The benefits of arterial conduits in coronary artery bypass grafting surgery (CABG) are well known [1, 2, 3]. Utilizing the internal thoracic (mammary) arteries (ITAs) for this purpose is highly recommended and improves long-term survival [1]. Harvesting both ITAs is well tolerated and associated with minimal side effects among younger and less morbid patients [4]. However, Long-term benefits and survival advantage are still a matter of ongoing debate [5, 6, 7].

Balancing the pros and cons of utilizing the internal thoracic arteries in the elderly and in patients with comorbidities is more complicated. Theoretically, as the pre-operative life expectancy shortens, the survival advantage of utilizing ITA grafts diminishes. A survival benefit for bilateral ITA harvesting in the elderly has not been confirmed [8]. The side effects of this practice become more relevant and less favorable for these patients. Complications, e.g., wound healing disorders and sternal instability, are seen at higher rates, especially in the presence of additional risk factors, e.g., obesity, female gender and comorbidities such as diabetes mellitus, chronic pulmonary disease, renal insufficiency. Hence, hesitancy still exists regarding the use of ITAs in CABG surgery in elderly patients [9].

The primary objective of this study is to evaluate the safety of harvesting both ITAs in the elderly by comparing the short-term results in patients who are 75 years of age and older who received either single left (sITA) or double (dITA) internal thoracic artery grafts.

During the period of the study, 1855 CABG operations that meet the study criteria were found. In patients undergoing sITA (n = 1114), 614 underwent a combination of LITA & 1$\times$ saphenous vein graft (VSM), while 500 underwent LITA & 2$\times$VSM. In patients undergoing dITA (n = 741), 329 underwent LITA & RITA, and 412 underwent LITA & RITA & 1$\times$VSM.

Comparing the unmatched sITA and dITA groups (Table 1A), there were slight differences in the patient characteristics age, BMI, past myocardial infarction and in NYHA class. These differences may suggest that dITA was selected for less complicated CABG surgeries.

Five hundred and nineteen (519) pairs of LITA-dITA were matched. Details of the logistic regression model are available in Table 2. The paired groups were very similar in their characteristics (Table 1B).

Year of surgery was a significant predictor for performing double ITA, indicating higher tendency for performing double ITA in the more recent years in our department (Fig. 3). Mortality however was not affected by this tendency.

Fig. 3.

The change of the ratio of performed dITA over the ratio. There was a tendency to perform more dITA in the recent years. Mortality for both sITA and dITA was relatively steady, except for a slight elevation in early years. The “zero” mortality of dITA in early years is a result of a low total number (less than 20 dITA for before 2007).

Significant differences were seen in median skin-to-skin operative time (181 vs 205 minutes; p 0.0001), in ventilation time (10 vs 9 hours; p = 0.020) and in the incidence of wound healing disorders, which was significantly higher when utilizing dITA (3.3 vs 6.9%; p = 0.011). There was no significant difference in the incidence of sternal wound healing disorders when comparing the skeletonized against the pedicled technique (4.2 vs 6.2%, p = 0.135. Calculation for the total number in both the matched sITA and dITA groups). The rest of the outcome parameters were similar. Table 3 lists the detailed results (Tables 3,4).

Shortly after the introduction of the use of ITA as a graft in CABG procedures in the 60s [14], it became clear that this practice is superior to utilizing vein grafts, e.g., the saphenous vein [15]. The saphenous vein graft, which is still widely used in bypass surgery due to its availability, is prone to progressive intimal hypertrophy and occlusion, leading to impaired long-term graft patency, which usually presents after the fifth postoperative year, but sometimes much earlier [16, 17, 18]. Studies in the 1980s and 1990s led to widespread utilization of the ITAs in CABG surgery. Very few other arteries in the human body can be harvested as easily and with minimal potential collateral damage as the ITAs. Anastomosing the LITA to the LAD artery became a standard surgical approach [19]. The logical next step was to harvest the ITA bilaterally, assuming that superior patency would correlate with improved long-term results.

Many observational studies emerged and reported better results for dITA compared to LITA in different patient groups [20, 21, 22]. The general impression of the superiority of dITA was questioned with the publishing of the primary and final results of the only available randomized controlled study (RCT) in this regard; the Arterial Revascularization Trial (ART) [6, 20, 23]. The ART investigator had conducted the largest multicenter randomized trial of bilateral versus single ITA grafts and followed 3102 patients in seven countries for up to 20 years. Surprisingly, they found no advantage of utilizing dITA compared to sITA in regard to mortality or the rate of cardiovascular events at both 5 years and 10 years of follow-up (5-years mortality 8.7% in dITA and 8.4% in sITA). However, they found higher rates of sternal wound complications in the dITA group (3.5% vs 1.9% in sITA). Multiple limitations may possibly have influenced the results, masking better results for dITA [7]. The discrepancy in the results between these studies may be partially attributed to the heterogeneity of techniques of anastomosing and reconstruction, ranging from in situ dITA to complex Y-anastomoses (Fig. 2). In addition, the classical LITA-LAD bypass is often protected by high blood flow, which may play a role in the better LITA results.

A long-term survival advantage in elderly is hard to prove as the life-expectancy shortens. Few groups have investigated this, and there has been no conclusive consensus on the matter. Navia et al. [24] conducted a retrospective analysis of 243 matched sITA-dITA pairs of patients $>$70 years of age and found survival advantage for dITA after 10 years of follow-up (66% vs 53.0%, p = 0.022). Medalion et al. [25] also found a survival advantage for dITA, but their study was unmatched and their compared group was not homogeneous. In contrast, both Mohammadi et al. [26] and Benedetto et al. [27] found no survival advantage in patients older than 65 years. Other studies included only small numbers of patients and have not been conclusive. In conclusion, the potential cut-off age for possible survival benefit of dITA remains to be determined. In this context, one should note the variability in the age cutoff defining the elderly groups in different observational studies. We chose a 75 year cutoff based on national demographic trend, as reported by the German Association for Thorax, Heart and Vascular surgery (DGTHG) [28], and based on our own statistics, that showed almost 30% of CABG was conducted on patients 75 years and older (Fig. 1).

Our study confirms the safety of harvesting both ITAs in the elderly without increasing the perioperative mortality but indicates a significant increase in the operation time and the incidence of sternal wound healing disorders. This suggests that the practice can be justified only if a significant long-term advantage is to be expected, which as discussed before, remains uncertain. Wound healing disorders have a severe negative effect on patients’ quality of life and can even impact short-term survival rate beyond the first 30 days. Nevertheless, sternal wound healing disorders can be avoided by careful patient selection, obesity, female gender and diabetes mellitus adversely influence the outcome. In addition, harvesting the ITA in a skeletonized way significantly reduces wound healing disorders [29, 30]. However, our study failed to establish such a correlation. Also, our study failed to find a statistically significant difference in the rate of sternal complications (0.6% vs 2.1%, p = 0.06). A bigger sample size is most-likely necessary to prove this as the total number of sternal complications (14 for both matched groups) was too low to avoid type II error.

Reviewing all the pros and cons for harvesting both ITAs in elderly patients that were reported by our study and others, we believe that harvesting both ITAs is safe and can be done in selected patients despite their age, whose overall life expectancy are good and whose risk factors for wound healing disorders are mild and when the viability of the saphenous vein is questionable.

Finally, study limitations include: (I) This study is prone to the usual limitation of a retrospective analysis, such as the quality of the data and their heterogeneity. Operative urgency for example, lacked a clear definition and was surgeon dependent, hence, it was not included in the propensity score matching calculation. (II) Considering the time-frame of almost 20 years, changes in surgical practice and patient care could have affected the results. More dITA and skeletonized harvesting were performed in the recent years (Fig. 3). The time factor can influence other variables in an unpredictable way. However, matching should minimize such effects. (III) Being a single-center study, there is risk of outcome data being affected by local experience in CABG practice; and (IV) data on exact classifications of wound healing disorders were not available to us, which is why we only considered wounds that have been treated surgically. (V) The decision for sITA or dITA was not standardized and hence remains subjective and dependent on the surgeon himself. Fig. 4 shows a clear trend for some surgeons to favor LITA over sITA. However, this is partially also due to the time effect as early surgeons favored sITA (Fig. 3). (VI) Comparing the unmatched sITA and dITA groups (Table 1), there were slight differences in the patient characteristics age, BMI, past myocardial infarction and in NYHA class. These differences may suggest that dITA was selected for less complicated CABG. However, propensity score matching should have compensated for this effect. (VII) Lastly, there are some limitations to the statistical methods that were used. Variables such as ventilation time and in-hospital stay were strongly skewed. While the median ventilation time was shorter in the double ITA group (9 h vs 10 h), the average was actually higher (24.8 vs 20.1 h). This skewness resulted from outliers from patients who had required significantly longer ventilation times (exceeding 2 months in some cases) and resulted in a positive Wilcoxon Signed Rank test (p = 0.02). The most significant prediction for longer ventilation time was re-thoracotomy (p = 0.004) and it was seen more often in the double ITA sample (6.2% vs 4.6%).

Fig. 4.

The ratio between sITA and dITA for the different surgeons in the study. The decision for sITA vs dITA was not standardized and hence individual differences were noticed between the surgeons.

In conclusion, harvesting both ITA in elderly patients is safe and feasible. However, it is associated with increase in the incidence of sternal wound healing disorders. When done, attention is required to control other factors influencing the wound healing. Long-term benefit of dITA grafting still needs to be proven.

ART, Arterial Revascularization Trail; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; EF, ejection fraction; ITA, internal thoracic artery; LAD, left descending coronary artery; NYHA, New York Heart Association; VSM, vena saphena magna.

The data related to this study are available on request from the corresponding author, RC.

Idea and design—RC; Data maintaining and analyzing—RC, AG; Statistical analysis—RC, AG; Manuscript writing—RC, AG; Critical review and supervising—HT. All authors read and approved the final manuscript.

This study was conducted in accordance with the relevant guidelines for good clinical and good scientific practice defined by the 1964 Helsinki declaration and with the approval of the university hospital of Mainz institutional board. According to the local regulation in Rheinland-Pfalz, reporting retrospective studies that analyze patient data anonymously to the official local ethic committee is not required. No Approval Numbers are given for this retrospective Study. Patient consent for anonymous use of their data in this retrospective study was obtained indirectly as a part of the hospital admission process.

The authors thank David K.C. Cooper MD, PhD, FRCS for reading the manuscript and providing his critical comments.

The authors declare that they did not receive any external funding for this work. Funding and material for this study was made available by the University Hospital of Johannes Gutenberg University Mainz. Ryan Chaban received funding from the German research foundation “Deutsche Forschungsgemeinschaft (DFG) - Project number 468485612” at the time of writing the manuscript.

The authors declare no conflict of interest.