In June, 2006, a 34-year-old woman presented to our diabetic outpatient clinic for routine follow up. She had been diagnosed with rheumatoid arthritis in February, 1997, and treated with penicillamine and prednisolone. In October, 1999, she developed diabetes mellitus. Steroid-induced diabetes was excluded because the patient remained diabetic despite discontinuing the steroid treatment, and she was treated with the oral hypoglycaemic medication, glibenclamide. 1 year later, her glycaemic control was poor, with a glycosylated haemoglobin (HbA1c) of 10·1% on maximum dose of oral medication. She was therefore started on 18 U daily of a biphasic isophane insulin and her diabetes control improved (HbA1c 8·6%). In September, 2002, after several disease-modifying anti-rheumatic drugs failed, she was treated with the anti-tumour necrosis factor (TNF)α antibody, adalimumab. This treatment resulted in substantial improvement in her symptoms. After starting anti-TNFα therapy our patient’s insulin requirement decreased to 8 U with a surprising improvement of her glycaemic control: her HbA1c in March, 2003, was 6·7%. When we reviewed her in 2006, we hypothesised that she had an autoimmune form of diabetes and that the anti-TNFα therapy aff ected the progression of the autoimmune T-cell response. Satisfactory endogenous insulin production and c-peptide concentrations after a glucagon stimulus indicated residual islet function, and further laboratory results confi rmed that this patient had a slowly progressive form of autoimmune or type 1 diabetes, known as latent autoimmune diabetes in adults (LADA). Our patient was seropositive for anti-cyclic citrullinated peptide (CCP) antibodies, anti-glutamic acid decarboxylase (GAD) autoantibodies, and weakly positive for anti-islet cell antibodies (ICA). T-cell responses were analysed with a cytokine ELISPOT to detect interferon-γ (a pro-infl ammatory cytokine) and interleukin-10 (IL-10) (an anti-infl ammatory cytokine). Her islet auto-reactive T cells had a suppressor function and produced IL-10 in response to several diabetes mellitus-associated, HLADR4 restricted auto-antigens (fi gure). Similar results are found in healthy controls. HLA type was A01,02; B08,51; DR1,. Residual intact insulin producing cells maintained an adequate preprandial and postprandial glucose level after discontinuing insulin for 1 month; at her most recent follow-up, in October, 2009, glycaemic control and insulin requirement were stable. Concomitant occurrence of rheumatoid arthritis and type 1 diabetes has been previously observed. Moreover, a specifi c association between autoimmune forms of diabetes and rheumatoid arthritis has recently been reported, especially in patients positive for anti-CCP antibodies. The diagnosis of LADA relies mainly on seropositivity for ICA or for autoantibodies to GAD. In addition, as in classical type 1 diabetes, patients with LADA commonly have T cells that react against islet autoantigens. HLA DR4 is associated with autoimmune diabetes, including LADA. We concluded that the antiTNF therapy may have directly or indirectly contributed to the induction of IL-10-secreting regulatory cells with a consequent resolution of the infl ammation in the pancreatic islets.
