Introduction

Neuromyelitis optica spectrum disorder (NMOSD) is a relapsing autoimmune neurological disease. For many years, it was a type of multiple sclerosis (MS), but it is now seen to be clearly different from MS [1, 2]. Antibodies against aquaporin-4 (AQP-4) are the hallmark of NMOSD, although not all patients present them. Presence of anti-AQP-4 autoantibodies relates to the unique immunopathology of the disease, but other autoantibodies like anti-myelin glycoprotein (anti-MOG) may be involved in the disease [3]. These autoantibodies have helped define a “disease spectrum” beyond the classical definition of acute attacks of transverse myelitis and optic neuritis [2]. In 2015, the International Panel for NMO Diagnosis established the criteria currently used for diagnosing NMOSD [4].

NMOSD is often a severe condition with disabling outcomes. Demyelination and loss of astrocytes are the main pathological findings in the central nervous system [5]. The most common presentations of the disease are severe optic neuritis and extensive transverse myelitis [6]. Brainstem syndromes are less frequent but equally disabling [6]. Current treatment strategies include use of corticosteroids, immunoglobulins, and plasmapheresis for acute attacks, and general or humoral immunosuppression for prevention of attacks [5, 7].

NMOSD is more prevalent among women [7, 8], with typical onset between the third and fourth decades of life [8,9,10]. Starting the disease after the age of 50 years is infrequent and seems to be associated with worse outcomes. Patients with initial signs and symptoms occurring beyond age 50 are considered to present “late-onset NMOSD” (LO-NMOSD) [11].

There are a few papers reporting on anecdotal cases of LO-NMOSD [11,12,13,14,15,16,17,18]. Case series are rare in the literature, with reports on 45 patients with LO-NMOSD from Korea [11], 108 cases from France, Germany, Turkey, and the UK [19], and 30 patients from China [20]. The aim of the present study is to report on 37 Brazilian patients with LO-NMOSD and to discuss treatment limitations to this condition.

Methods

Retrospective data were obtained from the medical charts of patients diagnosed with NMOSD in accordance with the 2015 criteria [4]. Only patients with at least 1 year of LO-NMOSD and records with complete information were included. The original study was approved by the Ethics Committee of Universidade Metropolitana de Santos, under the CAAE number 13994913.7.0000.5509. Additional ethical approval was obtained by each institution in accordance with its rules and regulations. Patients’ consent was given at the time of enrollment in the Demyelinating Diseases Specialized Centers. Patient identity remains confidential and will not be disclosed in open or shared databases.

The patients self-reported their ethnic background. Disability was assessed using the Expanded Disability Severity Scale (EDSS) [21]. Anti-AQP-4 antibodies were tested using the immunofluorescence technique [22] and no patients had access to anti-MOG antibody testing. The results were presented in a descriptive manner and, whenever required, statistical analyses were carried out using GraphPad Prism. The Shapiro–Wilk test was used to assess normality, Student t test was used to compare two means, while Chi-square test and Fisher exact test were used for independent criteria. Confidence intervals were established at 95% and differences were reported as significant when p < 0.05.

Results

A summary of the results of 37 patients with LO-NMOSD is shown in Tables 1, 2, and 3. Table 1 introduces patients’ demographic data, Table 2 presents their clinical data, and Table 3 shows clinical manifestations of LO-NMOSD, anti-AQP-4 status, and number of patients relapsing despite therapy. The ratio of women to men in the sample was 4.3:1. Gender, age of disease onset, ethnic background, and positive antibodies for AQP-4 were not associated with the number of relapses or disability outcomes. The patients were followed up for a median period of 4 years. Extensive longitudinal myelitis was the most frequent manifestation of LO-NMOSD, affecting 86% of the patients. Optic neuritis affected 70% of the patients during the course of the disease, while brainstem syndromes were present in only 16% of them. Three patients died (EDSS = 10) and six are currently using some type of support for walking or are wheelchair-bound (EDSS > 6.0).

Table 1 Sex, ethnic background, and associated diseases of 37 patients with late-onset neuromyelitis optica
Full size table
Table 2 Clinical information on 37 patients with late-onset neuromyelitis optica
Full size table
Table 3 Response to treatment, serological anti-AQP-4, and clinical manifestations in 37 patients with late-onset neuromyelitis optica spectrum disorders
Full size table

Sixty percent of the patients had relapses while using azathioprine (with or without corticosteroids) as their first-line therapy. Subsequently, either the azathioprine dose was increased or daily use of corticosteroids was added to these patients’ therapy, but 41% of them still had relapses. Through use of rituximab, immunoglobulin, pulses of plasma exchange, or therapeutic associations, LO-NMOSD was brought under control in all but six patients, who continued to have relapses. The neurologists in charge of these patients reported that the best therapeutic option was high doses of azathioprine in association with daily oral corticosteroids (which controlled relapses in 70% of the patients). Rituximab controlled the most aggressive cases but, being an off-label therapy, access to this drug was very limited in the public health system.

Discussion

Later age at onset and delay in diagnosing NMOSD have recently been described as independent factors for worse prognosis [8]. Differential diagnoses, age-related disease, and restriction on drug therapies pose particular challenges for these patients and their physicians. Chronic use of corticosteroids may be particularly difficult among individuals with diabetes, hypertension, gastrointestinal diseases, or osteoporosis. Pulses of methylprednisolone or dexamethasone, or plasma exchange, may be an option for elderly people, particularly those who do not have good tolerance for chronic use of corticosteroids [23]. Use of azathioprine has been correlated with higher risks of toxicity, infection, and malignancies [24]. Although high doses of azathioprine associated with daily oral prednisone was the first-line therapy for patients with LO-NMOSD in the present series, it is not easy to control the consequences of these drugs when administered chronically. On the other hand, the morbid consequences of underdosing or no treatment for these patients are well established. Despite the potential advantages of rituximab treatment, it cannot be used routinely for NMOSD in Brazil, since this is an off-label indication of the drug. The long-term risk–benefit relationship of rituximab is not well established for elderly individuals [25].

In comparing the present population of LO-NMOSD patients with data from a recent publication on Brazilian NMOSD patients of all ages [10], the more severe outcomes of older subjects are clear. In that series of 153 patients, the median age at NMOSD onset was 28 years. After a median follow-up of 7 years, EDSS was, on average, 5.4 ± 2.5. On the other hand, in the present population, LO-NMOSD led to significantly worse disability (p = 0.006) in just over half the time. This result is in line with what was reported by Seouk [11], Collongues [19], Mao [20], and Mealy [8]. Similar to the LO-NMOSD series of these authors, anti-AQP-4 antibodies were detected in 68% of the patients in the present study. As reported by those researchers, worse outcomes in LO-NMOSD cases were not associated with presence of anti-AQP-4 antibodies in the present series, either.

The present study has limitations. It is based upon a population of only 37 subjects and has a short follow-up period. Anti-AQP-4 was essentially assessed by an immunofluorescence technique which is not the gold standard method. Not all subjects had anti-AQP-4 assessed and anti-MOG was not easily available at the time the population was assessed. Antibodies anti-AQP-4 and anti-MOG are currently only available for private health care and are not covered or reimbursed by the public health system in Brazil.

Regarding therapeutic options, it is a challenge to receive a patient with LO-NMOSD in daily neuro-geriatric practice. The severity of the neurological disease and the age-related conditions the patient may present defy the physicians in charge of this individual. Patients with LO-NMOSD would have higher age-related risks of cardiovascular diseases, malignancies, and infections. They are also likely to make use of drugs to control chronic diseases, e.g., hypertension, diabetes, and dyslipidemia. To add corticosteroids, azathioprine, mycophenolate mofetil, and rituximab to the therapeutic scheme of these patients poses a neurotherapeutic challenge for which we have no guidelines yet.

Conclusion

LO-NMOSD is a severe disabling condition. Neurologists must work closely with specialists in geriatric diseases in order to use drugs that can lead to the best therapeutic response, good tolerability, and minimal safety issues.