Valproic acid and KD
In comparing the KD with VPA and PHT, it was of interest to examine potential synergistic effects. In particular, given the earlier finding that the KD worsened MES-induced seizures (33), we sought to determine the role of each drug in KD animals exposed to MES. Our rationale, therefore, was both to compare the KD with these standard AEDs and to examine KD–drug combination, both in a test in which the KD was protective (PTZ threshold) and in one in which it was detrimental (MES).
The elevation of PTZ seizure threshold in KD-fed animals is age dependent (16). In these experiments, the increase in seizure threshold afforded by the KD was greater in young animals (P28 at diet onset) than it was in older animals (P37) fed the same diet. Although the P28 and P37 rats were fed the KD for different durations (35–39 and 20 days, respectively), we do not feel that differences in the duration of feeding at 20 days and beyond compromise these conclusions. Appleton and DeVivo (13) described a plateau in electroconvulsive threshold after 20 days, and PTZ seizure threshold has been shown to be dependent on the age at which the diet was initiated (16) but not on the duration of feeding (15). Interestingly, the KD was as effective as a 300-mg/kg dose of VPA in younger animals (P28), but was only as effective as a 150-mg/kg VPA dose in older animals (P37). These results show that the KD is not as effective in animals that start the diet at an older age, support earlier findings that the efficacy of the KD is age dependent (12,16,17,34), and suggest that there is a “window of opportunity” during which KD treatment is maximally efficacious.
The effects of VPA, unlike treatment with the KD, were not dependent on age. If one measures the incremental PTZ-threshold increase induced by VPA (i.e., for each 150 mg/kg given), there is an increase in PTZ threshold of approximately +20mg PTZ/kg for all ages tested. If one considers the VPA-induced percentage change, the effects are not much different. These observations indicate that the threshold elevation by VPA is not age dependent.
Interpretation of the effects of diet and drugs on MES is more difficult. As noted previously (33), MES-induced seizures were more severe in KD-fed animals compared with those fed normal rodent chow, an unanticipated finding we cannot explain. Whereas seizure severity for saline-injected animals was not significantly greater than that for NCR animals in this study (likely due to the small numbers of control animals used), VPA and PHT each reduced seizure severity similarly in both diet groups. For the VPA experiments, a possible complication is that the same animals were used for both MES- and PTZ-induced seizures. It is possible that AED levels had not declined to zero in the 2-day interval between these two tests, but two considerations argue against a major role for residual effects of either VPA or PHT. First, the plasma half-life of VPA has been reported as from 2.57 to 3.05 h in rats aged 18–901 days for a dose of 400 mg/kg i.p. (35) and 2.3 ± 0.7 h for an oral dose of 600 mg/kg (36). Although it remains possible that a more lasting physiologic effect remained, the 2-day interval represents 16 times the half-life, suggesting a negligible residual level of VPA. For PHT the half-life is dose dependent, just >1 h for an i.v. dose of 10 mg/kg but nearly 4.5 h at 50 mg/kg (37). Extrapolation to a dose of 75 mg/kg (even though the relationship is nonlinear) suggests a half-life of 6 h, representing 7 times the half-life for the higher dose over the 2-day interval. Second, the marginally increased seizure threshold observed with PTZ infusion in KD-fed rats, relative to rodent chow–fed rats, was nearly the same in control (saline) injected animals as it was in VPA-injected animals. A similar relationship was observed for PHT. These results are consistent with minimal residual drug or physiologic effects and further suggest a minimal role for either residual VPA or PHT.
A major finding of the present study is that VPA and the KD exhibit synergistic threshold effects. These additive effects were seen in both age groups tested. As we did not measure plasma or brain levels of either VPA or PHT, we suggest that at least four possibilities might account for this finding. First, plasma levels of free fatty acids (FFAs) and free VPA have been shown to be positively correlated (38). As elevated plasma concentrations of fatty acids are found in both KD-fed patients (39) and rats (13), higher FFA levels in the blood likely compete for VPA-binding sites in plasma. Displacement of VPA associated with a subsequent increase in free VPA by FFAs may, at least in part, account for our observations (40). Second, high concentrations of medium-chain fatty acids are known to disrupt cell membranes and open tight junctions in the intestine, as shown by increased flux of cefoxitin (41). If high concentrations of FFA associated with the KD also disrupt tight junctions in the BBB, VPA entry into the brain would increase. Third, conditions that elevate BHB have been shown to induce an up-regulation of MCT1, a transporter for BHB across the blood–brain barrier (42). It has also been shown that VPA uptake is mediated by a medium-chain fatty acid transporter (40). Consumption of a ketogenic diet might lead to upregulation of the medium-chain transporter and increased VPA transport or, alternatively, competition from elevated levels of fatty acids could lead to decreased VPA transport. Which effect might dominate is unclear. A fourth possibility is that the KD and VPA share a common mechanism(s) of action.
It is important to note, however, that the KD has been shown to enhance hepatotoxicity when combined with VPA (43). These authors reported that the KD, perhaps via carnitine deficiency or inhibition of oxidative phosphorylation, enhanced VPA toxicity in human patients (43). Although our data show that treatment with both KD and VPA may improve seizure control in animals acutely, extrapolation of these data to humans is perilous, and the hepatoxicity studies cited suggest that this therapeutic strategy, especially if instituted over the long term, may not be warranted in human patients.
Phenytoin and the KD
The KD and PHT are effectively different. PHT treatment has been shown to attenuate MES seizure activity, but to be ineffective in elevating PTZ seizure threshold (23). Conversely, the KD only elevates PTZ seizure threshold and is ineffective against MES-induced seizure activity (15,33,44,45). Our results confirm these findings (Figs. 4 and 5). Indeed, KD animals tend to exhibit more severe MES-induced seizure activity (33). Despite an incomplete understanding of mechanisms of seizures and AEDs, this work provides evidence favoring a rational polytherapy between the “complementary”actions of KD (raising threshold) and PHT (reducing severity) if the elevated neurotoxicity noted later can be tolerated.
Diet- versus drug-induced neurotoxicity
There were neurobehavioral deficits for high doses of both VPA and PHT, but not for KD treatment alone. In accordance with work showing that ketogenic animals did not perform differently from controls on spatial learning or exploratory behavior tasks (18), these data support the findings that there are few KD-associated cognitive side effects (14). The rotorod is, perhaps, the most challenging of the six behavioral tests used, and the combination of KD and PHT resulted in significantly more frequent failures (animals falling off the rod in each of three 1-min trials). Although the highest dose of PHT exceeds at least one published TD50(23), rats fed rodent chow did not experience an increase in failures when injected with this dose of PHT. The basis of this apparent neurotoxic synergy between the KD and high-dose PHT is unknown. Whereas the seizure-protective effects of the two treatments are complementary, the neurotoxicity finding suggests caution in combining the KD with higher doses of PHT.
In summary, there are three major findings in this study. First, the KD is as effective as a high dose of VPA (i.e., 300 mg/kg) in young animals (P28 at onset). Second, the KD acts synergistically with VPA to markedly elevate PTZ seizure threshold. Third, the KD improves seizure control without the associated neurobehavioral contraindications that accompany administration of high doses of either VPA or PHT. Collectively, these results suggest that the KD ranks among VPA and PHT as an effective treatment for seizures, with no obvious drug-associated neurobehavioral effects. These findings may provide insights into future directions of rational use of the KD and its potential role—and limitations—in polytherapy. Importantly, however, expectations based on recognition that the KD acts synergistically with VPA and complementary to PHT should be tempered by findings of metabolic disturbance associated with VPA monotherapy (46,47), possible hepatotoxicity (43), and some evidence (Table 2) for enhanced neurotoxicity between the KD and high doses of PHT.