Autoimmune polyglandular syndromes result from the failure of multiple endocrine glands due to an autoimmune process. In autoimmune polyglandular syndrome type 3, autoimmune thyroiditis is associated with autoimmune diseases other than Addison’s disease and hypoparathyroidism.1

We report two identical twins with the rare combination of autoimmune thyroiditis with stiff person syndrome (SPS). SPS presents with progressive muscle rigidity and spasms of axial muscles and extremities.2 The majority of SPS patients have serum autoantibodies against glutamic acid decarboxylase (GAD)65,3 an enzyme responsible for the synthesis of γ-aminobutyric acid (GABA), the major inhibitory neurotransmitter of the central nervous system (CNS). GAD65 antibodies probably play a direct pathogenetic role in SPS because they are found within the cerebrospinal fluid (CSF) and block the activity of GAD65.4,5 The management of SPS includes GABAergic agents, such as diazepam. For unresponsive patients, intravenous immunoglobulins may represent an alternative.6

Twin A, a 34-year-old man, presented with a 5-year history of stiffness and spasms that involved the trunk and thigh muscles. The spasms affected walking, led to falls, contusions and a fracture. They never occurred during sleep, but were precipitated by several stimuli. The disease in twin B had begun 10 months before consultation. His spasms were less frequent and severe. In both brothers, electromyography had demonstrated continuous motor unit activation and abnormal exteroceptive reflexes in antagonistic muscles.7 Anti-GAD antibodies were positive in serum and CSF. Both twins had euthyroid Hashimoto’s thyroiditis with antithyroid peroxidase and antithyroglobulin antibodies. Diabetes mellitus, hypopituitarism, hypoparathyroidism and Addison’s disease were excluded.

In light of an inadequate response to diazepam, the humoral immune pathophysiology of SPS and the efficacy but expense of intravenous immunoglobulins, we aimed to improve the long-term course with rituximab, a monoclonal antibody that depletes B cells. After written consent, the patients were treated in a double-blind crossover design with either rituximab or saline. A prescheduled switch between rituximab and placebo arms was at week 18 and prescheduled follow-up visits at weeks 36 and 54. At week 54, both twins received an open-label second course of rituximab 2 weeks apart. Premedications were identical for rituximab and placebo (1 g paracetamol and 100 mg prednisolone). Twin A was randomly assigned to receive placebo in the first course and twin B to receive rituximab (fig 1). All infusions were tolerated well. The outcome (table 1)6 does not support a profound benefit from rituximab although rituximab may had prevented a more severe disease course.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Study design demonstrating the random allocation of rituximab or placebo to the identical twins at the beginning of the double-blind trial period (week 0). Black squares indicate the first of each set of two rituximab infusions (1000 mg) given 2 weeks apart. Open circles indicate placebo.

Rituximab was previously only used in one patient in whom a single low dose had induced a long-term clinical remission.8 In the reported woman, the GAD-antibody titres were, however, unusually low. It may also be necessary for rituximab to penetrate into the CSF. In relapsing–remitting multiple sclerosis, rituximab was effective in reducing CNS demyelination9 but the lack of inflammation in SPS may render it difficult for rituximab to enter the CNS.

In conclusion, our findings support a genetic factor in the aetiology of SPS but not a clinical benefit of rituximab in its treatment.