A 20-year experience with urgent percutaneous cardiopulmonary bypass for salvage of potential survivors of refractory cardiovascular collapse.

OBJECTIVE

In-hospital cardiac arrest or refractory shock carries a high mortality despite the use of advanced resuscitative measures. We have implemented an in-hospital, nurse-based, continuously available, percutaneous, venoarterial cardiopulmonary bypass system, also known as extracorporeal life support (ECLS), as an adjunct to resuscitation when initial measures are ineffective.

METHODS

In 1986, a system for the rapid initiation of ECLS, was created in which trained critical care nurses primed an ECLS circuit and in-house physicians percutaneously placed required cannulas. From a prospective registry, we assessed long-term survival (LTS) (> or =30 days, cardiopulmonary support weaned), short-term survival (<30 days, CPS weaned), or death on CPS.

RESULTS

One hundred fifty patients (age, 57 +/- 17 years) were urgently started on CPS for cardiac arrest (n = 127; witnessed, n = 124; unwitnessed, n = 3) and refractory shock (n = 23). Sixty-nine patients were weaned from CPS, and 81 could not be weaned. Overall, 39 (26.0%) patients achieved LTS with a subsequent Kaplan-Meier median survival of 9.5 years. Duration of CPS was 32 +/- 38 hours for LTS and 21 +/- 38 hours for non-LTS. LTS occurred in 29 (23.4%) of 124 patients started on CPS for witnessed cardiac arrest and 11 (47.8%) of 23 for refractory shock (P < .05). Among patients with CPS initiated in the cardiac catheterization laboratory, LTS was seen in 24 (50.0%) of 48 versus 15 (14.7%) of 102 in patients with CPS initiated in other locations (P < .001). Cardiopulmonary resuscitation times greater than or equal to 30 minutes were associated with lower LTS (P < .05). The most common cause of death during CPS was refractory cardiac dysfunction (39.5%), and the most common cause associated with short-term survival was neurologic/pulmonary dysfunction (53.6%). Seven patients were bridged to a left ventricular assist device, and 1 subsequently underwent heart transplantation. Multivariate analysis revealed only cardiac catheterization laboratory site of initiation as a significant independent predictor of LTS (P < .01). When dividing the 20-year experience in tertiles, recent recipients have had more common prearrest insertion. Rates of long-term survival have not changed.

CONCLUSION

Of patients started on CPS, 46% were weaned, and 26.0% were long-time survivors. Rapid initiation of CPS permits LTS for some inpatients with cardiovascular collapse when initial advanced resuscitation fails. Strategies to improve end-organ function associated with use of CPS should lead to greater LTS. This practical application of inexpensive available technology should be more widely used.