The current findings support the use of immunosuppressive therapy in moderate to severe COVID-19 using a combination of corticosteroids and a JAK1/2 inhibitor. All patients in this study also received remdesivir and dexamethasone in addition to baricitinib, so it is not possible to assess the individual contribution to recovery of each medication. This study spans a 6-month period during which SARS-CoV2 variants emerged in Vermont, USA.
Four patients required mechanical ventilation (8.9%) during their hospitalization. Of these, two received ventilator support prior to starting baricitinib therapy, and one of these was extubated shortly after initiating baricitinib. This contrasts with the average higher mechanical ventilation rate of 14.5% from a meta-analysis of 12,437 COVID-19 ICU admissions (24), and was considerably lower than our own experience in the early phases of the pandemic. The patient population in the current study was equally distributed between males and females, in contrast to many other studies. In addition, half were age 70 or older, which makes the favorable outcome all the more remarkable. Finally, there were six deaths (13.3%) among the 45 patients. Two of these, however, were complicated by sub-massive pulmonary emboli present at the time of admission, prior to initiation of baricitinib.
This study sought to assess the efficacy and safety of 7 days of baricitinib treatment, whereas other recent trials in COVID-19 patients treated for 14 days (16–19). The decision for a shorter treatment period was made out of a desire to balance the suppression of inflammation that might result in tissue damage with an avoidance of prolonged immunosuppression that might delay viral clearance or promote secondary infections. Delays in viral clearance have in fact been observed in other immunocompromised patients, resulting in the emergence of viral variants (25). Additionally, given the known risk of JAK inhibitor-induced thrombosis, in the context of the recognized coagulopathy risk in COVID-19, a shorter treatment course may be favorable and sufficient for the duration of cytokine release syndrome in these patients.
Half of the patients with moderate to severe COVID-19 had a BMI greater than 30. This is considerably higher than the 23.2% obesity prevalence for the general population in Vermont (https://www.cdc.gov/obesity/stateprograms/fundedstates/pdf/vermont-state-profile.pdf). Obesity is a known risk factor for severe COVID-19 infection (26). Obesity is also associated with a baseline inflammatory state (27). Adipose tissue supports the development of tissue resident T lymphocytes that upregulate gene expression for several inflammatory cytokines as well as for cytolytic activity, and express high levels of the checkpoint blocker programmed cell death protein-1 (PD-1) (28). A very similar phenotype of T cells is observed in bronchiolar lavage fluid of COVID-19 patients (29). Additionally our retrospective review echoes prior work by the Center for Disease Control that older age is associated with increased risk of hospitalization and poorer prognosis with COVID-19 infection.
No adverse effects were noted from use of baricitinib. In particular, there were no secondary infections. Despite concern for increased thrombotic risk with baricitinib, we did not observe clinical evidence of new clots during the brief course of baricitinib treatment, although two patients demonstrated significant clots on admission prior to initiation of baricitinib.
SARS-CoV-2 is known to suppress the initial IFN-I response, likely through the interaction of particular viral proteins with molecules of the IFN-I signaling pathway (30, 31). This allows the virus to rapidly replicate during the early stages of infection. The delayed immune response can then become hyperactive and result in considerable cell death of surrounding tissues. This could include tissues that are not known to support SARS-CoV-2 replication, such as liver inflammation observed in some cases of severe COVID-19 (32). The subsequent release of host RNA and DNA from damaged tissues can strongly activate, respectively, the retinoic acid-inducible gene 1 (RIG-I) and cyclic GMP-AMP synthase (cGAS) nucleic acid sensing pathways, leading to an augmented IFN-I response and persistent inflammation even in the absence of virus. This is consistent with studies showing that death of lung epithelium is due in some instances more to the immune response than to viral-mediated lysis (33). Emerging evidence in animal models of SARS and MERS has revealed that the initial IFN-I response has beneficial effects in the early phases of disease, but may become damaging in the latter phases (34).
Severe COVID-19 has close parallels with other seemingly unrelated syndromes that might collectively be classified as hyperinflammatory disorders. Chimeric antigen receptor T (CAR-T) cell therapy exposes patients to a large number of T cells that become activated upon contact with targeted tumor cells, often resulting in a highly inflammatory cytokine release syndrome that can include hypercoagulation and even acute respiratory distress syndrome (ARDS) (35–37). Toxic shock syndrome is a multiorgan inflammatory syndrome (38) in which tampons infected with Staphylococcus release an enterotoxin that acts as a superantigen by binding both the MHC class II molecule and the β-chain of several T cell receptors (39). This activates a significant portion of the T cell repertoire, similar to CAR-T therapy, resulting in injury to many organs including skin, liver, and lung, and can also be associated with coagulopathy and ARDS (38). Consistent with the view of hyperactivation of T cells in these disorders, individuals with HIV and low T cell counts have been noted to have less severe COVID-19 (40).
An additional parallel can be made between severe COVID-19 and hemophagocytic lymphohistiocytosis (HLH). HLH is a severe inflammatory syndrome characterized by fever, hepatitis, spleen and lymph node enlargement, and pancytopenia (41, 42). It is often observed secondary to certain viral infections as well as autoimmune syndromes such as juvenile inflammatory arthritis (41). An additional laboratory characteristic is elevated ferritin, which we observed in our severe COVID-19 cases. HLH is likely the result of strong T cell activation producing cytokines that activate macrophages to become highly phagocytic (41, 42). Consequently, anti-cytokine therapy has also been used to treat HLH, including IL-1 blockade as well as JAK inhibitors.
In other case series of patients with COVID-19, baricitinib treatment was associated with both an improvement in oxygenation and a reduction in select inflammatory markers (16–19). The largest of these, the ACTT-2 Study Group, randomized 1033 patients to receive remdesivir and either baricitinib for up to 14 days (515) or placebo control (518). Patients receiving baricitinib had a median time to recovery of 7 days compared to 8 days for the control group, and a 30% higher odds of improvement in clinical status at day 15. Patients receiving high-flow oxygen or noninvasive ventilation at enrollment had a time to recovery of 10 days with combination treatment and 18 days with control. The 28-day mortality was 5.1% in the combination group and 7.8% in the control group.
Limitations of the current study include its retrospective nature, the study site was a single center, and the study did not allow for comparison with untreated controls.