Increasing evidence supported the possible neuro-invasion potential of SARS-CoV-2. However, no studies were conducted to explore the existence of the micro-structural changes in the central nervous system after infection. We aimed to identify the existence of potential brain micro-structural changes related to SARS-CoV-2.
In this prospective study, diffusion tensor imaging (DTI) and 3D high-resolution T1WI sequences were acquired in 60 recovered COVID-19 patients (56.67% male; age: 44.10 ± 16.00) and 39 age- and sex-matched non-COVID-19 controls (56.41% male; age: 45.88 ± 13.90). Registered fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) were quantified for DTI, and an index score system was introduced. Regional volumes derived from Voxel-based Morphometry (VBM) and DTI metrics were compared using analysis of covariance (ANCOVA). Two sample t-test and Spearman correlation were conducted to assess the relationships among imaging indices, index scores and clinical information.
In this follow-up stage, neurological symptoms were presented in 55% COVID-19 patients. COVID-19 patients had statistically significantly higher bilateral gray matter volumes (GMV) in olfactory cortices, hippocampi, insulas, left Rolandic operculum, left Heschl's gyrus and right cingulate gyrus and a general decline of MD, AD, RD accompanied with an increase of FA in white matter, especially AD in the right CR, EC and SFF, and MD in SFF compared with non-COVID-19 volunteers (corrected p value <0.05). Global GMV, GMVs in left Rolandic operculum, right cingulate, bilateral hippocampi, left Heschl's gyrus, and Global MD of WM were found to correlate with memory loss (p value <0.05). GMVs in the right cingulate gyrus and left hippocampus were related to smell loss (p value <0.05). MD-GM score, global GMV, and GMV in right cingulate gyrus were correlated with LDH level (p value <0.05).
Study findings revealed possible disruption to micro-structural and functional brain integrity in the recovery stages of COVID-19, suggesting the long-term consequences of SARS-CoV-2.
In this prospective work, we recorded the detailed cerebral volumetric, DTI metrics 3 months after SARS-CoV-2 infection by applying DTI and 3D T1WI in 60 recovered COVID-19 patients and 39 age- and sex-matched normal volunteers. Overall, these recovered COVID-19 patients were more likely to have enlarged olfactory cortices, hippocampi, insulas, Heschl's gyrus, Rolandic operculum and cingulate gyrus, and a general decline of MD, AD, RD accompanied with an increase of FA in white matter, especially AD in the right CR, EC and SFF, and MD in SFF compared with non-COVID-19 volunteers. Global GMV, GMVs in left Rolandic operculum, right cingulate, bilateral hippocampus, left Heschl's gyrus, and Global MD of WM were found to correlate with memory loss. GMVs in right cingulate gyrus and left hippocampus were related to smell loss. MD-GM score, global GMV, and GMV in right cingulate gyrus were correlated with LDH level.
Given the olfactory and gustatory dysfunction in COVID-19 patients and evidence of olfactory epithelium invasion by SARS-CoV, the olfactory gyrus was thought to be the first functional area in CNS to be infected by SARS-CoV-2
To avoid cross-infection, patients were unable to undertake MRI scans during the acute phase, but we are still curious about whether any micro-structural changes exist in the recovery stage of COVID-19 and whether any clues were left to suggest the probable intracranial infection route.
According to our results, significant enlarged volumes were observed in the bilateral olfactory cortices, hippocampi, insulas, left Heschl's gyrus, left Rolandic operculum and right cingulate gyrus. All these structures mentioned above belonged to the central olfactory system. Among them, olfactory cortex, also named as piriform cortex, directly receives axonal projections from olfactory bulb (OB), is referred to as a part of ‘primary olfactory cortex’. Other structures are the cortical targets of primary olfactory cortex in the bilateral limbic lobe, temporal cortices, which were referred as ‘secondary olfactory cortex’. It was reported that the frequent olfactory loss during the course of upper respiratory tract infections (URTI) resulted in loss of stimulation and subsequent volume loss in the acute phase, while after olfactory recovery, the volumes of GM in the central olfactory system got enlarged subsequently. We analyzed the GMV difference between patients with or without olfactory loss and found the GMVs of the central olfactory system were generally smaller in patients with persistent olfactory loss compared with those without olfactory problems which was aligned with previous similar studies (Supplementary Table 2).
Several possible invasion routes of SARS-CoV-2 were raised including hematogenous, lymphatic and neuro retrograde routes, etc., yet the exact route was unknown. Based on our findings, the gray matter volumetric changes in the central olfactory system provided a speculation that SARS-CoV-2 might enter the CNS via an OB-mediated neuronal retrograde route. Two reasons were suspected to play a role in these GMV enlargement: the neurogenesis and functional compensation. Firstly, patients were suspected to experience neurogenesis. It is well accepted that neurogenesis in adults are restricted to two regions, the subventricular zone (SVZ) and the sub-granular layer of the dentate gyrus of the hippocampus. The neuroblasts from SVZ migrate along the rostral migratory stream, enter olfactory cortex first and finally replace interneurons (e.g., periglomerular cells, granular cells) in the OB. Therefore, the increasing neurons possibly resulted in the enlargement of the GMV in the olfactory system. Secondly, in order to compensate for impaired olfaction, the increased functional engagement of brain regions would be hypertrophy, which was proved to have enlarged neurons and increasing number of dendritic spines by experimental studies in sensory deprivation models. These two reasons might be an explanation of the enlarged GMV, while further pathological exploration was needed.
In general, lower diffusivity parameters (MD, AD, RD) and higher FA values were recognized in the white matter from COVID-19 cohort. In addition, the diffusivity (MD and AD) of right CR, EC and SFF decreased significantly. The CR, consisting massive bundle of projection fibers, connects the cortex to the brainstem and the thalamus in afferent and efferent manners. The EC and SFF are series of association fibers, connecting frontal, parietal, and temporal cortices. White matter was not the main target for neurotropic virus; however, the connecting fibers could act as the channel for intracranial viral transmission. Different from the elevated MD with compressed volume in fibers in hydrocephalus, the enlargement of white matter fibers, decreased MD values and elevated FA suggested a greater alignment of fibers and limited diffusion freedom, indicating a possible intrinsic re-construction (e.g. remyelination) process that occurred after infection.
It was interesting to notice that all the diffusivity abnormalities in the white matter restricted in the right hemisphere, without asymmetrical symptoms reported by COVID-19 patients. Since the blood flow rate and volume was higher in the hemisphere dominant for handedness, we hypothesized that the predominance of abnormal white matter diffusivities might be related to the difference of blood volume in bilateral hemispheres. However, after comparing these abnormal diffusivity indices between lefthanded and righthanded patients, no difference was found (Supplementary Table 1). The asymmetrical phenomenon was detected in other studies as well. The right-side predominance in odorant perception has been indicated by multiple olfactory functional studies which was not fully understood. The diffusivity changes on the right side might be related to the right-side odorant perception in which further exploration was needed.During SARS-CoV-2 infection, 41/60 (68.33%) patients had neurological symptoms and over 50% recovered patients still had symptoms 3 months later. It was interesting to find the GMV in hippocampi (a key part in the organization of memory) and cingulate gyri (an important part of limbic system) were negatively related to loss of smell during infection and loss of memory 3 month later, which could support our hypothesis of neurogenesis in these regions mentioned above. Tremor was found to be negatively related to FA_WM score both in the acute stage and at the 3-month follow-up point which indicated a destruction of WM fibers in both hemispheres, possibly resulted from SARS-CoV-2 induced cytokine storm.
According to WHO guideline, COVID-19 patients were divided into mild, severe and critical types based on their clinical information and laboratory results. No significant difference was observed between severe and non-severe groups. But the clinical types were positively related to MD values in bilateral cingulate gyri (r=0.271 and 0.272, p=0.035 and 0.036 respectively), implying the more severe the case was, the higher the MD value of bilateral cingulate gyri was presented. Cingulate gyrus usually plays an important role in attention, motivation, decision making, learning, and cost-benefit calculation, as well as conflict and error monitoring, which is frequently affected in limbic encephalitis. Therefore, the dysregulated cytokine response was speculated in severe cases.
After exploring the relationship between laboratory data and DTI metrics, the global GMV was significantly but slightly correlated with the LDH concentration in COVID-19 patients. LDH is one of the key enzymes in the glycolytic pathway, highly expressed in cells from kidney, heart, liver and brain. Elevated concentrations of LDH are observed in patients with encephalitis, ischemic stroke and head injuries. Higher concentration of serum LDH always follows tissue breakdown and is closely linked to the deterioration and poor outcome. The decreased global GMV in LDH-elevated patients might indicate an atrophy due to a severe inflammatory response.
Since the hemostatic abnormalities, including disseminated intravascular coagulation (DIC) and severe inflammatory response, were frequently observed in COVID-19 patients, individuals may predispose to cerebral-vascular events caused by infection and treatment. An index score system was introduced into our study to investigate the existence of micro-structural abnormalities due to micro-vessel diseases. The ischemic changes are known to be accompanied with lower FA value and higher MD value in the ischemic lesions. However, our findings showed that FA-WM score was higher, MD-GM and MD-WM scores were lower in the COVID-19 group compared with the control group, and the differences were insignificant. Thus, no obvious evidence of micro-vessel diseases was found.
It is important to investigate the relationship between abnormal anatomical brain areas and ACE-2 distribution. It is clarified that SARS-CoV-2 enters the host cell by attaching with ACE-2 via Spike (S) glycoprotein. Therefore, the more expression of ACE-2 might bring more severe abnormalities. The distribution of ACE2 was non-equivalent over the brain and was most frequently expressed in substantia nigra, followed by spinal cord, hippocampus, basal ganglia, limbic system and frontal cortex [17 ]. Our results suggested that various components in the limbic system were affected structures sharing possible high ACE-2 expression, which were partly aligned with the proposed ACE-2-riched regions. Although we were not able to observe the DTI metrics in substantia nigra since it was not included in the brain atlas we used, it was still very hard to support any relationship between ACE2 expression and affected brain areas.
At the time of writing, there was still no research to detect the cerebral micro-structural changes after SARS-CoV-2 infection from imaging or pathological aspects. Our study gave a hint to possible neurological changes after SARS-CoV-2 infection. The limitations of our study were listed as follows: 1) we did not enroll enough patients with neurological dysfunction or olfactory loss, therefore the relationship between GMV/diffusivity changes and olfactory symptoms would be missed; 2) as a single-centered study, a selection bias might result from limited ethnical and regional characteristics of the participants, and possible mutants of SARS-CoV-2 in other countries, and limit the generalization of the study; 3) the atlas we applied did not contain the structure in brainstem, therefore, we failed to obtain the volumetric and DTI information about the nucleus in brainstem, some of which were quite important, especially the solitary nuclei.
In this prospective study, volumetric and micro-structural abnormalities were detected mainly in the central olfactory cortices, partial white matter in the right hemisphere from recovered COVID-19 patients, providing new evidence to the neurological damage of SARS-CoV-2. The abnormalities in these brain areas might cause long-term burden to COVID-19 patients after recovery, which was thus worth public attention.
Reference & Source information: https://www.thelancet.com/
Read More on: