Cardiac troponin I for predicting right ventricular dysfunction and intermediate risk in patients with normotensive pulmonary embolism

Background Right ventricular dysfunction (RVD) and cardiac troponin I (cTnI) are important tools for risk stratification in pulmonary embolism (PE). We investigate the association of RVD and cTnI in normotensive PE patients and calculate a cTnI cut-off level for predicting RVD and submassive PE. Methods Clinical, laboratory, radiological and echocardiagraphic data were analysed. Patients were categorised into groups with or without RVD and compared focussing on cTnI. Effectiveness of cTnI for predicting RVD and submassive PE was tested. Results One hundred twenty-nine normotensive PE patients, 71 with and 58 without RVD, were included. Patients with RVD were older (75.0 years (61.3/81.0) vs. 66.0 years (57.7/75.1), P = 0.019). cTnI (0.06 ng/ml (0.02/0.23) vs. 0.01 ng/ml (0.00/0.03), P < 0.0001) and D-dimer values (2.00 mg/l (1.08/4.05) vs. 1.23 mg/l (0.76/2.26), P = 0.016) were higher in PE with RVD. cTnI was associated with RVD (OR 3.95; 95 % CI 1.95–8.02, p = 0.00014). AUC for cTnI diagnosing RVD was 0.79, and for submassive PE0.87. Cut-off values for cTnI predicting RVD and submassive PE were 0.01 ng/ml, with a negative predictive value of 73 %. cTnI was positively correlated with age, D-dimer and creatinine. Conclusions In normotensive PE patients, cTnI is helpful for risk stratification and excluding RVD. cTnI elevation is correlated with increasing age and reduced kidney function.

We aimed to investigate the association of RVD and cardiac troponin I (cTnI) in normotensive PE patients and the potential cTnI for exclusion of RVD. Moreover, a further objective was to investigate parameters that have an impact on elevation of cTnI levels.

Patients and methods
We performed a retrospective analysis of normotensive patients with confirmed acute PE, who were treated in the Internal Medicine Department of the St. Vincenz and Elisabeth Hospital Mainz, Germany, between May 2006 and June 2011.
Patients were identified by searching the hospital information system database (ICD diagnostic code of PE: I26).

Enrolled subjects
Patients were eligible for our analysis if they had a confirmed acute PE, were treated in the Internal Medicine Department of the hospital, were haemodynamically stable (systolic blood pressure ≥90 mmHg) according to the European Society of Cardiology (ESC) guidelines [4] and American Heart Association (AHA) scientific statement [21], an accurate transthoracic echocardiography (TTE) of acute phase was assessed and patients were at least 18 years of age. Diagnosis of acute PE was confirmed by an identified filling defect in the pulmonary artery system in a computed tomography pulmonary angiogram of the chest (CT) or positive venous ultrasound/phlebography of an extremity consistent with deep vein thrombosis (DVT) in patients with typical symptoms of PE (chest pain or dyspnoea) and a detected positive D-dimer or scintigraphic ventilationperfusion (V/Q) scan read as high probability for PE. If the diagnosis of PE was not confirmed or the criteria mentioned above were not fulfilled, patients were not included in this analysis. Studies with retrospective analysis of diagnostic standard data do not need an ethics statement in Germany.

Definitions
Per definition, cTnI values were elevated if they exceeded 0.1 ng/ml, and D-dimer values if they exceeded 0.110 mg/l. RVD in TTE was defined according to the ESC guidelines [4] and AHA scientific statement [21] as abnormal motion of the interventricular septum, right ventricular (RV) hypokinesis, pressure overload with tricuspid valve insufficiency or RV enlargement (quotient of end-diastolic lateral-septal RV diameter/end-diastolicleft ventricular lateral-septal diameter>0.9, referring to AHA scientific statement [21]).
According to the ESC guidelines [4] and AHA scientific statement [21] the PE patients of our study with RVD or pathological cTnI levels were classified as submassive PE with an intermediate risk and those without both were classified as low-risk PE.

Study groups
According to TTE results the normotensive PE patients were included in one of the two study groups: the PE group with RVD or the PE group without RVD.

Statistical analysis
We compared the groups especially focussing on cTnI. Effectiveness of cTnI to predict RVD was tested. Continuous distributed variables were described by median and interquartile range and tested with the Mann-Whitney U-test if they were not normally distributed or by mean, standard deviation (SD) and tested with the Student t-test if distribution was normal. Discrete variables were compared with the Chisquare test. A value of P<0.05 was considered to be statistically significant.
Receiver operating characteristic (ROC) curve and Youden index were calculated to test the effectiveness of cTnI to predict RVD as well as cTnI to predict submassive PE. Multivariate regression analysis was performed to test coherence of cTnI, gender, age, CK, creatinine, and D-dimer on RVD. Spearman's rank correlation was calculated for age, cTnI, CK, creatinine, D-dimer, systolic blood pressure, diastolic blood pressure, heart rate and systolic pulmonary artery pressure (sPAP).
R version 2.14.1 from R Development Core Team (2011) (R Foundation for Statistical Computing, Vienna, Austria) was used for data processing.   Table 1). cTnI was not higher in PE patients >65 years than in those ≤65 years (0.03 ng/ml (0.01/0.16) vs. 0.02 (0.00/0.08), P= 0.10). In the PE group without RVD, 10.3 % of patients showed positive cTnI values. Of the 129 PE patients, 52 were classified as low-risk PE and 77 as submassive PE (Table 1).
Our study confirms a strong association between cTnI and RVD in normotensive PE patients. The calculated OR of 3.95 is in accordance with the literature [27,32,33]. AUC for cTnI predicting RVD was 0.79, which is similar to published AUC values of other studies [34,35].
We calculated a cut-off value of 0.01 ng/ml for cTnI predicting RVD [24], which is in concordance with other study results [34]. Sensitivity, specificity, PPV and NPV were similar to other studies [34] and especially the PPV and NPV were high [24]. These results of AUC, cut-off value, sensitivity, specificity, PPV and NPV were previously published in an investigation for the comparison of three variations of RVD definitions [24]. But, in our opinion, these data are important to complete the general view on this topic and we have therefore quoted them once again.
Only 10.3 % of the PE patients without RVD showed an elevated cTnI value. These patients had myocardial damage without detected RVD and therefore were also categorised as submassive PE with intermediate risk.
At a first glance, calculation of an AUC for cTnI predicting the submassive stage of PE seems to make no sense. But, after careful consideration this calculated AUC could be used to evaluate the importance of the parameter cTnI for assessment of submassive PE. Especially the small difference between the AUC of cTnI predicting RVD and  In a previous investigation, in 179 PE patients of the identified 182 PE patients, we analysed the association of cTnI for risk stratification in haemodynamically unstable and normotensive PE patients together [36]. In the larger analysed group of PE patients, RVD diagnosis was made not only by TTE, but also by CT. In 3 of the 182 patients RVD was not detectable and these 3 patients were therefore not included in the analysis [36]. In these 179 PE patients the association between cTnI and RVD was similar to the present analysis (OR 3.98) [36]. AUC for cTnI predicting RVD was 0.81 with a cut-off level of 0.01 ng/ml [36]. Increasing age, reduced kidney function and higher activation of coagulation system are connected with elevated cTnI values.
In all patients with an acute PE event, anticoagulation therapy is recommended with the aim to prevent early death and to avoid symptomatic recurrent VTE, especially fatal PE [31]. Thrombolytic treatment is not the standard treatment regime in normotensive PE patients, but could be considered if signs of haemodynamic decompensation appear [31]. In the 129 normotensive PE patients of our study we started anticoagulation treatment, but thrombolytic treatment was not given in any of these patients.

Limitations and strength
Limitations of this study are the small number of included patients and the retrospective character. With these data, we were not able to draw conclusions about follow-up outcome. Despite these limitations we were able to find answers to the main questions of the analysis.

Conclusions
In normotensive PE patients, cTnI is helpful for risk stratification and especially for exclusion of RVD. Increasing age, reduced kidney function and higher activation of coagulation system are connected with elevated cTnI values.

Conflict of interest None.
Funding None.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.