Abstract 19160: Impact of Correction for Immortal Time Bias on Mortality After Inferior Vena Cava Filter Placement for Venous Thromboembolic Disease and a Contraindication to Anticoagulation
Introduction: Inferior vena cava filters (IVCF) are recommended in patients with venous thromboembolic disease (VTE) and a contraindication to anticoagulation therapy. Observational studies of IVCF have often failed to adjust for immortal time bias, which is the period of follow-up time before the IVCF is placed during which death cannot occur in the intervention arm but can occur in the control group.
Methods: Hospitalized patients with VTE plus a contraindication to anticoagulation were identified in the State Inpatient Databases from California (2005-2011), Florida (2005-2013), and New York (2005-2012). Propensity score matching was used to adjust for baseline differences in patients receiving IVCF and those who did not and was performed with and without adjustment for immortal time bias via exclusion of matched pairs where the control patient died prior to the day of filter placement in their matched partner. Multivariable Cox models were constructed with IVC filter use as a time-fixed variable in one model and as a time-dependent variable that adjusts for immortal time bias in a separate model.
Results: Of 126,030 VTE patients, propensity score matching identified 40,982 matched pairs of patients with VTE treated with or without IVCF placement. Patients who received an IVCF had a significant reduction in risk of 1-year mortality (HR 0.88; 95% CI:0.86-0.91). In contrast, the same propensity analysis adjusted for immortal time bias showed that IVCF placement was associated with increased risk of mortality (HR 1.05; 95% CI 1.02-1.08). In the Cox model with a time-fixed IVCF variable, patients receiving an IVCF had lower mortality risk (HR 0.88; 95% CI: 0.86-0.91). When IVCF status was analyzed as a time-dependent variable, patients with an IVCF had higher risk of mortality (HR 1.22; 95% CI 1.19-1.25).
Conclusions: The presence of immortal time in observational studies biases results in favor of intervention. Adjustment for immortal time bias by two different methods revealed that IVCF placement is associated with an increased hazard of death in patients with VTE and a contraindication to anticoagulation. Randomized clinical trials are needed to determine the efficacy of IVCF placement in VTE patients with a contraindication to anticoagulation.
Author Disclosures: T.E. Turner: None. M.J. Saeed: None. E. Novak: None. D.L. Brown: None.
- © 2016 by American Heart Association, Inc.