Intraoperative anticoagulation and limb amputations in patients with immune heparin-induced thrombocytopenia who require vascular surgery



    1. Departments of Pathology and Molecular Medicine
    2. Medicine, McMaster University, Hamilton, ON
    Search for more papers by this author
    • These authors are co-first authors and contributed equally to this report.

  • M. PAI,

    1. Departments of Pathology and Molecular Medicine
    2. Medicine, McMaster University, Hamilton, ON
    Search for more papers by this author
    • These authors are co-first authors and contributed equally to this report.

  • R. J. COOK

    1. Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, ON, Canada
    Search for more papers by this author

Theodore (Ted) E. Warkentin, Hamilton Regional Laboratory Medicine Program, Hamilton Health Sciences (Hamilton General Site), Room 1-270B, 237 Barton St. E., Hamilton, Ontario L8L 2X2 Canada.
Tel.: +1 905 527 0271 (ext. 46139); fax: +1 905 577 1421.

Immune heparin-induced thrombocytopenia (HIT) is characterized by an increased risk of thrombosis; HIT-associated limb artery thrombosis is one potential explanation for limb amputation, a complication that occurs in 5–15% of patients with HIT [1]. We are unaware of previous studies that have evaluated limb salvage and amputation outcomes in patients who have undergone vascular surgery in the setting of acute HIT or subacute HIT, where the latter denotes platelet count recovery, but persistence of platelet-activating antibodies.

We reviewed the medical records of patients over a 30-year period (from 1981 to 2010) at two hospitals (Hamilton General Hospital; McMaster University Medical Centre) with a diagnosis of HIT (>50% platelet count fall with proximate heparin exposure) who underwent vascular surgery at the time they tested positive for heparin-dependent platelet-activating antibodies by the serotonin-release assay (SRA). In most cases, vascular surgery was undertaken for acute arterial thrombosis. We classified patients into two groups, based upon their receiving intraoperative anticoagulation either with (i) unfractionated heparin (UFH) or (ii) any alternative non-UFH strategy. One patient contributed to both UFH and non-UFH groups, because the first vascular surgery was performed with UFH, but the second vascular surgery 33 days later (because of an infected graft, rather than because of rethrombosis) was performed using argatroban anticoagulation. Limb amputation (or amputation inevitability with intervening death) was the outcome of interest. An odds ratio and associated 95% confidence interval were calculated to measure the association between the type of intraoperative anticoagulation and limb amputation. This study was approved by the McMaster University Faculty of Health Sciences Research Ethics Board.

Thirteen patients were identified (mean age, 70 years; eight males) who underwent vascular surgery in the setting of HIT (mean platelet count nadir, 63 × 109 L−1; mean serotonin-release, 92% [range, 72–100%]). Baseline characteristics of patients receiving UFH and non-UFH were, respectively: age (72 vs. 55 years); female sex (40% vs. 25%); atrial fibrillation (50% vs. 0%); hypertension (40% vs. 25%); diabetes mellitus (0% vs. 50%); claudication (30% vs. 25%); active cancer (10% vs. 0%); prior myocardial infarction (20% vs. 0%); and prior stroke (10% vs. 0%). Ten patients received intraoperative UFH, and two (20%) of these required limb amputation. Four patients (including the aforementioned patient who underwent a second operation) received intraoperative anticoagulation with a non-UFH strategy (saline flushes, n = 1; danaparoid, n = 1; lepirudin, n = 1; argatroban, n = 1); three (75%) of these patients required limb amputation. The odds ratio for limb loss using UFH vs. non-UFH for intraoperative anticoagulation was 0.105 (exact 95% confidence interval, 0.001–2.054); another exact analysis stratified by presence of atrial fibrillation yielded an estimate of 0.120 (exact 95% CI, 0.001, 3.256).

Four representative case summaries are presented (the years are indicated as certain historical treatment approaches, e.g. ancrod, are no longer used) (Fig. 1).

Figure 1.

 Four patients who underwent vascular surgery at a time when HIT antibodies were present: cases A and B were performed using UFH for intraoperative anticoagulation, whereas case C was performed using danaparoid. For case D, the first vascular surgery was performed using UFH, and the second procedure was performed using argatroban. See text for further information. AAA, abdominal aortic aneurysm; AKA, above-knee amputation; DVT, deep vein thrombosis; fem, femoral; i.v., intravenous; LMWH, low-molecular-weight heparin; pop, popliteal; sc, subcutaneous; superf., superficial; UFH, unfractionated heparin.

A (1992): A 52-year-old man underwent appendectomy 2 weeks earlier at another institution where he was exposed to heparin. He presented with acute limb ischemia, but with a normal platelet count, and underwent embolectomy using UFH. The diagnosis of HIT was suspected post-surgery based upon a progressive platelet count fall to a nadir of 73 × 109 L−1. He subsequently received ancrod/warfarin anticoagulation, with venogram-documented deep vein thrombosis. He recovered uneventfully (no amputation).

B (1993): A 70-year-old woman underwent aorto-bifemoral bypass and left profundoplasty, and received postoperative UFH anticoagulation until day 6. She returned to hospital on postoperative day 13 with acute limb ischemia; emergency femoral thrombectomy and graft revision was performed with pre- and intraoperative UFH (total dose 12 000 U). Postoperatively, HIT was diagnosed, and although limb perfusion was excellent, ancrod was administered for antithrombotic prophylaxis. This was complicated by microvascular ischemia with severe pain and blistering, beginning a few hours after starting ancrod. Therapy was switched to danaparoid, and the patient had complete healing (no limb amputation). (This patient is case #3 in the 12-patient series describing the syndrome of ‘delayed-onset HIT’ [2].)

C (1995): A 64-year-old man with peripheral arterial occlusive disease and acute right femoral artery occlusion underwent emergency thrombectomy. Unfortunately, limb necrosis necessitated amputation on postoperative day 6. The next day, the patient evinced necrotizing skins lesion at the UFH injection sites; after one injection of low-molecular-weight heparin was given, the platelet count measured 96 × 109 L−1, and HIT was diagnosed. Danaparoid anticoagulation was given, but acute left femoral arterial thrombosis occurred, and the patient underwent emergency left femoral artery thrombectomy with danaparoid anticoagulation. Unfortunately, progressive ischemic necrosis necessitated left above-knee amputation 1 week later.

D (2009). A 75-year-old male with peripheral arterial occlusive disease underwent endovascular repair of abdominal aortic aneurysm; 26 days later he was readmitted with acute limb ischemia, and he underwent thrombectomy with UFH anticoagulation. As with case A, HIT was diagnosed postoperatively when abrupt thrombocytopenia occurred during/following vascular surgery. The patient had viable limbs, but the development of graft infection necessitated repeat vascular surgery and because of persisting SRA-positive status, surgery was performed with intraoperative argatroban anticoagulation; unfortunately, intraoperative thrombosis occurred (judged by surgeons as ‘argatroban failure’), and the patient developed bilateral limb gangrene and death.

The most surprising finding of our study was that 8 of 10 (80%) patients who received UFH for intraoperative anticoagulation during vascular surgery in the setting of HIT had successful outcomes, without the need for amputation; in contrast, three of four (75%) patients who received an alternative non-heparin anticoagulant strategy, such as danaparoid, lepirudin or argatroban, did require amputations. The 95% confidence interval of the odds ratio for the risk of amputation for a heparin vs. a non-heparin anticoagulant strategy crossed 1, indicating that this difference was not statistically significant. It should be noted that other factors relevant to limb loss were present in the patients who required limb amputation, for example warfarin use with acute HIT (UFH case) and technical failure to establish outflow (lepirudin case). Interestingly, four of five patients with atrial fibrillation did not require limb amputation (suggesting that presumed cardiac embolization might be associated with more favourable outcomes), whereas three of four patients with peripheral arterial occlusive disease ultimately required amputations (suggesting underlying arterial vasculopathy might predispose to unfavourable outcomes).

Our study’s findings require cautious interpretation. This is an observational study of a relatively small number of patients. Because patients were neither randomized nor matched, the baseline characteristics of patients who received UFH vs. non-UFH anticoagulation were not entirely similar. Additional factors besides the choice of intraoperative anticoagulation may have affected the risk of limb loss. However, our observations clearly show that UFH administration to acute HIT patients with SRA-positive status does not necessarily lead to adverse outcomes (8/10 had successful outcomes). Indeed, heparin has a paradoxical role in any patient with HIT: its use may increase platelet activation (through potentiation of heparin-dependent platelet-activating antibodies) but may also help achieve satisfactory anticoagulation. Whether heparin administration has a ‘net’ positive or negative effect could vary among patients. In emergency situations of acute vascular occlusion where thrombocytopenia suggests the possibility of HIT, but there is not sufficient time to perform diagnostic testing, it might still be reasonable to perform emergency vascular surgery with UFH anticoagulation.

Finally, two other points merit attention. First, 10 (77%) of the 13 patients in this study received UFH for intraoperative anticoagulation in the setting of acute HIT. Of these, seven received intraoperative anticoagulation with UFH at a time when HIT-associated platelet count decline had occurred, and HIT could potentially have been suspected; thus, the use of UFH in this setting presumably indicates that a diagnosis of HIT had not been reached prior to surgery. Nevertheless, six of these seven patients underwent surgery successfully – without the need for limb amputation – in spite of the intraoperative use of UFH. Second, even in 2011, there are no established protocols to guide intraoperative anticoagulation for a patient with acute HIT who requires vascular surgery, nor are any non-heparin anticoagulants approved for this indication; thus, even if the diagnosis of HIT is strongly suspected, it is not clear what the optimal choice of intraoperative anticoagulant (or its dosing) should be. This area merits further research, in order to establish evidence-based practise guidelines.


Supported by the Heart and Stroke Foundation of Ontario (operating grant #T6950).

Disclosure of Conflict of Interests

T.E. Warkentin has received lecture honoraria from GlaxoSmithKline, Pfizer Canada and Sanofi-Aventis, has provided consulting services to, and/or has received research funding from, Canyon Pharmaceuticals, DiaMed, Gen-Probe GTI Diagnostics, GlaxoSmithKline and Paringenix, and has provided expert witness testimony relating to HIT. M. Pai and R. J. Cook report no relevant conflicts of interest.