High-dose glucocorticoids in COVID-19 patients with acute encephalopathy: clinical and imaging findings in a retrospective cohort study

During the pandemic, acute encephalopathy (AE), in its most frequent clinical manifestation, delirium, affected up to 65% of hospitalized patients with COVID-19 (Helms et al. 2020) and was associated with a higher mortality (Mendes et al. 2021). The risk of developing delirium in COVID-19 was greater in older patients, in presence of a more severe form of the disease, and of a preexisting cognitive impairment (Wilke et al. 2022; Damanti et al. 2023). Several factors likely contribute to AE pathogenesis in COVID-19, with suggested mechanisms including exposure to intensive care unit procedures, a prothrombotic, microvascular dysfunction, and proinflammatory state (Pensato et al. 2021). Indeed, inflammation at both peripheral and central level has been associated with delirium (Bernard-Valnet et al. 2021; Fajgenbaum and June 2020) and a case series observed in the Geneva University Hospitals reported a good response to high-dose intravenous GC (Pugin et al. 2020). Based on these findings, in 2020, the interdisciplinary medical board of the Geneva University Hospitals approved an experimental protocol to identify COVID-19 patients with AE and, under specific conditions, treat them with high-doses of GC.

Here, we report the results of a retrospective clinical study, by describing clinical features of the cohort of COVID-19 AE patients who were included in a protocol evaluating the eligibility to a high-dose GC treatment. Within this group, we assessed differences in clinical and imaging features between patients who were treated by GC and those who did not receive treatment. We also evaluated whether imaging features indicative of inflammation (endotheliitis) or microangiopathy [cerebral small vessel disease (CSVD)] were associated with AE severity and clinical outcome at discharge.

This study was performed during the acute phase of the pandemic, as a follow-up to a previous study done in our center (Pugin et al. 2020). We report here the clinical and imaging features of COVID-19 patients with AE who were eligible to a treatment with high-dose intravenous GC, and compare clinical and imaging features with a non-treated group with spontaneous signs of remission. Our findings reveal that patients who were treated with high-dose GC had a similar disability at discharge, when compared to NT, despite signs of clinical and radiological vulnerability, as the lack of spontaneous improvement, longer hospitalization, longer AE duration, and higher HTNA score. Moreover, delirium severity was related to radiological features of microvascular damage but not of inflammation.

High-dose GC have been used to successfully treat patients with COVID-19 AE, with an improvement in the 48–72 h after administration (Pugin et al. 2020, Pizzato Tondo et al. 2021), and other neurological manifestations COVID-19 related (Pilotto et al. 2020). Indeed, an inflammatory process, involving the central nervous system in a direct or indirect way, through a cytokine storm, has been implied in the pathogenesis of neurological manifestations in COVID-19, including delirium. This has been corroborated by radiological and pathological findings showing signs of endotheliitis (Uginet et al. 2022b, Varga et al. 2020) and inflammatory syndromes involving the central nervous system (Uginet et al. 2022b, Paterson et al. 2021).

Based on these premises, following the case series (Pugin et al. 2020) done at our hospital and as a last resort given the lack of signs of clinical improvement, we decided to treat COVID-19 patients with AE with high-dose GC when eligible, based on specific criteria (Fig. 1). Our findings show that, though showing several signs of severity and lack of spontaneous remission, the GT rapidly improved after high-dose GC administration and showed similar functional outcome at discharge when compared to less severe AE patients not requiring high-dose GC treatment. Moreover, mortality rate was similar between groups. Both these observations lead us to suggest that GC treatment was justified in those patients and support inflammation as an etiological hypothesis in COVID-19 AE. This finding is corroborated by other more recent studies, showing that the use of high-dose GC in conditions associated with delirium can help to prevent it (Awada et al. 2022; Xiang et al. 2022) or to reduce its severity (Kluger et al. 2021). Further studies to properly define high-dose GC in preventing or treating delirium in populations at risk are needed.

In terms of brain imaging, we assessed whether imaging features of inflammation or microangiopathy were correlated to clinical severity of COVID-19 AE. The number of enhanced vessels was not related to the severity or clinical outcome of our patients, confirming our previous observation (Uginet et al. 2021). Possibly, vessels enhancement does not completely capture the inflammation process associated with COVID-19 AE, and more advanced imaging features reflecting parenchymal inflammation could better explain severity and predict outcome in COVID-19 AE.

Imaging features suggestive of CSVD, as microbleeds, WMH, and lacunes, have been described in COVID-19 patients and they have been associated with an increased mortality, critical illness, and worse functional outcome (Agarwal et al. 2020). The high rate of microbleeds observed in patients with COVID-19 could be a direct consequence of critical illness (Kirschenbaum et al. 2021). However, we found that the ICU admission rate was not different between the two groups and that patients admitted to ICU did not present with a higher rate of microbleeds. Thus, microbleeds together with WMH, lacunes, and EPVS could represent markers of preexisting CSVD. Indeed, mechanisms other than inflammation are likely to be involved in the severity of COVID-19 AE and brain endothelial damage has been widely described in patients with COVID-19. In our cohort, patients with a higher HTNA score had a more severe form of encephalopathy, independently of GC treatment. Thus, we can speculate that a preexisting frailty at a microvascular level can predispose to AE severity. The quantification of microvascular damage by use of imaging could shed light into the pathogenesis of COVID-19 AE, especially as the link between endothelial damage and neurovascular brain changes seen in hypertensive arteriopathy has already been linked to delirium in the postoperative phase, as well as cognitive decline in previous studies (Kant et al. 2017). In our entire cohort, AE clinical severity in the acute phase was also associated with a higher WMH load, especially in the periventricular regions, as well as deep EPVS. Presence of white matter hyperintensities, as well as EPVS, is a marker of CSVD and is associated with neurodegeneration and cognitive impairment (Matsuda et al. 2021). Therefore, these findings support the hypothesis that patients with imaging markers of an underlying neurodegenerative process were more prone to develop a more severe form of COVID-19 AE.

This raises the questions of an underlying vulnerability for severe AE in patients with recognized or unrecognized neurodegenerative conditions, but also for possible long-term effects of COVID-19 infection (Bommarito et al. 2023). Systemic inflammation, acute respiratory distress syndrome, and sepsis are known factors to contribute to long-term development of neurological consequences and cognitive decline, either accelerating an ongoing neurodegenerative process or being the onset of a new pathological process (Iwashyna et al. 2010). Future studies should investigate the correlation between COVID-19 AE and possible higher prevalence of cognitive decline in COVID-19 AE survivors (Heneka et al. 2020).

Although this sample of COVID-19 patients with brain imaging during the acute stage of their AE, as well as the thorough adherence to protocol regarding the selection of patients for treatment are a strength of this study, we should acknowledge some limitations. First, the small sample size of our cohort, due to an insufficient number of patients presenting with COVID-19 AE and responding to the inclusion criteria, limits the power of the study and thus may limit the generalization of the study’s findings. Second, the retrospective design of the study brings some limitations, including the incomplete inclusion of confounding factors that could have influenced AE severity, quantitative MRI data, and the absence of optimal control group(s). Specifically, as the data refer to the acute phase of the pandemic, first and second wave, we miss clinical and laboratory data, as a comprehensive list of precipitating factors for AE other than COVID-19, that could have helped to better explain AE duration and outcome. Moreover, we lack a control group of COVID-19 patients without AE which would enable us to specify if the white matter hyperintensities and enlarged perivascular spaces patterns described here are pathogenic to COVID-19 AE or not. We also lack a control group of AE patients without COVID-19, so we cannot conclude that our findings are specific to COVID-19 infection. Finally, the decision to reevaluate patients at 48 h specifically after the first neurological consult was based on experts’ opinions and “clinical improvement at 48 h” was a qualitative appreciation based on the clinical expertise of two physicians (one neurologist and one critical care physician), which could lead to biases because of the lack of quantitative criteria for improvement. A randomized-controlled trial, with a larger sample size, would allow us to confirm these preliminary observations.

This study suggests that severe COVID-19 AE patients treated with high-dose GC showed a favorable clinical outcome at discharge, similar to patients with spontaneous improvement, suggesting high-dose GC benefit as an acute treatment in some COVID-19 AE patients. White matter changes as well as other signs of CSVD both contribute to severity and clinical outcome of AE in COVID-19. Such findings can be useful in identifying patients at high risk of developing AE in COVID-19 or other predisposing conditions, and to refine treatment protocols.