Commentary: Widely Reported Study of Therapeutic Anticoagulation for COVID-19 Has Serious Flaws
Key findings
- A recently published research letter reports substantially longer in-hospital survival among mechanically ventilated patients with COVID-19 who received therapeutic anticoagulation than those who did not
- Limitations of the study include confounding by indication, additional unmeasured confounders and, most importantly, immortal time bias
- Many randomized, controlled trials of anticoagulation strategies are underway and will help determine the utility of this treatment
Subscribe to the latest updates from Pulmonary & Critical Care Advances in Motion
Multiple investigators have reported high rates of thrombotic events in patients with severe COVID-19. Even in the absence of documented thrombosis, therapeutic-dose anticoagulation has garnered attention as a potential treatment for COVID-19.
A research letter published online in the Journal of the American College of Cardiology reports an association between therapeutic anticoagulation and in-hospital survival among mechanically ventilated patients with COVID-19. The results were widely circulated in the lay press, with some journalists presenting them as evidence of a causal link between anticoagulation and improved survival.
In a commentary in Research and Practice in Thrombosis and Haemostasis, Jason H. Maley, MD, a clinician and research fellow in the Department of Medicine, and Corey Hardin, MD, PhD, physician in the Division of Pulmonary and Critical Care Medicine at Massachusetts General Hospital, and colleagues note several flaws in the report that must be considered when interpreting the data.
Study Summary
In the retrospective cohort study, 28% of 2,773 with COVID-19 patients received some form of systemic anticoagulation during their hospital stay. The key findings were:
- Mortality was nearly identical among patients who received therapeutic anticoagulation (22.5%) and those who did not (22.8%)
- Patients who received therapeutic anticoagulation were more likely than those who did not to require invasive mechanical ventilation (30% vs. 8%; P < .001)
- In-hospital mortality was 63% (median survival, 9 days) among mechanically ventilated patients who did not receive anticoagulation and 29% (median survival, 21 days) among those who did
Confounding by Indication
Anticoagulation was administered in this study based on a clinical indication. Patients who did not receive anticoagulation may not have had an indication or they may have had a contraindication.
Unfortunately, the research letter does not provide any baseline patient characteristics or any subgroup characteristics for mechanically ventilated patients who did and did not receive anticoagulation. Neither does it include information about the indications for anticoagulation in this cohort, such as institutional protocols or diagnoses of thrombosis.
Therefore, readers have no way to understand why some patients were treated and some were not, so they cannot assess the magnitude of the confounding by indication.
Additional Confounders
Reported mortality in the group that received anticoagulation is similar to mortality in other cohorts of critically ill patients with COVID-19, but mortality in the "no anticoagulation" group is strikingly higher. This suggests differences between the two groups beyond their indication for anticoagulation. Additional unmeasured confounders probably preclude the correct effect estimates.
Immortal Time Bias
The most serious flaw of the study is that immortal time bias certainly contributed to the treatment effect. Immortal time is the interval in an observational study between the time a patient enters a cohort (in this study, at hospital admission) and the time they receive the exposure of interest.
During the interval between admission and anticoagulation, death could not occur in the anticoagulation group because those patients had to survive long enough to receive treatment—they were "immortal" between admission and treatment. Deaths that occurred during immortal time could be attributed only to the "no anticoagulation" group.
The median time from admission to start of anticoagulation was two days (IQR, 0–5 days), so 50% of the treated cohort had two days of immortal time and 25% had at least five days. Once patients received anticoagulation, they became part of the anticoagulation group, and anticoagulation "gets the credit" for their survival prior to treatment despite playing no role in that survival.
Trial Data Awaited
It's understandable that the COVID-19 pandemic has produced a sense of urgency to report novel treatments, but data must be viewed critically, especially those from observational studies. Many randomized, controlled trials of anticoagulation strategies are underway and will help determine the utility of this treatment.
view original journal article Subscription may be required
View all COVID-19 updates
Learn more about research in the Division of Pulmonary & Critical Care