Abstract
Introduction: New techniques for supporting patients has led to a decline in the use of ECMO in neonatal and pediatric patients. Nineteen neonatal centers reported > 20 ECMO cases in '93 as compared to 16, 10 and 9 centers in '94, '95 and '96, respectively. To assess the impact frequency of usage of ECMO has on survival, an evaluation of the ELSO registry databank from 1993-96 was conducted. Methods:Approximately 100 ECMO centers submitted data to the ELSO registry during each of the yean 1993-96. Chi square tests of significance (p<.05) were computed to determine 1) if overall frequency of use of ECMO varied over time, 2) distribution, use and survival of neonatal and pediatric patients receiving ECMO and 3) if frequency of use of ECMO/center was associated with survival. Results: 1. Use of ECMO is daclining while survival is remaining steady. 1993 1994 1995 1996 Neonatal 1312 1237 1018 875 (% surv.) (79%) (79%) (78%) (77%) Pediatric 201 199 201 158 (%surv) (54%) (60%) (55%) (59%) 2. Neonatal centers performing >20 cases of neonatal ECMO/yr have a slightly higher survival (81%) than centers supporting <20 cases/year (77%) (p<.05). This is in the context of declining numbers of centers performing >20 ECMO over the years. 3. Centers that support <10 cases of pediatric ECMO/yr have survival rate of 54% vs. 73% in centers with 11-20 cases (p<.05). Conclusions:ECMO centers with larger numbers have improved survival. Whether this difference (1) relates to a selection bias in low frequency ECMO centers, (i.e. pts. who receive ECMO have failed lass invasive alternative therapies and are "sicker" when placed on ECMO than in high-frequency ECMO centers), or (2) indicates that higher frequency ECMO centers have improved survival secondary to increased proficiency/less complications than low frequency centers requires more evaluation. To prove that a center's ECMO frequency is truly related to mortality, data on number of admissions/center, use of ECMO and survival is needed. Correlation of risk of death (PRISM or similar), use of other therapies and overall outcome may be useful. Crit Care Med 1999 Vol. 27, No. 1 (Suppl.).