Urinary 11-Dehydro-Thromboxane B2 as a Predictor of Acute Myocardial Infarction Outcomes: Results of Leukotrienes and Thromboxane In Myocardial Infarction (LTIMI) Study.

BACKGROUND: Urinary 11-dehydro-thromboxane (TX)B2 has been described as a potential predictive biomarker of major adverse cardiovascular events (MACEs) in high cardiac risk patients. This part of LTIMI (Leukotrienes and Thromboxane In Myocardial Infarction) study aimed to evaluate the relationship between 11-dehydro-TXB2 and MACEs in patients with acute myocardial infarction (AMI). METHODS AND
RESULTS: LTIMI was an observational, prospective study in 180 consecutive patients with AMI type 1 referred for primary percutaneous coronary intervention. On admission and at follow-up visits (1 month, 1 year), 11-dehydro-TXB2 was measured in urinary samples by using high-performance liquid chromatography-tandem mass spectrometry. The primary outcome was occurrence of composite MACEs during 1-year after AMI. Left ventricular ejection fraction was assessed in echocardiography on admission and at 1-year follow-up. Analyses of 11-dehydro-TXB2 (pg/mg creatinine) were performed on log-transformed data and expressed as median with IQR (Q1-Q3). 11-Dehydro-TXB2 level on admission was 7.39 (6.85-8.01) and decreased at 1 month (6.73, 6.27-7.12; P<0.001) and 1-year follow-up (6.37, 5.91-6.94; P<0.001). In univariate analysis, baseline 11-dehydro-TXB2 was higher in patients with MACEs (n=60; 7.73, 7.07-8.60) compared with those without MACEs (n=119; 7.28, 6.68-7.79; P=0.002). In multivariate regression model, 11-dehydro-TXB2 and 3 other variables (diabetes, multivessel disease, and left ventricular ejection fraction) were found to be best 1-year cumulative MACE predictors with odds ratio for 11-dehydro-TXB2 of 1.58 (95% CI 1.095-2.33; P=0.017) and area under the curve (in receiver operating characteristic analysis of 0.8). Baseline 11-dehydro-TXB2 negatively correlated with both left ventricular ejection fraction on admission (R=-0.21; P=0.006) and after 1 year (R=-0.346; P<0.001).
CONCLUSIONS: 11-Dehydro-TXB2 predicts 1-year cumulative MACEs in AMI patients and provides prognostic information on the left ventricular performance.

Introduction

Thromboxane (TX)A2 is a potent platelet agonist and vasoconstrictor1 produced mostly by platelets, which play a pivotal role in thrombogenesis at sites of vulnerable plaque and provide a strong rationale for blocking their function in the setting of acute myocardial infarction (AMI).1, 2

However, platelets produce only 70% of TX in human body, and the remaining 30% is produced by extraplatelet sources such as monocytes and macrophages and may increase in acute inflammation.3 This unrecognized fact may translate into the unacceptable rates of clinical events such as myocardial infarction, cardiovascular death, and need for repeat revascularization, despite proper antiplatelet therapy.

TXA2 is an unstable compound with short plasma half‐life of ≈30 seconds and therefore is difficult to measure.4 TXA2 is rapidly inactivated to TXB2 and further undergoes dehydrogenation to 11‐dehydro‐TXB2.5, 6 The latter is a stable metabolite excreted in urine, and its measurement provides a reliable estimate of the total in vivo production of TXA2 including the extraplatelet sources.7, 8 11‐Dehydro‐TXB2 has been found to be elevated in several atherothrombotic diseases and to correlate with major adverse clinical events.9, 10, 11, 12, 13, 14 However, none of the studies have evaluated the association of 11‐dehydro‐TXB2 with patients’ long‐term outcomes in AMI.

The LTIMI (Leukotrienes and Thromboxane In Myocardial Infarction) study was a prospective, observational study that aimed to assess the relationship between arachidonic acid derivatives and major adverse cardiovascular events (MACEs) in patients with AMI. We report TX biosynthesis evaluated by urinary 11‐dehydro‐TXB2 excretion during AMI and after 1‐year observation period, analyzing its impact on patient's prognosis.

Methods

Study Population

The LTIMI study was conducted between July 2011 and September 2014. The study personnel evaluated 254 consecutive patients with ST‐segment elevation (STEMI) and non–ST‐segment elevation AMI (NSTEMI) who were transferred to interventional cardiology reference center for urgent coronary angiography and percutaneous coronary intervention (PCI). Only patients with MI type I (atherothrombotic, spontaneous) and symptoms lasting no longer than 24 hours were eligible for the study. In the setting of NSTEMI, most of the patients underwent immediate invasive strategy (0–2 hours) or early invasive strategy (<24 hours of diagnosis); none underwent delayed invasive strategy (within 24–72 hours of diagnosis) in the setting of NSTEMI. All patients with STEMI received immediate PCI without previous fibrinolysis.

Type of AMI was based on the Joint European Society of Cardiology/American College of Cardiology Foundation/American Heart Association/World Heart Federation Universal Definition15, 16 by 2 cardiologists who were aware of all patients’ clinical data including coronary angiography (CA) findings. Patients naive to antiplatelets were treated with 300 mg non–enteric‐coated aspirin and 300 to 600 mg of clopidogrel in the ambulance or the emergency department. If the patients were taking aspirin on a long‐term basis, they received a lower dose of 75 mg (not enteric coated); the same was true for clopidogrel (75‐mg dose was used). Exclusion criteria were late presentation of AMI (>24 hours from first symptoms), AMI other than type I, cardiogenic shock, history of coronary artery bypass graft surgery (CABG), severe valvular heart disease, symptoms of acute infection, asthma, chronic obstructive pulmonary disease exacerbation, use of antileukotriene medications, chronic kidney disease, liver cirrhosis, malignancy, patient's refusal to participate in the study (<2% of patients), and noncompliance. Patients who received antiplatelet drugs other than aspirin and clopidogrel before or after PCI or were enrolled in any other interventional trial were excluded from the study. The latter was the most common exclusion criterion. A study flow chart is presented on Figure 1.

Figure 1

Study flow chart. CABG indicates coronary artery bypass graft surgery; NSTEMI, non–ST‐elevation myocardial infarction; PPCI, primary percutaneous coronary intervention; STEMI, ST‐elevation myocardial infarction.

Finally, 180 patients who gave written informed consent were recruited to the study, and 171 completed the 1‐year follow‐up period. The study protocol complied with the Helsinki Declaration and was approved by the Jagiellonian University ethics committee.

Baseline Characteristics

Data on comorbidities, demographic characteristics, and history of presenting complaints were collected by research staff. Standard laboratory tests were performed including serial high‐sensitivity troponin T measurements (hs‐TnT, Roche Diagnostics). Blood and urine samples were collected from each participant before CA and stored immediately after centrifugation at −70°C for further analysis.

TX Measurements

11‐Dehydro‐TXB2 was measured at the presence of identical deuterated standard added to the urine sample before high‐performance liquid chromatography–tandem mass spectrometry measurement, as previously described.17 Results are expressed in picograms per milligram creatinine as natural logarithms. Detailed methods are described in Supplementary Materials.

Transthoracic Echocardiography

Standard transthoracic echocardiography was performed in all patients on admission because of AMI and at 1‐year follow‐up. Apical, parasternal, and subcostal views were used according to current guidelines18 with use of the Vivid 7 ultrasound system (GE Vingmed Ultrasound A/S). Left ventricular ejection fraction (LVEF) was calculated by using Simpson's method. Global left ventricular dysfunction was defined as LVEF <55%. For statistical comparisons, a patient's LVEF was stratified as slightly impaired (45–54%), moderately impaired (30–44%), or severely impaired (<30%). Two experienced cardiologists‐echocardiographers assessed images independently, unaware of 11‐dehydro‐TXB2 results.

CA and PCI

CA was performed by using the standard Seldinger technique, with the patient under local anesthesia, via puncture of the femoral or radial artery, with the use of typical diagnostic and guiding catheters. Arterial access was selected at the discretion of the operator. All patients were given the same type of low‐osmolar and nonionic iodinated contrast media (Visipaque 320 GE Healthcare Inc). Immediately after CA, the results were analyzed and patients were qualified for immediate PCI with angioplasty and stent implantation of the culprit vessel or referred for urgent CABG. The type of stent and its diameter and length were selected on the basis of angiographic measurements assisted with digital quantitative angiography (Axiom ZEE angiograph, computer software VC21B; Siemens). Control angiography was performed after stent implantation to document and assess obtained result.

Final Outcomes Measurements and Follow‐up

Study outcome measurements

The primary outcome was occurrence of a composite MACE that occurred during a 1‐year follow‐up period after AMI. This composite MACE consisted of 8 events: recurrent myocardial infarction, stroke/transient ischemic attack, cardiogenic shock, pulmonary edema, nonfatal cardiac arrest, need for rescue PCI or CABG, and cardiovascular death. If 2 MACEs occurred in 1 patient (eg, recurrent MI and death in follow‐up), it was counted as 1 cumulative MACE. Definitions are available in Supplemental Materials.

Secondary outcomes included (1) composite MACE at the following time points: discharge from hospital and at 1‐month follow‐up; and (2) LVEF measured on admission and at 1‐year follow‐up.

Two cardiologists‐echocardiographers adjudicated all events independently and were blinded to the 11‐dehydro‐TXB2 results while performing the clinical follow‐up.

Follow‐up

Patients were observed for a 1‐year period after enrollment in the study. Two visits were scheduled in the outpatient clinic, at 1 month and 1 year after AMI. The data on the occurrence of MACEs were collected by the research team and verified based on either hospital discharge notes or family member interrogation. Urine samples were collected at both visits for 11‐dehydro‐TXB2 analyses.

Statistical Analysis

Categorical variables were presented as counts (percentages), whereas continuous variables were reported as mean±SD or median with IQR (Q1–Q3) depending on their distribution. Urinary 11‐dehydro‐TXB2 data were naturally log‐transformed to approximate normal distribution for further analysis and expressed throughout the text as median with IQR (Q1–Q3). Assumption of normality was verified by using the Kolmogorov–Smirnov test. Categorical variables were compared between MACE and non‐MACE groups by using χ2 test or Fisher's exact test, and continuous variables were compared by using Student t test or Mann–Whitney U test as appropriate. Correlations between variables were analyzed by using Pearson's or Spearman's rank correlation, as appropriate for the data distribution. For analysis of 11‐dehydro‐TXB2 in urine at different time points during follow‐up, 1‐way ANOVA for repeated measures and the post‐hoc Tukey test were used.

For comparisons of LVEF with 11‐dehydro‐TXB2 levels, the Kruskal–Wallis test was used with post‐hoc Mann–Whitney comparison adjusted with the Bonferroni correction.

Associations between 11‐dehydro‐TXB2 and MACEs after PCI were estimated as an odds ratio (OR) and corresponding 95% CI) by using univariate logistic regression analysis.

The associations of 11‐dehydro‐TXB2, cardiac biomarkers, inflammatory parameters, and composite MACE were tested after division into quartiles according to the 11‐dehydro‐TXB2 distribution. The diagnostic performance of 11‐dehydro‐TXB2 for the prediction of cumulative MACEs was evaluated by using receiver operating characteristic curve analysis, and the optimal threshold was based on the point of combined best sensitivity/specificity results.

Finally, the association between 11‐dehydro‐TXB2 and cumulative MACE incidence was evaluated with use of the multiple logistic regression. Explanatory variables were chosen in stepwise approach on the basis of the Akaike information criterion based on the prespecified baseline clinical and laboratory characteristics. Variables included age, sex, body mass index, smoking (pack‐years), past history of MI, diabetes mellitus, hypertension, time from symptoms to PCI, multivessel disease, LVEF during hospitalization, maximal TnT level, C‐reactive protein (CRP), glomerular filtration rate, and 11‐dehydro‐TXB2.

All statistical analyses were performed by using R software, version 3.1.2 (R Development Core Team [2009]. R: A language and environment for statistical computing. R Foundation for Statistical Computing). P‐values <0.05 were considered statistically significant.

Results

Patients

One hundred eighty patients with type I AMI (50.6% with STEMI and 49.4% with NSTEMI) who underwent CA on hospital admission were found to be eligible for the study (Figure 1). Time from symptom onset to the PCI was 9.23±6.71 hours. Mean age of the study cohort was 66.48±11.87 years, and men outnumbered women (69.4% versus 30.6%). Detailed demographic parameters and cardiovascular risk factors are shown in Table 1. On admission, all patients had elevated hs‐TnT, 52.8% presented with increased leukocyte count >104/μL, and 46.1% had high‐sensitivity CRP elevated above the normal range. In 164 patients (91.1%), PCI was performed, and 16 patients (8.9%) required surgical revascularization (CABG). The majority of the patients (n=117; 65%) had at least 2‐vessel disease, and mean LVEF measured on admission was 44.7±13.8%.

Table 1

Baseline Characteristic of the Study Group by MACE and Non‐MACE Subgroup

Characteristic	Overall (n=180)	MACE (n=60)	Non‐MACE (n=120)	P Value
Demographic characteristics
Age, y (%)
<55	36 (20)	8 (13.3)	28 (23.3)	0.126
55–75	98 (54.4)	32 (53.3)	66 (55)
>75	46 (25.6)	20 (33.3)	26 (21.7)
Female sex	55 (30.6)	22 (36.7)	33 (27.5)	0.277
Body mass index, kg/m2 (%)				0.877
Normal weight (18.5–24.99)	60 (33.3)	19 (31.7)	41 (34.2)
Overweight (25–29.99)	76 (42.2)	25 (41.7)	51 (42.5)
Obesity (>30)	44 (24.4)	16 (26.7)	28 (23.3)
Medical history
Hypertension, n (%)	128 (71.1)	46 (76.7)	82 (68.3)	0.323
Diabetes mellitus, n (%)	61 (33.9)	33 (55)	28 (23.3)	<0.001
Dyslipidemia, n (%)	147 (86.5)	50 (89.3)	97 (85.1)	0.608
Myocardial infarction, n (%)	48 (26.7)	21 (35)	27 (22.5)	0.108
TIA/stroke, n (%)	13 (7.2)	5 (8.3)	8 (6.7)	0.762
PAD/claudication, n (%)	37 (20.6)	18 (30)	19 (15.8)	0.043
Family history of CAD, n (%)	78 (43.3)	29 (48.3)	49 (40.8)	0.425
Current smoker, n (%)	68 (37.8)	16 (26.7)	52 (43.3)	0.044
Clinical evaluation
Diagnosis				0.792
NSTEMI, n (%)	89 (49.4)	31 (51.7)	58 (48.3)
STEMI, n (%)	91 (50.6)	29 (48.3)	62 (51.7)
CAD, n (%)				<0.001
1‐Vessel disease	63 (35)	9 (15)	54 (45)
2‐Vessel disease	59 (32.8)	19 (31.7)	40 (33.3)
3‐Vessel disease	58 (32.2)	32 (53.3)	26 (21.7)
LVEF during AMI, n (%)				0.002
Normal (≥55)	55 (30.6)	11 (18.3)	44 (36.7)
Mildly impaired (45–54)	45 (25)	14 (23.3)	31 (25.8)
Moderately impaired (30–44)	57 (31.7)	20 (33.3)	37 (30.8)
Severely impaired (<30)	23 (12.8)	15 (25)	8 (6.7)

AMI indicates acute myocardial infarction; CAD, coronary artery disease; LVEF, left ventricular ejection fraction; MACE, major adverse cardiovascular event; NSTEMI, non–ST‐elevation myocardial infarction; PAD, peripheral artery disease; STEMI, ST‐elevation myocardial infarction; TIA, transient ischemic attack.

Aspirin and clopidogrel were administered to all patients at least 30 minutes before the urine sample collection. In the studied group, 54 (30%) patients received aspirin on a long‐term basis before admission, and 11 (6.1%) additional patients were taking clopidogrel. By discharge from the hospital, all patients were taking aspirin, and 94.4% were taking clopidogrel (except 10 patients scheduled for CABG).

Follow‐up MACEs

At 1‐year follow‐up, the composite MACE appeared in 60 patients (33.3%), and of this group, 14 (7.8%) patients died. All MACEs with their time occurrence are listed in Table S1. Patients with MACEs compared with patients without MACEs more frequently had multivessel disease, diabetes, and peripheral artery disease, whereas there were fewer current smokers. Additionally, in the MACE group, patients presented with higher admission levels of glucose (7.85 [6.27–10] versus 6.8 [5.8–8.3] mg/L; P=0.012) and lower triglyceride levels (1.05 [0.81–1.56] versus 1.3 [0.94–1.67] mmol/L; P=0.044). Comparisons of clinical and laboratory data are shown in Tables 1 and 2.

Table 2

Baseline Laboratory Parameters Characteristic of the Study Group on Admission

Parameter	Normal Values	Overall (n=180)	MACE (n=60)	Non‐MACE (n=120)	P‐Value
hs‐TnT on admission, μg/L	<0.014	0.14 (0.06–0.33)	0.19 (0.05–0.46)	0.14 (0.06–0.28)	0.14
hs‐TnT max, μg/L		1.25 (0.29–4.65)	0.92 (0.41–4.96)	1.28 (0.24–4.51)	0.88
CK‐MB on admission, U/L	0–24	21 (15–42.3)	23 (16.8–42.3)	20.5 (15–40.8)	0.41
CK‐MB max, U/L		66.5 (26–187)	52 (29.8–154.8)	71.5 (23–206.5)	0.76
CK on admission, U/L	0–190	182 (104–341.3)	188.5 (110.75–369.3)	179 (103–335)	0.4
CK max, U/L		623 (215–1756)	643.5 (282.3–1445.8)	605 (201.5–1816.5)	0.7
hs‐CRP, mg/L	≤3.0	2.8 (1.1–6.4)	3.95 (1.21–7)	2.5 (1.05–5.7)	0.16
WBC, ×103/μL	3.8–10	10.1 (8.1–12.2)	10.3 (8.4–13.7)	10.1 (7.9–11.9)	0.26
RBC, ×106/μL	4.2–6.0	4.7 (4.4–4.97)	4.59 (4.1–4.9)	4.8 (4.5–4.99)	0.07
Hb, g/dL	14–18	14.2 (13.2–15.1)	14 (12.8–15.2)	14.4 (13.4–15.1)	0.28
PLT, ×103/μL	140–440	221.5 (181.8–263.3)	218.5 (164–257.3)	222.5 (188.5–264.3)	0.43
GFR, mL/min per 1.73 m2	>60	74.8 (56.6–87.8)	67.3 (55.42–84.9)	78 (57.1–88.2)	0.08
Creatinine, μmol/L	62–106	84 (76.8–103)	86.5 (78–104.8)	83.5 (75–101)	0.33
Glucose, mmol/L	3.4–5.6	7 (6–8.6)	7.9 (6.27–10)	6.8 (5.8–8.3)	0.012
TC, mmol/L	3.1–5.0	4.7 (4.06–5.52)	4.42 (3.64–5.22)	4.85 (4.18–5.6)	0.15
LDL, mmol/L	<3.0	3.01 (2.4–3.9)	2.67 (2.0–3.6)	3.09 (2.5–3.9)	0.1
HDL, mmol/L	>1.0	1.2 (1.01–1.39)	1.2 (1.04–1.45)	1.2 (1.01–1.37)	0.3
TG, mmol/L	<1.7	1.2 (0.9–1.6)	1.1 (0.8–1.6)	1.3 (0.9–1.7)	0.044

CK indicates creatine kinase; CK‐MB, creatine kinase isoform; GFR, glomerular filtration rate; Hb, hemoglobin; HDL, high‐density lipoprotein; hs‐TnT, high‐sensitivity troponin T; hs‐CRP, high‐sensitive C‐reactive protein; LDL, low‐density lipoprotein; MACE indicates major adverse cardiovascular event; PLT, platelets; RBC, red blood cell; TC, total cholesterol; TG, triglycerides; WBC, white blood cell.

The baseline LVEF was 44.71±13.79% and increased at 1‐year follow‐up to 55.68±10.62% (P<0.001), being lower at both time points in patients with MACEs.

Urinary TX Metabolite Excretion

All analyses were performed on log‐transformed data. The highest level of 11‐dehydro‐TXB2 was on admission, 7.39 (6.85–8.01), and decreased in the follow‐up after 1 month to 6.73 (6.27–7.12) and after 1 year to 6.37 (5.91–6.94) pg/mg creatinine. The decline was statistically significant between each follow‐up level (1 month and 1 year) compared with the admission sample (P<0.001; Figure 2).

Figure 2

Urinary 11‐dehydro‐TXB 2 excretion on admission (baseline) and at 1‐month and 1‐year follow‐up. Average (points) and SD (whiskers) of 11‐dehydro‐TXB 2 on admission, at 1‐month and 1‐year follow‐up expressed in log‐transformed values. *P<0.001.

Patients taking aspirin and/or clopidogrel on a long‐term basis had lower admission 11‐dehydro‐TXB2 (7.12 [6.68–7.59] versus 7.57 [6.94–8.15] pg/mg creatinine; P=0.005).

11‐Dehydro‐TXB2 levels were higher in STEMI patients than in those with NSTEMI (7.59 [6.97–8.14] versus 7.24 [6.55–7.7] pg/mg creatinine, respectively; P=0.013). There were no differences in 11‐dehydro‐TXB2 levels adjusted for age and sex.

TX Correlations

Baseline 11‐dehydro‐TXB2 correlated weakly with maximal peri‐AMI levels of cardiac damage biomarkers (creatine kinase maximum [R=0.17; P=0.03], hs‐TnT [R=0.19; P=0.01]). 11‐Dehydro‐TXB2 also correlated with inflammatory parameters (white blood cell count >10×103/μL [R=0.23; P=0.002], high‐sensitivity CRP [R=0.31; P<0.001]). Both white blood cell count (9.46; 7.92–11.09 in Q1 versus 11.52; 9.34–14.51 ×103/μL in Q4; P=0.007) and high‐sensitivity CRP (1.22; 0.55–3.54 in Q1 versus 6.29; 1.29–23.4 mg/L in Q4; P<0.001; Figure S1) were significantly elevated in the highest quartile of 11‐dehydro‐TXB2 level.

Primary Outcomes Analysis

Baseline mean 11‐dehydro‐TXB2 measured during AMI was higher in patients who experienced MACEs at 1‐year follow‐up (7.73 [7.07–8.60]) compared with those who did not (7.28 [6.68–7.79]; P=0.002). Between the quartiles of 11‐dehydro‐TXB2, the incidence of cumulative MACEs increased by 266% in Q4 compared with the first quartile (OR 3.66 [95% CI 1.47–9.13]) (Figure 3; receiver operating characteristic curve—Figure S2).

Figure 3

Cumulative MACEs risk in 1 year following AMI according to 11‐dehydro‐thromboxane (TX)B 2 quartiles on admission. The ranges of quartiles intervals (Q1–Q4) expressed as log‐transformed values of 11‐dehydro‐TXB 2 in pg/mg creatinine. Ranges of log 11‐dehydro‐TXB 2 quartiles are as follows: Q1: <6.85; Q2: 6.85 to 7.39; Q3: 7.40 to 8.01; Q4: >8.01 (pg/mg creatinine). Top bars are the ORs for cumulative MACEs at 1 year from AMI, whiskers indicate CIs. AMI indicates acute myocardial infarction; MACE, major adverse cardiovascular event; OR, odds ratio.

In the multivariate analysis based on logistic regression, 11‐dehydro‐TXB2 and 3 other variables (diabetes, LVEF during AMI, and multivessel disease) were found to be the best predictors of 1‐year cumulative MACEs (Table 3). In this model, the increase in natural logarithm of 11‐dehydro‐TXB2 by 1 was associated with increased risk of developing MACEs within 1 year after AMI by 58.1%. Prognostic capability of 11‐dehydro‐TXB2 and other variables for predicting MACEs is presented on Figure 4 and has an area under the curve related to receiver operating characteristic curve of 0.806. Based on this multivariate model, we developed a mathematical formula for risk prediction, which is presented in Data S1. When long‐term aspirin treatment was included in the regression analysis, it did not influence the results.

Table 3

One‐Year Follow‐up Cumulative MACE Occurrence in Multivariate Analysis With 11‐Dehydro‐TXB2

Variable	OR	95% CI for OR		P Value
Diabetes mellitus	2.476	1.128–5.451		0.024
LVEF during hospitalization, %	0.978	0.95–1.006		0.133
CAD
1‐vessel disease	Reference level
2‐vessel disease	4.671	1.646–14.951		0.006
3‐vessel disease	9.141	3.308–29.023		<0.001
Log 11‐dehydro‐TXB2 on admission, pg/mg creatinine	1.581	1.095–2.329		0.017

CAD indicates coronary artery disease; LVEF, left ventricular ejection fraction; MACE, major adverse cardiovascular event; OR, odds ratio; TX, thromboxane. Variables included in the multiple regression model are: age, sex, body mass index, smoking (pack‐years), past history of myocardial infarction, diabetes mellitus, hypertension, time from symptoms to percutaneous coronary intervention, multivessel disease, LVEF during hospitalization, maximal troponin level, C‐reactive protein, glomerular filtration rate, and 11‐dehydro‐TXB2.

Figure 4

Receiver operating characteristic curve for predictive value of the multivariate regression model on the cumulative MACE occurrence at 1‐year follow‐up after AMI. The multivariate regression model includes 11‐dehydro‐thromboxane (TX)B 2 on admission, diabetes mellitus, MVD, and baseline LVEF. AMI indicates acute myocardial infarction; AUC, area under the curve; DM, diabetes mellitus; LVEF, left ventricular ejection fraction; MVD, multivessel disease.

Secondary Outcomes Analysis

LVEF and 11‐dehydro‐TXB2 level on admission

11‐Dehydro‐TXB2 on admission inversely correlated with both measurements of LVEF at initial hospitalization (R=−0.21; P=0.006) and at 1‐year observation (R=−0.346; P<0.001). Patients with most severely impaired LVEF (<30%) during AMI were found to have the highest levels of 11‐dehydro‐TXB2 compared with all others. 11‐dehydro‐TXB2 level on admission was also a predictor of LVEF after 1 year; it was higher in patients with impaired LVEF compared with patients with normal LVEF (Table 4).

Table 4

Association of Urinary 11‐Dehydro‐TXB2 on Admission With Baseline LVEF During Hospitalization and at 1‐Year Follow‐up

Systolic LV Function	n	Log 11‐Dehydro‐TXB2 on Admission, pg/mg Creatinine				P‐Value	Post‐hoca
		Median	Range	Q1	Q3
LVEF during hospitalization, %
Normal (≥55)	55	7.32	5.54–9.45	6.62	7.82	0.002	a
Mildly impaired (45–54)	45	7.1	4.35–9.66	6.71	7.67		a
Moderately impaired (30–44)	57	7.49	5.67–10.74	6.92	8.13		ab
Severely impaired (<30)	23	7.73	5.2–10.26	7.49	8.95		b
LVEF after 12 mo, %
Normal (≥55)	81	7.15	4.35–9.3	6.42	7.59	<0.001	a
Mildly impaired (45–54)	35	7.49	6.25–9.44	7.1	7.88		b
Moderately impaired (30–44)	21	7.73	5.67–10.13	7.27	8.68		b
Severely impaired (<30)	3	8.76	7.59–9.31	8.17	9.03		b

EF indicates ejection fraction; LV, left ventricular; Q1, quartile; Q3, quartile 3; TX, thromboxane.

Results of post‐hoc analysis. Groups identified with the same letter did not differ significantly.

The 11‐dehydro‐TXB2 levels measured at 1 month and 1 year after AMI did not correlate with LVEF.

11‐Dehydro‐TXB2 and MACEs at hospital discharge and at 1‐month follow‐up

Log‐transformed 11‐dehydro‐TXB2 on admission was predictor of cumulative MACEs for both time points: for discharge MACEs (OR 1.499, 95% CI 1.007–2.253) and for 1‐month MACEs (OR 1.57, 95% CI 1.11–2.257). Therefore, the increase in log‐transformed baseline 11‐dehydro‐TXB2 by 1 increased the risk of MACEs on discharge from hospital by 49.9% and by 57% at 1‐month follow‐up.

Discussion

In the present study we evaluated the association of systemic 11‐dehydro‐TXB2 production with clinical outcomes of patients with AMI. There are 2 main findings of the study. First, a higher urinary level of 11‐dehydro‐TXB2 on admission because of AMI significantly increases the probability of cumulative MACE occurrence at 1‐year follow‐up. Second, the same biomarker predicts LVEF performance post AMI.

The first observation is in line with the previous studies by Eikelboom et al in which 11‐dehydro‐TXB2 was found to be a good predictor of MACEs in a high cardiovascular risk population treated with aspirin.9, 10 The risk of MACE occurrence in these studies, when 2 extreme quartiles of 11‐dehydro‐TXB2 were compared, was between 1.66 and 1.8 times higher in the upper quartile. In the present study, univariate analysis revealed 3.66 times (95% CI 1.47–9.13) higher MACE risk in patients within the highest 11‐dehydro‐TXB2 quartile, compared with those in the lowest quartile. Importantly, when these data were reevaluated by using a multivariate analysis, the confounding factors, such as demographic features, treatment options, or main laboratory findings, did not influence performance of urinary 11‐dehydro‐TXB2 as a predictor of MACEs in AMI patients. Actually, in the multivariate model only 3 other variables contributed to 11‐dehydro‐TXB2 predictive properties: LVEF on admission, presence of diabetes mellitus, and multivessel disease. This model proved to have a good prediction capability with an area under the curve of 0.806 and OR for 11‐dehydro‐TXB2 of 1.581 (95% CI 1.095–2.329), results similar to those presented by Eikelboom et al.9, 10

There are several causes responsible for increased TX production in AMI. First, during the acute phase of AMI, more platelets are released and overactivated.1, 19 Newly formed platelets beside their constitutive cyclooxygenase (COX)‐1 expression, retain COX‐2 activity contributing to additional TX synthesis.1, 2, 11, 20, 21 Further, intense thrombotic and concomitant acute inflammatory processes present in AMI induce TX synthesis from sources other than platelets, mostly inflammatory (monocytes and macrophages) and endothelial cells.22 These cells produce TX mainly via the COX‐2 pathway and stimulate platelets despite their inhibition by aspirin.23 Aspirin in small doses blocks irreversibly COX‐1–dependent TX production; however, inhibition of COX‐2 pathway requires much higher doses of aspirin.2

These mechanisms may contribute to elevated TX synthesis and may overcome the aspirin suppression of COX‐1 pathway during AMI.24 It also explains why in patients taking aspirin on a long‐term basis, 11‐dehydro‐TXB2 production is still maintained.

Urinary excretion of 11‐dehydro‐TXB2 is assumed to reflect systemic TX biosynthesis most accurately25 and, in contrast to serum TXB2, can also detect the in vivo “extraplatelet” TX production.

In a recent study by Frelinger et al, the authors proved in a prospective study that patients presenting for diagnostic catheterization with elevated levels of serum TXB2 had higher incidence of MACEs despite aspirin treatment in a 2‐year observation period.26 It confirms the findings of the present study and highlights the importance of residual TX production by ASA suppressed platelets.

Extraplatelet pathways are described as responsible for ≈30% of 11‐dehydro‐TXB2 production27 and significantly increase in inflammatory and prothrombotic processes—which are present during AMI. Our findings support this hypothesis, as increased production of 11‐dehydro‐TXB2 during AMI correlated with inflammatory parameters: high‐sensitivity CRP and increased white blood cells. Further, 11‐dehydro‐TXB2 substantially decreased at both 1‐month and 1‐year follow‐up compared with the baseline measurements in the acute phase, which cannot be solely explained by the long‐term aspirin treatment after AMI. This is in parallel with the recent study by Cangemi et al, where the authors described TXB2 elevation in patients with pneumonia and subsequent AMI.28

Results of the present study demonstrated that patients with higher levels of 11‐dehydro‐TXB2 have decreased LVEF during AMI and increased levels of baseline 11‐dehydro‐TXB2 predicted decreased LVEF at 1‐year follow‐up. This relationship between elevated TX biosynthesis and reversible myocardial dysfunction may suggest impairment of microcirculation and may play a pivotal role in the pathogenesis of MI. Moreover, 11‐dehydro‐TXB2 on admission correlated with maximal values of biomarkers of myocardial damage, such as creatine kinase and hs‐TnT, which confirmed the results of Valles et al—that TXA2 was found to correlate with the myonecrosis in nonresponding to aspirin‐treated patients.29 Also, in a study by Nicolli et al, TXB2 production correlated with the severity and extent of atherosclerotic changes in coronary vessels despite aspirin treatment.30

In our study, we also found that 11‐dehydro‐TXB2 levels were higher in STEMI patients compared with NSTEMI patients, which can be explained by more intense inflammatory response and a crucial role of myonecrosis in pathogenesis of STEMI compared with NSTEMI.31

Complementary to our findings, it has been described by Nicolli et al that TXA2 measurement could be an independent predictor of no‐reflow syndrome following coronary intervention, which probably contributes to the short‐term complications after AMI and corresponds with our study findings.32 This could explain why patients with higher levels of 11‐dehydro‐TXB2 had a higher incidence of cumulative MACEs in a short‐term follow‐up after AMI (at hospital discharge and 1nmonth). However, these findings should be treated with caution, as the number of cumulative MACEs was relatively small at this time period.

To our best knowledge, no previous studies have examined 11‐dehydro‐TXB2 in AMI and its associations with future MACEs and left ventricular performance. The strengths of the present study include very low dropout of patients (3%) during follow‐up and strict inclusion criteria that made the studied group homogeneous. We evaluated only patients assumed to have AMI of type I who underwent PCI, whereas symptoms lasted not longer than 24 hours. In this cohort of 180 patients, a relatively large proportion of patients had cumulative MACEs at 1‐year follow‐up (60 patients), which enabled us to draw conclusions on the 11‐dehydro‐TXB2 predictive value as a biomarker. We used the most sensitive quantitative assay of combined high‐performance liquid chromatography–tandem mass spectrometry. This is of great importance, as several previous studies described problems with reliability with the monoclonal antibody ELISA, which detects not only 11‐dehydro‐TXB2 but also an additional metabolite like 11‐dehydro‐2,3‐dinor‐TXB2.4

Limitations of the Study

This study is a single‐center study with a limited number of patients. The study group was sufficient to evaluate the cumulative MACE outcomes but not large enough to evaluate each different MACE complication separately; thus, they had to be evaluated cumulatively. Being aware of this necessity when developing the study protocol, we designed the LTIMI study to evaluate cumulative MACEs as a primary outcome. Nevertheless, a future multicenter study is warranted to verify the outcomes in a larger cohort of patients with AMI. Another limitation of the study is that some patients were taking long‐term aspirin treatment before study enrollment. This bias imposed by long‐term treatment with aspirin was, however, balanced by the fact that all patients received aspirin immediately when suspected of having AMI and at least 30 minutes before sample collection for further analysis. After oral intake of uncoated aspirin, it reaches its concentration peak in blood at ≈30 minutes. This approach represents more of a “real‐life scenario” and, importantly, long‐term aspirin treatment did not influence the statistical models of 11‐dehydro‐TXB2 performance as a biomarker. Finally, several other factors not evaluated in this study could influence 11‐dehydro‐TXB2 measurements such as a patient's genetic variability of TX production or TX receptor synthesis and turnover. This study also did not aim to assess the interindividual resistance to aspirin or other antiplatelet drugs, which may also influence the TX production and MACE incidence.

In conclusion, 11‐dehydro‐TXB2 measured in AMI patients is an independent predictor of major cardiovascular outcomes that occur during the first year after the incidence and may provide prognostic information on left ventricular performance. These findings shed additional light on the pathological role of TXA2 in AMI complications.

Sources of Funding

This work has been supported by a grant from the National Science Center “OPUS” (grant 2011/03/B/NZ5/01450). The publication cost is supported by the Leading National Research Center (KNOW), Faculty of Medicine, Jagiellonian University Medical College, Krakow, Poland.

Disclosures

None.

Table S1. Incidence of MACE and Cumulative MACE During Hospitalization, at 1‐Month and 1‐Year Follow‐up

Figure S1. Association of hs‐CRP with quartile distribution of log‐transformed 11‐dehydro‐TXB2.

Figure S2. Receiver operating characteristic curve for 11‐dehydro‐TXB2 on admission as a biomarker of cumulative MACE occurrence at 1 year post‐AMI in univariate analysis

Data S1. Supplementary Data

Click here for additional data file.