The predictive value of four serum biomarkers for major adverse events

in patients with small abdominal aortic aneurysm

Jonathan Golledge, MB, BChir, MA, MChir, FRACS,a,b,c Ramesh Velu, FRACS,b Frank Quigley, FRACS,d
Jason Jenkins, FRACS,e and Tejas P. Singh, MBBS, MPH,a,b Townsville and Brisbane, Queensland, Australia

ABSTRACT
Background: The primary aim of this study was to test which of a group of four inflammation and thrombosis biomarkers
were independently predictive of major adverse cardiovascular events (MACE) in patients with small abdominal aortic
aneurysm (AAA).
Methods: A total of 471 participants with a 30- to 54-mm AAA had serum C-reactive protein (CRP), fibrinogen, neutrophil-
lymphocyte ratio (NLR), and homocysteine measured. The primary outcome was MACE, which was defined as the first
occurrence of myocardial infarction, stroke, or cardiovascular death. The association of biomarkers with events was
assessed using Kaplan-Meier and Cox proportional hazard analyses. The net improvement in risk of event categorization
with addition of a biomarker to clinical risk factors alone was assessed using net reclassification index.
Results: Participants were followed for a median of 2.4 years (interquartile range, 0.8-5.4 years), and 102 (21.7%) had a
MACE. The incidence of MACE was 13.2% in participants with CRP >3.0 mg/L, compared with 10.1% in those with CRP
#3.0 mg/L at 2.5 years (P ¼ .047). After adjusting for other risk factors, higher CRP was associated with a significantly
higher risk of MACE (hazard ratio, 1.19; 95% confidence interval, 1.05-1.35). None of the other biomarkers were associated
with the risk of MACE. According to the net reclassification index, CRP significantly improved the risk classification of
MACE compared with clinical risk factors alone.
Conclusions: CRP can assist in classification of risk of MACE for patients with small AAA. (J Vasc Surg 2023;77:1037-44.)
Keywords: Abdominal aortic aneurysm; Biomarkers; Major adverse cardiovascular events; Risk factors; Surgery

In many countries abdominal aortic aneurysm (AAA) is and 5.6% at 5 years, respectively.6 Predicting which pa-
now identified when small and asymptomatic as a result tients are most at risk of MACE might facilitate targeted
of incidental imaging or screening programs.1 Current intensive medical management to reduce these events.
guidelines recommend that small AAAs (<55 mm in Thrombosis and inflammation are key mechanisms
men and <50 mm in women) enter repeat imaging sur- involved in cardiovascular events, evidenced by the
veillance.2,3 The focus during surveillance is the identifica- recent findings of the benefits of rivaroxaban and
tion of AAAs that reach intervention threshold diameter or interleukin-1 inhibition in reducing the risk of MACE.7,8
become symptomatic, at which stage surgical repair may A range of novel anti-inflammatory and anti-
be considered.2,3 Patients with small AAAs are, however, thrombotic drugs are now being introduced or are under
also at high risk of major adverse cardiovascular events development to provide more intensive secondary pre-
(MACE), including myocardial infarction, stroke, and car- vention in order to reduce the risk of MACE.9,10 Blood
diovascular death.4-6 A recent study among patients markers able to identify patients with the highest pro-
with small AAA reported incidences of cardiovascular inflammatory and pro-thrombosis potential would be
death, myocardial infarction, and stroke of 13.8%, 11.5%, very valuable in personalizing the application of these

From the Queensland Research Centre for Peripheral Vascular Disease, College Correspondence: Jonathan Golledge, MB, BChir, MA, MChir, FRACS, Director,
of Medicine and Dentistry, James Cook University, Townsvillea; the Depart- Queensland Research Centre for Peripheral Vascular Disease, College of Med-
ment of Vascular and Endovascular Surgery, The Townsville University Hospi- icine and Dentistry, James Cook University, Townsville, Queensland, Australia,
tal, Townsvilleb; the Australian Institute of Tropical Health and Medicine, 4811 (e-mail: Jonathan.Golledge@jcu.edu.au).
James Cook University, Townsvillec; the Mater Hospital, Townsvilled; and the The editors and reviewers of this article have no relevant financial relationships to
Royal Brisbane and Women’s Hospital, Brisbane.e disclose per the JVS policy that requires reviewers to decline review of any
This study was supported by funding from the National Health and Medical manuscript for which they may have a conflict of interest.
Research Council (1180736), Medical Research Futures Fund (2015999), Heart 0741-5214
Foundation (Strategic grant), Townsville Hospital and Health Services (Serta) Copyright Ó 2022 by the Society for Vascular Surgery. Published by Elsevier Inc.
and the Queensland Government (Australia). J.G. holds a Senior Clinical https://doi.org/10.1016/j.jvs.2022.12.001
Research Fellowship from the Queensland Government.
Author conflict of interest: none.

1037
1038 Golledge et al Journal of Vascular Surgery
April 2023

new treatments. There has, however, been no prior study

investigating the accuracy of biomarkers in predicting ARTICLE HIGHLIGHTS
MACE within patients with small AAA. d
Type of Research: Multi-center prospective cohort
Fibrinogen, neutrophil-lymphocyte ratio (NLR), C-reactive study
protein (CRP), and homocysteine are circulating markers d
Key Findings: High C-reactive protein significantly
of inflammation and thrombosis that are commonly improved the risk classification of major adverse car-
measured in routine practice, have established methods diovascular events compared with clinical risk factors
to accurately quantify in most pathology laboratories, alone in 471 participants with 30- to 54-mm abdom-
and have been associated with MACE in other popula- inal aortic aneurysms.
tions.11-14 The primary aim of this study was to examine d
Take Home Message: C-reactive protein can assist in
whether fibrinogen, NLR, CRP, and homocysteine were classification of risk of major adverse cardiovascular
independently associated with the risk of MACE in pa- events for patients with small abdominal aortic
tients with small AAA. The study also aimed to examine aneurysms.
whether the four biomarkers significantly improved strati-
fying of the risk of MACE over standard risk factors alone. It
was hypothesized that fibrinogen, NLR, CRP, and homo- calculated using the Chronic Kidney Disease-
cysteine were independently associated with the risk of Epidemiology Collaboration group (CKD-EPI) formulae.16
MACE and would improve the risk stratification of these
Risk factors and medications. Clinical characteristics
events in patients with small AAA.
collected at study entry included age, sex, diabetes, hy-
pertension, smoking, coronary heart disease (CHD),
METHODS stroke, and currently prescribed medications, as previ-
Study design and participants. This study was part of ously described.15,16 Smoking classification was based
an ongoing prospective cohort investigation that on smoking history and defined as current smoker
commenced in 2002 and aimed to identify risk factors (smoked within the last month), previous smoker, or
associated with the outcome of vascular disease.15,16 To never smoked. Hypertension, diabetes, and stroke were
be eligible for the current study, participants were diag- defined by documented history or medical treatment for
nosed with an asymptomatic AAA with a maximum these conditions. CHD was defined by a history of
diameter of 30 to 54 mm. Patients presenting with myocardial infarction, angina, or coronary revasculariza-
ruptured, symptomatic, large, or previously repaired AAA tion. Blood pressure and fasting low-density lipoprotein
were excluded. Patients were also excluded if they did (LDL) and high-density lipoprotein (HDL) were measured as
not provide a fasting blood test for biomarker assess- previously described.6,19 Prescriptions of aspirin and statins
ment. Patients were recruited from the outpatient were obtained from medical records at the time of
vascular services at The Townsville University Hospital, recruitment. Maximum anterior to posterior and transverse
the Mater Hospital Townsville, and The Royal Brisbane infra-renal aortic diameters were measured by experi-
and Women’s Hospital in Queensland, Australia between enced sonographers using ultrasound machines
March 2002 and July 2019. Written informed consent employed in the vascular laboratories at each site using a
was obtained from all participants upon entry into the standard protocol, as described previously.20,21 Aortic
study. The study was performed in accordance with the diameter was measured from outer wall to outer wall of the
Helsinki declaration, and ethical approval was granted artery in the orthogonal plane.20,22 The reproducibility of
from institutional ethics committees (HREC/12/QTHS/202 aortic diameter measurements was assessed in each
and HREC/14/QTHS/203). vascular laboratory, with inter-observer reproducibility co-
efficients being less than 4 mm as previously reported.20-22
Measurement of biomarkers. All blood markers were
tested at the time of recruitment in an accredited pa- Definition and assessment of outcomes. The primary
thology laboratory monitored as part of both an internal outcome was the first occurrence of a MACE defined as
quality assurance program and Royal College of Patholo- including myocardial infarction, stroke, or cardiovascular
gists of Australasia quality assurance program. Homocys- death.19,20 Secondary outcomes were myocardial infarc-
teine and CRP were measured using chemiluminescent tion alone, cardiovascular death alone, all-cause mortal-
immunoassays on the Abbott Alinity platform as previ- ity, and AAA-related events including AAA rupture or
ously described.17,18 Neutrophil and lymphocyte requirement for repair. Outcome data were recorded
numbers and fibrinogen concentration were measured during clinical reviews on prospectively defined case
on the Sysmex platform using flow cytometry and the report forms. Hospital charts and electronic records were
Von Clauss method, respectively. Serum creatinine was also reviewed by a vascular specialist at the time of re-
measured using a spectrophotometry method in line view. Outpatient reviews were planned yearly for patients
with established guidelines as previously described.16 with AAA measuring <40 mm and 6-monthly for AAAs
Estimated glomerular filtration rate (eGFR) was measuring 40 to 54 mm. Outcome data were obtained
Journal of Vascular Surgery Golledge et al 1039
Volume 77, Number 4

from clinical reviews and linked hospital admission re- CRP (>3.0 mg/L) having a MACE during follow-up. Par-
cords from the Queensland Hospital Admitted Patient ticipants were grouped into low (#3.0 mg/L) and high
Data Collection and death register, which is regularly CRP (>3.0 mg/L) groups in line with previously reported
audited to minimize inaccuracies.19,20 CRP cutoffs reported by the Centers for Disease Control
and Prevention and the American Heart Association.23
Sample size. It was aimed to have adequate power to
The negative and positive predictive values of a CRP
test the hypothesis that the four biomarkers were associ-
>3 mg/L for predicting future MACE were calculated as
ated with the risk of MACE. Previous studies suggest that
previously described.24 Data were analyzed using the
approximately 25% of people with small AAA have a
SPSS v25 (IBM, Armonk, NY) and Stata v16.1 (StataCorp LP,
MACE during short term follow-up.5 Monte-Carlo simu-
College Station, TX) software packages. P values of <.05
lations suggest that a multivariable regression model is
were accepted to be significant for all of these analyses.
powered sufficiently when 10 outcome events per de-
gree of freedom of the predictor variables are observed.21
Assuming an incidence of MACE of 25% and planning to
RESULTS
include an individual biomarker, age, sex, diabetes,
Association of biomarkers with MACE. A total of 471
hypertension, current smoking, CHD, stroke, eGFR, aortic
participants were included and followed for a median
diameter, aspirin, and statins in the regression models, it
of 2.4 years (IQR, 0.8-5.4 years). During this time 102 par-
was estimated that a sample size of 480 participants
ticipants (21.7%) had a MACE. Participants who later
would be required to have adequate power to test the
had a MACE were significantly more likely to have hyper-
main hypothesis.
tension, be smokers and have a prior history of stroke or
Data analysis. Continuous data that were not normally CHD at entry than those who did not (Table I). Forty-eight
distributed, as confirmed using the Shapiro-Wilk test, of 172 patients (27.9%) who had an AAA repair experi-
were presented as median and interquartile range (IQR). enced a MACE during follow-up, whereas 54 of 299 pa-
Statistical comparisons between groups were conducted tients (18.1%) that did not have an AAA repair
using the Mann-Whitney U and the Pearson c2 tests. Cox experienced a MACE during follow-up (P ¼ .012). There
proportional hazard analyses assessed the association of were no significant differences in CRP between partici-
the biomarkers with outcome events adjusted for age, pants with CHD compared with those who did not have
sex, diabetes, hypertension, current smoking, CHD, CHD (2.0; IQR, 1.0-5.1 vs 3.1; IQR, 1.0-5.9 mg/L for partici-
stroke, eGFR, initial aortic diameter, aspirin, and statin pants with CHD and no CHD respectively; P ¼ .123). Par-
prescription. To prevent violation of model assumptions, ticipants having a MACE had significantly higher NLR,
aortic diameter was converted to medians as previously CRP, and homocysteine, but not fibrinogen, at entry
recommended.22 The Cox proportional hazard analyses compared with those who did not (Table I). By Kaplan-
examined the association of an increase in the concen- Meier analysis, the proportion of participants having a
trations of the biomarkers of approximately one standard MACE at 1 and 2.5 years were 4.6% and 10.1% in those
deviation within the population. Results were presented who had a CRP #3.0 mg/L compared with 8.2% and
as hazard ratios (HRs) and 95% confidence intervals (CIs). 13.2% in those with a CRP >3.0 mg/L (log-rank test; P ¼
A sub-analysis was performed that was restricted to .047) (Fig). Participants with a CRP >3.0 mg/L were
participants in whom blood pressure, LDL, and HDL data significantly more likely to have a MACE than those with
were available (n ¼ 421). In this analysis, the Cox propor- a CRP #3.0 mg/L (HR, 1.48; 95% CI, 1.00-2.18). After
tional hazard analysis examining the association of CRP adjusting for other risk factors, higher CRP (per standard
with MACE was adjusted for age, sex, diabetes, hyper- deviation increase) was associated with a significantly
tension, current smoking, systolic blood pressure, LDL, higher risk of MACE (HR, 1.19; 95% CI, 1.05-1.35). CRP
HDL, CHD, stroke, eGFR, aortic diameter, aspirin, and >3.0 mg/L had a negative predictive value of 28.42% (95%
statin prescription. A further sub-analysis was performed CI, 24.35%-32.49%) and a positive predictive value of
where participants that had a MACE within 30 days of 82.92% (95% CI, 79.52%-86.32%) in predicting future
AAA repair were excluded. The net improvement in risk MACE. The other biomarkers were not significantly asso-
of event categorization with addition of a biomarker to ciated with the risk of MACE or myocardial infarction in
clinical risk factors alone (ie, age, sex, diabetes, current the adjusted analyses (Table II). In the sub-analysis, the
smoking, hypertension, CHD, previous stroke, aspirin, association between high CRP and MACE remained
statin, aortic diameter, and eGFR) was assessed using net robust after additional adjustment for systolic blood
reclassification index (NRI).16 These analyses were pressure, LDL, and HDL (Table III). Four patients had a
restricted to biomarkers shown to be significantly inde- MACE within 30 days of AAA repair. Exclusion of these
pendently associated with the event in question within patients from the analysis did not significantly change the
the Cox Proportional hazard analyses. Kaplan-Meier final result. After adjusting for risk factors, higher CRP (per
curves with log rank test were used to compare the standard deviation increase) was associated with signifi-
proportion of participants with low (#3.0 mg/L) and high cantly higher risk of MACE (HR, 1.18; 95% CI, 1.04-1.35).
1040 Golledge et al Journal of Vascular Surgery
April 2023

Table I. Risk factors at recruitment in participants with small abdominal aortic aneurysm (AAA) that subsequently did and
did not have a major adverse cardiovascular event (MACE)
Risk factor at entry MACE (n ¼ 102) No MACE (n ¼ 369) P value
Age, years 74 (67-79) 74 (68-78) .763
Male 86 (84) 84 (310) .941
Diabetes 27 (26) 86 (23) .508
Hypertension 89 (87) 267 (72) .002
Smoking <.001
Current 41 (40) 66 (18)
Former 48 (47) 246 (67)
Never 13 (13) 57 (15)
Previous stroke 14 (14) 26 (7) .032
CHD 66 (65) 144 (39) <.001
Statin 69 (68) 255 (69) .778
Aspirin 63 (62) 205 (56) .262
eGFR, mL/minute/1.73 m2 67 (46-82) 74 (61-86) .002
Fibrinogen, g/L 3.2 (2.9-3.8) 3.2 (2.8-3.7) .129
NLR 2.5 (1.9-3.2) 2.2 (1.6-3.0) .033
CRP, mg/L 4.0 (1.9-6.0) 2.0 (1.0-5.0) .002
High CRP (>3 mg/L) 54 (53) 136 (37) .003
Homocysteine, mM 15.0 (11.7-18.3) 12.7 (10.6-15.1) <.001
CHD, Coronary heart disease; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; NLR, neutrophil lymphocyte ratio.
Data are presented as number (%) and median (interquartile range).

1.20; 95% CI, 1.01-1.43), cardiovascular death (HR, 1.24; 95%

CI, 1.09-1.41), and all-cause mortality (HR, 1.18; 95% CI, 1.06-
1.32). None of the other biomarkers were significantly
associated with the risk of myocardial infarction or car-
diovascular death in the adjusted analyses (Table II).
Higher homocysteine, but not fibrinogen or NLR, was
associated with a significantly higher risk of all-cause
mortality (Table II).

Association of biomarkers with AAA-related events. A

total of 172 participants underwent AAA repair (endovas-
cular, 109; open surgery, 63), and four men had fatal
aneurysm rupture. The last recorded AAA diameter
was <55 mm in three of these men (47, 49, and
53 mm). In one man who presented with a 45-mm
Fig. Kaplan-Meier curves illustrating the freedom from
major adverse cardiac events (MACE) for participants with AAA, the aneurysm diameter gradually increased over
unrepaired abdominal aortic aneurysms (AAAs) who have 6 years, but he was deemed unfit for repair. The aneu-
low (#3.0 mg/L) and high (>3.0 mg/L) serum concentra- rysm ruptured with a last recorded diameter of 76 mm.
tion of C-reactive protein (CRP) at entry. A significant High homocysteine was associated with a reduced risk
difference between groups was found on log-rank test
of AAA-related events in the analysis adjusted for other
(P ¼ .047).
risk factors (HR, 0.75; 95% CI, 0.60-0.93) but not in the
unadjusted analysis (Table II). No other biomarker was
associated with the risk of AAA-related events.
Association of biomarkers with myocardial infarction
and death. A total of 51 (10.8%), 21 (4.5%), 57 (12.1%), and Ability of biomarkers to improve risk classification of
136 (28.9%) participants had a myocardial infarction, events. CRP significantly improved the ability to catego-
stroke, cardiovascular, or all-cause death during follow- rize participants for the risk of MACE, cardiovascular
up. After adjusting for other risk factors, higher CRP death, and all-cause mortality over clinical risk factors
(per standard deviation increase) was associated with alone (Table IV). Homocysteine significantly improved
significantly higher risks of myocardial infarction (HR, the ability to categorize participants for the risk of all-
Journal of Vascular Surgery Golledge et al 1041
Volume 77, Number 4

Table II. Independent association of biomarkers with Table II. Continued.

major adverse events among participants with small
Biomarker HR 95% CI P value
abdominal aortic aneurysm (AAA)
Fibrinogen (per 0.9 g/L) 0.99 0.87-1.12 .876
Biomarker HR 95% CI P value
NLR (per 1.4 units) 0.92 0.80-1.07 .290
Major adverse cardiovascular events
CRP (per 13.5 mg/L) 1.09 0.96-1.23 .188
Unadjusted analysis
Homocysteine (per 5.6 mM) 0.90 0.76-1.05 .187
Fibrinogen (per 0.9 g/L) 1.09 0.95-1.24 .234
Adjusted analysisa
NLR (per 1.4 units) 1.02 0.87-1.19 .840
Fibrinogen (per 0.9 g/L) 1.06 0.93-1.22 .378
CRP (per 13.5 mg/L) 1.19 1.07-1.33 .002
NLR (per 1.4 units) 0.99 0.83-1.18 .922
Homocysteine (per 5.6 mM) 1.29 1.13-1.48 <.001
CRP (per 13.5 mg/L) 1.10 0.97-1.25 .129
Adjusted analysisa
Homocysteine (per 5.6 mM) 0.75 0.60-0.93 .011
Fibrinogen (per 0.9 g/L) 1.09 0.94-1.28 .258
CI, Confidence interval; CRP, C-reactive protein; HR, hazard ratio; NLR,
NLR (per 1.4 units) 0.98 0.81-1.17 .790 neutrophil lymphocyte ratio.
a
CRP (per 13.5 mg/L) 1.19 1.05-1.35 .008 Adjusted for age, sex, diabetes, hypertension, current smoking, coro-
nary heart disease, stroke, estimated glomerular filtration rate, aortic
Homocysteine (per 5.6 mM) 1.05 0.88-1.25 .619 diameter, and aspirin and statin prescription. Hazard ratios are pro-
vided relative to approximately a one standard deviation increase in
Myocardial infarction the biomarkers.
Unadjusted analysis
Fibrinogen (per 0.9 g/L) 0.99 0.80-1.22 .908
NLR (per 1.4 units) 0.94 0.72-1.22 .631 cause mortality over clinical risk factors alone. No bio-
CRP (per 13.5 mg/L) 1.17 1.00-1.37 .048 markers were found to significantly improve the risk
Homocysteine (per 5.6 mM) 1.12 0.87-1.44 .366 stratification for myocardial infarction and AAA-related
Adjusted analysisa events.
Fibrinogen (per 0.9 g/L) 1.02 0.81-1.28 .886
DISCUSSION
NLR (per 1.4 units) 0.89 0.66-1.20 .456
Multiple studies have associated larger aortic diameter
CRP (per 13.5 mg/L) 1.20 1.01-1.43 .041 with a higher risk of MACE.25 Despite this, the high risk of
Homocysteine (per 5.6 mM) 0.80 0.57-1.12 .195 MACE and cardiovascular death in patients with small
Cardiovascular death AAA has received relatively limited attention. This study
Unadjusted analysis confirms the high risk of MACE in these patients. In just
Fibrinogen (per 0.9 g/L) 1.07 0.90-1.28 .448 over 2 years, approximately 22% had a MACE, including
NLR (per 1.4 units) 1.06 0.86-1.29 .598 a 12% incidence of cardiovascular death. This highlights
CRP (per 13.5 mg/L) 1.23 1.09-1.39 .001 the need to focus not only on identifying the need for
Homocysteine (per 5.6 mM) 1.28 1.06-1.54 .009 AAA repair but also on reducing MACE in patients with
Adjusted analysis a small AAA. Personalizing the application of new anti-
inflammatory and anti-thrombotic treatments could be
Fibrinogen (per 0.9 g/L) 1.13 0.91-1.40 .259
beneficial for patients with small AAA. Biomarkers have
NLR (per 1.4 units) 0.99 0.77-1.27 .950
the potential to add to clinical risk factors in selecting pa-
CRP (per 13.5 mg/L) 1.24 1.09-1.41 .001
tients for new cardiovascular drugs. This study found that
Homocysteine (per 5.6 mM) 1.21 0.93-1.57 .161 CRP was useful in predicting MACE and death in pa-
All-cause mortality tients with small AAA. This suggests, as in patients who
Unadjusted analysis have recently had a myocardial infarction, CRP maybe
Fibrinogen (per 0.9 g/L) 1.04 0.92-1.16 .536 useful in selecting patients with small AAA for new
NLR (per 1.4 units) 1.02 0.89-1.16 .821 anti-inflammatory drugs.8
CRP (per 13.5 mg/L) 1.19 1.07-1.32 .001 Our literature search identified no prior studies exam-
Homocysteine (per 5.6 mM) 1.31 1.17-1.47 <.001 ining circulating biomarkers for MACE among patients
Adjusted analysisa with small AAA. It is possible that biomarkers identified
Fibrinogen (per 0.9 g/L) 1.06 0.93-1.21 .404
to be associated with MACE within other populations,
such as those linked with athero-occlusive disease,
NLR (per 1.4 units) 0.96 0.82-1.12 .620
would be equally applicable to patients with AAA. AAA
CRP (per 13.5 mg/L) 1.18 1.06-1.32 .002
has a different biomarker profile to athero-occlusive dis-
Homocysteine (per 5.6 mM) 1.31 1.11-1.55 .001
ease, likely resulting from its unique pathological fea-
AAA event (AAA repair or rupture) tures, such as marked aortic inflammation and large
Unadjusted analysis intra-luminal thrombus volume.26 Large AAA thrombus
(Continued) volume has been previously associated with increased
1042 Golledge et al Journal of Vascular Surgery
April 2023

Table III. Independent risk factors for major adverse cardiovascular events (MACE) among participants with small
abdominal aortic aneurysm (AAA)
Biomarker HR 95% CI P value
Major adverse cardiovascular events
Complete cohort
Age (per 1 year) 1.03 0.99-1.06 .105
Female sex 0.56 0.32-0.99 .048
Diabetes 1.20 0.75-1.91 .439
Hypertension 1.82 0.98-3.37 .056
Current smoking 3.29 2.09-5.17 <.001
CHD 2.54 1.63-3.96 <.001
Prior stroke 1.17 0.62-2.20 .623
Aspirin 0.99 0.65-1.52 .967
Statin 0.56 0.36-0.89 .015
eGFR (per 1 mL/min/1.73m2) 0.98 0.97-0.99 .003
Aortic diameter (per 1 mm) 0.96 0.64-1.44 .830
CRP (per 13.5 mg/L) 1.19 1.05-1.35 .008
Sub-analysisa
Age (per 1 year) 1.02 0.98-1.06 .344
Female sex 0.63 0.34-1.17 .146
Diabetes 1.15 0.70-1.90 .571
Hypertension 1.70 0.84-3.44 .138
Current smoking 3.82 2.34-6.26 <.001
Systolic blood pressure (per 1 mmHg) 1.01 1.00-1.02 .124
LDL (per 1 mMol/L) 0.98 0.77-1.25 .873
HDL (per 1 mMol/L) 0.90 0.49-1.64 .721
CHD 2.62 1.60-4.30 <.001
Prior stroke 1.03 0.50-2.11 .936
Aspirin 1.03 0.64-1.64 .914
Statin 0.55 0.33-0.92 .024
eGFR (per 1 mL/min/1.73m2) 0.98 0.97-0.99 .005
Aortic diameter (per 1 mm) 1.01 0.65-1.55 .973
CRP (per 14.2 mg/L) 1.21 1.06-1.38 .006
CHD, Coronary heart disease; CI, confidence interval; CRP, C-reactive protein; HDL, high-density lipoprotein; HR, hazard ratio; LDL, low-density
lipoprotein.
a
Comprised of 421 participants in whom low density lipoprotein, high density lipoprotein, and systolic blood pressure data were available. HRs are
provided relative to approximately a one standard deviation increase in the biomarkers.

risk of MACE, suggesting the potential for unique bio- different co-variant adjustment.14 The current study ex-
markers of events in patients with small AAA.5 We, there- tends these studies by demonstrating that CRP is inde-
fore, felt it appropriate to examine whether biomarkers pendently associated with the risk of cardiovascular
established in other populations, such as CRP, were events in patients with small AAA. Furthermore, accord-
also predictive of events among patients with small AAA. ing to NRI, CRP significantly improved risk stratification
CRP is the most established circulating biomarker of of events compared with clinical risk factors alone. This
cardiovascular event risk that is commonly used in clin- suggests CRP has independent value in categorizing
ical practice.8,14 CRP was, for instance, used to select par- risk, although the extent of the additive value is unclear
ticipants for inclusion in a prior trial showing the benefit as interpretation of NRI is controversial. It is possible
of interleukin-1 inhibition, although patients with small that other less commonly measured inflammatory
AAA were not included in this trial.8 Prior systematic re- markers, such as interleukins (IL), may be more useful
views have reported consistent associations of CRP with in predicting MACE but this was not examined in the
MACE in patients with occlusive artery disease, although current study. Higher IL-6 has been associated with
the independent association has not been clear due to increased risk of myocardial infarction in healthy com-
the difficulty in combining data that has received munity populations.27 High circulating concentrations
Journal of Vascular Surgery Golledge et al 1043
Volume 77, Number 4

Table IV. Discrimination and reclassification using bio- referral to vascular outpatient services. The findings may
markers plus clinical risk factors as compared with clinical therefore not be generalizable to patients with screening
risk factors alone for predicting major adverse cardiac
identified asymptomatic AAA. Finally, no cost-
events (MACE) among participants with small abdominal
aortic aneurysm (AAA) effectiveness analysis was performed.

Models NRI (95% CI) P-value

CONCLUSIONS
Major adverse cardiovascular events In conclusion, this study suggests that CRP is useful in
Clinical risk factors only Reference stratifying risk of MACE in patients with small AAA. CRP
Clinical risk factors þ CRP 0.36 (0.14-0.58) .001 may be useful in identifying patients with small AAA for
Myocardial infarction intensive medical management, including new anti-
Clinical risk factors only Reference inflammatory drugs.
Clinical risk factors þ CRP 0.21 (0.08 to 0.50) .162
Cardiovascular death AUTHOR CONTRIBUTIONS
Clinical risk factors only Reference Conception and design: JG, RM, FQ, JJ, TS
Clinical risk factors þ CRP 0.47 (0.19-0.74) .001 Analysis and interpretation: JG, TS
All-cause mortality Data collection: JG, RM, FQ, JJ, TS
Writing the article: JG
Clinical risk factors only Reference
Critical revision of the article: JG, RM, FQ, JJ, TS
Clinical risk factors þ CRP 0.41 (0.21-0.61) <.001
Final approval of the article: JG, RM, FQ, JJ, TS
Clinical risk factors þ 0.31 (0.11-0.51) .003
Statistical analysis: JG, TS
homocysteine
Obtained funding: JG
AAA events (ie, AAA repair or
rupture) Overall responsibility: JG
Clinical risk factors only
Clinical risk factors þ 0.12 (0.07 to 0.30) .226 REFERENCES
homocysteine 1. Guirguis-Blake JM, Beil TL, Senger CA, Coppola EL. Primary care
screening for abdominal aortic aneurysm: updated evidence report
CI, confidence interval; CRP, C-reactive protein; NRI, Net reclassification and systematic review for the US preventive services task force. JAMA
index. 2019;322:2219-38.
Clinical risk factors included age, sex, diabetes, current smoking, 2. Wanhainen A, Verzini F, Van Herzeele I, Allaire E, Bown M,
hypertension, coronary heart disease, previous stroke, aspirin, statin,
Cohnert T, et al. Editor’s choice - European Society for Vascular
aortic diameter and estimated glomerular filtration rate.
Surgery (ESVS) 2019 clinical practice guidelines on the manage-
ment of abdominal aorto-iliac artery aneurysms. Eur J Vasc
Endovasc Surg 2019;57:8-93.
of IL-6 have been consistently associated with AAA diag- 3. Chaikof EL, Dalman RL, Eskandari MK, Jackson BM, Lee WA,
Mansour MA, et al. The Society for Vascular Surgery practice guide-
nosis but whether high IL-6 is associated with MACE in lines on the care of patients with an abdominal aortic aneurysm.
patients with AAA is not known.28 J Vasc Surg 2018;67:2-77.e2.
In contrast, the other risk factors tested had no or 4. Bath MF, Gokani VJ, Sidloff DA, Jones LR, Choke E, Sayers RD, et al.
Systematic review of cardiovascular disease and cardiovascular
inconsistent associations with events. None of the bio- death in patients with a small abdominal aortic aneurysm. Br J Surg
markers significantly improved the risk stratification of 2015;102:866-72.
AAA-related events, although this was not the primary 5. Parr A, McCann M, Bradshaw B, Shahzad A, Buttner P, Golledge J.
Thrombus volume is associated with cardiovascular events and
outcome of this study. aneurysm growth in patients who have abdominal aortic aneurysms.
The current study has a number of strengths and weak- J Vasc Surg 2011;53:28-35.
nesses. Strengths included the relatively large sample of 6. Nastasi DR, Norman R, Moxon JV, Quigley F, Velu R, Jenkins J, et al.
The potential benefits and costs of an intensified approach to low
a uniform population, the consistent assessment of density lipoprotein cholesterol lowering in people with abdominal
markers and clinical risk factors, and the variety of statis- aortic aneurysm. Eur J Vasc Endovasc Surg 2021;62:643-50.
tical analyses employed. Weaknesses included the 7. Eikelboom JW, Connolly SJ, Bosch J, Dagenais GR, Hart RG,
Shestakovska O, et al. Rivaroxaban with or without aspirin in stable
limited number of biomarkers examined, the limited cardiovascular disease. N Engl J Med 2017;377:1319-30.
follow-up, and the lack of data on repeat aortic imaging. 8. Ridker PM, Everett BM, Thuren T, MacFadyen JG, Chang WH,
The incidence of MACE was slightly lower than included Ballantyne C, et al. Antiinflammatory therapy with canakinumab for
atherosclerotic disease. N Engl J Med 2017;377:1119-31.
in our sample size calculation, and so it is likely our study 9. Ridker PM, Devalaraja M, Baeres FMM, Engelmann MDM,
was underpowered. Also, biomarkers were only Hovingh GK, Ivkovic M, et al. IL-6 inhibition with ziltivekimab in pa-
measured on one occasion. Furthermore, all patients tients at high atherosclerotic risk (RESCUE): a double-blind, rando-
mised, placebo-controlled, phase 2 trial. Lancet 2021;397:2060-9.
were recruited from hospitals, and therefore, the findings 10. Chatterjee M, Ehrenberg A, Toska LM, Metz LM, Klier M, Krueger I,
may not be relevant to people with less severe disease et al. Molecular drivers of platelet activation: unraveling novel targets
not referred to hospital. There is no national screening for anti-thrombotic and anti-thrombo-inflammatory therapy. Int J
Mol Sci 2020;21:7906.
program for AAA in Australia, and thus the included par- 11. Kremers B, Wubbeke L, Mees B, Ten Cate H, Spronk H, Ten Cate-
ticipants were identified through incidental imaging and Hoek A. Plasma biomarkers to predict cardiovascular outcome in
1044 Golledge et al Journal of Vascular Surgery
April 2023

patients with peripheral artery disease: a systematic review and 21. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR.
meta-Analysis. Arterioscler Thromb Vasc Biol 2020;40:2018-32. A simulation study of the number of events per variable in logistic
12. Saenz-Pipaon G, Martinez-Aguilar E, Orbe J, Gonzalez Miqueo A, regression analysis. J Clin Epidemiol 1996;49:1373-9.
Fernandez-Alonso L, Paramo JA, et al. The role of circulating bio- 22. MacCallum RC, Zhang S, Preacher KJ, Rucker DD. On the practice of
markers in peripheral arterial disease. Int J Mol Sci 2021;22:3601. dichotomization of quantitative variables. Psychol Methods 2002;7:
13. Krishna SM, Moxon JV, Golledge J. A review of the pathophysiology 19-40.
and potential biomarkers for peripheral artery disease. Int J Mol Sci 23. Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO 3rd,
2015;16:11294-322. Criqui M, et al. Markers of inflammation and cardiovascular disease:
14. Singh TP, Morris DR, Smith S, Moxon JV, Golledge J. Systematic re- application to clinical and public health practice: a statement for
view and meta-analysis of the association between C-reactive pro- healthcare professionals from the Centers for disease control and
tein and major cardiovascular events in patients with peripheral prevention and the American Heart Association. Circulation
artery disease. Eur J Vasc Endovasc Surg 2017;54:220-33. 2003;107:499-511.
15. Golledge J, Cronin O, Iyer V, Bradshaw B, Moxon JV, 24. Trevethan R. Sensitivity, specificity, and predictive values: founda-
Cunningham MA. Body mass index is inversely associated with tions, pliabilities, and pitfalls in research and practice. Front Public
mortality in patients with peripheral vascular disease. Atherosclerosis Health 2017;5:307.
2013;229:549-55. 25. Sidloff DA, Saratzis A, Thompson J, Katsogridakis E, Bown MJ. Editor’s
16. Golledge J, Ewels C, Muller R, Walker PJ. Association of chronic kid- choice - infra-renal aortic diameter and cardiovascular risk: making
ney disease categories defined with different formulae with major better use of abdominal aortic aneurysm screening outcomes. Eur J
adverse events in patients with peripheral vascular disease. Athero- Vasc Endovasc Surg 2021;62:38-45.
sclerosis 2014;232:289-97. 26. Golledge J, Tsao PS, Dalman RL, Norman PE. Circulating markers of
17. Ferguson CD, Clancy P, Bourke B, Walker PJ, Dear A, Buckenham T, abdominal aortic aneurysm presence and progression. Circulation
et al. Association of statin prescription with small abdominal aortic 2008;118:2382-92.
aneurysm progression. Am Heart J 2010;159:307-13. 27. Danesh J, Kaptoge S, Mann AG, Sarwar N, Wood A, Angleman SB,
18. Smith RK, Quigley F, Tosenovsky P, Velu R, Bradshaw B, Buettner P, et al. Long-term interleukin-6 levels and subsequent risk of coronary
et al. Serum homocysteine is associated with the severity of primary heart disease: two new prospective studies and a systematic review.
chronic venous disease. Phlebology 2016;31:409-15. PLoS Med 2008;5:e78.
19. Thomas Manapurathe D, Moxon JV, Krishna SM, Rowbotham S, 28. Harrison SC, Smith AJ, Jones GT, Swerdlow DI, Rampuri R, Bown MJ,
Quigley F, Jenkins J, et al. Cohort study examining the association et al. Interleukin-6 receptor pathways in abdominal aortic aneurysm.
between blood pressure and cardiovascular events in patients with Eur Heart J 2013;34:3707-16.
peripheral artery disease. J Am Heart Assoc 2019;8:e010748.
20. Golledge J, Rowbotham S, Velu R, Quigley F, Jenkins J, Bourke M, Submitted Sep 10, 2022; accepted Dec 1, 2022.
et al. Association of serum lipoprotein (a) with the requirement for a
peripheral artery operation and the incidence of major adverse
cardiovascular events in people with peripheral artery disease. J Am The CME exam for this article can be accessed at http://
Heart Assoc 2020;9:e015355. www.jvascsurg.org/cme/home.