Temporal biomarker concentration patterns during the early course of acute coronary syndrome

Study population

This is an observational post-hoc study of patients included in the PLATelet inhibition and patient Outcomes (PLATO) trial (ClinicalTrials.gov identifier: NCT00391872). PLATO was an international, multicenter, double‐blind, randomized control trial comparing ticagrelor and clopidogrel in addition to aspirin for the prevention of cardiovascular events in 18,624 patients with ACS [10]. Patients were included in the PLATO trial if they were hospitalized for STEMI or NSTE-ACS with symptom onset during the previous 24 h and lasting ≥10 min while at rest. For NSTE-ACS, at least two of three criteria were required: ST-segment depression or transient elevation ≥1 mm in two or more contiguous leads, a positive biomarker result indicating myocardial necrosis, or one additional risk factor (age >60 years, prior myocardial infarction or coronary artery bypass grafting, carotid artery disease, previous ischemic stroke, transient ischemic attack, carotid stenosis or cerebral revascularization, diabetes mellitus, peripheral artery disease, or chronic kidney disease). For STEMI, study inclusion also required a planned primary percutaneous coronary intervention. Among exclusion criteria, the most important included fibrinolytic therapy within 24 h, need for oral anticoagulation therapy, need for dialysis, and clinically important anemia or thrombocytopenia [10].

The PLATO trial and the current investigation were conducted in adherence to the Declaration of Helsinki and had been approved by Ethical Review Boards. All patients provided written informed consent to participate and received routine medical care with the exception of randomized treatment (ticagrelor or clopidogrel).

Biomarker analysis

Ethylenediaminetetraacetic acid (EDTA) plasma samples were collected at randomization which always took place before invasive procedures. After acquisition, samples were stored in aliquots at −70 °C until analysis which was performed centrally at the Uppsala Clinical Research Center (UCR) Laboratory (Uppsala University, Uppsala, Sweden). Concentrations of hs‐cTnT, NT‐proBNP and GDF‐15 were measured on Cobas Analytics Immunoanalyzers (Roche Diagnostics, Basel, Switzerland). Hs-cTnI and CRP concentrations were measured using the Architect platform (Abbott Diagnostics, Abbott Park, IL, USA).

Statistical analysis

Categorical variables are reported as frequencies and percentages, and continuous variables as medians with interquartile ranges.

Linear regression models were fitted to study the association of each biomarker with the time interval from symptom onset to blood sampling. Symptom onset had been defined in PLATO as the date and time of the onset of the most recent cardiac ischemic symptoms. Biomarker concentrations were log-transformed before inclusion as outcome in the models. To capture non-linear associations, time since symptom onset was included as a restricted cubic spline with five knots, placed at the 5th, 27.5th, 50th, 72.5th, and 95th percentiles of the sample distribution of time. The models also included sex, type of ACS (STEMI or NSTE-ACS), and age as a continuous linear term, as well as the interaction of each of these variables with time. To save degrees of freedom, only the linear part of the spline representation of time was included in the interactions. Accordingly, the linear regression models can heuristically be represented as: log(biomarker) ∼ rcs(time) + sex + age + acstype + time × sex + time × age + time × acstype. The associations between biomarker concentrations and time elapsed from symptom onset are presented graphically by plotting predictions from these models for combinations of sex, type of ACS and three arbitrarily chosen ages (50, 60 and 70 years). The overall test of any interaction with time was assessed by testing the simultaneous null hypothesis of all three interaction terms in the respective models being zero. The interactions of time with the type of ACS, age or sex, respectively, on biomarker concentrations were also tested. As a sensitivity analysis, models were fitted by adding body mass index, smoking status, hypertension, dyslipidemia, diabetes, chronic renal disease, previous myocardial infarction, previous ischemic stroke, previous heart failure, and peripheral vascular disease as covariates.

The software package R 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria) was used for the statistical analyses.

Cardiac troponin T and I

The concentrations of hs-cTnT and hs-cTnI exhibited strong associations with time from symptom onset. Highest concentrations in NSTE-ACS patients were noted if sampling was performed after 12–14 h. In STEMI in contrast, hs-cTn concentrations did not differ considerably when sampling was performed after 15 h. Accordingly, hs-cTn concentrations seem to reach a plateau phase at this time point, likely explained by ongoing cTn release from the infarcted myocardium due to early remodeling [11].

Hs-cTnI concentrations were higher compared to hs-cTnT regardless of the type of ACS. This finding corresponds with data from smaller cohorts demonstrating earlier increases and higher peak concentrations for hs-cTnI compared to hs-cTnT [1], [2], [3]. This difference appears to be caused by a faster proteolytic degradation of cTnI in necrotic myocardial tissue which results in a quicker release of measurable cTnI fragments into the circulation [12]. cTnT in contrast, remains mostly bound to cardiac myofibrils where it to a greater extent, is degraded by phagocytes [12]. Consequently, only a smaller proportion of cTnT reaches the circulation and at a later stage compared to cTnI.

Hs-cTn concentrations did not differ by age. However, men were found to have higher hs-cTn concentrations than women, potentially owing to a greater proportion of women with unstable angina included in PLATO [13]. The sex-discrepancy in hs-cTn concentrations increased over time in both STEMI and NSTE-ACS, as illustrated by the diverging spline curves. The cause of this intriguing finding is not clear. No similar patterns emerged for the other investigated biomarkers. As such, the observed sex-difference for hs-cTn appears not to reflect variations in hemodynamic stress or inflammation. Differences in cardiovascular risk factors or comorbidities do not appear to matter either, as suggested by the results of our sensitivity analysis. Greater cardiac muscle mass in men [14] with potentially more releasable cTn may be one contributing factor for the findings reported here. However, final infarct size, as assessed by magnetic resonance imaging, has been shown to be similar in men and women [15, 16]. Further research on sex-differences in hs-cTn release during ACS and the underlying explanations is clearly needed.

N-terminal pro-B-type natriuretic peptide

The concentrations of NT-proBNP were higher along with longer delay time from symptom onset. Interestingly, NSTE-ACS patients undergoing early blood sampling had higher NT-proBNP concentrations compared to those with STEMI. This likely reflects higher prevalence of pre-existing conditions associated with elevated NT-proBNP in these patients, e.g. congestive heart failure and hypertension [17]. The highest NT-proBNP concentrations were noted in patients undergoing sampling at 12–14 h from symptom onset. This indicates that a peak is reached at this time point. Previous data on early NT-proBNP changes during NSTE-ACS are somewhat ambiguous since timing referred to study inclusion which occurred at least one day after the ischemic event [4, 5]. Our study thus, provides clarifying evidence in this regard.

In STEMI, NT-proBNP concentrations continued to increase across patients with longer delay time to blood sampling, and concentrations reached higher levels compared to NSTE-ACS. This reflects the association of this biomarker with greater hemodynamic stress and infarct size in these patients [6]. Our data are consistent with other studies that have studied NT-proBNP kinetics using serial samples and demonstrating peak concentrations at 24–48 h in STEMI [5, 6]. A second peak may occur on day 2–5 indicating the event of adverse remodeling [5, 6].

NT-proBNP concentrations were associated with higher age in both STEMI and NSTE-ACS. Women had higher concentrations of NT-proBNP. Regardless of age and sex however, NT-proBNP concentrations followed similar temporal patterns as indicated by paralleling spline curves. The differences observed here appear thus, not to reflect intrinsic age- or sex-related variations in the pathophysiology of ACS but rather the well-known associations between NT-proBNP, age and female sex [18, 19]. Underlying explanations include age-related changes affecting the heart, renal function, and natriuretic peptide metabolism, greater prevalence of diastolic dysfunction in women and the effect of sex hormones [18, 19]. Taken together, our data emphasize that neither age nor sex need to be considered in ACS for management decisions based on the magnitude of early NT-proBNP concentration changes.

C-reactive protein

The concentrations of CRP increased with longer time from symptom onset, regardless of age and sex. These data correspond with previous findings reporting peak CRP concentrations at 40–54 h in NSTE-ACS and 48–72 h in STEMI [7], [8], [9], i.e. at timepoints beyond our observation period.

Initial CRP concentrations were higher in patients with NSTE-ACS compared to those with STEMI. This may be explained by more pronounced vascular inflammation in our NSTE-ACS subset or alternatively, a higher prevalence of comorbidities associated with higher CRP concentrations such as diabetes mellitus [20] or renal dysfunction [21]. However, the association between CRP concentrations and time from symptom onset was stronger in STEMI patients who also had higher concentrations when undergoing sampling at later timepoints. CRP is a marker of the intensity of the inflammation in the infarcted myocardium [22] which explains why CRP elevations were more pronounced and longer-lasting in STEMI [8, 9].

Growth-differentiation factor-15

Concentrations of GDF-15 were higher in patients with more advanced age and those having STEMI. There was no sex-difference. Moreover, we observed no temporal variation in GDF-15 concentrations.

In animal models, GDF-15 is induced in the myocardium in response to pathological stress associated with ischemia-reperfusion injury or inflammation [23]. However, the data presented here provide no convincing evidence that GDF-15 might be involved in the early activation of pathways during ACS in humans. Accordingly, GDF-15 cannot be used for diagnostic purposes in this setting. Given the relative stability of its concentrations over time, GDF-15 is instead well suited as a marker of risk [24, 25] and might be useful as decision support concerning treatment measures. This is supported by recent data from our research group, identifying GDF-15 as the strongest contributor to adverse outcome in the recently published ABC-ACS ischemia score [26].

The importance of timing of biomarker measurement

Our findings emphasize the importance of timing aspects for the interpretation of biomarker results. This is an important issue when assessing populations with different manifestations of ACS. Many studies apply biomarker results obtained at fixed timepoints, e.g. upon admission or randomization into controlled trials [1, 3, 5, 8, 9]. Such a strategy does however not, take into account the dynamic nature of biomarker changes in ACS, as evidenced by our data. It is thus, essential to consider the delay time since onset of chest pain when interpreting differences in biomarker concentrations between individual patients or in cohorts with different manifestations of ACS. This applies not only to the cardiac troponins but to any biomarker exhibiting early concentration changes following myocardial ischemia, e.g. NT-proBNP and CRP. It should be emphasized that the assessment of true biomarker concentration changes during ACS requires serial sampling. However, current guidelines recommend such an approach only for diagnostic purposes and using cTn [27].

Strengths and limitations

Our investigation has several strengths. We investigated a broad and large international cohort of patients with ACS patients with high-quality data from a pivotal trial. This enabled us to study several sizable subcohorts of clinical relevance. Multiple biomarkers were assessed in parallel which provided the possibility to demonstrate and compare the dynamics of several key pathophysiological processes in ACS. The plasma samples were obtained before invasive procedures which allowed us to model the “natural course” of the disease without interference from interventional attempts to restore coronary blood flow.

There are also some limitations that deserve comment. Our cohort stems from a randomized controlled trial. Accordingly, it probably represents a somewhat selected subset of the general ACS population based on some of the exclusion criteria related to the tested study treatment. The diagnoses of the index events as STEMI and NSTE-ACS were defined by the local PLATO investigators without central adjudication. We focused on established biomarkers and GDF-15 which has an emerging role in cardiovascular medicine [24], [25], [26]. However, our findings cannot be translated to other biomarkers reflecting similar pathological processes. Temporal biomarker patterns were visualized based on a single blood sample in different patients with different delay times rather than on serial sampling in the same patient. Although our single-sample approach may infer some uncertainty concerning biomarker kinetics patterns, the confidence intervals of the spline curves were rather narrow which is explained by the very large sample size. We cannot exclude the possibility that biomarker concentration patterns might have been affected by early pharmacological treatments. Data on hemodynamics, coronary status and cardiac function had not been documented within the PLATO trial. We are for this reason not able to further explore the associations between temporal biomarker concentration patterns and these entities. Finally, it should be emphasized that the history on onset of symptoms is a rather imprecise estimate of the ischemic event in NSTE-ACS.