The Incidence of Recognized Heparin- Induced Thrombocytopenia in a Large, Tertiary Care Teaching Hospital*

CHEST The Incidence of Recognized Heparin- Induced Thrombocytopenia in a Large, Tertiary Care Teaching Hospital* Maureen A. Smythe, PharmD, FCCP; John M. Koerber, PharmD; and Joan C. Mattson, MD Original Research ANTITHROMBOTIC THERAPY Background: Heparin-induced thrombocytopenia (HIT) is estimated to occur in up to 5% of patients receiving unfractionated heparin. The goal was to determine the incidence of HIT within our 1,061-bed tertiary care institution. Methods: A retrospective review of three hospital database systems (ie, admission, pharmacy, and laboratory) was undertaken for a 1-year period ending in March 2004. The pharmacy database was queried to identify patients who received heparin and those who received a direct thrombin inhibitor (DTI). The medical records of patients receiving a DTI were reviewed to categorize the indication for DTI therapy. The laboratory system database was queried to retrieve heparin platelet factor 4 immunoassay results. Results: A total of 58,814 patient admissions occurred with an estimated 24,068 patients being exposed to unfractionated heparin. DTI therapy was administered to 133 patients. Of these, 49 new HIT cases and 15 cases of suspected HIT (unconfirmed) were identified. The overall incidence of recognized new HIT was 0.2%. New HIT occurred in 0.76% of patients receiving therapeutic-dose IV heparin and in < 0.1% of patients receiving antithrombotic prophylaxis (subcutaneous heparin). Forty-nine percent of all new HIT cases were in coronary artery bypass and/or valve replacement surgery patients, while no cases were identified in hip/knee arthroplasty patients. Conclusions: The incidence of recognized HIT in a large teaching institution was 0.2%, with a 0.76% incidence in those patients receiving therapeutic-dose IV heparin. The low incidence likely reflects a brief duration of heparin exposure for many patients. Approximately half of all new HIT cases were recognized in the cardiovascular surgery population. (CHEST 2007; 131:1644 1649) Key words: anticoagulation; heparin; heparin-induced thrombocytopenia; thrombosis Abbreviations: DTI direct thrombin inhibitor; ELISA enzyme-linked immunosorbent assay; HIT heparininduced thrombocytopenia; HPF4 heparin platelet factor 4; OD optical density; SQ subcutaneous Heparin-induced thrombocytopenia (HIT) is a severe immune-mediated adverse effect of heparin, which can result in life-threatening venous and arterial thrombosis. Patients suspected of having HIT typically present with a 50% drop in platelet count beginning 5 to 14 days after heparin initiation and/or the development of a thrombotic event while receiving therapeutic anticoagulation. Rapid-onset HIT (ie, a drop in platelet count within 24 h in patients with previous heparin exposure) and delayed-onset HIT have also been described. 1 Rare presentations of HIT include skin lesions at the heparin injection sites or an acute systemic reaction immediately following an IV bolus dose of heparin. 1 HIT occurs as the result of the development of IgG antibodies, which can activate platelets in the presence of heparin (unfractionated or low-molecular weight heparin). Activated platelets release microparticles, which promote thrombin generation and result in a subsequent prothrombotic state. 2,3 The diagnosis of HIT should consider both the clinical presentation and the results of laboratory testing. Although two types of laboratory tests are available for confirming the diagnosis of HIT (enzyme-linked 1644 Original Research

immunosorbent assays [ELISAs] and functional or activation assays), limitations in laboratory testing can be problematic. Although ELISAs have a high sensitivity, they fail to reach 100% sensitivity and have a low specificity. Functional assays, while having a higher specificity, are not routinely available due to their complexity. The risk of HIT is influenced by many factors, including the route, dose, duration, and type of For editorial comment see page 1620 heparin therapy; patient population; gender; as well as a history of heparin exposure. 1,4,5 HIT occurs in up to 5% of postoperative patients receiving unfractionated heparin for thromboprophylaxis. 1 Incidence rates as high as 5% were reported primarily in older orthopedic studies 6,7 in which the duration of heparin therapy was longer (10 to 14 days) than is typically used today. Thromboprophylaxis with lowmolecular-weight heparin has been shown to significantly reduce the development of HIT in surgical patients when compared to unfractionated heparin. 7 In patients receiving IV unfractionated heparin for the treatment of thrombosis, the risk of HIT is estimated to be approximately 1%. 1 At the time of this study, unfractionated heparin was the primary parenteral anticoagulant and thromboprophylaxis agent of choice in our institution. As part of a safety assessment, our institution previously evaluated the rate of major bleeding with unfractionated heparin and found it to be comparable to that found in the published literature. 8 In order to further evaluate the safety of unfractionated heparin therapy, the goal of this study was to identify the *From the Department of Pharmaceutical Services (Drs. Smythe and Koerber), and the Division of Hematopathology (Dr. Mattson), Department of Clinical Pathology, William Beaumont Hospital, Royal Oak, MI. This article was presented at the International Society of Thrombosis and Haemostasis Meeting, Sydney, NSW, Australia, August 2005. All authors meet the criteria for authorship. Drs. Smythe and Koerber had full access to all the study data, and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr. Smythe has been a consultant and speaker for GlaxoSmithKline and a recipient of an educational grant from GlaxoSmithKline. Drs. Koerber and Mattson have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Manuscript received August 27, 2006; revision accepted January 31, 2007. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Maureen A. Smythe, PharmD, FCCP, Department of Pharmacy Practice, Suite 2190, Wayne State University, 259 Mack Ave, Detroit, MI 48201; e-mail: msmythe@beaumont.edu DOI: 10.1378/chest.06-2109 incidence of recognized HIT at our institution, a 1,061-bed, community-based, tertiary care teaching facility. Methods and Materials This study was a retrospective review of three hospital database systems (admission, pharmacy, and laboratory). This study was approved by the Human Investigation Committee (institutional review board), which waived the requirement for informed consent. The study period was 1 year, ending March 31, 2004. The hospital admission database was queried to identify the number of initial and repeat hospital admissions. The pharmacy database system was queried for a 6-month period from October 2003 through March 2004. Patients receiving any heparin product dispensed by the pharmacy were identified. Heparin administered during cardiac catheterization or surgery is not dispensed by the pharmacy and therefore was not routinely captured. Heparin administration to surgical patients preoperatively or postoperatively was captured. Heparin exposure during this 6-month period was used to estimate the number of patients exposed to heparin annually at our institution. Patients receiving a direct thrombin inhibitor (DTI) during the entire study period (April 2003 through March 2004) were also identified using the pharmacy database. The laboratory system database was used to retrieve all heparin platelet factor 4 (HPF4) assay results (PF4 ELISA Assay; GTI; Waukesha, WI) in patients in whom a DTI was prescribed. A positive assay result was defined as an optical density (OD) of 0.4 with 50% inhibition of the positive reaction by administration of 100 U/mL heparin during a confirmation step. The medical records of those patients receiving a DTI were reviewed to categorize the indication for DTI therapy into one of the following categories: New HIT case: clinical diagnosis of HIT with positive HPF4 antibody result; Suspected HIT (unconfirmed): patient management consistent with a new diagnosis of HIT (sufficient clinical probability to initiate and continue DTI therapy); the ELISA result for HPF4 antibody was negative, equivocal (ie, the inhibition of a positive reaction by 50% in the presence of excess heparin 9 ), or not performed; the medical record indicated probable, likely, or possible HIT and did not include patients found to have a negative serotonin release assay result (performed as a sendout test by an external laboratory); History of HIT: a documented history of HIT in the medical record; and Other: DTI started for suspected HIT and then HIT was ruled out; and DTI started for an indication other than HIT. Statistical Analysis Data are presented for both new HIT cases and those patients with suspected HIT (unconfirmed). Heparin exposure was verified by reviewing the pharmacy database and/or the medical record for every new case of HIT. The number of new HIT cases was divided by the number of patients exposed annually to unfractionated heparin to identify the incidence of HIT at our institution. The percentage of patients with isolated HIT vs HIT with thrombosis (HITTS) was determined. The incidence of HIT was determined for both subcutaneous (SQ) and therapeuticdose IV heparin exposure (defined as any continuous infusion or a therapeutic bolus). The overall patient hospital admission classification (ie, surgical or medical) of new HIT cases was www.chestjournal.org CHEST / 131 / 6/ JUNE, 2007 1645

determined. As orthopedic surgery and cardiovascular surgery patients are often reported as high-risk populations, the incidence of HIT was determined in patients undergoing hip/knee joint arthroplasty and coronary artery bypass/valve replacement surgery. Results A total of 58,814 patient admissions occurred during the study period. From October 2003 through March 2004, 12,034 patients were exposed to a heparin product. Fifty percent of patients receiving heparin were medical patients, 47.8% were surgical/procedural patients; and 2.2% were unable to be classified. SQ heparin (only) was prescribed in 59.7% of patients, IV heparin in 22.5% of patients, other forms of heparin exposure (ie, flush and dialysis) in 5.8% of patients, and more than one form of heparin in 12% of patients. Surgical patients were more likely to receive only SQ heparin compared to medical patients (67.9% vs 51.5%, respectively). Medical patients were more likely to receive IV therapeutic treatment doses of heparin compared to surgical patients (28.6% vs 16.5%, respectively). During the study period, 1,080 patients (9%) were reexposed to a heparin product within 90 days, while an additional 154 patients (1.3%) were reexposed between 90 and 180 days. The number of hospitalized inpatients exposed to heparin on an annual basis was projected at 24,068 or approximately 41% of hospital admissions. DTI therapy was administered to 133 patients; 114 received argatroban, 15 received lepirudin, and 4 patients received both DTIs during the study period. DTI use was categorized as 49 patients with a new diagnosis of HIT (all with antibody confirmation), 15 patients with suspected HIT (unconfirmed), 15 patients with a history of HIT, and 54 patients classified as other. The antibody status of the 15 cases of suspected HIT (unconfirmed) were as follows: negative, 80% (12 of 15 cases); equivocal, 13.3% (2 of 15 cases); and not tested, 6.7% (1 of 15 cases). A hematology/ oncology consultant service followed 87.8% of new HIT patients and 86.7% of those patients with suspected HIT (unconfirmed). Demographic information is provided in Table 1. The overall incidence of HIT in this study was 0.2% (49 new HIT cases per 24,068 patients exposed to heparin). The incidence of new HIT cases in those patients receiving therapeutic-dose IV heparin was 0.76% (41 new cases per 5,415 patients exposed to heparin). The incidence of HIT in patients receiving SQ heparin for thromboprophylaxis with or without flush was 0.1% (6 new HIT cases per 14,368 patients exposed to heparin). HIT developed in two patients as a result of receiving only heparin flushes. Nineteen new HIT patients received both IV and SQ heparin. One patient of the 19 patients was classified as having HIT from receiving SQ heparin as HIT was believed to be present at the time of the initiation of therapy with IV heparin. Sixteen of the 19 patients had clinical pictures consistent with IV heparin being the causative form of heparin. Two patients in the group with HIT resulting from IV heparin were Table 1 Patient Information* Variables New HIT Patients (n 49) Suspected HIT (n 15) All Patients (n 64) Age, yr 72.5 12.5 67.3 14.1 71.3 13.0 Weight, kg 81.0 19.4 77.1 20.4 80.1 19.6 Gender, No. Male 23 10 33 Female 26 5 31 Medical patients, % 30.6 53.5 25 Surgical patients, % 69.4 46.7 75 Antibody positive, % 100 0 76.6 HPF4 OD Overall 1.11 0.65 0.36 0.60 0.94 0.71 Isolated HIT 1.02 0.59 0.22 0.27 0.83 0.63 HITTS 1.25 0.73 0.60 0.97 1.12 0.80 HITTS 20 (40.8) 6 (40) 26 (40.6) Patients with OD 1.0, No. 9 4 13 Patients with OD 1.0, No. 11 1 12 Duration of DTI therapy, h 228 186 198 155 221 178 *Values are given as the mean SD or No. (%), unless otherwise indicated; HITTS HIT with thrombosis. For patients with multiple HPF4 results, the highest OD was used. One patient was not tested. One patient received a single bolus of lepirudin and was not included in calculation. 1646 Original Research

Table 2 HIT in Coronary Artery Bypass Surgery and/or Valve Replacement Patients Variables New HIT Patients* (n 24) Suspected HIT (n 5) All Patients* (n 29) HITTS, No. (%) 7 (29.2) 3 (60) 10 (34.5) Patients with OD 1.0 4 2 6 Patients with 1.0 3 1 4 Heparin use during cardiac catheterization within 30 d of surgery, % 70.8 80 72.4 Preoperative heparin administration, % 54.2 40 51.7 Postoperative heparin administration, % SQ ( 15,000 U/d) 12.5 0 10.3 IV infusion 33.3 20 31.0 Both IV and SQ 8.3 0 6.9 Heparin flush 4.2 0 3.4 *Two patients did not have a cardiac catheterization prior to surgery. One patient had a cardiac catheterization at an outside institution, and the use of heparin was unable to be confirmed. See Table 1 for expansion of abbreviation not used in text. Heparin administration within 30 days prior to surgery excluding heparin used during cardiac catheterization. IV heparin therapy adjusted to maintain a therapeutic clotting time. initially treated with an SQ heparin product (heparin in one patient and dalteparin in another patient). The role of the prior SQ treatment in precipitating HIT, however, was unclear, and these patients were classified as having HIT caused by IV heparin. Twenty-four of the 49 new HIT cases (49%) occurred in patients undergoing coronary artery bypass surgery and/or valve replacement surgery. The majority of HIT cases in cardiac surgery patients were diagnosed postoperatively. Three patients underwent an additional surgical procedure (one patient had abdominal surgery, and two patients had nonjoint replacement orthopedic surgery). The majority of new HIT cases in cardiovascular surgery patients involved either preoperative or postoperative heparin exposure (Table 2). At the time of this study, our institution did not have a standard protocol for thromboprophylaxis in cardiovascular surgery patients. Our institution performed 1,163 coronary artery bypass/valve replacement surgeries during the study period, resulting in an incidence of new HIT of 2.1% in this population. No cases of HIT were identified in patients undergoing hip/knee joint arthroplasty (replacement or revision) even though our institution typically performs 1,700 of these procedures per year. At the time of this study, our primary thromboprophylaxis regimen for orthopedic surgery was adjusted-dose SQ heparin as a bridge to warfarin therapy. An audit of heparin use in 50 randomly selected joint arthroplasty patients during this study period is summarized in Table 3. The combined overall incidence of new HIT and suspected HIT (unconfirmed) in our institution was 0.27% (64 of 24,068 cases). The combined incidence of new HIT and suspected HIT (unconfirmed) from IV heparin exposure was 1.0% (54 of 5,415 patients exposed to heparin), while the combined incidence for SQ heparin with or without flushing was 0.1% (8 of 14,368 patients exposed to heparin) The combined incidence of new HIT and suspected HIT in cardiovascular surgery patients was 2.5% (29 of 1,163 patients). Discussion The American College of Chest Physician Guidelines 1 on HIT report that the risk of HIT is 1 to 5% in patients receiving unfractionated heparin for 1 to 2 weeks. The incidence of HIT varies depending on a number of different factors. Overall, the use of unfractionated heparin is associated with a higher frequency of HIT than is low-molecular-weight heparin. Since unfractionated heparin is the predominant parenteral agent used for inpatient anticoagu- Table 3 Summary of SQ, Adjusted-Dose, Unfractionated Heparin Exposure in 50 Joint Arthroplasty Patients* Variables Values Initial daily dose, U/d Mean SD 11,040 947 Range 9,000 15,000 Concurrent warfarin therapy, % of patients 92 Time from end of surgery to heparin therapy initiation, h Mean SD 6.9 3.0 Range 1.5 11.5 Duration of therapy, h Mean SD 72.7 43.6 Range 24 284 Heparin dose adjusted, % of patients 28 *Adjustments were allowed to maintain the activated partial thromboplastin time between 30 and 40 s (inclusive). Only one patient received 12,000 U/d during heparin therapy. Four of 50 patients received 96 h of heparin therapy. www.chestjournal.org CHEST / 131 / 6/ JUNE, 2007 1647

lation and antithrombotic prophylaxis at our institution, we recognized the need to evaluate the frequency of HIT in our institution. This study retrospectively identified the percentage of patients who had been exposed to heparin and developed new HIT and were subsequently managed with a DTI. Overall, HIT developed in 0.2% of patients exposed to heparin. HIT occurred more frequently in surgical patients compared to medicine patients (69.4% vs 30.6%, respectively). Surgical patients are often considered to be at higher risk for HIT. The incidence of new HIT cases in those patients receiving therapeutic-dose IV heparin was 0.76% vs 0.1% in those receiving SQ heparin therapy. The mean ( SD) OD of the ELISA result for new HIT cases was 1.11 0.65. The ELISA used in this study is capable of detecting antibodies of the IgG class as well as those of the IgM and IgA class. 9 The thrombocytopenia and thrombosis of HIT are thought to be mediated primarily by IgG antibodies. The pathogenicity of IgM and IgA antibodies in patients with HIT is unclear; however, recent data have suggested that these antibodies are much less likely to be pathogenic. 10 Although this may have led to an overestimation of true HIT, we feel that it is likely that cases of HIT were missed. Fifteen cases of suspected HIT that were not confirmed with antibody testing were identified. The overall combined incidence of new HIT and suspected HIT was 0.27%. Approximately 23% of patients receiving treatment with a DTI for HIT either did not have a positive antibody or did not have an antibody test performed. Although this rate is high, there are three likely contributing factors, as follows: (1) an in-house functional assay was not available; (2) clinicians are aware that although the presence of HIT with a negative ELISA result is uncommon, it is possible, so not treating a patient with HIT can have devastating consequences; and (3) the clinician s suspicion of HIT either was too high to confidently rule out HIT and/or was high enough to warrant full treatment with a DTI (Table 1). A hematology/ oncology consultant was following up 86% of the suspected HIT cases. Several recent publications suggest orthopedic surgery patients who receive thromboprophylaxis with heparin are at greater risk for HIT than cardiac surgery patients who receive intraoperative /- postoperative heparin, 5% vs 1 3% respectively. 6,11 13 Data supporting a higher incidence of HIT in the orthopedic surgery population originate from older studies 6,7 in which heparin was administered for 10 to 14 days or until hospital discharge. Unfortunately, recent reviews 13 on HIT have failed to acknowledge that the data used to categorize orthopedic surgery patients as high risk involved longer courses of heparin therapy than are typically administered today. At the time of this study, most hip/knee joint arthroplasty patients at our institution received lowdose SQ heparin ( 15,000 U/d) as a bridge to warfarin therapy. No new cases of HIT were identified in this patient population. We believe the lack of HIT found in the joint arthroplasty patients in our study is a result of a shorter duration of heparin exposure compared to the duration of heparin exposure a decade ago. In the United States, the average hospital length of stay for primary hip and knee replacement surgery has decreased over the last decade and is now 5 days. The average length of stay for the revision of knee replacements is also 5 days, while the mean length of stay for hip revision and partial hip replacement is approximately 5.5 days. 14 Therefore, the duration of heparin exposure for many patients undergoing hip and knee arthroplasty in the United States may be insufficient to illicit the immune response necessary for the development of HIT. In contrast, coronary artery bypass surgery and/or valve replacement surgery patients were found to have an incidence of HIT of 2.1%. At the time of this study, there was not a standardized approach to thromboprophylaxis in this patient population. Pharmacologic prophylaxis with unfractionated heparin was left to physician discretion. The majority of patients had documented heparin exposure either preoperatively or postoperatively. This study took a practical approach to identifying the frequency of HIT in a large hospital in which unfractionated heparin therapy is commonly administered. The overall incidence of recognized HIT was 0.2%, which is well below the commonly cited range of1to5%. 1 Cases of HIT may have been missed as a result of lack of recognition, transfer to hospice before the establishment of a diagnosis, or death prior to the initiation of DTI therapy. The short duration of heparin therapy likely accounted for the absence of HIT in orthopedic patients undergoing joint arthroplasty. The incidence of HIT with IV heparin therapy was 0.76% and is similar to that reported for the treatment of thrombosis. 1 Understanding the safety of heparin therapy in our hospital was an important goal. This study provided useful information for the evaluation of more expensive alternative anticoagulant and antithrombotic therapeutic agents. This study did have a few design limitations. We were unable to consistently capture heparin use for inpatients whose only heparin exposure was during cardiac catheterization or surgery. The identification of a new HIT case required the initiation of DTI therapy to be recognized. The lack of an in-house 1648 Original Research

functional assay was also a limitation and prevented the definitive resolution of 15 cases of suspected HIT (unconfirmed). The overall incidence of recognized HIT in our institution was 0.2%. The low incidence likely reflects a brief duration of heparin exposure for many patients. The incidence was 0.76% in those patients receiving IV treatment doses of heparin and 0.1% in those receiving SQ heparin therapy. Of the surgical patients in whom HIT developed, coronary artery bypass surgery and/or valve replacement surgery patients were at the greatest risk (2.1% of all new HIT cases). Platelet count monitoring in this patient population in highly recommended. The lack of HIT diagnoses in joint arthroplasty patients is in contrast to recent review articles, 13 which continue to identify the orthopedic surgery population as being at high risk. Our data suggest that the incidence of HIT with low-dose SQ unfractionated heparin is very low if the duration of heparin exposure is limited to 5 days. References 1 Warkentin TE, Greinacher A. Heparin induced thrombocytopenia: recognition, treatment and prevention: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004; 126(suppl):311s 337s 2 de Maistre E, Gruel Y, Lasne D. Diagnosis and management of heparin induced thrombocytopenia. Can J Anaesth 2006; 53:S123 S134 3 Davoren A, Aster AH. Heparin induced thrombocytopenia and thrombosis. Am J Hematol 2006; 81:36 44 4 Prandoni P, Siragusa S, Girolami B, et al. The incidence of heparin induced thrombocytopenia in medical patients treated with low molecular weight heparin: a prospective cohort study. Blood 2005; 106:3049 3054 5 Warkentin TE, Sheppard JA, Sigouin CS, et al. Gender imbalance and risk factor interactions in heparin-induced thrombocytopenia. Blood 2006; 108:2937 2941 6 Warkentin TE, Sheppard JI, Horsewood P, et al. Impact of the patient population on the risk for heparin induced thrombocytopenia. Blood 2000; 96:1703 1708 7 Warkentin TE, Levine MN, Hirsh J, et al. Heparin induced thrombocytopenia in patients treated with low molecular weight heparin or unfractionated heparin. N Engl J Med 1995; 332:1330 1335 8 Smythe MA, Koerber JK, Markstrom D, et al. The rate of major bleeding in heparinized patients [abstract]. Pharmacy World Sci 2005; PT135:A131 9 PF4 ELISA Assay for the detection of antibodies directed against platelet factor 4 (PF4): polyvinyl sulfonate (PVS) complex [package insert]. Waukesha, WI: GTI, 2002 10 Eichler JD, Lubenow N, Strobel U 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 heparin induced thrombocytopenia. Eur J Haematol 2006; 76:420 426 11 Greinacher A, Eichler P, Lietz T, et al. Replacement of unfractionated heparin by low molecular weight heparin for post-orthopedic surgery antithrombotic prophylaxis lowers the overall risk of symptomatic thrombosis because of a lower frequency of heparin induced thrombocytopenia. Blood 2005; 106:2921 2922 12 Warkentin TE. New approaches to the diagnosis of heparininduced thrombocytopenia. Chest 2005; 127(suppl):35s 45s 13 Arepally GM, Ortel TL. Clinical practice: heparin-induced thrombocytopenia. N Engl J Med 2006; 355:809 817 14 US Department of Health & Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. National hospital discharge survey 1998 2004. Available at: http://www.cdc.gov/nchs/about/major/hdasd/ nhds.htm. Accessed January 11, 2007 www.chestjournal.org CHEST / 131 / 6/ JUNE, 2007 1649