Fibrinolysis Shutdown and Thrombosis in a COVID-19 ICU.

Abstract

To the Editor: Thank you for your interest in our work and correspondence. We appreciate some of the confusion that the term “fibrinolytic shutdown” may evoke and would like to address some of the misunderstanding regarding its use. We also recognize that viscoelastic testing has limitations, but would caution against dismissing a potentially useful diagnostic signal in a widely available test in favor of less established methods. The concept of diminished activity of the fibrinolytic system evolved from work more than six decades ago. Lakner and Merskey reported decreased fibrinolytic activity as measured by the euglobulin and blood clot lysis times in patients who suffered a myocardial infarction in 1960 (1). This was expanded upon by other investigators in other patient populations including trauma patients. The work pointed to a common theme of some patients having a biphasic fibrinolytic response to injury—initially being increased, but then followed by greatly diminished activity for some period of time. Chakrabarti et al. (2) are largely credited with the specific phrase “fibrinolytic shutdown,” which was mentioned in their 1969 paper looking at the phenomenon in patients undergoing surgery, electroconvulsive therapy, and post myocardial infarction. Specifically, they demonstrated the euglobulin lysis time, a marker of fibrinolysis where longer times indicate less activity, peaked on day 4 after the myocardial infarction and then diminished to control levels over the next 10 days. It is therefore an incorrect assertion that “the term fibrinolytic shutdown suggests there is no fibrinolytic activity whatsoever.” What the term does refer to is a phenotype of patients that have evidence of prior fibrinolytic activation with subsequent low current activity. This is excellently illustrated in the recent review on the subject by Moore et al. (3) It is well known that COVID-19 patients have extremely elevated levels of d-dimers, and the cohort of patients in our study was no exception. It is certainly correct to point out that endogenous fibrinolytic activity was likely very high at some point. However, the typical half-life of d-dimers is 8 h, and can be longer in patients with compromised renal function and in the elderly (4). Therefore, it would be expected that patients experiencing fibrinolytic shutdown would still have high levels of this biomarker present despite decreased plasmin activity. Our findings were consistent with Ibanez et al. (5), who also found diminished fibrinolytic activity on ROTEM (Instrumentation Laboratories, Bedford, Mass) despite high d-dimer levels. To be clear, high d-dimer levels do not indicate that fibrinolysis is currently occurring, but rather that fibrinolysis had taken place at some earlier time point. As suggested, the main source of fibrinolysis may be the lungs, but definitive evidence of this is currently lacking. We disagree that viscoelastic tests cannot be used to identify impaired fibrinolytic activity. Moore et al. have produced a large body of literature on this subject using thromboelastography (Haemonetics, Boston, Mass). We admit identifying hypofibrinolysis is less well-defined using rotational thromboelastometry (e.g. ROTEM) and does represent one of the mentioned limitations in our report (6). However, our defined cut-off for fibrinolytic shutdown of a maximum lysis < 3.5% came from Gomez-Builes et al. who derived this value from 550 injured patients. While there are likely undefined differences between trauma patients and COVID-19 patients, the fact that within our cohort, 8 of the 11 patients experiencing fibrinolytic shutdown suffered a venous thrombotic event would suggest there is some value in this definition (6). With respect to diagnostics more ideal to assess hypofibrinolysis, we utilized the tools readily available to our clinical ICU teams. Plasma-based clot lysis assays such as the euglobulin lysis time, which takes anywhere from 4 to 6 h to perform, were not among them. We do not perform non-standard viscoelastic testing outside of a research setting at our institution, so cannot comment on the clinical usefulness of adding Thrombolytics to viscoelastic testing for COVID-19 patients. We look forward to further research in this area.