,000 patient years (95%CI: 10.7-14.2) and in patients without prostate cancer was 9.5/100,000 patient years (95% CL: 9.0-10.0). Prostate cancer was associated with a 30% higher risk of incident ITP in unadjusted analyses (hazard ratio, 95%CI: 1.30, 1.12-1.51, p=0.001), but the difference was non-significant after multivariable adjustments (adjusted HR, 95%CI: 1.11, 0.95-1.30, p=0.18). Men with prostate cancer were older than controls (mean age 69 vs 59). Patients with ITP were more likely to have lupus (2.0% vs. 0.2%), lymphomas (7.3% vs. 0.8%) and rheumatoid arthritis (3.6% vs. 1.8%).This analysis provided a broad correlation between ITP and prostate cancer but lacked sensitivity due to the specificity of the single ICD-9 code. In our chart review of 3973 veterans with prostate cancer, we identified six cases of ITP (30/100,000 patient years). Among 4801 veterans without prostate cancer, none developed ITP. These numbers were too small to demonstrate clinical significance. The prostate cancer patients in our local institution were older than the controls (67.6 vs 57.7), similar to the age discrepancy in the large VA database. In none of the patients with ITP, thrombocytopenia was attributable to prior irradiation or chemotherapy. Two patients had severe ITP episodes preceding prostate cancer progression. Four patients had ITP worsening with stable prostate cancer and one patient had stable ITP with prostate cancer progression. Two of the ITP patients never required treatment, three required and responded to steroids and immunosuppressive drugs, and of these one required splenectomy.

CONCLUSION:

ITP, identified by the ICD-9 code 287.31 in a large multi-institutional survey, occurs more often in patients with prostate cancer, but this association may be due to confounding variables such as age. The true incidence is likely higher than this. At the Memphis VAMC, utilizing broader ICD.9 codes, including both ITP and Thrombocytopenia NOS, followed by an in-depth chart review, there is a strong suggestion of an increased incidence of ITP among prostate cancer patients, but the numbers are too small to be clinically significant. The clinical course of ITP in prostate cancer patients is similar to that seen with other disorders and does not clearly parallel the course of the prostate cancer.