The fourth week, high-dose methylprednisolone was initiated because of signs of organising pneumonia (a late-phase complication of severe COVID-19) on a chest CT scan (figure 1), after which his respiratory status improved. Isoflurane inhalation was started in order to wash out the residual sedation and low-dose diazepam and methadone were given to prevent withdrawal symptoms. The prednisolone dose was quickly reduced to limit neurological side effects. Sodium and urea had nearly normalised and his liver function was not impaired. However, our patient remained encephalopathic.

In the fifth week, magnetic resonance imaging (MRI) of the brain showed very mild periventricular leucoaraiosis though, interestingly, no specific white matter lesions matching COVID-19 encephalopathy (figure 2). A second opinion from a neuro-intensivist from another hospital followed, offering no new insights or alterations in therapy. On day 35 the patient suddenly showed signs of rapid improvement: he was completely weaned from ventilation in one day and he was first able to follow motor commands the following day. After removing the tracheostomy tube, and before discharge from the ICU, his confusion and agitation persisted. The psychotropic drugs were gradually lowered as far as possible.

During his stay on the pulmonary medicine ward he recovered from his critical illness polyneuropathy and was able to walk without support. Two months after hospital admission, the patient showed signs of neurological improvement and he was able to communicate adequately. He was no longer confused or agitated when he was discharged to a medical rehabilitation clinic. After two weeks, he was able to go home with an impaired short-term memory as the only remaining neurological issue.

Discussion
This case is unusual; there were no brain imaging abnormalities in a patient with severe and prolonged encephalopathy, where pulmonary recovery from COVID-19 preceded neurological recovery by a full month. This may suggest that the severity of the encephalopathy might not only be explained by the severity of the COVID-19 pneumonia and there might be an inflammatory neurological response. Multiple factors, including hypoxaemia, metabolic disequilibrium, residual sedation, sepsis, delirium and multi-organ failure may have contributed to the prolonged encephalopathy. However, COVID-19 seems to be the major contributing factor to the encephalopathy, since it remained after successfully treating other causes. Furthermore, the continuing need for mechanical ventilation due to impaired consciousness suggests that it was not solely prolonged delirium causing the impaired consciousness. A post-anoxic cause of encephalopathy seems unlikely, considering many COVID-19 patients presenting with very severe (‘happy’) hypoxia do not regularly develop encephalopathy. COVID-19 encephalopathy, however, remains a diagnosis of exclusion and cannot be properly distinguished from other causes of encephalopathy, exclusively based on symptoms.

Another key point that distinguishes this patient from other cases is that the severity and duration of the clinical symptoms did not correspond with the lack of MRI abnormalities (figure 2). Frequently, patients who have COVID-19 encephalopathy display cortical or subcortical white matter hyperintensity on neuroimaging. They may also show microbleeds or leukoencephalopathy.[10] Interestingly, our patient did not show these imaging abnormalities, hence the severity of the clinical symptoms did not correspond with the normal brain imaging details. However, Uginet et al. reported that a small minority (8%) of patients with COVID-19 encephalopathy do not show MRI abnormalities and 15% of the patients with COVID-19 encephalopathy lack intracranial vessel gadolinium enhancement, similar to our case.[11] These findings suggest that a minority of patients with COVID-19 are encephalopathic without displaying MRI abnormalities. Furthermore, it should be taken into account that the number of patients with MRI-negative COVID-19 encephalopathy is possibly underestimated. Patients hospitalised with COVID-19 are not systematically screened by a neurologist, but evaluated only on indication by their treating physicians: patients with subtle signs may be missed. An earlier case report described a useful role of brain perfusion single-photon emission computed tomography (SPECT)/CT in the evaluation of COVID-19 associated with nonspecific neurological manifestations. [12] In our case, not all currently available imaging techniques were used to exclude other neurological pathologies. Brain perfusion scintigraphy is one such imaging technique and may be particularly considered in patients for diagnostic work-up.

A limitation of this case report is that neither a lumbar puncture nor EEG were performed; consequently, meningitis or epilepsy as causes of encephalopathy cannot be completely ruled out. However, these causes were deemed unlikely in the absence of fever, seizures, meningism or nystagmus.

Our patient still showed signs of impaired short-term memory at the time of discharge from the rehabilitation clinic. This is consistent with the outcomes described in earlier case reports concerning neurological prognosis of post-COVID-19 encephalopathy. Heneka et al. suggest that the survivor is at risk of developing long-term neurological consequences, either by aggravating a pre-existing neurological disorder or by initiating a new disorder. This concern is supported by findings that show that at the time of discharge one third of patients have evidence of cognitive impairment and motor deficits.[13]

With this case report, we present an example of late neurological recovery in COVID-19; in this patient prolonged treatment was necessary.More extensive research of COVID-19 encephalopathy is necessary, to give care providers, patients and family clear indications for treatment and information on prognosis.

Conclusions
The late neurological recovery suggests that the severity of encephalopathy may not only be related to the severity of the COVID-19 pneumonia. Multiple factors, including hypoxaemia, metabolic disequilibrium, residual sedation, sepsis and multi-organ failure possibly contributed to the prolonged encephalopathy. However, COVID-19 infection may also exert a severe influence on the central nervous system without showing imaging abnormalities. The aim of this case report is to give an example of late neurological recovery in COVID-19 in a patient in whom prolonged treatment was indicated.

Disclosures
Written informed consent was obtained from the patient for publication of this case report and any accompanying images. All authors declare no conflict of interest. No funding or financial support was received.