**4. Diagnosis**

202 Venous Thrombosis – Principles and Practice

Thrombosis occurs in 30% to 70% of cases of HIT, depending on the population, and can occur without the presence of significant thrombocytopenia (Warkentin, 2007). Thrombotic events may occur days prior to the onset of thrombocytopenia (Greinacher et al., 2005). In patients who develop a thrombotic event who are either on or have recently completed heparin therapy, the possibility of HIT should be considered (Levine et al., 2006). In addition, it is important to realize that small amounts of heparin, such as those used to perform heparin

When HIT is associated with thrombosis approximately 20-30% of cases are fatal and an additional 20-30% result in permanent disability (Greinacher, 1995). DVT (50%) and PE (25%) are the most common thrombotic events related to HIT. Arterial thrombosis, infrarenal aortic thrombosis (Karkos et al., 2011), acute myocardial infarct (Iqbal et al., 2007), cerebral ischemia (Meyer-Lindenberg et al., 1997), limb ischemia (Kreidy & Hatem, 2004), acute adrenal insufficiency (Poulain et al., 2008) and bilateral adrenal hemorrhage (Ernest & Fisher, 1991; Rosenberg et al., 2011) are less common. Skin lesions, which may or may not be necrotic, can be seen at the heparin injection site in 10% to 20% of patients who develop HIT (Jang et al., 2005). Skin lesions appear to be more common in patients with higher levels of platelet-activating IgG (Warkentin, 1996). Despite the presence of thrombocytopenia, bleeding complications related to HIT are uncommon (Selleng et al.,

Systemic or anaphylactoid reactions can rapidly occur after an IV bolus of UFH. In addition to a marked decline in platelet count, patients may develop fever, chills, respiratory symptoms that may simulate a pulmonary embolus (Hartman et al., 2006; Popov et al., 1997), cardiac arrest, gastrointestinal symptoms such as nausea, vomiting and diarrhea, or even neurologic symptoms such as ischemia or transient global amnesia (Warkentin et al., 1994b; Warkentin & Greinacher, 2009). It is important to note that these reactions are due to the immune-mediated response to heparin therapy (Warkentin & Greinacher, 2009); additional cases of anaphylactic reactions due to contaminated or over-sulfated heparins also exist (Liu et al., 2009). This type of systemic reaction is illustrated in the clinical vignette

In hemodialysis patients, a unique set of complications may emerge as a result of HIT. For example, there may be clotting of the extracorporeal circuit or failed arteriovenous fistulae. If an IV bolus of UFH or LMWH is given prior to hemodialysis and systemic reactions occur, it is important to consider HIT as a cause (Syed and Reilly, 2009). Increased circuit pressures, formation of a clot in the drip chambers, clotted dialyzer fibers, or an acute thrombocytopenia with at least a 20% decrease in platelet counts may also be suggestive of

Thrombocytopenia is frequently encountered in critically ill patients and can have a variety of etiologies. ICU patients often receive UFH or LMWH for either prophylaxis or treatment of venous thromboembolic disease or for treatment of a variety of other conditions, such as cardiac ischemia or atrial fibrillation. The question of HIT is therefore frequently raised (Sakr, 2011). However, HIT is actually an uncommon cause of thrombocytopenia in this patient population with an incidence of less than 1% (Crowther et al., 2010; Verma et al., 2003). Some more common causes of thrombocytopenia in critically ill patients include

flushes, can also cause significant manifestations of HIT (Refaai et al., 2007).

**3.4 Complications** 

2007).

at the beginning of the chapter.

HIT (Yamamoto et al., 1996).

**3.5 Alternative causes** 

Recent data from the CATCH (Complications After Thrombocytopenia Caused by Heparin) registry suggests that less than 10% of patients who develop thrombocytopenia receive a diagnostic evaluation for HIT (Oliveira et al., 2008). Furthermore, many do not receive diagnostic attention for possible HIT until after a thromboembolic event has occurred (Crespo et al., 2009). Given the potential for significant morbidity and mortality, timely diagnosis is of utmost importance. Diagnosis of HIT should be based on both clinical judgment and laboratory assessment. The patient's presentation and history provide the most important information for initial determination of likelihood of HIT, but verification of anti-heparin/PF4 antibodies through serum or plasma analysis is a significant step for guiding treatment and follow-up.

## **4.1 Clinical diagnosis**

The Four T's (4T) score can be used to calculate the pre-test probability of HIT in patients experiencing signs or symptoms of the disorder (Warkentin & Heddle, 2003). The 4T score takes into account four domains and generates a score of 0–8 points depending on the patients' signs and symptoms; a total score of 0 to 3 indicates that HIT is unlikely; 4 to 5 indicates an intermediate probability; 6 to 8 indicates high likelihood of HIT. In addition to the severity of thrombocytopenia, the score also takes into account the timing of the fall in platelet count, the occurrence of thrombosis or other sequelae, and the presence of other potential causes of the thrombocytopenia. In general, studies have found that the 4T score has high sensitivity for diagnosis (>95%), but specificity is low, especially in ICU patients (Lo et al., 2006; Pouplard et al., 2007). Most authors suggest that those patients with a high likelihood of HIT based on 4T score should be immediately withdrawn from heparin therapy, be treated with an alternative, and be monitored closely while laboratory tests for HIT antibodies are performed. Withdrawal of heparin should not wait until laboratory results are obtained in any patient where HIT is being strongly considered.

Several other sets of criteria and scoring systems have been developed for determining pretest probability of HIT in both general and specific populations (e.g. Messmore et al., 2011). The HIT Expert Probability Score has recently been developed based on a panel of expert opinions (Cuker et al., 2010), but data regarding its psychometric properties are still lacking. Scoring systems for patients on hemodialysis (Yamamoto et al., 1996) have been created owing to the unique factors associated with HIT in this population. It has also been noted that the use of some scoring systems, such as the 4T score, should be modified in the setting of a critically ill patient; in such cases more emphasis should be placed on ruling out other causes of thrombocytopenia (Hall et al., 2010).

Heparin-Induced Thrombocytopenia 205

antibodies, regardless of their ability to activate platelets. Immunologic assays, such as polytypic ELISA, IgG-specific ELISA, and particle gel immunoassay (PGI) have a sensitivity of over 95%, but the specificity is sub-optimal and true positive diagnosis is identified in as few as half of all cases. Due to the pathogenicity of IgG, assays that are specific to this antibody are more likely to yield a true positive diagnosis than the general ELISA assays. Immunologic tests are relatively rapid, producing results within hours. Because of the high sensitivity and the fast turn-around, they are often used to rule-out HIT, but a positive test needs to be interpreted in conjunction with clinical data (Amiral, 1999; Amiral & Vissac, 1999). Mistakenly diagnosing and treating for HIT based on positive assay in the setting of a low pre-test probability score could result in serious consequences, such as venous limb

Functional assay tests, such as serotonin release assay (SRA), flow cytometric detection and heparin-induced platelet aggregation (HIPA) assay, have a slightly lower sensitivity (approximately 90%), but a much better specificity (>90%) than the immunologic assays. Therefore, the overall accuracy of the test for correctly identifying HIT is high (Warkentin, 2002). However, these tests are technically more demanding and are only offered at a

The HemosIL AcuStar HIT-IgG (specific for IgG anti-PF4/heparin antibodies) and the HemosIL AcuStar HIT-Ab (for detecting IgG, IgA and IgM anti-PF4/heparin antibodies) are newer laboratory tests available for diagnosing HIT. These two semi-quantitative chemiluminescent immunoassays provide results in 30 minutes and can be run for single sample testing. Legnani and colleagues (2010) reported on initial data showing a 100% sensitivity and negative predictive value for these tests, making them ideal for ruling out HIT. Specificity for the HIT-IgG test was 96.5% and for the HIT-Ab test was 81.2%. These tests, especially the IgG specific test, may gain more use in the future depending on results

A recent article advocates the use of a colorimetric test to detect HIT (Prechel et al., 2011). The test uses a tetrazolium-based indicator dye that reacts to the activity of platelets. When contact is made with inactivated or mildly activated platelets, the dye metabolizes to a dark color. When the platelets are highly activated (e.g. due to the presence of HIT antibodies) the dye is unable to metabolize and remains light in color. Preliminary analyses have shown this platelet activation assay test to have between 96% and 100% agreement with the

In general, pairing the findings of a clinical diagnostic score with the results of a laboratory test significantly increases the sensitivity and specificity of the diagnostic yield (Demma et

When HIT is suspected or proven, all heparin and LMWH should be discontinued. A careful investigation for heparin exposure from sources such as catheter flushes and hemodialysis is necessary. Even the low doses used to flush a catheter can lead to worsening thrombocytopenia and a systemic reaction in patients with HIT, as in the case presented above. Once heparin therapy is discontinued, platelet count should begin to rise within three days, though this is dependent on the amount of antibodies present in the system

minority of laboratories. Therefore, they are not relevant in most clinical settings.

gangrene or fatal hemorrhage (Smythe et al., 2011).

of further studies.

**5. Treatment** 

functional assay C-SRA for HIT diagnosis.

al., 2011; Ruf et al., 2011; Kim et al., 2011).

(Seleng et al., 2007; Kelton, 2002).


Table 1. Non-Chemotherapeutic Agents other than Heparin and Low Molecular Weight Heparin that may cause Thrombocytopenia in the Hospitalized Patient

#### **4.2 Laboratory diagnosis**

Both functional and immunologic assay tests are available for the detection of HIT antibodies. Functional tests work by detecting antibodies that induce heparin-dependant platelet activation whereas immunologic tests detect circulating anti-heparin/PF4

Antibiotics

 Vancomycin Penicillins Cephalosporins Ciprofloxacin Clarithromycin Pentamidine Daptomycin Meropenem Rifampin Linezolid Nitrofurantoin Ganciclovir

Trimethoprim/sulfamethoxazole

Abciximab Eptifibatide

Clopidogrel Ticlopidine

Phenytoin Carbamazepine

Procainamide Amiodarone

 Salicylates Acetaminophen Nonsteroidal Anti-Inflammatory Drugs (NSAIDS)

 Furosemide Thiazides Spironolactone Acetazolamide

Ranitidine Cimetidine

Table 1. Non-Chemotherapeutic Agents other than Heparin and Low Molecular Weight

Both functional and immunologic assay tests are available for the detection of HIT antibodies. Functional tests work by detecting antibodies that induce heparin-dependant platelet activation whereas immunologic tests detect circulating anti-heparin/PF4

Heparin that may cause Thrombocytopenia in the Hospitalized Patient

Glycoprotein IIb/IIIa inhibitors

Tirofiban

Thienopyridines

Antiepileptics

Valproate

Antiarrhythmics

Diuretics

Captopril Digoxin Morphine Haloperidol

Quinidines

Anti-Inflammatory Drugs

Histamine 2 Receptor Blockers

Sulfonylurea Drugs

**4.2 Laboratory diagnosis** 

Fluconazole Valganciclovir hydrochloride

antibodies, regardless of their ability to activate platelets. Immunologic assays, such as polytypic ELISA, IgG-specific ELISA, and particle gel immunoassay (PGI) have a sensitivity of over 95%, but the specificity is sub-optimal and true positive diagnosis is identified in as few as half of all cases. Due to the pathogenicity of IgG, assays that are specific to this antibody are more likely to yield a true positive diagnosis than the general ELISA assays. Immunologic tests are relatively rapid, producing results within hours. Because of the high sensitivity and the fast turn-around, they are often used to rule-out HIT, but a positive test needs to be interpreted in conjunction with clinical data (Amiral, 1999; Amiral & Vissac, 1999). Mistakenly diagnosing and treating for HIT based on positive assay in the setting of a low pre-test probability score could result in serious consequences, such as venous limb gangrene or fatal hemorrhage (Smythe et al., 2011).

Functional assay tests, such as serotonin release assay (SRA), flow cytometric detection and heparin-induced platelet aggregation (HIPA) assay, have a slightly lower sensitivity (approximately 90%), but a much better specificity (>90%) than the immunologic assays. Therefore, the overall accuracy of the test for correctly identifying HIT is high (Warkentin, 2002). However, these tests are technically more demanding and are only offered at a minority of laboratories. Therefore, they are not relevant in most clinical settings.

The HemosIL AcuStar HIT-IgG (specific for IgG anti-PF4/heparin antibodies) and the HemosIL AcuStar HIT-Ab (for detecting IgG, IgA and IgM anti-PF4/heparin antibodies) are newer laboratory tests available for diagnosing HIT. These two semi-quantitative chemiluminescent immunoassays provide results in 30 minutes and can be run for single sample testing. Legnani and colleagues (2010) reported on initial data showing a 100% sensitivity and negative predictive value for these tests, making them ideal for ruling out HIT. Specificity for the HIT-IgG test was 96.5% and for the HIT-Ab test was 81.2%. These tests, especially the IgG specific test, may gain more use in the future depending on results of further studies.

A recent article advocates the use of a colorimetric test to detect HIT (Prechel et al., 2011). The test uses a tetrazolium-based indicator dye that reacts to the activity of platelets. When contact is made with inactivated or mildly activated platelets, the dye metabolizes to a dark color. When the platelets are highly activated (e.g. due to the presence of HIT antibodies) the dye is unable to metabolize and remains light in color. Preliminary analyses have shown this platelet activation assay test to have between 96% and 100% agreement with the functional assay C-SRA for HIT diagnosis.

In general, pairing the findings of a clinical diagnostic score with the results of a laboratory test significantly increases the sensitivity and specificity of the diagnostic yield (Demma et al., 2011; Ruf et al., 2011; Kim et al., 2011).
