Interact CardioVasc Thorac Surg 2007;6:323-327. doi:10.1510/icvts.2006.146076
© 2007 European Association of Cardio-Thoracic Surgery
Institutional report - Assisted circulation |
Association between prothrombin activation fragment (F1.2), cerebral ischemia (S-100ß) and international normalized ratio (INR) in patients with ventricular assisted devices ,
Ashish Joshia,
Laurence S. Magdera,
Zachary Konb,
Seeta Kallamb,
Michael Kwonb,
Rupali Sangrampurkarb,
Richard Piersonb and
Robert Postonb,*
a Department of Epidemiology and Preventive Medicine, University of Maryland, Baltimore, Maryland, USA
b Division of Cardiac Surgery, University of Maryland School of Medicine and Veterans Affairs Medical Center at Baltimore, N4W94 22 S. Greene St, Baltimore, MD 21201, USA
Received 3 October 2006;
received in revised form 9 February 2007;
accepted 12 February 2007
 Presented at the joint 20th Annual Meeting of the European Association for Cardio-thoracic Surgery and the 14th Annual Meeting of the European Society of Thoracic Surgeons, Stockholm, Sweden, September 10–13, 2006. 
This work was funded in part by grants from the Office of Naval Research, the Thoracic Surgery Foundation for Research and Education, World Heart, Inc., the Maryland Cigarette Restitution Fund Program, and the American Heart Association.
Robert S. Poston and Richard N. Pierson III disclose that they have received grant support from World Heart, Inc. Robert S. Poston discloses that he has received grant support and honoraria from Bayer, Inc.
*Corresponding author. Tel.: +1-410-328-5842; fax: +1-410-328-2750.
E-mail address: rposton{at}smail.umaryland.edu (R. Poston).
 |
Abstract
|
|---|
Prothrombin time, expressed as international normalized ratio (INR) and activated partial thromboplastin time (aPTT), are standard methods of monitoring coumadin and heparin administration. Prothrombin activation fragment (F1.2) is an index of in vivo thrombin generation. We hypothesized that F1.2 would provide a better surrogate of thromboembolism risk than standard coagulation assays during ventricular assist device (VAD) support. INR, PTT and F1.2 were analyzed in 31 patients after implantation of a left-sided VAD daily during hospitalization and weekly after discharge. Thromboembolic events (TE) were defined by evidence of neurological injury revealed by plasma levels of S-100ß. The relationships between F1.2, INR for patients on coumadin and aPTT for patients on heparin were evaluated from 1250 observations of blood samples. S-100ß was positively correlated with F1.2, but not with INR and aPTT. Correlation between S-100ß and F1.2 is significantly higher than with the other two markers (P<0.0001). Higher values of aPTT and INR were not associated with TE. Compared to conventional coagulation assays, the F1.2 level provides a single endpoint that is a more accurate predictor of TE after VAD implantation. Further trials that incorporate the F1.2 marker into anticoagulation algorithms may help reduce adverse events in this high-risk population.
Key Words: Heart-assist device; Thrombosis; Coagulation; Stroke
|
1. Introduction
|
|---|
Ventricular assist devices (VADs) are most common and reliable tools to provide mechanical circulatory support for patients with heart failure. Long-term success of VAD is limited due to several factors: multiple organ failure (MOF), bleeding, thromboembolism and infection [1]. Several reasons include contact between blood components and foreign surfaces of assist device system, altered rheologic conditions with different blood flow velocities and blood stasis in the recipient heart [2]. Thromboembolic event (TE) incidences (8.1% and 47%) vary widely between centers [3]. There are few clinical studies investigating mechanisms of TE in these VAD patients. S-100ß, a brain protein released from glial cells after injury, provide a more sensitive measure of TE.
The prothrombin time (PTT), international normalized ratio (INR), and platelet count remain standard of care for monitoring the status of coagulation system after VAD implantation. Abnormalities in platelet function have been detected after VAD placement [4], but little correlation with TE is seen. TE events occur despite therapeutic ranges for PTT and INR [5]. Prothrombin fragment F1.2, is a sensitive assay for in vivo thrombin production and a common pathway for thrombus development.
There is evidence of greater ability to predict TE in VAD patients using F1.2 assay as compared to the other available tests [6].
The main objective of this study was to assess associations between TE defined by S-100ß and markers of coagulation activity: F1.2, INR and PTT.
|
2. Materials and methods
|
|---|
2.1. Patient enrollment
Patients aged 18 years or older and who provided informed consent (UM IRB protocol No. H-22995) were scheduled for implantation of a left-sided VAD at a single institution (JarvikTM, n=6, NovacorTM, n=14, ThoratecTM, n=2, VentricorTM, n=4 or HeartmateTM, n=5) between July 2004 and September 2005. Twenty-five patients had VAD implantation via median sternotomy while six patients underwent left thoracotomy using cardiopulmonary bypass support.
In all patients, blood flow within pulsatile VADs (NovacorTM, HeartmateTM and ThoratecTM) was recorded every four hours until hospital discharge, and also during all subsequent inpatient visits.
Antithrombotic therapy and perioperative blood product usage were guided by data from thrombelastography (TEGTM, Haemoscope; Niles, IL) [7]. Aspirin (325 mg/day) was administered within 6 h of surgery, and titrated thereafter to a maximum dose of 650 mg/day as necessary [8]. When transitioning to oral anticoagulation, coumadin was titrated to an INR of 2.0–3.0.
2.2. Heparin management
Intravenous heparin was initiated at a rate of 5 USP/kg/h, and increased to achieve target aPTT between 50–65 s.
2.3. Platelet function
Platelet function was defined by maximum amplitude (MA) of TEG trace and impedance change on whole blood aggregometry (WBA, Chronolog; Hawerton, PA). Impedance changes ( ) on WBA assessed responsiveness to antiplatelet therapy at 6 min following addition of 1 and 5 µg/ml collagen (an assay of the aspirin response [9]) and 3 µM ADP.
2.4. Coagulation activity
Citrated blood samples were analyzed by a series of coagulation assays including PTT and INR by clinical laboratory and F1.2 levels using an ELISA kit (Enzygnost F1.2 Micro, Dade-Behring; Deerfield, IL). These blood samples were obtained prior to skin incision (baseline), after protamine administration (postoperation), daily during acute hospitalization and weekly after discharge. A twofold or greater increase in F1.2 levels above a baseline of  0.5 nM/mg protein was defined as a clinically important elevation based on prior reports [10].
2.5. Platelet factor four antibody
These antibodies have been associated with an increased risk of TE in prior VAD studies [11]. Anti-PF4 absorbance 0.4 OD was defined as a positive result with this assay [12].
2.6. Detecting thromboembolic events (TE)
Daily neurological physical examinations and routine preoperative and monthly head computed tomography (CT) scanning was used to screen patients for clinical stroke. Findings were confirmed by neurologist consultation.
S-100ß, a marker of cerebral ischemia [13] and evidence of neurological injury was measured using commercially available ELISA kit (SYN•X Pharma Inc.; Toronto, Ontario). Based on observations from our pilot data, a clinically important rise in S-100ß was defined as  twofold increase above baseline (0.5 nM S-100ß/mg protein) and this rise in S-100ß was defined as TE.
2.7. Statistical methods
Descriptive statistics for continuous and categorical variables were reported on 1250 observations. To assess the degree to which S-100ß correlated with coagulation biomarkers for each patient, we used Pearson correlation coefficients. To assess the degree to which various variables predicted TE events, we used conditional logistic regression, conditioning on patient. This is an approach to model multiple event data that appropriately accounts for the correlation between repeated events within an individual. This approach also implicitly controls for fixed characteristics of the individuals such as age, race, gender etc. We used mixed effect model to compare means of F1.2, S-100, INR and PTT on days with/without TE to account for repeated measures within each patient. All analysis was performed using SAS version 9.1 and all reported P-values are two-sided.
|
3. Results
|
|---|
Thirty-one VAD patients were analyzed. Eight patients got heart transplant and eight other patients died (three due to perioperative respiratory failure, one due to intracranial bleeding after a fall, one due to ischemic small bowel, and four due to MOF). Fifteen patients are currently on VAD support (Fig. 1). The follow-up time varied between patients from 18–432 days with approximately 50% of the patients being followed for 50–150 days. Given that biomarkers were assessed daily while the patient was in the hospital and weekly after discharge, the number of assessments per patient varied from 7–80.
3.1. Socio-demographics
The average age of the VAD patients was 49 years (S.D.=18). More than half of these subjects were males (68%), whites (55%) and an average body mass index of 29 (S.D.=7) (Tables 1 and 2)
3.2. Preoperative analysis
Preoperative analysis was performed for the laboratory variables S-100ß, F1.2 and INR. Preoperatively, average S-100ß levels was 0.45 Nm/mg protein (S.D.=0.28; range=1.14), F1.2 was 0.35 Nm/mg protein (S.D.=0.30; range=1.04). INR was 1.64 (S.D.=0.80; range=3.30) and PTT was 40 s (S.D.=8.43; range=28).
3.3. Thromboembolism
One hundred and seventy-six episodes of TE (as defined by a two-fold rise in S-100ß) were seen in our cohort. There were ten patients (32%) with no TE, two patients (6%) with one TE and nineteen patients (62%) with more than one TE (mean 4±3 events, range of 10). Six patients had VAD related infections, but this event showed no association with TE. Based on the days with/without TE, S-100ß, F1.2, INR, PTT and blood flow within the VAD were compared across both the groups. The average S-100ß and F1.2 levels during TE was 1.07 (S.D.=1.22) and 0.89 (S.D.=1.50) and were higher compared to days without a TE (Table 3).
View this table:
[in this window]
[in a new window]
|
Table 3 Distribution of mean, standard deviation and range of F1.2, S-100ß, INR and PTT levels during all the patient days of observations and also during days with/without TE. Proc mixed method was used to compare difference in means in F1.2, S-100ß, INR and PTT in two groups on days when there was TE compared to those days when there was no TE
|
|
3.4. Correlation between S-100ß and coagulation markers
Fig. 2 shows a box plot of patient specific Pearson correlation coefficients between surrogates for TE S-100ß and coagulation biomarkers (F1.2, PTT, and INR). It is evident that for the large majority of patients, S-100ß tends to be positively correlated with F1.2, whereas this is not the case for the other two biomarkers. The mean correlation between S-100ß and F1.2 is significantly higher than with the other two markers (P<0.0001). Of all 31 VAD patients, five patients developed stroke based on clinical and CT examinations. One patient had four strokes while others had one stroke each. Each stroke was associated with a concurrent increase in S-100ß (Fig. 3).
View larger version (6K):
[in this window]
[in a new window]
|
Fig. 2. Box plots showing the distribution of patient-specific correlation coefficients between S-100ß and each biomarker. In these plots, the boxes represent the location of the middle 50% of the correlation coefficients, and the lines in the boxes represent the medians.
|
|
3.5. Predicting TE
Using the conditional logistic regression models that implicitly control for fixed characteristics of individuals, we found that higher values of F1.2 were significantly associated with TE events (OR=2.3 per difference of one S.D. in the F1.2 value, P<0.0001). However, higher values of aPTT and INR were not statistically associated with TE (OR=1.1 for each variable).
From a clinical standpoint, it would be useful to know if previous values or previous changes in F1.2 were significantly associated with a future TE. To assess this, we explored whether high values of F1.2 on a given day predicted a TE on the next day. Using a conditional logistic regression, and limiting the analysis to visits one day apart, we did find that when the F1.2 value was more than double the baseline value, the odds of a TE on the next day was increased by a factor of 2.2 (95% confidence interval 1.0–4.7, P=0.044). Previous high values of INR and aPTT were not significantly predictive of TE on the next day: For PTT, the OR=1.6, P=0.55, while for INR the OR=1.0, P=1.0).
|
4. Discussion
|
|---|
Despite strict adherence to evidence-based anticoagulation protocols, TE was a persistent problem in our VAD cohort. Our results showed that F1.2 was a significant independent predictor of TE. Studies have shown wide variety of coagulation changes, including higher F1.2 levels, in VAD patients.
A sensitive analysis of the value of F1.2 requires a surrogate marker of TE risk because clinical TE is uncommon. We defined neurologic injury by using S-100ß as a proxy for subclinical injury. Doppler sonography is used to predict clinical stroke [12] by detecting microembolic signals presence in cerebral vessels. While this technique is accurate, it is not widely available. MRI examinations are not feasible in the presence of a metallic VAD. Although it is convenient to use S-100ß to assess neurologic injury, it is confounded in the perioperative cardiac surgery patient by the additional presence of this marker in mediastinal tissue [14]. Therefore S-100ß values measured within 24 h of surgery were excluded in order to minimize the influence of retransfused mediastinal blood. Given this study design, the changes detected in S-100ß were highly likely to be induced by and specific for neurological injury in our VAD patients, as shown by other studies [15].
An INR target of 2–3 was chosen as range less likely to provoke bleeding in VAD patients who are routinely on dual antiplatelet therapy. Maintaining an INR 2.5–3.5 by AHA guidelines for mechanical prostheses was made in context of sole coumadin therapy without antiplatelet agents. The lack of correlation between INR and S-100ß in our study provides support for this approach. Thirty percent of INR observations were outside target range of 2.0–3.0. In reports [2], traditional coagulation tests (aPTT, INR) failed to detect clinical situations where TE risk was elevated. We speculate that platelet and thrombin interactions, both of them together are necessary for detectable thrombin burst to occur and direct clinical phenotype with respect to TE risk.
The study is limited by interpatient differences in important characteristics such as the medical history, post operative management strategy and variability in surgical interventions. In addition, the small number of VAD subjects that are available at a single institution limits the statistical power of our results. ‘Asymptomatic’ rise in S-100ß levels need to be further evaluated in the future or less obvious clinical abnormalities (e.g. neurocognitive dysfunction, depression, reduction in quality of life, etc.)
Our data strongly support prospective evaluation of F1.2 as a surrogate marker for TE risk in VAD patients, and as an independent point-of-care test used to direct the intensity of antithrombotic for patients at increased risk of TE complications. When coupled with more accurate, physiologically relevant assays of platelet function, such a strategy may minimize the risk of both bleeding and TE in these patients and unlock the potential of VAD therapy to alleviate the current physiologic and financial burden of end-stage congestive heart failure.
|
Conference discussion
|
|---|
Dr. G. Laufer (Innsbruck, Austria): I do not really understand if you are correlating this fibrinogen fragment at the time when the thromboembolic event occurs, or do you also investigate the titer of the fragment one day or two days before the thromboembolic event occurs?
It's not surprising that in case of a thromboembolic event a fragment of the fibrinogen molecule is elevated. I think this is something which is self-understanding.
Dr. Joshi: Absolutely.
What we did was we excluded the first 24 h postoperative where we do realize that S-100 could be elevated because of the mediastinal release. So that particular 24 h was actually excluded from the analysis.
The analysis was done starting postoperative day 1 comparing F1.2 to S-100. And the purpose was to look at the data and know the particular reasons that F1.2 was elevated and caused thromboembolic events even though that the anticoagulation protocol was followed. We are currently doing a study, where we would like to compare the change in the INR variability.
Dr. M. Pasic (Berlin, Germany): In a patient with a clinically evident infection, such as wound infection or caval infection, do you find any correlation or influence on your results?
Dr. Joshi: This was not really accounted for now into the analysis. The only thing what we wanted to know is like what frequency of the infections were available in the data. So in the ongoing study, we're looking at all possible variables which were not accounted for at this particular period of time for this analysis.
|
References
|
|---|
- Williams MR, Oz MC. Indications and patient selection for mechanical ventricular assistance. Am Thorac Surg 2001; 71:S86–S91. discussion S114–S115.[CrossRef][Medline]
- Fries D, Innerhofer P, Streif W, Schobersberger W, Margreiter J, Antretter H, Hormann H. Coagulation monitoring and management of anticoagulation during cardiac assist device support. Ann Thorac Surg 2003; 76:1593–1597.[Abstract/Free Full Text]
- Navia JL, McCarthy PM, Hoercher KJ, Feng J, Jeevantham R, Smedira NG, Banbury MK, Blackstone EH. Do left ventricular assist device (LVAD) bridge-to-transplantation outcomes predict the results of permanent LVAD implantation? Ann Thorac Surg 2002; 74:2051–2062. discussion 2062–2063.[Abstract/Free Full Text]
- Schmid C, Weyand M, Nabavi DG, Hammel D, Deng MC, Ringelstein EB, Scheld HH. Cerebral and systemic embolization during left ventricular support with the Novacor N100 Device. Ann Thorac Surg 1998; 65:1703–1710.[Abstract/Free Full Text]
- Matsubayashi H, Fastenau DR, McIntyre JA. Changes in platelet activation associated with left ventricular assist system placement. J Heart Lung Transplant 2000; 19:462–468.[CrossRef][Medline]
- Wilhelm CR, Ristich J, Knepper LE, Holubkov R, Wisniewski SR, Kormos RL, Wagner WR. Measurement of hemostatic indexes in conjugation with transcranial Doppler sonography in patients with ventricular assist devices. Stroke 1999; 30:2554–2561.[Abstract/Free Full Text]
- Brister SJ, Ofosu FA, Buchanan MR. Thrombin generation during cardiac surgery: is heparin the ideal anticoagulant? Thromb Haemost 1993; 70:259–262.[Medline]
- Craft RM, Chavez JJ, Bresee SJ, Wortham DC, Cohen E, Carroll RC. A novel modification of the Thrombelastograph assay, isolating platelet function, correlates with optical platelet aggregation. J Lab Clin Med 2004; 143:301–309.[CrossRef][Medline]
- Poston RS, White C, Gu J, Read K, Gammie J, Brown J, Lee A, Griffith B, Pierson R III. Endothelial injury and acquired aspirin resistance promote regional thrombin formation and early vein graft failure after coronary artery bypass. JTCVS 2006; 131:122–130.
- Haude M, Hafner G, Jablonka A, Rupprecht HJ, Prellwitz W, Meyer J, Erbel R. Guidance of anticoagulation after intracoronary implantation of Palmaz-Schatz stents by monitoring prothrombin and prothrombin fragment 1+2. Am Heart J 1995; 130:228–238.[CrossRef][Medline]
- Koster A, Loebe M, Sodian R, Potapov EV, Hansen R, Muller J, Mertzlufft F, Crystal GJ, Kuppe H, Hetzer R. Heparin antibodies and thromboembolism in heparin-coated and noncoated ventricular assist devices. J Thorac Cardiovasc Surg 2001; 121:331–335.[CrossRef][Medline]
- Craft RM, Chavez JJ, Bresee SJ, Wortham DC, Cohen E, Carroll RC. A novel modification of the Thrombelastograph assay, isolating platelet function, correlates with optical platelet aggregation. J Lab Clin Med 2004; 143:301–309.[CrossRef][Medline]
- Hall RI. Serum S-100ß protein and postoperative neurological dysfunction – ready for prime time? Can J Anesth 2004; 51:645–648.[Free Full Text]
- Potapov EV, Loebe M, Abdul-Khaliq H, Koster A, Stein J, Sodian R, Kopitz M, Hausmann H, Noon GP, DeBakey ME, Hetzer R. Postoperative course of S-100ß protein and neuron-specific enolase in patients after implantation of continuous and pulsatile flow LVADs. J Heart Lung Transplant 2001; 20:1310–1316.[CrossRef][Medline]
- Gurbel PA, Bliden KP, Guyer K, Cho PW, Zaman KA, Kreutz RP, Bassi AK, Tantry US. Platelet reactivity in patients and recurrent events post-stenting: results of the PREPARE POST-STENTING Study. J Am Coll Cardiol 2005; 46:1820–1826.[Abstract/Free Full Text]
|
|
Top
Abstract
1. Introduction
