C-reactive protein and incidence of symptomatic peripheral vascular disease (sPVD): Findings from the multi-ethnic study of atherosclerosis (MESA) Authors: Atef Akoum, Christopher Wang, Bahaa El Deen Wehbeh, Yugene Y Guo, Lee Bockus, Joey Chiang, Kai Chen, Younghoon Kwon, Honghuang Lin, Rajeev Malhotra, Allan Walkey, Jason P Li Published: 2026 ABSTRACT AbstractEarly identification of individuals at risk for symptomatic peripheral vascular disease (sPVD), including abdominal aortic aneurysm (AAA) and peripheral artery disease (PAD), is essential to reduce the substantial burden of associated morbidity and mortality. Given the central role of systemic inflammation in the pathogenesis and progression of atherosclerosis, we prospectively evaluated high sensitivity C-reactive protein (CRP) as a potential predictive biomarker for the long-term development of these specific vascular conditions within a large, population-based cohort free of baseline cardiovascular disease. Among 6161 participants followed over a median of 10.0 years, elevated CRP at or above 2.0 mg/L was independently associated with a more than two-fold higher incidence of sPVD after adjustment for traditional risk factors (HR 2.70; 95% CI 1.64 to 4.43; p < 0.001). Collectively, these findings suggest that elevated CRP might be an independent marker of long-term peripheral vascular risk and support further investigation of its role in primary prevention risk stratification.Keywords: Symptomatic peripheral vascular disease, Abdominal aortic aneurysm, Peripheral artery disease, Subclinical atherosclerosis, C-reactive protein, Inflammation Background Peripheral vascular disease (PVD), a manifestation of systemic atherosclerosis, affects arteries outside the heart and brain. Peripheral artery disease (PAD), a common yet frequently underdiagnosed form of atherosclerosis affecting more than 15% of older adults, is associated with an increased risk of cardiovascular events and mortality [1]. Abdominal aortic aneurysm (AAA), another atherosclerotic-associated condition, is often detected incidentally through screening, or presents as a life-threatening rupture. In this study, PVD denotes the pre-specified composite outcome used by the MESA cohort, encompassing both PAD and AAA, two conditions that share a common atherosclerotic and inflammatory pathophysiology despite differing clinical presentations. Diagnosis of symptomatic PVD (sPVD) relies on clinical findings in conjunction with established diagnostic modalities such as the ankle-brachial index (ABI), angiography, or ultrasound, yet clinical risk stratification remains inconsistent, warranting better biomarkers. High-sensitivity C-reactive protein (CRP), a marker of systemic inflammation, has been implicated in the process of atherothrombosis; however, the relationship between low-grade systemic inflammation and sPVD remains poorly characterized. Several prior cohort studies have linked elevated CRP to incident PAD, including the Physicians' Health Study and the EPIC-Norfolk prospective cohort [2], [3]. However, these analyses were limited by single-ancestry populations, non-adjudicated outcomes, or incomplete adjustment for contemporary risk factors such as statin therapy. To address this gap, we investigated the association between CRP levels and incident sPVD in a large, multi-ethnic cohort with physician-adjudicated endpoints and full adjustment for current preventive therapies. Methods We performed a secondary analysis of the Multi-Ethnic Study of Atherosclerosis (MESA). Of 6814 initial participants free of cardiovascular disease at baseline, 6161 were analyzed, excluding those with: missing baseline CRP levels (n = 52) and data on incident sPVD (n = 29), likely acute-phase CRP elevation (CRP greater than 10.0 mg/L; n = 568) [4], and sPVD at baseline (n = 4). This approach is consistent with epidemiologic conventions, as CRP values above 10.0 mg/L were considered to reflect likely acute-phase elevation and were excluded, while values at or below this threshold were considered to reflect the chronic, low-grade inflammatory state relevant to atherosclerosis [4], [5]. Baseline demographic and clinical characteristics were assessed at visit 1. Continuous variables were summarized as median [IQR], and categorical variables as frequencies; Mann-Whitney and chi-square test were used for comparison, respectively. Baseline ABI was calculated as the ratio of the highest ankle to the highest brachial systolic blood pressure (BP). CRP was measured using a BNII nephelometer (Dade Behring Inc., Deerfield, IL) and categorized using established cardiovascular risk cut-offs: <2 mg/L (n = 3482, low risk) and ≥2 mg/L (n = 2679, high risk). This categorization follows established cardiovascular risk thresholds from the CDC/AHA joint scientific statement and aligns with the enrollment criterion used in the JUPITER primary prevention trial [5], [6]. The primary outcome was incident sPVD. The MESA definition of sPVD was physician adjudicated with incident sPVD classified as definite, probable, or absent. It included symptomatic conditions such as lower extremity claudication, atherosclerosis, arterial embolism and/or thrombosis, and AAA. Probable sPVD required a physician-documented symptomatic diagnosis. Definite sPVD required at least one of the following criteria: ultrasound evidence of obstruction, positive exercise claudication test, revascularization, amputation for ischemia, ABI ≤ 0.8, new aortic aneurysm, or vascular procedure for AAA. Smoothing splines analysis assessed the relationship between continuous CRP levels and incident sPVD. Kaplan-Meier curves and Cox regression were performed for unadjusted and adjusted risk estimates of incident sPVD based on CRP groups. The Cox model was adjusted for age, sex, body mass index (BMI), smoking, diabetes (per 2003 ADA fasting criteria), hypertension (average 2nd-3rd systolic BP ≥ 130 mmHg or diastolic BP ≥ 80 mmHg or antihypertensive use), low-density lipoprotein cholesterol (LDL-C) levels, and statin utilization. The proportionality assumption was verified by Schoenfeld residual testing. A 2-sided p-value <0.05 was considered statistically significant. Follow-up time was administratively censored at 10 years for the primary analysis to preserve adequate numbers at risk throughout the observed period and ensure stable effect estimates. Data analyses and figure generation were conducted using R version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria). Kaplan-Meier survival curves, Schoenfeld residual plots, and cumulative hazard plots were generated using the survival (v3.5-5) and survminer (v0.4.9) packages. Smoothing splines were fitted using the npreg package (v0.87-9). To further evaluate the association between CRP and PAD specifically, we conducted a pre-specified sensitivity analysis restricting the outcome to incident PAD defined by an ankle-brachial index of ≤0.90 at visit 3 or visit 5. Participants with missing or an abnormal baseline ABI, defined as ≤0.90 or > 1.40 (n = 184), and those with missing or ABI >1.40 at any follow-up visit (n = 858) were excluded, yielding an analytic sample of 5119 participants. Participants with ABI above 1.40 were excluded because this finding reflects arterial incompressibility due to medial arterial calcification rather than flow-limiting luminal obstruction, rendering ABI unreliable for the assessment of obstructive PAD [1]. Follow-up time was divided into six discrete intervals corresponding to study assessment periods, and the occurrence of PAD during each interval was modeled as a binary outcome using a pooled logistic regression model adjusted for age, sex, BMI, LDL-C, smoking status, statin use, hypertension, and diabetes mellitus. On the other hand, a parallel sensitivity analysis for AAA alone was not feasible, as AAA incidence in MESA is captured only through the composite physician-adjudicated outcome and cannot be independently ascertained using imaging or procedural data available to this analysis. This study was approved by the UMass Institutional Review Board (STUDY00002864). Results In our study, 56.5% of participants had CRP < 2.0 mg/L (Fig. 1). The low CRP group had similar median age (62 [53–70] vs. 63 [54–70] yo, p = 0.06), more males (56.3% vs. 39.9%, p < 0.001), lower BMI (26.0 [23.4–28.9] vs. 29.2 [26.2–33.0] kg/m2, p < 0.001), lower LDL-C (115 [96–135] vs. 118 [97–139] mg/dL, p = 0.002), and lower CRP levels (0.86 [0.51–1.32] vs. 3.84 [2.79–5.62] mg/L, p < 0.001) compared to the high CRP group. ABI values were normal in both low and high CRP groups (1.16 [1.10–1.22] and 1.15 [1.08–1.21]). While diabetes, hypertension, and smoking were more common in the high CRP group (14.0% vs. 10.5%, 64.6% vs. 54.4%, 15.5% vs. 10.8%, respectively, p < 0.001 for all), statin utilization was not statistically different (14.2% vs. 15.8%, p = 0.083). Histogram displaying the distribution of study participants across baseline CRP levels. A majority of participants (n = 5372; 87.2%) had CRP concentrations between 0.15 and 5.0 mg/L, while a smaller subset (n = 789; 12.8%) had CRP concentrations greater than 5.0 and equal to or less than 10.0 mg/L. CRP- High-sensitivity C-reactive protein. With a median follow-up of 10.0 years, there were more sPVD events in the high CRP group (n = 52 vs. 28, p < 0.001), with an incidence of 1.9% vs. 0.8% (Fig. 2A). Elevated CRP was significantly associated with increased sPVD risk in unadjusted (HR = 2.47, 95% CI 1.56–3.90, p < 0.001) and adjusted analyses (HR = 2.70; 95% CI 1.64–4.43, p < 0.001). Cardiovascular risk factors - including smoking status, hypertension, and DM - as well as age and sex, were significantly associated with incidence of sPVD. In contrast, elevated LDL-C levels, BMI and statin use were not significantly associated with sPVD in our analysis (Fig. 2B). Furthermore, the probability of developing sPVD increased progressively with rising CRP levels (Fig. 3). Association of low and high CRP groups with sPVD. A. Kaplan-Meier Curve for the Cumulative Incidence of sPVD by CRP Groups. Unadjusted time-to-event analysis showing a significantly higher cumulative hazard of developing symptomatic peripheral vascular disease among participants with high CRP (≥2 mg/L) compared with low CRP (<2 mg/L) (log-rank p < 0.001). CRP = high-sensitivity C-reactive protein; sPVD = symptomatic peripheral vascular disease. B. Multivariable Cox Regression of Risk Factors and Incident sPVD. After adjustment for clinical risk factors, older age, elevated CRP, smoking, and hypertension were significantly associated with higher risk of incident sPVD. Elevated CRP was associated with more than a two-fold higher risk of sPVD (HR 2.70, 95% CI 1.64–4.43; p < 0.001), BMI, LDL-C, and statin use were not significantly associated with sPVD in the adjusted model. CI = confidence interval; BMI = body mass index; HR = Hazards ratio; LDL-C = low-density lipoprotein cholesterol; CRP = high-sensitivity C-reactive protein; HTN- hypertension; DM- diabetes mellitus; sPVD = symptomatic peripheral vascular disease. Association between High sensitivity C-reactive protein (CRP) levels and incidence of symptomatic peripheral vascular disease (sPVD). Spline analysis demonstrating a mostly linear increase in sPVD incidence as CRP rises from 0 to 5.0 mg/L, followed by a subsequent gradual upward trend between 5.0 and 10.0 mg/L. The solid black line represents sPVD incidence and the shaded yellow area indicates the 95% confidence interval. sPVD- symptomatic peripheral vascular disease. CRP- High-sensitivity C- reactive protein. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) In the sensitivity analysis restricted to ABI-defined incident PAD at visits 3 and 5 (n = 5119; 102 PAD events), elevated CRP between 2.0 and 10.0 mg/L remained independently associated with incident PAD after full multivariable adjustment (OR 1.86; 95% CI 1.21 to 2.85; p = 0.005), consistent with the primary analysis. Discussion To our knowledge, this study represents one of the largest analyses evaluating the association between CRP levels and incident sPVD among individuals free of known cardiovascular disease at baseline. Prior studies have linked CRP to incident PAD, including the Physicians' Health Study, which first demonstrated this association in apparently healthy men [2], and the EPIC-Norfolk cohort, which confirmed it across more than 18,000 participants [3]. Our analysis extends this literature by evaluating physician-adjudicated symptomatic outcomes in a large multi-ethnic cohort with adjustment for contemporary therapies including statins. In our analysis, elevated CRP (≥2 mg/L) was independently associated with more than a two-fold higher risk of incident sPVD. This association persisted after adjustment for risk factors (Fig. 2B), suggesting that low-grade inflammation contributes to symptomatic peripheral vascular events beyond the cumulative burden of traditional atherosclerotic risk factors. Moreover, restricted cubic spline analysis demonstrated a continuous positive association between CRP levels and sPVD incidence (Fig. 3). Notably, incidence of sPVD did not diminish among participants with CRP concentrations between 5.0 and 10.0 mg/L (n = 789, 12.8% of the cohort); within this range, incidence remained higher than among participants with CRP ≤ 5.0 mg/L, 20 (2.5%) vs 60 (1.1%) events, indicating that the observed association is not an artifact of limited representation at higher CRP concentrations. The observed association is biologically plausible. High sensitivity CRP reflects chronic low-grade inflammatory activation, which promotes endothelial dysfunction, plaque instability, and thrombotic susceptibility, central mechanisms to both PAD and AAA [2]. Accordingly, CRP may serve as an early marker of sPVD risk, potentially identifying susceptible individuals before ABI abnormalities become apparent, as observed ABI values across both CRP groups were normal. Beyond systemic inflammation, our analysis confirms the substantial impact of traditional clinical drivers on sPVD risk. Among other risk factors, smoking carried the highest risk, followed by diabetes and hypertension, aligning with previously reported data [7]. While hyperlipidemia is a well-documented risk factor for PVD, the evidence regarding obesity remains heterogenous, with prior studies suggesting a potential protective effect of higher BMI, frequently termed as “obesity paradox” [8]. In our analysis, BMI was not associated with incident sPVD, aligning with this theory and suggesting that the relationship between BMI and peripheral vascular outcomes may not be as linear as the direct correlation observed with inflammatory markers like CRP. While obesity clearly contributes to a pro-inflammatory state, the clinical manifestation of sPVD likely reflects a more complex interaction of multifactorial risks rather than a simple adiposity-driven model. Furthermore, statin therapy was not associated with a lower incidence of sPVD in our study population. Baseline statin utilization was similar between the low and high CRP groups, reducing, though not eliminating, concerns about confounding by indication. Although statins reduce CRP and LDL-C levels and improve outcomes in patients with established cardiovascular disease [1], [9], their role in primary prevention of PVD remains inadequately defined, as existing evidence is largely extrapolated from cardiovascular outcomes such as stroke and myocardial infarction rather than direct sPVD endpoints [1], [10], [11]. The persistent risk of sPVD associated with elevated CRP in our cohort suggests that inflammation may reflect complex and multifactorial mechanisms not fully captured by traditional risk modifications. Consequently, reducing CRP, as demonstrated in the CANTOS trial, may reduce cardiovascular events through pathways independent of lipid lowering therapies [12]. Nonetheless, the observational design cannot fully separate a true lack of statin effect from residual indication bias, and this finding should be interpreted with caution. Sex differences were also observed. Male sex was associated with more than a two-fold higher risk of incident sPVD. Although contemporary data suggest that females may present with more severe manifestations, historically, PAD was considered predominantly a male disease, partly due to atypical presentations and diagnostic approaches in females [7], [13]. Since MESA was a non-contemporary trial, our findings are in line with previous literature. Finally, the observed sPVD incidence of 1.3% was lower than previously reported PAD incidence rates of 3.5% [10]. This discrepancy can likely be attributed to the strict definition of sPVD employed in the MESA cohort. Limitations First, the MESA cohort defines sPVD as a composite endpoint combining symptomatic PAD and AAA. Although both conditions fall within the spectrum of peripheral vascular disease, they differ in underlying biology, natural history, and clinical presentation. Consequently, aggregating them into a single outcome may obscure disease-specific associations and limits direct comparison with studies evaluating each condition separately. Moreover, individual adjudicated event counts are not available within the MESA outcome framework, precluding stratified analysis by disease subtype. To partially address this limitation, we conducted a pre-specified sensitivity analysis restricted to ABI-confirmed incident PAD, which yielded a consistent association. Second, CRP was measured only once at baseline. Although a single measurement may not fully capture longitudinal inflammatory burden or temporal variability in inflammation over time, serial CRP measurements were not consistently available across follow-up visits for the majority of participants, precluding a formal analysis of longitudinal CRP trends. Furthermore, this non-differential misclassification of CRP exposure would bias the observed association toward the null, suggesting the true relationship between CRP and sPVD risk may be stronger than reported. Nonetheless, the incidence of sPVD remained elevated among participants with baseline CRP between 5.0 and 10.0 mg/L (n = 789, 12.8% of the cohort) relative to those with CRP ≤ 5.0 mg/L, supporting the stability of the association despite reliance on a single measurement. Lastly, as with all observational studies, residual confounding cannot be excluded despite multivariable adjustment for major risk factors. Conclusion Among participants without clinically apparent cardiovascular disease, elevated CRP levels were significantly associated with a higher incidence of sPVD. These findings identify elevated CRP as an independent marker of long-term sPVD risk and support further investigation of CRP-guided risk stratification in primary prevention. Clinical perspective What Is New? •This is the largest study to evaluate the incidence of symptomatic peripheral vascular disease (sPVD) with CRP levels. This is the largest study to evaluate the incidence of symptomatic peripheral vascular disease (sPVD) with CRP levels. •Elevated CRP levels were independently associated with over a two-fold increased risk of developing sPVD. Elevated CRP levels were independently associated with over a two-fold increased risk of developing sPVD. What Are the Clinical Implications? •CRP may serve as a useful inflammatory biomarker to identify individuals at higher long-term risk for developing symptomatic peripheral vascular disease. CRP may serve as a useful inflammatory biomarker to identify individuals at higher long-term risk for developing symptomatic peripheral vascular disease. •Early identification of elevated CRP levels could help guide targeted prevention strategies in asymptomatic individuals before the onset of clinically evident peripheral vascular disease. Early identification of elevated CRP levels could help guide targeted prevention strategies in asymptomatic individuals before the onset of clinically evident peripheral vascular disease. CRediT authorship contribution statement Atef Akoum: Writing – original draft, Software, Methodology, Formal analysis, Conceptualization, Resources, Writing – review & editing. Christopher Wang: Writing – original draft, Methodology, Formal analysis, Conceptualization. Bahaa El Deen Wehbeh: Writing – review & editing, Methodology, Formal analysis. Yugene Y. Guo: Writing – original draft, Methodology, Formal analysis, Conceptualization. Lee Bockus: Writing – original draft, Methodology, Conceptualization. Joey Chiang: Writing – original draft, Methodology, Formal analysis. Kai Chen: Writing – review & editing, Methodology, Conceptualization. Younghoon Kwon: Writing – review & editing, Methodology, Conceptualization. Honghuang Lin: Writing – review & editing, Methodology, Resources, Validation, Visualization. Rajeev Malhotra: Writing – review & editing, Methodology, Conceptualization, Supervision, Visualization. Allan Walkey: Writing – review & editing, Methodology, Investigation, Supervision. Jason P. Li: Writing – review & editing, Validation, Supervision, Formal analysis, Data curation, Conceptualization, Methodology, Project administration, Resources, Writing – original draft. Ethical statement This study was conducted using data from the Multi-Ethnic Study of atherosclerosis (MESA), a prospective multicenter cohort study designed to investigate the prevalence, correlates, and progression of subclinical cardiovascular disease in a diverse population. The original MESA study protocol was approved by the Institutional Review Boards (IRB) of all participating field centers, and all participants provided written informed consent at enrollment. The present study was conducted in accordance with the ethical principles of the Declaration of Helsinki and all applicable institutional and federal guidelines for research involving human subjects. The data analyzed were de-identified prior to access and analysis to maintain participant confidentiality and privacy. In addition to the original MESA approvals, the current secondary analysis received approval from the UMass Institutional Review Board (STUDY00002864). The requirement for additional informed consent was waived due to the retrospective nature of the study and the use of de-identified data. Declaration of Generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used ChatGPT (OpenAI) to assist with language editing and manuscript refinement. After using this service, all authors reviewed and edited the content as needed, and take full responsibility for the content of the published article. Source of funding Dr. Malhotra was supported by the National Heart, Lung, and Blood Institute (R01HL159514 and R01HL162928), the Leducq Foundation, and the American Heart Association (25TPA1474605). Declaration of competing interest The authors have no relevant conflicts of interest to disclose. Acknowledgments None. References