RESULTS
During the study period, 83 patients underwent LTHS dLVAD placement. Of these, 50 did not require tMCs, while 22 (26%) were bridged with IABP and 11 (13%) with ECMO. Preoperative data are summarized in Table 1.
Baseline demographics were similar among groups. Indication for dLVAD was primarily bridge to candidacy. Two ECMO patients (18%) and none in the other two groups required mechanical ventilation (p=0.001). All ECMO patients were INTERMACS 1, while IABP recipients were predominantly INTERMACS 2 (45%), and non-tMCS patients INTERMACS 3 (56%) (p<0.001). ECMO patients were hospitalized for median 14 days prior to dLVAD, versus 10 for IABP and 2.5 for no tMCS (p=0.002). Median ECMO duration was longer than for IABP (10 vs. 6 days, p=0.02).
Prior to dLVAD, ECMO patients had lower albumin and hemoglobin and higher AST, while IABP patients had higher creatinine. Echocardiographic variables were similar among the groups. Both tMCS groups displayed worse baseline hemodynamics (higher PCWP, CVP, and PAD; lower PA pulsatility index [PAPi]) compared to non-tMCS patients.
Table 2 compares pre- and post-tMCS end-organ function and net fluid balance during tMCS between IABP and ECMO recipients. Pre-tMCS ECMO patient values tended to be higher for all variables except BUN. Post-tMCS, IABP patients had significant decreases in BUN, creatinine, bilirubin, and MELD-XI, while ECMO patients had significantly decreased creatinine and MELD-XI. Decrease in creatinine was significantly greater for ECMO patients. Net fluid balance for the duration of tMCS and net fluid balance per day were not significantly different.
Operative data are presented in Table 3. Operative and cardiopulmonary bypass times did not differ among the groups. In addition to ECMO removal, 7 (64%) ECMO patients underwent concomitant procedures. These included: percutaneous septostomy closure in three; and septostomy closure with femoral artery repair, VT ablation, femoral thrombectomy, and veno-venous (VV) ECMO in one each. Fewer IABP (1 [5%]) and non-tMCS (4 [8%]) patients had concomitant procedures (p<0.001). In these patients, procedures were: VT ablation; splenectomy (for infarct) and femoral artery repair; surgical PFO closure with RVAD/VV ECMO; aortic valve plication; and left atrial appendage closure with VV ECMO.
One IABP and one non-tMCS patient required conversion to sternotomy, the former for coronary graft injury and the latter for hypoxemia and ventricular arrhythmias in the setting of an open PFO. That patient was one of two who required an RVAD. The other was an ECMO patient with intractable inotrope-dependent RV failure who underwent late RVAD insertion. ECMO patients more frequently required ≥4 units intraoperative red blood cells (64%, vs. 23% IABP and 4% no tMCS, p<0.001). No ECMO patient underwent transfusion-free implant. Despite this, no ECMO patient required reoperation for bleeding or delayed chest closure.
Post-dLVAD inotropic support duration did not differ among groups (median 7-10 days). Median ICU length of stay was slightly higher for ECMO (9 days vs. 6 in the other groups, p=0.08). Median hospital length of stay ranged from 16 days for no tMCS to 19 days for ECMO (p=0.21). Two operative deaths (18%) occurred in the ECMO group, one from multiorgan failure and one from a hemorrhagic CVA. There were no operative deaths in the IABP group and 2 (4%) in the no tMCS group (p=0.06).
Figure 1 compares rates (in events/100 patient-months [EPPM]) of postoperative adverse events. In the early phase (<3 months), major and GI bleeds were most common. Rates were similar in ECMO and non-tMCS groups. Device-related infections and CVAs were rare, and no ECMO patient suffered an early infection or ischemic CVA. One early hemorrhagic CVA occurred in an ECMO patient.
There were also no significant differences in rates of late (>3 months post-dLVAD) complications. In this phase, device-related infections were predominant. These occurred at a higher rate in IABP patients, although the trend is not significant (p=0.14). No late hemorrhagic CVAs occurred in ECMO or IABP patients, versus 0.93 EPPM in non-tMCS patients. Likewise, no ECMO patient had an ischemic CVA, compared to 0.62-0.71 EPPM in the other groups.
Figure 2 illustrates Kaplan-Meier survival and Figure 3 competing outcomes for the three groups. Overall survival curves are not different among groups (p=0.45). For ECMO patients, no late mortality occurred after two operative deaths, yielding 81% 3- and 12-month survival. By 12 months, 41% were transplanted. In IABP recipients 3- and 12-month survival were 100% and 88%. By 12 months, 36% were transplanted. In the no-tMCS cohort, 3- and 12-month survival were 94% and 86%, and 32% were transplanted by 12 months.