Highly sensitive serum cardiac troponin T and cardiovascular events in patients with systemic lupus erythematosus (TROPOPLUS study)

Abstract

Introduction

Overall mortality in SLE has largely improved in the past 50 years with a 5-year survival rate of over 90% since 1990. However, mortality rate is still increased by a factor of 2–5 in SLE as compared with the general population, mainly due to cardiovascular events (CVEs) [1–3]. Although traditional cardiovascular risk factors contribute to early-onset atherosclerosis in SLE, the phenomenon is not fully explained by a higher frequency of smoking habit, hypertension or dyslipidaemia. Because the Framingham risk equation usually underestimates the 10-year cardiovascular risk in this population [4–9], additional markers are needed to better stratify cardiovascular risks in SLE patients.

Cardiac troponin (cTnT) is a marker of myocyte necrosis and injury in the early phase of acute myocardial infarction [10]. Measured with high-sensitivity (HS) assays, high sensitivity cTnT (HS-cTnT) has proven predictive value for coronary heart disease, heart failure and mortality in the general population at apparent low-risk for CVE [11]. In a previous study, our group showed that HS-cTnT concentration was associated with subclinical atherosclerosis in SLE patients [12]. The aim of this study was to determine whether HS-cTnT was associated with incident CVE in SLE (TROPOPLUS study).

Patients and methods

Patients

The TROPOPLUS study is an ancillary study of the previously published PLUS study [13]. The PLUS Study was a randomized, double-blind, placebo-controlled, multicentre trial conducted from June 2007 through August 2010 at 37 centres in France involving 573 SLE patients. Details of the PLUS study methods including data collection and definitions of comorbid conditions have been previously published. For the TROPOPLUS study, subjects were considered to have hypertension if they repeatedly had a systolic blood pressure of at least 140 mm or a diastolic blood pressure of at least 90 mm. Subjects were considered to have dyslipidaemia if they had low-density lipoprotein-cholesterol >116 mg/dl [14]. The diagnosis of APS was based on a history of venous and/or arterial thromboses or recurrent miscarriages in the presence of aPL antibodies in accordance with published criteria [15]. Lupus nephritis diagnosis was based on International Society of Nephrology/Renal Pathology Society classification [16]. SLE patients with prior CVE at PLUS study entry were excluded. Supplementary Fig. S1 (available at Rheumatology online) shows the numbers of PLUS participants with adequate follow-up period (at least 20 months) and no past history of CVE at baseline for inclusion in the TROPOPLUS study. The PLUS Study was approved by the Comité de Protection des Personnes, St Louis Hospital, Paris. The TROPOPLUS study was approved by the Comité de Protection des Personnes SUD-EST II, Lyon. All participants gave written informed consent to participate at the time of study enrolment.

HS-cTnT measurement

All measurements were performed on serum samples stored at −70°C to −80°C and thawed just before testing (maximum of three freeze-thaw cycles) in October 2017. HS-cTnT measurements were performed on the e602 immunomodule of the cobas 8000 analyser (Roche Diagnostics, Meylan, France) using the HS-cTnT Elecsys^®2010 immunoassay. This assay is based on a one-step sandwich principle, with electrochemiluminescent revelation. The total duration of the assay is 18 min, and 50 μL of the sample is incubated with an anti-cTnT monoclonal antibody labelled with ruthenium and with a biotinylated monoclonal anti-cTnT antibody. According to the manufacturer, the measurement range of the assay is 3–10 000 ng/l; the limit of detection is 2.05 ng/l (internal study from manufacturer, according to EP17-A2 protocol from Clinical and Laboratory Standards Institute); the 99th percentile value is 14 ng/l, and the 10% coefficient of variation value is 13 ng/l (data from the manufacturer). Our laboratory coefficient of variation during the study measurements was 2.6% (at 29.6 ng/l) and 2.4% (at 2196 ng/l) as previously reported [12].

Primary outcome

The primary outcome was the occurrence of CVE over follow-up. Incidents of CVE were ascertained, blinded to HS-TnT concentration, by physician interview using a standardized questionnaire and through examination of medical records. CVE included coronary heart disease, (i.e. hospitalization for unstable angina, coronary revascularization or myocardial infarction), stroke, revascularization procedure for other atherosclerotic cardiovascular diseases and sudden cardiac death. All CVE that occurred through March 2019 were considered for analysis. For patients who had >1 incident CVE, the first event was considered for analysis.

Statistical analysis

Descriptive statistics were conducted; continuous variables were expressed as median [interquartile range (IQR)] and categorical variables were expressed as number and percentage. When HS-cTnT concentration was undetectable (<3 ng/l), the value was imputed by the limit of quantification divided by two (1.5 ng/l). Univariate and multivariate Cox proportional-hazards models were used to estimate hazard ratios (HRs) for CVE with 95% CIs. Univariate Cox models were used to evaluate the association between clinical, biological and therapeutic variables and the risk of CVE. Factors identified in univariate analysis with a P-value <0.2 were used in a multivariate Cox regression model to identify those independently associated with CVE. The ‘Maximally Selected Log-Rank statistic’ method [17] was used to calculate the optimal threshold of HS-cTnT concentration to predict a major CVE. The HS-cTnT concentration that maximised the standardised test statistic was selected (Supplementary Fig. S2, available at Rheumatology online). The risk of CVE stratified by the calculated HS-cTnT concentration threshold was analysed using the Kaplan–Meier method and groups were compared using a log-rank test.

Results

Characteristics of SLE patients

In total, 573 patients included in the PLUS study at 37 centres in France from June 2007 through August 2010 were eligible for the study. Among them, 127 patients were not included in the TROPOPLUS study because they were lost to follow-up (n = 126) or had a CVE prior PLUS inclusion (n = 1). In addition, four patients were excluded because data regarding the occurrence of CVE during follow-up were not exploitable. Overall, 442 SLE patients were included in the TROPOPLUS study and analysed for the primary outcome with a median follow-up of 110 (IQR: 99–120) months (Supplementary Fig. S1, available at Rheumatology online). At inclusion, 400 (90.5%) patients were female with a median age of 37 (IQR: 29–48) years. A total of 99 (22.4%) were current smokers, 60 (13.6%) had hypertension, 51 (11.6%) had a BMI >30; 34 (7.7%) had dyslipidaemia and 11 (2.5%) had diabetes. In keeping with the exclusion criteria of the PLUS study, no patients had a chronic kidney disease, as defined by an eGFR<60 mL/min/1.73 m². The median duration of SLE disease at baseline was of 84 (IQR: 36–144) months and only 47 (10.6%) had a SLEDAI >4. Phospholipid antibodies were found in 217 (49.1%) patients including 71 (16.1%) with an APS. Hydroxychloroquine, steroid, immunosuppressive drugs, antiplatelet, biologics and statin treatments have ever been prescribed in 98.9%, 86.9%, 50.7%, 28.1%, 8.4% and 6.6% of patients, respectively (Table 1).

CVEs during follow-up

According to the Framingham score, the estimate absolute risk for CVEs within the next 10 years for the overall cohort was 0.1%. However, 29 (6.6%) patients experienced at least one CVE that occurred at a median of 67 (IQR: 31–91) months after inclusion. CVEs included coronary heart disease (n = 14) that required percutaneous coronary revascularization in all but two cases, nonfatal ischaemic stroke (n = 11), peripheral arterial disease requiring revascularization treatment (n = 2), dissection of an aortic aneurysm requiring surgery (n = 1) and sudden cardiac death (n = 1). Six patients had more than one CVE.

Association of HS-cTNT with CVE in SLE patients

At inclusion, the median level of serum HS-cTnT was 3.3 (IQR: 1.5–5.1) ng/l and serum HS-cTnT was detectable (i.e. >3 ng/l) in 236 (53.4%) patients. HS-cTnT concentration was higher in SLE patients in whom a CVE eventually occurred during follow-up [4.9 (IQR: 1.5–9.5) vs 3.21 (IQR: 1.5–4.9) ng/l, P =0.0067] (Fig. 1). HS-cTnT concentration was even higher in SLE patients who developed more than one CVE event during follow up [11.3 (IQR: 6.8–16.7) as compared with 4.3 (1.5–7.4) ng/l in those with 1 CVE only, P =0.0058].

Fig. 1

HS-cTnT concentration in SLE patients

Serum level of HS-cTnT at baseline is higher in SLE patients who present CVE during follow-up as compared to those who did not (no CVE). Box plots of median level of HS-cTnT in both groups are represented. Analysis was performed on 440 patients (HS-cTnT was missing in two patients without incident CVE); P =0.0067). CVE: cardiovascular event; HS-cTnT: high-sensitivity cardiac troponin T.

Open in new tabDownload slide

Univariate analysis of risk factors for CVE (Table 2) revealed that age [HR 1.79 (95% CI: 1.37, 2.34), P <0.0001], dyslipidaemia [HR 4.88 (95%CI: 2.44, 9.76), P <0.0001], hypertension [HR 2.89 (95%CI: 1.39, 6 P =0.0045], duration of SLE disease [HR 1.06 (95% CI: 1.02, 1.10), P =0.0024] and HS-cTnT [HR 1.01 (95% CI: 1.00, 1.02), P =0.0066] were significantly associated with CVE in SLE. As expected, duration of SLE disease strongly correlated with age (Pearson r 0.32, P <0.0001). In a multivariate Cox regression model with CVE as the dependent variable, dyslipidaemia [HR 3.14 (95% CI: 1.25, 7.88), P =0.0005], age [HR 1.65 (95% CI: 1.25, 2.19), P =0.0150] and HS-cTnT [HR 1.13 (95% CI: 1.04, 1.23), P =0.0044] were still associated with the occurrence of CVE (Table 3). By using maximally selected Log-Rank statistic, the cut point in HS-cTnT value that provided the best separation between the SLE patients who developed CVE and those who did not was 4.27 ng/l (Supplementary Fig. S2, available at Rheumatology online). Accordingly, Kaplan–Meier analysis showed that a concentration of HS-cTnT > 4.27 ng/l in serum at inclusion was significantly associated with the occurrence of CVE during follow up [HR 2.7 (95% CI: 1.3, 5.6), P =0.0061] (Fig. 2).

Fig. 2

Kaplan–Meier curves reflecting cumulative proportion of SLE patients free of CVE according to a baseline HS-cTnT

Analysis was performed on 440 patients (HS-cTnT was missing in two patients without incident CVE). CVE: cardiovascular event; HS-cTnT: high-sensitivity cardiac troponin T.

Open in new tabDownload slide

Discussion

We show that detectable serum cTnT, measured with a highly sensitive assay, is independently associated with incident CVE in SLE patients at otherwise apparent low risk for CVE. Hence, HS-cTnT appeared as a unique biomarker that might help predict long-term CVE in SLE.

Based on traditional cardiovascular risk factors (low to very low frequencies of hypertension, tobacco use, overweight, dyslipidaemia and diabetes) and SLE characteristics (no patients with eGFR< 60 mL/min/1.73 m² at inclusion) [18], SLE patients in our cohort appeared to be at a very low risk for CVE. Over time, CVE occurred in 29 patients (6.9%). The fact that such CVE rate was >60 times the estimate by the Framingham score (0.1% for the overall cohort) clearly shows that prediction models based on traditional cardiovascular risk factors are inaccurate in SLE.

HS-cTnT levels – measured in patients who had no symptoms or diagnosis of CVE upon study entry – likely reflect latent ongoing cardiac myocyte damage. Detectable HS-cTnT was previously shown to be associated with coronary heart disease, mortality and hospitalization for heart failure in individuals from a general population without known coronary heart disease or stroke [11]. In a population-based cohort of Dallas County, the prevalence of detectable HS-cTnT was 25% in 3546 individuals aged 30–65 years [19]. Interestingly, the frequency of detectable HS-cTnT in our study (i.e. 53.4%) was closer to the rate observed in patients older than 60 [19, 20]. In the same line, a HS-cTnT level ≥10 ng/l was found in 24 subjects (2.2%) with the highest value of 20 ng/l in a work site-based population of 1072 middle-aged males (mean age 44 years) without any history or presence of CVE [21]. In our series of SLE patients (90.5% female, median age 37), HS-cTnT levels ≥10 ng/l were found in 35 subjects (7.9%) with the highest value being 289 ng/l.

Our study has several strengths. First, it is an ancillary longitudinal study of a large prospective and multicentre trial including >400 SLE with an extended follow-up. Second, the unbiased samples strategy likely allows an accurate estimation of the prevalence of detectable HScTnT among SLE patients. Third, and in contrast to most studies assessing the risk for CVE in SLE, our primary end point did not focus on surrogate markers of subclinical atherosclerosis but on incident CVE. Our study also has several limitations. First, primary outcome was not available for all PLUS participants. Second, primary outcome was not adjudicated. Third, only a single baseline measurement – both of HScTnT level and other covariates – was performed. Finally, the observed number of CVE was small, limiting statistical power.

In conclusion, we show that detectable serum HS-cTnT is independently associated with incident CVE in asymptomatic SLE patients at low risk for cardiovascular disease according to traditional risk factors. HS-cTnT could be a biomarker useful for risk stratification and primary prevention of CVE in SLE patients.

See Acknowledgements section for a list of collaborators (PLUS study group).

Acknowledgements

We are thankful to Mrs Ada Clarke for help with study organization and to Mrs Claire Fernandez for technical assistance.

J.C. was supported by a research grant from the Fondation pour la Recherche Medicale (FRM DEA20170638077).

List of collaborators (PLUS study group): Leonardo Astudillo, Cristina Belizna, Nadia Belmatoug, Olivier Benveniste, Audrey Benyamine, Holly Bezanahary, Patrick Blanco, Benoît Brihaye, Patrice Cacoub, Emmanuel Chatelus, Judith Cohen-Bittan Richard Damade, Eric Daugas, Christian de-Gennes, Jean-François Delfraissy, Céline Delluc, Aurélien Delluc, Alain Dupuy, Isabelle Durieu, Hang-Korng EA, Dominique Farge, Camille Frances, Christian Funck-Brentano, Frédérique Gandjbakhch, Justine Gellen-Dautremer, Bertrand Godeau, Cécile Goujard, Catherine Grandpeix, Claire Grange, Gaëlle Guettrot-Imbert, Loïc Guillevin, Eric Hachulla, Jean-Robert Harle, Julien Haroche, Pierre Hausfater, Jean Sebastien Hulot, Moez Jallouli, Jean Jouquan, Gilles Kaplanski, Homa Keshtmand, Mehdi Khellaf, Olivier Lambotte, David Launay, Du Le Thi Huong, Gaelle Leroux, Hervé Levesque, Olivier Lidove, Frederic Liote, Eric Liozon, Kim Ly, Matthieu Mahevas, Kubéraka Mariampillai, Xavier Mariette, Karin Mazodier, Marc Michel, Luc Mouthon, Rokiya Ngack, Jacques Ninet, Eric Oksenhendler, Jean-Luc Pellegrin, Olivier Peyr, Anne-Marie Piette, Jean-Charles Piette, Vincent Poindron, Fabienne Roux, David Saadoun, Sabrinel Sahali, Laurent Sailler, Bernadette Saint-Marcoux, Yoland Schoindre, Jérémie Sellam, Damien Sene, Jacques Serratrice, Pascal Seve, Jean Sibilia, Claude Simon, Amar Smail, Christelle Sordet, Jérôme Stirnemann, Benjamin Terrier, Salim Trad, Jean-François Viallard, Pierre-Jean Weiller, Noël Zahr.

Funding: The PLUS study was funded by a grant from the French PHRC 2005 (number 05–125) Ministère de la Santé. The TROPOPLUS study was funded by a grant from the French CRC 2017 (number 17–068) Ministère de la Santé.

Disclosure statement: The authors have declared no conflicts of interest.

Supplementary data

Supplementary data are available at Rheumatology online.

References

Author notes