Brain Metabolic PET Findings on the Long-Term Effects of COVID-19

Abstract

TO THE EDITOR: We would like to thank Meyer et al. for their impressive systematic review on brain PET and SPECT findings on the acute and long-term effects of coronavirus disease 2019 (COVID-19) (1). We anticipate that this article will constitute an important reference for this topic, especially for long COVID, also by identifying roadmap points for further studies. Moreover, we believe that the authors' recommendations are reasonable and correspond to our own practice, namely, possible use of PET/SPECT: (a) for differential diagnosis in selected cases after clinical evaluation within the framework of existing authorizations and recommendations (2), particularly for encephalitis and neurodegenerative diseases; and (b) when neurologic disorders have persisted several months, or-in the event of worsening-for cerebral assessment of such patients after a clinical evaluation to confirm such impairments, which we believe cannot be limited to the cognitive domain and consequently only to neuropsychiatric testing (e.g., dysautonomia).We would like to clarify several points concerning our previous publications, on which recommendations for long COVID are partly based.As accurately highlighted by our colleagues (1), the inclusion criteria of time spans from initial infection have fluctuated in our studies (3,4), albeit in accordance with the fluctuations of the French and international definitions of the condition (3 wk, 1 mo, and now 3 mo). The definition used at the time of publication was justified in our articles (3,4). We fully recognize the possible impact of this delay on PET findings and the need for further standardized studies based on the current clinical definition of long COVID. In this line, we recently showed, in a multicentric study including 143 patients, a consensual profile of brain hypometabolism on visual interpretation for approximately one half of patients with suspected neurologic long COVID approximately 11 mo after symptom onset, whereas the second half of patients had normal brain PET metabolism (5). We also agree that recommendations for the clinical use of PET imaging must take into account a delay of confirmed persistent symptoms (.3-6 mo for Meyer et al. ( 1)).Meyer et al. suggested that our PET results were unjustifiably obtained with 2 distinct statistical thresholds in the 2 studies ("P , 0.05, FWE-corrected [familywise-error-corrected] in adults; P , 0.001, uncorrected in children") (1). The same statistical thresholds were in fact used for the 2 studies (3,4). The reader can refer to the methods and Table 2 of the 2 studies (P [voxel] , 0.001; P [cluster] , 0.05, familywise-error-corrected) (3,4).Meyer et al. mentioned that we reported a "weak" negative association between the number of complaints and the PET metabolism of the brain stem and cerebellum ("r 2 5 0.1 and 0.34, respectively") (1). Similarly, the reader can refer to the results of our study: the r 2