- Email: [email protected]
heparin antibodies before cardiac surgery
Surgery
Incidence and clinical relevance of anti–platelet factor 4/heparin antibodies before cardiac surgery Sixten Selleng, MD, a Birgit Malowsky, MD, b Till Itterman, BSc, c Jessica Bagemu ˙˙hl, a Antje Wessel, a b d,e Hans-Georg Wollert, MD, Theodore E. Warkentin, MD, and Andreas Greinacher, MD a Greifswald and Karlsburg, Germany; and Hamilton, Ontario, Canada
Background
Heparin-induced thrombocytopenia (HIT) is caused by anti–platelet factor 4/heparin (PF4/H) immunoglobulin (Ig) G antibodies, which activate platelets. In some patients, anti-PF4/H antibodies are already detectable before cardiac surgery. Whether preoperative presence of antibodies confers adverse prognosis and which particular antibody classes (IgG, IgA, IgM) might be implicated are unknown.
Methods
We prospectively screened 591 patients undergoing cardiopulmonary bypass surgery for heparin-dependent antibodies by PF4/H immunoassay (separately for IgG, IgA, and IgM) and platelet activation test at preoperative baseline and at days 6 and 10. All patients received heparin or low–molecular-weight heparin postsurgery regardless of antibody status and were followed for postoperative complications, frequency of HIT, length of hospital stay, and 30-day mortality.
Results Anti-PF4/H antibodies of any class were detected at preoperative baseline in 128 (21.7%) of 591 patients: IgG n = 44 (7.4%), IgA n = 36 (6.1%), and IgM n = 79 (13.4%); some patients had N1 antibody class. Neither IgG nor IgA was a risk factor for any adverse outcome parameter. However, preoperative presence of IgM antibodies was associated with an increased risk for nonthromboembolic complications (all complications combined: hazard ratio 1.73, 95% CI 1.15-2.61) and a longer in-hospital stay (P = .02), but without evidence for increased risk of thrombotic complications or subsequent HIT. Conclusions Patients with preoperative anti-PF4/H antibodies of IgG and IgA class are not at increased risk for thrombotic or nonthrombotic adverse events, whereas those with baseline anti-PF4/H IgM had an increased risk of nonthrombotic adverse outcomes but not of subsequent HIT or thrombosis. Because IgM antibodies do not cause HIT, they could represent a surrogate marker for other heparin-independent risk factors. (Am Heart J 2010;160:362-9.)
Heparin-induced thrombocytopenia (HIT) is a serious adverse effect of heparin treatment that is usually caused by platelet-activating antibodies that recognize multimolecular complexes of platelet factor 4 bound to heparin (PF4/H). Several studies have shown that, in cardiac surgery patients, these antibodies are present in 4% to 20% of patients before surgery.1-3 It is generally assumed that preceding treatment with heparin is the main explanation for the presence of these antibodies,
From the aInstitut für Immunologie und Transfusionsmedizin, Ernst-Moritz-Arndt Universität, Greifswald, Germany, bKlinik für Herz-, Thorax- und Gefäßchirurgie, Herz- und Diabeteszentrum Mecklenburg/Vorpommern, Karlsburg, Germany, cInstitut für Community Medicine, Ernst-Moritz-Arndt Universität, Greifswald, Germany, dDepartment of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, Canada, and e Department of Medicine, McMaster University, Hamilton, Ontario, Canada. Submitted December 14, 2009; accepted May 11, 2010. Reprint requests: Sixten Selleng, MD, Institut für Immunologie und Transfusionsmedizin, Ernst-Moritz-Arndt Universität, Sauerbruchstraße, D-17475 Greifswald, Germany. E-mail: [email protected] 0002-8703/$ - see front matter © 2010, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.05.026
although none of the studies has examined whether antibody presence is associated with previous heparin exposure, particularly in the recent past. Furthermore, any pathobiological role of these antibodies is controversial, especially whether their presence predicts adverse outcomes. Whereas some authors found no correlation between the preoperative presence of these antibodies and the rate of postoperative thromboembolic events,3,4 others identified patients with preexisting anti-PF4/H antibodies to be at risk for a higher incidence of thromboses,5 increased length of hospital stay,6 and other adverse events.2,6 There is growing consensus that anti-PF4/H antibodies of the immunoglobulin (Ig) G class, which have the potential to activate platelets through their Fcγ receptors, are the primary cause of HIT,7,8 whereas the non–plateletactivating IgM and IgA class antibodies are of minor or no pathophysiologic relevance. All but one4 of the previously reported studies on the role of anti-PF4/H antibodies before cardiac surgery, however, did not differentiate between these antibody classes. Therefore, it remains uncertain whether the antibodies found to be predictive
American Heart Journal Volume 160, Number 2
for adverse outcomes are related to those that might cause HIT (ie, anti-PF4/H IgG antibodies showing platelet activation in vitro) or other antibodies (IgM, IgA, non– platelet-activating IgG) that might represent a surrogate marker for adverse prognosis because of a common factor such as inflammation, which has been shown to be associated with formation of anti-PF4/H antibodies.9 This is a crucial distinction because only the former pathophysiologic explanation infers the potential for improved clinical outcomes through periprocedural substitution of heparin with alternative nonheparin anticoagulants. In this study, therefore, we systematically screened almost 600 patients undergoing cardiac surgery at one institution10 for preoperative presence of anti-PF4/H antibodies of the individual immunoglobulin classes (IgG, IgM, IgA) and platelet-activating antibodies to assess their impact on postoperative adverse events.
Methods Study design This prospective, observational, single-center study was conducted at the Heart Center Mecklenburg/Westpomerania, Karlsburg, between January 1 and December 31, 2007, and included all consecutive patients who fulfilled the following criteria: (a) adult patients (≥18 years old), (b) patients scheduled for cardiac surgery requiring cardiopulmonary bypass, and (c) obtaining patient written informed consent. There were no additional exclusion criteria. In this study, we determined the number of patients testing positive for anti-PF4/ H antibodies after cardiac surgery and how the antibody status of patients correlates with platelet count patterns and HIT. The results of this analysis have been published previously.10 We now report the prespecified secondary analysis. The primary end point of this secondary analysis was to test whether preoperative presence of anti-PF4/H antibodies (combined for IgG, IgA, IgM) detected by an antigen assay is a risk factor for HIT or thrombotic complications, nonthrombotic complications, or prolonged in-hospital stay. The secondary end points of this analysis were to test whether any single class of anti-PF4/H antibodies (IgG, IgM, or IgA) detected by an antigen assay before surgery is a risk factor for HIT or thrombotic complications, nonthrombotic complications, or prolonged in-hospital stay. Power analysis was based on a previous study showing a significant effect of a positive precardiac surgery anti-PF4/H antibody status on the risk of postsurgery complications that enrolled 466 patients of which 13% had been antibody positive.2 The treating physician and the on-site study personnel were unaware of the results of antibody tests, which were performed in batches following patient discharge and follow-up assessments. In addition, the laboratory personnel were blinded to the clinical course of the patient. All patients received heparin for cardiopulmonary bypass (CPB), and all patients received either unfractionated heparin (UFH) or low–molecular-weight heparin for postsurgery anticoagulation. However, when HIT was clinically suspected, an independent sample was sent to the routine laboratory for HIT antibody testing as per usual clinical routine.
Selleng et al 363
The study was approved by the ethics committee of the ErnstMoritz-Arndt University, Greifswald; was supported by the German Federal Ministry for Education and Research (CAN04/ 006; ZIK-HIKE 03Z2CK1) and the “Forschungsverbund Molekulare Medizin” of the Ernst-Moritz-Arndt University Greifswald (FOMM 2007-06); and was done within the Department of Cardiovascular Medicine at the Medical Faculty, Ernst-MoritzArndt-University, Greifswald. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Data collection Data obtained included age, sex, height, weight, type (UFH and/or low–molecular-weight heparin) and duration of preoperative heparin treatment, underlying cardiac disease, type of surgery, the preoperative risk EURO score,11 duration of extracorporeal circulation, daily platelet counts up to day 10 (for analyzing platelet count patterns as described by Pouplard12), and the patient's outcome. Outcome parameters included 30-day mortality, length of hospital stay of the survivors (before discharge to rehabilitation center), and postsurgery complications. Postsurgery complications assessed comprised symptomatic thromboembolic complications, namely, stroke, myocardial infarction, peripheral arterial thrombosis or embolism, and venous thrombosis (deep vein thrombosis and/or pulmonary embolism), as well as nonthromboembolic complications including postoperative infections, hemorrhagic or cardiogenic shock, psychosis, respiratory failure, renal failure, and any other complications judged to have resulted in prolonged hospitalization or death. All clinically suspected thromboembolic complications during the 30-day follow-up period were objectively confirmed or excluded by appropriate imaging methods. All outcome parameters during hospital treatment were identified through review of the patient chart, with 30-day follow-up status of outpatients ascertained through standardized telephone interview with the treating physician at the rehabilitation center including the following questions: “Did the patient show any thromboembolic complication; specifically, were there any symptoms for venous and/or arterial thromboses, myocardial infarction, or stroke, or was limb amputation necessary? Were there any nonthromboembolic complications, including any requiring referral to an acute care hospital?” When any of these questions were answered in the affirmative, further details including the date and the hospital at which this complication was treated were obtained; and further information was gathered.
Laboratory assays Sera were collected at day 0 (=immediately before surgery) and at day 6 and day 10 post–cardiac surgery. Sera were stored at −30°C until testing. After the patient had been discharged from hospital, all sera obtained from one patient were thawed and tested for PF4/H antibodies on the same day. Sera were screened by an in-house anti-PF4/H enzyme immunoassay (EIA) separately for IgG, IgM, and IgA, with a cutoff of 0.5 optical density (OD) units, as described13 and by the heparin-induced platelet activation assay (HIPA).14 The correlation of our in-house EIA and a commercial PF4/H antibody assay has been shown to be high in cardiac surgery
American Heart Journal August 2010
364 Selleng et al
patients.15 We did not use the high heparin maneuver16 in the EIA to subclassify positive-reacting sera. The HIPA test is a washed platelet activation assay with very similar specificity/ sensitivity characteristics as the platelet serotonin-release assay17 and was considered positive if N2 of 4 donor platelets showed platelet activation in the presence of low (0.2 IU/mL), but not high (100 IU/mL), heparin concentrations, with a lag time of b30 minutes. Diagnostic sensitivity of combined use of the EIA and the HIPA for HIT is N95%.7,13
Figure 1
Definitions and diagnosis of HIT Typical HIT was considered definite when all 3 of the following criteria were met: (1) a platelet count fall by N50% (from the postoperative peak platelet count using the platelet counts obtained daily after surgery) and/or a thrombotic complication occurring between postoperative days 5 and 10 (inclusive; day of cardiac surgery = day 0),18 (2) a positive antiPF4/H IgG test result, and (3) a positive HIPA result.
Statistical analysis Demographic, clinical, and laboratory data of patients testing either positive or negative for anti-PF4/H antibodies by EIA were compared. Univariate comparison of data was performed using the Wilcoxon rank sum test with continuous and Fisher exact test with categorical variables. Cox regression hazard analysis was performed to determine the impact of antibody status before surgery on patient's outcome. For these analyses, we examined all of the 3 antibody classes (IgG, IgA, and IgM) individually as well as combined, in which patients who tested positive (OD ≥0.5) by anti-PF4/H EIA of any antibody class were compared with patients who tested antibody negative (OD b0.5). Outcome parameters involved incidence of thromboembolic and/or other (nonthromboembolic) complications and 30-day mortality. The correlation between preoperative occurrence of anti-PF4/H antibodies (EIA OD ≥0.5) and length of hospital stay of patients who survived until day 30 was calculated using linear logistic regression. The model was adjusted for age, sex, and preoperative EURO score.11 Furthermore, we also correlated the EIA level before surgery as a continuous variable (estimated by the EIA OD) to the risk for thromboembolic and/or other complications, and 30-day mortality. The model was adjusted for age, sex, and preoperative EURO score.11 All statistical analyses were performed using SAS 9.1 (SAS Institute Inc, Cary, NC). A P value ≤ .05 was considered statistically significant.
Results Patients, preoperative anti-PF4/H antibodies, and postoperative outcome A total of 591 patients were enrolled into the study. Of those, one hundred twenty-eight patients (128/591 = 21.7%) tested positive by EIA for any antibody class; and 6 patients (6/591 = 1.0%) tested positive in the HIPA. Comparing patients testing positive for anti-PF4/H antibodies of any class (IgG, IgA, and IgM; n = 128) before surgery with those testing negative (primary end point),
This Venn diagram provides numbers and percentages of patients testing positive for multiple anti-PF4/H antibody classes.
there was no difference in the risk for thromboembolic complications (5/128 = 3.9% vs 25/463 = 5.4%, hazard ratio [HR] = 0.72, 95% confidence interval (CI) = 0.281.89) and nonthromboembolic complications (38/128 = 29.7% vs 115/463 = 24.8%, HR = 1.29, 95% CI = 0.901.87), no longer in-hospital length of stay (15.5 vs 15.1 days, P = .70), and no higher 30-day mortality (3/128 = 2.3% vs 17/463 = 3.7%, HR = 0.64, 95% CI = 0.18-2.22). When we analyzed the outcomes separately for the single immunoglobulin classes, 44 patients (44/591 = 7.4%) tested positive in the day 0 (preoperative) blood sample for anti-PF4/H IgG (OD N0.5), 36 patients (36/ 591 = 6.1%) for IgA, and 79 patients (79/591 = 13.4%) for IgM. Results for multiple antibody classes are given in Figure 1. At 30-day follow-up, patients testing positive for antiPF4/H IgG before surgery had, compared with those testing negative, neither a significantly increased risk for thromboembolic complications (3/44 = 6.8% vs 27/547 = 4.9%, HR = 1.39, 95% CI = 0.42-4.60) or for other postoperative complications (11/44 = 25.0% vs 142/547 = 26.0%, HR = 0.90, 95% CI = 0.49-1.64) nor a prolonged inhospital length of stay (15.9 vs 15.2 days, P = .95). However, there was a slightly, though not significantly, higher 30-day mortality rate (3/44 = 6.8% vs 17/547 = 3.1%, HR = 1.49, 95% CI = 0.41-5.44). The 3 deaths among the patients who tested IgG positive at preoperative baseline were caused by cardiogenic shock (day 4 and day 6 in 2 patients) and by hemorrhagic shock (day 15 in the third one) and did not appear to be related to rapidonset HIT. None of the 7 patients showing PF4/H IgG OD N1.0 before surgery had an adverse outcome or a prolonged in-hospital stay. A detailed comparison of
American Heart Journal Volume 160, Number 2
Selleng et al 365
Table I. Comparison of demographic and clinical characteristics between patients testing positive and negative, respectively, for anti-PF4/H IgG before cardiac surgery Anti-PF4/H IgG negative n = 547 Characteristics Male Age (y) Height (cm) Weight (kg) EURO score (additive) EURO score (logistic) Recent heparin treatment Duration of recent heparin treatment (if performed; d) Type of surgery Isolated CABG Valve repair (with or without CABG) Other⁎ Duration of extracorporeal circulation (min)
n 397
Anti-PF4/H IgG positive n = 44
%
n
(72.58)
32
68.06 170.34 84.24 5.45 6.09 292
(72.73) 67.42 168.41 81.78 6.09 9.21
(53.38)
28
9.08
317 161 69
%
(63.64) 14.75
(57.95) (29.43) (12.61)
18 18 8
92.80
(40.91) (40.91) (18.18) 98.05
P value 1.00 .66 .18 .25 .23 .23 .21 .01
.02 .13 .35 .34
CABG, Coronary artery bypass grafting. ⁎ Other combined major cardiac procedure and surgery on thoracic aorta.
Table II. Results of Cox regression hazard analysis assessing the impact of presurgery anti-PF4/H antibody EIA OD results as continuous variables on the risk for post–cardiac surgery adverse events until postoperative day 30
Anti-PF4/H IgG Anti-PF4/H IgA Anti-PF4/H IgM
Thromboembolic events
Adverse events (other than thromboembolic)
30-d mortality
0.54 (0.06; 4.66) 0.48 (0.04; 5.26) 0.64 (0.12; 3.41)
0.66 (0.26; 1.65) 0.68 (0.25; 1.85) 1.94 (1.12; 3.36)⁎
2.09 (0.29; 14.82) 1.49 (0.19; 11.90) 0.17 (0.01; 2.99)
Data are expressed as hazard ratio (95% CI); the model was adjusted for age, sex, and EURO score. ⁎ P b .05.
demographic and clinical characteristics between patients testing positive and negative, respectively, for anti-PF4/H IgG at preoperative baseline is given in Table I. Patients testing positive for anti-PF4/H IgA before surgery had no higher risk for thromboembolic complications (0/36 = 0.0% vs 30/555 = 5.4%, HR = 0.03, 95% CI = 0.00-2.27) or other postoperative complications (8/36 = 22.2% vs 145/555 = 26.1%, HR = 0.76, 95% CI = 0.361.64), no significantly longer in-hospital length of stay (15.3 vs 14.7 days, P = .84), and no higher 30-day mortality (1/36 = 2.3% vs 19/555 = 3.4%, HR = 0.91, 95% CI = 0.12-6.79). At 30-day follow-up, patients testing positive for antiPF4/H IgM before surgery had no higher risk for thromboembolic complications (3/79 = 3.8% vs 27/512 = 5.3%, HR = 0.76, 95% CI = 0.23-2.50). However, the risk for nonthromboembolic complications was increased (30/79 = 38.0% vs 123/512 = 24.0%, HR = 1.73, 95% CI = 1.15-2.61). The nonthromboembolic complications were a heterogeneous group including neurologic complications (7/79 = 8.9% vs 21/512 = 4.1%), postoperative infectious complications (5/79 = 6.3% vs 17/512 = 3.3%),
and respiratory failure (6/79 = 7.6% vs 19/512 = 3.76%). These patients had a significantly longer in-hospital length of stay (16.4 vs 15.1 days, P = .021). However, 30-day mortality was not increased (0/79 = 0.0% vs 20/512 = 3.9%, P = .09). To test for a potential effect of PF4/H antibody strength (according to EIA OD), we included anti-PF4/H EIA ODs into a Cox logistic regression model as a continuous variable. Immunoglobulin G or IgA antibodies were again not associated with an increased risk for thromboembolic or other postoperative complications, or a higher 30-day mortality, whereas IgM antibodies were again associated with an increased risk for nonthromboembolic complications (Table II).
Preoperative anti-PF4/H antibodies and HIT Of the 6 patients who tested positive in the HIPA test at baseline, 5 had an uneventful course postsurgery. But one of these patients presented with early and profound thrombocytopenia that would be consistent with “rapidonset” HIT. However, apart from thrombocytopenia, the
366 Selleng et al
American Heart Journal August 2010
Figure 2
Anti-PF4/H antibody levels, estimated by EIA OD units, are presented separately for IgG (A), IgA (B), and IgM (C) as median and 25% and 75% quartiles. Irrespective of antibody class, at all time points—day 0 = day of surgery and day 6 and day 10 postsurgery—anti-PF4/H antibody levels of patients testing positive already before surgery (black bars) were statistically significantly higher compared with those of the patients testing negative at baseline (white bars).
clinical course was uneventful; and there was no evidence of any thromboembolic or other complications. In this patient, thrombocytopenia prompted a test for anti-PF4/H antibodies independently from the study samples during the first postoperative week to exclude
HIT for safety reasons. When the result for this test was reported to be positive, anticoagulation was switched from UFH to hirudin at postoperative day 4; but platelet counts did not recover subsequently until day 10. This suggests that, in this patient, early-onset and persisting
American Heart Journal Volume 160, Number 2
thrombocytopenia may have been due to non-HIT factors, as we have reported in detail elsewhere.10 As also reported elsewhere,10 in the entire study, only 3 patients developed “typical-onset HIT”, that is, a platelet count fall by N50% between postoperative days 5 and 10, with detection of platelet-activating (HIPApositive) anti-PF4/H IgG antibodies (EIA ODs 2.064, 1.567, 1.013). It is noteworthy that all 3 patients tested negative for anti-PF4/H IgG, IgA, and IgM, as well as in the HIPA, before surgery.
Correlation between pre- and postoperative anti-PF4/H antibodies Anti-PF4/H IgG levels (estimated by mean EIA OD) in patients testing positive for IgG before surgery (n = 44) was 0.792 ± 0.255 at day 0 and increased to 0.859 ± 0.345 at day 6 and to 1.062 ± 0.369 at day 10. Patients who tested negative before surgery also showed a progressive increase in antibody levels between days 0 and 10 (day 0: 0.218 ± 0.107, day 6: 0.402 ± 0.288, day 10: 0.619 ± 0.372). Antibody levels were significantly higher at all 3 time points in the patients testing positive at baseline (P b .001; Figure 2, A provides the medians). A similar pattern was observed for anti-PF4/H IgA and IgM antibodies (Figure 2, B and C provide the medians).
Discussion This prospective study confirms a relatively high frequency of anti-PF4/H antibodies (IgG, IgA, and/or IgM) in patients before cardiac surgery.1–3,5 We found these antibodies in N20% of patients. All patients irrespective of their preoperative antibody status received heparin during and after cardiac surgery; but unlike some previous studies enrolling similar numbers of patients, we did not find an association between the presence of PF4/H antibodies of any class (ie, combined IgG, IgA, and IgM) with adverse outcomes. Furthermore, when we analyzed the different immunoglobulin classes separately, we did not find any evidence for an increased risk of adverse outcomes in patients testing positive at preoperative baseline for anti-PF4/H IgG or IgA antibodies. However, presence of anti-PF4/H IgM antibodies before cardiac surgery was associated with an increased risk for the combined incidence of nonthrombotic adverse outcomes and also for an increased length of stay. Kress et al6 also found in a 1,209-patient study that anti-PF4/H antibody–positive patients had longer mean postoperative in-hospital length of stay and higher incidences of prolonged mechanical ventilation. However, they also reported an increased rate for acute limb ischemia and renal and gastrointestinal complications. Bennett-Guerrero et al2 identified preoperative anti-PF4/ H antibodies as an independent predictor for postoperative complications (death or prolonged hospital stay) in a 466-patient study. Probably the strongest association of
Selleng et al 367
anti-PF4/H antibodies and adverse outcomes has been reported by Mattioli et al.5 In a 500-patient study, using the McNemar χ2 test, they found that patients with preexisting anti-PF4/H antibodies before cardiac surgery have a highly significantly increased risk for death (10/ 131 vs 14/369, P = .001), pulmonary embolism (10/131 vs 14/369, P = .001), myocardial infarction (9/131 vs 15/ 369, P = .01), and stroke (4/131 vs 7/369, P = .01) at day 30 (however, applying categorical analyses using the Fisher exact test, these P values are all N .05). They also found PF4/H antibody–positive patients to be at risk for a higher incidence of thrombotic events over time (ie, 36 months).5 Three other studies (enrolling 111,1 299,3 and 368 patients4), however, found no correlation between antibody levels before cardiac surgery and the rate of postoperative thromboembolic complications. Our study is only the second (along with that of Gluckman et al4) to have analyzed patient sera separately for anti-PF4/H IgG, IgM, and IgA classes and for presence of platelet-activating antibodies. The finding that anti-PF4/ H antibodies of IgM class were (weakly) predictive for adverse outcomes is surprising. However, these IgM antibodies were not a risk factor for later seroconversion to functionally reactive IgG antibodies and HIT. Even among the 3 patients who developed typical-onset HIT during this study, none had a positive preoperative test result for anti-PF4/H antibodies (including for IgM), thus indicating that they seroconverted from a negative baseline antibody status. Several studies have shown that anti-PF4/H IgM antibodies (as well as IgA antibodies) are not capable to engage and to cross-link the platelet FcγIIa receptors on platelets,19,20 which are considered the key mechanisms underlying platelet activation in HIT.21 As PF4/H IgM antibodies do not cause the HIT syndrome, which requires intravascular activation of platelets via the Fcγ receptor, it seems unlikely that the nonthromboembolic complications associated with the presence of IgM antibodies could be prevented by avoiding intra- and postoperative heparin. Furthermore, as the use of nonheparin anticoagulation for cardiac surgery patients is not well established, its wider use could cause increased bleeding.22 Indeed, given the relatively high number of patients with precardiac surgery PF4/H antibodies (≈15%-20%), deliberate use of alternative anticoagulants because of preoperative detection of antibodies might itself result in significant complications (although promising results regarding the use of bivalirudin on CPB have been recently published).23,24 Further studies evaluating the risk of bleeding vs thrombosis when using nonheparin anticoagulation on CPB are urgently needed. Considering the findings of our and other similar studies,2-4 anti-PF4/H antibody testing before cardiac surgery remains indicated only if a patient belongs to the 2 groups in whom these antibodies have been shown to be clinically relevant for peri- and postsurgical
American Heart Journal August 2010
368 Selleng et al
management.22 The first group consists of patients with typical clinical signs of HIT, for example, a new-onset platelet count fall bearing a temporal relationship with heparin administered preoperatively for acute coronary syndrome while waiting for urgent coronary bypass surgery. In these patients, detection of platelet-activating anti-PF4/H IgG antibodies likely requires nonheparin anticoagulation for CPB.22 The second group represents patients with a history of HIT, typically within the previous 4 to 12 weeks, who now require urgent cardiac surgery. In these patients, the absence of plateletactivating anti-PF4/H antibodies allows the use of UFH during CPB.25,26 Our study has several limitations. Only a minority of patients (ie, ≈20%) had preexisting anti-PF4/H antibodies, thus limiting the number of patients at risk. Furthermore, certain outcome parameters like thromboembolic events and post–cardiac surgery death occurred in only 5.1% and 3.4%, respectively. Finally, the observation period of 30 days is not suitable to detect any long-term effects. A definite answer of a potential minor association of anti-PF4/H antibodies before cardiac surgery with adverse outcomes will require a large prospective trial enrolling several thousand patients ideally with a longer observation period than 30 days. The association of anti-PF4/H antibodies with nonthrombotic adverse events might indicate that anti-PF4/H IgM antibodies could represent a surrogate marker for other risk factors. However, another possibility is that we may not have included all potential confounders for these nonthrombotic complications in our Cox regression analysis. All analyses were adjusted for age, sex, and EURO score. The EURO score, however, was developed to predict mortality but not to assess the risk for prolonged length of hospital stay or other complications. Finally, the association between antibody presence and length of hospital stay excluded some patients with a poor outcome, namely, patients who died before day 30. Therefore, the potential association of anti-PF4/H IgM antibodies with adverse outcomes needs to be interpreted with caution. In summary, anti-PF4/H antibodies (of any immunoglobulin class) are present before cardiac surgery in approximately 20% of patients. The presence of such antibodies was associated with a low risk (at most, 1/ 128 or b1%) of rapid-onset HIT during the early postoperative course and was not associated with an increased risk of typical-onset HIT. Our study does not confirm earlier reports on a relatively strong association of these antibodies with the risk for HIT, new thrombosis, or mortality. However, we did detect an association between increased risk of nonthromboembolic morbidity and increased length of hospital stay among those patients who had detectable anti-PF4/H antibodies of IgM class. Inasmuch as this class of
antibodies is not associated with HIT, our study suggests that IgM antibodies could represent a surrogate marker for risk of non–HIT-related adverse events (eg, through a common factor such as pre/ perioperative inflammation). Because it seems unlikely that nonheparin anticoagulation would avoid these complications, heparin should remain as the anticoagulant of choice during CPB in otherwise asymptomatic patients who test positive for anti-PF4/H antibodies in a clinical context that does not suggest acute HIT. Our findings are also consistent with the recommendations of the American College of Chest Physicians Consensus Conference guidelines (eighth edition)22 that routine testing for anti-PF4/H antibodies before cardiac surgery should be avoided.
Acknowledgements We thank the staff of the Heart Center Mecklenburg/ Westpommerania, Karlsburg, for collecting the blood samples.
Disclosures The authors state that they have no conflict of interest.
References 1. Bauer TL, Arepally G, Konkle BA, et al. Prevalence of heparin-associated antibodies without thrombosis in patients undergoing cardiopulmonary bypass surgery. Circulation 1997; 95:1242-6. 2. Bennett-Guerrero E, Slaughter TF, White WD, et al. Preoperative antiPF4/heparin antibody level predicts adverse outcome after cardiac surgery. J Thorac Cardiovasc Surg 2005;130:1567-72. 3. Everett BM, Yeh R, Foo SY, et al. Prevalence of heparin/platelet factor 4 antibodies before and after cardiac surgery. Ann Thorac Surg 2007;83:592-7. 4. Gluckman TJ, Segal JB, Schulman SP, et al. Effect of anti–platelet factor-4/heparin antibody induction on early saphenous vein graft occlusion after coronary artery bypass surgery. J Thromb Haemost 2009;7:1457-64. 5. Mattioli AV, Bonetti L, Zennaro M, et al. Heparin/PF4 antibodies formation after heparin treatment: temporal aspects and long-term follow-up. Am Heart J 2009;157:589-95. 6. Kress DC, Aronson S, McDonald ML, et al. Positive heparin–platelet factor 4 antibody complex and cardiac surgical outcomes. Ann Thorac Surg 2007;83:1737-43. 7. Greinacher A, Juhl D, Strobel U, et al. Heparin-induced thrombocytopenia: a prospective study on the incidence, platelet-activating capacity and clinical significance of antiplatelet factor 4/heparin antibodies of the IgG, IgM, and IgA classes. J Thromb Haemost 2007; 5:1666-73. 8. Warkentin TE, Sheppard JA, Moore JC, et al. Laboratory testing for the antibodies that cause heparin-induced thrombocytopenia: how much class do we need? J Lab Clin Med 2005;146:341-6. 9. Paparella D, Scrascia G, Galeone A, et al. Formation of anti–platelet factor 4/heparin antibodies after cardiac surgery: influence of perioperative platelet activation, the inflammatory response, and
American Heart Journal Volume 160, Number 2
10.
11.
12.
13.
14.
15.
16.
17.
histocompatibility leukocyte antigen status. J Thorac Cardiovasc Surg 2008;136:1456-63. Selleng S, Malowsky B, Strobel U, et al. Early-onset and persisting thrombocytopenia in post-cardiac surgery patients is rarely due to heparin-induced thrombocytopenia even when antibody tests are positive. J Thromb Haemost 2010;8:30-6. Nashef SA, Roques F, Michel P, et al. European system for cardiac operative risk evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9-13. Pouplard C, May MA, Regina S, et al. Changes in platelet count after cardiac surgery can effectively predict the development of pathogenic heparin-dependent antibodies. Br J Haematol 2005; 128:837-41. Juhl D, Eichler P, Lubenow N, et al. Incidence and clinical significance of anti-PF4/heparin antibodies of the IgG, IgM, and IgA class in 755 consecutive patient samples referred for diagnostic testing for heparininduced thrombocytopenia. Eur J Haematol 2006;76:420-6. Greinacher A, Michels I, Kiefel V, et al. A rapid and sensitive test for diagnosing heparin-associated thrombocytopenia. Thromb Haemost 1991;66:734-6. Schenk S, El Banayosy A, Morshuis M, et al. IgG classification of antiPF4/heparin antibodies to identify patients with heparin-induced thrombocytopenia during mechanical circulatory support. J Thromb Haemost 2007;5:235-41. Whitlatch NL, Perry SL, Ortel TL. Anti–heparin/platelet factor 4 antibody optical density values and the confirmatory procedure in the diagnosis of heparin-induced thrombocytopenia. Thromb Haemost 2008;100:678-84. Savi P, Chong BH, Greinacher A, et al. Effect of fondaparinux on platelet activation in the presence of heparin-dependent antibodies: a blinded comparative multicenter study with unfractionated heparin. Blood 2005;105:139-44.
Selleng et al 369
18. Warkentin TE, Kelton JG. Temporal aspects of heparin-induced thrombocytopenia. N Engl J Med 2001;344: 1286-92. 19. Denomme GA, Warkentin TE, Horsewood P, et al. Activation of platelets by sera containing IgG1 heparin-dependent antibodies: an explanation for the predominance of the Fc gammaRIIa “low responder” (his131) gene in patients with heparin-induced thrombocytopenia. J Lab Clin Med 1997;130:278-84. 20. Rauova L, Zhai L, Kowalska MA, et al. Role of platelet surface PF4 antigenic complexes in heparin-induced thrombocytopenia pathogenesis: diagnostic and therapeutic implications. Blood 2006;107:2346-53. 21. Kelton JG, Sheridan D, Santos A, et al. Heparin-induced thrombocytopenia: laboratory studies. Blood 1988;72:925-30. 22. Warkentin TE, Greinacher A, Koster A, et al. Treatment and prevention of heparin-induced thrombocytopenia: American College of Chest Physicians evidence-based clinical practice guidelines (8th edition). Chest 2008;133:340S-80S. 23. Dyke CM, Smedira NG, Koster A, et al. A comparison of bivalirudin to heparin with protamine reversal in patients undergoing cardiac surgery with cardiopulmonary bypass: the EVOLUTION-ON study. J Thorac Cardiovasc Surg 2006;131:533-9. 24. Koster A, Dyke CM, Aldea G, et al. Bivalirudin during cardiopulmonary bypass in patients with previous or acute heparin-induced thrombocytopenia and heparin antibodies: results of the CHOOSEON trial. Ann Thorac Surg 2007;83:572-7. 25. Pötzsch B, Klovekorn WP, Madlener K. Use of heparin during cardiopulmonary bypass in patients with a history of heparin-induced thrombocytopenia. N Engl J Med 2000;343:515. 26. Selleng S, Haneya A, Hirt S, et al. Management of anticoagulation in patients with subacute heparin-induced thrombocytopenia scheduled for heart transplantation. Blood 2008;112:4024-7.
Incidence and clinical relevance of anti–platelet factor 4/heparin antibodies before cardiac surgery Sixten Selleng, MD, a Birgit Malowsky, MD, b Till Itterman, BSc, c Jessica Bagemu ˙˙hl, a Antje Wessel, a b d,e Hans-Georg Wollert, MD, Theodore E. Warkentin, MD, and Andreas Greinacher, MD a Greifswald and Karlsburg, Germany; and Hamilton, Ontario, Canada
Background
Heparin-induced thrombocytopenia (HIT) is caused by anti–platelet factor 4/heparin (PF4/H) immunoglobulin (Ig) G antibodies, which activate platelets. In some patients, anti-PF4/H antibodies are already detectable before cardiac surgery. Whether preoperative presence of antibodies confers adverse prognosis and which particular antibody classes (IgG, IgA, IgM) might be implicated are unknown.
Methods
We prospectively screened 591 patients undergoing cardiopulmonary bypass surgery for heparin-dependent antibodies by PF4/H immunoassay (separately for IgG, IgA, and IgM) and platelet activation test at preoperative baseline and at days 6 and 10. All patients received heparin or low–molecular-weight heparin postsurgery regardless of antibody status and were followed for postoperative complications, frequency of HIT, length of hospital stay, and 30-day mortality.
Results Anti-PF4/H antibodies of any class were detected at preoperative baseline in 128 (21.7%) of 591 patients: IgG n = 44 (7.4%), IgA n = 36 (6.1%), and IgM n = 79 (13.4%); some patients had N1 antibody class. Neither IgG nor IgA was a risk factor for any adverse outcome parameter. However, preoperative presence of IgM antibodies was associated with an increased risk for nonthromboembolic complications (all complications combined: hazard ratio 1.73, 95% CI 1.15-2.61) and a longer in-hospital stay (P = .02), but without evidence for increased risk of thrombotic complications or subsequent HIT. Conclusions Patients with preoperative anti-PF4/H antibodies of IgG and IgA class are not at increased risk for thrombotic or nonthrombotic adverse events, whereas those with baseline anti-PF4/H IgM had an increased risk of nonthrombotic adverse outcomes but not of subsequent HIT or thrombosis. Because IgM antibodies do not cause HIT, they could represent a surrogate marker for other heparin-independent risk factors. (Am Heart J 2010;160:362-9.)
Heparin-induced thrombocytopenia (HIT) is a serious adverse effect of heparin treatment that is usually caused by platelet-activating antibodies that recognize multimolecular complexes of platelet factor 4 bound to heparin (PF4/H). Several studies have shown that, in cardiac surgery patients, these antibodies are present in 4% to 20% of patients before surgery.1-3 It is generally assumed that preceding treatment with heparin is the main explanation for the presence of these antibodies,
From the aInstitut für Immunologie und Transfusionsmedizin, Ernst-Moritz-Arndt Universität, Greifswald, Germany, bKlinik für Herz-, Thorax- und Gefäßchirurgie, Herz- und Diabeteszentrum Mecklenburg/Vorpommern, Karlsburg, Germany, cInstitut für Community Medicine, Ernst-Moritz-Arndt Universität, Greifswald, Germany, dDepartment of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, Canada, and e Department of Medicine, McMaster University, Hamilton, Ontario, Canada. Submitted December 14, 2009; accepted May 11, 2010. Reprint requests: Sixten Selleng, MD, Institut für Immunologie und Transfusionsmedizin, Ernst-Moritz-Arndt Universität, Sauerbruchstraße, D-17475 Greifswald, Germany. E-mail: [email protected] 0002-8703/$ - see front matter © 2010, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.05.026
although none of the studies has examined whether antibody presence is associated with previous heparin exposure, particularly in the recent past. Furthermore, any pathobiological role of these antibodies is controversial, especially whether their presence predicts adverse outcomes. Whereas some authors found no correlation between the preoperative presence of these antibodies and the rate of postoperative thromboembolic events,3,4 others identified patients with preexisting anti-PF4/H antibodies to be at risk for a higher incidence of thromboses,5 increased length of hospital stay,6 and other adverse events.2,6 There is growing consensus that anti-PF4/H antibodies of the immunoglobulin (Ig) G class, which have the potential to activate platelets through their Fcγ receptors, are the primary cause of HIT,7,8 whereas the non–plateletactivating IgM and IgA class antibodies are of minor or no pathophysiologic relevance. All but one4 of the previously reported studies on the role of anti-PF4/H antibodies before cardiac surgery, however, did not differentiate between these antibody classes. Therefore, it remains uncertain whether the antibodies found to be predictive
American Heart Journal Volume 160, Number 2
for adverse outcomes are related to those that might cause HIT (ie, anti-PF4/H IgG antibodies showing platelet activation in vitro) or other antibodies (IgM, IgA, non– platelet-activating IgG) that might represent a surrogate marker for adverse prognosis because of a common factor such as inflammation, which has been shown to be associated with formation of anti-PF4/H antibodies.9 This is a crucial distinction because only the former pathophysiologic explanation infers the potential for improved clinical outcomes through periprocedural substitution of heparin with alternative nonheparin anticoagulants. In this study, therefore, we systematically screened almost 600 patients undergoing cardiac surgery at one institution10 for preoperative presence of anti-PF4/H antibodies of the individual immunoglobulin classes (IgG, IgM, IgA) and platelet-activating antibodies to assess their impact on postoperative adverse events.
Methods Study design This prospective, observational, single-center study was conducted at the Heart Center Mecklenburg/Westpomerania, Karlsburg, between January 1 and December 31, 2007, and included all consecutive patients who fulfilled the following criteria: (a) adult patients (≥18 years old), (b) patients scheduled for cardiac surgery requiring cardiopulmonary bypass, and (c) obtaining patient written informed consent. There were no additional exclusion criteria. In this study, we determined the number of patients testing positive for anti-PF4/ H antibodies after cardiac surgery and how the antibody status of patients correlates with platelet count patterns and HIT. The results of this analysis have been published previously.10 We now report the prespecified secondary analysis. The primary end point of this secondary analysis was to test whether preoperative presence of anti-PF4/H antibodies (combined for IgG, IgA, IgM) detected by an antigen assay is a risk factor for HIT or thrombotic complications, nonthrombotic complications, or prolonged in-hospital stay. The secondary end points of this analysis were to test whether any single class of anti-PF4/H antibodies (IgG, IgM, or IgA) detected by an antigen assay before surgery is a risk factor for HIT or thrombotic complications, nonthrombotic complications, or prolonged in-hospital stay. Power analysis was based on a previous study showing a significant effect of a positive precardiac surgery anti-PF4/H antibody status on the risk of postsurgery complications that enrolled 466 patients of which 13% had been antibody positive.2 The treating physician and the on-site study personnel were unaware of the results of antibody tests, which were performed in batches following patient discharge and follow-up assessments. In addition, the laboratory personnel were blinded to the clinical course of the patient. All patients received heparin for cardiopulmonary bypass (CPB), and all patients received either unfractionated heparin (UFH) or low–molecular-weight heparin for postsurgery anticoagulation. However, when HIT was clinically suspected, an independent sample was sent to the routine laboratory for HIT antibody testing as per usual clinical routine.
Selleng et al 363
The study was approved by the ethics committee of the ErnstMoritz-Arndt University, Greifswald; was supported by the German Federal Ministry for Education and Research (CAN04/ 006; ZIK-HIKE 03Z2CK1) and the “Forschungsverbund Molekulare Medizin” of the Ernst-Moritz-Arndt University Greifswald (FOMM 2007-06); and was done within the Department of Cardiovascular Medicine at the Medical Faculty, Ernst-MoritzArndt-University, Greifswald. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Data collection Data obtained included age, sex, height, weight, type (UFH and/or low–molecular-weight heparin) and duration of preoperative heparin treatment, underlying cardiac disease, type of surgery, the preoperative risk EURO score,11 duration of extracorporeal circulation, daily platelet counts up to day 10 (for analyzing platelet count patterns as described by Pouplard12), and the patient's outcome. Outcome parameters included 30-day mortality, length of hospital stay of the survivors (before discharge to rehabilitation center), and postsurgery complications. Postsurgery complications assessed comprised symptomatic thromboembolic complications, namely, stroke, myocardial infarction, peripheral arterial thrombosis or embolism, and venous thrombosis (deep vein thrombosis and/or pulmonary embolism), as well as nonthromboembolic complications including postoperative infections, hemorrhagic or cardiogenic shock, psychosis, respiratory failure, renal failure, and any other complications judged to have resulted in prolonged hospitalization or death. All clinically suspected thromboembolic complications during the 30-day follow-up period were objectively confirmed or excluded by appropriate imaging methods. All outcome parameters during hospital treatment were identified through review of the patient chart, with 30-day follow-up status of outpatients ascertained through standardized telephone interview with the treating physician at the rehabilitation center including the following questions: “Did the patient show any thromboembolic complication; specifically, were there any symptoms for venous and/or arterial thromboses, myocardial infarction, or stroke, or was limb amputation necessary? Were there any nonthromboembolic complications, including any requiring referral to an acute care hospital?” When any of these questions were answered in the affirmative, further details including the date and the hospital at which this complication was treated were obtained; and further information was gathered.
Laboratory assays Sera were collected at day 0 (=immediately before surgery) and at day 6 and day 10 post–cardiac surgery. Sera were stored at −30°C until testing. After the patient had been discharged from hospital, all sera obtained from one patient were thawed and tested for PF4/H antibodies on the same day. Sera were screened by an in-house anti-PF4/H enzyme immunoassay (EIA) separately for IgG, IgM, and IgA, with a cutoff of 0.5 optical density (OD) units, as described13 and by the heparin-induced platelet activation assay (HIPA).14 The correlation of our in-house EIA and a commercial PF4/H antibody assay has been shown to be high in cardiac surgery
American Heart Journal August 2010
364 Selleng et al
patients.15 We did not use the high heparin maneuver16 in the EIA to subclassify positive-reacting sera. The HIPA test is a washed platelet activation assay with very similar specificity/ sensitivity characteristics as the platelet serotonin-release assay17 and was considered positive if N2 of 4 donor platelets showed platelet activation in the presence of low (0.2 IU/mL), but not high (100 IU/mL), heparin concentrations, with a lag time of b30 minutes. Diagnostic sensitivity of combined use of the EIA and the HIPA for HIT is N95%.7,13
Figure 1
Definitions and diagnosis of HIT Typical HIT was considered definite when all 3 of the following criteria were met: (1) a platelet count fall by N50% (from the postoperative peak platelet count using the platelet counts obtained daily after surgery) and/or a thrombotic complication occurring between postoperative days 5 and 10 (inclusive; day of cardiac surgery = day 0),18 (2) a positive antiPF4/H IgG test result, and (3) a positive HIPA result.
Statistical analysis Demographic, clinical, and laboratory data of patients testing either positive or negative for anti-PF4/H antibodies by EIA were compared. Univariate comparison of data was performed using the Wilcoxon rank sum test with continuous and Fisher exact test with categorical variables. Cox regression hazard analysis was performed to determine the impact of antibody status before surgery on patient's outcome. For these analyses, we examined all of the 3 antibody classes (IgG, IgA, and IgM) individually as well as combined, in which patients who tested positive (OD ≥0.5) by anti-PF4/H EIA of any antibody class were compared with patients who tested antibody negative (OD b0.5). Outcome parameters involved incidence of thromboembolic and/or other (nonthromboembolic) complications and 30-day mortality. The correlation between preoperative occurrence of anti-PF4/H antibodies (EIA OD ≥0.5) and length of hospital stay of patients who survived until day 30 was calculated using linear logistic regression. The model was adjusted for age, sex, and preoperative EURO score.11 Furthermore, we also correlated the EIA level before surgery as a continuous variable (estimated by the EIA OD) to the risk for thromboembolic and/or other complications, and 30-day mortality. The model was adjusted for age, sex, and preoperative EURO score.11 All statistical analyses were performed using SAS 9.1 (SAS Institute Inc, Cary, NC). A P value ≤ .05 was considered statistically significant.
Results Patients, preoperative anti-PF4/H antibodies, and postoperative outcome A total of 591 patients were enrolled into the study. Of those, one hundred twenty-eight patients (128/591 = 21.7%) tested positive by EIA for any antibody class; and 6 patients (6/591 = 1.0%) tested positive in the HIPA. Comparing patients testing positive for anti-PF4/H antibodies of any class (IgG, IgA, and IgM; n = 128) before surgery with those testing negative (primary end point),
This Venn diagram provides numbers and percentages of patients testing positive for multiple anti-PF4/H antibody classes.
there was no difference in the risk for thromboembolic complications (5/128 = 3.9% vs 25/463 = 5.4%, hazard ratio [HR] = 0.72, 95% confidence interval (CI) = 0.281.89) and nonthromboembolic complications (38/128 = 29.7% vs 115/463 = 24.8%, HR = 1.29, 95% CI = 0.901.87), no longer in-hospital length of stay (15.5 vs 15.1 days, P = .70), and no higher 30-day mortality (3/128 = 2.3% vs 17/463 = 3.7%, HR = 0.64, 95% CI = 0.18-2.22). When we analyzed the outcomes separately for the single immunoglobulin classes, 44 patients (44/591 = 7.4%) tested positive in the day 0 (preoperative) blood sample for anti-PF4/H IgG (OD N0.5), 36 patients (36/ 591 = 6.1%) for IgA, and 79 patients (79/591 = 13.4%) for IgM. Results for multiple antibody classes are given in Figure 1. At 30-day follow-up, patients testing positive for antiPF4/H IgG before surgery had, compared with those testing negative, neither a significantly increased risk for thromboembolic complications (3/44 = 6.8% vs 27/547 = 4.9%, HR = 1.39, 95% CI = 0.42-4.60) or for other postoperative complications (11/44 = 25.0% vs 142/547 = 26.0%, HR = 0.90, 95% CI = 0.49-1.64) nor a prolonged inhospital length of stay (15.9 vs 15.2 days, P = .95). However, there was a slightly, though not significantly, higher 30-day mortality rate (3/44 = 6.8% vs 17/547 = 3.1%, HR = 1.49, 95% CI = 0.41-5.44). The 3 deaths among the patients who tested IgG positive at preoperative baseline were caused by cardiogenic shock (day 4 and day 6 in 2 patients) and by hemorrhagic shock (day 15 in the third one) and did not appear to be related to rapidonset HIT. None of the 7 patients showing PF4/H IgG OD N1.0 before surgery had an adverse outcome or a prolonged in-hospital stay. A detailed comparison of
American Heart Journal Volume 160, Number 2
Selleng et al 365
Table I. Comparison of demographic and clinical characteristics between patients testing positive and negative, respectively, for anti-PF4/H IgG before cardiac surgery Anti-PF4/H IgG negative n = 547 Characteristics Male Age (y) Height (cm) Weight (kg) EURO score (additive) EURO score (logistic) Recent heparin treatment Duration of recent heparin treatment (if performed; d) Type of surgery Isolated CABG Valve repair (with or without CABG) Other⁎ Duration of extracorporeal circulation (min)
n 397
Anti-PF4/H IgG positive n = 44
%
n
(72.58)
32
68.06 170.34 84.24 5.45 6.09 292
(72.73) 67.42 168.41 81.78 6.09 9.21
(53.38)
28
9.08
317 161 69
%
(63.64) 14.75
(57.95) (29.43) (12.61)
18 18 8
92.80
(40.91) (40.91) (18.18) 98.05
P value 1.00 .66 .18 .25 .23 .23 .21 .01
.02 .13 .35 .34
CABG, Coronary artery bypass grafting. ⁎ Other combined major cardiac procedure and surgery on thoracic aorta.
Table II. Results of Cox regression hazard analysis assessing the impact of presurgery anti-PF4/H antibody EIA OD results as continuous variables on the risk for post–cardiac surgery adverse events until postoperative day 30
Anti-PF4/H IgG Anti-PF4/H IgA Anti-PF4/H IgM
Thromboembolic events
Adverse events (other than thromboembolic)
30-d mortality
0.54 (0.06; 4.66) 0.48 (0.04; 5.26) 0.64 (0.12; 3.41)
0.66 (0.26; 1.65) 0.68 (0.25; 1.85) 1.94 (1.12; 3.36)⁎
2.09 (0.29; 14.82) 1.49 (0.19; 11.90) 0.17 (0.01; 2.99)
Data are expressed as hazard ratio (95% CI); the model was adjusted for age, sex, and EURO score. ⁎ P b .05.
demographic and clinical characteristics between patients testing positive and negative, respectively, for anti-PF4/H IgG at preoperative baseline is given in Table I. Patients testing positive for anti-PF4/H IgA before surgery had no higher risk for thromboembolic complications (0/36 = 0.0% vs 30/555 = 5.4%, HR = 0.03, 95% CI = 0.00-2.27) or other postoperative complications (8/36 = 22.2% vs 145/555 = 26.1%, HR = 0.76, 95% CI = 0.361.64), no significantly longer in-hospital length of stay (15.3 vs 14.7 days, P = .84), and no higher 30-day mortality (1/36 = 2.3% vs 19/555 = 3.4%, HR = 0.91, 95% CI = 0.12-6.79). At 30-day follow-up, patients testing positive for antiPF4/H IgM before surgery had no higher risk for thromboembolic complications (3/79 = 3.8% vs 27/512 = 5.3%, HR = 0.76, 95% CI = 0.23-2.50). However, the risk for nonthromboembolic complications was increased (30/79 = 38.0% vs 123/512 = 24.0%, HR = 1.73, 95% CI = 1.15-2.61). The nonthromboembolic complications were a heterogeneous group including neurologic complications (7/79 = 8.9% vs 21/512 = 4.1%), postoperative infectious complications (5/79 = 6.3% vs 17/512 = 3.3%),
and respiratory failure (6/79 = 7.6% vs 19/512 = 3.76%). These patients had a significantly longer in-hospital length of stay (16.4 vs 15.1 days, P = .021). However, 30-day mortality was not increased (0/79 = 0.0% vs 20/512 = 3.9%, P = .09). To test for a potential effect of PF4/H antibody strength (according to EIA OD), we included anti-PF4/H EIA ODs into a Cox logistic regression model as a continuous variable. Immunoglobulin G or IgA antibodies were again not associated with an increased risk for thromboembolic or other postoperative complications, or a higher 30-day mortality, whereas IgM antibodies were again associated with an increased risk for nonthromboembolic complications (Table II).
Preoperative anti-PF4/H antibodies and HIT Of the 6 patients who tested positive in the HIPA test at baseline, 5 had an uneventful course postsurgery. But one of these patients presented with early and profound thrombocytopenia that would be consistent with “rapidonset” HIT. However, apart from thrombocytopenia, the
366 Selleng et al
American Heart Journal August 2010
Figure 2
Anti-PF4/H antibody levels, estimated by EIA OD units, are presented separately for IgG (A), IgA (B), and IgM (C) as median and 25% and 75% quartiles. Irrespective of antibody class, at all time points—day 0 = day of surgery and day 6 and day 10 postsurgery—anti-PF4/H antibody levels of patients testing positive already before surgery (black bars) were statistically significantly higher compared with those of the patients testing negative at baseline (white bars).
clinical course was uneventful; and there was no evidence of any thromboembolic or other complications. In this patient, thrombocytopenia prompted a test for anti-PF4/H antibodies independently from the study samples during the first postoperative week to exclude
HIT for safety reasons. When the result for this test was reported to be positive, anticoagulation was switched from UFH to hirudin at postoperative day 4; but platelet counts did not recover subsequently until day 10. This suggests that, in this patient, early-onset and persisting
American Heart Journal Volume 160, Number 2
thrombocytopenia may have been due to non-HIT factors, as we have reported in detail elsewhere.10 As also reported elsewhere,10 in the entire study, only 3 patients developed “typical-onset HIT”, that is, a platelet count fall by N50% between postoperative days 5 and 10, with detection of platelet-activating (HIPApositive) anti-PF4/H IgG antibodies (EIA ODs 2.064, 1.567, 1.013). It is noteworthy that all 3 patients tested negative for anti-PF4/H IgG, IgA, and IgM, as well as in the HIPA, before surgery.
Correlation between pre- and postoperative anti-PF4/H antibodies Anti-PF4/H IgG levels (estimated by mean EIA OD) in patients testing positive for IgG before surgery (n = 44) was 0.792 ± 0.255 at day 0 and increased to 0.859 ± 0.345 at day 6 and to 1.062 ± 0.369 at day 10. Patients who tested negative before surgery also showed a progressive increase in antibody levels between days 0 and 10 (day 0: 0.218 ± 0.107, day 6: 0.402 ± 0.288, day 10: 0.619 ± 0.372). Antibody levels were significantly higher at all 3 time points in the patients testing positive at baseline (P b .001; Figure 2, A provides the medians). A similar pattern was observed for anti-PF4/H IgA and IgM antibodies (Figure 2, B and C provide the medians).
Discussion This prospective study confirms a relatively high frequency of anti-PF4/H antibodies (IgG, IgA, and/or IgM) in patients before cardiac surgery.1–3,5 We found these antibodies in N20% of patients. All patients irrespective of their preoperative antibody status received heparin during and after cardiac surgery; but unlike some previous studies enrolling similar numbers of patients, we did not find an association between the presence of PF4/H antibodies of any class (ie, combined IgG, IgA, and IgM) with adverse outcomes. Furthermore, when we analyzed the different immunoglobulin classes separately, we did not find any evidence for an increased risk of adverse outcomes in patients testing positive at preoperative baseline for anti-PF4/H IgG or IgA antibodies. However, presence of anti-PF4/H IgM antibodies before cardiac surgery was associated with an increased risk for the combined incidence of nonthrombotic adverse outcomes and also for an increased length of stay. Kress et al6 also found in a 1,209-patient study that anti-PF4/H antibody–positive patients had longer mean postoperative in-hospital length of stay and higher incidences of prolonged mechanical ventilation. However, they also reported an increased rate for acute limb ischemia and renal and gastrointestinal complications. Bennett-Guerrero et al2 identified preoperative anti-PF4/ H antibodies as an independent predictor for postoperative complications (death or prolonged hospital stay) in a 466-patient study. Probably the strongest association of
Selleng et al 367
anti-PF4/H antibodies and adverse outcomes has been reported by Mattioli et al.5 In a 500-patient study, using the McNemar χ2 test, they found that patients with preexisting anti-PF4/H antibodies before cardiac surgery have a highly significantly increased risk for death (10/ 131 vs 14/369, P = .001), pulmonary embolism (10/131 vs 14/369, P = .001), myocardial infarction (9/131 vs 15/ 369, P = .01), and stroke (4/131 vs 7/369, P = .01) at day 30 (however, applying categorical analyses using the Fisher exact test, these P values are all N .05). They also found PF4/H antibody–positive patients to be at risk for a higher incidence of thrombotic events over time (ie, 36 months).5 Three other studies (enrolling 111,1 299,3 and 368 patients4), however, found no correlation between antibody levels before cardiac surgery and the rate of postoperative thromboembolic complications. Our study is only the second (along with that of Gluckman et al4) to have analyzed patient sera separately for anti-PF4/H IgG, IgM, and IgA classes and for presence of platelet-activating antibodies. The finding that anti-PF4/ H antibodies of IgM class were (weakly) predictive for adverse outcomes is surprising. However, these IgM antibodies were not a risk factor for later seroconversion to functionally reactive IgG antibodies and HIT. Even among the 3 patients who developed typical-onset HIT during this study, none had a positive preoperative test result for anti-PF4/H antibodies (including for IgM), thus indicating that they seroconverted from a negative baseline antibody status. Several studies have shown that anti-PF4/H IgM antibodies (as well as IgA antibodies) are not capable to engage and to cross-link the platelet FcγIIa receptors on platelets,19,20 which are considered the key mechanisms underlying platelet activation in HIT.21 As PF4/H IgM antibodies do not cause the HIT syndrome, which requires intravascular activation of platelets via the Fcγ receptor, it seems unlikely that the nonthromboembolic complications associated with the presence of IgM antibodies could be prevented by avoiding intra- and postoperative heparin. Furthermore, as the use of nonheparin anticoagulation for cardiac surgery patients is not well established, its wider use could cause increased bleeding.22 Indeed, given the relatively high number of patients with precardiac surgery PF4/H antibodies (≈15%-20%), deliberate use of alternative anticoagulants because of preoperative detection of antibodies might itself result in significant complications (although promising results regarding the use of bivalirudin on CPB have been recently published).23,24 Further studies evaluating the risk of bleeding vs thrombosis when using nonheparin anticoagulation on CPB are urgently needed. Considering the findings of our and other similar studies,2-4 anti-PF4/H antibody testing before cardiac surgery remains indicated only if a patient belongs to the 2 groups in whom these antibodies have been shown to be clinically relevant for peri- and postsurgical
American Heart Journal August 2010
368 Selleng et al
management.22 The first group consists of patients with typical clinical signs of HIT, for example, a new-onset platelet count fall bearing a temporal relationship with heparin administered preoperatively for acute coronary syndrome while waiting for urgent coronary bypass surgery. In these patients, detection of platelet-activating anti-PF4/H IgG antibodies likely requires nonheparin anticoagulation for CPB.22 The second group represents patients with a history of HIT, typically within the previous 4 to 12 weeks, who now require urgent cardiac surgery. In these patients, the absence of plateletactivating anti-PF4/H antibodies allows the use of UFH during CPB.25,26 Our study has several limitations. Only a minority of patients (ie, ≈20%) had preexisting anti-PF4/H antibodies, thus limiting the number of patients at risk. Furthermore, certain outcome parameters like thromboembolic events and post–cardiac surgery death occurred in only 5.1% and 3.4%, respectively. Finally, the observation period of 30 days is not suitable to detect any long-term effects. A definite answer of a potential minor association of anti-PF4/H antibodies before cardiac surgery with adverse outcomes will require a large prospective trial enrolling several thousand patients ideally with a longer observation period than 30 days. The association of anti-PF4/H antibodies with nonthrombotic adverse events might indicate that anti-PF4/H IgM antibodies could represent a surrogate marker for other risk factors. However, another possibility is that we may not have included all potential confounders for these nonthrombotic complications in our Cox regression analysis. All analyses were adjusted for age, sex, and EURO score. The EURO score, however, was developed to predict mortality but not to assess the risk for prolonged length of hospital stay or other complications. Finally, the association between antibody presence and length of hospital stay excluded some patients with a poor outcome, namely, patients who died before day 30. Therefore, the potential association of anti-PF4/H IgM antibodies with adverse outcomes needs to be interpreted with caution. In summary, anti-PF4/H antibodies (of any immunoglobulin class) are present before cardiac surgery in approximately 20% of patients. The presence of such antibodies was associated with a low risk (at most, 1/ 128 or b1%) of rapid-onset HIT during the early postoperative course and was not associated with an increased risk of typical-onset HIT. Our study does not confirm earlier reports on a relatively strong association of these antibodies with the risk for HIT, new thrombosis, or mortality. However, we did detect an association between increased risk of nonthromboembolic morbidity and increased length of hospital stay among those patients who had detectable anti-PF4/H antibodies of IgM class. Inasmuch as this class of
antibodies is not associated with HIT, our study suggests that IgM antibodies could represent a surrogate marker for risk of non–HIT-related adverse events (eg, through a common factor such as pre/ perioperative inflammation). Because it seems unlikely that nonheparin anticoagulation would avoid these complications, heparin should remain as the anticoagulant of choice during CPB in otherwise asymptomatic patients who test positive for anti-PF4/H antibodies in a clinical context that does not suggest acute HIT. Our findings are also consistent with the recommendations of the American College of Chest Physicians Consensus Conference guidelines (eighth edition)22 that routine testing for anti-PF4/H antibodies before cardiac surgery should be avoided.
Acknowledgements We thank the staff of the Heart Center Mecklenburg/ Westpommerania, Karlsburg, for collecting the blood samples.
Disclosures The authors state that they have no conflict of interest.
References 1. Bauer TL, Arepally G, Konkle BA, et al. Prevalence of heparin-associated antibodies without thrombosis in patients undergoing cardiopulmonary bypass surgery. Circulation 1997; 95:1242-6. 2. Bennett-Guerrero E, Slaughter TF, White WD, et al. Preoperative antiPF4/heparin antibody level predicts adverse outcome after cardiac surgery. J Thorac Cardiovasc Surg 2005;130:1567-72. 3. Everett BM, Yeh R, Foo SY, et al. Prevalence of heparin/platelet factor 4 antibodies before and after cardiac surgery. Ann Thorac Surg 2007;83:592-7. 4. Gluckman TJ, Segal JB, Schulman SP, et al. Effect of anti–platelet factor-4/heparin antibody induction on early saphenous vein graft occlusion after coronary artery bypass surgery. J Thromb Haemost 2009;7:1457-64. 5. Mattioli AV, Bonetti L, Zennaro M, et al. Heparin/PF4 antibodies formation after heparin treatment: temporal aspects and long-term follow-up. Am Heart J 2009;157:589-95. 6. Kress DC, Aronson S, McDonald ML, et al. Positive heparin–platelet factor 4 antibody complex and cardiac surgical outcomes. Ann Thorac Surg 2007;83:1737-43. 7. Greinacher A, Juhl D, Strobel U, et al. Heparin-induced thrombocytopenia: a prospective study on the incidence, platelet-activating capacity and clinical significance of antiplatelet factor 4/heparin antibodies of the IgG, IgM, and IgA classes. J Thromb Haemost 2007; 5:1666-73. 8. Warkentin TE, Sheppard JA, Moore JC, et al. Laboratory testing for the antibodies that cause heparin-induced thrombocytopenia: how much class do we need? J Lab Clin Med 2005;146:341-6. 9. Paparella D, Scrascia G, Galeone A, et al. Formation of anti–platelet factor 4/heparin antibodies after cardiac surgery: influence of perioperative platelet activation, the inflammatory response, and
American Heart Journal Volume 160, Number 2
10.
11.
12.
13.
14.
15.
16.
17.
histocompatibility leukocyte antigen status. J Thorac Cardiovasc Surg 2008;136:1456-63. Selleng S, Malowsky B, Strobel U, et al. Early-onset and persisting thrombocytopenia in post-cardiac surgery patients is rarely due to heparin-induced thrombocytopenia even when antibody tests are positive. J Thromb Haemost 2010;8:30-6. Nashef SA, Roques F, Michel P, et al. European system for cardiac operative risk evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9-13. Pouplard C, May MA, Regina S, et al. Changes in platelet count after cardiac surgery can effectively predict the development of pathogenic heparin-dependent antibodies. Br J Haematol 2005; 128:837-41. Juhl D, Eichler P, Lubenow N, et al. Incidence and clinical significance of anti-PF4/heparin antibodies of the IgG, IgM, and IgA class in 755 consecutive patient samples referred for diagnostic testing for heparininduced thrombocytopenia. Eur J Haematol 2006;76:420-6. Greinacher A, Michels I, Kiefel V, et al. A rapid and sensitive test for diagnosing heparin-associated thrombocytopenia. Thromb Haemost 1991;66:734-6. Schenk S, El Banayosy A, Morshuis M, et al. IgG classification of antiPF4/heparin antibodies to identify patients with heparin-induced thrombocytopenia during mechanical circulatory support. J Thromb Haemost 2007;5:235-41. Whitlatch NL, Perry SL, Ortel TL. Anti–heparin/platelet factor 4 antibody optical density values and the confirmatory procedure in the diagnosis of heparin-induced thrombocytopenia. Thromb Haemost 2008;100:678-84. Savi P, Chong BH, Greinacher A, et al. Effect of fondaparinux on platelet activation in the presence of heparin-dependent antibodies: a blinded comparative multicenter study with unfractionated heparin. Blood 2005;105:139-44.
Selleng et al 369
18. Warkentin TE, Kelton JG. Temporal aspects of heparin-induced thrombocytopenia. N Engl J Med 2001;344: 1286-92. 19. Denomme GA, Warkentin TE, Horsewood P, et al. Activation of platelets by sera containing IgG1 heparin-dependent antibodies: an explanation for the predominance of the Fc gammaRIIa “low responder” (his131) gene in patients with heparin-induced thrombocytopenia. J Lab Clin Med 1997;130:278-84. 20. Rauova L, Zhai L, Kowalska MA, et al. Role of platelet surface PF4 antigenic complexes in heparin-induced thrombocytopenia pathogenesis: diagnostic and therapeutic implications. Blood 2006;107:2346-53. 21. Kelton JG, Sheridan D, Santos A, et al. Heparin-induced thrombocytopenia: laboratory studies. Blood 1988;72:925-30. 22. Warkentin TE, Greinacher A, Koster A, et al. Treatment and prevention of heparin-induced thrombocytopenia: American College of Chest Physicians evidence-based clinical practice guidelines (8th edition). Chest 2008;133:340S-80S. 23. Dyke CM, Smedira NG, Koster A, et al. A comparison of bivalirudin to heparin with protamine reversal in patients undergoing cardiac surgery with cardiopulmonary bypass: the EVOLUTION-ON study. J Thorac Cardiovasc Surg 2006;131:533-9. 24. Koster A, Dyke CM, Aldea G, et al. Bivalirudin during cardiopulmonary bypass in patients with previous or acute heparin-induced thrombocytopenia and heparin antibodies: results of the CHOOSEON trial. Ann Thorac Surg 2007;83:572-7. 25. Pötzsch B, Klovekorn WP, Madlener K. Use of heparin during cardiopulmonary bypass in patients with a history of heparin-induced thrombocytopenia. N Engl J Med 2000;343:515. 26. Selleng S, Haneya A, Hirt S, et al. Management of anticoagulation in patients with subacute heparin-induced thrombocytopenia scheduled for heart transplantation. Blood 2008;112:4024-7.