Anatomical Distribution Patterns of Peripheral Arterial Disease in the Upper Extremities According to Patient Characteristics: A Retrospective Cohort Study

Introduction

Peripheral arterial disease (PAD) includes various noncoronary arterial syndromes wherein the blood flow to the head, visceral organs, and limbs is restricted. Atherosclerosis is the most common disease of the aorta and its ramified arteries.^1,2 PAD is considered chronic when it progresses to an advanced atherosclerotic form. It has several peripheral clinical presentations, including intermittent claudication, resting pain, ischemic ulcers, gangrene, and functional impairment.^3,4

The limb-related manifestations of PAD are devastating. Furthermore, PAD can often manifest in any component of the vascular system, such as the cerebrovascular system (manifesting as ischemic stroke) and coronary arteries (manifesting as myocardial infarction).^5–8 The quality of life of patients with PAD, particularly in terms of cardiovascular morbidity and mortality, may be significantly impaired.^3,4,9,10

Large-artery atherosclerosis is the primary cause of lower extremity ischemia, whereas small-artery occlusive disease of the palmar and digital arteries remains the primary cause of upper extremity ischemia. PAD is a “silent” disease that manifests in only 5% of the patients with ischemic upper extremities.^1,2,11

The prevalence of PAD varies according to the ethnicity and geographic location, with a higher prevalence reported in Australian, African-American, and Hispanic populations.^12,13 Between 2000 and 2010, the global prevalence of PAD increased by 5% (approximately 200 million patients).¹⁴ The increase in prevalence was more commonly observed in low- and middle-income countries than in high-income countries, with the highest prevalence observed in the Southeast Asian and Western Pacific regions.^14–16

Tobacco use, diabetes mellitus (DM), aging, dyslipidemia, hypertension (HTN), obesity, and chronic kidney disease (CKD) are risk factors for PAD.^12,17,18 A single-center study found that HTN, DM, and obesity were the most prevalent risk factors for PAD in patients from Jeddah (Saudi Arabia).¹⁷

The 2017 European Society of Cardiology Guidelines on the diagnosis and treatment of patients with PAD recommend computed tomography angiography (CTA) and magnetic resonance angiography as the imaging modalities for assessing the structural morphology and perfusion in the regions affected by PAD.¹⁹ The effectiveness of these tools lies in their ability to both distinguish between anterograde and retrograde perfusion and estimate stenosis severity.¹⁹ Nonetheless, digital subtraction angiography remains the imaging modality of choice and is primarily used in conjunction with endovascular therapy.¹⁹

Risk factor control and the best medical therapy were initially advocated for controlling symptomatic upper extremity PAD and reducing its pertinent risks.^18,19 Revascularization is recommended for significant stenosis or occlusion; it can be achieved through both open surgery and endovascular procedures. Treatment decisions should always be made in consultation with a healthcare provider and should take into consideration each patient’s individual condition.^20–23

To the best of our knowledge, no studies have investigated the patterns of PAD in the upper extremities, the circulation in the upper extremity segments affected by PAD, and how specific factors affect the degree of disease severity and the extent of the affected area.

The primary purpose of this cohort study was to provide the best available evidence on the topographical distribution patterns of PAD in the upper extremities in terms of the variable heritable patient characteristics and associated comorbidities. An improved understanding of the underlying clinical relationships could considerably improve screening protocols, thereby leading to early disease detection, chronicity reduction, and a better understanding of the silent complications of PAD in the upper extremities. This in turn could allow practitioners to save diseased upper limbs from unnecessary amputation.

Materials and Methods

This study was performed in accordance with the “Strengthening the Reporting of Observational Studies in Epidemiology” guidelines for cohort studies.

Study Setting and Design

This retrospective chart review was performed at the Radiology Department of the King Fahd Hospital of the University (KFHU; Khobar City, Saudi Arabia) between January 2015 and January 2021. KFHU is one of the largest academic hospitals in Al Khobar City (Eastern Province of Saudi Arabia), and its Radiology Department has been operational since 1981.

Study Population and Participants

The sample size was determined using G*Power v3.1.9.7. (RRID: SCR_013726, available at http://www.gpower.hhu.de/) to ensure sufficient statistical power (1–β = 0.80) and to prioritize outcomes based on parameters for independent means (α error: 0.05, effect size: 0.5).

The inclusion criteria were as follows: age >18 years, presentation to our health facility under an initial clinical suspicion of upper extremity PAD, and presumptive diagnosis of PAD confirmed using upper limb computed tomography (CT). The CT examinations were performed at the Radiology Department of KFHU using a commercially available 64-slice CT scanner (SOMATOM Definition AS, Siemens™, Germany).

The exclusion criteria were as follows: presence of nonatherosclerotic lesions (eg, dissection; n = 8), iatrogenic etiologies (including previous vascular interventions or presence of surgically created arteriovenous fistulas for dialysis vascular access [n = 13] and upper extremity amputations [n = 10]), age <18 years (n = 10), and poorly documented or incomplete data (n = 7). The participant selection flowchart is shown in Figure 1. Data from other imaging modalities for PAD, such as ultrasound imaging and magnetic resonance imaging, were not included owing to the study scope.

Figure 1 Patient selection flowchart displays the included and excluded cases as well as the final sample size.

Study Variables

The collected data were categorized into patient characteristics (independent variables) and segmental artery involvement patterns (dependent variables). Data on patient characteristics comprised basic demographic information (eg, age and sex) and factors suspected to be associated with PAD (eg, risk factors and metabolic comorbidities).

Data on segmental artery involvement patterns comprised the sites of the affected arterial segments (ie, the name of the artery and its segmental localization [proximal, middle, or distal]) and the severity of stenosis (mild = 30%, moderate = 30–70%, severe = 70–99%, and total occlusion = 100%). This classification system for stenosis and occlusion was derived not only from existing literature in endovascular and interventional radiology but also from a 2011 diagnostic study.²⁴ Napoli et al graded PAD using data from a similar 64-section CT angiography machine.²⁴ Patients with a current or previous smoking history were defined as “smokers”.

Data Collection

Vascular and interventional radiology consultants (AMA, KSZ, and IAG) reviewed the radiological findings and reported on the presence of stenotic atherosclerotic lesions. These lesions were classified according to the following three anatomical partitions after referring to the methodology of similar previous studies on lower extremity PAD.^25–27

1. Proximal:

Shoulder (proximal quarter): subclavian and axillary arteries.

2. Intermediate:

Arm (second quarter): superficial and deep brachial arteries.

Forearm (third quarter): ulnar and radial arteries.

3. Distal:

Hand (distal quarter): superficial and deep palmar arches and digital arteries.

Statistical Analysis

Statistical analyses were performed using SPSS version 23.0 (IBM Corp., Armonk, NY, USA). Patient characteristics recorded during the first radiological examination were subjected to a descriptive analysis. Categorical variables were expressed as percentages; these were compared using the chi-square test. Continuous variables were expressed as mean ± standard deviation; these were compared using an unpaired t-test. The hypotheses were tested at a significance level of 5%.

A univariate logistic regression analysis was performed to evaluate the relationships between patient characteristics (independent variables) and the anatomical distribution patterns of PAD (dependent variables). Significant variables from the univariate analysis were entered into a parsimonious binary forward stepwise logistic regression model to identify the independent predictors of PAD distribution. Nonsignificant variables were excluded from the analysis. A multivariate logistic regression analysis was also performed to assess the correlations between PAD distribution in both upper limbs and patient characteristics. The confidence interval was set at 95%; statistical significance was determined at p <0.05.

Ethical Consideration

The Institutional Review Board of Imam Abdulrahman Bin Faisal University granted ethical approval for this study to be conducted at the King Fahd Hospital of the University (approval no.: IRB-2023-01-325).

Results

Demographics

A total of 150 patients with PAD presented to our facility during the study period. Among these, 102 patients (204 upper limbs) who presented between January 2018 and January 2023 met our eligibility criteria and were included for further analyses. Their CTA data were analyzed for PAD (Figure 1).

An average of 19 cases of upper extremity PAD were newly diagnosed annually at our hospital. The clinical characteristics of the patients are summarized in Table 1. The mean age of the patients was 55.45 ± 27.17 years (range, 28–83 years). Most patients were male (66.7%), with a male-to-female ratio of 2:1.

Table 1 Most Common Clinical Presentations in the Study Population (Listed in the Descending Order of Prevalence)

Comorbidity Analysis

No family history of PAD was observed in any case. The most prevalent comorbidities were DM and HTN, followed by hyperlipidemia and obesity (body mass index [BMI] >30 kg/m²). Data on CKD severity (based on dialysis requirement) and obesity are summarized in Figure 2. Findings on the correlations between segmental involvement and DM, HTN, hyperlipidemia, obesity (BMI >30 kg/m²), CKD (dialysis or no dialysis), connective tissue disease, ischemic heart disease, smoking status, and family history are summarized in Table 2.

Table 2 Patient Characteristics According to the Segments Affected by Peripheral Arterial Disease

Figure 2 Prevalence and severity of chronic kidney disease (a) and obesity (b) in patients with upper extremity peripheral arterial disease.

In patients with DM, the occlusive form of PAD prevailed in the second quarter (ie, the arm; four segments [18.18%]); this indicated a trend towards the involvement of the intermediate part. However, the stenotic form of PAD prevailed in the third quarter (ie, the forearm; 16 segments [72.72%]). Thus, both the ulnar and radial arteries were evenly involved in patients with DM (eight segments [36.36%]; Figure 3).

Figure 3 Radiological image from a 47-year-old man with type 2 diabetes mellitus and upper extremity peripheral arterial disease. Three-dimensional, reformatted, volume-rendered computed tomography angiography image shows the high origin of the brachioradial branch (arrow), which represents a higher bifurcation point of the brachial artery. The atherosclerotic and multifocal atherosclerotic segments (arrowheads), leading to attenuated hypoperfusion of the palmar arcs (asterisks), are observed.

In patients with HTN, the occlusive form of PAD was prevalent in the second quarter (ie, the arm; 10 segments [45.45%]); this indicated a trend towards the involvement of the intermediate part. The stenotic form of PAD also tended to occur in the intermediate part in these patients (20 segments [90.90%]). The brachial artery was the most involved artery in PAD in patients with HTN (10 segments [45.45%]; Figure 4). The upper limb PAD profile, severity, and sites of arterial involvement in patients with DM and in those with HTN are summarized in Table 3.

Table 3 Characterization of the Distribution Pattern and Severity of Peripheral Arterial Disease Involvement in Patients with Diabetes Mellitus and in Those with Hypertension

Figure 4 Radiological images from a 45-year-old man with hypertension and upper extremity peripheral arterial disease. The patient reported that the right upper limb felt cold and pained for a week prior to presentation. Upon examination, the radial pulse was absent. (a) A three-dimensional, reformatted, volume-rendered computed tomography angiography (CTA) image acquired from the same patient, with the left arm abducted and externally rotated, shows atherosclerotic changes (arrowheads) with an abrupt cut in the distal part of the brachial artery just before the antecubital fossa (arrow). (b and c) Catheter-directed angiogram confirms the CTA findings of an occluded brachial artery (arrowheads). (d) Further characterization of the distal reconstitution of the radial artery by small collaterals. (e) The thrombus is crossed successfully. (f) The interventionalist used catheter-directed thrombolysis to lyse the thrombus and prevent a distal embolic shower through endovascular manipulation. (g) Final angiographic examination reveals restored flow in the upper right extremities; the patient’s symptoms resolved subsequently. However, a partial filling defect (indicating residual thrombosis) is seen to have persisted in the brachial artery (arrow); this required a more conservative management.

Laterality and Pattern Analyses

Compared with the upper right limb segments, the upper left limb segments were more affected by PAD (42 vs 63 diseased segments); the left-to-right upper extremity involvement ratio was 3:2. Compared with the subclavian artery in the left shoulder, the subclavian artery in the right shoulder was more affected by PAD (four vs eight segments); conversely, the remaining upper limb arterial segments (ie, the axillary, brachial, deep brachial, radial, ulnar, superficial, deep palmar, and digital arteries) on the left side were more affected by PAD than were those on the right side.

The stenotic form of PAD was the most common among all patients (62 segments). Furthermore, the forearm (arterial segments in the third quarter) was the most affected by PAD in all patients, followed by the arm (58 segments [3.15%]). However, the arm was the most affected in terms of the occlusive form of the disease (14 segments), followed by the forearm (one segment). Tables 2 provides a detailed summary of these findings.

Analysis of PAD Predictors

Univariate logistic regression analysis revealed DM and HTN as the significant predictors of PAD according to the anatomical arterial segments involved (p = 0.046 for both). Furthermore, multivariate logistic regression analysis revealed that DM and HTN were the independent predictors of total occlusion in the intermediate part of the upper extremities. In patients with DM, the forearm was particularly affected (radial artery: odds ratio [OR] = 3.053, 95% confidence interval [CI] = 1.313–11.710, p = 0.010). Furthermore, in patients with HTN, the arm was particularly affected (superficial brachial artery: OR = 2.419, 95% CI = 0.969–6.039, p = 0.041; deep brachial artery: OR = 4.063, 95% CI = 1.393–11.850, p = 0.020; Table 4).

Table 4 Independent Predictors of Total Occlusion Secondary to Peripheral Arterial Disease (PAD) from Among the Patient Characteristics Correlated with Disease Distributiona

Discussion

This study is a continuum sub-study of another observational cohort study conducted by Alghanimi et al on lower extremity involvement and anatomical distribution of PAD with different comorbidities in the Saudi population (2023).²⁵ The present study focused on elucidating the relationship between anatomical sites or distribution patterns in upper extremity PAD and the associated risk factors and comorbidities.

Our findings are consistent with existing literature in that most of our patients with upper extremity PAD were male (male-to-female ratio: 2:1).²⁸ This can be explained by the fact that women have higher levels of estrogen and progesterone than do men. Estrogen plays a protective role against PAD by improving the breakdown of low-density lipoproteins and cholesterol; furthermore, the elevated progesterone levels are thought to play an anti-atherosclerotic role, both directly in the blood vessel walls and indirectly through antioxidant activity.²⁸

Interestingly, Pham et al studied the upper extremity vasculature using ultrasound mapping and found that the mean diameters of the brachial artery in the elbow and the radial artery in the wrist were significantly higher in men than in women.²⁸ Thus, we speculate that the incidence of upper extremity PAD in women may have been underestimated and may actually be higher than that in men. Further studies must investigate the true incidence of PAD in women; in the interim, women should be made aware of their risk of developing the disease.^28–30

Regarding the laterality of disease involvement, our study found that the left upper extremity vessels were generally the sites most affected by PAD (left-to-right ratio, 3:2); this is consistent with the findings of a previous analysis of the Proximal Upper Limb Artery Registry (2021).³¹ We hypothesize that having a slightly lower blood pressure and vessels of a longer course^32–34 contributes to the accumulation of immune cells and uptake of circulating oxidized lipoproteins, which lead to the foundation for atheroma and fibroatheroma formation in PAD.^32,35 However, further research is required to validate this hypothesis.

Our study also revealed that the subclavian artery in the right shoulder is more affected than the corresponding artery in the left shoulder. The remaining arterial segments in the upper extremities (ie, the axillary, brachial, deep brachial, radial, ulnar, superficial and deep palmar, and digital arteries) were more affected on the left side than on the right side; this may be attributed to anatomical reasons.

The left subclavian artery is relatively long with a larger diameter at its aortic origin; the diameter progressively narrows distally. Conversely, the right subclavian artery has a shorter length, a smaller vascular diameter at its origin, and a slightly larger diameter at its distal end. This may explain the higher incidence of PAD in the proximal right subclavian artery than in the proximal left subclavian artery.^31–36

Our study revealed that HTN and DM were significant risk factors for PAD, whereas tobacco use, aging, dyslipidemia, obesity, and CKD were not (Table 2). In patients with DM, the intermediate part was the site most involved in upper extremity PAD; this is consistent with the observation made by Alghanimi et al that the femoropopliteal segment is the site most involved in PAD in patients with DM.²⁵ The present study found that PAD coexisting with HTN and DM likely manifests in a stenotic form rather than as an occlusive disease.

In patients with HTN, the intermediate part was the site most affected by PAD as well. In these patients, the total occlusion form of PAD was more likely to affect the brachial artery segments, as in patients with DM. However, unlike the superficial brachial artery, the deep brachial artery was more frequently involved in patients with HTN than in those with DM (Table 3).

Based on current evidence, we strongly recommend the use of ultrasound guidance to obtain vascular access because of multiple complications related to the access site. In 2022, Levin et al reported that for endovascular interventions, radial artery access was associated with a lower rate of postoperative complications than were other access points.³⁷ In 2021, Nardai et al reported that performing interventions targeting the upper extremities under ultrasound guidance using a distal radial approach was safe.³¹ Therefore, ultrasound imaging (instead of pulse palpation) is essential for obtaining endovascular access, particularly in diseased vessels.

In 2019, Jahic et al reported that the technical success rates of percutaneous endovascular procedures were higher in patients with the stenotic form of PAD than in those with the occlusive form of PAD.³⁸ Our findings support this: the forearm segments were the most affected in all patients with the stenotic form of PAD.

We believe that the findings of the present study have considerable relevance in clinical practice because they allow us to advocate for the use of ultrasound guidance to safely obtain arterial access in ipsilateral and contralateral upper extremities during interventional procedures tailored to individual patient characteristics and the affected anatomical regions.^31,38 Finally, recent evidence suggests that obtaining arterial access through a distal radial or snuff box is a safe and feasible technique with a more tolerable experience for patients (including pediatric patients).^39,40

Strengths and Limitations

Our study is one of the few Saudi Arabian studies on the anatomical distribution patterns of PAD and their correlation with risk factors and comorbidities. However, this study has certain limitations. First, because of its retrospective design, it was subject to substantial observer bias, which may have contributed to the underestimation of some clinical data due to the inaccurate reporting of medical histories. Another possible source of bias is the nature of our study population, which was relatively small and included patients from a single academic care center.

Recommendations for Future Studies

Confirming the effects of risk factors on the anatomical distribution of atherosclerosis can enable the development of targeted screening guidelines for the early detection of PAD and prevention of adverse outcomes (such as limb amputation). Given the substantial cardiovascular comorbidity associated with PAD, identifying patients at the highest risk of cardiovascular events is pivotal for preventing such events.⁴¹ The authors suggest performing a prospective screening study using Doppler ultrasound for the positive diagnosis of PAD while collecting more data on the associated risk factors. Furthermore, randomized controlled interventional trials on the treatment of PAD in patients with highly calcified vessels are required to obtain further insights into this condition. Regular screening of the upper extremity vessels in patients with DM is also recommended.

Conclusion

This cohort study enabled us to distinguish among the various patterns of upper extremity PAD. Men, in whom the anti-atherosclerotic roles of estrogen and progesterone are absent, are more likely to be affected by PAD (with a predilection for the left upper extremities). Our findings also suggest a link between the underlying pathophysiology of this disease and certain patient characteristics in terms of the disease distribution patterns. HTN and DM were the most prevalent comorbidities in patients with upper extremity PAD. Angiographically, whether with cross-sectional imaging or with catheterized contrast-enhanced fluoroscopy imaging, PAD in patients with HTN and DM is probably observed as stenosis rather than as occlusion. HTN and DM were identified as independent predictors of PAD in the arms and forearms, respectively. This is an important finding for interventionalists who opt for upper extremity access (eg, radial artery access) in patients with upper extremity PAD. The methodology of the current study and the concepts discussed herein may pioneer future multicenter studies with larger sample sizes.