Acute disseminated encephalomyelitis (ADEM) is a rare immune-mediated inflammatory demyelinating disorder of the central nervous system (CNS) (1, 2) that was first described in a patient after smallpox infection 250 years ago (3). Although it occurs at all ages, because it is commonly preceded by viral infections or vaccinations, ADEM is more common in children than in adults (2). With the ongoing COVID-19 pandemic and subsequent vaccination use, the incidence of ADEM may increase (4).

Early and accurate diagnosis is important to start prompt treatment and improve outcomes. Due to the lack of specific biomarkers, clinical features play a vital role in the diagnosis of ADEM (5). Other demyelinating disorders, such as multiple sclerosis (MS), may be indistinguishable from ADEM at initial presentation (6). Recently, the International Pediatric Multiple Sclerosis Study Group (IPMSSG) proposed a consensus definition of ADEM for children (7). Due to discrepancies in clinical features between children and adults, this consensus definition may not be suitable for diagnosis in adults. To date, no specific diagnostic criteria have been established for adults.

Recently, several studies have focused on the clinical features of ADEM in adults and found that adult patients had worse outcomes than children (8, 9). However, most of these studies were single-centre studies with small sample sizes. Worldwide data on the clinical features of ADEM in adults are still inconclusive. A systematic evaluation using an evidence-based approach is urgent. To better characterize this rare clinical entity, we conducted a systematic meta-analysis to investigate the clinical features and outcomes of ADEM in adults.

The lack of available discrete biomarkers remains a challenge in the diagnosis of ADEM. Currently, the diagnostic criteria of the International Paediatric Multiple Sclerosis Society Group (IPMSSG), which are evaluated by clinical and neuroimaging features, have been partly successful in diagnosing paediatric ADEM. Without alternative criteria for the adult population, the IPMSSG criteria have also been applied in adults in recent studies (28). However, the clinical and imaging profile of ADEM differs substantially according to age (29). Applying the IPMSSG criteria to all adults diagnosed with ADEM leaves more than half of cases without a diagnosis (18, 20). To guide clinical diagnosis and interventions for ADEM adults, we tried to combine our main findings of clinical features and the IPMSSG criteria to propose a summary of the key observations in ADEM adults ( Table 6 ).

Among 437 ADEM adults included in 12 studies, ADEM could occur in any age group, with a mean age at onset of 37.1 years. The ratio of males to females with ADEM changes with advancing age. A male predominance has been shown in most studies of paediatric cohorts (30, 31). For adults, we found that the ratio of females was higher than that of males, which is consistent with other inflammatory demyelinating diseases, such as multiple sclerosis (MS) (32).

ADEM symptoms may initially present after a prodromal period of several days or weeks (mean 12.5 days), which may not occur in MS (21, 33). A history of a precipitating event (previous infection or vaccinations) was not a prerequisite for the diagnosis of ADEM, which occurred in half of adult patients. The relationship between precipitating events and the occurrence of ADEM remains controversial, and possible mechanisms may include either molecular mimicry or direct inflammatory damage to myelinated neurons (34).Seasonal variation in the ADEM frequency (with peaks in spring and winter) supports its infectious aetiology (9). Patients with prior infection commonly had respiratory or gastrointestinal viral infections, such as rubella, mumps, measles, varicella, smallpox, and Epstein-Barr virus and infrequently had post bacterial infections, such as Mycoplasma pneumonia (35, 36). The severity of ADEM may be closely associated with different types of prior infection (37). A recent study included 30 ADEM cases after COVID-19 with a predominance of adult cases, possibly because adults are more likely to be infected by COVID-19 (38). For these adult cases, the average age was 50 years, which is older than those in previous reports of ADEM adults (39). More severe cases occur in older adults who often have multiple complications (40). Notably, ADEM can also occur in patients with asymptomatic COVID-19 infection (41). Postvaccination ADEM has occurred with various vaccines, including hepatitis B, hepatitis A, influenza, yellow fever, rubella, and poliomyelitis tetanus vaccines (22–24). In addition, a few recent studies reported that patients developed ADEM after being vaccinated for COVID-19 (42, 43). As postvaccination ADEM is rare, no definitive conclusions can be drawn about the association between a specific vaccine and the real risk of ADEM (44).

The clinical presentation is heterogeneous depending on the area involved in the demyelinating process. Four-fifths of the patients had a polyfocal clinical presentation associated with multifocal neurological deficits. Patients could show prodromal symptoms such as fever, headache, nausea, and vomiting. Fever was less common in adults than in paediatric patients, implying that age-related changes in the immune response, especially inflammatory cytokine reactions, might partially explain this difference (8). As fever is an unusual symptom in other demyelinating diseases, such as MS, this symptom may support the diagnosis of ADEM (21). The acute phase occurs with encephalopathy, which may be subtle, especially in the early course, and is often noted as “irritability” or “sleepiness” rather than confusion or obtundation (45). Studies in paediatric cohorts have shown encephalopathy, with rates ranging from 58.3% to 74% (8, 9, 18). In adults, the occurrence of encephalopathy is clinically important, as its absence may help identify other diseases, such as MS, which is less likely to present with encephalopathy (21). However, the frequency of encephalopathy in adults is relatively low (43.7%), suggesting that the absence of encephalopathy should not discourage neurologists from diagnosing this disease in adults. Encephalopathy is essential for the diagnosis of ADEM according to the IPMSSG criteria. These criteria may be overly restrictive for adults. We found that motor deficits, including hemiplegia, paraplegia and quadriplegia, were the most common clinical features, which occurred in more than three-fifths of patients, implying that pyramidal tract lesions are involved in most cases. Another common clinical feature is ataxia, which is rare in MS (46). Atypical symptoms, including meningeal signs, seizures and neuropsychiatric symptoms, may resemble CNS infectious diseases, highlighting a certain variability in the clinical presentation of ADEM. Thus, further auxiliary examinations including MRI imaging and CSF examinations are needed (5).

A previous study suggested that the clinical diagnosis of ADEM is mainly based on neuroimaging findings in clinical practice (47). MRI is a highly sensitive technique for detecting white matter abnormalities and is the preferred examination method for the diagnosis of ADEM (48). FLAIR and T2-weighted MR imaging is suitable for lesions detection. Lesions in most of ADEM patients show vasogenic cerebral edema on diffusion weighted imaging (DWI), which is beneficial to establish a differential diagnosis with other diseases (cytotoxic edema), such as acute cerebral infarction (49). ADEM is characterized by multifocal inflammatory demyelinating lesions in white matter (87.1%). Lesions are typically large and localized asymmetrically and have poorly defined margins. The deep grey matter is frequently involved (32.4%), often symmetrically, while cortical areas (23.8%) are involved to a lesser extent (29). The frequency of gadolinium-enhancing lesions is highly variable among studies, ranging from 30.0% to 95.2%, which may depend on the time of MRI evaluation. Notably, nearly one-tenth of ADEM adults present with a normal MRI or a delay of a few days to a few weeks between symptom onset and the appearance of MRI abnormalities (29). Thus, a normal MRI in the first days after symptom onset does not rule out a diagnosis of ADEM, and a follow-up MRI is needed for these patients. Compared with children, adults present with more frequent involvement of the periventricular areas and less often with involvement of the basal ganglia (8, 9). In addition, neuroimaging is useful to distinguish CNS demyelination diseases from other causes, such as MS and neuromyelitis optica spectrum disorder (NMOSD). Lesions in the thalamus and basal ganglia are more typical of ADEM than MS (50). Callosal lesions have been reported to occur in more than half of patients with MS (51, 52) but only in nearly one-fifth of adults with ADEM. Furthermore, a recent study compared callosal lesions on MRI among adults with CNS demyelinating diseases and found that radial lesions, defined as strip-shaped lesions spreading vertically from the lower surface to the upper surface of the corpus callosum, may be characteristic of MS and NMOSD but are rarely observed in ADEM (53). Moreover, a paediatric cohort study found that T1 hypointense lesions and more than two periventricular lesions may be MRI characteristics for distinguishing MS from ADEM, which requires further confirmation in an adult population (54). The spinal cord is also commonly involved in ADEM; however, an isolated spinal cord lesion without supratentorial involvement is unusual, which is different from NMOSD (55, 56).

CSF examinations in suspected ADEM should be performed rapidly after hospital admission to differentiate ADEM from other disorders, such as infection, MS and NMOSD. ADEM often presents as nonspecific inflammatory changes in most ADEM patients, including pleocytosis (51.8%) and elevated protein levels (39.1%) (29). Compared with the high incidence of CSF oligoclonal bands (OCBs) (>80%) in MS, positive CSF OCBs occurred in only one-fifth of ADEM adults (57). Among ADEM patients with positive OCBs, identical serum and CSF OCBs (‘mirror pattern’) were the most common pattern, suggesting predominant systemic immune activation, which is likely a consequence of infection or vaccination. However, most patients had CSF-restricted OCBs in MS. These results indicate that the detection of OCBs in both CSF and serum could help discriminate ADEM from MS (58). Autoantibodies targeting myelin oligodendrocyte glycoprotein (anti-MOG-IgG) are a marker of several central nervous system inflammatory demyelinating disorders, including ADEM, bilateral optic neuritis, transverse myelitis, and brainstem encephalitis. MOG-IgG positive children presented as ADEM in approximately 50% of cases, whereas adults in less than 10% (59). In children, high MOG-IgG at onset with declining antibody levels may be associated with a monophasic disease course and more likely to have a favorable long-term prognosis (60). While, persistent MOG-IgG may imply a recurrent disease course (61). Due to the limited ADEM cases in adults, further larger-scale longitudinal studies are needed to evaluate the value of MOG antibody in ADEM adults.

EEG is useful for the evaluation of brain function, especially if neuroimaging and CSF findings are negative (19). Diffuse background slowing is the most common finding, which may provide an alternative explanation for the encephalopathy observed in these patients (3, 26). In addition, the presence of focal slowing or epileptiform discharges may support seizures in ADEM patients.

Functional outcomes of ADEM in paediatric patients are generally favourable. Adults had a more unfavourable functional outcome and slower recovery: nearly one-tenth of adult patients died, and residual deficits occurred in nearly half of the patients during the follow-up. In addition, two-fifths of adult patients required admission to the intensive care unit (ICU). Moreover, the duration of hospitalization was longer for adults than for children (2, 9). The discrepancy in functional outcomes may be related to the reduced plasticity of the ageing brain and differences in immune responses, which has been hypothesized previously in a mouse model (62). Our analysis of the follow-up time-specifc mortality indicated that half of deaths occurred after the first 3 months. This finding stresses the importance of follow-up.

Previous studies suggested that impaired consciousness and seizures were associated with poor outcomes, while patients with fever, nonreactive CSF and a low lesion load on MRI were more likely to have a good outcome (8, 22, 26). However, due to the different risk factors and endpoints among these studies, it is difficult to merge and analyze the common risk factors among these studies. Moreover, recurrence may occur in adults, affecting 0% to 22.6% of patients according to different studies, and follow-up visits may be necessary (8, 18, 23, 25) As the prognosis of ADEM in adults is worse, more rapid and aggressive treatment may contribute to a better outcome in these patients (9). No randomized trials have identified the best treatment for ADEM, and small observational studies and expert opinions are the main basis for therapy. The most effective therapy for ADEM remains unclear. Based on the hypothesis of immune-mediated damage, the standard treatment is unspecific immunosuppressive therapy. High doses of corticosteroids are the first-line treatment and are used in most patients. Intravenous immunoglobulins (IVIGs) and plasma exchange (PLEX) are considered second-line treatment in patients with resistance or contraindications to steroids (3, 63, 64). Notably, PLEX could also be effective as rescue therapy in paediatric ADEM (65). Moreover, patients with contrast-enhancing lesions or mass effect may have the greatest benefit from PLEX (66). In addition, for a few patients with intracranial mass effects, decompressive craniectomy may be a lifesaving intervention for such cases (67, 68).

This study has several limitations. First, most of the included studies were retrospective with small sample sizes and some auxiliary examinations could not be available for all patients. Second, there were possible causes of heterogeneity among the included studies based on the inclusion of different races, studies conducted in different countries. The lack of continuity in the diagnostic criteria used throughout the past decades could also cause heterogeneity, an uniform diagnostic criteria is urgent, thus we proposed a summary of the key observations in ADEM adults. Fourth, only a few included studies evaluated prognostic factors, which should be investigated in further studies. Fifth, the examination of MOG antibody is absent in the original articles, which should be conducted in further studies. Finally, no randomized trials for the treatment of ADEM were found, and which treatment may be the best is unclear. Due to the limitations mentioned above, these data should be interpreted with caution, and further large and well-designed studies are needed to confirm the conclusions of this study.

The outcomes of ADEM in adults are worse than those in children. Considering the differences in the clinical characteristics of ADEM between adults and children, to distinguish ADEM from other CNS demyelination diseases, diagnostic criteria specialized for adults are needed. In addition, controlled trials are needed to evaluate the effectiveness and safety of different immunotherapies for this disease.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

JW, XD and KL had the idea for the study. KL, ML, LW and QW collected the data. All authors drafted and revised the paper, and approved the final version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.