Primary thrombolysis for free flap surgery in head and neck reconstruction: a case report and review

Indication

The indications for flap thrombolysis vary in the literature (Table 1). Panchapakesan et al. [7] proposed that the criterion for using thrombolytic therapy was failure to re-establish satisfactory venous outflow despite restoring good arterial inflow. This is logical as it suggests the presence of significant intra-flap thrombosis obstructing venous outflow [1], and is adopted by other authors including Bui et al. [1], Rinker et al. [5], and Bonde et al. [8]. Other authors recognize late-explored flaps as an indication. Yii et al. [9] suggested that whereas thrombosis that is recognized early is often treated successfully with anastomotic revision alone, cases recognized or re-explored late (with a longer warm ischemic time) tend to have increased thrombus formation and “no-reflow phenomenon,” requiring additional intervention such as thrombolysis. Similarly, Bui et al. [1] also recognized late thrombosis, as evidenced by extensive, established clots (usually white), as an indication. Venous thrombosis (as opposed to arterial) has also been identified as an indication, as proposed by Serletti et al. [10] and Casey et al. [3]. Serletti et al. suggest that this might be because venous thrombosis tends to be detected later, resulting in microcirculatory changes and a poorer response to anastomotic revision alone. Based on the same rationale, Chang et al. [11] routinely perform thrombolysis for all cases of venous thrombosis; in addition, they also perform thrombolysis for isolated arterial thrombosis if there is visualized distal clot propagation from the perianastomotic area into the flap.

In our case, pedicle thrombosis was limited to the artery, and the cause was likely a technical issue resulting in thrombus formation at the anastomosis. This was dislodged distally by attempts at milking, into the branches of the flap artery (muscle branches and perforators), making the thrombi difficult to retrieve by Fogarty thrombectomy. Fogarty thrombectomy is indicated in extensive pedicle thrombosis which cannot be removed through simple extraction at the anastomotic site [12], but without propagation into the perforators of the flap, as thrombosis at this level is too distal for mechanical retrieval. Distal clot propagation (as described by Chang et al.) was the main indication for thrombolytic therapy [11]. Without thrombolysis, the revision of the arterial anastomosis alone would unlikely be successful as the clots would continue to obstruct vascular inflow, or propagate further distally and result in intra-flap thrombosis.

Hence, we suggest that the main indications for thrombolysis are (1) arterial thrombosis with evidence of distal clot propagation (as described by Chang et al.), (2) venous thrombosis with failure to restore venous outflow despite thrombectomy and reestablishment of arterial inflow (as this suggests the presence of intra-flap thrombosis), and (3) late thrombosis (as evidenced by extensive clots, usually white.) This is illustrated in our proposed algorithm (Fig. 7), which may be applied to both revision cases as well as primary surgery. This case also demonstrates that milking an arterial anastomosis (empty-refill test) to check for its patency should be avoided; instead, the arterial anastomosis may be partially opened and the clot removed by gentle manual extraction or Fogarty catheter if it is more distal [4].

Route of administration

Intra-arterial injection of thrombolytic agent remains the most common route of administration. This may be performed through a small arteriotomy in the main pedicle, through a side branch of the artery, or at the site of original arterial anastomosis [1,7]. Either way, the recipient artery (proximal to the injection site) and the draining vein(s) are clamped prior to injection, to allow the thrombolytic agent to remain in the flap circulation [1]. The agent is then allowed to drain through an open flap vein, to prevent its flow into the systemic circulation. In cases of end-to-side arterial anastomosis, the recipient artery distal to the anastomosis should also be clamped prior to injection to prevent systemic absorption [10]. In our case, injection was performed at the site of the original arterial anastomosis as this had already been taken down and provided an open lumen for easy access, avoiding the need for additional arteriotomy or side branch dissection.

The temporal order of thrombolysis versus revision of anastomosis also deserves discussion. In our case, injection was performed prior to revision of anastomosis; this is similar to the protocol suggested by Chang et al. [11] for arterial thrombosis with distal clot propagation. The presumable advantage is that thrombolysis begins earlier, giving the thrombus less time to expand or propagate. The reverse order has also been used, where the arterial anastomosis is first revised prior to thrombolysis [4,5,8]. This allows for easier assessment of flap perfusion following thrombolysis. Nonetheless, it is still possible to assess for response to treatment prior to re-establishing arterial inflow, by gauging the resistance to injection caused by the presence of thrombi. Serletti et al. [10] described their observation that a moderate amount of manual pressure was required on beginning the infusion of thrombolytic agent, but this gradually lessened and reached normal pressures as flap and venous blood color improved. A similar gradual reduction of injection pressures was also noted in our case, reflecting the successful removal of obstructing thrombi within the flap.

Dose

There remains no consensus as to the optimal dose of rTPA for flap thrombolysis, with varying practices ranging from as low as 2 mg to as high as 50 mg (Table 1). We adopt an initial dose of 10 mg for multiple reasons: (1) this dose is similar to that used by Bui et al. (5–20 mg) and Yii et al. (15 mg), which are two of the largest cohorts studied to date (Table 1); (2) we believe that using a higher dose would be more effective in achieving successful thrombolysis than repeated lower doses of 2–2.5 mg; and (3) a dose of 10 mg is an easy number to remember in a stressful situation of flap compromise. In our case, as resistance to injection of heparinized saline remained high after the first dose, we repeated a second dose of 10 mg, which resulted in reduction of resistance to injection.

The risk of systemic absorption and bleeding is also very low at such a dose. For comparison, the usual total dosage of rTPA to treat myocardial infarction is 100 mg given intravenously; a dose of 150 mg is associated with increased risk of intracranial bleeding [4]. In flap thrombolysis, most of the rTPA is flushed out through the open flap vein prior to venous anastomosis, and only a small proportion will end up in the systemic circulation. In addition, the half-life of rTPA is only 5 minutes, which means that there is only a very short time window in which there is an increased bleeding risk. While our patient did develop mild periorbital bruising on postoperative day 1, this was likely due to the administration of intravenous heparin rather than intra-flap rTPA.

Efficacy

Evidence supporting the use of flap thrombolysis in improving salvage rates remains limited (Table 1). Casey et al. [3] and Serletti et al. [10] reported excellent salvage rates of 100%, though their sample size was small (11 and 5 patients, respectively), limited to use in venous thrombosis only, and there was no control group (without thrombolysis) for comparison. Rinker et al. [5] reported higher salvage rates in flaps receiving thrombolysis compared to those without (67% vs. 29%, respectively), although this was not statistically significant. Other retrospective reviews also failed to demonstrate any statistically significant difference in salvage rates, with some reporting an even lower salvage rate with thrombolysis [7]. This could be explained by the fact that flaps requiring thrombolysis tend to be explored later and have more extensive thrombosis [7]. Further studies adjusting for time to exploration would be useful for more accurate analysis of efficacy. In our case, the flap was successfully “salvaged” with thrombolysis. Given the presence of multiple irretrievable thrombi within the flap, revision of arterial anastomosis alone would have been unlikely to achieve the same outcome, as it would only push the clots further distally and compromise flap perfusion. Successful salvage in our case was also likely contributed by the fact that thrombolysis was performed early, and so it is easier to dissolve the clot.

In conclusion, this case illustrates the successful use of rTPA in treating flap thrombosis during primary free flap reconstruction of a head and neck defect. It is important to regularly check flap perfusion even though it may be patent immediately following microvascular anastomosis, as pedicle thrombosis may occur most frequently in the first 24 hours and even intraoperatively. Intraoperative detection and anastomotic revision provide a better chance at flap salvage. If an arterial thrombus is identified, attempts at milking should be avoided to prevent distal clot propagation towards the flap. If there is propagation of clots into the flap, thrombolysis is indicated to restore perfusion and improve the chance of successful flap salvage.