Thread by @nickmmark

Retrospective series of n=4,389 pts w/ #COVID19 @mountSinai in @JACCJournals comparing prophylactic #anticoagulation (AC) vs therapeutic AC vs no AC:
*Prophylactic AC reduced mortality compared to no AC *Therapeutic AC no better than prophylactic AC
➡️ bit.ly/3gBpvmo
1/

The *unadjusted* mortality is quite different:
*Prophylactic AC 21.6%
*Therapeutic AC 28.6%
*No AC 25.6%

But one of the challenges of a retrospective study like this is there are usually significant differences between the groups, as we can see in Table 1:
2/

They use a logistic regression model to adjust mortality for differences in age, comorbidities, labs, vitals, etc

This is reasonable but you must be careful: it can overcorrect, residual confounders can persist, & it provides many opportunities to find a model that you like
3/

They argue that after adjusting mortality for potential biases there is a trend towards benefit w/ therapeutic AC “even though not statistically significant.”

Even with a very large n & adjusting for risk factors they could not find a significant benefit to therapeutic AC.
4/

Therapeutic AC also increased the risk of bleeding (3% vs 1.7%) compared to Ppx AC.

Overall, I’m pretty skeptical that Tx AC is better than PPx AC for hospitalized pts w/ COVID.

There may be subgroups that benefit from Tx dose AC, but this paper doesn't tell me who they are.
5/

There are currently dozen trials of AC in COVID listed on clinicaltrials.gov, testing prophylactic, treatment, & intermediate doses of AC.

Hopefully, these RCTs will elucidate if therapeutic AC is beneficial in specific subgroups of COVID pts.
➡️ bit.ly/34GAJUo
6/

🥡Points:
*COVID patients appear to do better with prophylactic AC than with no AC (this is already standard of care)
*even after adjusting for severity of illness, therapeutic AC is not superior to prophylactic AC
*I'm not going to use therapeutically AC based on this study

7/7