PROMISE I: Early feasibility study of the LimFlow System for percutaneous deep vein arterialization in no-option chronic limb-threatening ischemia: 12-month results

Objective: We report the 6- and 12-month outcomes of the PROMISE I early feasibility study after treatment of no-option chronic limb-threatening ischemia (CLTI) with percutaneous deep vein arterialization (pDVA) using the LimFlow System. Methods: Thirty-two patients with no-option CLTI, previously offered major amputation, were enrolled in this single-arm early feasibility study of the LimFlow pDVA System. No-option CLTI was de ﬁ ned as being ineligible for surgical or endovascular arterial revascularization. Patients were assessed for clinical status, pain, wound healing, and duplex ultrasound at 30 days, 6 months, and 12 months post-treatment. Primary endpoint analysis was amputation-free survival (AFS) at 30 days and 6 and 12 months. AFS was de ﬁ ned as freedom from above-ankle amputation of the index limb and freedom from all-cause mortality. Secondary endpoints evaluated included technical success of the procedure, and wound healing at 6 and 12 months. Results: Of 32 enrolled patients, 31 (97%) were successfully treated with the LimFlow System at the time of the procedure, and two (6.3%) were lost to follow-up. The 30-day, 6-month, and 12-month AFS rates were 91%, 74%, and 70% respectively. The wound healing status of fully healed or healing was 67% at 6 months, and 75% at 12 months. Reintervention was performed in 16 patients (52%) with 14 (88%) of the maintenance reinterventions occurring within the ﬁ rst 3 months. The majority of reinterventions (n ¼ 12; 75%), involved the arterial in ﬂ ow tract proximal to the stented LimFlow circuit, and no in-stent stenoses were determined to have been the cause of reintervention. Conclusions: The LimFlow pDVA System was utilized in treating patients with no-option CLTI. A high technical success rate was observed, with a signi ﬁ cant percentage of patients surviving free of major amputation at 12 months. These results suggest early safety and provide an initial assessment of the ef ﬁ cacy of the LimFlow pDVA System that supports the expansion of carefully executed studies to determine whether this is a viable option that can be used in this critically disadvantaged and growing patient population. (J Vasc Surg 2021;74:1626-35.)

5 years, 4-6 and a reported major amputation rate of 25% to 30% at 1 year. 2,7 Above-ankle amputation has far reaching effects beyond the initial procedure and recovery. It is associated with a cascading series of negative events that results in a mortality of 50% by 1 year and 75% by 5 years. 8 Conventional treatment of CLTI includes both endovascular and open arterial reconstruction and is dependent on a variety of anatomic variables and patient comorbidities. A significant proportion, up to 20%, of patients with CLTI are deemed "unreconstructable" (no-option) with arterial revascularization techniques due to the absence of a viable distal target vessel, viable conduit, or other comorbidities. 9,10 The recognition of this patient population has given impetus to alternative techniques to include venous arterialization (ie, diversion of arterial flow into the venous vasculature), first reported by Halstead and Vaughan in 1912. 9 Open surgical venous arterialization failed to gain clinical acceptance secondary to technical challenges and high surgical morbidity. 11 This no-option group of patients faces a poor prognosis, with a reported 40% major amputation rate and 20% mortality at 6 months. 2 For those patients who do undergo amputation, the risk of peri-procedural and 1-year mortality is very high: above-the-knee and below-the-knee amputation are two of the five surgical procedures most associated with the greatest number of perioperative deaths. 12 In an effort to improve limb salvage and mortality, endovascular deep vein arterialization of the foot has been attempted in multiple settings using the tools and techniques currently available (ie, "off-label"), given the absence of any purposebuilt and approved devices for achieving effective venous arterialization of the foot. These reports demonstrate a broad range of variable results but have not produced a consistent procedure and technique that would allow this technique to become a useful and accepted option. 13,14 The LimFlow System for percutaneous deep vein arterialization (pDVA) represents a purpose-built product and standardized approach to treating CLTI in the no-option population. Initial clinical studies of the LimFlow System were performed in the European Union and Singapore, with CE mark obtained in October 2016. 10 The PROMISE I early feasibility study (EFS) was launched in the U.S. in 2017. Early results of the first 10 patients to be enrolled in the study have previously been reported. 15 Presented here are the 1-year results of the full 32-patient cohort.

METHODS
Trial design and participants. PROMISE I is a prospective, multicenter, single-arm, early feasibility study of the LimFlow pDVA approach to treating no-option CLTI conducted under investigational device exemption from the U.S. Food and Drug Administration. The objective of the study is to assess the feasibility, safety, and effectiveness of the LimFlow stent graft system. Between July 2017 and April 2019, 32 patients were enrolled at seven trial centers in the United States. Trial follow-up will continue through 2 years. The target population included adult patients with no-option CLTI, defined as being ineligible for conventional surgical or endovascular revascularization on the basis of absence of pedal artery target or suitable vein conduit and a salvageable foot with either open wounds or gangrene (ie, Rutherford 5/6) based upon review by an independent review committee as outlined below. As there is no standard definition of a "salvageable" limb and the investigators were experienced in the treatment of patients with critical limb ischemia and wounds, the determination of whether a limb was salvageable or not was left to the determination performed by the investigator with oversight by the independent review committee. Patients were excluded if they were New York Heart Association Class IV, had any ipsilateral revascularization procedure within 30 days prior to index, or had uncontrolled infection (ie, osteomyelitis, deep abscess). All patients underwent baseline examination consisting of physical examination, calibrated photographic wound assessment, and assessment of pain. Baseline ultrasound, to assess for venous bypass conduit, and angiographic images, to assess for arterial surgical or interventional target, were carefully reviewed by an independent review committee (IRC; Syntactx, New York, NY) of vascular experts using strict guidelines to confirm lack of standard revascularization options and thus suitability for the pDVA. Patients who had occlusions above the tibial arteries were included if the vessel above the tibial arteries and the origin of a tibial vessel could be recanalized, allowing a cross-over to the deep vein system in a tibial vessel. These inflow procedures had to be completed prior to the evaluation by the IRC, and additionally, these revascularizations had to have been performed at least 30 days prior to acceptance into the ARTICLE HIGHLIGHTS d Type of Research: Multicenter prospective singlearm study d Key Findings: Of 32 patients with "no-option" critical limb-threatening ischemia, 31 underwent percutaneous deep vein arterialization with the LimFlow System, resulting in 74% and 70% amputation-free survival at 6 and 12 months respectively. Core labadjudicated wound healing status of fully healed or healing was 75% at 12 months. d Take Home Message: The LimFlow percutaneous deep vein arterialization system was effective in treating no-option critical limb-threatening ischemia, resulting in a high percentage of treated patients healing their wounds and surviving free of major amputation at 12 months.
trial. All participating sites obtained central Institutional Review Board approval before enrollment, and each patient provided trial-specific informed consent before treatment. The ClinicalTrials.gov identifier for this study is NCT03124875, and the trial was funded by LimFlow (LimFlow Inc, San Jose, Calif).
Device and procedure. The LimFlow System consists of arterial and venous catheters that allow ultrasonographic determination of crossover direction and location (Fig 1); an antegrade, over-the-wire valvulotome (Fig 2); and selfexpanding stent grafts specifically designed to divert blood flow from the tibial (donor) artery into the tibial and pedal (recipient) venous system (Fig 3). The arterial and venous catheters allow orientation of the cross-over needle in the arterial catheter to direct the needle into the appropriate vein. The LimFlow procedure starts with confirmation of adequate pedal venous anatomy via ultrasound venous mapping and/or venography. Adequate pedal venous anatomy consists of confirmation of patent venous vessels of at least 2-mm diameter on the dorsal and plantar surface of the foot without contained thrombus and with ability to be compressed at the time of ultrasound. Pedal venous access facilitates uneventful wire access of the target vessel and recipient tibial vein. Ipsilateral antegrade access of the common femoral artery is obtained, and a sheath is advanced to the popliteal artery. A radiopaque marker tape serves as a reference for crossover location and stent graft length. Arterial and venous imaging is performed simultaneously to locate a proper crossover point, with care taken to preserve the native arterial system and collaterals; thus, in some instances, recanalization of a portion of occluded tibial artery is performed so that flow that is currently providing arterial flow through collaterals to the lower limb and into the foot is maintained. The maintenance of the arterial flow noted prior to performance of the deep vein arterialization is critical to the early success of this procedure. The venous LimFlow catheter is advanced cephalad in the recipient tibial vein and positioned at the crossover point. The arterial LimFlow catheter is advanced into the donor artery, most frequently the posterior tibial artery, to the crossover point. The fluoroscopy tube is rotated to obtain an angle orthogonal to the plane of the artery and vein. The Lim-Flow ultrasound system is used to obtain radial alignment for arterio-venous crossing when the venous and arterial catheters are at the same longitudinal position. Once the longitudinal and radial orientation are   confirmed, the crossing needle of the arterial catheter is deployed into the target vein, and a guidewire is advanced from the artery into the vein. The crossover point is dilated, and the LimFlow valvulotome is advanced through the arterial sheath and activated to render all venous valves from the crossing point to the midfoot (typically the lateral plantar vein) incompetent. This antegrade, over-the-wire valvulotome passes over a 0.018-inch wire and is contained within a catheter. The valves on the device are directed somewhat inward to avoid damage to the sheath, but will engage the venous valves and with forward pressure will lyse the valve cusps. Low pressure balloon inflation to 4 to 5 mm confirms absence of stenosis in the recipient tibial and plantar veins. LimFlow straight covered stents (5-mm diameter and 6, 10, and 15 cm in length) are then deployed from the level of the calcaneus to the arterio-venous crossover point, and a LimFlow tapered covered stent (3.5-to 5.0-mm taper and 4.0-to 5.0-mm taper; both 6-cm length) is placed to bridge from the recipient vein into the donor artery. Balloon angioplasty of the implanted stent graft system confirms nominal diameter and freedom from restrictive lesions. Completion arteriography confirms adequacy of flow into the stent graft conduit and flow through the pedal venous loop. Access site closure is managed at the discretion of the operator. Postprocedural anticoagulation therapy was left to the discretion of the investigators.
Postprocedural follow-up. Patients were assessed for clinical and duplex ultrasound status at months 1, 3, 6, and 12. Evaluations included an assessment of pulse, pain, Rutherford Classification, transcutaneous oxygen pressure (TcPO2), wound healing status, Wound, Ischemia, and foot Infection (WIfI) scoring, 16 medication changes, and any procedures or complications related to the index limb occurring since the last visit. Wound care was provided either by the investigators themselves or at dedicated wound care centers depending upon the standard care provided at each investigational site. All wounds were photographed with a calibrated wound evaluation photography system (eKare inSight; eKare Inc, Fairfax, VA) and were qualitatively and quantitatively assessed by an independent wound care lab (Syntactx). Beyond the initial 6 months, follow-up was performed with duplex ultrasound and investigator wound assessment along with calibrated wound photography at 12 and 24 months. Duplex ultrasound of the treated vasculature was performed at all visits by registered vascular technologists and interpreted by the investigators. A medical monitor was used to interpret the duplex ultrasound results. Trial follow-up will continue with annual clinical follow-up to 2 years. Occurrence of adverse events was assessed at each follow-up visit. Safety was reviewed at regular intervals. All study deaths and above-ankle amputations were adjudicated for composite endpoint of amputation-free survival (AFS) determination.
Data and statistical analysis. The safety endpoint of AFS was analyzed at 30 days and at 6 and 12 months. AFS was defined as freedom from above-ankle amputation of the index limb and freedom from all-cause mortality. Secondary endpoints included procedural and technical success and wound healing. Procedural success was defined as the combination of technical success and absence of all-cause death, above-ankle amputation, or clinically driven major reintervention of the stent graft at 30 days. Technical success was defined as completion with the LimFlow pDVA procedure with placement the covered stent with immediate filling of the pedal-venous channels. Analyses are reported in a descriptive fashion, given the limited sample size of the

EFS. Primary safety analysis reports survival rates with
Kaplan-Meier analysis. Statistical analysis was performed by Syntactx.

RESULTS
Baseline characteristics of the trial population. Thirtytwo patients were deemed eligible for the LimFlow pDVA procedure by the IRC based on review of baseline ultrasounds and angiograms. Information regarding 51 patients was submitted to the IRC, and of these, 38 were approved for inclusion. Of the approved 38 patients, 32 underwent the LimFlow pDVA procedure and were enrolled in the trial. Baseline characteristics are outlined in the Table. The patient population was 66% men with a mean age of 70.8 years. As is typical of chronic limb ischemia populations, there was a high percentage of diabetics (69%), with 75% of these on insulin. Approximately one-half of patients (59.4%) were overweight (body mass index >25). Rates of hypertension and hyperlipidemia were high (88% and 91%, respectively), and approximately one-third of patients (34%) had renal insufficiency. Of the 32 patients enrolled, 28 were noted to have no viable target for revascularization, and four were noted to have inadequate vein conduit for bypass.
Trial outcomes. Of 32 enrolled patients, LimFlow pDVA technical success was seen in 31 cases (96%). The single technical failure was related to an inability to achieve venous access beyond the ankle in the first case attempted by one of the investigators. Two subjects were lost to study follow-up (one at 30 days and one prior to 6 months). Procedural success was 75% (24/32), with one failed procedure, zero deaths, four reinterventions, and three amputations (one of which also underwent reintervention) occurring within the first 30 days. The 30-day, 6-month, and 12-month Kaplan-Meier estimate AFS rates were 91%, 74%, and 70%, respectively (Fig 4).Of the eight subjects who failed to meet the 6-month AFS endpoint, there were two deaths (one related to sepsis and one related to intracranial hemorrhage) and six major, aboveankle amputations. A single patient expired from systolic heart failure between 6 and 12 months with the limb intact. Of the 21 subjects who met the study endpoint of AFS at 6 months, core lab-adjudicated wound healing status of "fully healed" or "healing" was 67% (n ¼ 14) at 6 months, whereas of the 20 patients who remained alive and free of amputation at 12 months, the core labadjudicated wound healing status of "fully healed" or "healing" was 75%; there was a trend towards area reduction at both time points (Figs 5 and 6). Fully healed was defined as all surfaces of the wound fully epithelialized, and healing was defined as evidence of granulation tissue formation, apparent wound edge epithelialization, and evident contraction of wound edges.
Secondary outcomes in the trial included recording of TcPO2 and WIfI status. The TcPO2 recordings were inconsistent across sites and time points and were extremely variable, making interpretation impossible. WIfI status was fairly consistently recorded across sites and is presented graphically in Fig 7. Although the numbers are small, there is a decreasing average WIfI score for each aspect of the score. Particularly of note, the mean ischemia score decreases across the time period of the study.
Within the group of patients with limb salvage, there were a number of amputations below the ankle level performed to eliminate gangrenous or infected tissue and allow healing. In total, 19 minor amputations were performed in 15 patients, including seven toe amputations, two ray amputations, and 10 trans-metatarsal amputations.
Reintervention in the form of plain balloon angioplasty, drug-coated balloons, drug-eluting stents, cutting balloon, or atherectomy was performed in 52% of patients (16/31). The pattern of reintervention performed occurred temporally early in the follow-up, with the majority of the cases having reintervention within the first 3 months, and the majority (75%; 12/16) of these involved reinterventions to the arterial inflow tract proximal to the  LimFlow stented circuit. The CLTI population has, by definition, diseased arteries, and it is to be expected that maintenance of the inflow arteries will be a necessary key to success in those patients with extensive inflow disease. DISCUSSION PROMISE I demonstrates that pDVA utilizing a the Lim-Flow System can be performed with good technical success and encouraging AFS rates at 6 and 12 months. The patient population of PROMISE I is unique with respect to the severity (ie, no-option as confirmed by independent review) of their arterial disease and the degree of ischemic damage (ie, Rutherford 5/6). There was only one patient in the current group admitting to being a current smoker; however, there was a typical number of smokers (current and former, n ¼ 22/32) in the population. These patients, largely uncharacterized with respect to outcomes, represent the most advanced stage of PAD and CLTI. Previous reports have estimated 8% to 50% of patients with CLTI as having no conventional option for revascularization. 2,4,17,18 All of the patients in PROMISE I were deemed to have no other option beyond medical management, and the treating physicians estimated a high likelihood (>90%) of above-ankle amputation within 12 months.
This study is also the first published U.S. report of 12month outcomes of a specific purpose-built and exclusively percutaneous system to create perfusion of the pedal venous arch and achieve effective oxygenation of previously ischemic tissue in the foot. As such, this is an early report of the technique for performing these procedures, and interventionalists continue to gain knowledge and skill in creating this new route of tissue perfusion in the foot. The venous anatomy of the foot is distinctly different from the arterial vasculature, with parallel, paired veins for each major artery and multiple perforating branches and interconnections. The pedal venous system is specialized beyond the features described above, with a dominant and large plantar vein (ie, lateral plantar vein) that facilitates venous drainage during ambulation. The results of this initial study are encouraging, given the rapidly evolving understanding of how to most effectively create and manage the arterial inflow in the pedal veins. This EFS has been helpful in determining that communications can be made from tibial vessels; lysis of venous valves with a novel antegrade valvulotome allows reversal of flow in the venous system, and oxygenated blood flowing into the pedal venous loop does appear to improve wound healing and limb salvage in these patients who all have high presumed likelihood of limb loss. The results of this study also demonstrate the utility and effectiveness of a purpose-built pDVA system with dedicated arterio-venous crossing tools, stent grafts that vary in length and shape, and an antegrade valvulotome that allows for a fully percutaneous procedure. Improvements in perfusion effectiveness and durability would be expected as operators gain a better understanding of intervention and management of flow to effect maximal effective tissue perfusion.
There remain opportunities to further refine the procedure. Although the covered stent graft extends arterial pressure and flow to the level of the ankle, it is optimal to direct flow in the foot most effectively to the ischemic tissue (ie, nutritive flow) and away from proximal venous return (ie, non-nutritive flow). Without the presence of this long covered stent, the flow would preferentially be diverted through branches into venous channels in the calf and returned centrally prior to reaching the foot itself and the area of ischemia. Other challenges include the management of venous spasm in the setting of intervention and navigation of complex pedal venous system. The determination of the venous anatomy of the foot can be performed with ultrasound or venography and will likely aide in case planning.
The primary endpoint in this study of AFS provides both a measure of safety (mortality) and some insight into the effectiveness (limb salvage) of this procedure as a method to maintain the lower limb in these patients. The AFS of 70% is indicative of the potential for this method to save limbs that are felt to be heading to amputation. The results of this early study compare favorably with other reports regarding similar patient populations, although it is difficult to assure that the patients in other trials have as advanced disease as those noted here, as all of these patients had tissue loss as an inclusion criterion and the independent verification of no-option status may be more strict than the sitedetermined status most often reported in studies. 17,19 There is an additional critical aspect to providing care for these patients that entails provision of medical therapy and attentive wound care to assure that limb loss is not a function of a failed wound care rather than a failure of the revascularization strategy. The physicians managing patients in this study were experienced in the care of patients with CLTI, and the successful outcomes outlined here are clearly a reflection of detailed wound care delivered in concert with effective revascularization. A comprehensive multi-disciplinary approach to wound care is an essential aspect of the care pathway for these patients.
Secondary interventions in this patient population were relatively high when compared with conventional arterial revascularization. 4 Most reinterventions were directed at the arterial inflow to the arterio-venous conduit. All of these patients have advanced and diffuse atherosclerotic disease affecting the inflow donor artery of the pDVA procedure. This, coupled with efforts to preserve all existing patent tibial arteries and the increased flow demands of the arterio-venous circuit, make intraprocedural and postprocedural arterial inflow management critical. As present with surgically created arteriovenous connections, a rapid evolution is expected in response to the sudden change in physiologic circumstance in the newly arterialized veins, with both the main lateral plantar vein and newly opened branches evolving in the first days and weeks after the procedure. For these reasons, close follow-up of clinical symptoms and physical presentation are critical to judgments considering reintervention. As we continue to improve our understanding of the postprocedure physiological changes, the hope is that AFS rates continue to increase. It is reasonable to expect, as with other arterio-venous connections and procedures, that management of the maturation process to achieve maximum effectiveness and minimal ischemic complications will be more frequent than in conventional open and endovascular arterial reconstruction.
The limitations of this study include its small size, the lack of a control group, and the limitations of 12-month follow-up. This initial trial provides a positive signal for the potential for this technique. Observations from the study included high rates of wound healing and limb preservation after pDVA with the LimFlow System. There is significant likelihood for improved results as greater experience is obtained in all phases of the procedure, and this EFS clearly indicates a need for further evaluation of this procedure in a larger study with additional patients.

CONCLUSIONS
This initial trial provides a positive signal regarding the potential for the deep vein arterialization technique. Observations from this study included high rates of wound healing and limb preservation after pDVA with the LimFlow system. This EFS clearly indicates a need for further evaluation of this procedure in a larger study with additional patients.