Thromboprophylaxis Strategies for Hospitalized Patients With COVID-19.

To the Editor:

We read with interest the updated guidelines for thromboprophylaxis in patients with COVID-19 in this issue of CHEST. As stated in this and the previous version, the intent of these guidelines is to provide guidance regarding the optimal thromboprophylaxis strategies for VTE. However, although most of the recommendations align with the available evidence, it is our opinion that the first recommendation regarding the subgroup of moderately ill (noncritical) hospitalized patients does not. Although the risk ratio for VTE was 0.48 in favor of therapeutic anticoagulation (TA), additional meta-analyses of the risk difference for the same data show that, compared with prophylactic anticoagulation, TA was associated with a reduction of only 1.3% (95% CI, 0.6%-2.5%) in the risk of VTE with an associated increase in the risk of major bleeding of 0.9% (95% CI, 0.6%-1.7%). In addition, a meta-analysis of the reported main outcome for each of the four studies shows a risk ratio of 0.8 (95% CI, 0.5-1.2), suggesting no advantage of the intervention across all trials with the caveat that the main outcomes designed for all four studies were not the same. Therefore, the recommendation favoring TA was made on the basis of the reported advantage on organ support-free days for the multiplatform trial only, which amounted to only 4% (95% CI, 0.5%-7.2%). In our opinion, this deviates from the original intent of the guidelines, making their adoption confusing and potentially harmful.

In addition, the generalizability of the findings for the four trials is questionable as, overall, they included only 11% of more than 32,000 screened patients. In particular, the risk of bleeding is significant in these patients as shown by a systematic review and meta-analysis, including more than 18,000 hospitalized patients with COVID-19, reporting a pooled incidence of total and major bleeding of 7.8% and 3.9%, respectively, with the highest pooled incidence estimate among patients receiving intermediate- or full-dose anticoagulation (21.4%). In another cohort study of 1,965 patients receiving either intermediate or therapeutic anticoagulation, major bleeding events occurred in 5.7%. These figures differ from those reported in the four trials included in the guidelines, suggesting again a potential issue of generalizability.

We strongly believe that despite emerging data, TA strategies should not be considered standard of care yet, and if they are used outside of randomized trials, they must be executed very carefully on a case-by-case basis with multidisciplinary input and careful documentation.