Outcomes of IVC Filter Use in Heparin-Induced Thrombocytopenia: A Retrospective Study

Introduction Heparin-induced thrombocytopenia (HIT) is a prothrombotic state characterized by prior exposure to heparin products and decreased platelets. Inferior vena cava filters (IVCFs) are often placed in patients with deep venous thrombus (DVT) or pulmonary emboli (PE) that are not candidates for anticoagulation. Patients with HIT are in a hypercoagulable state and presumably may have high rates of thrombotic events after IVCF placement. Recommendations against insertion of an IVCF in patients with acute HIT or acute isolated HIT have been supported by a single, small case series (n=10) and a few case reports of IVC-related thrombotic events in patients with IVCF placement either before or during time of HIT diagnosis.¹ However, there have not been any larger studies looking specifically at patients with HIT and IVCF placement and subsequent outcomes.

Methods We retrospectively reviewed patient charts from the Kaiser Permanente Northern California electronic medical record database (over 4 million patients) from 2006-2015 and cross-referenced patients with both an ICD-9 code for IVCF placement and for HIT. Chart review was done on the patients who met both criteria. The patients were divided into two subgroups- those with HIT diagnosis confirmed (SRA positive or HIT ab >1) and those with HIT diagnosis likely (HIT ab <1), as well as sub-groups based on the timing of IVCF placement relative to HIT diagnosis (within 14 days or not). Sub-groups were analyzed for thrombotic events after placement of IVCF, including new or extension of DVT, new or extension of PE, IVC thrombosis, lower extremity phlegmasia cerulea dolens, critical limb ischemia, and mortality.

Results During this ten-year period, 3,934 patients had IVCFs placed and 814 patients were identified with HIT diagnostic codes, with 30 patients meeting criteria for both an ICD-9 code for HIT and IVCF placement. Of these 30 patients, 4 were excluded as they had IVCF placement well after diagnosis of HIT (>14 days). Of these patients (n=26), a total of 4 patients (15.4%) had thrombotic progression/complications, either a single complication or multiple complications. Overall, 11.5% (n=3) had progression of DVT noted either on ultrasound or clinically, 3.9% (n=1) had IVC thrombosis, and 3.9% (n=1) had lower extremity phlegmasia. Of these 26 patients, we further examined those who had an IVCF placed up to 14 days before or after HIT diagnosis (n=17). Of these patients, 11.8% (n=2) had progression of DVT, 5.9% (n=1) had IVC thrombosis, and no patients had lower extremity phlegmasia. There were 2 deaths (7.7%) related to thrombotic events. Six-month all-cause mortality was 23% (n=7) with similar mortality rates in patients who had HIT confirmed vs. HIT likely. Patients who had an IVCF placed within 30 days of HIT diagnosis (27% mortality) and after 30 days of HIT diagnosis (25% mortality) did not have significant difference in six-month all-cause mortality. Our data tracked mortality up to 12 months (27%).

Conclusion Our study found a thrombotic rate of 15.4% in HIT patients with IVCF placement, while a rate of thrombosis after IVCF placement in patients without HIT has been reported between 2-30%.² The rate of new thrombotic events in HIT alone is reported to be 23-35%.³ Our six-month all-cause mortality rate was 26.9%, with mortality directly related to thrombotic events at 7.7%. Six-month mortality associated with HIT without IVCF placement has been reported at 20-30%.⁴ One patient developed lower extremity phlegmasia, which is reported to be rare in the literature. Our results do not show an increased risk of mortality, DVT, PE, IVC thrombosis, or lower extremity phlegmasia in the setting of IVCF placement in patients diagnosed with HIT, as compared to rates of complications in patients with only HIT or in patients with only IVCF placement. While HIT is associated with an increased risk of thrombosis, and a published small case series (n=10) recommended against placing IVCFs in these patients, our larger, retrospective study does not demonstrate increased thrombotic events in this patient population beyond patients with only HIT or only IVCFs.

Full references available upon request

1. Jung, 2011

2. Milovanovic, 2015

3. Bruce, 2003

4. Benjamin, 2016