The present study is the first to evaluate the risk of delirium associated with usage of antiepileptic drugs using data from a large sample; however, the data are self-reported. Nevertheless, our results indicated that antiepileptic drugs were associated with delirium. Moreover, even after adjustment for possible confounding factors, lacosamide, lamotrigine, levetiracetam, and VPA were detected as signals for delirium in each antiepileptic drug.
Previous reviews have revealed that patients who consume antiepileptic drugs are at a medium-to-high risk of delirium development [14]. Wierenga et al. [15] reported that 26% of their target patients had antiepileptic drug-associated delirium. Our results support these findings and suggest that use of antiepileptic drugs may pose a non-uniform risk for delirium.
In the previous evaluation of each antiepileptic drug that detected signals, two case reports of delirium were associated with levetiracetam [10, 16]. Similar case reports have been identified for lamotrigine; however, they also include reports on lamotrigine administered in combination with VPA [17, 18]. Furthermore, some case reports on VPA-induced delirium are available [9, 19]. Similar to the evidence in these case reports, our findings also show that antiepileptic drugs may induce delirium. In contrast, for some antiepileptic drugs, a signal for delirium was not detected in the presence of these risk factors when stratified by other risk factors for delirium, such as age ≥ 70 years or the use of BZAs (Table 3). This indicates that even if antiepileptic drugs increase the risk of delirium, they may not be as strong a risk factor as age ≥ 70 years or BZA usage.
Some of these antiepileptic drugs have been reported to have caused psychotic disorders not limited to delirium. Particularly, psychotic disorders caused by levetiracetam are well known [7, 8]. One study noted that lacosamide also induced psychotic disorders; however, the incidence of these disorders was not as high as that of levetiracetam-induced psychotic disorders [20]. The same study also identified depression and paranoid ideation as the psychological adverse events associated with lacosamide consumption. The association between depression and delirium has been reported [21], and antiepileptic drug-induced psychotic disorders may also include delirium.
It is also important to know that some antiepileptic drugs are effective against delirium. In particular, VPA has shown efficacy against delirium in patients admitted to the intensive care unit [22]. Some studies have shown a statistically significant reduction in the incidence of agitation and delirium after VPA therapy [11, 23]. VPA is a therapeutic agent for bipolar disorder and is particularly effective for acute mania; therefore, it is expected to suppress agitation [24]. However, in many studies that showed a relationship between delirium and VPA, therapeutic agents (such as antipsychotic drugs and dexmedetomidine) were used concomitantly with VPA for delirium management; therefore, it is unclear whether VPA alone is effective for delirium. Our findings showed that VPA had a significantly higher ROR for delirium even when adjusted for known risk factors. This was also the same when stratified in those of age ≥ 70 years. Based on previously published case reports on VPA-induced delirium as well as our results, VPA may be a risk factor for delirium. These conclusions require additional research.
The strengths of this study are its large sample size, use of the more informative JADER database, and careful adjustment for potential confounders (such as concomitant medications and comorbidities). However, this study also has several limitations. First, the results were obtained from the JADER database, which operates on a voluntary reporting system; thus, they should be interpreted with caution. Most importantly, because the database is updated via a passive reporting system, many biases are present (including underreporting, overreporting, missing data, exclusion of healthy participants, lack of denominators for incidence estimates, and presence of confounding factors) [25]. When using an adverse event database, the "true" risk of delirium cannot be assessed without information on the total number of patients treated with antiepileptic drugs. Second, the accuracy of delirium diagnosis may be poor. Tools such as the Delirium Rating Scale and the Confusion Assessment Method are available for the evaluation of delirium. However, the basis for the diagnosis of delirium in the JADER database is unknown. Finally, the Food and Drug Administration Adverse Event Reporting System notes that adverse event reporting increases during the first 2 years after drug approval; this is followed by a rapid decline in the reporting rates [26]. Therefore, we adjusted our analyses for the reporting year to minimize these confounding factors.