Clinical and therapeutic decision making in paediatric autoimmune and inflammatory neurological disease

ABSTRACT 1. Background. Paediatric neuroimmunology is a rapidly evolving field both as regards clinical-radiological phenotyping, biomarker development, and therapeutic possibilities. In this latter aspect, while a growing armamentarium of treating agents is becoming available, this is not mirrored by quality evidence and definite recommendations on treatment strategies, drugs’ efficacy and tolerability. This is especially true in paediatric age, where most data is derived from adult studies. Objective. To investigate clinical and therapeutic aspects of decision making in paediatric autoimmune and immune-mediated inflammatory conditions. In particular, the aims of this work include: exploring the available immune therapeutic agents and their mechanisms of action; investigating the use of immune therapy in autoimmune encephalitis; investigating the use, efficacy and tolerability of individual immune therapeutic agents in different clinical situations in paediatric neurology (intravenous immunoglobulin in Sydenham’s chorea; intravenous immunoglobulin in paediatric neurology; therapeutic plasma exchange in anti-N-methyl-D-aspartate receptor (anti-NMDAR) encephalitis; rituximab in neuromyelitis optica spectrum disorders; mycophenolate mofetil, azathioprine and methotrexate in anti-NMDAR encephalitis; mycophenolate mofetil in paediatric autoimmune and immune-mediated central nervous system (CNS) conditions). 2. Methods. The present PhD thesis is articulated into sub-projects carried out at the Department of Women's and Children's Health in Padua, Italy, and at the Children's Hospital at Westmead, Sydney, Australia. Study designs include six literature reviews (available immune therapeutic agents and their mechanisms of action; immune therapy in autoimmune encephalitis; immune therapy in herpes simplex virus-induced anti-NMDAR encephalitis; intravenous immunoglobulin in Sydenham’s chorea; therapeutic plasma exchange in paediatric anti-NMDAR encephalitis; mycophenolate mofetil, azathioprine and methotrexate in paediatric anti-NMDAR encephalitis) and four original studies with observational retrospective design (immune therapy in paediatric anti-NMDAR encephalitis; intravenous immunoglobulin in paediatric neurology; rituximab in neuromyelitis optica spectrum disorders; mycophenolate mofetil in paediatric autoimmune and immune-mediated CNS conditions). Of these latter original studies, one includes an Italian population (immune therapy in paediatric anti-NMDAR encephalitis), one a single-center Australian population (intravenous immunoglobulin in paediatric neurology), and two include an international cohort of paediatric patients (rituximab in neuromyelitis optica spectrum disorders; mycophenolate mofetil in paediatric autoimmune and immune-mediated CNS conditions). Most of the projects have been concluded, whereas two are in their final phases (mycophenolate mofetil, azathioprine and methotrexate in paediatric anti-NMDAR encephalitis; mycophenolate mofetil in paediatric autoimmune and immune-mediated CNS conditions). 3. Results. Key findings are presented for the main study objectives. 3.1 Immune therapeutic agents and their mechanisms of action. 3.1.1 Immune therapeutic agents and their mechanisms of action (literature review). First-line treatments typically include corticosteroids, intravenous immunoglobulin, and plasmapheresis, while for severe disease second-line ‘induction’ agents such as rituximab or cyclophosphamide are used. Steroid-sparing agents such as mycophenolate mofetil, azathioprine or methotrexate are often used in potentially relapsing or corticosteroid-dependent diseases. Lessons from adult neuroimmunology and rheumatology could be translated into pediatric autoimmune CNS disease in the future, including the potential utility of monoclonal antibodies targeting lymphocytes, adhesion molecules for lymphocytic migration, cytokines or their receptors, or complement. Finally, many agents used in other fields have multiple mechanisms of action, including immunomodulation, with potential utility in neuroimmunology, such as antibiotics, psychotropic drugs, probiotics, gut health, and ketogenic diet. 3.2 Immune therapy in autoimmune encephalitis. 3.2.1 Autoimmune encephalitis with antibodies targeting neuronal surface antigens (systematic literature review). Most studies on immune therapy in autoimmune encephalitis associated with antibodies to cell surface antigens are retrospective cohorts, and there are no randomised controlled trials. Most clinicians use first-line therapy (steroids, intravenous immunoglobulin, plasma exchange), and if severe or refractory, second-line therapy (rituximab, cyclophosphamide). When present, tumours should be removed. There are common therapeutic themes emerging. Firstly, patients given immune therapy do better and relapse less than patients given no treatment. Secondly, patients given early treatment do better. And thirdly, when patients fail first-line therapy, second-line therapy improves outcomes and reduces relapses. Given the retrospective uncontrolled data, the literature has inherent bias, including severity and reporting bias. 3.2.2 Clinical and therapeutic aspects of the Italian cohort of paediatric anti-NMDAR encephalitis (national retrospective observational study). We described a new case series of 20 children (50% females), with anti-NMDAR encephalitis referred by 13 Italian centers (mean age at onset 8 years, range 3-17). Onset was with neurological symptoms in 70%, and with behavioral/psychiatric disturbances in 30%. Most patients developed a severe clinical picture (90%), and 41% experienced medical complications; children 12-18 years old seemed to be more severe and symptomatic than younger patients. All children received first-line immune therapy; second-line treatment was administered to 45%. Relapses occurred in 15%. At last follow-up (mean 23.9 months, range 5-82), 85% patients had modified Rankin Scale (mRS) 0-1; this rate was higher among older patients, and in those receiving first immune therapy within 1 month. 3.2.3 Herpes simplex virus-induced anti-NMDAR encephalitis (systematic literature review). 43 patients with herpes simplex encephalitis (HSE) followed by anti-NMDAR encephalitis were identified in the literature (31 children). Latency between HSE and anti-NMDAR encephalitis was significantly shorter in children than adults (median 24 vs. 40.5 days; p=0.0057). Compared to the HSE phase, anti-NMDAR encephalitis was characterized by significantly higher frequency of movement disorder (2.5% in HSE, vs. 75% in anti-NMDAR encephalitis; p<.0001), and by significantly lower rate of seizures (70% in HSE, vs. 30% in anti-NMDAR encephalitis; p=0.0011). Compared to adults, during anti-NMDAR encephalitis children had significantly more movement disorder (86.7% in children, vs. 40% in adults; p=0.0064) and less psychiatric symptoms (41.9% in children, vs. 90% in adults; p=0.0251). Children also had a slightly higher median mRS than adults during the acute phase of anti-NMDAR encephalitis (5 vs. 4; p=0.0146). During anti-NMDAR encephalitis, 84.6% patients received acyclovir (for ≤7 days in 22.7%; long-term antivirals in 18% only), and 92.7% immune therapy, but none had recurrence of HSE clinically or using CSF HSV-PCR (median follow-up 7 months). 3.3 Modes of use, efficacy and tolerability of individual immune therapeutic agents in different clinical situations in paediatric neurology. 3.3.1 Intravenous immunoglobulin in Sydenham’s chorea (systematic literature review). The studies reviewed on intravenous immunoglobulin in Sydenham’s chorea demonstrate a short-term benefit in symptomatic improvement. However, they do not clarify an optimum timing and duration for use of intravenous immunoglobulin, and do not provide data on the effect on long-term neurological and psychiatric complications. 3.3.2 Intravenous immunoglobulin in paediatric neurology (single-center retrospective observational study). 196 children received intravenous immunoglobulin for neuroimmunological indications at the Children’s Hospital at Westmead, Australia, between 2000 and 2014 (28.1% had Guillain-Barré syndrome) (15.5% of all hospital indications). In total, 1669 intravenous immunoglobulin courses were administered (total 57221 g, median 78 g/patient, range 12-5748 g). Highest median number of courses was in chronic inflammatory demyelinating polyneuropathies, opsoclonus-myoclonus-ataxia, suspected immune-mediated epilepsies and Rasmussen’s encephalitis. Adverse reactions occurred in 25.5%, mostly minor. Outcome at follow-up was best in anti-NMDAR encephalitis, Guillain-Barré syndrome and myasthenia gravis, and worst in Rasmussen’s encephalitis and epilepsies. The total cost for intravenous immunoglobulin was 2,595,907 American dollars (median $3,538/patient, range $544-260,766). 45.4%-57.1% patients received intravenous immunoglobulin for ‘weak’ indications or ‘not listed’ in international guidelines. Some entities frequently treated with intravenous immunoglobulin in current practice, such as anti-NMDAR encephalitis and transverse myelitis, are not listed in most guidelines. 3.3.3 Therapeutic plasma exchange in anti-NMDAR encephalitis (systematic literature review). 71 articles were identified (mostly retrospective), reporting a total of 242 children treated with therapeutic plasma exchange for anti-NMDAR encephalitis (73.2%, 93/127 females; median age at onset 12 years, range 1-18). Median time to immunotherapy was 21 days (range 0-190). In most cases, therapeutic plasma exchange was given with steroids and intravenous immunoglobulin (69.5%, 89/128), or steroids only (18%, 23/128); in a minority, it was associated with intravenous immunoglobulin only (7%, 9/128), or was the only first-line treatment (5.5%, 7/128). In 54.5% (65/119), therapeutic plasma exchange was the third treatment after steroids and intravenous immunoglobulin, in 31.1% (37/119) the second after steroids or intravenous immunoglobulin; only in 14.3% (17/119) was it the first treatment. Second-line immunotherapies were administered in 71.9% (100/139). Higher rates of full/substantial recovery at follow-up were observed with immunotherapy given ≤30 days from onset (69.4%, 25/36) compared to later (59.2%, 16/27), and when therapeutic plasma exchange was associated with steroids (66.7%, 70/105) rather than not (46.7%, 7/15). Significant adverse reactions to therapeutic plasma exchange were reported in 6 patients. 3.3.4 Rituximab in paediatric neuromyelitis optica spectrum disorders (international retrospective observational study). 16 patients treated with at least two courses of rituximab for neuromyelitis optica were included (14 females; mean age 9.6 years, range 1.8-15.3). The patients had a mean of 6.1 events (range 1-11) during a mean follow-up of 6.1 years (range 1.6-13.6), and received a total of 76 rituximab courses (mean 4.7, range 2-9) in 42.6-year cohort treatment. Before rituximab, 62.5% received azathioprine, mycophenolate mofetil or cyclophosphamide. Mean time from rituximab to last documented B cell depletion and first repopulation was 4.5 and 6.8 months respectively, with large inter-patient variability. Earliest repopulations (2.7 and 2.9 months) occurred with the lowest rituximab doses. Significant reduction between pre and post rituximab annualized relapse rate (ARR) was observed (p=0.003). During rituximab, 6 patients were relapse-free, although 21 relapses occurred in 10 patients, including 13 ‘repopulation’, 3 ‘depletion’, and 4 ‘depletion failure-related relapses’. Of the 13 ‘repopulation relapses’, 4 had CD19 10-50x106cells/L, 10 inadequate monitoring (≤1 CD19 in the 4 months before relapses), and 5 delayed re-dosing ≥10 days after repopulation detection. 3.3.5 Mycophenolate mofetil, azathioprine and methotrexate in anti-NMDAR encephalitis (systematic literature review). 76 patients treated with mycophenolate mofetil/azathioprine/methotrexate for paediatric-onset anti-NMDAR encephalitis were included (age range at onset 0.8-18 years; 69.7% females; 49.1% had ≥1 relapse), reported in 37 articles. Mycophenolate mofetil was used in 53.9%, azathioprine in 25%, methotrexate in 15.8%; an additional 5.3% received two among mycophenolate mofetil/azathioprine/methotrexate. Mycophenolate mofetil/azathioprine/methotrexate were not preceded by any second-line therapy (rituximab/cyclophosphamide) in 47.7%, and were administered only after relapses in 46.8%. Among the subgroup treated with mycophenolate mofetil/azathioprine/methotrexate after the first event, relapses occurred in 8.3% only. Time on mycophenolate mofetil/azathioprine/methotrexate was median 9 months (range 1-48). Median annualised relapse rate was 0.45 (mean 1, range 0-6.67) before mycophenolate mofetil/azathioprine/methotrexate (excluding onset), and 0 (mean 0.06, range 0-1.3) during/after mycophenolate mofetil/azathioprine/methotrexate. Adverse reactions were reported only for mycophenolate mofetil (cytomegalovirus colitis and respiratory infection; grade 3 Common Terminology Criteria for Adverse Events v4.0). Relapse rate was significantly higher in patients started on first immune therapy (any) >30 days after onset (85.7%) compared to those treated early (31.2% (p=0.0272). 3.3.6 Mycophenolate mofetil in paediatric autoimmune and immune-mediated central nervous system conditions (international retrospective observational study). 44 children were included (30/44, 68.2% females). 43.2% (19/44) had proven or suspected autoimmune encephalitis, 31.8% (14/44) autoimmune inflammatory demyelinating CNS diseases, and 25% (11/44) other autoimmune/immune-mediated CNS conditions. Worst mRS was median 4 (range 2-6). Disease course was relapsing in 52.3% (23/44), monophasic in 38.6% (17/44), and chronic/chronic-progressive in 9.1% (4/44). Before mycophenolate mofetil, all patients received first-line (steroids: 44/44, 100%; intravenous immunoglobulin: 23/44, 52.3%; plasma exchange: 14/44, 31.8%) and 38.6% (17/44) second-line immune therapies (cyclophosphamide: 12/44, 27.3%; rituximab: 6/44, 13.6%). Median age at mycophenolate mofetil commencement was 9.3 years (range 1.4-16.4). Mycophenolate mofetil was started at median 9.5 months from onset (range 1-127; ≤6 months in 31.8%, 14/44). In 55% (22/40) of patients, mycophenolate mofetil was started only after ≥2 events had occurred. Median duration of treatment with mycophenolate mofetil was 18 months (mean 23.2, range 0.3-73). Median annualised relapse rate (excluding patients with chronic/chronic-progressive disease) was 0.52 (mean 0.86, range 0-3) before mycophenolate mofetil (excluding first events), and 0 (mean 0.36, range 0-4.64) during mycophenolate mofetil. 20.5% (8/39) patients relapsed during mycophenolate mofetil; compared to patients who did not relapse (31/49, 79.5%), these patients were younger (median age at onset 4.2 years versus 7.6), were more frequently females (8/8, 100% versus 21/31, 67.7%), had lower rate of second-line treatments before mycophenolate mofetil (1/8, 12.5% versus 15/31, 48.4%), a later commencement of mycophenolate mofetil (>6 months after onset in 7/8, 87.5% versus 22/35, 58.1%), and more frequently they were started on mycophenolate mofetil only after ≥2 events had occurred (7/8, 87.5% versus 14/35, 45.2%). Adverse reactions to mycophenolate mofetil occurred in 18.2% (8/44) of cases (6/8: grade 2, 2/8: grade 3 Common Terminology Criteria for Adverse Events v4.0). 4. Conclusion. The present thesis is a collection of ten works exploring several aspects of the clinical and therapeutic decision-making in paediatric autoimmune and immune-mediated inflammatory conditions. A growing array of immune therapies are becoming available in paediatric neurology, also derived from the experience in immune modulation from other fields of paediatrics and in adult patients. While quality data and definite recommendations are generally lacking, there are common themes emerging, such as the utility of early and aggressive immune therapy in certain clinical situations, such as in autoimmune encephalitis. The use of immune therapy is still characterised by a great heterogeneity between physicians in many neurological conditions, for examples as regards therapeutic plasma exchange and steroid sparers in anti-NMDAR encephalitis, reflecting not only the lack of definite recommendations, but also different treating habits and potential practical difficulties, such as with therapeutic plasma exchange in children and uncooperative patients. Even when recommendations do exist, such as for the use of intravenous immunoglobulin in neurology, current practice is not always adherent to the guidelines, suggesting both the need for greater adherence to existing recommendations and the need for recommendations to be updated to accommodate emerging indications. In other cases, finally, the utility and safety of treatments such as steroid sparing agents warrants further investigations in several fields of paediatric neurology, such as in anti-NMDAR encephalitis. In all cases, both currently accepted and future potential agents have adverse effects, which can be severe. A comprehensive understanding of the therapeutic aspects should not go without the ability to understand each clinical situation in all its facets, taking into considerations not only potential effects, adverse reactions and mechanisms of treatment agents, but also the pathophysiology, the severity of the acute disease, the risk of relapses and of permanent disability, in a complex ‘risk-versus-benefit’ determination and a tailored approach.