Prevalence of Heparin-Associated Antibodies Without Thrombosis in Patients Undergoing Cardiopulmonary Bypass Surgery
Background Patients with cardiovascular disease almost invariably receive heparin before cardiopulmonary bypass surgery, which places them at risk of developing heparin-associated antibodies with a risk of thromboembolic complications. This study was designed to determine the prevalence of heparin-induced antibodies in patients before and after cardiopulmonary bypass surgery.
Methods and Results Plasma from 111 patients was tested before surgery and 5 days after surgery for heparin-dependent platelet-reactive antibodies with a 14C-serotonin–release assay (SRA) and for antibodies to heparin/platelet factor 4 complexes with an ELISA. Heparin exposure after surgery was minimized. Heparin-dependent antibodies were detected before surgery in 5% of patients with SRA and 19% of patients with ELISA. By the fifth postoperative day, there was a marked increase in patients positive on the SRA or ELISA (13% and 51%, respectively; P<.01 for each). Patients who had received heparin therapy earlier in their hospitalization were more likely to have a positive ELISA before surgery (35%; P=.017) and a positive ELISA (68%; P=.054) or SRA (30%; P=.002) after surgery. However, there was no difference in the prevalence of thrombocytopenia or thromboembolic events between the antibody-positive and -negative groups.
Conclusions Approximately one fifth of patients undergoing cardiopulmonary bypass surgery have heparin-induced platelet antibodies detectable before the procedure as a result of prior heparin exposure, and many more develop antibodies after surgery. The absence of an association between these antibodies and thromboembolic complications in this study may be, in part, attributable to careful avoidance of heparin after surgery. The high prevalence of heparin-induced antibodies in this setting suggests that these patients may be at risk of developing thrombotic complications with additional heparin exposure.
Heparin is one of the most commonly used drugs in patients requiring therapeutic intervention for cardiovascular disease. Heparin is administered to almost all patients during cardiac catheterization, angioplasty, cardiopulmonary bypass, and the treatment of unstable angina and myocardial infarction. The clinical advantages of heparin include its immediate onset after intravenous administration, relative ease of titration, and rapid reversibility with protamine. Indeed, in some situations, such as cardiopulmonary bypass, no substitute for heparin is available for general use.
Despite its many desirable characteristics, the use of heparin has some serious side effects. One of the most common side effects is heparin-induced thrombocytopenia (HIT). Thrombocytopenia occurs in ≈1% to 4% of patients,1 some of whom develop multiple and recurrent arterial and venous thromboemboli.1 2 Thus, heparin can cause the same devastating complications that it is administered to prevent. HIT is mediated by antibodies that bind to complexes between a heterologous mucopolysaccharide, heparin, and platelet factor 4 (PF4), a tetrameric protein that is stored in the α-granules of platelets and released when platelets are activated.3 PF4 binds to the surface of activated platelets and to heparin-like endothelial cell proteoglycans,4 where it may become a target for these autoantibodies.5 6 Patients who are reexposed to even small amounts of heparin through arterial lines and bonded catheters can experience devastating complications from immune-mediated platelet activation and vascular injury.7 8 9
Patients undergoing cardiopulmonary bypass surgery have been exposed to heparin on one or more occasions before surgery. During bypass, patients are then exposed to high doses of heparin, and at the same time, platelets are activated.10 Exposure to heparin may continue in the postoperative setting. Thrombocytopenia and thrombotic complications have been reported to occur in this setting.11 12 13 14 However, the prevalence of heparin-induced antibodies and their relation to thrombotic complications after coronary artery bypass graft surgery have not been reported. To address this issue, we performed a prospective study of 111 patients undergoing cardiopulmonary bypass and determined the prevalence of heparin-dependent platelet-reactive antibodies by using both the classic 14C-serotonin–release assay (SRA) and the newly described ELISA to detect anti-heparin/PF4 antibodies. Our data suggest that such antibodies are more common in this population than has been appreciated previously, suggesting the need to avoid prolonged exposure to heparin in these patients.
Prospective Clinical Trial
All patients undergoing cardiopulmonary bypass during an 18-month period were offered the option of participating in this study. Of the 150 patients entered in the study, preoperative and postoperative blood samples were both available in only 111 patients, who form the basis of this study. Patients receiving heparin within 4 hours of surgery were excluded because residual heparin in the plasma might interfere with the SRA. Blood samples for SRA and ELISA were obtained in sodium citrate (0.32% final concentration) just before induction of anesthesia and on the fifth postoperative day. Platelet-poor plasma was isolated within 1 hour of collection, heated at 56°C for 1 hour, and stored at −70°C for subsequent analysis. Clinical data were obtained through a review of each patient's chart. All exposures to heparin before and after surgery, whether intravenously or subcutaneously, were noted. Aspirin use was defined as the use of any aspirin-containing products by the patient within the 7 days preceding the bypass procedure. A complete blood count, including a platelet count, was obtained on the day of surgery and on postoperative days 1, 2 and 5. All other laboratory tests were obtained as indicated clinically. The heparin/PF4 ELISA and the SRA were usually performed after the patients had been discharged, and the results of these tests did not affect management decisions. The protocol was approved by the Institutional Review Board of Jefferson Medical College, and informed written consent was obtained from all participants.
Anesthetic and Surgical Management
A radial artery catheter and a heparin-coated Swan-Ganz catheter (Baxter-Edwards) were placed in each patient before surgery. High-dose fentanyl (50 to 100 μg/kg, Elkins-Sinn) was used to induce and maintain anesthesia. Beef lung heparin (UpJohn) was administered at a dose of 300 U/kg before bypass. Additional heparin was given by bolus as needed to achieve a whole blood activated clotting time of >480 seconds. Anticoagulation was maintained at this level throughout bypass. The cardiopulmonary bypass circuit consisted of a membrane oxygenator (Baxter-Bentley), closed venous reservoir (Baxter-Bentley), and centrifugal pump (Medtronic-Biomedicus). Moderate systemic hypothermia, hemodilution, and blood cardioplegia were used. Approximately 25% of the patients received ε-aminocaproic acid (150 μg/kg, Lederle Laboratories) to minimize postoperative blood loss. At the end of bypass, heparin activity was reversed with protamine. All heparin administration, including flushes of indwelling arterial catheters and exposure through heparin-bonded Swan-Ganz catheters, was discontinued after the first postoperative day in >80% of patients.
Detection of Antibodies to Heparin/PF4
Preoperative and postoperative plasma samples were tested with the use of Asserachrom H PF4 ELISA kits (Diagnostica STAGO). All reagents were included in the kit with the exception of 3 mol/L sulfuric acid (JT Baker Chemical). Solutions were reconstituted, and the ELISA protocol was followed according to the manufacturer's directions. Briefly, plasma was diluted 1:100 in buffer (10% goat serum in phosphate-buffered saline). Duplicate samples (0.2 mL) were incubated for 1 hour at 22°C in ELISA wells covalently coated with heparin/PF4 complexes.15 The plates were washed and incubated with peroxidase-conjugated goat anti-human IgG/IgA/IgM (0.2 mL) for an additional hour at 22°C. Enzymatic activity was developed by the addition of orthophenylenediamine and urea peroxide tablets dissolved in distilled water. Ten minutes later, the absorbance was measured at 492 nm (Anthos Reader 2001, Anthos Labtec Instruments). Each assay included a known positive control (heparin/PF4 antibody positive) and negative control (no heparin/PF4 antibodies) run in parallel. The results were interpreted according to the manufacturer's recommendations: a negative result was defined as an A492nm of <0.25, an intermediate result was defined as an A492nm of ≥0.25 and <0.5, and a positive result was defined as an A492nm of ≥0.5. These cutoff values recommended by the manufacturer are based on studies of the PF4/heparin ELISA in patients with established HIT and various control populations.15 An absorbance of 0.5 is >8 SD above the values obtained from healthy normal subjects and are values that are characteristically seen in patients with HIT.15 Samples that generate absorbances in the intermediate range of detection (≥0.25 to <0.5) are <8 SD necessary for a positive result and >2 SD above those of healthy normal subjects.15
Heparin-dependent platelet-reactive antibodies were detected with a two-point SRA as described previously,16 except that heat-inactivated plasma was used. Each sample was run in triplicate, and the results were averaged. The test was considered positive when there was >20% serotonin release at 0.1 U/mL and <20% release at 100 U/mL heparin but no release when patient plasma was added to the platelets in the absence of heparin.
Comparisons of assay results by subgroups, such as sex, were made with Fisher's exact test. For paired data, such as a comparison of two assays or preoperative/postoperative comparisons, we used McNemar's χ2 test. Quantitative comparisons of ELISA results were performed with the use of Wilcoxon rank sum statistics. Regressions of platelet counts were analyzed after log transformation. All P values reported are for the results of two-tailed tests.
The frequency of a positive ELISA or SRA did not vary with patient age, the use of aspirin in the week before bypass, cross-clamp or total bypass time, total amount of heparin administered during bypass, the lot of heparin used, or ε-aminocaproic acid administration (data not shown). Although there was no significant difference in ELISA or SRA positivity between men and women (Table 1⇓), the percentage increase in antibody titer after surgery was higher in women, with a median increase of 190% for women versus 65% for men (P=.014). In the 13 patients who identified themselves as African-American or Hispanic, there was a greater preoperative (3 of 13) and postoperative (6 of 13) SRA positivity compared with whites (P=.011 for preoperative; P=.001 for postoperative).
Prevalence of Antibodies to Heparin/PF4 complexes
Twenty-one patients (19%) tested positive for anti-heparin/PF4 antibodies by ELISA before surgery, and 57 (51%) tested positive after bypass (P<.001 for preoperative versus postoperative). Furthermore, 82 patients (74%) had an increase in antibody titer, measured as an increase in absorbance. Patients with a negative ELISA before surgery had an increase in mean absorbance from 0.148±0.062 to 0.751±0.792 (mean±SD); 23 of 54 patients (43%) developed a positive test, and another 12 (22%) developed an intermediate test result after surgery. Patients with an intermediate ELISA before surgery had an increase in mean absorbance from 0.347±0.068 to 0.897±0.857; 19 of 36 patients (53%) developed a positive test, whereas eight tests (22%) reverted to normal after surgery. Of the 21 patients testing positive before surgery, there was an increase in mean absorbance from 0.893±0.480 to 1.271±0.975, with 15 of 21 (71%) continuing to have a positive test, whereas five (24%) tests reverted to an intermediate value. Anti-heparin/PF4 antibodies were not detected in only 1 of the 21 samples collected after surgery in this group of patients. Overall, plasma from 35 patients (32%) gave an A492nm that was >1.0 after surgery.
Additional experiments were performed to determine whether the apparent increase in antibody titer after surgery could be the result of preexisting antibodies that formed immune complexes with heparin or PF4 released during the bypass, which were subsequently detected by ELISA. Plasma samples from 2 patients with a positive ELISA before and after surgery, 1 patient whose ELISA converted to positive after surgery, and 2 patients whose ELISA remained negative after surgery were studied. Preoperative plasma samples from each patient, diluted 1:100 in 10% FCS, were incubated with heparin (0.2 to 1.0 U/mL) or with recombinant human PF4 (20 ng/mL to 1 μg/mL),17 and the ELISA was repeated. In none of the samples tested was an increase in absorbance detected. Furthermore, soluble immune complexes containing PF4 were not detected in the IgG fractions of postoperative specimens by SDS-PAGE or Western blot analysis (data not shown). Therefore, the increase in titer likely reflects additional antibody formation.
Prevalence of Antibodies Detected With the SRA
Five patients (5%) had heparin-dependent platelet-reactive antibodies detected with SRA before surgery, whereas 14 (13%) were positive after surgery (P=.007 for preoperative versus postoperative). One patient who had a positive SRA before surgery had a negative test afterwards. An additional 10 patients (9%) developed positive SRAs after surgery. Most patients (87%) had a negative SRA both before and after surgery. Patients testing positive by ELISA after surgery were more likely than those testing negative to be positive by SRA (Table 2⇓). Twelve of the 57 samples collected after surgery that were positive by ELISA were also positive by SRA; only 2 of 54 samples with a negative or intermediate ELISA result were positive by SRA (P<.001). Furthermore, the mean absorbance of samples positive by SRA was higher in the ELISA (1.86±1.09) than that of the samples that were negative in the SRA (1.27±0.71; P<.001), suggesting an association between the antibody titer and platelet activation. However, 5 of 19 samples (26%) positive by SRA were negative by ELISA.
Patients who had received heparin earlier in the hospitalization during which the surgery was performed were more likely to have a positive ELISA both before and after surgery than were patients whose only exposure to heparin had been on previous hospitalizations (35% versus 13% before surgery, P=.017; 68% versus 46% on day 5 after surgery, P=.054) (Table 1⇑). Similarly, of the patients who developed a positive SRA after surgery, 71% had received heparin earlier in the hospitalization (P=.002).
Correlation Among Heparin-Induced Antibodies, Platelet Counts, and Thromboembolic Complications
Platelet counts fell on the first postoperative day in all patients undergoing coronary artery bypass, from a mean of 228 730±65 492 μL−1 (mean±SD) to 137 223±54 493 μL−1. The mean platelet count decreased further to 118 615±45 685 μL−1 on day 2 and recovered to 249 200±113 815 μL−1 on day 5. There was no difference in mean platelet count between the antibody-positive and antibody-negative groups by SRA or ELISA on any day. Multivariate analysis indicated an inverse correlation between platelet counts on postoperative day 5 and increasing age (P<.01), time on bypass pump (P=.006), and female sex (P=.018) and a positive correlation with preoperative platelet count (P<.001). Only 2 patients had thromboembolic complications, neither of whom had a positive SRA (Table 1⇑). Nor was there a correlation between these adverse events and the results of the ELISA (Table 1⇑). Three patients died, none of thromboembolic events.
In the present study, we evaluated the prevalence of heparin/PF4 antibodies in patients before and after cardiopulmonary bypass surgery. A surprising number of patients had already developed antibodies to heparin-PF4 complexes before undergoing cardiopulmonary bypass surgery. Antibody titers were further increased in three fourths of the patients after the procedure. The only preoperative characteristic that we were able to identify that correlated with preoperative or postoperative antibody titer was a history of exposure to unfractionated heparin during the hospitalization in which the bypass took place. The odds ratio of forming heparin-dependent antibodies was 3.19 with the ELISA and 9.38 with the less sensitive SRA when patients who had recently received heparin were compared with those whose exposure was more remote.
Recently, antibodies to heparin/PF4 complexes have been identified in the plasma of >90% of patients with HIT.3 15 18 However, low titers of anti-heparin/PF4 antibodies have been detected in ≈20% of patients receiving heparin for medical therapy who are not thrombocytopenic.15 18
The prevalence of antibodies to heparin/PF4 complexes in this heavily treated population was considerably higher. Fifty-one percent of these patients were positive or intermediate positive before surgery, and the prevalence increased to 83% by the fifth day after surgery, which is comparable to a study recently reported by others.19 In our study, the prevalence of a positive SRA also significantly increased after bypass surgery. This finding suggests that the heterologous mucopolysaccharide heparin is considerably more immunogenic than has been previously appreciated.20 Whether low-molecular-weight heparins will prove to be substantially less immunogenic in this population remains to be determined.21
The apparent increase in antibody titer after surgery was not the result of immune complexes formed between residual heparin that had not been neutralized and PF4 released from platelets during bypass. The addition of heparin or PF4 to preoperative samples (negative or positive) in vitro had no effect on the ELISA results. Therefore, it is likely that the increase in absorbance seen 5 days after surgery represents an anamnestic response in previously sensitized individuals. Moreover, 53% of patients whose preoperative ELISA fell in the intermediate range became positive after surgery, which is consistent with the hypothesis that the antibody repertoires in many, but not all, of these patients undergoes affinity maturation with additional exposure to heparin. The apparent decrease in the ELISA after surgery in a few patients may have been the result of transient consumption of antibody, although serial studies would have been necessary to evaluate this possibility.
What is most remarkable is that given the very high prevalence of heparin-induced antibodies in this setting, there was no associated thrombocytopenia or thrombosis that could be attributed to these antibodies, even among those patients in whom antibodies were detected before surgery. Several factors may contribute to this seemingly anomalous result. First, not all antibodies detected with the heparin/PF4 ELISA may have comparable pathogenic potential. For example, the clinical impact of the IgM and IgA antibodies detected by ELISA is unknown. These antibodies may differ in affinity and antigenic specificity and/or exert differing effects on platelet and vascular functions. Second, only some of the epitopes expressed by heparin/PF4 complexes bound on plastic wells may be expressed when this complex is bound to platelets and endothelial cells in vivo. Third, thrombotic events may have gone unrecognized since there was no systematic investigation for subclinical events. Fourth, our sample size was relatively small. In a previous study, only 11 of 1155 patients (0.95%) developed serious complications after cardiopulmonary bypass attributable to HIT.11 Similar results have been reported by others.22 Therefore, a much larger sample size would have been required to determine whether a relation exists between the in vitro test results and adverse clinical events. Finally, there might be a significant biological variation in the host response to vascular injury mediated by antibodies.
These results suggest that heparin-induced antibodies are relatively common but do not invariably cause severe clinical sequelae. One explanation for the low frequency of heparin-associated complications in the present study lies in our policy of minimizing heparin exposure after surgery. The only postoperative exposure to heparin in >80% of our study patients was the use of a heparin-coated pulmonary artery catheter and the use of heparin to flush arterial catheters, both of which were discontinued within 24 hours of surgery. Our protocol thereby differs from that of the two other prospective studies reported to date, in which patients continued to receive heparin therapy for several days in the postoperative period.19 22 In those studies, thrombotic complications were noted.
The results of this study also pose an important diagnostic dilemma. Although the recently developed ELISA, which uses a polyclonal antibody to human IgG, IgA, and IgM, is more sensitive than the SRA, the presence of such antibodies did not predict an increase in heparin-associated complication, even when they were present at a high titer. Furthermore, although the numbers were small, SRA positivity also did not identify patients at risk for thrombosis under the specific clinical conditions present in the study. Prospective clinical trials are therefore needed to address the risk of reexposing patients with and without detectable antibodies to additional heparin and to define the pathogenic role of these antibodies in HIT. Nevertheless, it is clear that unfractionated heparin is far more immunogenic than has been appreciated and that other means to provide anticoagulation during cardiovascular procedures should be sought.
This work was supported in part by a Special Purpose Grant from Thomas Jefferson University Hospital; by grants HL-44956 (Dr Konkle), HL-09163 (Dr Shapiro), HL-50100 (Dr Shapiro), HL-54749 (Drs Poncz and Cines), HL-40387 (Dr Cines), HL-50790 (Dr Cines), HL-49517 (Dr Cines), HL-37419 (Dr Poncz), and HL-56265 (Dr Poncz); and by a Fellowship Award from the American Heart Association, Southeast Pennsylvania Affiliate (Dr Arepally).
*Dr Amiral is employed by SERBIO, a research subdivision of the STAGO group, the manufacturer of the PF4/heparin ELISA kit used in this study. This study was performed with kits provided by the manufacturer for research purposes only.
- Received May 8, 1996.
- Revision received October 14, 1996.
- Accepted October 23, 1996.
- Copyright © 1997 by American Heart Association
Rybak ME, Gimbrone MA Jr, Davies PF, Handin RI. Interaction of platelet factor four with cultured vascular endothelial cells. Blood. 1989;73:1534-1539.
Zucker MB, Katz IR. Platelet factor 4: production, structure, and physiologic and immunologic action. Proc Soc Exp Biol Med. 1991;198:693-702.
Heeger PS, Backstrom JT. Heparin flushes and thrombocytopenia. Ann Intern Med. 1986;105:143.
Kalangos A, Relland JYM, Massonet-Castel S, Acar C, Carpentier A. Heparin-induced thrombocytopenia and thrombosis following open heart surgery. Eur J Cardiothorac Surg. 1994;8:199-203.
Sheridan D, Carter C, Kelton JG. A diagnostic test for heparin-induced thrombocytopenia. Blood. 1986;67:27-30.
Park KS, Rifat S, Eck H, Adachi K, Surrey S, Poncz M. Biologic and biochemic properties of recombinant platelet factor 4 demonstrate identity with the native protein. Blood. 1990;75:1290-1295.