We included 68 patients seen at our outpatient hematology clinic between 2010 and 2015 for an initial workup of thrombocytopenia of unclear etiology (Table 1). The median age at thrombocytopenia diagnosis was 60 (range, 17–90) years, and the majority of patients (64.7%) were male. The median platelet count at diagnosis was 91 (range, 3–146)×10⁹/L: 30 patients (44.1%) with a platelet count of 100–146×10⁹/L, 26 (38.2%) with a platelet count of 50–100×10⁹/L, and 12 (17.6%) with a platelet count of <50×10⁹/L.

Although 37 patients (54.4%) and 27 patients (39.7%) underwent evaluation for autoimmune causes and infectious diseases, respectively, only the following few test results were positive: anti-glycoprotein platelet antibodies (N=2; although routine use of this testing is not recommended [4]), anti-nuclear antibodies (N=4), elevated rheumatoid factor level (N=1), and anti-IgM Epstein-Barr virus antibody (N=1). None had hepatitis B or C, or HIV despite routine use of these tests. Bone marrow biopsies performed on 16 patients (23.5%) revealed abnormalities in 2 patients (both with coexisting anemia and leukopenia at the time of the thrombocytopenia workup): one with hairy cell leukemia and one with myelodysplastic syndrome (refractory anemia with excess blasts-2).

Based on the results of these tests, 27 patients (39.7%) were diagnosed with primary immune thrombocytopenia, and 23 patients (33.8%) were diagnosed with thrombocytopenia due to possible multifactorial causes, 5 (7.4%) suspected to be due to drugs (tacrolimus, methotrexate, bosentan, amitriptyline, or adalimumab), 4 (5.9%) due to hypersplenism, 4 (5.9%) due to unclear etiology, and 2 (2.9%) due to hematologic malignancies; 1 patient (1.5%) each was diagnosed with thrombocytopenia secondary to lupus, liver cirrhosis secondary to non-alcoholic steatohepatitis associated with hypersplenism, and gestational thrombocytopenia.

Patients with thrombocytopenia are often recommended to undergo a variety of tests, much of which may not be indicated and/or diagnostically helpful. In this series, we found that although a large percentage of patients underwent multiple laboratory tests and pathologic evaluation, these tests often did not offer significant insight into the etiology of the thrombocytopenia. This is congruent with results from other studies specifically focusing on bone marrow biopsies in the workup of thrombocytopenia [12] and further adds to the literature with regard to autoimmune and infectious workup. Additionally, nearly half of the patients in this series had a platelet count >100×10⁹/L, which may in some cases be considered a normal variant [3].

In summary, evaluation of thrombocytopenia should vary based on individual patient history and risk factors rather than a one-size-fits-all testing panel. Our study was based on a small cohort of patients but is likely representative of clinical practice at many large tertiary care centers. Large-scale prospective studies may be helpful to define the optimal workup for thrombocytopenia and formulate a protocol for cost-effective workup by stratification of clinical risk factors.