Ironically, although thrombosis was the first manifestation of the HIT syndrome, first recognized almost 50 yr ago (Weismann and Tobin, 1958), widespread recognition that thrombosis was a common complication of HIT did not occur until recently. Indeed, until 1995 no study of HIT had compared the frequency of thrombosis with a matched control population (Warkentin et al., 1995). This study quantitated the strength of the association between HIT and thrombosis and further noted that the more unusual the thrombotic event (e.g., bilateral deep venous thrombosis, pulmonary embolism), the stronger the association with HIT (see Chapter 2).
Table 6 summarizes the thrombotic events that have been observed during prospective studies of HIT. The major observation is that thrombosis is relatively common in HIT patients, occurring in approximately one-third of medical patients and about one half of postoperative surgical patients. The data also support findings from a prior retrospective study (Boshkov et al., 1993) that found the type of thrombotic event complicating HIT was influenced by the patient population. Table 6 suggests that the ratio of arterial to venous thrombosis is about 1: 1 in medical patients, many of whom might have had arterial disease as their basis for hospitalization. Additionally, the therapeutic-dose heparin used in many of these studies may have partially protected against VTE, although it may not have prevented platelet-mediated arterial occlusion. In contrast, there appears to be a strong predisposition to VTE in postoperative orthopedic patients who have developed HIT (venous:arterial ratio at least 14:1) (Table 6).
The retrospective identification of patients with serologically confirmed HIT permits analysis of large groups of HIT patients (Table 7). This provides an alternative assessment of the spectrum of thrombotic complications in HIT. Three large studies (Warkentin and Kelton, 1996; Nand et al., 1997; Wallis et al., 1999) showed a predominance of venous thrombosis complicating HIT. Indeed, pulmonary embolism was even more frequent than all types of arterial thromboses combined.
In contrast, a different spectrum of thrombotic complications was reported by investigators at the University of Missouri-Columbia Health Sciences Center (Silver et al., 1983; Laster et al., 1987; Almeida et al., 1998b). Arterial, rather than venous, thromboembolism predominates in these patient series. Because this work is from the perspective of a vascular surgery service, it is possible that patients with arterial thrombosis are either more likely to be recognized as having HIT, or greater numbers of patients with preexisting arteriopathy are treated with heparin and thus at higher risk for developing arterial thrombosis if HIT develops.
Another pattern that emerges from the Missouri series is a progressively decreasing frequency of reported thrombotic or hemorrhagic complications, from 61% in 1983, to 23% in 1987, then to 7.4% in 1998. The authors believe this to be the result of earlier recognition of HIT. However, an alternative explanation could be greater awareness of HIT over time, and thus a higher likelihood of identifying patients with less severe HIT. Indeed, a study by Wallis and colleagues (1999) suggests that earlier recognition of HIT may not reduce the risk of thrombosis (Table 7).
A progressive reduction in HIT-associated mortality over time was also observed by the Missouri group (Table 7). However, early discontinuation of heparin was not associated with significantly lower mortality in another study (Wallis et al., 1999). This issue is complicated by the observation that deaths apparently unrelated to thrombosis are relatively common in patients with HIT (Warkentin and Kelton, 1996; Greinacher et al., 1999).
It is possible that nonthrombotic mortality may be higher than expected by chance in patients with HIT. This speculation is based on the observation that only a minority of patients who form anti-PF4/heparin antibodies develop HIT (discussed subsequently); a corollary to this statement is that comorbid factors that tend to result in increased pathogenicity of heparin-dependent antibodies may also independently contribute to increased patient morbidity and mortality (i.e., patients with septicemia or multisystem organ failure may be more likely to have platelet activation in the presence of HIT antibodies than "well" patients). Alternatively, because the patients develop thrombocytopenia they are tested for heparin-dependent antibodies, and non-pathogenic antibodies are detected.
Was this article helpful?