The expanding range of disease modifying therapies (DMT) for relapsing-remitting multiple sclerosis (RRMS) has led to increased interest in the relative effects of different DMTs. Previous mixed treatment comparisons (MTCs) have used different methods to address similar questions highlighting the need for a consistent approach to the assessment of treatments in RRMS.

We compared the methodology of six published MTCs of DMTs for RRMS identified by a systematic search of the literature. We assessed sources of evidence, DMTs included, outcomes reported and methods of data synthesis.

All six MTCs were based on systematic reviews that included randomized controlled trials (RCTs). MS relapse was reported as the rate ratio based on annualised relapse rates (four MTCs) and as odds ratios or relative risk (one MTC each) based on the proportion with relapse. The analysis of relapse included between sixteen and twenty-seven RCTs and seven to twenty DMTs in different MTCs. One MTC reported both disability progression confirmed after three months (CDP3M) and disability progression confirmed after six months (CDP6M) as hazard ratios. One MTC combined CDP3M and CDP6M as a single outcome. One MTC reported only CDP3M based on hazard ratios. Two MTCs reported only CDP6M as either odds ratios or risk ratios (one MTC each). In one MTC the definition of disability progression was not reported. The analysis of disability included between seven and twenty-six RCTs and between six and nineteen DMTs in different MTCs. All six MTCs fitted a random effects MTC model using either Bayesian (four MTCs) or frequentist (two MTCs) methods.

There is substantial heterogeneity between published MTCs in RRMS with regard to inclusion criteria, outcome definitions, effect measures and statistical methods. There is a clear requirement for a consistent approach to health technology assessment of DMTs for RRMS.

Symptoms of obsessive–compulsive disorder (OCD) have been described in neuropsychiatric syndromes associated with streptococcal infections. It is proposed that antibodies raised against streptococcal proteins cross-react with neuronal proteins (antigens) in the brain, particularly in the basal ganglia, which is a brain region implicated in OCD pathogenesis.

To test the hypothesis that post-streptococcal autoimmunity, directed against neuronal antigens, may contribute to the pathogenesis of OCD in adults.

Ninety-six participants with OCD were tested for the presence of anti-streptolysin-O titres (ASOT) and the presence of anti-basal ganglia antibodies (ABGA) in a cross-sectional study. The ABGA were tested for with western blots using three recombinant antigens; aldolase C, enolase and pyruvate kinase. The findings were compared with those in a control group of individuals with depression (n = 33) and schizophrenia (n = 17).

Positivity for ABGA was observed in 19/96 (19.8%) participants with OCD compared with 2/50 (4%) of controls (Fisher's exact test P = 0.012). The majority of positive OCD sera (13/19) had antibodies against the enolase antigen. No clinical variables were associated with ABGA positivity. Positivity for ASOT was not associated with ABGA positivity nor found at an increased incidence in participants with OCD compared with controls.

These findings support the hypothesis that central nervous system autoimmunity may have an aetiological role in some adults with OCD. Further study is required to examine whether the antibodies concerned are pathogenic and whether exposure to streptococcal infection in vulnerable individuals is a risk factor for the development of OCD.

Obsessions and compulsions may occur in the post-streptococcal disorders Sydenham's chorea and paediatric autoimmune neuropsychiatric disorders associated with streptococcus (PANDAS). The proposed mediators are anti-basal ganglia antibodies (ABGA)

We tested the hypothesis that post-streptococcal autoimmunity may have a role in ‘idiopathic’ obsessive–compulsive disorder (OCD)

We examined 50 children with OCD for ABGA using enzyme-linked immunosorbent assay (ELISA) and western immunoblotting. The findings were compared with paediatric autoimmune (n=50), neurological (n=100) and streptococcal (n=40) controls.

The mean ABGA binding on ELISA was elevated in the patient cohort compared with all control groups (P<0.005 in all comparisons). Western immunoblotting revealed positive antibody binding (as seen in Sydenham's chorea) in 42% of the patient cohort compared with 2–10% of control groups (P<0.001 in all comparisons)

Our findings support the hypothesis that central nervous system autoimmunity may have a role in a significant subgroup of cases of OCD. Further study is required to examine whether the antibodies concerned are pathogenic.