Background: Patients with end-stage heart failure placed on a cardiac assist device show at least some degree of improvement of cardiac function. In a subgroup of selected patients, some hearts recovered considerable function. In these patients the device was removed and cardiac transplantation was no longer necessary. We report our long-term experience with these weaned patients.
Methods: As of today, 512 cardiac assist devices of various types (Berlin Heart, Berlin, Germany; Novacor, World Heart, Ottawa, Ontario, Canada; TCI, ThermoCardio Systems, Inc, Woburn, MA; DeBakey, Micromed Technology Inc, Houston, TX) were implanted in patients with end-stage heart failure in our institution. Of these, 95 patients belonged to a subgroup of patients with nonischemic, idiopathic, dilated cardiomyopathy who were implanted with a left ventricular support system (Novacor 84, TCI 10, Berlin Heart 1) between 1994 and 2000. All were routinely examined by echocardiography for improvement of cardiac function. The left ventricular diameter in diastole (LVIDd) and left ventricular ejection fraction (LVEF) served as the main parameters to assess changes in cardiac performance. Under the conditions of a running device, an LVIDd below 60 mm and an LVEF above 40% were the criteria to do further echocardiographic studies when the pump was turned off for up to 20 minutes.
Results: Twenty-eight patients (26 men, 2 women; ages 18 to 64 yrs; history of heart failure, 1 to 17 yrs) fulfilled the criteria of improved cardiac performance and were weaned from the device. Since then, 16 patients have continued "normal" heart function with follow-up times ranging from 1 month to 5.5 years (group B). Three patients died of noncardiac causes (group C). Eight patients were transplanted from 1 to 17 months later and one died on the waiting list (group A). Statistically significant differences between groups A and B were calculated for the duration of heart failure (9 versus 2 years, p = 0.0002). Differences in LVIDd before removal of the device (57 versus 51 mm, p = 0.0420), LVEF after 2 months of unloading (30 versus 49%, p = 0.0300), and LVEF preexplantation (43 versus 52%, p = 0.0001) were significant. Overall, 17% of the cohort of 95 patients were weaned successfully.
Conclusions: Weaning from cardiac assist devices is feasible for selected patients; it saves donor hearts and is preferred to cardiac transplantation. However, as of today no reliable parameter predicts outcome after weaning and none determines the possibility of device removal before implantation in advance.