Early and mid-term results of minimally invasive direct coronary artery bypass (MIDCAB) surgery

INTRODUCTION

Coronary artery bypass grafting (CABG) has delivered excellent outcomes in terms of morbidity and long-term mortality, establishing a benchmark for medical therapy and percutaneous coronary intervention (PCI)^[1-3]. Despite the advancements in drug-eluting stents (DES) and the development of P2Y12 inhibitors, the significant long-term survival benefits associated with grafting the left internal thoracic artery (LITA) to the left anterior descending artery (LAD) remain unchallenged^[4,5]. Less invasive CABG techniques emerged in the 1990s and have nowadays been proven as safe and effective therapeutic options^[1]. In particular, the minimally invasive direct coronary artery bypass (MIDCAB), which involves a sternal-sparing, off-pump procedure via a small thoracotomy, typically for LAD revascularization, is a uniquely appealing strategy by reason of its reduced invasiveness and the reported low in-hospital morbidity and mortality, and good long-term survival^[6-9]. Notwithstanding, MIDCAB remains underutilized, accounting for only 0.5% of cases in 2022, according to the Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACDS)^[10]. Despite consistent evidence supporting MIDCAB, its adoption has been slow, possibly due to the class 1A recommendations for both CABG and PCI in treating isolated LAD disease in current guidelines^[11], and the limited understanding of MIDCAB's reproducibility, which is still primarily documented in specialized centers^[6-9].

The aim of this study was to evaluate the early and medium-term outcomes of MIDCAB for isolated LAD disease, providing insights into its safety, reproducibility, and potential benefits. By addressing this gap in knowledge, our findings may help guide clinical decision making and encourage broader adoption of this less invasive approach. We also evaluated the incidence of events at follow-up based on perioperative myocardial damage.

METHODS

This study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki. Approval for the study was granted by the IRCCS Istituto Tumori Giovanni Paolo II - Bari, Review Board (1431/CEL - 03/11/2023). Preoperative, intraoperative, and postoperative data were collected anonymously and recorded in our prospectively maintained institutional database, which is utilized for internal quality control purposes. As such, the requirement for informed consent was waived. Follow-up was performed through direct telephone interviews conducted by adequately trained research personnel.

RESULTS

Follow-up

The median follow-up duration was 36 months [12-48], with a 100% completion rate. The primary endpoint was reached by 21 patients (11.4%). Follow-up mortality was reported in 14 patients (7.6%), with cardiac-related mortality accounting for 7 cases (3.6%). Seven patients (3.8%) required rehospitalization for percutaneous revascularization, and in 3 of these patients, a malfunction of the LITA-LAD bypass was identified. No patient required surgical revascularization during follow-up.

The Kaplan-Meier analysis revealed a 5-year revascularization-free survival rate of 84.2% [77.4-91.7] [Figure 1] and an overall 5-year survival rate of 90.5% [85.1-96.2] [Figure 2]. The 5-year survival free from Cardiac Death and LITA-LAD graft failure was 94.0% [89.7-98.6] [Figure 3]. In the multivariable model, preoperative NYHA Class III status (HR 5.898 [1.187-29.311], P = 0.030) and postoperative atrial fibrillation (HR 3.405 [1.232-9.415], P = 0.018) were identified as significant negative predictors of revascularization-free survival during follow-up [Table 3].

Figure 1. Kaplan-Meier analysis of mid-term revascularization-free survival.

Figure 2. Kaplan-Meier analysis of mid-term mortality.

Figure 3. Kaplan-Meier analysis of mid-term cardiac-related mortality and/or LITA-LAD graft failure.

Table 3

Multivariate cox proportional hazard analyses for the primary endpoint (revascularization-free survival)

Variable	Estimate	Hazard ratio [95%CI]	P-value	Estimate	Hazard ratio [95%CI]	P-value
	Initial model			Stepwise elimination model
NYHA class II	0.584	1.794 [0.605-5.320]	0.292	0.672	1.972 [0.690-5.635]	0.205
NYHA class III	1.506	4.510 [0.787-25.837]	0.091	1.775	5.898 [1.187-29.311]	0.030
Preoperative hematocrit (%)	-1.030	0.357 [0.025-5.165]	0.450
cTnI (ng/L)	0.291	1.338 [0.012-144.153]	0.903
Postoperative AF	1.177	3.244 [1.153-9.126]	0.026	1.225	3.405 [1.232-9.415]	0.018

AF: Atrial fibrillation; cTnI: cardiac troponin I.

We also examined the optimal cutoff values for peak cTnI levels as a potential predictor of the primary endpoint. ROC curve analysis identified the best cutoff value according to Youden’s index at 0.87 ng/L, with a sensitivity of 55% and specificity of 62% [Figure 4]. Comparing groups based on this cutoff, 11 patients (55.0%) in the primary endpoint group and 51 patients (37.8%) in the non-primary endpoint group had cTnI peaks exceeding this threshold (P = 0.221).

Figure 4. Receiver operating characteristic curve (ROC) for the Primary Endpoint based on the Cardiac Troponin I Cutoff as per Youden J Statistic (0.87 ng/L).

DISCUSSION

According to the figures provided by the STS ACDS, CABG has been consistently accounting for over half of all cardiac operations performed per annum^[10]. Less-invasive alternatives to CABG via the canonical full sternotomy have been available for decades^[12,13], yet their broader adoption has been limited, particularly compared to the more widespread acceptance of less-invasive valve surgeries, and these procedures remain predominantly within the domain of specialized centers.

This study presents the mid-term results of MIDCAB surgery conducted via a left mini-thoracotomy. Our series highlights favorable early postoperative outcomes, including no mortality within 30 days, a low incidence of complications, and a reduced ICU stay, in agreement with the outcomes reported in other series in the literature^[6-9]. This fosters the notion of safety of the MIDCAB procedure, and provides evidence of its reproducibility.

In addition, we report a 2% occurrence of postoperative LITA graft defects, detected using coronary CT scans in each of the first 100 cases and concomitant angiography assessment during hybrid revascularization cases. This finding is in line with observations from other large series, such as those from Brescia, Leipzig, and Ottawa^[6,7,9]. However, despite the already limited reports on MIDCAB, even fewer studies have systematically evaluated graft patency using angiography or CT scan^[14,15]. This makes comparisons between existing case series challenging, highlighting the need for prospective, multicenter studies with standardized imaging controls to provide more robust data.

Other complications were also infrequent in our cohort. Notably, we observed no cerebrovascular events, a significant finding compared to the STS ACSD report, which documented a permanent stroke rate of 1.5% among over 150,000 CABG cases performed in 2021^[3]. The absence of aortic manipulation for cross-clamping, arterial cannulation, and proximal anastomoses likely reduces the risk of cerebrovascular complications in MIDCAB. Additionally, the postoperative atrial fibrillation rate in our cohort was 14.2%, substantially lower than the 26% benchmark reported in the STS registry^[3]. This lower incidence may be attributed to the reduced surgical trauma of the sternal-sparing approach and the avoidance of CPB, which likely diminishes the inflammatory response and subsequently the risk of postoperative atrial fibrillation.

However, we report three cases of thoracotomy wound dehiscence, resulting in an incidence of 1.9% in our series of 155 cases, higher than the 0.8% rate of deep sternal wound infection or mediastinitis reported by the STS ACSD^[3]. These cases involved female patients with inframammary incisions. A recent randomized study comparing periareolar and inframammary approaches in minimally invasive cardiac surgery found that surgical site infections were more common in the inframammary incision group compared to the periareolar approach^[16].

This study is not the first to document the favorable outcomes associated with MIDCAB surgery. Recent case series from Brescia, Leipzig, New York, and Ottawa have reported promising results. Davierwala et al. presented data from the largest MIDCAB series globally, covering a 22-year period from 1996 to 2018, with an in-hospital mortality rate of 0.9% among 2,667 patients^[7]. The estimated overall survival rates at 10, 15, and 20 years were 77.7% ± 0.9%, 66.1% ± 1.2%, and 55.6% ± 1.6%, respectively^[7]. Repossini et al. shared findings from a cohort of 1,060 patients who underwent MIDCAB between 1997 and 2016^[6]. The early mortality rate was similarly low at 0.8%, with survival estimates at 5, 10, and 15 years being 87.1% [81.0-92.5], 84.3% [77.1-91.4], and 79.8% [72.2-87.3], respectively. Patel et al. reported a crude 9-year survival of 96.2% in a retrospective series of 158 MIDCAB patients compared with 158 propensity-matched patients who underwent PCI using a second-generation DES^[9]. Interestingly, MIDCAB patients were less likely to require repeat LAD reintervention (13.4% vs. 2.5%, P < 0.0001). More recently, Guo et al. reported on 283 patients prospectively followed up after undergoing MIDCAB at the University of Ottawa Heart Institute over a 17-year period, alongside another 283 patients who underwent minimally invasive multivessel coronary bypass grafting (MICS CABG)^[8]. Thirty-day mortality was observed in 1 patient, accounting for an overall mortality rate of 0.2%. The 12-year survival rate was 82.2% ± 2.6%, with no significant differences observed among patients receiving one, two, three, or more grafts^[9].

MIDCAB has demonstrated complication rates and long-term patency comparable to conventional sternotomy in multiple studies. A propensity score-adjusted comparison between MIDCAB and sternotomy for isolated LAD revascularization revealed no significant differences in operative mortality and complications^[17]. Furthermore, there was no observed improvement in late survival or a reduction in the risk of repeat revascularization^[17]. Similar findings were reported by Stanislawski et al. in a propensity score matching analysis involving 193 patients who underwent either MIDCAB or off-pump coronary surgery via conventional sternotomy for isolated LAD disease^[18]. Notably, all these studies indicated a trend toward improved short-term outcomes with MIDCAB. Recently, the Ruhr-University Bochum in Germany published their experience with 388 elective patients, of whom 229 underwent MIDCAB and 159 underwent off-pump LIMA-to-LAD grafting via full sternotomy^[19]. After applying inverse probability of treatment weighting, long-term outcomes did not differ significantly between the groups, with survival rates at 1, 5, and 10 years being 98.4%, 87.8%, and 71.7% for the sternotomy group, and 98.4%, 87.7%, and 68.7% for the MIDCAB cohort, resulting in a non-statistically significant relative risk (RR) of 1.24 [0.87-1.86]. Similar outcomes were observed for other endpoints, such as freedom from stroke, with an RR of 0.52 [0.25-1.09] (P = 0.06), and repeat revascularization, with an RR of 0.73 [0.47-1.16] (P = 0.22)^[19].

MIDCAB has also been compared to percutaneous coronary intervention (PCI) for the treatment of isolated LAD disease, a particularly relevant comparison given the class 1A recommendations for both CABG and PCI in current guidelines. The Leipzig group randomized 130 patients with isolated LAD disease to receive either MIDCAB or sirolimus-eluting stents^[20]. After a median follow-up of 7.3 years [5.7-8.3], with 99.3% completeness, there were no significant differences between the groups in the primary composite endpoint of death, myocardial infarction, and target vessel revascularization (22% PCI vs. 12% MIDCAB; P = 0.17), or in the individual endpoints of death (14% vs. 17%; P = 0.81) and myocardial infarction (6% vs. 9%; P = 0.74). However, target vessel revascularization was significantly more frequent in the PCI group (20% vs. 1.5%; P < 0.001)^[20]. More recently, a meta-analysis including 7,710 patients who underwent either MIDCAB surgery or PCI with drug-eluting stents (DES) confirmed these findings. The analysis showed no significant difference in mortality rates (pooled OR 0.92 [0.65-1.32], P = 0.66), but a significantly higher rate of target vessel revascularization in the DES group (pooled OR 0.27 [0.16-0.45], P < 0.0001) compared to MIDCAB^[21]. In this context, the results from Manuel et al., who followed 271 MIDCAB patients for 20 years, demonstrated that patient survival returned to levels comparable to their age/gender/year-matched counterparts in the general population, strongly suggesting that MIDCAB should be considered as an alternative to PCI when advising patients with isolated LAD disease^[22].

This study offers valuable insights into MIDCAB outcomes, but several limitations must be acknowledged. Although data were prospectively collected, the retrospective follow-up and single-center design limit the generalizability of the findings. The sample size is relatively small, reducing the study's statistical power and potentially affecting the robustness of the multivariable analyses. Additionally, the median follow-up of 36 months, while adequate for mid-term outcomes, may not fully capture long-term complications or graft failures. Moreover, hard clinical outcomes, such as spontaneous myocardial infarction and stroke, were not systematically assessed during follow-up. Furthermore, there was no formal adjudication of clinical events, which may impact the accuracy of reported outcomes. Information on secondary prevention medications, such as antiplatelet agents and statins, which are associated with long-term prognosis, was also not available, limiting the ability to make comprehensive comparisons with other studies. The evolution in surgical techniques during the study period, from direct vision to video-assisted thoracoscopy for LITA harvesting, introduces variability that could influence outcomes. Additionally, the learning curve associated with the MIDCAB technique may have impacted some results. Furthermore, the study did not assess quality of life, focusing solely on clinical and survival outcomes. Finally, graft patency via coronary CT or angiography was systematically evaluated in only a subset of patients, limiting the ability to generalize these findings.

In conclusion, our study demonstrates that MIDCAB for the treatment of isolated LAD disease is safe and yields satisfactory postoperative outcomes with excellent mid-term results. Our findings align with existing literature, underscoring the reproducibility of results achieved by other institutions. Non-sternotomy approaches represent both the present and future of coronary surgery, offering significant benefits to patients and advancing our specialty. However, it is crucial to acknowledge the technical challenges and steep learning curve associated with these procedures, which require substantial experience and confidence, particularly when addressing complex multivessel disease or multifactorial pathology^[23,24]. Future studies with larger, multicenter, and randomized designs are necessary to confirm these results, validate the safety and effectiveness of MIDCAB across diverse patient populations, and provide a more robust comparison with other revascularization strategies.