Results
Baseline data on the 739 subjects are shown in Table 1 . These comprised 362 head trauma subjects without TBI, 297 with mild TBI and 80 with moderate-to-severe TBI. Altogether 42 patients developed new-onset seizure(s) during the follow-up period, all attributable to structural or metabolic causes except for one subject who had genetic epilepsy. Alcohol-related seizures were observed in 19 out of the 42 subjects (45.2%), and were most frequent among those with no TBI. Most subjects developing alcohol-related new-onset seizures (17/19, 89.5%) had been intoxicated at the time of the index trauma; only one with moderate-to-severe TBI and another without TBI were not.
As expected, the subjects with moderate-to-severe TBI developed new-onset seizures significantly more frequently than others during the follow-up (p=0.001), and mortality during the follow-up was also significantly higher among those with moderate-to-severe TBI initially than in those without TBI (HR 3.64, 95% CI 2.35 to 5.64, p<0.001). On the other hand, those with mild TBI had a lower risk of death than those without TBI (0.62, 0.39 to 0.99, p=0.047). Of the 42 subjects with new-onset seizures, 14 (33%) died during the follow-up (including 5 with alcohol-related seizures), but the time between the first seizure and death ranged from 0.26 to 9.54 years (median 3.5 years). There was significant heterogeneity in age and sex between the groups (p<0.01), the subjects with mild TBI being the youngest, followed by those with no TBI. The proportion of men was greatest among the subjects with no TBI and smallest among those with mild TBI, and an alcohol-related index injury was most common among the subjects with no TBI.
Significant and independent risk factors predicting new-onset seizures were identified using the Cox proportional hazard regression model ( Table 2 ). After adjusting for age and sex, we found alcohol-related index injury, moderate-to-severe TBI as the index trauma and a preceding psychiatric disease to be significant predictors of a new-onset seizure. We also analysed the subjects separately according to the severity of their head trauma, whereupon the significant independent predictors of a new-onset seizure were a preceding psychiatric disease (HR 4.32, 95% CI 1.24 to 15.12, p=0.022) and an alcohol-related index injury (3.47, 1.26 to 9.56, p=0.016) among subjects without TBI and an alcohol-related index injury (5.08, 1.07–24.23, 0.041) and age (1.04 per year, 1.01 to 1.08, p=0.027) among the subjects with mild TBI, while none of the parameters was a significant predictor of new-onset seizures among the subjects with moderate-to-severe TBI, perhaps because of the smaller number of subjects. The 21 subjects with a preceding psychiatric disease included four who suffered from new-onset seizures, two of whom had alcohol withdrawal as one reason for the appearance of a seizure problem.
Cumulative rates of occurrence of new-onset seizures according to severity of the index head trauma are shown in Figure 1, where again significant heterogeneity existed between the groups (p=0.001). New-onset seizures occurred most frequently among the subjects with moderate-to-severe TBI, but they also occurred more frequently among those with no TBI than among those with mild TBI, although the difference between these two groups was not significant. None of those with moderate-to-severe TBI developed a new-onset seizure problem after January 2004.
(Enlarge Image)
Figure 1.
Kaplan-Meier curves showing cumulative rates of new-onset seizures by severity of the head trauma.
Cumulative curves for the occurrence of alcohol-related new-onset seizures stratified by the presence and absence of TBI in connection with the index trauma are shown in Figure 2. Interestingly, the subjects without TBI developed alcohol-related seizures significantly more frequently (p=0.029). In a separate multivariable Cox analysis adjusted for age, sex, severity of the index head trauma and a history of psychiatric disease, alcohol-related index injury appeared to be the only independent predictor of the occurrence of an alcohol-related new-onset seizure problem (HR 12.13, 95% CI 2.70 to 54.50, p=0.001).
(Enlarge Image)
Figure 2.
Kaplan-Meier curves showing the cumulative occurrence of alcohol-related new-onset seizures by presence/absence of traumatic brain injury.