Diabetes mellitus and peripheral artery disease

Abstract

Atherosclerotic peripheral artery disease (PAD), that is, arteriosclerosis obliterans, is pathologically rooted in atherosclerosis, similar to other cardiovascular diseases. In addition to smoking, hypertension, and dyslipidemia, diabetes mellitus is a major risk factor. People with diabetes mellitus have an elevated risk of developing PAD. PAD in turn increases the risk of diabetic foot ulcers and gangrene in the population. Rest pain, nonhealing ulcers, and gangrene associated with chronic ischemia are known as chronic limb-threatening ischemia (CLTI). This article gives an overview of the link between atherosclerotic PAD, particularly CLTI, and diabetes mellitus. First, the clinical impact of CLTI among patients with diabetes mellitus is presented. Second, its clinical features, including prognosis, comorbidity, occurrence, and seasonality, are mentioned. The clinical management of CLTI is also discussed. Diabetes mellitus has notable clinical impact on CLTI and vice versa. CLTI has different clinical features from those of other atherosclerotic cardiovascular diseases. Its clinical profile also differs between individuals with both diabetes mellitus and CLTI and general people with diabetes mellitus. There is considerable room for improvement in CLTI treatment and management. Clinical measures taken before revascularization, including CLTI risk assessment, prompt diagnosis, and expedited referral to vascular specialists, may enhance CLTI outcomes. Further research is warranted to obtain more evidence.

Introduction

Atherosclerotic peripheral artery disease (PAD), that is, arteriosclerosis obliterans, is pathologically rooted in atherosclerosis, similar to other cardiovascular diseases such as coronary artery disease (CAD) and cerebrovascular disease. It is well recognized that along with smoking, hypertension, and dyslipidemia, diabetes mellitus is a major risk factor. Patients with diabetes mellitus are at a higher risk of developing PAD.

PAD increases the risk of developing diabetic foot ulcers and gangrene. The clinical symptoms of PAD are graded as per the Fontaine classification as follows: no or atypical symptoms (stage I), intermittent claudication (stage II), rest pain (stage III), and nonhealing ulcers and gangrene (stage IV). Of these, rest pain, nonhealing ulcers, and gangrene that are associated with chronic ischemia are distinctively called chronic limb-threatening ischemia (CLTI). Therefore, CLTI is a clinical phenotype that is closely linked to diabetic foot disease.

However, physicians treating patients with diabetes mellitus rarely encounter CLTI. A domestic nationwide survey reported that the prevalence of lower limb gangrene and amputation in outpatients with diabetes was 0.9% and 0.4%, respectively, which was as low as that of other major diabetes-related complications, including renal failure on dialysis and blindness [1]. Its prevalence was far lower than that of myocardial infarction (4.9%) in the same population. In addition, diabetic foot disease includes not only the ischemic type but also the neuropathic one, and previous studies have suggested that cases associated with limb ischemia account for only about half of the cases of diabetic foot disease [2]. Taken together, it is rare for physicians to gather experience in the management of ischemic foot disease, i.e., CLTI, in clinical practice. In addition, clinical evidence remains insufficient, and limited data regarding the clinical features of CLTI are available. Accordingly, my collaborators and I conducted clinical studies to determine the clinical features of PAD, particularly CLTI.

Clinical impact of PAD in diabetes mellitus

Avoiding a decline in health-related quality of life (QOL) is a key goal in the management of diabetes mellitus. We surveyed thousands of Japanese patients with diabetes mellitus, and observed that clinical symptoms relevant to PAD, including intermittent claudication, foot ulcers, gangrene, and above-the-knee amputation, were strongly linked with reduced QOL [3]. The QOL decline related to those symptoms, especially tissue loss and above-the-knee amputation, was not less than that of other major diabetes-related complications, including symptomatic neuropathy, blindness, renal failure on dialysis, cardiac symptoms, and sequelae of stroke. Symptomatic PAD, particularly CLTI, is an important complication in individuals with diabetes mellitus.

Clinical influence of diabetes mellitus in PAD

The prognosis in patients with CLTI is extremely poor, even after they undergo timely revascularization. We previously conducted a retrospective single-center study to reveal the association between diabetes mellitus and prognosis in 278 patients who underwent endovascular therapy for CLTI. In this study, diabetes mellitus was significantly associated with an elevated risk of incident above-the-knee amputation; the adjusted hazard ratio (HR) was 3.1 (95% confidence interval, 1.3–7.6) [4]. Furthermore, in patients with both diabetes mellitus and CLTI, hemoglobin A1c levels had an independent association with an elevated risk of incident above-the-knee amputation; the adjusted HR was 1.3 (95% confidence interval, 1.1–1.7) per 1% increase. Our findings support the idea that optimal glycemic control is important for the improvement of limb-related outcomes in patients with both CLTI and diabetes mellitus, as recommended by clinical guidelines for PAD [5, 6]. Another finding of our study was that diabetes mellitus did not have any significant association with mortality risk (P = 0.72) [4]. This lack of a significant association between diabetes mellitus and mortality risk was corroborated by our retrospective multicenter studies including 995 patients with CLTI who underwent endovascular therapy [7] and 459 patients with CLTI who underwent endovascular therapy or surgical reconstruction [8] and our prospective multicenter study including 520 patients with CLTI who underwent endovascular therapy or surgical reconstruction [9] (all P > 0.05). It has been generally recognized that diabetes mellitus is a major risk factor for mortality in people with CAD [10, 11]. Atherosclerosis underlies both CLTI and CAD, explaining the similarities in their clinical manifestations. Despite that, our findings suggest that CLTI has different clinical aspects from CAD.

Patient profile

CLTI also differs from CAD in terms of patient profile. We compared the clinical profile between patients undergoing endovascular therapy for PAD and those undergoing percutaneous coronary intervention for CAD, using nationwide databases of percutaneous interventions for PAD and CAD between 2012 and 2017, where 41,718 cases undergoing endovascular therapy for CLTI and 516,134 cases undergoing coronary percutaneous intervention for acute coronary syndrome (ACS) were registered [12]. The prevalence of diabetes mellitus was 65.5% in CLTI cases, which was much higher than that in ACS cases, being 37.9%. The odds ratio of CLTI cases relative to ACS cases reached 3.12 (95% confidence interval, 3.05–3.19). Similarly, CLTI cases had a higher prevalence of advanced age (≥ 75 years) and end-stage renal disease than ACS cases, with odds ratios of 1.85 (95% confidence interval, 1.82–1.89) and 18.7 (95% confidence interval, 18.2–19.1), respectively. On the other hand, the prevalence of dyslipidemia was lower in CLTI cases, with an odds ratio of 0.45 (95% confidence interval, 0.44–0.46). The heterogeneity in patient profile between CLTI cases and ACS cases was so evident that patients with CLTI and those with ACS could be easily distinguished from each other by their basic patient profile, with the C-statistic as high as 0.833 (95% confidence interval, 0.831–0.836).

Furthermore, individuals with diabetes mellitus and comorbid CLTI are distinct in profile from general people with diabetes mellitus. The mean age of Japanese patients with diabetes mellitus in clinical settings is generally 55–65 years, with a mean duration of diabetes mellitus of 10–15 years and a mean body mass index of approximately 25 kg/m² [13, 14]. Patients with diabetes mellitus and comorbid CLTI usually exhibit different features; they are typically older (70–75 years on average) and have a longer duration of diabetes mellitus (20–25 years on average) and lower body mass index (approximately 22 kg/m² on average) [15, 16]. Having a long duration of diabetes mellitus suggests a high likelihood of other diabetes-related comorbidities. Our survey of patients with diabetes mellitus and comorbid CLTI presenting tissue loss revealed that 83% (95% confidence interval, 76–89%) of the patients had one or more advanced microangiopathies (insensateness at all examined podalic sites, proliferative retinopathy, and renal failure on dialysis) [16]. Furthermore, these advanced microangiopathies were more likely to be clustered in patients having a longer duration of diabetes mellitus (P = 0.004). However, a more important finding of this survey is that the prevalence of advanced microangiopathies reached as high as two-thirds even in the subgroup with a duration of diabetes mellitus shorter than 10 years [16]. One possible explanation would be that the duration of diabetes mellitus merely refers to the time from its clinical diagnosis. Most patients would not have undergone health checkups regularly and would be incidentally diagnosed as diabetes mellitus. Their condition would have remained untreated for many years until diagnosis. Consequently, they frequently had advanced microangiopathies despite their short duration of diabetes mellitus. Another possibility is that diabetes mellitus may not have been adequately controlled after its diagnosis, which would accelerate the progression of microangiopathies. There might be considerable room for improvement in the management of diabetes mellitus before CLTI onset. Earlier diagnosis of diabetes mellitus and more appropriate control after its diagnosis might lower the risk of developing CLTI.

Onset of CLTI

The Fontaine classification is one of the most familiar grading systems for PAD. However, it should be remembered that the classification system was originally developed to describe the severity of clinical symptoms, and not to grade in severity of ischemia. The severity of clinical symptoms does not always mean the severity of limb ischemia. That is why the Fontaine grading does not always accurately predict the linear progression of PAD; asymptomatic patients may skip intermittent claudication and directly display CLTI. Our multicenter study showed that 50% (95% confidence interval, 46–55%) of patients with CLTI did not have a history of intermittent claudication; they stayed asymptomatic, or at least free from typical symptoms before they developed CLTI [17]. Factors associated with absent claudication history included non-ambulatory status, diabetes mellitus, and end-stage renal disease requiring dialysis. The association between absent claudication history and non-ambulatory status is reasonable as intermittent claudication is a symptom that appears during walks. Non-ambulatory people living sedentary or bed-bound lives would seldom experience intermittent claudication, even when their lower limbs are severely ischemic. The association of diabetes mellitus and end-stage renal disease requiring dialysis with absent claudication history may be partially explained by distal atherosclerotic disease involvement [18]. Claudication often occurs in the thighs and calves, for which arterial lesions in suprapopliteal segments are usually responsible [19–21]. Individuals with diabetes mellitus and those with renal failure requiring dialysis typically suffer from arterial lesions localized to infrapopliteal segments, which will cause claudication less often. Furthermore, in individuals with diabetes mellitus, diabetic neuropathy impairs the perception of pain, which may also contribute to the absence of claudication [22].

The fact that CLTI is not always developed from intermittent claudication means that claudication is not a useful predictor for CLTI. People with no or atypical symptoms, especially those with non-ambulatory status, diabetes mellitus, and renal failure on dialysis, who may develop CLTI directly, should be objectively screened for limb ischemia. The most popular objective screening tool for lower extremity ischemia is the ankle brachial index (ABI). However, the ABI is elevated falsely when tibial arteries have severe calcification. Furthermore, the ABI cannot be used to assess the blood flow distal to the ankle, although diseased pedal arteries also cause foot ischemia. Arterial calcification and pedal artery disease are common in patients with diabetes mellitus and renal failure requiring dialysis [18]. False elevation of the ABI will not be rare in such patients. In our previous report, 22% of patients requiring revascularization for CLTI had an ABI > 0.90 [23]. Severe lower limb ischemia can be overlooked when using the ABI.

The drawbacks of the ABI can be overcome by using the toe-brachial index (TBI) [5, 6, 22]. The TBI can be used to detect more distal blood-flow impairment than the ABI. In addition, the TBI seems less likely to be falsely elevated, because the toe arteries undergo medial arterial calcification and incompressibility less frequently than tibial arteries [24, 25]. Lower limb ischemia can be identified from decreased TBI, even when the ABI is normal. Our retrospective study involving 869 patients with diabetes mellitus demonstrated that the TBI was more markedly decreased than the ABI in older patients, those having a longer duration of diabetes mellitus, and those with a lower body mass index [26]. Our finding suggests a potential risk of undiagnosed lower limb ischemia in these subgroups. Interestingly, the clinical features of these subgroups, namely, older age, longer duration of diabetes mellitus, and lower body mass index, are the same as those of patients with diabetes mellitus and comorbid CLTI versus a general population with diabetes mellitus [16], as discussed in the “Patient profile” section. The clinical features could be easy indicators of CLTI risk.

Another important issue related to CLTI onset would be injury, which triggers unhealed ischemic tissue loss. Our study analyzed data from 450 patients with CLTI and tissue loss enrolled in a multicenter prospective registry [27]. While poorly fitting shoes and mechanical trauma were the primary causes of injury leading to non-healing tissue loss, as much as 52.0% (95% confidence interval, 47.3–56.6%) of the population lacked a discernible preceding injury. Some may criticize the fact that most of our data came from self-reports and therefore were at risk of inaccuracy, especially in patients suffering neuropathy, who are more likely to be unaware of injuries. However, our subsequent analysis confirmed that the absence of clear precipitating injuries was inversely associated with neuropathy. Patients with intact protective sensation often lacked a distinct initial injury leading to tissue loss. Although avoiding walking barefoot and wearing well-fitting shoes are important strategies to avoid pedal trauma [5, 28], our findings support the idea that avoiding pedal trauma does not perfectly prevent ischemic feet from developing unhealing tissue loss.

Seasonality of CLTI

The incidence, severity, and outcomes of CLTI exhibit seasonal fluctuations. In our previous study of 1568 CLTI patients treated with endovascular therapy, we identified seasonal variation: lowest incidence during summer–autumn and highest during winter–spring [29]. The severity of symptoms (evaluated using the Rutherford classification) and the risk for incident above-the-ankle amputation also exhibited a seasonal pattern, showing a peak in winter and a trough in summer. Similar findings were observed in patients with ACS [30]. However, further analysis using nationwide databases of percutaneous interventions showed that cases with CLTI and those with ACS exhibited different presentation patterns in the following two aspects [31]. First, the seasonal change was more pronounced in cases with CLTI than in those with ACS (P < 0.001); the peak-to-trough ratio in volumes in CLTI and ACS cases were 1.75 (95% confidence interval, 1.71–1.80) and 1.21 (95% confidence interval, 1.20–1.22), respectively. This finding indicates that seasonal fluctuations in hospital bed demand and medical staff workload are more pronounced for CLTI compared to ACS. Second, the appearance of the peak in cases with CLTI delayed by 1.37 (95% confidence interval, 1.25–1.49) months compared to that in those with ACS (February–March versus January–February). This delay may simply come from the etiological difference between CLTI, a chronic disease, and ACS, an acute disease. Even if the two diseases were to have the same time of onset, CLTI could lag behind in revascularization time because of this difference in etiology. However, CLTI is usually defined as ischemic rest pain or tissue loss that lasts for at least two weeks. A lag of 1.37 months seems long for its definition. Further studies are required to reveal the mechanisms underlying these lags.

Delay in referral to vascular centers

CLTI ultimately requires timely revascularization for limb salvage. To achieve timely revascularization, patients who develop CLTI should be promptly referred to a vascular center. We previously analyzed the database of a multicenter prospective study where 428 patients suffering from CLTI with ischemic wounds were registered between January 2012 and March 2013. The time from onset of tissue loss to referral to a vascular center exceeded one month in 58.2% (95% confidence interval, 53.2–63.1%) of the patients, and exceeded three months in 15.9% (95% confidence interval, 12.4%–19.4%) [32]. No clinical features, including diabetes mellitus, had any significant association with the time to referral. Furthermore, the time to referral had a positive association with wound severity and a negative association with survival free from wound recurrence and above-the-ankle amputation after revascularization. In other words, patients with a longer time to referral had more severe tissue loss and a dimmer prognosis. If ischemic tissue loss is left untreated, it would progressively deteriorate and thus counteract the beneficial effects of revascularization. If patients were promptly referred to a vascular center, they might have a better prognosis. Thereafter, we re-examined the time to referral in another population with CLTI presenting with tissue loss between October 2017 and June 2020 [27]. Consequently, patients referred to vascular centers more than 3 months after tissue-loss occurrence accounted for 16.8% (95% confidence interval, 12.8%–20.7%) of the population, indicating that delayed referral did not drastically decrease during recent years. Although the management of CLTI has drastically improved in recent years [5, 6, 22], there is still considerable room for improvement in the cooperation between home doctors and vascular specialists.

Lack of awareness about the disease may be a significant contributor to delayed referrals, and it unfortunately remains a practical and unsolved issue [33, 34]. This would be true even in the field of the management of diabetes mellitus. Our previous survey demonstrated that only 31% (95% confidence interval, 22–39%) of patients with diabetes mellitus who developed CLTI with tissue loss underwent ABI screening before its onset [16]. This finding suggests that lower limb ischemia is not frequently evaluated, even in populations at a high risk of CLTI.

Conclusions

The clinical impact of PAD, particularly CLTI, was overviewed. A bidirectional relationship exists between diabetes mellitus and CLTI, with each condition having a great clinical impact. Clinical features differ considerably between patients with CLTI and those with other atherosclerotic cardiovascular diseases. They are also different between people with diabetes mellitus and comorbid CLTI and general people with diabetes mellitus. There appears to be considerable room for improvement in CLTI management. Clinical measures taken before revascularization, including CLTI risk assessment, prompt diagnosis, and expedited referral to vascular specialists, may enhance CLTI outcomes. Further research is warranted to obtain more evidence.

Author contribution

Mitsuyoshi Takahara developed the idea for this article, performed the literature search, and wrote the review.

Funding

This study was supported by the Japan Society for the Promotion of Science (KAKENHI, Grant Number JP22K08623), which covered the English editing fees.

Data availability

All data are provided in the manuscript.

Declarations

Conflict of interest

The author declares that that he has no conflict of interest.

Ethics approval and consent to participate

Not applicable.

Human or animal rights

This article does not include data collected from any studies involving human or animal subjects.