We analyzed the outcome of allogeneic hematopoietic stem cell transplantation (HSCT) over the past 2 decades. Between 1992 and 2009, 953 patients were treated with HSCT, mainly for a hematologic malignancy. They were divided according to 4 different time periods of treatment: 1992 to 1995, 1996 to 2000, 2001 to 2005, and 2006 to 2009. Over the years, many factors have changed considerably regarding patient age, diagnosis, disease stage, type of donor, stem cell source, genomic HLA typing, cell dose, type of conditioning, treatment of infections, use of granulocyte-colony stimulating factor (G-CSF), use of mesenchymal stem cells, use of cytotoxic T cells, and home care. When we compared the last period (2006-2009) with earlier periods, we found slower neutrophil engraftment, a higher incidence of acute graft-versus-host disease (aGVHD) of grades II-IV, and less chronic GVHD (cGHVD). The incidence of relapse was unchanged over the 4 periods (22%-25%). Overall survival (OS) and transplant-related mortality (TRM) improved significantly in the more recent periods, with the best results during the last period (2006-2009) and a 100-day TRM of 5.5%. This improvement was also apparent in a multivariate analysis. When correcting for differences between the 4 groups, the hazard ratio for mortality in the last period was 0.59 (95% confidence interval [CI]: 0.44-0.79; P < .001) and for TRM it was 0.63 (CI: 0.43-0.92; P = .02). This study shows that the combined efforts to improve outcome after HSCT have been very effective. Even though we now treat older patients with more advanced disease and use more alternative HLA nonidentical donors, OS and TRM have improved. The problem of relapse still has to be remedied. Thus, several different developments together have resulted in significantly lower TRM and improved survival after HSCT over the last few years.

The prevalence and titre of epitope-specific autoantibodies to glutamic acid decarboxylase (GAD65) in 155 insulin-dependent diabetic (IDDM) and 9 GAD65 antibody (Ab)-positive healthy children were determined using four GAD65/67 chimaeric molecules which discriminate among the N-terminal (N), middle (M) and C-terminal (C) epitopes of GAD65. Radioligand binding assays for IgGAb used immunoprecipitation of in vitro translated 35S-GAD. We found autoantibodies to GAD65 in 116 of 155 (75%), to GAD67 in 19 of 155 (12%) (p < 0.0001) and to the GAD65-N-67 chimaera in 25 of 155 (16%) (p < 0.0001) IDDM sera. GAD67Ab were found almost exclusively (17 of 19, 89%) in GAD65Ab-positive sera and the levels of GAD67Ab correlated with those of GAD65Ab (r2 = 0.5913; p = 0.009). GAD65Ab directed to GAD65-M were found in 104 of 155 (67%), to GAD65-C in 104 of 155 (67%) and to GAD65-M + C in 116 of 155 (75%) of IDDM sera, and indicated reactivity to at least two distinct epitopes. Among the nine GAD65Ab-positive healthy children, two (22%) were also positive with GAD67, nine (100%) with GAD65-M + C, seven (78%) with GAD65-M, eight (89%) with GAD65-C and two (22%) with GAD65-N-67. Titres of GAD65Ab (p = 0.007), GAD65-C-Ab (p = 0.002) and GAD65-C + M-Ab (p = 0.003), but not of GAD65-M-Ab (p = 0.101) were significantly higher in IDDM than in healthy children. We conclude that GAD65Ab in IDDM and healthy children are directed to middle and C-terminal epitopes, and propose that levels of antibodies specifically directed to the carboxy-terminal end of GAD65 may distinguish IDDM from healthy children.

We studied the risk for diabetes of glutamate decarboxylase (GAD65Ab) and islet cell (ICA) autoantibodies in non-diabetic pregnant mothers and their children. Pregnancy and cord blood sera were collected in 1970-87 from about 35,000 mothers who delivered a child in the city of Malmö, Sweden. A total of 42 mothers were identified in 1988 who, 1-18 years after their pregnancies, had developed either insulin-dependent (n = 22) or non-insulin dependent (n = 20) diabetes mellitus. First, in 123 pregnant mothers selected as controls, 0.8% had GAD65Ab and 0.8% ICA. Second, among the mothers with non-insulin dependent diabetes, 7/20 (35%) had GAD65Ab eight months to 13 years, 10 months before clinical diagnosis. Third, in mothers who later developed insulin-dependent diabetes, 12/22 (55%) had GAD65Ab and 10/22 (45%) had ICA in pregnancies preceding the clinical diagnosis by 13 months to 9 years, 4 months. In 1996, none of the children born to the 42 mothers have developed diabetes. GAD65Ab and ICA in non-diabetic pregnancies may predict insulin-dependent diabetes in the mother but not necessarily in the offspring.