In March 2019, OrthoBuzz covered a JBJS study by Rudasill et al. that found a progressively increasing risk of bleeding requiring transfusion among total knee arthroplasty (TKA) patients who had a preoperative International Normalized Ratio (INR) >1. (INR is a standardized measure of how long it takes blood to clot—the higher the number, the longer the clotting time.) These authors also found a significantly increased risk of infection in TKA patients with INR >1.5. and an increased risk of mortality within 30 days of surgery among those with an INR >1.25 to 1.5.
In the January 2, 2020 issue of JBJS, the same team of researchers report findings from a similarly designed NSQIP-based study of patients undergoing total hip arthroplasty (THA). The authors evaluated data from >17,500 patients who underwent a primary THA between 2005 and 2016 and who also had an INR value documented within 2 days prior to joint replacement. Rudasill et al. stratified these patients into 4 groups based on preoperative INRs: ≤1, >1 to <1.25, 1.25 to <1.5, and ≥1.5).
After adjustment, the authors found a significant, independent effect between increased preoperative INR and increased bleeding requiring transfusion and mortality. Specifically, bleeding risk became evident at INR ≥1.25, and patients with INR ≥1.5 were at a significantly increased risk of mortality. The length of hospital stay also increased significantly as INR class increased.
The authors suggest that “current INR targeting [INR <1.5 for elective orthopaedic surgery] may not be strict enough to minimize adverse outcomes for patients undergoing primary total hip arthroplasty.” While admitting that these findings are not likely to change the day-to-day practice of orthopaedic surgeons, the authors say they “may influence preoperative risk stratification for those patients with elevated INR.”
Big data is increasingly important in identifying real-world outcomes of an intervention or risk factor beyond the Phase III clinical trials, with their restricted and highly defined study population. The ACS National Surgical Quality Improvement Program (NSQIP) database offers a unique opportunity to review outcomes during post-marketing surveillance and to identify new phenomena and challenge common assumptions. A highly relevant and practical use of NSQIP is its Surgical Risk Calculator (https://riskcalculator.facs.org/RiskCalculator/), which can help with surgical risk assessment and discussions during the consent process as well as decision making during patient optimisation.
Rudasill et al. use the same NSQIP database to challenge the conventional idea that a target INR of <1.5 is adequate to reduce risks from “warfarinisation.” They concluded that "preoperative INR of 1.25 to <1.5 and INR [greater than or equal to] 1.5 were associated with significantly increased bleeding necessitating transfusion following primary total hip arthroplasty, compared with patients with INR [less than 1.0].”
There are significant flaws related to this study. Specifically, the authors fail to account for the patient who requires bridging anticoagulation when ceasing warfarin dosing to allow a much faster drug washout time before surgery. Bridging often entails use of low molecular weight heparin (LMWH) within 12 to 24 hours of surgery or unfractionated heparin (UH) infusion within 6 hours.
Both LMWH and UH can increase INR and could result in readings close to INR of 1.5 during bridging anticoagulation, even when warfarin is no longer bioavailable in the body. This may confound an assumption that INR of >1.0 means the prevailing warfarin in blood is the cause of this result.
Why does it matter what is causing INR >1.0? It is possible that LMWH and UH may have far higher activity in soft tissues than is reflected by INR results (particularly when relying on the presumed drug half-life for washout of anticoagulation activity) and hence result in higher-than-expected chances of bleeding and associated complications. Another consideration: patients who require bridging anticoagulation often have high CHADS2 scores or a high risk of recurrent thromboembolism (including venous thrombosis and pulmonary embolism), which could have contributed to the adverse outcomes in this study.
The study of INR in surgical patients is far more complex than assumed, and due diligence should be made when formulating future research.