Background Extracorporeal membrane oxygenation (ECMO) has been utilized for patients in critical condition, such as those with life-threatening respiratory failure or postcardiotomy cardiogenic shock. This study analyzed the outcomes of patients treated with ECMO and identified the relationship between prognosis and the Acute Kidney Injury Network (AKIN) scores obtained at pre-ECMO support (AKIN 0-hour ); and at post-ECMO support 24 hours (AKIN 24-hour ) and 48 hours (AKIN 48-hour ). Methods This study reviewed the medical records of 102 critically ill patients on ECMO support at a specialized intensive care unit at a tertiary care university hospital between March 2002 and January 2008. Demographic, clinical, and laboratory variables were retrospectively collected as survival predicators. Results The overall mortality rate was 57.8%. The most common condition requiring ECMO support was cardiogenic shock. Goodness-of-fit was good for AKIN 0-hour , AKIN 24-hour , and AKIN 48-hour criteria. The AKIN 0-hour , AKIN 24-hour , and AKIN 48-hour scoring systems also had excellent areas under the receiver operating characteristic curve (0.804 ± 0.046, 0.811 ± 0.045, and 0.858 ± 0.040, respectively). Furthermore, multiple logistic regression analysis indicated that AKIN 48-hour , age, and Glasgow Coma Scale score on the first day of intensive care unit admission were independent risk factors for hospital mortality. Finally, cumulative survival rates at 6-month follow-up after hospital discharge differed significantly ( p < 0.05) for AKIN 48-hour stage 0 versus AKIN 48-hour stages 1, 2, and 3; and AKIN 48-hour stage 1 and 2 versus AKIN 48-hour stage 3. Conclusions During ECMO support, the AKIN 48-hour scoring system proved to be a reproducible evaluation tool with excellent prognostic abilities for these patients.