Diagnosis, management and prevention of thrombotic complications in microsurgical breast reconstruction: a review of the literature

Hereditary thrombophilia

While there have been individual case reports of thrombosis with flap loss in Factor V Leiden patients undergoing microsurgical breast reconstruction, they cannot be used to accurately estimate thrombotic risk^[9-11]. In a retrospective study of 2032 consecutive free flaps (not limited to breast reconstruction), 58 of which were performed on patients with prior macrovascular thrombosis and/or known thrombophilia, Wang et al. found significantly higher rates of flap thrombosis and flap failure among the hypercoagulable group^[12]. However, Pannucci et al. noted that this study failed to recognize that flap thrombosis occurred only among hypercoagulable patients with prior history of macrovascular thrombosis or another acquired hypercoagulable disorder^[13]. Flap thrombosis did not occur in patients who had known hereditary thrombophilia without any additional history, suggesting that hereditary thrombophilias are less predictive of flap outcomes than acquired thrombophilias or prior history of thrombosis.

Based on these findings, Pannucci et al. recommend preoperative screening according to personal and family history of thrombosis, acquired risk factors, and Caprini score; if the patient has elevated risk determined by the screening, they should be referred to a hematologist^[13]. This approach deviates from the algorithm previously proposed by Friedman et al., who suggest that surgeons should order thrombophilia testing if there is a concern for thrombosis risk and refer to hematology only if the testing is positive^[14]. Pannucci et al. argue that decisions on thrombophilia testing should be deferred to the hematologist, because there is no evidence supporting hereditary thrombophilia as a risk factor for flap thrombosis^[13]. At our institution, all patients with a history of VTE or hereditary thrombophilia are routinely evaluated by Hematology for risk optimization and operative clearance. Typically, a prophylactic regimen of either an injectable low molecular weight heparin (LMWH) or an oral, direct factor Xa inhibitor is recommended for 1-4 weeks postoperatively. All patients are placed on heparin prophylaxis intraoperatively.

Obesity

Obesity in patients undergoing microsurgical breast reconstruction has been associated with increased risks of partial flap necrosis, fat necrosis, and venous congestion^[15-17]. In a study of 936 free transverse rectus abdominis muscle (TRAM) flap cases, Chang et al. found that obese and overweight patients had a significantly higher overall flap complication rate of 39.1% (compared to 20.4% among normal-weight patients), which included a total flap loss, hematoma, seroma, and skin necrosis^[18]. Notably, they did not find any difference in the rate of vessel thrombosis. Hanwright et al. found similar results in their analysis of free flap breast reconstruction cases taken from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database^[17]. When classified into class I (BMI of 30 to < 35), class II (BMI of 35 to < 40) and class III (BMI ≤ 40) obesity, Fischer et al. found that class III obesity patients had significantly higher rates of flap loss and trended toward higher rates of intraoperative arterial thrombosis^[19]. Similarly, Schaverien et al. found that class III obesity was associated with a significantly higher risk of complete flap failure, donor-site complications, and overall complications^[20].

Given the increased risk of complications associated especially with morbid obesity, careful patient selection is necessary and patients with class III obesity may be advised to reduce their weight prior to surgery^[19]. While no studies demonstrate a specific BMI that provides an acceptable risk to proceed with surgery, statistically significant differences in complications tend to increase proportionally as BMI increases^[21]. In our practice, we do not use a specific BMI cutoff to assess surgical candidacy, but we candidly discuss the increased risks of partial or total flap failure with all class III patients seeking microvascular breast reconstruction.

Tobacco use

Despite experimental evidence on the detrimental effects of tobacco smoke exposure on thrombogenicity, clinical studies on free flap transfers in breast reconstruction have demonstrated conflicting results, with the majority suggesting a less significant effect^[22-24]. Khouri et al. found that there was no significant effect of tobacco use on flap outcome^[25]. Masoomi et al. and Arnez et al. found no significance in flap loss or vascular thrombosis rates in smokers compared to non-smokers^[7,26]. Despite studies showing a weak association between flap thrombosis and smoking, patients should still be advised to cease smoking a minimum of 3 weeks prior to surgery, a widely advocated practice due to the established risk of poor wound healing^[27].

Radiation

Patients seeking breast reconstruction following post-mastectomy radiation therapy have become increasingly common. Although radiotherapy is known to impair wound healing, its effect on microanastomoses remains an area of ongoing study^[28]. Findings from animal studies on irradiated microanastomoses have been variable, with some demonstrating significant change in patency due to intimal hyperplasia as well as increased thrombosis risk, and others showing no such effects^[29-32]. Fracol et al. and Fosnot et al. found a significantly higher risk of any intraoperative vascular complication in radiated fields compared to non-radiated fields, but no significant differences in arterial or venous thrombosis rates both intra- and postoperatively, and no overall difference in the rate of flap loss associated with radiation^[33,34].

Complication rates have been shown to be lower in patients who delay breast reconstruction until after radiation is complete^[35,36]. Though there is little consensus in the literature regarding the optimal timing of autologous reconstruction following radiation therapy, the majority of surgeons report waiting for 4-6 or 7-12 months after the end of radiation, with patient preference and desire to optimize aesthetic outcomes being the primary drivers of the timing selected^[37,38]. Notably, Baumann et al. found that among patients receiving delayed abdominal free flap breast reconstruction, flap loss and reoperation rate was higher following reconstruction within 12 months of completion of radiation therapy^[39]. At our institution, we routinely delay breast reconstruction for at least 6-12 months after the last radiation treatment, depending on the total radiation dosage, the patient’s symptoms, and the effects noted on physical examination.

Hormonal therapy

Anti-estrogen therapies used in the adjuvant treatment of hormone-sensitive breast cancers, most notably tamoxifen, have been shown to be associated with venous thromboembolism^[40-42]. However, results from studies examining the effect of hormone therapy on flap thrombosis are conflicting^[43-47]. Although there are studies suggesting discontinuation of hormone therapy 2 to 4 weeks prior to breast reconstruction, there is no consensus in the literature on whether cessation is necessary^[44,46-50]. Until a more definitive conclusion is reached, at our institution, we typically recommend holding hormone therapy for a period of 2 weeks before and after surgery, given the low oncologic risk of short-term cessation.

Heparin

Heparin binds and enhances the activity of antithrombin III, which in turn inhibits the coagulation cascade and effectively blocks clot formation and growth. Although animal studies have demonstrated improvement in microvascular thrombosis rates with heparin, clinical findings have been conflicting^[51,52]. Lighthall et al. and Zhou et al. found no significant differences in flap failure rates between patients with postoperative heparin and patients with no postoperative anticoagulants^[53,54]. Multiple studies have also found no significant differences in microvascular thrombosis rates between cases performed with and without intraoperative heparin^[25,55,56]. However, an earlier study by Kroll et al. found that free flap patients dosed with postoperative heparin had lower pedicle thrombosis rates and a trend toward lower flap loss compared to patients with no postoperative anticoagulant agents^[56].

Aspirin

Aspirin inhibits the production of thromboxane A2 by platelets, which prevents further platelet activation and aggregation. Similar to heparin, the effectiveness of aspirin for flap thrombosis prevention is unclear despite its widespread use. When used alone, aspirin has not been found to be effective for thromboprophylaxis and may be associated with higher complication rates^[53]. Interestingly, Ashjian et al. found in a retrospective review of 505 microvascular free flap patients that rates of microvascular thrombosis and flap loss were equivalent between patients who received a postoperative 5-day daily regimen of 325 mg of aspirin and patients who received 5,000 units of LMWH until ambulating^[57].

Combination Therapy

Although many recommendations for thromboprophylaxis have been proposed, there is no consensus on any single regimen. Based on experience and literature review, Conrad et al. proposed a protocol consisting of aspirin dosed 1.4 mg/kg/day administered pre- and postoperatively for 2 weeks, with intraoperative heparin as a bolus and local topical agent^[58]. Stephan et al. and Brinkman et al. do not recommend aspirin and instead adhere to heparin monotherapy^[59,60]. Overall, current evidence seems to suggest that a more conservative approach to prophylactic antithrombotics is warranted. In a recent systematic review, Liu et al. concluded that postoperative antithrombotics including aspirin, dextran, and heparin had no significant effects on flap thrombosis or flap failure, and may increase the risk for hematoma regardless of regimen^[61].

At our institution, patients undergoing microsurgical breast reconstruction receive a bolus of heparin 5,000 units subcutaneously, or enoxaparin (40 mg or 0.5 mg/kg if BMI exceeds 40 kg/m²) subcutaneously intraoperatively. Postoperatively, patients receive heparin 5000 units subcutaneously every 8 h, or enoxaparin (same dosing scheme as previously stated) subcutaneously and aspirin 121.5 mg (half a baby aspirin) once per day. In patients deemed at high risk for microvascular thrombosis, enoxaparin is continued for 3-4 weeks postoperatively.

Flap monitoring techniques

It is well established in the literature that early detection of and intervention for microvascular thrombosis maximizes the chance of flap salvage^[71,72]. Therefore, having a reliable means of flap monitoring is critical. In 1975, Creech and Miller described the ideal flap monitoring technique as one that does not cause harm to the patient or flap and is rapid, accurate, reliable, cost-effective, and applicable to all flap types^[73]. Jones further proposed that the ideal monitor be objective, simple to use for inexperienced personnel, and capable of continuous and prolonged monitoring^[74]. While no one flap monitoring technique embodies all of these qualities, multiple technologies have since emerged, often used in combination with conventional techniques.

Physical examination

Physical examination is a commonly practiced method for assessing flap viability. The physical examination should include an assessment of flap color, temperature, turgor, and capillary refill^[74,75]. Flap temperature has previously been assessed via touch, temperature probe, temperature-sensitive tape, or handheld contactless thermometer. However, surface temperature monitoring has not been routinely recommended in larger perforator flaps due to its inability to detect changes prior to flap failure or predict reoperation in DIEP flap breast reconstructions^[76]. Bleeding is also an important component of the physical examination in flap monitoring. This can be ascertained via needle prick or skin incision; however, both of these techniques can result in transient ecchymosis that may affect accurate flap assessment^[74,76].

The physical examination can aid in determining the etiology of flap thrombosis. An arterial thrombosis can be characterized by a cool and pale appearing skin paddle, diminished turgor, and delayed capillary refill. On the contrary, venous thrombosis is often characterized by an edematous and mottled appearing skin paddle with brisk (< 1-2 seconds) capillary refill, increased turgor, and bleeding on needle prick^[75].

Vascular flow

Physical examination is typically accompanied by a more objective assessment of vascular flow. Currently, the most widely used techniques include those that monitor vascular flow and those that monitor tissue ischemia [Table 1]. The most commonly used objective assessment is intermittent interrogation of blood flow with an acoustic Doppler. While handheld acoustic Doppler sonography is not capable of performing continuous monitoring, it is a widely available, cost-efficient, and non-invasive method for monitoring vascular flow that can be easily operated by house staff^[77].

Additional Doppler technologies have since been developed to allow for continuous flap monitoring, including the Cook-Swartz implantable Doppler, flow coupler implantable Doppler, and laser Doppler flowmetry^[78]. Given the invasive nature of implantable Doppler, some authors suggest that vessel compression and anastomotic injury by the implanted cuff or wire should be considered^[79]. In addition, laser Doppler is a promising non-invasive option but lacks consensus values for detection and thus remains experimental^[80].

At our institution, we recommend regularly spaced intervals of acoustic Doppler sonography, ideally in combination with continuous tissue oximetry-based monitoring. However, implantable Doppler is used for continuous monitoring when a skin paddle is not available, such as in buried skin flaps or muscle flaps without a skin paddle.

Tissue ischemia

Near-Infrared Spectroscopy (NIRS) tissue oximetry is an important tool that has been shown to detect flap compromise before it is clinically apparent, decrease rates of flap loss, and improve rates of flap salvage compared to conventional techniques^[81]. While more expensive upfront than continuous Doppler techniques, NIRS has demonstrated an overall potential cost benefit across multiple studies^[75,82-84]. Pelletier et al. found an average reduction of $1,937 per patient when monitored on the surgical floor with NIRS tissue oximetry compared to the surgical intensive care unit (ICU)^[82]. Additionally, given the quantitative output of NIRS compared to Doppler technology, an automated text message alert system has been developed, allowing for rapid notification of the surgical team^[85]. The potential for decreased time to re-exploration, a critical factor in flap salvage, makes NIRS a promising technology. While NIRS is a valuable tool to continuously monitor flaps with a skin paddle, no single monitoring device should supersede a thorough physical examination and individual clinical experience.

Flap monitoring protocols

Currently, there is no universally recognized protocol or standardized practice for flap monitoring following microsurgical breast reconstruction. Historically, flaps have been monitored in an intensive care or step-down setting for 1 or more days postoperatively, given that the majority of complications occur within the first 24-48 h after surgery^[85,86]. With advancements in flap monitoring technologies, many institutions have altered their protocols to allow for early disposition to the floor without increasing the risk of flap failure or postoperative complications^{[82,83,87-89]}. In line with the available literature, we present our institution's flap monitoring protocol in [Figure 2], adapted from Khansa et al. to reflect our institution’s recommendation for timing and location of monitoring, and criteria for takeback^[90]. Nonetheless, we recognize that ultimately a surgeon’s approach to flap monitoring should take into account individual patient factors, institutional resources, and the evolving literature.

Figure 2. Algorithm for postoperative flap monitoring. This figure is adapted from Khansa et al. published in Microsurgery by Wiley Periodicals, Inc., copyright 2013^[90]. Adapted with permission from John Wiley and Sons.

Intraoperative management of arterial insufficiency

Upon intraoperative detection of signs of arterial insufficiency, the anastomosis should promptly be re-examined. The vessel should be inspected for extrinsic compression, vessel spasm, and positional issues such as kinking^[91]. The anastomosis should be assessed for the presence or absence of flow with clinical examinations such as the milking test and acoustic Doppler sonography. If arterial thrombosis is suspected, one or more salvage modalities may be attempted^[90]. A detailed algorithm for our approach to the intraoperative management of arterial insufficiency is available in [Figure 3A], adapted from Khansa et al. to reflect our institution’s use of papaverine^[90].

Figure 3. Algorithm for intraoperative management of free flap vessel insufficiency. (A) arterial insufficiency; (B) venous insufficiency; This figure is adapted from Khansa et al. published in Microsurgery by Wiley Periodicals, Inc., copyright 2013^[90]. Adapted with permission from JohnWiley and Sons.

Arterial flow present

Upon exploration of the pedicle, should arterial Doppler flow be present, a careful clinical examination of the flap and the entire pedicle should be performed. The pedicle should be inspected for any areas that may be prone to twisting, kinking, or external compression. The use of fat grafting over the pedicle can help to maintain the optimal vessel lie. Should the flap appears clinically improved - including the presence of normal capillary refill, turgor, and dermal edge bleeding - it may be carefully re-inset. ICG angiography could be considered to evaluate the flap after inset to ensure the adequacy of flow. If flow is confirmed, close clinical observation in the postoperative period is recommended^[90].

Arterial flow diminished

If flow is present but diminished, etiologies can include partial microvascular thrombosis, vasospasm, or suboptimal vessel positioning. Vasospasm is best treated through the avoidance of peripheral vasopressors and the local application of topical vasodilators^[92-95]. Topical treatments include a wide variety of vasodilators, including alpha antagonists (eg., phentolamine), calcium channel blockers (eg., nicardipine), direct vasodilators (eg., hydralazine), local anesthetics (eg., lidocaine), and phosphodiesterase inhibitors (eg., papaverine)^[96]. As multiple vasodilators have been proven to be efficacious, the precise type of vasodilator and the dosing used is more often based on surgeon experience and availability^[95-98]. At our institution, vasospasm is often treated with an adventitial injection of papaverine and warm heparinized saline. If there is a specific point of vasospasm identified, these injections can be combined with careful milking of the pedicle using microforceps or the surgeon’s pinched fingers, a technique that may provide sufficient intraluminal pressure to break the spasm. Should arterial flow not improve after treatment for vasospasm, the anastomosis should be re-explored, as described in the Arterial Flow Absent section.

Arterial flow absent

Time to re-exploration and anastomotic revision is critical to flap survival. If the arterial flow is absent or does not respond to treatment for vasospasm, the anastomosis should be opened and promptly explored. If a thrombus is identified upon opening the arterial anastomosis, heparin irrigation and mechanical thrombectomy or chemical thrombolysis may be necessary in addition to revision of the anastomosis. In cases of arterial thrombosis, heparinized saline is used liberally in a 100 I.U./mL concentration to flush the flap and the anastomosis. A review by Couteau et al. supports this practice, with 9 of 11 animal studies showing improved free flap survival rates with the use of intraoperative heparin irrigation compared to saline^[99].

The simplest form of mechanical thrombectomy is the direct removal of the thrombus with standard microforceps. If the thrombosis is detected prior to propagation, direct removal at the proximal end of the anastomosis can be sufficient. However, if thrombus is suspected to be in the distal pedicle, Fogarty catheter-assisted thrombectomy may be necessary^[100]. We typically use a Fogarty catheter with a 1- or 2-mm balloon. The catheter may be introduced via the proximal lumen or a distal side branch if one of sufficient size is available for cannulation. The catheter should be carefully passed until it reaches the perforating vessels entering the flap to ensure that the entire vessel is cleared. Prior to withdrawal, the balloon is typically inflated to half its total capacity to minimize damage to the perforator. Multiple passes may be needed to completely eliminate the propagated thrombus^[101]. Once the mechanical thrombectomy is complete, heparinized saline flushes can be used to assess the flap’s resistance to flow. If the pedicle can be flushed easily and venous return of the saline is confirmed, a revision of the anastomosis should be attempted. Nevertheless, if resistance to flow is detected, chemical thrombolysis may be needed. At our center, a catheter clearing dose of 1-2 mg of tissue plasminogen activator (tPA) is used. To avoid systemic thrombolysis, we ensure that the flap is isolated from the systemic circulation during the injection of the tPA. After the thrombolytic is allowed to dwell within the flap for several minutes, the flap is again flushed with 300-500 milliliters of heparinized saline to minimize the introduction of systemic tPA after reanastomosis. Furthermore, should an arterial thrombus not be identified after opening the anastomosis, the venous anastomosis should be further explored per the Intraoperative Management of Venous Insufficiency guidelines.

Although the above algorithm is used for complete loss of inflow, it is also possible to have partial loss of arterial flow. If the clinical examination or ICG angiography demonstrates inadequate flow to only a portion of the flap, the presence of partial flap thrombosis should be considered. Partial flap thrombosis, especially if it is thought to be intra-flap thrombosis, is often treated with medical management. At our center, we typically attempt to treat partial intra-flap thrombosis with a combination of chemical thrombolysis (e.g., tPA), anticoagulation and/or simple debridement of the thrombosed portion of the flap.

Reoperation

If arterial thrombosis is suspected in the postoperative period, expeditious return to the operating room to expose the anastomosis is the most appropriate next step^[90]. Time to reoperation is consistently shown to be associated with salvage rates after arterial and venous thrombosis^[71,72,102]. Likely secondary to delays in management, the rate of flap salvage in postoperative compromise is less than that of intraoperative compromise^[90]. It has been shown that the use of careful continuous postoperative monitoring is associated with a decreased time to diagnosis and operative management of flap thrombosis, and thus an increase in the rate of salvage^[81]. The approach to anastomotic assessment and revision is discussed in the intraoperative management section above. Our novel algorithm for the approach to postoperative management of thrombotic complications is available in [Figure 4].

Figure 4. Algorithm for postoperative treatment of microvascular thrombosis.

Systemic anticoagulation

In the setting of postoperative arterial thrombosis, the use of systemic anticoagulation varies by surgeon and institution. Many authors report urgent administration of a 5,000-unit bolus of intravenous heparin at the time of reoperation^[71,103]. Others report the use of weight-based dosing to achieve an institution-determined therapeutic PTT level^[58,103]. Given higher rates of hematoma with systemic heparin use in free tissue transfer, we typically recommend weight-based dosing of intravenous heparin without the use of a bolus^[87]. Once the risk of postoperative hemorrhage is deemed sufficiently low, the patient may be transitioned to a low molecular weight heparin injection (eg., enoxaparin). The duration of anticoagulation after flap thrombosis is often driven by surgeon experience. Although there is no definitive data on the optimal duration of therapeutic anticoagulation in this cohort, Khansa et al maintain systemic anticoagulation for at least 4-7 days after reoperation, which is consistent with what other authors report^[58,71,90].

Venous congestion and venous thrombosis

The development of venous congestion can be attributed to several etiologies, the most common of which is venous thrombosis^[104]. Mechanical factors such as unfavorable flap position, or kinking and compression of the vascular pedicle are common causes of venous insufficiency. Venous thrombosis can occur secondary to a hypercoagulable state, a technical error at the anastomosis, or from prolonged venous congestion or insufficiency from one of the above mechanical factors^[105,106]. Finally, venous insufficiency can result from anatomic variability within the flap, especially if the flap contains portions of two perforasomes. Anatomical studies of DIEP and TRAM flaps have shown that normal venous drainage of the lower anterior abdominal subcutaneous tissue and skin occurs primarily through the superficial inferior epigastric vein (SIEV), which is connected to the deep inferior epigastric vein (DIEV) by choke vessels composed of the vena comitantes of the perforators of the deep inferior epigastric artery (DIEA)^[107]. Although the majority of DIEP flaps may survive based on the outflow from the DIEV system alone, the venous outflow of some DIEP flaps may be superficially dominant. In these cases, both the DIEV and the SIEV may require anastomosis for adequate venous drainage^[107].

Preoperative imaging with CT angiogram, especially among patients with prior surgery in the region of planned flap harvest, may be beneficial in perforator selection and evaluation of venous anatomy. The presence of the SIEV and its caliber should also be evaluated radiologically and intraoperatively. If the SIEV is noted to be of good caliber (> 1.5 mm), it is prudent to preserve adequate length on this vessel to allow for anastomosis.

Intraoperative management of venous insufficiency

As described by Heller and Levin, obstruction of venous outflow can lead to red blood cell extravasation, endothelial breakdown, microcirculation thrombosis, and flap death^[108]. Signs of venous congestion in the flap may include rapid capillary refill on the skin paddle (< 1-2 seconds), more profuse dermal bleeding of a darker color, loss of venous Doppler signals in the pedicle and perforators, greater flap turgor, and enlarged secondary veins such as the SIEV. Our algorithm for intraoperative management of suspected venous insufficiency is available in [Figure 3B], adapted from Khansa et al. to reflect our institution’s use of leeching and de-epithelialization^[90].

Suspicion for venous congestion requires release of insetting sutures and diligent assessment of vessel position, flap and pedicle orientation within the breast pocket, and presence of hematoma and/or edema requiring evacuation. This should be paired with the assessment of anastomotic flow using Doppler sonography. Secondary veins, such as the SIEV in the case of a DIEP flap, should be examined closely. These veins should be opened to allow for assessment of their outflow. If the secondary veins are noted to have robust outflow, they should be used for venous supercharging of the flap as described below (Venous Flow Present).

Flaps or pedicles can be repositioned to achieve a more favorable lie without kinking. Autologous fat grafts may be used to cushion the pedicle or maintain its position without twists or kinks. Re-evaluation of the venous anastomosis or coupler is critical. A milking test can be used to ensure flow across the anastomosis and throughout the length of the pedicle. If the flap vein was twisted or kinked relative to the recipient vein during anastomosis, the anastomosis may need revision to avoid propagation. Venous vasospasm can cause global congestion, which can usually be resolved by irrigating the vessels with vasodilators (papaverine, lidocaine, verapamil, and nitroglycerin mixture) and warm salinex^[109].

Venous flow diminished or absent

If there is absent or diminished flow in the anastomosis after mechanical factors and simple vasospasm have been ruled out, the anastomosis should be taken down and inspected for thrombi. As venous supercharging may be necessary, secondary veins should be examined closely as described below. The artery should also be carefully inspected for signs of diminished flow. If the artery is noted to have abnormal flow, this vessel should be treated using the algorithm above (See intraoperative management of arterial thrombosis). If thrombus is noted in the proximal vein, direct thrombectomy with microforceps should be attempted. If the thrombus is too extensive or distant for this approach, a 1-3 mm Fogarty catheter can be used to attempt thrombus removal in the distal pedicle. Although there is concern that the use of Fogarty catheters on microvasculature can increase the risk of endothelial denudation and thrombogenesis, studies on complication rates following Fogarty catheterization so far have been conflicting and limited by small study samples. While some studies have reported successful flap salvage using the Fogarty catheter, others have found a higher rate of failure in flaps undergoing Fogarty catheter thrombectomy^[101,110]. However, as there have been no large-scale studies on the use of Fogarty catheters in flap salvage, and all previous studies are subject to significant selection bias, we believe that Fogarty catheters have an important role in the armamentarium of the reconstructive microsurgeon - especially when attempting to salvage flaps with more extensive or proximal thrombi that are not accessible for direct thrombectomy. After the Fogarty catheter is used, the vessels are typically flushed copiously with heparinized saline as per our arterial algorithm.

If no thrombus is visualized upon taking down the anastomosis, it is possible that there is evidence of intra-flap venous insufficiency or thrombosis. Although ICG angiography is typically used to examine inflow to the flap, venous outflow can be assessed using a washout phase. After the arterial flow is confirmed using the ICG, a second examination can be done after a 2-3 min delay. If ICG dye remains in portions of the flap after this delay, it is likely that these portions may be experiencing venous insufficiency. In cases of intra-flap thrombosis, whether arterial or venous, chemical thrombolysis may be needed (see the arterial treatment algorithm above). If only a portion of the flap is determined to suffer from venous insufficiency, this non-viable tissue may simply be debrided.

Venous flow present

Evidence of persistent flap congestion in the setting of venous patency indicates the need for venous supercharging, or additional venous flow augmentation. Of the two most common autologous breast reconstructions, DIEP flaps are more likely than free TRAM flaps to be complicated by venous congestion requiring flow augmentation, likely due to DIEP flaps having fewer perforators^{[107,111,112]}.

A broad range of techniques for augmenting venous outflow in abdominally-based autologous breast reconstruction has been widely reported in the literature. The vena comitantes of the ipsilateral DIEA, vena comitantes of the contralateral DIEA, ipsilateral SCIV, ipsilateral SIEV, and contralateral SIEV serve as potential sources for donor veins in venous super-drainage, with the ipsilateral SIEV being the most common^[113]. The most frequently used recipient vessel is the second internal mammary vein (if available), or the retrograde inframammary vein (IMV) due to their location allowing for optimal flap positioning on the chest. At our institution, for all DIEP flaps, we routinely preserve sufficient length on the SIEV and retrograde IMV to allow for anastomosis if necessary. Given the ease of supercharging the flap during the initial microsurgery, it is prudent to perform this anastomosis early if there is any level of concern for superficial venous dominance. Other potential recipients include the intercostal perforating vein, thoracodorsal vein, cephalic vein, thoracoacromial vein and lateral superficial thoracic vein - however, all of these systems require more time for dissection and may necessitate significant flap rotation to allow for anastomosis^[113,114].

According to cadaveric and imaging studies, over 75 percent of females will have both lateral and medial vena comitantes to the inframammary artery (IMA) present above the lower border of the 4th intercostal space (ICS). If the lateral IMV is of adequate caliber for anastomosis, this may be used as the recipient vessel for venous supercharging^[115]. Multiple studies have demonstrated that the caudal end of the IMV can accommodate retrograde flow, and have promoted the assumption that the IMV is valveless^[116-119]. However, an anatomic study by Mackey and Ramsey on 32 human cadavers found that 1 to 3 valves were present in the IMV of 44% of female cadavers and 42% of male cadavers^[120]. Additionally, valves were found between the preferred point of distal anastomosis and the next draining vein in 9% of 2nd ICS and 5% of 3rd ICS. While it is possible that valvular incompetence allowed for retrograde flow in prior studies, the findings by Mackey and Ramsey indicate that retrograde flow may not be guaranteed in the caudal end of the IMV, and that more dynamic studies are required to validate this technique^[120]. In our experience, the lateral IMV is often diminutive, even if present. In these cases, we routinely preserve sufficient length on the retrograde end of the medial IMV and utilize this for secondary venous anastomosis. Although some authors may raise concern that retrograde outflow may be diminished due to the presence of valves, we have found that there is sufficient collateralization from the intercostal system and IMV perforators to allow for adequate outflow.

If no other recipient veins are available, it may be possible to perform venous turbo-drainage via an intra-flap anastomosis. Rohde and Keller have described a turbo-drainage technique in which a superficial to deep venous loop is created within the flap by anastomosing the ligated SIEV to the proximal end of one of the vena comitantes of the DIEA^[121]. This allows blood from the superficial system to drain directly into the deep system via anterograde venous flow through the vena comitantes, and eventually through the original DIEV-IMV anastomosis. This technique requires minimal additional accommodations for vessel length or flap positioning. It is also suitable for cases in which the superficial venous system is overdominant.

If venous congestion persists despite venous outflow augmentation or there are no viable alternate recipient veins, mechanical leeching can be considered. This method entails intraoperatively placing an angiocatheter in the dominant vein, which is brought up to the skin as a venostomy for controlled manual drainage. The angiocatheter should be flushed periodically and aspirated at hourly intervals for the next 3-6 days based on clinical examination of the flap^[122,123]. Bank et al. have also reported a case of successful resolution of venous congestion with mechanical leeching guided by ViOptix measurements^[124]. Once congestion resolves, the angiocatheter can be removed and the vein can be allowed to clot. Although this method eliminates the infection risks associated with leech therapy, it still requires blood products due to volume depletion by aspiration. Furthermore, studies on mechanical leeching for venous congestion have so far reported high success rates, but given the limited number of studies and their small sample sizes, they are highly likely to be subject to publication bias.

Postoperative management of venous insufficiency

Delayed venous thrombosis in the postoperative period is most likely to occur within 3 days of the initial operation^[125]. As previously discussed, early detection of a possible venous thrombosis allows for the best chance of ensuring flap survival. Since prolonged venous congestion is often a precursor to venous thrombosis, flap monitoring and timely detection of signs of venous congestion is essential to the prevention of this complication^[104].

If clinical signs of venous congestion are present, initial management involves addressing sources of extrinsic compression that may be contributing to poor venous outflow. Common troubleshooting techniques include loosening of surgical bra, removing tight dressings, or removing compressive sutures^{[113,126-128]}. Topical nitroglycerin paste causes both arterial and venous dilation; it may be used as an adjunct to remove extrinsic compression^{[127,129,130]}. If conservative methods fail to relieve congestion, or if venous thrombosis is suspected, reoperation offers the best potential for flap salvage.

The strategy for addressing postoperative venous compromise in the operating room follows a pattern similar to that seen with intraoperative venous insufficiency. One notable exception is the presence of a hematoma that may be compressing the pedicle. Hematoma may be seen in the presence or absence of venous congestion^[131]. However, it is more often seen concurrently with venous congestion. When hematoma is suspected as the cause of venous compromise, the hematoma should be evacuated on an emergent basis to avoid further compression of the pedicle^[103,131]. In cases of delayed venous insufficiency (i.e., greater than 3 postoperative days) and/or when re-exploration of anastomosis and surgical revision may be impossible, venous insufficiency may be managed medically^[128].

While the majority of venous thrombosis events occur within the immediate postoperative period, delayed venous insufficiency and/or thrombosis have been documented up to 5 weeks after initial operation^[125,128]. In these later presentations, successful salvage without re-exploration of anastomosis is more common^[128]. Yoon and Jones suggest a critical time period for flap survival whereby flaps with delayed thrombosis have a higher rate of survival due to neovascularization and angiogenesis that has already taken place^[132].

Systemic anticoagulation in conjunction with reoperation

Heparin prevents clot formation by activating antithrombin III, which ultimately prevents the formation of fibrin^[133]. While some have utilized antiplatelet therapy in addition to systemic anticoagulation, there is well-documented evidence to show that heparin is favorable to antiplatelet therapy in cases of microvascular thrombosis^[134]. Several methods have been reported on the use of systemic heparin in cases of venous thrombosis or congestion, but timing, dosage, and routes of administration vary depending on the institution. Most authors report using the same protocol for systemic anticoagulation in both venous and arterial thrombotic complications. There is currently no standardized recommendation, and no single protocol has been proven to be superior. At our institution, we typically recommend a continuous weight-based heparin infusion titrated with a PTT of 60-80. This can be transitioned to weight-based low molecular weight heparin injections once the patient is deemed to have a sufficiently low bleeding risk. The duration of the treatment may range from 1 week to 4 weeks, depending on our level of concern for thrombosis.

Alternative venous drainage

Several methods exist for the medical management of venous insufficiency in free flaps, with varying levels of success demonstrated in the literature^[127]. Local injection of subcutaneous heparin has been demonstrated to be effective in several studies. More recently, the use of subcutaneous heparin was discussed by Perez et al, who showed that local subcutaneous injection of LMWH is an effective method for flap salvage in cases of venous congestion^[135].

Relief of venous congestion may be further facilitated by pricking or de-epithelialization of the flap. Pricking the flap with a needle allows blood loss from the congested area, thereby reducing venous compromise^[136]. In a similar manner, de-epithelializing a portion of the flap allows for venous drainage. Heparin may be injected into the de-epithelialized area or a heparin-soaked gauze may be applied to the de-epithelialized area to further increase venous outflow^[136,137].

Hirudotherapy, the use of medicinal leeches, may be used in cases of irreparable venous insufficiency and/or flap necrosis secondary to venous compromise. The application of leeches provides temporary relief of venous congestion while a more reliable network for venous drainage is being established^[127,138]. The effectiveness of medicinal leech therapy in decreasing venous congestion is two-fold; the initial blood meal by the leech allows for active drainage of ~5-15 mL of congested blood, after which passive blood loss from the bite injury continues to occur. Leech-mediated release of vasoactive substances allows for further local thrombolysis and anticoagulation^[139,140]. While leech therapy for free flap salvage has reported success rates ranging from 60-80%, it may be less effective for higher volume flaps such as TRAM or DIEP flaps^[127,138]. Primary complications of leech therapy include infection and anemia^[139]. The evidence for medicinal leech therapy is limited to case series and retrospective studies. While Pannucci et al. found that leech therapy in microvascular breast flaps was associated with higher flap loss rates, this is likely secondary to significant selection bias^[141]. Current evidence indicates that leech therapy should be used with discretion and in consideration of patient-specific risk factors^[141]. In our experience, leech therapy should be considered as an adjunct in cases with significant intra-flap venous insufficiency that does not respond adequately to other therapies.

Veno-cutaneous catheterization presents another option for the relief of venous congestion. An angiocatheter is placed into a superficial vein at the margin of the flap. Distilled heparin solution is injected into the vein. The catheter is left in place with a valve such that venous drainage may occur as needed. When clinical signs of congestion improve, the catheter may be removed^[142-144]. In comparison to leech therapy, veno-cutaneous catheterization is less costly. Further, Mozafari et al. showed that veno-cutaneous catheter use is associated with decreased blood loss, wound dehiscence, and flap necrosis compared to leech therapy. It is also associated with high rates of nursing and patient satisfaction^[145]. All reported protocols indicate that the catheter must be placed in the operating room, which is a notable disadvantage of this technique^[127]. In our experience, venocutaneous catheterization can be difficult to maintain for more than 1-2 days, given the high likelihood of catheter thrombosis with intermittent use.

Negative pressure therapy has also been reported in the literature for the management of venous congestion. However, its use in practice is still rare. Negative pressure therapy is thought to reduce congestion by decreasing edema, increasing drainage and local venous flow, and increasing the rate of neovascularization^[146-148]. Negative pressure may also have a compressive effect, making the overall benefit of this therapy difficult to accurately assess^[127].

Management of ischemia-reperfusion injury

Ischemia reperfusion injury is an important consideration for microsurgeons as tissue damage can persist well after the flow is re-established. Restoration of blood flow to a flap promotes the release of proinflammatory cytokines and reactive oxygen species (ROS), leading to tissue inflammation, coagulation, and necrosis. This cascade can ultimately result in partial or complete flap loss and fat necrosis as well as adverse patient outcomes and healthcare costs^[149]. The most dreaded outcome in this scenario is the “no-reflow” phenomenon, whereby tissue damage is so severe that the flap does not perfuse despite the patency of the anastomosis. Several factors have been implicated in an increased risk for ischemia-reperfusion injury, including tissue type, surgical technique, temperature, and ischemia time^[150].

Given the pathogenesis of ischemia-reperfusion injury, immunomodulators, antioxidants, and anticoagulants have each been proposed as potential therapeutics. While these therapies have shown promise in animal models, the data on their utility in human patients is unclear^[151-157]. For example, while statins have theoretical anti-inflammatory and antioxidant activity, Koolen et al. and van den Heuvel et al. did not find such benefits in breast microsurgery^[158,159]. Additionally, in a retrospective study, Coriddi et al. found no significant difference in lost vs salvaged flaps and patients who received intra/postoperative steroids or therapeutic anticoagulation for ischemia-reperfusion injury prophylaxis^[103]. Ultimately, more research in this area, including randomized controlled clinical trials, is needed before further therapeutic recommendations are made.

When flap salvage is not feasible

When considering approaches to tertiary reconstruction, Baumeister et al. recommend the following steps: a sensitive psychosocial approach to the patient and family, an analysis of the cause of the first flap failure, reconsideration of the need for vascularized free tissue transfer, and a change in microsurgical strategy^[160]. An investigation into the cause of flap failure should include careful consideration of the following: the preoperative preparations, the recipient vessels and anastomosis, the patient’s risk for hypercoagulability and thrombosis, the postoperative care, and the surgeon’s individual expertise. Baumeister et al. provide a thorough checklist to consider in [Table 2]^[160].

Hamdi et al. broadly classify the causes of flap failure into “technical” (anastomosis errors, pedicle kinking, anatomical variations, and quality and choice of recipient vessels and/or perforator of the nourishing pedicle) and “nontechnical” (one or more hypercoagulability disorders) etiologies^[161]. In the event of “nontechnical” flap failure, alternative options, including a pedicled flap, should be strongly considered, given the high risk of another failure. However, for patients whose free flap failed due to a presumed technical error, another free flap may be reasonably considered.

In the rare case of non-salvageable total flap failure, an in-depth and empathetic discussion with the patient and family is essential. A description of possible alternative forms of breast reconstruction will provide necessary reassurance. We recommend debriding all non-viable tissue in the operating room soon after the diagnosis is assured. The mastectomy skin flaps should be closed if possible. If the skin flaps cannot be closed primarily, a negative pressure therapy dressing may be used temporarily. In a case of partial flap failure, debridement of the non-viable tissue should take place only after demarcation. The timing of future efforts at breast reconstruction should be dictated by the patient’s preferences, psychosocial needs, and the state of the wound after flap debridement.

Flap alternatives

Much of the decision-making regarding the next steps following flap failure depends on what was found following troubleshooting of the prior failed flap, anatomical limitations of the patient, and the patient’s preferences. The decision to pursue another reconstruction should be made only after a thorough reassessment of the patient’s medical and familial history for hypercoagulability and other potential risk factors. Following their review of 14 patients who underwent tertiary breast reconstruction after a prior failed reconstruction, Hamdi et al. recommend that, based on their experience, the latissimus dorsi flap and the thoracodorsal artery perforator flap with or without an implant are associated with lower morbidity compared to free flaps, and should be considered if the patient is at high risk of complications^[161]. At our center, pedicled options such as the latissimus flap are essential for patients at high risk for microsurgical thrombosis. If the patient displays a strong preference for a free flap and they are deemed a candidate for a second attempt at free tissue transfer, preoperative planning should include CT and color Duplex imaging to assess alternative donor sites and viable recipient vessels, hematologic consultation for assessment of thromboembolism risk and application of thromboprophylaxis measures, and preparation of secondary options in case the second free flap fails.