Sources

A comprehensive search of PubMed and SCOPUS was performed for all studies published prior to November 30, 2013, using the search terms “encephalitis” and “teratoma”, which yielded 192 articles in PubMed and 274 in SCOPUS. A systematic review of these papers was performed, and after removal of repeated articles from both searches and cases involving male patients, the titles and abstracts of all articles were evaluated to determine whether case reports with ovarian teratoma were included. There were no language restrictions, although for articles in Chinese or Japanese, only abstracts were consulted.

Study selection

All publications reporting one or more cases with encephalitis and ovarian teratoma, including suspicious cases in which the tumor had not been found, were selected for review and full publication analysis. Altogether, 119 reports containing a total of 173 cases of anti-NMDAR encephalitis and ovarian teratoma were identified from 1997 until November 30, 2013. To these, we have added our case attended in 2012-2013 and recently published [9], making a total of 174 cases of ovarian teratoma-associated anti-NMDAR encephalitis.

Moreover, 7 of the aforementioned papers and another 32 papers published between 2009 and 2013 described 40 additional cases of anti-NMDAR encephalitis and suspicious of teratoma in which, following ultrasound, CT or MR, an ovarian teratoma had not been found. Other 11 pediatric cases (seven in females) of anti-NMDAR encephalitis without teratoma [8], and 20 female cases of ovarian teratoma with brainstem-cerebellar manifestation with or without opsoclonus-myoclonus syndrome reported by Armangue et al. [10] (including three cases previously reported [11]-[13]) have also been analyzed or independently commented on. Flow diagram for the study is shown in Figure 1 as supplemental data.

The following data were collected from all of the articles containing any case of anti-NMDAR encephalitis and ovarian teratoma: authors, journal and year of publication, country of birth or hospital service where the case was studied, number of cases included with or without ovarian tumor, patient age, clinical presentation, type of ovarian teratoma, tumor size and laterality, medical treatment, time to surgery, surgery performed, outcome and time to recovery. These data were tabulated in a Supplemental Data File (Additional file 1: Table S1), and were also entered into SPSS statistics (IBM, Spain) and statistically analyzed. Times to discharge, partial recovery or full recovery were considered following the case reported if it was mentioned.

We also analysed the case series of anti-NMDAR encephalitis, with or without teratoma, in which the cases were not individually reported, and the most comprehensive reviews of paraneoplastic limbic encephalitis for discussion.

Statistical analysis

Data were introduced in a SPSS-15 file being statistically analyzed. Percentages, mean and standard deviation, and median with minimal and maximal values were obtained. Comparison of independent means (Student's T-test), comparison of two proportions (relative risk -RR-), a 95% confidence interval (CI) and the chi-squared test were used to compare the groups. RSigma (Systat Software, San Jose, California, USA) and PEDro (Physiotherapy Evidence Database, Sidney, Australia) software were also used. All p values reported are 2 tailed and p value of 0.05 or less was considered significant.

Strengths and weaknesses

We have included in the systematic review all detailed cases of ovarian teratoma-associated anti-NMDAR encephalitis appearing in PubMed and/or Scopus and whose data had been reported as individual case report. Observational cohorts' studies without individual description of cases have not been included in the analysis of cases.

Weaknesses

1. We suspect that this pathological association is underdiagnosed and infrareported. Moreover, the case has not been analyzed if the publishing journal is not in PubMed or Scopus. 2. Teratoma is not always found, especially among girls and adolescents. 3. There may be extraovarian teratoma. 4. The ovarian teratoma may also cause a novel type of encephalitis negative for anti-NMDAR, with brainstem-cerebellar manifestation and with or without opsoclonus-myoclonus syndrome [10].

Main findings

The association of anti-NMDAR encephalitis and ovarian teratoma is a severe, prolonged and potentially fatal pathologic condition of young women that has been increasingly reported in developed countries since its identification and publication by Dalmau et al. [4]. But is uncommon in girls ≤14 years of age [25], Only 4 females younger than 12 years -6%- had a tumor in Titulaer et al. [26] study.

The most frequent clinical presentations are neuropsychiatric symptoms (especially, behavioural changes), although the neuropsychiatric manifestations may often be preceded by nonspecific prodromal symptoms. Moreover, the typical presentation of the cases include memory impairment and, later but very evocative, abnormal movements of the face, mouth and tongue and symptoms of dysautonomia. General practitioners and gynecologists should be cognizant of available imaging modalities that may be employed in the early stage and of the need for the eventual discovery of a dermoid cyst because early tumor resection with immunotherapy will facilitate patient recovery. It should also be noted that at the onset of symptoms, the results of anti-NMDA antibody testing may be negative, even in CSF [21]. Nevertheless, it has been suggested that in some patients with anti-NMDAR encephalitis the antibodies were detectable only in the CSF [27]; and Gresa-Arribas et al. [23] have reported that the sensitivity of NMDA receptor antibody testing is higher in CSF than in serum, that the titres were higher in patients with poor outcome and that antibody titres can complement clinical assessment.

Interpretation and alternative explanations for observed results

Even small ovarian teratomas containing nervous tissue may express NMDAR subunits that react with a patient's antibodies, thereby triggering a cascade of symptoms that is recognized as anti-NMDAR encephalitis and that mainly affects the hippocampus/forebrain regions. In such cases, tumor resection and immunotherapy result in improvement or full recovery. However, the diagnosis is difficult and time-consuming, and patients may succumb to neurological deterioration, thus necessitating a high index of suspicion and prompt intervention, even in the absence of confirmatory studies.

From this systematic review, it is apparent that awareness of the association of ovarian teratoma (usually benign or dermoid cyst) and anti-NMDAR encephalitis is low in many parts of the world, possibly because practitioners may not entertain such a diagnosis, thus resulting in a low rate of diagnosis even among affected patients. Some women may be transferred from one hospital to another or from one service to another, which may contribute to the fatality rate of approximately 7%, with the teratoma occasionally found only at autopsy. Therefore, heightened awareness will benefit patients through a more rapid identification of this association, especially in Africa, Asia and Central and South America, and particularly among gynecologists. Few reported cases have emerged from these regions, possibly because of their respective levels of health care and socio-economic development.

However, from the point of view of etnicity, Titulaer et al. [26] in a cohort multi-institutional observational study of 577 anti-NMDAR encephalitis patients have observed that Asian and African-American patients were more likely to have a teratoma (45% and 48% respectively) than Caucasian (31%) or Hispanics (27%), being p < 0.007.

On the other hand, Tachivana et al. [28] noted that in the vast majority of encephalitis cases, patients have a prodromal infection before the onset of the disease. Should this infection affect either the normal ovary or ovarian teratoma, it may trigger the expression of NMDAR-related epitopes in oocytes and lead to the development of limbic encephalitis disease [29]. Indeed, ovarian teratomas that contain large amounts of neural tissue may present an increased risk for the development of this type of life-threatening encephalitis. Subsequently, the diagnosis can be established by identification of anti-NMDA antibodies in the serum and CSF. When these antibodies are positive, even in the absence of teratoma, it has been suggested that proceeding to exploratory laparotomy may increase the possibility of identifying microscopic teratoma and may improve the outcome in patients who are refractory to immunotherapy [30], however it is unclear whether this procedure is the most appropriate.

Generalization of the conclusions

It now appears relevant to raise the questions posed to gynecologists by the 2009 report of Kort et al. [31].

Incidence and prevalence of this condition

Only eight cases with paraneoplastic limbic encephalitis and teratoma, some of which were recurrent [20], were described before 2005, the year in which Dalmau et al. introduced the determination of anti-NMDAR antibodies in the serum and CSF [3],[6],[32]. Since 2007, the year in which Dalmau et al. published their work in the Annals of Neurology [4], there has been a progressive increase in the number of articles and case reports regarding this entity. Thus, it appears that the incidence of this condition is higher than previously estimated and will continue to increase with improvements in health care, knowledge of the condition and improved socio-economic development, especially in many of the larger countries of Asia, Africa and South-America where cases have not yet been reported. General practitioners and gynecologists should be aware that the underlying cause of anti-NMDAR encephalitis may involve a simple ovarian dermoid cyst.

Clinical presentation

Neuropsychiatric symptoms, rapid neurological deterioration, ICU admission and ventilator support are often initiated before the diagnosis is established. Patients often require long periods of support, and they may also develop tracheotomy complications related to hypersalivation and may develop life-threatening hyperthermia before teratoma-associated encephalitis is even suspected. Other frequent prodromal symptoms are, e.g., earache, headache, dysuria and malaise, which may be attributed to various benign etiologies. And as indicated, are also very evocative, abnormal movements of the face, mouth and tongue and symptoms of dysautonomia. Although it is not common practice to perform a gynecological examination at the time of catatonia or status epilepticus, if an X-ray, CT, MR or ultrasound demonstrates ovarian pathology or calcification, it is imperative to consider the association of ovarian teratoma and encephalitis.

Diagnosis

Serum and CSF evaluation for anti-NMDAR antibodies and imaging tests, particularly transvaginal or transrectal ultrasound in female patients to identify small ovarian teratomas, are essential for diagnosis. In cases of brainstem cerebellar syndrome with or without teratoma-associated opsoclonus-myoclonus syndrome, anti-NMDAR antibody studies may be negative [10].

Treatment

The combination of tumor removal and immunotherapy (ivIg, corticosteroids and plasma exchange) yields the best therapeutic results and a more rapid recovery than immunotherapy alone [4]. In the cohort study of Titulaer et al. [26] it was observed that patients without a tumor had a higher frequency of relapses than those with a tumor. Therefore, tumor resection should be performed as soon as possible after psychomotor and hemodynamic stabilization of the patient. In some cases, a delay in tumor excision has led to death [4],[33], although other patients have recovered without surgery [34]. Boeck et al. [17] performed ovarectomy despite negative imaging in anti-NMDAR encephalitis. Histological examination revealed an occult teratoma and they pointed out that removal of the teratoma even 11 months after a very severe course is still therapeutically effective.

Role of the dermoid cyst and need for its excision

Regardless of the histological type (most mature teratomas), teratomas exhibit elements of the three blastodermic layers and contain neural tissue that triggers an immune response resulting in over-production of anti-NMDAR antibodies. Dalmau et al. [35] demonstrated that all dermoid tumors examined in anti-NMDAR encephalitis patients contained neural tissue, and 100% of specimens tested were positive for NMDA receptors. Similarly, it was noted that the presence of NR2B immunoreactivity in normal human ovary oocytes may account for the occurrence of anti-NMDAR encephalitis in predominantly young females and its absence in girls [28],[29]. In a case [36], a bilateral oophorectomy was performed despite the absence of a teratoma on imaging tests after the patient had failed to recover from anti-NMDA encephalitis over a 6-week period. Histology demonstrated a small, 0,7 cm mature teratoma and patient recovered afterwards. Thus, if all imaging tests are negative, it has been recommended to proceed with exploratory laparotomy and a large wedge resection of the ovary to increase the likelihood of microscopic teratoma identification and to improve the outcome in patients who are refractory to immunotherapy and have anti-NMDAR antibodies [17],[30], but this issue remains controversial. We believe that a focused transvaginal or transrectal ultrasound may help decision-making better than a wedge resection, and if results are negative, all immunotherapeutic protocols should be used before proceeding to bilateral salpingo-oophorectomy that could result in ovaries macroscopically and histologically normal and menopause [37],[38].

Finally, we must also remember that patients with ovarian teratoma could develop other forms of encephalitis without NMDAR antibodies and that any adolescent or young adult, particularly female, who develops subacute brainstem-cerebellar symptoms or opsoclonus-myoclonus that is suspected to be immune-mediated (because of the rapid onset of symptoms and/or CSF pleocytosis) should be investigated for teratoma of the ovary (or testes in male patients) [10].

However, the findings of Mangler et al. [39] support the current clinical practice, in which systematic screening for anti-NMDAR antibodies in teratoma patients is not indicated.

Guidelines for future research

Heightened recognition of behavioral changes as the initial symptoms of this entity must be emphasized, followed by diagnosis through TVU and subsequent tumor removal in addition to diagnostic confirmation through the presence of anti-NMDAR antibodies. However, in some cases, the teratoma could present following the encephalitis or be microscopic. Besides, some patients with ovarian teratoma may develop other forms of encephalitis that are not anti-NMDAR-related [10] or have extra-ovarian teratomas, notably mediastinal teratomas [40],[41]. These aspects should guide future investigation among encephalitis and teratoma.