Purpose: The association between extracorporeal membrane oxygenation (ECMO) use and the development of thrombocytopenia is widely presumed yet weakly demonstrated. We hypothesized that longer duration of ECMO support would be independently associated with worsened thrombocytopenia. Methods: We performed a single-center retrospective cohort study of 100 adults who received ECMO for acute respiratory failure. We used generalized estimating equations to test the association between days on ECMO and daily percentage of platelets compared to the first post-cannulation platelet count. We constructed a multivariable logistic regression model with backwards stepwise elimination to identify clinical predictors of severe thrombocytopenia (≤50,000/μL) while on ECMO. Results: Days on ECMO was not associated with a decrease in platelet count in the unadjusted analysis (β −0.85, 95 % CI −2.05 to 0.36), nor after considering and controlling for days hospitalized prior to ECMO, APACHE II score, platelet transfusions, and potential thrombocytopenia-inducing medications (β −0.83, 95 % CI −1.9 to 0.25). Twenty-two subjects (22 %) developed severe thrombocytopenia. The APACHE II score and platelet count at the time of cannulation predicted the development of severe thrombocytopenia. The odds of developing severe thrombocytopenia increased 35 % for every 5-point increase in APACHE II score (OR 1.35, 95 % CI 0.94–1.94) and increased 35 % for every 25,000/μL platelets below a mean at cannulation of 188,000/μL (OR 1.35, 95 % CI 1.10–1.64). Conclusions: Duration of ECMO is not associated with the development of thrombocytopenia. The severity of critical illness and platelet count at the time of cannulation predict the development of severe thrombocytopenia while receiving ECMO for respiratory failure. Future studies should validate these findings, especially in cohorts with more venoarterial ECMO patients, and should characterize the association between thrombocytopenia and bleeding events while on ECMO.