Risk Factors for Transition From Normal Platelet Count to Severe Thrombocytopenia Among Patients with Chronic Liver Disease.

Abstract

Abstract 3518Poster Board III-455 IntroductionThrombocytopenia [TCP] is a complication of advanced chronic liver disease [CLD]. We examined clinical factors related to the development of severe TCP among patients with initially normal platelet counts using laboratory, clinical, and administrative data from the Veterans Administration Healthcare System, New England region. MethodsAll patients with a diagnosis of CLD or hepatitis C who were seen in a VA hospital or in a VA outpatient clinic between 10/2002 and 9/2008 were included in this analysis. CLD diagnosis definitions included: 1) ICD9-CM code of 51.0 – 571.9, excluding 571.1at a clinic visit or hospital discharge 2) positive hepatitis C [Hep-C] antibody, and 3) elevated ALT (> 84 IU/dL) on two occasions at least 60 days. Subjects were required to have normal platelet counts (platelet count >150,000/μL) at baseline. Subjects were excluded if they had the following: 1) hepatorenal syndrome, coagulation disorders, immune TCP purpura, lupus erythematosis, rheumatoid arthritis, pernicious anemia, aplastic anemia or septicemia. Severe TCP was defined as a platelet count < 50,000/μL. Risk factors evaluated for the development of TCP included age, comorbidities (diabetes, alcoholic and/or drug dependence, chronic kidney disease, Hepatitis-C [Hep-C] status, hepatorenal syndrome, congestive heart failure [CHF], coronary artery disease, hypertension), resource utilization (infectious disease, gastrointestinal [GI] nephrology and/or hematology clinic visit), and current laboratory values (hemoglobin, ALT and MELD Score). All potential predictors were treated as time-varying exposures over the follow-up and updated at each interval. Relative risk was estimated using logistic regression with pooled repeated observations, using 3-month periods where risk factors at the beginning of a 3-month interval were used to predict severe TCP in that interval. Results6,102 subjects with Hep-C contributed a median follow-up of 1,414 days and had 82 events (1.34%) during the study. 5,358 Non Hep-C subjects contributed a median follow-up of 1,238 days and had 97 events (1.81%) during the study. The median platelet count among CLD patients at baseline was 238,000/μL (IQR 203,000/μL- 282,000/μL) and among Hep-C patients at baseline was 240,000 (IQR 205,000/μL- 285,000/μL). After multivariate adjustment, alcohol dependence [Odds ratio = 2.24, 95% confidence interval: (1.14, 4.39)], CHF [2.6 (1.05,6.43)], 5 unit change in MELD [1.13 (1.05,1.21)], GI clinic visits [1.89 (0.93,3.87)] and anti retroviral therapy [3.84 (1.60, 7.61)] increased the odds of severe TCP; while drug dependence [0.62 (0.34, 1.13)] and increasing hemoglobin [0.74 (0.65,0.84)] reduced the odds of severe TCP. In non Hep-C subjects, alcohol dependence [2.88 (1.58,5.25)], CHF [1.81 (0.9, 3.65)], 5 unit change in MELD [1.12 (1.03, 1.23)], hematology clinic [3.64 (1.91, 6.93)] current anti-retroviral therapy [6.65 (1.96, 22.6)] and ALT 42-84g/dL vs. ALT < 42g/dL [1.55 (0.78, 3.06)] increased the odds of severe TCP; while drug dependence [0.46 (0.17, 1.22)], increasing hemoglobin [0.68 (0.60, 0.78)] and ALT > 84g/dL vs. ALT < 42 reduced the odds of severe TCP. ConclusionsIn a population of veterans with chronic liver disease or hepatitis C infection with normal platelet counts > 150,000/μL at baseline, we estimate the incidence of development of severe thrombocytopenia to be between 1 and 2% over an average of 3 years of follow-up. Key clinical predictors of development of severe TCP included prevalent CHF diagnosis and changing MELD score. Specialist care referrals to hematologists and gastroenterologists, also identified patients likely to become severe TCP cases. As expected, antiretroviral therapies were associated with an increased risk of severe TCP over the follow-up. AcknowledgmentsGlaxoSmithKline provided research support for this project, with no restrictions on publication. The study was conducted using resources of the VA Cooperative Studies Program and VA Boston Healthcare System. Disclosures:Lawler:GlaxoSmithKline: Research Funding. Horblyuk:GlaxoSmithKline: Employment. Grotzinger:GlaxoSmithKline: Employment, Research Funding.