We investigated the association of serum antibodies to glutamic acid decarboxylase (GADab) with early start of insulin treatment (< or = 1 year from diagnosis, or < or = 2 years from diagnosis) using data from a representative sample of 374 adult-onset insulin-treated individuals from the Tasmanian Diabetes Register. Furthermore, we examined whether this association was stronger than the phenotypic characteristics (age at diagnosis, sex, family history of diabetes, level of obesity, duration of diabetes) often used for diabetes classification. In this cohort, 35.9% of males and 38.5% of females were GADab positive. Within the first year from diagnosis, 78.4% of GADab positive people compared to 44.0% of GADab negative people (p < 0.001) had started insulin treatment. Univariate associations with insulin treatment < or = 1 year from diagnosis included GADab positivity, no family history of diabetes, lower BMI for men, and GADab positivity and lower BMI for women. In multivariate models, significant associations with insulin treatment < or = 1 year from diagnosis included a family history of diabetes (OR = 0.47, 95% CI = 0.23-0.95) and GADab positivity (OR = 2.19, 95% CI = 1.01-4.73) for men, but only GADab positivity (OR = 7.53, 95% CI = 3.09-18.30) for women. Age at diagnosis was not associated with insulin treatment < or = 1 year or < or = 2 years from diagnosis for either sex. These findings indicate that a positive GADab test result is strongly associated with start of insulin treatment within 1 or 2 years from diagnosis, more so than characteristics such as level of obesity and age at diagnosis.