We thank Sander et al for their letter [1] commenting on our recent article [2], in which we reported that patients with type 1 diabetes mellitus are at an increased risk of developing epilepsy. Sander et al point out the strengths and challenges of the study. First, our researchers conducted a retrospective population-based cohort study using claims data from the Taiwan National Health Insurance Research Database (NHIRD). The coverage rate of the Taiwan National Health Insurance programme (NHI) reached 99% of the 23 million residents of Taiwan in 1998. The NHIRD is renewed annually and includes registries for beneficiaries, disease diagnosis records, medical prescriptions and other medical services. It is a large, population-based sample effectively including all Taiwanese residents. The registries of NHI claims primarily serve the purpose of administrative billing and do not undergo verification for scientific purposes. Therefore, we were unable to contact the patients directly to obtain more information, because of the anonymity assured by the identification numbers.Second, metabolic abnormalities of type 1 diabetes, such as hyper-and hypoglycaemia, may have a damaging effect on the central nervous system (CNS) and be associated with significant long-term neurological sequelae. Sander et al [1] point out that our data included hypoglycaemia as a potential confounder, although hyperglycaemia is not considered in the statistical analysis. This was because we used an in-hospital diagnosis (ICD-9-CM [www.icd9data.com/2007/Volume1] 251.0-251.2 and 775.6), whereby a patient was deemed to have hypoglycaemia occurring during follow-up. As suggested by Sander et al, patients with hyperglycaemia might have been underestimated, because this may not translate into a recoded diagnosis.Third, we found that younger age was associated with an increased risk of developing epilepsy; those aged at least 6 years were 40 percent less likely to have epilepsy than those under 6 years. This is consistent with previous studies indicating that young age is an important risk factor for electroencephalographic abnormalities [3].Fourth, 59 people were reported to have 'developed' epilepsy over the course of the study, but 27% of these (16 of 59) had a prior history of epilepsy, which would generally make these prevalent rather than incident cases. Actually, epilepsy is characterised by spontaneous recurrent seizures, but can be a pervasive condition of which seizures are just one expression [4]. Several mechanisms explain how epilepsy and comorbidities are associated, including shared risk factors and bidirectional * Chia-Hung Kao