Recurrent Anti-NMDA Receptor Encephalitis After Mycoplasma Pneumonia Infection

Since its formal designation as a disease in 2007, autoimmune anti-NMDA receptor (anti-NMDAR) encephalitis has been increasingly recognized for its neuropsychiatric presentation. It is currently estimated to affect 1.5 million people worldwide each year (1). Common signs and symptoms include rapid onset of autonomic instability, altered level of consciousness (including catatonia), speech and/or movement disorders, behavioral and cognitive dysfunction, and seizures (2). Diagnosis of the disease is confirmed by the detection of IgG antibodies to the GluN1 (NR1) subunit of the NMDA receptor in the patient’s CSF (2). However, because results of this test are often delayed, patients who exhibit characteristic signs and symptoms should be treated presumptively when they meet certain criteria and other reasonable diagnoses have been excluded (Table 2). Detection of antibodies in a patient’s serum, while less sensitive than in CSF, is also diagnostic (3).

Classically, anti-NMDAR encephalitis has been associated with ovarian teratomas, with a predilection for women of reproductive age. However, only about 50% of cases in this population are associated with teratomas, and the percentage is lower in men and in women under age 18 (4). This gap has led to an interest in other etiologies. Infection with several agents, including herpes simplex virus, human herpesvirus 7, SARS-CoV-2, and Mycoplasma pneumoniae, have been associated with the development of anti-NMDAR antibodies in small subsets of patients (5, 6). Unlike patients who have tumors that can be resected, patients with postinfectious anti-NMDAR encephalitis are at higher risk of relapse after their initial symptoms have resolved (4).

The patient we describe here presented at 16 weeks’ gestation with neuropsychiatric symptoms and was diagnosed with anti-NMDAR encephalitis. No tumor was identified. On chart review, the patient was found to have a history of M. pneumoniae infection complicated by similar neuropsychiatric symptoms. She also experienced a relapse of anti-NMDAR encephalitis after delivery. A literature search regarding M. pneumoniae and NMDAR encephalitis performed in PubMed, PsycINFO, and MEDLINE returned a total of three articles, consisting of two case reports and one case series. The case reports described cases in children (ages 9 and 13) where CSF was positive for anti-NMDAR antibodies with concurrent serum studies positive for Mycoplasma IgM antibodies (7, 8). In the case series of 10 patients with anti-NMDAR encephalitis, only two were found to have a teratoma. Of the remaining eight patients, half (40% total) were found to have serologic evidence of M. pneumoniae infection (5).

M. pneumoniae infects an estimated 2 million Americans annually, although the actual number could be higher given the prevalence of subclinical infections (9). Although classically associated with respiratory disease, there are several commonly reported extrapulmonary manifestations, including neurologic disease. CNS involvement is the most common extrapulmonary illness, affecting up to 10% of patients hospitalized with M. pneumoniae infection, through various pathogenic mechanisms, including direct infection, vascular occlusion, and autoimmunity (10, 11). It is proposed that NMDAR autoantibodies after M. pneumoniae infection develop in a manner similar to other known postinfectious neurologic syndromes, such as Sydenham’s chorea and pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections (PANDAS). Specifically, after M. pneumoniae infection, a patient can develop autoantibodies against the neuronal membrane, which, through molecular mimicry, can come to be associated with receptors in the CNS, like the NDMA receptor (10). The number of cases of anti-NMDAR encephalitis that develop after M. pneumoniae is unknown, although given the high incidence of infections each year, the severity of neuropsychiatric symptoms that emerge in cases of anti-NMDAR encephalitis, and the relative paucity of reports on cases of anti-NMDAR encephalitis resulting from infections as opposed to tumors, it is important to recognize this as a potential mechanism of disease.

Our patient had developed seizures, behavioral and cognitive changes, and abnormal speech patterns 2 years before presenting to our hospital. At that time, she was found to have both IgG and IgM antibodies to M. pneumoniae in her serum, but not CSF, and was treated with antibiotics and immune suppression for presumed Mycoplasma encephalitis. She received a course of intravenous methylprednisolone, followed by an oral steroid pulse, with symptomatic improvement. A CSF autoimmune panel was not performed. However, reviewing the accepted criteria for probable anti-NMDAR encephalitis presented in Table 1, we suspect that she had anti-NMDAR encephalitis at that time and recovered after being treated with steroids. Supporting this hypothesis is the fact that studies comparing the presentation of infectious encephalitis to that of autoimmune encephalitis find that fever and headache are the most common presenting symptoms for a patient with infectious encephalitis, while seizures, psychiatric symptoms (especially psychosis), language dysfunction, and autonomic instability are all more common in autoimmune disease (6). Our patient’s prior presentation was notable for seizures, auditory hallucinations, agitation, vital sign instability, and speech abnormalities, but there is no record of fever or subjective headaches. Even if her first episode of neuropsychiatric symptoms was due to Mycoplasma encephalitis, rather than an autoimmune encephalitis, the infection still provides a compelling link between past illness and current presentation in this otherwise healthy patient without evidence of malignancy, through the development of postinfectious autoantibodies.

For classic tumor-related cases of anti-NMDAR encephalitis, treatment with first-line immunomodulation (Table 2) plus tumor resection is typically successful. Because patients with postinfectious symptoms are more likely than those with tumor-related cases to experience relapse, this patient’s history of M. pneumoniae infection has important implications for long-term management (12). Specifically, her treatment plan included rituximab, a second-line immunosuppressant therapy, after pregnancy. This is consistent with the case report of a 9-year-old boy with M. pneumoniae and subsequent anti-NMDAR encephalitis for whom treatment with both high-dose steroids and IVIG failed and who did not have sustained improvement until initiation of rituximab (8). The symptomatic management of neuropsychiatric manifestations of anti-NMDAR encephalitis, even when the diagnosis is confirmed and immunotherapy is initiated, is particularly challenging. Our patient’s psychotic and catatonic symptoms were slow to resolve with medication management using benzodiazepines, antipsychotics, and memantine. Although the common and disruptive psychotic and behavioral symptoms associated with anti-NMDAR encephalitis often warrant an antipsychotic trial, there is some evidence that anti-NMDAR encephalitis increases the risk of adverse effects of antipsychotics, such as neuroleptic malignant syndrome and seizures (13, 14). Similarly, catatonic symptoms often warrant treatment with benzodiazepines, which can lead to worsening delirium and hypoventilation. There is a small body of case report literature discussing the safety and efficacy of ECT as an alternative treatment for the neuropsychiatric symptoms of anti-NMDAR encephalitis, including psychosis and catatonia, given its rapid onset of action (13–15). As mentioned above, this was discussed with the patient’s family as an option because of the severity and duration of her catatonic and psychotic symptoms, but ultimately not agreed upon. The use of ECT early in a course of treatment for anti-NMDAR encephalitis is an area that warrants further study.

Notably, while SARS-CoV-2 infections have been associated with a myriad of neuropsychiatric symptoms, the patient’s prior illness and onset of presenting symptoms preceded her SARS-CoV-2 infection, so we do not think it was a relevant factor in her course of illness. Likewise, while pregnancy has also been associated with cases of autoimmune encephalitis in women of reproductive age, this too does not explain our patient’s course of illness, since she had episodes before, during, and after pregnancy. The evaluation, management, and outcomes of pregnancies complicated by anti-NMDAR encephalitis are also important but have been well documented elsewhere and are beyond the scope of this article.