1Department of Electrocardiology, Institute of Cardiology, Jagiellonian University Medical College, Kraków, Poland

2Department of Electrocardiology, The John Paul II Hospital, Kraków, Poland

3Faculty of Medicine, Jagiellonian University Medical College, Kraków, Poland

4Department of Nephrology, Transplantology and Internal Medicine, Medical University of Gdansk, Gdańsk, Poland

5Department of Experimental Cardiac Surgery, Anesthesiology and Cardiology, Institute of Cardiology, Jagiellonian University Medical College, Kraków, Poland

6Krakow Centre for Medical Research and Technologies, The John Paul II Hospital, Kraków, Poland

Editorial

by Krittayaphong

see p. 1058

N-terminal pro-B-type natriuretic peptide (NT-proBNP) and high-sensitivity cardiac troponin have been identified as independent predictors of TE in AF [12]. It has been shown that the combination of the two biomarkers with basic clinical data may improve stroke risk assessment in AF [13] while mortality in AF can be predicted by combining age and heart failure status with growth differentiation factor-15 (GDF-15), NT-proBNP, and high-sensitivity cardiac troponin T (cTnT-hs) [14]. Cystatin C, a glomerular filtration marker associated with cardiovascular diseases, has been recently demonstrated as a predictor of acute ischemic stroke [15]. Renal function may affect circulating levels of biomarkers [16]. Landmark AF trials, which assessed biomarkers, generally excluded subjects with creatinine clearance below 30 ml/min [9, 17]. Therefore, little is known about the prognostic value of biomarkers in this subset of AF patients with CKD. It has been suggested that increased density of fibrin networks and prolonged clot lysis can predict TE and major bleeding events in anticoagulated AF patients [18–20]. Although we have recently shown that stage 4 CKD in AF patients is associated with unfavorable fibrin clot phenotype [5], its predictive value in terms of clinical outcomes, in this subgroup of patients, has not been evaluated yet.

We tested the hypothesis that GDF-15, cTnT-hs, and cystatin C, along with prothrombotic variables can predict complications of AF in patients with stage 4 CKD during anticoagulant therapy.

METHODS

We included 180 adults with a history of AF and stage 4 CKD, defined as 2 or more eGFR values in the range of 15–29 ml/min/1.73 m2 assessed using the CKD Epidemiology Collaboration (CKD-EPI) formula [10], enrolled in Kraków, Poland, between January 2014 and October 2017, with complete clinical and laboratory data. The inclusion and exclusion criteria were presented previously [10]. Briefly, the exclusion criteria were acute infection, liver injury, acute coronary syndrome in the preceding 12 months, and known active malignancy. The study was approved by the Bioethical Committee of the Jagiellonian University. All patients provided written informed consent.

Using a standardized questionnaire, we collected data on demographics, cardiovascular risk factors, comorbidities, and medications used. Definitions of risk factors and comorbidities were described previously [10].

Laboratory studies

Fasting venous blood samples were taken from the antecubital vein from the patients off anticoagulation. The patients on VKA were switched to a low-molecular-weight heparin, and blood samples were collected at least 12 hours after the last heparin injection. In the group of patients receiving rivaroxaban, or apixaban, blood samples were taken at least 24 h after the last dose. Routine laboratory tests were performed by standard laboratory techniques. The Clauss method was used to determine fibrinogen. Plasminogen activator inhibitor-1 (PAI-1) and thrombin-activatable fibrinolysis inhibitor were assessed by ELISAs (American Diagnostica, Stamford, CT, USA and Chromogenix, Lexington, MA, USA, respectively) [21]. Biomarkers such as GDF-15 and NT-proBNP were measured using electrochemiluminescence immunoassays (Roche Diagnostics, Mannheim, Germany) [21]. Endogenous thrombin potential (ETP) was quantified using calibrated automated thrombography, as previously described [21]. Fibrin clot permeability (Ks), reflecting the average size of fibrin clot network pores, was calculated using the following equation: Ks = Q × η × L / t × A × Δp, where Q indicates the flow rate in percolating time (t), η reflects the viscosity of liquid in the poise, L is the length of a fibrin gel, Δp is a differential pressure in dyne/cm2, while A indicates the cross-sectional area in cm2. Clot lysis time (CLT), indicating fibrin clot susceptibility to lysis was assessed as described [21]. CLT was measured from the midpoint of the clear-to-maximum-turbid transition to the midpoint of the maximum-turbid-to-clear transition.

Follow-up

The participants were followed up via phone calls at least twice a year or through clinic visits till January 2019. The occurrence of stroke/transient ischemic attack was the primary study outcome that was diagnosed based on the World Health Organization criteria. Major bleeding or clinically relevant non-major bleeding (CRNMB) were the secondary outcomes, which were defined as described previously [22, 23]. Deaths were also recorded.

Statistical analysis

Continuous variables are shown as the means (standard deviations) or medians (interquartile ranges) and were compared using the Student’s or Welch’s t-test or Mann-Whitney U test, as appropriate. P-values <0.05 were considered statistically significant. Statistical analyses were performed with the use of IBM SPSS Statistics for Windows (version 26, IBM Corp., Armonk, NY, USA) or STATISTICA (version 13.3, TIBCO Software Inc., Palo Alto, CA, USA), for Youden indexes and areas under the curves (AUC) calculations. For details see supplemental material.

RESULTS

Patients

A total of 180 AF patients (men 65%, median age; 71.0 [64.0–75.0] years) with stage 4 CKD (median eGFR, 24.0; interquartile range 21.0–25.0 ml/min/1.73 m2) and a median CHA2DS2-VASc score of 3.0 (2.0–4.0) were studied (Table 1).

The majority of the patients had hypertension, while diabetes mellitus was diagnosed in 35.6% of them, which were the main causes of CKD in this cohort (Table 1). None of the patients had proteinuria above 1 g per day. No lupus nephritis cases were included. Half of the patients were treated with VKA, while the other half was on apixaban or rivaroxaban (Table 1). Among the patients on VKA, the mean time in the therapeutic range was 65.6 (16.5)% while the percentage of patients with time in the therapeutic range ≥70.0% was 42.2% (n = 38).

Thromboembolism

During the median follow-up at 25.5 (23.0–29.0) months, thromboembolic events occurred in 18 (10%) patients (4.7% per year). There were 16 (8.9%) ischemic cerebrovascular events and 2 (1.1%) systemic thromboembolic episodes. There were no differences in the variables studied in relation to these events (Table 1 and Table 2).

However, in univariable and multivariable analyses, increased age and decreased Ks were associated with thromboembolism (Table 3).

Bleeding

Twenty-seven (15%) patients experienced bleeding (13 major and 14 clinically relevant non-major bleedings) while on anticoagulation (7.1% per year). No differences in the variables studied were observed between the 27 subjects and the remaining participants except for the increased prevalence of prior major bleeding and less frequent beta-blocker use (Table 1). The patients with a history of major bleeding were over 8-fold more likely to develop gastrointestinal (n = 12, 3.2% per year) bleeding (HR, 8.29; 95% CI, 2.49–27.62; P = 0.001) as compared to the others. Past bleeding, use of beta-blockers, increased GDF-15, cystatin C, and cTnT-hs were associated with major bleeds and CRNMB, along with prolonged CLT (Table 1 and Table 2), which, however, did not independently predict bleeding in contrast to previous bleeding and the three biomarkers (Table 3). The optimal cut-off points for the occurrence of bleeding estimated for, GDF-15, cTnT-hs and cystatin C were 1930.0 pg/ml (Youden index = 0.74; AUC, 0.885; 95% CI, 0.828–0.942; P <0.0001), 9.3 ng/l (Youden index = 0.62; AUC, 0.823; 95% CI, 0.747–0.899; P <0.0001) and 1.16 mg/l (Youden index = 0.48; AUC, 0.740; 95% CI, 0.657–0.823; P <0.0001), respectively (Supplementary material, Figure S1). The Kaplan-Meier survival analysis showed that GDF-15, cTnT-hs, and cystatin C discriminated AF patients with stage 4 CKD at high risk of bleeding on anticoagulation, with the highest value observed for GDF-15 ≥1930 pg/ml (Figure 1).

Mortality

Twenty-five (13.9%) patients died during the follow-up (6.5% per year), including 9 (5.0%) cardiovascular deaths. None of the four biomarkers and hemostatic parameters differed between non-survivors and survivors except for the fibrinogen concentration, which was slightly higher in the latter group. Mortality analysis is presented in Supplementary material, Table S1–S3. Regression analysis adjusted for age and comorbidities showed that solely cystatin C was independently associated with poor survival (HR, 3.95; 95% CI, 1.08–14.37; P = 0.04; Supplementary material, Table S3, Panel B).

DISCUSSION

In this cohort study, we have shown for the first time that in AF patients with stage 4 CKD, decreased fibrin clot permeability, indicating the formation of more compact fibrin networks, along with increased age, independently predict thromboembolic events that occur despite oral anticoagulation, while cystatin C, cTnT-hs, and GDF-15, together with past bleeding can independently predict clinically significant bleeding. Cystatin C was identified as the only independent predictor of all-cause death. The current findings expand our knowledge on the prediction of adverse events by suggesting that in AF patients with stage 4 CKD, emerging biomarkers might help identify the patients, who require closer ambulatory surveillance and education, as well as those who might particularly benefit in terms of anticoagulation safety from the measurement of biomarkers, especially GDF-15.

Biomarkers and prothrombotic state

We provided evidence that concentrations of GDF-15 in these high-risk AF patients are positively associated with a key marker of thrombin generation measured in plasma, i.e. ETP, and a marker of impaired fibrinolysis, i.e. CLT, which extends previous findings [5]. Similar observations were made in AF patients with a median eGFR of 73.0 ml/min/1.73 m2 [21]. There was no association between ETP and cardiac troponin I in patients with a median eGFR of 70.0 ml/min/1.73m2 [24]. The 3 biomarkers, which could be useful in risk stratification in AF patients and also acute pulmonary embolism [25], reflect, to some extent, the prothrombotic state typical of this arrhythmia, but their associations with other markers and prothrombotic state parameters at least in part seem to be diminished in severe CKD [5, 21, 24, 26].

Prediction of TE

Of importance, we found that low fibrin clot permeability, a key measure of fibrin clot density [18, 19], can predict TE outcomes in AF patients with stage 4 CKD, which extends our previous findings [18, 19] and highlights the role of fibrin properties in the pathophysiology of thromboembolism in AF. In contrast to Ks, we failed to observe a similar impact of hypolysability on this outcome, which suggests that changes in fibrin networks are more sensitive to most likely posttranslational modifications in severe CKD and in this subgroup have a stronger effect on the TE risk [20]. It might be speculated that therapies aimed at ‘normalizing’ fibrin clot structure might be useful in TE prevention [27].

Prediction of bleeding

We demonstrated that apart from prior major bleeding, cystatin C, GDF-15, and cTnT-hs can predict clinically relevant bleeding (not only major bleeds) also in AF patients with severe CKD, which supports previous observations on bleeding risk stratification [14, 28–30]. The 2020 European Society of Cardiology guidelines on the management of AF stated that biomarkers-based bleeding risk scores outperform clinical scores, for example, HAS-BLED, but there is insufficient evidence to implement them in practice [2]. Our study might suggest that in patients with low eGFR the predictive value of the novel biomarkers (but not eGFR assessed using the CKD-EPI formula) might be higher as compared to patients with better renal function, which warrants further studies.

Prediction of death

Limitations of the study

The current cohort was relatively small. We measured biomarkers and hemostatic parameters only once. The assessment of proteinuria in the current study was based on the records from nephrologists. It is unclear whether mild proteinuria can impact the hemostatic variables assessed in this study. Our results might not be extrapolated to patients on dialysis.

CONCLUSIONS

Supplementary material

Supplementary material is available at https://journals.viamedica.pl/kardiologia_polska.

Article information

Acknowledgments: This work was supported by a grant from Jagiellonian University Medical College (K/ZDS/007717, to A.U). P.T.M. was supported by the Ministry of Science and Higher Education stipend for outstanding young scientists.

Conflict of interest: AU received lecture honoraria from Bayer, Boehringer Ingelheim, and Pfizer, PTM received speech honorarium from Boehringer Ingelheim, while the rest of the authors declared no conflict of interest.

Open access: This article is available in open access under Creative Common Attribution-Non-Commercial-No Derivatives 4.0 International (CC BY-NC-ND 4.0) license, allowing to download articles and share them with others as long as they credit the authors and the publisher, but without permission to change them in any way or use them commercially. For commercial use, please contact the journal office at kardiologiapolska@ptkardio.pl.

How to cite: Matusik PT, Leśniak WJ, Heleniak Z, et al. Thromboembolism and bleeding in patients with atrial fibrillation and stage 4 chronic kidney disease: Impact of biomarkers. Kardiol Pol. 2021; 79(10): 1086–1092, doi: 10.33963/KP.a2021.0088.

REFERENCES