Overrepresentation of epilepsy in children with type 1 diabetes is declining in a longitudinal population study in Finland

AIM

The aim was to determine temporal changes in increased risk of epilepsy among children with type 1 diabetes.

METHODS

The incidence of epilepsy up to age 15 in children with prior type 1 diabetes was analysed regarding the general Finnish child population using data from the Finnish nationwide hospital register. Type 1 diabetes and epilepsy were identified by the International Classification of Diseases 9th and 10th revision codes. Epilepsy was defined according to ILAE guidelines. The analyses were done using negative binomial regression models.

RESULTS

Preceding type 1 diabetes was diagnosed in 6162 (0.91%) of the 679 375 general children population. Incidence rate of new-onset epilepsy among children with type 1 diabetes was higher than in controls (140 vs 82 per 100 000 person-years at risk, respectively). The excess incidence diminished with time (P = 0.033 for diabetes to birth cohort interaction), from over twofold in birth cohort 1990-1993 [incidence rate ratio 2.2 (95% CI 1.7-2.9)] to 40% in birth cohort 1998-2000 [1.4 (95% CI 1.001-1.9)].

CONCLUSION

In a population study setting, children with type 1 diabetes had an increased, but slowly declining risk of developing epilepsy. Future research may elucidate the underlying mechanisms.

Neurocognition in childhood epilepsy: Impact on mortality and complete seizure remission 50 years later

OBJECTIVE

To study associations of the severity of impairment in childhood neurocognition (NC) with long-term mortality and complete seizure remission.

METHODS

A population-based cohort of 245 subjects with childhood onset epilepsy was followed up for 50 years (median = 45, range = 2-50). Childhood NC before age 18 years was assessed as a combination of formal intelligence quotient scores and functional criteria (school achievement, working history, and psychoneurological development). Impaired NC was categorized with respect to definitions of intellectual functioning in International Classification of Diseases, 10th revision (R41.83, F70-F73). The outcome variables, defined as all-cause mortality and 10-year terminal remission with the 5 past years off medication (10YTR), were analyzed with Cox regression models.

RESULTS

Of the 245 subjects, 119 (49%) had normal childhood NC, whereas 126 (51%) had various degrees of neurocognitive impairment. During the 50-year observation period, 71 (29%) of the subjects died, 13% of those with normal and 44% of those with impaired NC. The hazard of death increased gradually in line with more impaired cognition, reaching significance in moderate, severe, and profound impairment versus normal NC (hazard ratio [Bonferroni corrected 95% confidence interval] = 3.3 [1.2-9.2], 4.2 [1.2-14.2], and 5.5 [2.4-12.3], respectively). The chance for 10YTR was highest among subjects with normal NC (61%), whereas none of those with profound impairment reached 10YTR. In the intermediate categories, the chance was, however, not directly related to the increasing severity of impairment.

SIGNIFICANCE

The severity of neurocognitive impairment during childhood shows a parallel increase in the risk of death. In comparison with normal NC, subjects with lower childhood NC are less likely to enter seizure remission. However, normal NC does not guarantee complete remission or prevent premature death in some individuals with childhood onset epilepsy.

Remission in epilepsy: How long is enough?

OBJECTIVE

The International League Against Epilepsy (ILAE) has proposed to expand the definition of remission to 10 years seizure-free with the last 5 years off antiepileptic drugs (AEDs). We examined if a 10-year remission is needed to predict the lowest recurrence risk.

METHODS

The population-based study cohort consisted of 148 patients with new-onset childhood epilepsy living in the catchment area of Turku University Hospital. They were prospectively followed for 44 years (median). Patients in first remission were prospectively followed for the duration of remission or possible relapse at 2 years in remission with the last year without antiepileptic drugs (AEDs), at 5 years in remission with the last 2 years without AEDs, and at 10 years with the last 5 years without AEDs. For comparison of the proportions of relapsed patients within each remission category exact Clopper Pearson 95% confidence intervals were used.

RESULTS

The magnitude of the relapse rate estimates off AEDs did not significantly improve when remission increased from 2 years (2YR) to 5 years (5YR) and further to 10 years (10YR). However, 10YR was a more sensitive measure of no relapse than 2YR. Among patients with remission on or off AEDs, the ability to predict lower relapse rate increased markedly from 2 to 5 years, and again from 5 to 10 years. The risk of relapse was virtually the same estimated after 2YR off AEDs as after 10YR on or off AEDs, except for patients with generalized epilepsy whose 2YR off AEDs was a weaker predictor than 10YR on or off AEDs.

SIGNIFICANCE

Given the modest differences in relapse rates between the 5 years seizure-free with last 2 years off medications definition and the 10 years seizure-free with last 5 years off medications, and the adverse impact of not being considered in remission, we propose that a return to the 5-year definition may be warranted.