Diabetes mellitus (DM) is considered a risk equivalent to atherosclerotic cardiovascular disease, including peripheral artery disease (PAD). Prior studies evaluating standard medications that reduce blood glucose failed to show cardiovascular benefits or improved survival. The multicenter Canagliflozin Cardiovascular Assessment Study (CANVAS) showed significant reduction in death from cardiovascular causes, nonfatal myocardial infarction (MI) and nonfatal stroke in patients with Type 2 diabetes at increased risk of cardiovascular disease when treated with canagliflozin. However, canagliflozin was associated with a significant doubling in the risk of toe amputations, a new finding for which the mechanism is unknown. Therefore, a translational overview of the microcirculation in the context of DM and the development of PAD is needed. Microcirculatory concerns in DM are not new, and there is robust literature regarding amputations and retinal changes in people with diabetes. Not until recently has more specialized testing become available to evaluate possible etiologies of microvascular disease. Virchow’s triad describes three broad categories of potential mechanisms for vascular thrombosis.
Introduction: Diabetes and PAD
PAD is currently defined by exertional non-joint-related leg symptoms (e.g., intermittent claudication) and an abnormal resting ankle-brachial index (ABI) of ≤0.90, an abnormal exercise ABI or an abnormal toe-brachial index (TBI) of ≤0.70.1 However, historically PAD has been variably defined in the literature by intermittent claudication or resting ABI ≤0.90. In 2010, PAD affected an estimated 202 million people worldwide—an estimated rate of increase of 23.5% overall, with the most rapid increase in lower-income countries (28.7%).2 Not only does PAD worsen quality of life, but it also increases the 10-year risk of cardiovascular death nearly sixfold.3 While cigarette smoking is associated with more proximal (e.g., aortoiliac) PAD, DM results in more infrapopliteal PAD.4 Infrapopliteal PAD is characterized by more complex, calcific and diffuse atherosclerosis and often involves chronic total occlusions of the runoff arteries. The severe, diffuse nature of infrapopliteal PAD also leads to an increased association with critical limb ischemia (CLI), with up to 50% of CLI patients having concurrent DM. In addition to the increased cardiovascular mortality, 30% of CLI patients will undergo major amputation. Despite advances in surgical and percutaneous vascular interventions, patency rates remain lower than with aortoiliac and iliofemoral PAD interventions.5
DM is considered a risk equivalent to atherosclerotic cardiovascular disease, but it is also one of the most powerful predictors of PAD. Prior studies have demonstrated odds ratios of 1.89 to 4.05.2 In population studies from the Framingham, MESA, Rotterdam and NHANES cohorts, the odds ratios of PAD in DM patients were 1.89-2.71, nearly as high as the odds ratios for cigarette smoking.6-9 Furthermore, in a meta-analysis of community-based studies, the overall odds ratio of PAD in DM was 1.68, second only to current or prior cigarette smoking.2 When comparing diabetic with non-diabetic patients, over 20% of patients with elevated blood glucose have abnormal ankle-brachial index (ABI) versus only 9.5% of patients with impaired glucose tolerance and 7.0% of patients with normal glucose tolerance.10,11 DM severity may also be associated with an increased risk of PAD. The United Kingdom Prospective Diabetes Study reported that for every 1% increase in HbA1c, there was a >25% increase in PAD.12 The prevalence of symptomatic PAD in patients with diabetes approaches 10%, with CLI developing in 1.3%.13 Furthermore, the impact of PAD extends beyond limb amputation and quality of life. PAD patients are at a significantly increased risk of MI, approaching 30% in those with prior limb revascularization, as well as a nearly fourfold increase in those with acute limb ischemia.14
The pathophysiology of PAD involves the imbalance of the circulatory supply of nutrients and oxygen to the metabolic demands of the skeletal muscle. In addition to the abnormalities induced in the skeletal muscle from chronic ischemia (e.g., switch to anaerobic metabolism and decline in exercise tolerance), factors regulating blood flow to the limb are key to the pathobiology. Flow-limiting lesions seen in PAD are appreciable on angiographic imaging, but the microvascular sequela of accentuated vasoconstriction, impaired vasodilation and abnormal rheology equally contribute to claudication, ischemia and risk of limb loss.15
Pathophysiologic Associations between DM and PAD
The Hagen-Poiseuille equation describes the rate of blood flow to organs (Figure 1). The most clinically important factors affecting blood flow are pressure gradient, radius of the residual lumen and blood viscosity. Virchow’s triad describes three major factors closely related to vascular thrombosis: 1) vascular endothelial injury, 2) changes in the blood composition and 3) blood flow alterations.
1. Endothelial Injury
The endothelial lining of blood vessels is a highly active, single-cellular layer that performs multiple functions. It mediates the interaction between blood cells and the vascular wall, which affects modulation of blood flow, nutrient delivery, coagulation and thrombosis, and leukocyte diapedesis.16 Most patients with DM (>80%) have a history of hypertension,17,18,19 and the increased shear stress seen in hypertensive states is a major contributor to long-term endothelial injury. The endothelial injury from mechanical forces sets the stage for increased microvascular dysfunction and thrombosis.
One of the more important chemicals released by the endothelium for vascular homeostasis is nitric oxide (NO). NO has antiplatelet activity and indirectly reduces vascular smooth muscle cell movement and proliferation. Elevated glucose decreases the bioavailability of NO and decreases prostacyclin (PGI 2) while increasing synthesis of vasoconstricting prostanoids and endothelin (ET-1) via multiple mechanisms.20 The loss of NO and PGI 2 leads to impaired vasodilation. The shift in profile of the vasoactive substances results in a net vasoconstrictor effect, which reduces flow in the microvasculature, impairs tissue perfusion and increases risk for thrombosis and limb loss.15
Moreover, reduced NO levels cause increased inflammation,21 which leads to increased plasminogen activator inhibitor-1 (PAI-1) levels, resulting in decreased fibrinolytic activity from a reduction in plasmin levels (Figure 2).22,23 Hyperglycemia also leads to increased expression of von Willebrand factor and elevated fibrinogen and PAI-1 levels, which further contribute to vascular injury-mediated thrombosis.
Lastly, impeding the anticoagulant activity of antithrombin-III in response to oxidative stress from DM can lead to thrombin hyperactivity. Thrombin has proinflammatory effects that include induction of nuclear factor-kappa B (NF-kB). NF-kB increases transcription of proinflammatory mediators: cytokines (tumor necrosis factor-alpha) and interleukins (IL-1β, IL-6 and IL-8), which further drive the shift toward thrombus formation and subsequent thrombotic events.24,25,26
2. Changes in Blood Composition
In patients with poor glycemic control, plasma viscosity is increased partly due to structural alterations of the red blood cells (RBCs). In vitro studies have shown changes in RBC deformability, presumably due to membrane protein glycation, that subsequently result in increased whole blood viscosity (WBV).27,28 RBCs normally shift into an elliptical shape, align in direction of flow and move with rapid velocity, which allows blood flow to be maintained even with a relatively high hematocrit that approaches 50%. At higher flow rates, RBCs have more fluid properties, but as blood flow is reduced, blood viscosity significantly increases (Figure 2).29 RBC deformability becomes crucial in the microcirculation where the capillary is 4-9 µm and RBC size averages 8 µm. Thus, the deformation properties of RBCs are critical to allow them to pass through the capillary and maintain blood flow, oxygen and nutrient supply.30
Other important high molecular weight particles are fibrinogen, IgM and alpha 2 macroglobulin. The relative effect on aggregation is highest with fibrinogen and less with alpha 2 macroglobulin, haptoglobin and albumin; platelets and white blood cells have little impact on blood viscosity.31 Fibrinogen, factor VIIa and elevated plasminogen activator inhibitor-1 are all significant contributors to the hypercoagulable state of DM, and elevated fibrinogen further increases blood viscosity as well.32
3. Blood Flow Alterations
Blood viscosity has long been associated with vascular thrombosis. In the Edinburgh Artery Study, a five-year study of 4,860 healthy men ages 45-59, 20% of individuals with the highest blood viscosity were matched to the 20% with the lowest blood viscosity. Those in the lowest blood viscosity group had fewer cardiovascular events than those in the highest group, resulting upon a 3.2-fold increased risk.33 DM is accompanied by increases in blood viscosity, with nearly a twofold increase in measured WBV in diabetic versus non-diabetic patients.34,35 Mathematical models have suggested that for every 100 mg/dL increase in blood glucose concentration, there is a linear increase in blood and plasma flow times corresponding to an increase in WBV.29
As blood viscosity increases, blood flow slows and consequently increases the tendency for atherosclerosis (Figure 3).36 The physiologic responses to increased WBV are based upon the Hagen-Poiseuille principle and thus include increases in blood pressure or vascular dilatation. Increased radial forces produce more stretch of smooth muscle cells in the tunica media and, in time, the increased stimulation to the smooth muscle cells will lead to protective hypertrophy. Increased radial forces can also cause endothelial cell activation and damage. Cellular leakage, inflammation and apoptosis of endothelial cells exposing tissue factor, collagen and other cellular components lead to a prothrombotic state, demonstrating the interplay between endothelial injury and vascular thrombosis (Figure 4).37
Important Clinical Considerations
Microcirculatory concerns in DM are not a novel concept, with amputations and retinal changes having been previously studied. Only in recent times has more specialized testing become available to evaluate possible etiologies of microvascular disease.
One of the early vascular papers by Bailey et al. identified a patient’s preoperative hemoglobin as a predictor of outcome of diabetic amputations. Fifty-nine consecutive diabetes patients required local amputations in the foot and survived for at least one month. Forty patients had digital amputations, and 19 had a metatarsal or transmetatarsal as their primary surgery. In patients who healed after surgery (18 amputations), lower initial hemoglobin levels (<12 g/dl) were beneficial. Thus, they concluded that the effects of increased viscosity from higher hemoglobin concentrations could be the major reason. A 2.7-g/dl increase in hemoglobin would increase blood viscosity about 25% at high shear rates and even more at lower shear rates.37
Recently, the CANVAS trial investigated the benefit of the sodium-glucose cotransporter 2 (SGLT2) inhibitor canagliflozin on cardiovascular outcomes. Prior studies evaluating standard medications that reduce blood glucose failed to show cardiovascular benefits or improved survival. SGLT2 inhibitors have been shown to reduce hyperglycemia, blood pressure and body weight. CANVAS was a randomized, double-blind, placebo-controlled multicenter study involving 10,142 patients with Type 2 diabetes and a history of (or at high risk for) cardiovascular disease that was designed to assess cardiovascular safety and efficacy of canagliflozin. At a mean follow-up of 188 weeks, there was a significant reduction in the risk of cardiovascular death, nonfatal myocardial infarction and nonfatal stroke.38 However, canagliflozin was associated with a significant doubling in the risk of amputations (6.3 versus 3.4 participants with amputation per 1,000 patient-years; HR 1.97), with 71% of affected participants having their highest amputation at the level of the toe or metatarsal. Recent evidence has also found increased risk for reduced peripheral flow to toes with loop diuretics and canagliflozin (Figure 5). The increased rate of amputation is a new finding for which the mechanism is not known. However, understanding the pathophysiology of PAD and the impact of underlying DM with its effects upon fluid dynamics (Hagen-Poiseuille principle) and vascular thrombosis (Virchow’s triad) may help us understand the potential mechanisms for this observed finding.
Three important factors relating to enhanced peripheral atherothrombotic risk in diabetes mellitus are vessel wall injury, hypercoagulable state and stasis/impaired blood flow. Indeed, diabetes patients are at very high risk for developing PAD. Current consensus statements advocate routine screening for PAD in all diabetes patients over age 50 as well as diabetes patients younger than 50 with an additional atherosclerotic risk factor. Patients demonstrating reduced peripheral vascular circulation would potentially be at higher risk for atherothrombotic events. Given the potential for increased risk for lower extremity amputation, careful discussion about the risk and benefits of SGLT2 inhibitor treatment is reasonable prior to starting therapy. Future studies would include identifying a definite relationship between canagliflozin use and lower extremity amputations and other potential therapies that might lower amputation risk through improved viscosity, organ flow and endothelial function.39,40
Access additional resources and practical information to enhance the care and treatment of your diabetes patients.
About our experts:
Charles Lin, MD<sup>1</sup>, Adam Obaidi, MD<sup>2</sup>, James Blair<sup>3</sup>, Son Pham, MD<sup>4</sup> and Robert Chilton, DO<sup>4</sup>
<sup>1</sup>Interventional Cardiology, University of Texas Health Science Center at San Antonio, San Antonio, Texas; <sup>2</sup>Department of Cardiology, San Antonio Military Medical Center, San Antonio, Texas; <sup>3</sup>Long School of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Texas; <sup>4</sup>Interventional Cardiology, Audie L. Murphy Memorial Veterans Affairs Hospital, San Antonio, Texas