Follow-up results
Long-term mortality was obtained from the National Demographic Census with a 100% completion of data. Overall, 13% of the patients died during the follow-up period. The causes of death were as follows: Cardiac 35%, Neurological 13%, Cancer 8%, Renal 7%, Respiratory 7%, Sepsis 6%, Vascular 1%, Gastrointestinal 1% and Other 7%. We were unable to identify the cause of death in 15% of the cases.
Overall one-year mortality was 2%. It was 2% in the SIMA group and 1% in the BIMA group (p = 0.22). Overall five-year mortality was 10% and in this case it was significantly higher in the SIMA group (16% vs. 5%, p < 0.001). (Table 5)
Age (p = 0.001, HR 1.1, 95% CI 1.1 – 1-2), COPD (p = 0.004, HR 2.0, 95% CI 1.2-3.2) and Logistic EuroScore (p = 0.002, HR 1.0, 95% CI 1.1- 1.2) were identified by the Cox regression as risk factors for long-term mortality.
Mean survival calculated by the Kaplan-Meier curve was 11.0 ± 0.2 years (95% CI 10.7 – 11.4) in the BIMA group and 10.0 ± 0.3 years (95% CI 9.5 – 10.6) in the SIMA group.
The Kaplan-Meier survival curves and number of patients at risk for each period are displayed in Figure 1 and annexed table. The survival curve for the non-matched population is displayed in Figure 2.
Since the two groups still differed in age, with the BIMA group having significantly younger patients, a Cox model was applied to account the age difference. Subjects who received BIMA still had better survival compared to those who received SIMA (HR 0.63 (0.4 – 0.99), p 0.047).
The Cox analysis was also applied to the LV function subgroups, as they remained significantly different after the propensity matching. BIMA conferred a survival benefit for patients with good LVEF (HR 0.31 (0.18 – 0.54), p = 0.0001) and fair LVEF (HR 0.54 (0.3 – 0.99), p = 0.048), but we did not demonstrate a significant survival benefit when using BIMA for patients with poor LVEF (HR 1.61 (0.27 – 9.59, p = 0.6).
The follow-up visits and/or telephone interview revealed an excellent functional class years after the operation. Only 8% of the patients reported residual angina being in CCS class II - III and 4% of the patients described dyspnea in NYHA Class III-IV, which was more likely attributed to concomitant aortic stenosis or advanced COPD. The incidence of myocardial and cerebrovascular events during the following years after the operation was 3%. There were no differences in the distribution of the complications or residual symptoms amongst the groups.
Thirty-two patients (6%) were lost to follow-up. (Table 6)
Thirty-nine patients (7%) underwent a repeated angiogram years after the operation. The most common reason was residual angina, although 20% of them were done as part of the preoperative workup due to severe aortic stenosis.
The angiogram confirmed patent grafts in 36% of the cases. Graft failure was reported in 33% cases (13 patients), most commonly due to the SV graft failure (62%, 8 cases) but arterial grafts were also affected (LIMA 15%, 2 cases; RIMA 15%, 2 cases; RA 8%, 1 case). Two patients showed combination of venous and arterial graft failure. In 8% of the cases (3 patients), the angiogram showed progression of the native disease in vessels not previously grafted, and in 15% of the cases (6 patients) we were not able to determine the result since the angiogram was not performed in our institution but in a local hospital.
Only 17 patients (5%) required further revascularization with PCI, and only 0.4% of them were required during the first year after the operation. (Table 6)