Non-severe ischaemia is characterised by findings indicative of PAD, such as absent pulses, monophasic or absent pedal Doppler waveforms, ankle pressure 100 mmHg, or toe pressure 60 mmHg, but not meeting the threshold for severe ischaemia [8, 14]. In this situation, vascular consultation should form part of comprehensive patient management. Studies report limb salvage rates of 80% to 85% and ulcer healing rates of over 60% at 12 months following revascularisation [18], indicating benefits in appropriately selected patients with diabetes and hemodynamically significant PAD. However, no specific timing for intervention is provided in these guidelines, and a period of observation of ulcer progression over several weeks will often inform the clinical need for revascularisation. Severe ischaemia is defined by an ABI 0.4, ankle pressure 50 mmHg, toe pressure 30 mmHg, with monophasic or absent pedal Doppler waveforms, and an urgent vascular opinion should be sought to consider revascularisation [8].

Delaying revascularisation for more than two weeks in DM increases the risk of limb loss [19], and a shorter time to revascularisation (8 weeks) is associated with a higher probability of DFU healing [20]. Moreover, an analysis of 11,398 patients undergoing non-elective revascularisation for CLTI across the UK also suggests that national-level service factors impact the ability to revascularise early, such as the day of the week the patient is admitted and the case volume of the admitting hospital [21]; this reiterates the need for prompt referral to a specialist unit upon recognition of severe ischaemia in DM. As such, UK national best practice suggests a target time from referral to revascularisation of 5 days and 14 days for in-patients and out-patients, respectively, although this remains aspirational in many centres [22].

In patients who present with foot infection or gangrene, the priority is urgent assessment and consideration of immediate drainage of infection, since a delay in treatment can result in rapid tissue destruction and life-threatening sepsis [17]. For individuals with an abscess or infection of a deep foot compartment requiring urgent drainage, or where gangrene must be removed to control infection, immediate surgical intervention should be prioritised [23], in conjunction with broad-spectrum antibiotic therapy, tailored to tissue culture results. Once sepsis is controlled and the patient stabilises, prompt revascularisation should then be considered within a few days.

This population is at higher risk of poor outcomes due to the combined effects of diabetes-related vascular complications and progressive PAD. Three prognostic models have been proposed for predicting survival benefits after revascularisation in those with CLTI: Bypass versus Angioplasty in Severe Ischaemia of the Leg (BASIL) Weibull [24]; PREVENT-III [25]; and FinVasc [26] (Table 2). However, their performance across the majority of validation studies suggests these models are less appropriate on an individual level, which is reflected in the Global Vascular Guidelines, which recommend that no specific tool and model can be recommended in preference to others [8].

Ultimately, patients may not be suitable for revascularisation when there is no realistic prospect of wound healing, when major amputation is unavoidable, or if achieving a functional foot is unlikely due to limited life expectancy. In these cases, revascularisation may offer no benefit and could cause harm, especially in patients with high peri-procedural risk due to advanced disease and multiple co-morbidities. Specific groups that may be unsuitable include those who are frail, have poor functional status, are bed-bound, or have extensive tissue loss that makes the foot unsalvageable. Decisions should be personalised, considering patient preferences and involving the multidisciplinary team [17].

The Global Limb Anatomic Staging System (GLASS) provides a mechanism through which disease severity can be stratified to describe the likelihood of success of revascularisation (Fig. 1; Ref. [8, 27]). In general, disease patterns with a low GLASS score, usually short-segment stenosis or occlusions with otherwise patent vessels, can be managed with endovascular solutions. On the other hand, a higher GLASS score representing anatomically complex disease, such as multiple or long-segment occlusions, requires open or hybrid procedures to achieve in-line flow with acceptable rates of patency [8]. Unfortunately, disease patterns frequently fall into an indeterminate range in which decision-making is more challenging and may be influenced by locally available expertise.

Fig. 1.

Simplified illustrative example of the GLASS classification system [8] using disease of the superficial femoral artery (SFA). GLASS, Global Limb Anatomic Staging System. Created in BioRender ([27]; https://BioRender.com/t95n377).

An alternative suggestion for using the anatomy of the PAD to guide the treatment modality is the angiosome theory. Angiosomes are an anatomical unit of tissue supplied by a specific source artery and drained by corresponding veins. In the ankle and foot, angiosomes can be topographically mapped into six territories [28] (Fig. 2; Ref. [29]), acting as anatomical targets for revascularisation. Direct revascularisation, implying revascularisation into an artery supplying the angiosomal area of tissue loss, may provide some benefit over indirect revascularisation (targeting an artery not directly supplying areas of tissue loss). A systematic review has found mixed results in this domain, with improved one-year healing rates for direct revascularisation through endovascular interventions (66.9–96.4% vs 54.6–85.7%) and open bypass (81–90.9% vs 41–68.5%). However, while most studies reported improved outcomes with direct revascularisation, results are heterogeneous, and the certainty of the evidence was found to be very low. Similarly, inconclusive results were found for amputation-free survival, overall survival, and limb salvage [30].

Fig. 2.

Anatomical distribution of the angiosomes of the foot and lower leg. Created in BioRender ([29]; https://BioRender.com/l73h427).

Surgical bypass can be accomplished by using an autologous vein or prosthetic material. The presence of a suitable venous conduit is a favourable factor when planning to undertake a bypass over an endovascular approach. The ipsilateral great saphenous vein (GSV) is most commonly used, with studies highlighting patency up to 80% at 5 years [31]. Other native vessels may be considered, including contralateral GSV, short saphenous vein, or upper limb veins. A Cochrane review found moderate-quality evidence favouring an improved long-term (5-year) primary patency for autologous vein grafts when compared to prosthetic materials for above-knee bypasses. However, for below-knee bypasses, there remains limited evidence that directly compares these conduit choices [32].

Multiple studies have attempted to address whether surgical bypass or angioplasty is better. The BASIL trial randomised 452 participants with CLTI to either a surgery-first approach using vein, or an endovascular-first revascularisation strategy, for infrainguinal disease. 58% of participants had DM. While 30-day outcomes were similar, surgery-first was associated with improved amputation-free survival beyond 2 years [33]. This was reflected in a cost-effectiveness analysis, demonstrating an increased cost of a surgery-first strategy ( $\$$ 34,378 vs $\$$ 25,909) at 1 year, which diminished at 3 years ( $\$$ 45,322 vs $\$$ 39,801), largely attributed to the need for reintervention. A surgery-first strategy provided a negligible benefit of increasing quality-adjusted life-years by just 11 days when studied for 3 years post-intervention [34].

The BEST-CLI trial explored whether angioplasty or open surgery was superior in managing infrainguinal disease, measured by a composite primary outcome of major adverse limb events or death. There was a high proportion of participants with diabetes (69.3% of 1830). Participants with and without suitable autologous veins were in separate cohorts. There was diversity of anatomical levels in which bypasses were performed, including above-knee popliteal (15.3%), below-knee popliteal (24.7%), and tibial and pedal vessel (51.4%) anastomoses, reflecting the diversity of disease patterns often seen in DM. This identified a reduction in the composite outcome in the surgical group in the presence of a suitable vein (42.6% vs 57.4%; hazard ratio [HR] 0.68; p 0.001). Furthermore, major reinterventions, defined by the need for a new or revised bypass graft, thrombectomy or thrombolysis, were higher in the endovascular-first group compared to the surgery-first group (23.5 % vs 9.2%). Major lower limb amputations were lower with a surgery-first strategy (10.4% vs 14.9%) [35]. Moreover, sub-analysis of the cohort with DM with a suitable vein demonstrated superiority of surgery-first in preventing a major adverse limb event or death (HR 0.57; 95% confidence interval [CI] 0.10–0.78). However, if no suitable venous conduit was available, outcomes for endovascular intervention and bypass were equivalent (HR 1.03; 95% CI 0.70–1.53) [30].

Infrapopliteal disease represents a challenging pattern of disease strongly associated with comorbid DM. The BASIL-2 trial compared a bypass-first strategy against a balloon angioplasty-first strategy for the management of infrapopliteal disease in 345 patients, of which almost 70% had DM. This identified a significant improvement in amputation-free survival (47% vs 37%; HR 1.35; p = 0.037), largely driven by reduced mortality, in the endovascular-first group. However, challenges in patient recruitment and the prolongation of follow-up time led to the capture of mortality that was possibly not related to the intervention. Indeed, amputation rates and quality of life scores were similar. Further, TBI, WIfI scores and re-intervention rates were lower in the bypass-first cohort, suggesting that surgical intervention may provide improvements in peripheral perfusion and reduction in the need for subsequent admission or intervention [36]. While this trial generally supports an endovascular-first strategy, there may be a role for a bypass-first strategy in a carefully selected, low-risk subset of patients with severe ischaemia.

Whilst an aortoiliac pattern of disease is less associated with DM, many individuals also smoke tobacco and suffer from hyperlipidaemia, resulting in a less common but important cohort whose ischaemia is compounded by aortoiliac disease. The anatomic distribution of such disease is usually described using the TransAtlantic Inter-Society Consensus (TASC) classification. Traditionally, TASC C and D patterns were managed by open surgery. However, technological advances mean that many individuals are now treated with endovascular techniques. While open surgery has been associated with improved patency rates and limb salvage, the morbidity associated with open revascularisation of these lesions is high, leading many to prefer an endovascular first strategy [37]. No randomised control studies have yet been published in this domain; the Open revascularisation in severe oCClusive aorto-iliac disease (EVOCC) trial is recruiting participants to address this knowledge gap [38].

In the absence of conclusive randomised trials for either a surgery-first or an endovascular-first strategy, the ultimate decision to revascularise, and thus which method to use, relies upon a multi-disciplinary team approach with input from surgeons and interventionalists who can offer either treatment modality. Moreover, the ability to offer the optimal current method of patient selection, local expertise, availability of resources, and patient preference will also influence this decision-making.