Transit time flow measurement of coronary bypass grafts before and after protamine administration

Intra-operative assessment of graft patency is crucial to perform high-quality coronary artery bypass graft (CABG) procedures. Occlusion rates as high as 20.0 and 8.0% have been reported at 1-year follow-up for venous and arterial conduits, respectively [1, 2]. Graft failure within the first year of surgery may be in part attributed to technical errors [3, 4], which can lead to graft kinking, overstretching, lumen occlusion and anastomotic stenosis, thus resulting in incomplete revascularization [5].

Transit-Time Flow Measurement (TTFM) is an intra-operative tool to assess the patency and the quality of a graft in order to prevent errors from going unnoticed. TTFM relies on specific flow parameters, namely mean graft flow (MGF), pulsatility index (PI), diastolic filling (DF) and backflow (BF). Importantly, no single-parameter measurements have been demonstrated to predict the quality and long-term patency of a graft.

TTFM has been reported to have a 2–4% detection rate of intra-operative errors, and despite the recommendation for its use (class IIa) in the 2018 European guidelines for myocardial revascularization, its adoption has not yet become widespread [6, 7] with an estimate of a global average use of TTFM of ~ 30% but with much regional variance [Medistim data]. This can be attributed in part to a lack of clear cut-off values and the varying sensitivity and specificity of each parameter to predict graft patency.

Performing a retrospective review of the large, multicenter REQUEST (REgistry for QUality AssESsmenT with Ultrasound Imaging and TTFM in Cardiac Bypass Surgery) registry [8], we investigated the variation of the most commonly used TTFM parameters before and after the administration of protamine to identify new benchmark parameters for graft quality assessment, with the hope to increase both the practice and caliber of intraoperative quality control. No study has to this day attempted to investigate the change in TTFM parameters before and after the administration of protamine. Additionally, there are no guidelines as to whether TTFM measurements must be performed both before and after protamine administration or if one measurement will suffice.

Graft and patient selection as previously described are illustrated in Fig. 1. The baseline characteristics and intra- and post-operative in-hospital variables for the 368 patients with at least a single graft with pre- and post-protamine ACI above 30 and same MAP range pre- and post-protamine are reported in Table 1. Mean age was 67 with a majority (84.5%) of male patients. Slightly more than half of the procedures (56.8%) were performed on-pump. Adverse event rates was extremely low with a mortality of 0.5% and a stroke rate of 1.4%.

The pre- and post-protamine TTFM variables for the eligible grafts are reported in Table 2. MGF increased post-protamine for venous grafts from 30 to 33 mL/min but remained unchanged for arterial grafts. PI decreased for both venous (2.3 to 2.0) and arterial grafts (2.3 to 2.2).

There were statistically significant differences in pre- vs. post-protamine MAPs in the 768 grafts (with the MAP generally higher post-protamine. See Additional file 1 for detailed MAP table).

Secondary analysis, including only grafts with pre- and post-protamine ACI above 30% and MAP within the same range pre- and post-protamine, showed that MGF increased post-protamine for venous grafts from 30 to 32 mL/min (P <  0.001), but remained unchanged for arterial grafts (28 mL/min, P = 0.44). Concerning the single grafts with the same MAP range pre and post protamine, the 95th percentile for MGF exceeded 25 mL/min. No correlation was found between the outliers (i.e., grafts with a reduction of more than 25 mL/min in flow) and graft or anastomosis revision (as defined in the original REQUEST trial, Table 1 [8]).

PI decreased in venous (2.4 to 2.1, P <  0.001), and was numerically unchanged, but statistically significant for arterial grafts (2.2, p = 0.02). DF decreased post-protamine for arterial grafts (71% pre-protamine vs 70% pre-protamine, p <  0.001) but was unchanged for venous grafts (61% pre-protamine vs 60% post-protamine, p = 0.18). Finally, BF was unchanged for both arterial (0.9% vs 1.1%, P = 0.78) and venous grafts (0.4% vs. 0.5%, p = 0.96) (Table 3 and Fig. 2).

In this novel study comparing pre- and post-protamine measurements, we found that MGF was unchanged for arterial but slightly increased for venous grafts (30 mL/min vs. 32 mL/min, respectively, P <  0.001) after protamine administration, once MAP had been corrected for. The 95th percentile for MGF was 25 mL/min (i.e., in 95% of all single grafts after accounting for MAP, the MGF increased, stayed the same, or decreased in less than 25 mL/min). PI slightly decreased (especially for venous grafts. Of note, though the PI of the arterial grafts did not change, this finding was statistically significant since the distribution of the difference of pre- and post-protamine PI had a tendency towards negative values representing higher values pre-protamine compared to post-protamine. However, these differences were mostly symmetrically distributed around zero (see figure 1 in the Additional file 1). These differences were, in our opinion, not clinically significant.

TTFM can provide detailed information regarding graft flow and patency; however, proper handling technique is paramount. Graft patency and flow are assessed by four variables: MGF, PI, BF%, and DF% [9]. Each variable adds its own unique piece of information, and no single variable can be taken in isolation during decision making for graft revision.

Despite its clinical utility, TTFM has not yet achieved widespread acceptance. Part of this can be explained by the small size of the cohorts and studies used to research TTFM and their differing cutoff values for predicting graft failure. The ESC/EACTS guidelines include a recommendation for TTFM graft assessment (class of recommendation IIa, level of evidence B) [7] with cutoff values for MGF of > 20 mL/min and a PI < 5 being the most widely cited values [6, 10].

The diagnostic accuracy of TTFM is still a matter of debate with different values of sensitivity and specificity reported [6, 11]. Furthermore, the lack of universally accepted cut off values results in divergent values used in different studies.

Using a patency prediction model based on virtual machine learning, Mao et al. have recently demonstrated that the implementation of TTFM parameters with clinical and hemodynamic characteristics significantly increased its sensitivity and specificity values [12]. Though this has not yet been clinically applied, it appears that the addition of high frequency ultrasound (HFUS) increases the sensitivity and specificity to detect a problematic graft dramatically [13].

The recently published REQUEST trial [8] was the first prospective, multicenter trial to examine the influence of intraoperative graft assessment with TTFM and/or HFUS on decision-making in multivessel CABG. In the trial, changes in surgical strategy were performed in 25.2% of patients, with most of these based only on an abnormal TTFM or HFUS. These changes included a revision of 7.8% of completed grafts. Very low rates of peri-operative adverse events, particularly mortality, were observed. Of note, 37.5% of patients in the trial were above the age of 70, which is reflective of current real-world practice. Previous reports have shown that TTFM parameters in the elderly population is similar to that of the younger population [14].

Mean arterial pressure (MAP) has been known to effect MGF. Except at the extremes of MAP, MGF is directly proportional to MAP when coronary vascular resistance is unaltered. For instance, Balacumaraswami et al. [15], when reporting on TTFM parameters in on- vs. off-pump CABG, devised the flow/pressure ratio which was calculated as a ratio of MGF-to-MAP in order to account for the effects of MAP on graft flow. In order to account for the effect of MAP, after our first analysis of TTFM parameters (702 patients with a total of 1335 grafts) we conducted a secondary analysis of TTFM parameters only in the 368 patients with 567 grafts that had the same MAP reported before and after the administration of protamine.

As previously mentioned, despite TTFM’s potential to provide valuable data in coronary surgery (especially when combined with HFUS), it is underutilized. Using a retrospective analysis of the REQUEST trial, we aimed to add another parameter to TTFM and to improve TTFM’s sensitivity and specificity. Based on clinical observations of a change in TTFM parameters after the administration of protamine, we retrospectively analyzed the REQUEST trial to quantify this change in parameters. To reduce the effects of MAP on graft flow we compared the various TTFM parameters using only single grafts with similar MAP range before and after the administration of protamine. Sequential grafts were excluded to minimize confounding, since the effect of protamine on single grafts had not yet been quantified. We found no reduction in flow in the post-protamine measurements (there was even a small, albeit clinically insignificant, rise in flow in venous grafts and arterial grafts). There was also some improvement in the PI post–protamine (mainly in venous grafts but not in arterial grafts). DF decreased post-protamine for arterial grafts but was unchanged for venous grafts and BF was unchanged for both graft types. Of note, reports on flows in arterial vs. venous conduits usually report higher flows in venous conduits [16], but this is not always true when comparing the radial artery to venous grafts [17]. We did not differentiate between radial grafts and other arterial grafts which might have yielded slightly different results.

Protamine causes a well-known series of transient effects during reversal of heparin anticoagulation, including systemic hypotension, pulmonary hypertension, bradycardia and bronchoconstriction [18]. Belboul et al. have also investigated the effects of protamine on the epicardial microflow [19]. They demonstrated a dynamic effect: after transient vasodilation (resulting in an improvement of MGF), vasospasm was detected (thus making vascular resistance higher and decreasing mean flow locally). Protamine has been demonstrated to induce vasoconstriction by inducing leucocyte degranulation [20] and to alter the viscosity of blood by activating platelet adherence to intact endothelium of arterioles thereby inducing thrombus formation [21]. Of note, a recent publication from Korea did not show any association between blood viscosity and TTFM [22]. However, these (mostly) old data have neither been subsequently confirmed in vivo nor been linked to TTFM values after protamine administration in CABG.

We found no clinically significant difference in TTFM parameters before and after the administration of protamine. Furthermore, even when examining outliers (i.e., grafts with a reduction of MGF > 25 mL/min with protamine administration), we did not find a higher rate of graft revision. This finding might obviate the need for performing TTFM before and after protamine administration since a single TTFM measurement may suffice to achieve reliable data on each graft’s performance. Depending on the specific clinical situation and intraoperative changes, more measurements may be informative (i.e., with changes in MAP, volume status of the patient, and more).