Introduction
In patients with stable coronary artery disease (SCAD), systemic biomarkers may improve
risk classification and should be selected according to their pathogenetic significance.
In the present study, we selected fibrinogen and neopterin, which both respond to
inflammation, known to be involved in thromboatherogenesis.
Fibrinogen is a large, soluble glycoprotein made up of three different polypeptides
(Aα, Bβ and γ-chains), and is mainly synthesized by hepatocytes.[1] It forms bridges between platelets during thrombus formation and is a precursor
of fibrin in which they are trapped, but it also regulates plasma viscosity by promoting
red blood cell aggregation.[2] Fibrin contributes to the growth of and is abundantly present in atheromatous lesions.
Subintimal fibrin attracts leukocytes, stimulates smooth muscle cell proliferation
and affects endothelial permeability and vascular tone.[2] Fibrinogen is an acute-phase protein and a marker of ongoing inflammation. Its synthesis
is stimulated by IL-6 and other proinflammatory cytokines,[1] and elevated levels have been reported in a variety of thrombotic and inflammatory
conditions, including cardiovascular disease.[1] Fibrinogen levels have been found to be independently associated with coronary artery
disease (CAD) among both healthy subjects and cardiovascular patients.[2]
[3]
[4]
[5] Its increase is usually considered to be secondary, but causality is also being
debated.[6]
Neopterin is an aromatic pteridine which is produced by the enzyme GTP cyclohydrolase
I in activated monocytes or macrophages within atherosclerotic plaques.[7] It is regarded as a sensitive marker of activation of cell-mediated immunity and
phagocytic activity,[7]
[8] which constitute important parts in the transformation from relatively stable to
more vulnerable lesions and unstable atheromatous plaques.[9]
[10] Furthermore, neopterin and its derivatives act as pro-oxidants and may play a role
in the inflammatory process related to atherosclerosis.[11] They also stimulate the expression of thromboplastin (CD 142),[12]
[13]
[14] and thereby promote thromboatheromatous growth. Higher levels of neopterin have
been found in many clinical conditions involving increased activity of monocytes or
macrophages, such as infectious diseases; malignancies and autoimmune-, renal- and
cardiovascular diseases.[15] Several reports link neopterin to the acute coronary syndrome (ACS),[16]
[17]
[18]
[19]
[20] more complex and extensive atherosclerotic lesions,[21]
[22] and to a higher probability of major adverse cardiac events.[23]
[24]
[25]
[26]
[27] In patients with SCAD, it has been found to be of some prognostic value.[28]
[29]
Neopterin's pro-oxidant properties and its association with inflammation, together
with that of fibrinogen, led us to assess the prognostic utility of these biomarkers
in an established SCAD population. Furthermore, their prognostic utility with respect
to future events was also assessed in patient groups stratified according to the level
of high-sensitivity troponin T (hsTnT), considered to be a risk factor in chronic
CAD patients,[30] and in relation to gender and smoking habits. Whether these markers interact or
act independently among patients with suspected SCAD has previously not been investigated.
We hypothesize that fibrinogen and neopterin or the two combined may correlate with
the long-term outcome in patients with SCAD.
Materials and Methods
Study Design and Patient Population
Patients were collected from the BECAC (Bergen Coronary Angiography Cohort) and the
WENBIT trials (Western Norway B Vitamin Intervention Trial, ClinicalTrials.gov Identifier:
NCT00354081).[31] Participants (n = 4,166) were included between 2000 and 2004, as previously described,[31]
[32] and all patients underwent coronary angiography. We lacked fibrinogen measurements
in 621 patients (mostly due to administrative reasons), leaving 3,545 SCAD patients
to be enrolled in the present study. However, patients without available fibrinogen
were statistically described as similar to the investigated cohort with respect to
baseline characteristics and clinical endpoints. Among the 3,545 SCAD patients, there
were 18 missing neopterin measurements. Written informed consent was obtained from
all participants. The study was approved by the Regional Committee for Medical Research
Ethics and the Norwegian Data Inspectorate, and conducted in accordance with the Declaration
of Helsinki.
Primary and secondary endpoints in the present study were acute myocardial infarction
(AMI; both fatal and nonfatal) and total mortality. AMI was classified according to
the revised definition of myocardial infarction from 2000.[33] Procedure-related nonfatal AMI occurring within the first 24 hours was excluded.
Follow-up of incident AMI lasted throughout 2009 and of total mortality throughout
2012. Information on AMI was collected from the Cardiovascular Disease in Norway project
(www.cvdnor.no)[34] and death from the Cause of Death Registry at Statistics Norway (www.ssb.no) and from the Western Norway Cardiovascular Registry.
Baseline Parameters
The patients underwent interview, clinical examination and blood sampling at baseline.
Clinical and laboratory parameters, as presented in the baseline tables and elsewhere,[32] were collected by study personnel. In addition, hospital journals were searched
for confirmation of reported data. Smoking was classified based on self-reports where
non-smokers include both ex-smokers and never-smokers. Left ventricular ejection fraction
(LVEF) was estimated either by echocardiography or by ventriculography during cardiac
catheterization. Coronary angiography was performed and evaluated by trained invasive
cardiologists. The angiograms were analysed in orthogonal views, and a significant
stenosis was defined as greater than 50% luminal diameter narrowing, and the extent
of CAD was described by adding the number of vessels with significant stenosis (0–3
vessels).
Blood Sampling and Laboratory Measurements
Blood samples were drawn at baseline (before angiography) and were collected and processed
by study personnel. Routine blood analyses were performed by the hospital laboratories.[31] Aliquots of serum and citrated plasma were immediately frozen at −80°C for further
analysis.
Fibrinogen was measured using quantitative determination according to Clauss, utilizing
the STA—Liquid Fib kit cat. number 00673 (Diagnostica Stago S.A.S., France), with
a lower detection limit of 0.4 g/L and an intra-assay and inter-assay coefficients
of variations (CVs) of 2.1 to 4.9% and 2.1 to 3.2%, respectively. Plasma concentrations
of neopterin were analysed by liquid chromatography-tandem mass spectrometry at Bevital
A/S (www.bevital.no, Bergen, Norway).[35] The cardiac troponin T was analysed using a high-sensitivity cardiac troponin T
(hsTnT) assay on Modular E170 from Roche Diagnostics, with a lower detection limit
of 3 ng/L.[36] High-sensitivity C-reactive protein (hsCRP) was determined in serum by an ultrasensitive
immunoassay, using the Behring nephelometer II system (CV: 8.1–11.4%; N Latex CRP
mono; Behring Diagnostic, Germany).
Statistics
Baseline characteristics are reported for all and for quartile (Q) subsets of patients
arranged according to fibrinogen and neopterin values. Normally distributed variables
were summarized as means and standard deviations (SD), whereas variables with skewed
distributions were reported as medians and quartiles. The chi-square and the Kruskal–Wallis
tests were applied to test for differences in categorical variables and for differences
in continuous variables across quartile groups, respectively. The Student t-test or Mann–Whitney test was used to test for differences in fibrinogen and neopterin
levels in subgroups according to gender and smoking habits depending on whether the
variable was normally distributed or not. Differences in logarithmically transformed
levels of the biomarkers in these subgroups were further evaluated in a linear regression
model including age.
Kaplan–Meier plots according to quartiles of fibrinogen and neopterin were used for
estimating the survival curves for times to event. The log-rank test was used to test
for differences in the survival curves. Cox proportional hazard regression analysis
was used to obtain hazard ratios (HRs; 95% confidence intervals [CI]) for incident
AMI per 1 SD increase of logarithmically transformed fibrinogen and neopterin levels,
in a crude model, a model adjusted for age and gender (multivariable Model 1), in
addition to a model adjusted for age, gender, diabetes, current smoking, hypertension,
estimated glomerular filtration rate (eGFR), serum levels of triglycerides, apolipoprotein
A1 and apolipoprotein B (multivariable Model 2). Potential effect modifications were
investigated according to strata (below or above the median) of fibrinogen, neopterin
or hsTnT, as well as according to gender and smoking status. Assumption for the proportional
hazard was confirmed by investigating the Schoenfeld residuals. Differences of the
area under the curve (AUC) based on Cox regression models were assessed using the
DeLong's test for two correlated receiver operator curves (ROC). A p-value <0.05 was considered significant. The statistical analysis were performed using
the statistical package SPSS version 23.0 and R version 3.3.1.[37]
Results
Baseline Characteristics
Baseline patient characteristics across quartiles of fibrinogen and neopterin are
presented in [Tables 1] and [2], respectively, and evaluation of correlates between these and the baseline variables
are shown in [Supplementary Tables S1] and [S2] (online only). Patients with higher fibrinogen levels had higher levels of neopterin,
hsTnT, hsCRP, leucocytes and thrombocytes, whereas eGFR, HDL cholesterol and apolipoprotein
A1 were lower. Furthermore, these patients were older, and a higher proportion were
current smokers, had known hypertension, diabetes mellitus, hypercholesterolemia,
known history of vascular disease (CAD, cerebrovascular disease and peripheral artery
disease) and heart failure, and a higher proportion used medication such as warfarin
and angiotensin-converting enzyme inhibitors, angiotensin receptor blockers and diuretics.
Similar associations were observed across quartiles of neopterin, except that neopterin
was inversely associated with smoking and hypercholesterolemia.
Table 1
Baseline characteristics according to fibrinogen quartiles
|
Variable
|
n
|
Total
|
Quartiles of fibrinogen
|
|
|
Cohort
|
Q1
|
Q2
|
Q3
|
Q4
|
p-Value[a]
|
|
Age, y
|
3,545
|
62.0 (55.0, 70.0)
|
59.0 (53.0, 67.0)
|
61.0 (54.5, 69.0)
|
64.0 (56.0, 71.0)
|
64.0 (57.0, 71.0)
|
<0.001
|
|
Male gender, n (%)
|
3,545
|
2,503 (70.6)
|
632 (74.5)
|
579 (70.9)
|
669 (69.5)
|
623 (67.9)
|
0.017
|
|
Risk markers at baseline
|
|
Fibrinogen (g/L)
|
3,545
|
3.60 (3.20, 4.10)
|
2.90 (2.70, 3.00)
|
3.30 (3.20, 3.50)
|
3.80 (3.70, 3.90)
|
4.40 (4.20, 4.80)
|
<0.001
|
|
Neopterin (nmol/L)
|
3,527
|
8.22 (6.69, 10.49)
|
7.62 (6.41, 9.40)
|
8.03 (6.71, 10.05)
|
8.33 (6.72, 10.57)
|
9.04 (6.96, 12.20)
|
<0.001
|
|
hsTnT (ng/L)
|
3,475
|
5.00 (3.00, 10.00)
|
3.00 (3.00, 7.00)
|
4.00 (3.00, 9.00)
|
5.00 (3.00, 11.00)
|
6.00 (3.00, 13.00)
|
<0.001
|
|
hsCRP (mg/L)
|
3,542
|
1.79 (0.87, 3.69)
|
0.92 (0.53, 1.79)
|
1.44 (0.77, 2.64)
|
1.88 (1.03, 3.62)
|
3.93 (2.05, 8.22)
|
<0.001
|
|
Leucocytes (×106/mL)
|
3,543
|
6.8 (5.7, 8.2)
|
6.3 (5.2, 7.4)
|
6.5 (5.5, 7.7)
|
7.0 (5.8, 8.2)
|
7.7 (6.4, 9.3)
|
<0.001
|
|
Monocyte fraction
|
3,438
|
0.08 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.07 (0.06, 0.09)
|
0.308
|
|
TRC (109/L)
|
3,521
|
239 (204, 281)
|
233 (201, 270)
|
234 (198, 274)
|
239 (203, 277)
|
255 (215, 302)
|
<0.001
|
|
Risk factors
|
|
Current smoking, n (%)
|
3,538
|
816 (23.1)
|
149 (17.6)
|
165 (20.2)
|
212 (22.1)
|
290 (31.7)
|
<0.001
|
|
Hypertension, n (%)
|
3,545
|
1,677 (47.3)
|
331 (39.0)
|
358 (43.8)
|
476 (49.5)
|
512 (55.8)
|
<0.001
|
|
IDDM, n (%)
|
3,545
|
36 (1.0)
|
10 (1.2)
|
8 (1.0)
|
10 (1.0)
|
8 (0.9)
|
<0.001[b]
|
|
NIDDM, n (%)
|
3,545
|
394 (11.1)
|
73 (8.6)
|
81 (9.9)
|
99 (10.3)
|
141 (15.4)
|
<0.001[b]
|
|
eGFR (mL/min per 1.73 m2)
|
3,539
|
90.0 (78.0, 99.0)
|
92.0 (82.0, 101.0)
|
90.0 (80.0, 99.0)
|
89.0 (77.0, 98.0)
|
88.0 (73.0, 98.0)
|
<0.001
|
|
Hypercholesterolemia, n (%)
|
3,545
|
1,892 (57.4)
|
424 (53.3)
|
432 (55.5)
|
530 (59.9)
|
506 (60.5)
|
0.007
|
|
BMI, kg/m2
|
3,542
|
26.0 (24.0, 29.0)
|
25.0 (23.0, 28.0)
|
26.0 (24.0, 28.0)
|
26.0 (24.0,29.0)
|
26.0 (24.0,29.0)
|
<0.001
|
|
Serum lipids and apolipoproteins
|
|
Total cholesterol (mmol/L)
|
3,544
|
4.90 (4.30, 5.70)
|
5.00 (4.23, 5.80)
|
5.00 (4.30, 5.70)
|
4.90 (4.30, 5.70)
|
4.90 (4.28, 5.70)
|
0.690
|
|
LDL (mmol/L)
|
3,543
|
2.97 (2.40, 3.70)
|
2.93 (2.30, 3.70)
|
3.00 (2.40, 3.70)
|
2.95 (2.40,3.70)
|
2.90 (2.40, 3.70)
|
0.865
|
|
HDL (mmol/L)
|
3,545
|
1.20 (1.00, 1.50)
|
1.30 (1.10, 1.50)
|
1.30 (1.10, 1.50)
|
1.20 (1.00, 1.50)
|
1.20 (1.00, 1.50)
|
<0.001
|
|
Triglycerides (mmol/L)
|
3,542
|
1.49 (1.07, 2.11)
|
1.51 (1.05, 2.20)
|
1.42 (1.06, 2.19)
|
1.51 (1.10, 2.08)
|
1.46 (1.05, 1.98)
|
0.545
|
|
ApoB, g/L
|
3,543
|
0.87 (0.73, 1.05)
|
0.85 (0.71, 1.04)
|
0.86 (0.73, 1.03)
|
0.87 (0.74, 1.04)
|
0.88 (0.75, 1.07)
|
0.023
|
|
ApoA1, g/L
|
3,542
|
1.31 (1.14, 1.50)
|
1.34 (1.18, 1.52)
|
1.34 (1.18, 1.51)
|
1.29 (1.14, 1.49)
|
1.27 (1.11, 1.47)
|
<0.001
|
|
History of cardiovascular disease
|
|
Previous AMI, n (%)
|
3,545
|
1,386 (39.1)
|
306 (36.1)
|
299 (36.6)
|
382 (39.7)
|
399 (43.5)
|
0.005
|
|
Previous CBV, n (%)
|
3,545
|
254 (7.2)
|
43 (5.1)
|
55 (6.7)
|
59 (6.1)
|
97 (10.6)
|
<0.001
|
|
Previous PAD, n (%)
|
3,545
|
312 (8.8)
|
51 (6.0)
|
47 (5.8)
|
96 (10.0)
|
118 (12.9)
|
<0.001
|
|
Previous PCI, n (%)
|
3,545
|
665 (18.8)
|
162 (19.1)
|
152 (18.6)
|
169 (17.6)
|
182 (19.8)
|
0.646
|
|
Previous CABG, n (%)
|
3,545
|
399 (11.3)
|
90 (10.6)
|
90 (11.0)
|
111 (11.5)
|
108 (11.8)
|
0.870
|
|
Known CHF, n (%)
|
3,545
|
242 (6.8)
|
37 (4.4)
|
38 (4.7)
|
63 (6.5)
|
104 (11.3)
|
<0.001
|
|
LVEF, %
|
3,545
|
66 (60, 70)
|
69 (60, 70)
|
68 (60, 70)
|
67 (60, 70)
|
65 (59, 70)
|
<0.001
|
|
Extent of CAD at baseline
|
|
No stenotic vessels, n (%)
|
3,545
|
949 (26.8)
|
284 (33.5)
|
239 (29.3)
|
229 (23.8)
|
197 (21.5)
|
<0.001[c]
|
|
One-vessel disease, n (%)
|
3,545
|
811 (22.9)
|
192 (22.6)
|
178 (21.8)
|
227 (23.6)
|
214 (23.3)
|
<0.001[c]
|
|
Two-vessel disease, n (%)
|
3,545
|
773 (21.8)
|
168 (19.8)
|
175 (21.4)
|
231 (24.0)
|
199 (21.7)
|
<0.001[c]
|
|
Three-vessel disease, n (%)
|
3,545
|
1,012(28.5)
|
204 (24.1)
|
225 (27.5)
|
275 (28.6)
|
308 (33.6)
|
<0.001[c]
|
|
Treatment following baseline coronary angiography
|
|
No specific treatment
|
3,544
|
237 (6.7)
|
91 (10.7)
|
80 (9.8)
|
44 (4.6)
|
22 (2.4)
|
<0.001[d]
|
|
Medical therapy only
|
3,544
|
1,432 (40.4)
|
346 (40.8)
|
315 (38.6)
|
376 (39.1)
|
395 (43.1)
|
<0.001[d]
|
|
PCI
|
3,544
|
1,120 (31.6)
|
256 (30.2)
|
246 (30.1)
|
329 (34.2)
|
289 (31.5)
|
<0.001[d]
|
|
CABG
|
3,544
|
671 (18.9)
|
141 (16.6)
|
159 (19.5)
|
189 (19.6)
|
182 (19.8)
|
<0.001[d]
|
|
Combination of surgical treatments
|
3,544
|
84 (2.4)
|
14 (1.7)
|
17 (2.1)
|
24 (2.5)
|
29 (3.2)
|
<0.001[d]
|
|
Medication prior to inclusion
|
|
Acetylsalicylic acid, n (%)
|
3,545
|
2,869 (80.9)
|
668 (78.8)
|
656 (80.3)
|
799 (83.1)
|
746 (81.3)
|
0.130
|
|
ADP receptor antagonist, n (%)
|
3,545
|
551 (15.5)
|
134 (15.8)
|
126 (15.4)
|
155 (16.1)
|
136 (14.8)
|
0.882
|
|
Warfarin, n (%)
|
3,545
|
196 (5.5)
|
25 (2.9)
|
36 (4.4)
|
49 (5.1)
|
86 (9.4)
|
<0.001
|
|
Statin, n (%)
|
3,534
|
2,806 (79.4)
|
642 (75.9)
|
631 (77.4)
|
783 (81.5)
|
750 (82.2)
|
0.016
|
|
Beta-blocker, n (%)
|
3,545
|
2,548 (71.9)
|
578 (68.2)
|
567 (69.4)
|
729 (75.8)
|
674 (73.4)
|
0.001
|
|
ACE-inhibitor, n (%)
|
3,545
|
725 (20.5)
|
141 (16.6)
|
154 (18.8)
|
187 (19.4)
|
243 (26.5)
|
<0.001
|
|
ARB, n (%)
|
3,545
|
413 (11.7)
|
81 (9.6)
|
90 (11.0)
|
112 (11.6)
|
130 (14.2)
|
0.023
|
|
Loop diuretics, n (%)
|
3,545
|
394 (11.1)
|
49 (5.8)
|
69 (8.4)
|
108 (11.2)
|
168 (18.3)
|
<0.001
|
|
Thiazide diuretic, n (%)
|
3,545
|
253 (7.1)
|
44 (5.2)
|
56 (6.9)
|
63 (6.5)
|
90 (9.8)
|
0.002
|
Abbreviations: ACE, angiotensin-converting enzyme; ADP, adenosine diphosphate; AMI,
acute myocardial infarction; ApoA1, apolipoprotein A1; ApoB, apolipoprotein B; ARB
angiotensin receptor blocker; BMI, body mass index; BNP, B-type natriuretic peptide;
CABG, coronary artery bypass graft; CBV, cerebrovascular disease; CHF, congestive
heart failure; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate;
HDL, high-density lipoprotein cholesterol; IDDM, insulin-dependent diabetes mellitus;
LDL, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction;
NIDDM, non–insulin-dependent diabetes mellitus; PAD, peripheral artery disease; PCI,
percutaneous coronary intervention; TNT, troponin T; TRC, platelets.
a
p-Values for chi-square tests of difference in categorical variables across quartile
groups, and for Kruskal–Wallis tests for differences in continuous variables across
quartile groups.
b
p-Value for chi-square test comparing IDDM, NIDM and no history of diabetes mellitus.
c
p-Value for chi-square test comparing no stenotic vessels—one, two and three-vessel
disease.
d
p-Value for chi-square test comparing the different treatments given following baseline
coronary angiography.
Note: Continuous variables are reported as median (25th–75th percentile), and categorical variables as counts (percentage of total).
Table 2
Baseline characteristics according to neopterin quartiles
|
Variable
|
n
|
Total
|
Quartiles of neopterin[a]
|
p-Value[b]
|
|
Cohort
|
Q1
|
Q2
|
Q3
|
Q4
|
|
Age, y
|
3,545
|
62.0 (55.0, 70.0)
|
58.0 (51.0, 64.0)
|
59.5 (53.0, 66.0)
|
64.0 (57.0, 71.0)
|
69.0 (60.0, 75.0)
|
<0.001
|
|
Male gender, n (%)
|
3,545
|
2503 (70.6)
|
699 (79.3)
|
649 (73.6)
|
598 (67.8)
|
547 (62.0)
|
<0.001
|
|
Risk markers at baseline
|
|
Neopterin (nmol/L)
|
3,527
|
8.22 (6.69, 10.49)
|
5.91 (5.30, 6.32)
|
7.41 (7.02, 7.80)
|
9.19 (8.65, 9.69)
|
12.72 (11.45, 15.38)
|
<0.001
|
|
Fibrinogen (g/L)
|
3,545
|
3.60 (3.20, 4.10)
|
3.50 (3.10, 4.00)
|
3.50 (3.10, 4.00)
|
3.60 (3.20, 4.00)
|
3.80 (3.30, 4.40)
|
<0.001
|
|
hsTnT (ng/L)
|
3,475
|
5.00 (3.00, 10.00)
|
3.00 (3.00, 7.00)
|
3.00 (3.00, 8.00)
|
5.00 (3.00, 10.00)
|
8.00 (3.00, 17.00)
|
<0.001
|
|
hsCRP (mg/L)
|
3,542
|
1.79 (0.87, 3.69)
|
1.46 (0.76, 2.69)
|
1.69 (0.79, 3.26)
|
1.82 (0.91, 3.70)
|
2.48 (1.16, 13.22)
|
<0.001
|
|
Leucocytes (×106/mL)
|
3,543
|
6.8 (5.7, 8.2)
|
6.8 (5.7, 8.2)
|
6.7 (5.6, 7.9)
|
6.8 (5.8, 8.0)
|
6.9 (5.8, 8.4)
|
0.044
|
|
Monocyte fraction
|
3,438
|
0.08 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.07 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.08 (0.06, 0.09)
|
0.010
|
|
TRC (109/L)
|
3,521
|
239 (204, 281)
|
242 (206, 282)
|
243 (207, 281)
|
238 (202, 282)
|
238 (201, 280)
|
0.476
|
|
Risk factors
|
|
Current smoking, n (%)
|
3,538
|
816 (23.1)
|
268 (30.5)
|
212 (24.1)
|
184 (20.9)
|
149 (16.9)
|
<0.001
|
|
Hypertension, n (%)
|
3,545
|
1677 (47.3)
|
381 (43.2)
|
384 (43.5)
|
404 (45.8)
|
499 (56.6)
|
<0.001
|
|
IDDM, n (%)
|
3,545
|
36 (1.0)
|
13 (1.5)
|
7 (0.8)
|
5 (0.6)
|
11 (1.2)
|
0.328[c]
|
|
NIDDM, n (%)
|
3,545
|
394 (11.1)
|
98 (11.1)
|
94 (10.7)
|
89 (10.1)
|
108 (12.2)
|
0.328[c]
|
|
eGFR (mL/min per 1.73 m2)
|
3,539
|
90.0 (78.0, 99.0)
|
97.0 (90.0, 105.0)
|
93.0 (86.0, 101.0)
|
87.0 (76.0, 96.0)
|
75.0 (61.0, 88.0)
|
<0.001
|
|
Hypercholesterolemia, n (%)
|
3,545
|
1892 (57.4)
|
519 (61.8)
|
474 (57.2)
|
453 (55.4)
|
439 (55.2)
|
0.023
|
|
BMI, kg/m2
|
3,542
|
26.0 (24.0, 29.0)
|
26.0 (24.0, 29.0)
|
26.0 (24.0, 29.0)
|
26.0 (23.0,28.0)
|
26.0 (23.0,28.0)
|
0.003
|
|
Serum lipids and apolipoproteins
|
|
Total cholesterol (mmol/L)
|
3,544
|
4.90 (4.30, 5.70)
|
5.00 (4.30, 5.80)
|
4.90 (4.30, 5.70)
|
4.90 (4.30, 5.70)
|
4.80 (4.20, 5.70)
|
0.139
|
|
LDL (mmol/L)
|
3,543
|
2.97 (2.40, 3.70)
|
3.00 (2.48, 3.70)
|
2.97 (2.40, 3.70)
|
3.00 (2.40,3.70)
|
2.90 (2.30, 3.70)
|
0.152
|
|
HDL (mmol/L)
|
3,545
|
1.20 (1.00, 1.50)
|
1.20 (1.10, 1.50)
|
1.20 (1.00, 1.50)
|
1.30 (1.10, 1.50)
|
1.20 (1.00, 1.50)
|
0.004
|
|
Triglycerides (mmol/L)
|
3,542
|
1.49 (1.07, 2.33)
|
1.60 (1.12, 2.33)
|
1.53 (1.09, 2.19)
|
1.40 (1.05, 1.95)
|
1.43 (1.04, 1.98)
|
<0.001
|
|
ApoB, g/L
|
3,543
|
0.87 (0.73, 1.05)
|
0.88 (0.74, 1.05)
|
0.86 (0.74, 1.05)
|
0.87 (0.72, 1.04)
|
0.87 (0.74, 1.06)
|
0.777
|
|
ApoA1, g/L
|
3,542
|
1.31 (1.14, 1.50)
|
1.31 (1.16, 1.48)
|
1.31 (1.15, 1.48)
|
1.33 (1.15, 1.53)
|
1.28 (1.12, 1.50)
|
0.005
|
|
History of cardiovascular disease
|
|
Previous AMI, n (%)
|
3,545
|
1,386 (39.1)
|
338 (38.4)
|
343 (38.9)
|
316 (35.8)
|
383 (43.4)
|
0.011
|
|
Previous CBV n (%)
|
3,545
|
254 (7.2)
|
45 (5.1)
|
42 (4.8)
|
59 (6.7)
|
107 (12.1)
|
<0.001
|
|
Previous PAD n (%)
|
3,545
|
312 (8.8)
|
64 (7.3)
|
57 (6,5)
|
88 (10.0)
|
102 (11.6)
|
<0.001
|
|
Previous PCI, n (%)
|
3,545
|
665 (18.8)
|
174 (19.8)
|
174 (19.7)
|
147 (16.7)
|
170 (19.3)
|
0.288
|
|
Previous CABG, n (%)
|
3,545
|
399 (11.3)
|
80 (9.1)
|
96 (10.9)
|
108 (12.2)
|
113 (12.8)
|
0.063
|
|
Known CHF, n (%)
|
3,545
|
242 (6.8)
|
33 (3.7)
|
44 (5.0)
|
69 (7.8)
|
95 (10.8)
|
<0.001
|
|
LVEF, %
|
3545
|
66 (60, 70)
|
69 (60, 70)
|
66 (60, 70)
|
66 (60, 70)
|
65 (56, 70)
|
<0.001
|
|
Extent of CAD at baseline
|
|
No stenotic vessels, n (%)
|
3,545
|
949 (26.8)
|
235 (26.7)
|
232 (26.3)
|
258 (29.3)
|
221 (25.1)
|
0.010[d]
|
|
One-vessel disease, n (%)
|
3,545
|
811 (22.9)
|
218 (24.7)
|
217 (24.6)
|
187 (21.2)
|
182 (20.6)
|
0.010[d]
|
|
Two-vessel disease, n (%)
|
3,545
|
773 (21.8)
|
206 (23.4)
|
188 (21.3)
|
193 (21.9)
|
183 (20.7)
|
0.010[d]
|
|
Three-vessel disease, n (%)
|
3,545
|
1,012(28.5)
|
222 (25.2)
|
245 (27.8)
|
244 (27.7)
|
296 (33.6)
|
0.010[d]
|
|
Treatment following baseline coronary angiography
|
|
No specific treatment
|
3,544
|
237 (6.7)
|
63 (7.2)
|
61 (6.9)
|
62 (7.0)
|
50 (5.7)
|
0.134[e]
|
|
Medical therapy only
|
3,544
|
1,432 (40.4)
|
339 (38.5)
|
344 (39.0)
|
365 (41.4)
|
378 (42.9)
|
0.134[e]
|
|
PCI
|
3,544
|
1,120 (31.6)
|
283 (32.1)
|
280 (31.7)
|
286 (32.4)
|
263 (29.9)
|
0.134[e]
|
|
CABG
|
3,544
|
671 (18.9)
|
173 (19.6)
|
186 (21.1)
|
147 (16.7)
|
163 (18.5)
|
0.134[e]
|
|
Combination of surgical treatments
|
3,544
|
84 (2.4)
|
23 (2.6)
|
11 (1.2)
|
22 (2.5)
|
27 (3.1)
|
0.134[e]
|
|
Medication prior to inclusion
|
|
Acetylsalicylic acid, n (%)
|
3,545
|
2,869 (80.9)
|
733 (83.2)
|
742 (84.1)
|
702 (79.6)
|
677 (76.8)
|
<0.001
|
|
ADP receptor antagonist, n (%)
|
3,545
|
551 (15.5)
|
138 (15.7)
|
137 (15.5)
|
134 (15.2)
|
137 (15.5)
|
0.994
|
|
Warfarin, n (%)
|
3,545
|
196 (5.5)
|
42 (4.8)
|
26 (2.9)
|
53 (6.0)
|
75 (8.5)
|
<0.001
|
|
Statin, n (%)
|
3,534
|
2,806 (79.4)
|
710 (80.7)
|
721 (81.8)
|
679 (77.3)
|
680 (77.5)
|
0.165
|
|
Beta-blocker, n (%)
|
3,545
|
2,548 (71.9)
|
627 (71.2)
|
637 (72.2)
|
609 (69.0)
|
660 (74.8)
|
0.056
|
|
ACE inhibitor, n (%)
|
3,545
|
725 (20.5)
|
157 (17.8)
|
148 (16.8)
|
171 (19.4)
|
246 (27.9)
|
<0.001
|
|
ARB, n (%)
|
3,545
|
413 (11.7)
|
86 (9.8)
|
102 (11.6)
|
107 (12.1)
|
117 (13.3)
|
0.140
|
|
Loop diuretics, n (%)
|
3,545
|
394 (11.1)
|
45 (5.1)
|
56 (6.3)
|
93 (10.5)
|
200 (22.7)
|
<0.001
|
|
Thiazide diuretic, n (%)
|
3,545
|
253 (7.1)
|
40 (4.5)
|
60 (6.8)
|
69 (7.8)
|
82 (9.3)
|
0.001
|
Abbreviations: ACE inhibitor, angiotensin-converting enzyme; ADP, adenosine diphosphate;
AMI, acute myocardial infarction; ApoA1, apolipoprotein A1; ApoB, apolipoprotein B;
ARB angiotensin receptor blocker; BMI, body mass index; BNP, B-type natriuretic peptide;
CABG, coronary artery bypass graft; CBV, cerebrovascular disease; CHF, congestive
heart failure; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate;
HDL, high-density lipoprotein cholesterol; IDDM, insulin-dependent diabetes mellitus;
LDL, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction;
NIDDM, non–insulin-dependent diabetes mellitus; PAD, peripheral artery disease; PCI,
percutaneous coronary intervention; TNT, troponin T; TRC, platelets.
a Patients with valid measurements of neopterin.
b
p-Values for chi-square tests of difference in categorical variables across quartile
groups, and for Kruskal–Wallis tests for differences in continuous variables across
quartile groups.
c
p-Value for chi-square test comparing IDDM, NIDM and no history of diabetes mellitus.
d
p-Value for chi-square test comparing no stenotic vessels—one-, two- and three-vessel
disease.
e
p-Value for chi-square test comparing the different treatments given following baseline
coronary angiography.
Note: Continuous variables are reported as median (25th–75th percentile), and categorical variables as counts (percentage of total).
Baseline fibrinogen and neopterin levels were modestly positively correlated (Spearman's
rho = 0.18, p-value < 0.001). Fibrinogen was strongly correlated with hsCRP (Spearman's rho = 0.50,
p-value < 0.001), and modestly to age (r = 0.18), hsTnT (r = 0.17), thrombocyte count (r = 0.14). Fibrinogen correlated negatively to eGFR (r = − 0.13). Neopterin was strongly positively associated with eGFR (r = 0.52, p-value < 0.001), and modestly also to age (r = 0.36), hsTnT (r = 0.26) and hsCRP (r = 0.19).
Patients with higher serum hsTnT had more extensive CAD, as assessed by baseline coronary
angiography, when compared with those with lower hsTnT ([Supplementary Table S3] [online only]). In Q1 of hsTnT, 36.5% (n = 542) had no significant lesions and 18.4% (n = 274) had three vessel disease, in contrast to 14.1% (n = 115) versus 43.2% (n = 353), respectively, in the upper quartile (p < 0.001). A similar relationship with CAD was noted across fibrinogen quartiles.
These data are displayed in [Supplementary Tables S4] and [S5] (online only).
Supplementary Subgroup Baseline Information
Differences according to gender and smoking status are depicted in [Supplementary Table S6] (online only). Females had higher levels of fibrinogen (p = 0.008) and neopterin (p < 0.001), and lower levels of hsTnT level (p < 0.001), as compared with males. Smokers had a higher median concentration of fibrinogen
(p < 0.001) but lower levels of neopterin (p < 0.001) and hsTnT (p < 0.001), as compared with non-smokers. The proportion of female and male smokers
was not statistically different, but current smokers were younger than non-smokers.
In females, fibrinogen and neopterin levels were statistically higher in the presence
of significant CAD (≥1 vessel disease, as compared with patients with no significant
lesions; p < 0.001 and p = 0.001, respectively), whereas in men such difference was noted only for fibrinogen
(p < 0.001). Median concentrations of fibrinogen were also higher, irrespective of smoking
habits, when comparing patients with no stenotic vessels to those with one- to two-vessel
disease, whereas a similar relationship was not found for neopterin.
Acute Myocardial Infarction
A total of 543 patients suffered an AMI during a median (25th and 75th percentiles)
follow-up time of 7.3 (6.3, 8.6) years. Kaplan–Meier plots for patient groups divided
according to quartiles of fibrinogen and neopterin are presented in [Fig. 1]. Patients with fibrinogen or neopterin levels in the upper quartile had a shorter
time to incident AMI (p-value for log-rank test for survival curves is < 0.001 for both biomarkers).
Fig. 1 Kaplan–Meier plots for the cumulative risk of AMI (A and B) and total mortality (C and D) during follow-up stratified by fibrinogen and neopterin quartiles. Numbers at risk
displayed for each time point. p-Value represents log-rank.
In the univariable analysis, fibrinogen ([Table 3]) and neopterin ([Table 4]) were strongly associated with the primary endpoint AMI (HRs [95% CIs] per 1 SD
increment of logarithmically transformed values: 1.30 [1.20, 1.42], p < 0.001 and 1.31 [1.23, 1.40], p < 0.001, respectively). This association remained statistically significant for both
biomarkers after adjusting for possible confounders in our multivariable models ([Tables 3] and [4]). For both biomarkers, the association with AMI risk was essentially linear across
the whole distribution ([Fig. 2]). The AUC for the ROC for covariates applied to multivariable Model 2, without fibrinogen
and neopterin, was 0.650, which increased to 0.665 after inclusion of both fibrinogen
and neopterin (p = 0.007; [Table 5]; [Supplementary Figs. S1] and [S2] [online only]).
Table 3
Association of fibrinogen and covariates with incident acute myocardial infarction
|
Univariable[a]
|
Multivariable Model 1[a]
|
Multivariable Model 2[a]
|
|
HR (95% CI)
|
p-Value
|
HR (95% CI)
|
P-value
|
HR (95% CI)
|
p-Value
|
|
Fibrinogen[b]
|
1.30 (1.20, 1.42)
|
<0.001
|
1.26 (1.15, 1.37)
|
<0.001
|
1.18 (1.08, 1.29)
|
<0.001
|
|
Age[c]
|
|
|
1.40 (1.28, 1.54)
|
<0.001
|
1.40 (1.25, 1.57)
|
<0.001
|
|
Gender, male
|
|
|
1.41 (1.16, 1.72)
|
<0.001
|
1.34 (1.09, 1.65)
|
0.006
|
|
IDDM
|
|
|
|
|
2.64 (1.37, 5.08)
|
0.004
|
|
NIDDM
|
|
|
|
|
1.45 (1.15, 1.84)
|
0.002
|
|
Current smoker
|
|
|
|
|
1.65 (1.35, 2.01)
|
<0.001
|
|
Hypertension
|
|
|
|
|
1.11 (0.93, 1.32)
|
0.266
|
|
TG[c]
|
|
|
|
|
1.05 (0.97, 1.14)
|
0.185
|
|
ApoB[c]
|
|
|
|
|
1.01 (0.92, 1.10)
|
0.861
|
|
ApoA1[c]
|
|
|
|
|
0.91 (0.83, 1.01)
|
0.064
|
|
eGFR[c]
|
|
|
|
|
0.89 (0.80, 0.98)
|
0.016
|
Abbreviations: ApoA1, apolipoprotein A1; ApoB, apolipoprotein B; eGFR, estimated glomerular
filtration rate; IDDM, insulin-dependent diabetes mellitus; NIDDM, non–insulin-dependent
diabetes mellitus; SD, standard deviation; TG, triglycerides.
a Estimated by Cox proportional hazard regression analysis.
b Per 1 SD increase in logarithmically transformed fibrinogen levels.
c Per 1 SD increase.
Table 4
Association of neopterin and covariates with incident acute myocardial infarction
|
Univariable[a]
|
Multivariable Model 1[a]
|
Multivariable Model 2[a]
|
|
HR (95% CI)
|
p-Value
|
HR (95% CI)
|
p-Value
|
HR (95% CI)
|
p-Value
|
|
Neopterin[b]
|
1.31 (1.23, 1.40)
|
<0.001
|
1.24 (1.16, 1.34)
|
<0.001
|
1.21 (1.10, 1.33)
|
<0.001
|
|
Age[c]
|
|
|
1.35 (1.23, 1.48)
|
<0.001
|
1.44 (1.29, 1.62)
|
<0.001
|
|
Sex, male
|
|
|
1.44 (1.18, 1.76)
|
<0.001
|
1.33 (1.08, 1.65)
|
0.007
|
|
IDDM
|
|
|
|
|
2.30 (1.18, 4.49)
|
0.014
|
|
NIDDM
|
|
|
|
|
1.51 (1.19, 1.92)
|
<0.001
|
|
Current smoker
|
|
|
|
|
1.73 (1.42, 2.10)
|
<0.001
|
|
Hypertension
|
|
|
|
|
1.13 (0.94, 1.34)
|
0.189
|
|
TG[c]
|
|
|
|
|
1.04 (0.96, 1.13)
|
0.318
|
|
ApoB[c]
|
|
|
|
|
1.01 (0.92, 1.10)
|
0.847
|
|
ApoA1[c]
|
|
|
|
|
0.91 (0.83, 1.00)
|
0.059
|
|
eGFR[c]
|
|
|
|
|
1.00 (0.89, 1.13)
|
0.988
|
Abbreviations: ApoA1, apolipoprotein A1; ApoB, apolipoprotein B; eGFR, estimated glomerular
filtration rate; IDDM, insulin-dependent diabetes mellitus; NIDDM, non–insulin-dependent
diabetes mellitus; SD, standard deviation; TG, triglycerides.
a Estimated by Cox proportional hazard regression analysis.
b Per 1 SD increase in logarithmically transformed neopterin levels.
c Per 1 SD increase.
Table 5
Model discrimination
|
Model[a]
|
AMI
|
Total mortality
|
|
AUC (95% CI)
|
p-Value[b]
|
AUC (95% CI)
|
p-Value[b]
|
|
Model 2
|
0.650 (0.584, 0.643)
|
|
0.750 (0.724, 0.776)
|
|
|
Model 2 + fibrinogen
|
0.657 (0.629, 0.685)
|
0.155
|
0.755 (0.729, 0.780)
|
0.141
|
|
Model 2 + neopterin
|
0.661 (0.661, 0.689)
|
0.006
|
0.757 (0.732, 0.782)
|
0.011
|
|
Model 2 + fibrinogen and neopterin
|
0.665 (0.637, 0.693)
|
0.007
|
0.760 (0.735, 0.785)
|
0.008
|
Abbreviations: AMI, acute myocardial infarction; ApoA1, apolipoprotein A1; ApoB, apolipoprotein
B; eGFR, estimated glomerular filtration rate.
a Model 2 includes age, gender, diabetes mellitus, smoking status, hypertension, triglycerides,
ApoA1, ApoB and eGFR.
b
p-Value represents comparison of gains in AUC between Model 2 and Model 2 + fibrinogen,
Model 2 and Model 2 + neopterin, and Model 2 and Model 2 + both fibrinogen and neopterin.
Fig. 2 Dose–response relationship between risk of acute myocardial infarction (A and B) or all-cause mortality (C and D) and fibrinogen or neopterin. Estimated by generalized additive Cox regression with
penalized smoothing spline (multivariable model 2). The solid lines represent hazard
ratios, and the shaded areas represent the 95% confidence intervals. The x-axes are trimmed, excluding the lower and upper 2.5 percentile. Density plots are
superimposed along the x-axes, with vertical lines displaying (from left) the 25th, 50th and 75th percentile
of the distribution.
There were no statistically significant interactions across subgroups of fibrinogen
and neopterin dichotomized at median levels. However, there was a statistically significant
interaction between neopterin and hsTnT groups (below/above the median), such that
the association for neopterin with AMI was greater when hsTnT levels were above median
(p = 0.032). This interaction remained statistically significant in Model 1 adjusting
for age and gender, but not after adjusting for all covariates in our multivariable
Model 2 (p-value for interaction is 0.056).
Acute Myocardial Infarction According to Gender and Smoking Status
In both males and females, fibrinogen and neopterin were associated with incident
AMI in univariable and multivariable models ([Supplementary Table S7] [online only]). The association for neopterin was stronger among females than among
males (p = 0.006 for interaction). Both fibrinogen and neopterin were associated with incident
AMI in the simple model, irrespective of smoking habits, remaining statistically significant
in the fully adjusted model for non-smokers ([Supplementary Table S7] [online only]). However, no statistically significant interactions were found between
fibrinogen or neopterin and smoking status with respect to future AMI.
All-Cause Mortality
Median (25th and 75th percentiles) follow-up time with respect to the secondary outcome
measure of total mortality was 10.2 (9.2, 11.4) years, during which a total of 769
patients died. Kaplan–Meier plots for quartile of fibrinogen and neopterin are presented
in [Fig. 1]. Patients in the upper quartile of both fibrinogen and neopterin had a worse outcome,
as compared with patients of the lower quartile (log-rank test, p < 0.001). In the univariable and multivariable Cox regression models,[1]
[2] both fibrinogen and neopterin were associated with all-cause death during follow-up.
The HR (95% CI) per 1 SD (logarithmically transformed) was 1.22 (1.13, 1.31) for fibrinogen
and 1.24 (1.15, 1.34) for neopterin in the fully adjusted Model 2 (p < 0.001 for both; [Supplementary Tables S8] and [S9] [online only]; [Fig. 2]). The AUC for the ROC curve with respect to all-cause mortality was 0.750, based
on multivariable Model 2, not including fibrinogen and neopterin, increasing to 0.760
when adding both these biomarkers to the model (p = 0.008; [Table 5], [Supplementary Figs. 3] and [4] [online only]).
For all-cause mortality, there was a statistically significant interaction between
neopterin and fibrinogen groups (arranged according to below and above the median,
respectively) (p-value for interaction is 0.019). We observed a reduced association of neopterin with
risk among patients with fibrinogen levels above the median (HR: 1.39 [95% CI: 1.31,
1.46], p < 0.001), below the median (HR: 1.61 [95% CI: 1.44, 1.80], p < 0.001). However, this interaction was no longer significant after multivariable
adjustment. Furthermore, no significant interaction was found between fibrinogen and
neopterin across subgroups of hsTnT below or above median (p-value for interaction is 0.126–0.725).
All-Cause Mortality According to Gender and Smoking Status
For gender, higher levels of both fibrinogen and neopterin correlated with increased
risk of total mortality ([Supplementary Table S7] [online only]). Neopterin was more strongly related to death in women (p = 0.021 for interaction).
Fibrinogen and neopterin were associated with death in both smokers and non-smokers
in our univariable model, and this relationship remained statistically significant
after adjusting for possible confounding factors in Model 2 ([Supplementary Table S7] [online only]). In addition, a significant interaction (p = 0.027) was found between neopterin and smoking status, where the prognostic utility
of neopterin was greater among non-smokers.
Discussion
Principal Findings
We found that higher values of fibrinogen and neopterin were independently associated
with increased risk of AMI and total mortality in a population of SCAD patients, supporting
our hypothesis. We also observed that the association with total mortality for neopterin
was higher at hsTnT values above the median, in women and in non-smokers.
Fibrinogen
In the total cohort, fibrinogen was related to both incident AMI and all-cause mortality,
but it did not add additional significant prognostic information, as compared with
traditional risk parameters, based on comparison of AUC in the ROC analysis. The extent
of CAD as judged angiographically increased across the quartiles of fibrinogen, from
Q1 to Q4, which is in accordance with a recent study by Tabacki et al.[38] Fibrinogen is a marker of inflammation, and showed expected positive correlation
with hsCRP. Our findings with respect to future AMI and mortality are essentially
in accordance with a report on patients with angiographically significant CAD by Sinning
et al.[39]
Neopterin
We found that neopterin was a significant risk marker for both incident AMI and total
mortality. In the ROC analysis, this biomarker significantly improved the AUC for
both AMI and total mortality, as compared with traditional risk markers (Model 2).
Neopterin has previously been shown to yield prognostic information related to both
incident AMI and total mortality.[29] In a mixed angiographic population consisting of 1,083 SCAD and 718 ACS patients
with a median follow-up of 8.0 years, neopterin levels were associated with both total
and cardiovascular mortality.[40] The results of our study, which included only SCAD patients, are essentially in
agreement with that study. Neopterin may also serve as a useful risk marker for death
or recurrent acute coronary events among ACS patients, as demonstrated in the TIMI
22 study by Ray et al,[25] in which blood samples were collected 1 week after an acute coronary event.
Interactions
Both fibrinogen and neopterin were independently associated with outcome in patients
with SCAD, but no interactions were observed between the two biomarkers with respect
to both incident AMI and total mortality.
Small increases in troponins may reflect active plaques generating microembolisms
associated with asymptomatic minor myocardial injury in SCAD.[41] Lyngbæk et al have previously reported an association between increasing hsTnT for
both total mortality and AMI, evaluated as separate endpoints.[30] In the present study, we noted that the association of neopterin with incident AMI
was enhanced in patients with hsTnT above the median (limited to Model 1), whereas
this interaction was not observed for total mortality. Thus, the prognostic utility
of neopterin may be improved among high-risk SCAD patients, as defined by their hsTnT
levels. Notably, to the best of our knowledge, a potential effect modification between
neopterin and hsTnT has not been evaluated previously.
Subgroups
Circulating concentrations of both fibrinogen and neopterin were higher in females
than in males, also after adjustment for age. In a previous report on healthy subjects,
the level of neopterin was similar for men and women,[42] whereas in one study with unstable angina pectoris patients the level was higher
in females.[22] We also observed that neopterin in women was associated with the presence of significant
CAD, whereas this association was not noted in males. To our knowledge, similar findings
have previously not been reported. Our main findings related to prognosis were valid
in both genders. According to an interaction analysis, the risk associations for neopterin
were more pronounced in the female subgroup.
Current smokers had higher levels of fibrinogen but lower levels of neopterin as compared
with non-smokers. These findings for fibrinogen are consistent with previous reports,[43] whereas data on the influence of smoking on neopterin are inconsistent.[29]
[40]
[44]
[45] In our study, the association with respect to total mortality for neopterin was
stronger among non-smokers, but not modified by smoking status for fibrinogen. These
observations on effect modification by smoking have not previously been reported.
