FormalPara Key Points

Classic enzyme-inducing drugs such as phenytoin and carbamazepine show significant adverse effects on bone density and bone turnover markers over periods of only 6 or 12 months.

A prospective evaluation of osteodensitometry as well as indicators of bone metabolism showed no significant changes after 12 months of treatment with eslicarbazepine acetate, indicating the relative safety of this treatment, in contrast to the related drug carbamazepine.

1 Introduction

Epilepsy patients often suffer from bone metabolism disturbances due to reduced physical mobility forced by physical handicap, psychological movement restrictions, and avoidance of sports for fear of injury [1]. Additionally, treatment with anti-seizure medications (ASMs) may exert negative effects on bone metabolism, and thus contribute to an increased risk of fractures [2]. Given the required long-term treatment of epilepsy patients, osteopathy is a problem in all age groups, not only in the elderly and in postmenopausal women [3, 4].

Several studies evaluated the effects of classic enzyme-inducing drugs like phenytoin (PHT) and carbamazepine (CBZ) on bone density and bone turnover markers, often showing significant unwanted effects on bone health over periods as short as 6 or 12 months [3, 5,6,7,8]. The high risk of these older drugs with regard to bone health is thus widely acknowledged. The prevailing notion is that those hepatic cytochrome P450 (CYP450) inducers promote the metabolism of 25-hydroxyvitamin D (25-OHD) to biologically less active metabolites. This results in decreased bone mineralization, decreased intestinal calcium absorption, increased calcium mobilization from the skeleton to maintain eucalcemia and decreased bone density [9]. Whereas there is strong evidence for unwanted effects of these strong enzyme inducers, for oxcarbazepine (OXC), which is a less potent inducer of CYP450 than CBZ [10], findings are controversial, with several studies reporting reduced 25-OHD levels and elevated biomarkers suggesting increased bone turnover [5, 6, 11], but not all [12, 13]. Analogous to the effects postulated for CBZ and PHT, also for OXC it is has been suggested that it may cause minimal dose-dependent induction of the hepatic enzyme system P450 [10, 14]. Additionally, direct effects on bone cells may also contribute to the detrimental effects on bone health, in a similar manner as postulated for CBZ and PHT [15].

Eslicarbazepine acetate (ESL) is a third-generation ASM with improved pharmacokinetic properties compared with CBZ and OXC and possibly less metabolic effects, as recently shown with regard to lipid profiles [16, 17]. Effects of ESL on bone density have so far not been evaluated systematically.

The primary objective of the Bone Density Assessment in Patients with Partial Epileptic Seizures (BONAPARTE) study was to investigate whether adverse effects of ESL on bone density could be measured after a 12-month exposure period. In addition, the effects of ESL on bone turnover were investigated using laboratory indicators of bone metabolism.

2 Methods

This was a prospective, investigator-initiated, longitudinal observational study in patients with focal-onset seizures with or without secondary generalization who received treatment with ESL as either adjunctive treatment or monotherapy. The study was conducted at two sites, Freiburg Epilepsy Center and Marburg Epilepsy Center, after approval of the local ethics committees. The study was registered in the German register for clinical studies under the number DRKS00010430 with the official name BONAPARTE. All patients gave informed consent prior to inclusion in the study. The study was sponsored by the pharmaceutical company BIAL. Subjects were enrolled at Freiburg Epilepsy Center and Marburg Epilepsy Center between February 2018 and July 2020.

Inclusion criteria were:

  1. o

    Male or female patients ≥ 18 years of age

  2. o

    Patients with focal-onset seizures with or without secondary generalization

  3. o

    New treatment with eslicarbazepine acetate, either in monotherapy or as an adjunctive ASM

  4. o

    Written informed consent to participate in the trial and ability to adhere to the protocol

Exclusion criteria were:

  1. o

    Concomitant medication with phenobarbital, primidone, carbamazepine, phenytoin, oxcarbazepine, valproate, or corticosteroids

  2. o

    Known or planned pregnancy

  3. o

    Immobilized patients

  4. o

    Pre-existing manifest osteoporosis

BMD (g/cm2 and T-Score) was assessed twice, at the time of introduction of ESL and 1 year after continuous therapy. BMD of the lumbar spine and femur was assessed using the DEXA (dual energy X-ray absorptiometry) technique. At the Freiburg site, the device used was GE Healthcare, Prodigy (2020), S/N: UC2341LU03; at the Marburg site, Hologic Horizon-W, S/N: 301261M. Normal T-score is defined as a value − 1.0 or higher, osteopenia as a T-score value between − 1.0 and − 2.5, and osteoporosis as a T score value of − 2.5 or lower. Biochemical markers of bone metabolism [vitamin D3 (ng/mL), calcium (mmol/L), phosphorus (mmol/L), bone alkaline phosphatase (ALP) (µg/L), osteocalcin (µg/L) and parathyroid hormone (PTH) (pg/mL)] were also measured at the beginning and after 12 months of treatment.

2.1 Statistical Analysis

The SPSS 29 package program was used for the statistical analysis. Comparisons between biochemical and densitometric parameters at the beginning and after 12 months of treatment were performed using the paired samples t test. A p value of ≤ 0.05 was considered to be statistically significant.

The power was 1 − β 0.8 for paired t tests and an assumed effect size of d = 0.5 with an N of 26 and an α of 0.05.

3 Results

Originally, 45 patients had given their consent to participate in the study. Data from 19 patients could not be included in the analysis, 17 patients had incomplete data sets due to premature discontinuation of ESL treatment for clinical reasons, and 2 patients were “lost to follow-up”. The 26 patients with complete records from a 1-year follow-up period collected at the Freiburg and Marburg epilepsy centres ranged in age from 21 to 59 years. Five patients were postmenopausal women (patients 1, 6, 8, 18, and 24). Six of the 26 patients (5 men, 1 woman; mean age 45 ± 12 years; mean disease duration 9.3 ±14.6 years) had pathologic osteodensitometry findings (osteopenia) already at baseline (Figs. 1,2).

Fig. 1:
figure 1

Intra-individual comparison of bone density measurements in the femoral neck, at baseline and 12 months follow-up. Reduction in bone density with de novo osteopenia was observed in two patients. 12M FUP 12 months follow-up, BMD bone mineral density

Full size image
Fig. 2:
figure 2

Intra-individual comparison of bone density measurements in the lumbar spine, at baseline and 12 months follow-up. 12M FUP 12 months follow-up, BMD bone mineral density

Full size image

Of these six patients, three had a history of treatment with enzyme-inducing ASMs (two with carbamazepine, one with oxcarbazepine). At least one metabolic parameter was outside the standard range in 24/26 patients at baseline and in 23/26 patients at 12-months follow-up.

The mean number of concomitantly administered ASMs was 1.19 (± 0.98); maximum three ASMs. No patient took ASMs known to have enzyme-inducing effects. Nine patients received ESL as monotherapy. None of the patients received comedication with other drug classes that increase the risk of osteoporosis, such as selective serotonin reuptake inhibitors (SSRIs) or proton pump inhibitors. For demographic characteristics, see Table 1.

Table 1: Demographic characteristics and clinical variables
Full size table

The mean values for BMD at the femoral neck were 0.98 (± 0.15) g/cm2 at baseline and 0.96 (± 0.15) g/cm2 after 12 months of treatment. The T-scores were − 0.09 (± 1.14) and − 0.1 (± 1.14), respectively.

The mean values for BMD at the lumbar spine were 1.17 (± 0.16) g/cm2 at baseline and 1.16 (± 0.16) g/cm2 after 12 months of treatment. The T-scores were − 0.14 (± 1.27) and − 0.2 (± 1.27), respectively. At the group level, differences between osteodensitometry findings at baseline and after 12 months of ESL therapy were minimal and non-significant (Table 2).

Table 2: Statistical comparison of bone mineral density and bone metabolism parameters between baseline and after 1 year of ESL treatment. There were no statistically significant differences, even without correction for multiple testing
Full size table

Two patients (P12 and P18) developed de novo osteopenia measured at the femoral neck (Fig. 1): Patient 12, male, 31 years of age, mobile, ESL dosage 1600 mg/day under co-medication with brivaracetam 200 mg/d and perampanel 6 mg/d; bone density measurements were at baseline 0.96 g/cm2, T-score − 0.8, and after 12 months of treatment: 0.80 g/cm2, T-score − 2.0. The second patient (P18), female, 58 years of age, mobile, ESL monotherapy at a dosage of 1000 mg/day; bone density measurements were at baseline 0.96 g/cm2, T-score − 0.2; and after 12 months 0.81 g/cm2, T-score − 1.5. In patient 12, there was also a corresponding decrease in 25-hydroxycholecalciferol levels from 32.6 to 8.2 ng/mL (normal range > 20.0 ng/m), while parathyroid hormone levels increased from 35 to 50 pg/mL. In patient 18, vitamin D levels were decreased at both baseline and 12 months follow-up, while parathyroid hormone levels rose from 41 to 68 pg/mL (normal range 15–65).

In contrast, statistical comparison of group mean values for various blood bone metabolism marker levels did not show any significant differences (Table 2). No clinical manifestations of altered bone structure like fractures were reported during the study period.

4 Discussion

As previous studies have postulated a possible causal relationship between lower BMD and ASMs, not only for enzyme-inducing drugs but also for e.g. valproic acid [18, 19], gabapentin [20] and in animal models also with third generation ASMs such as levetiracetam [21], lamotrigine and topiramate [22], an assessment of the effects on bone health of a compound such as ESL appears interesting from a clinical perspective.

In this longitudinal prospective study investigating the effects of ESL on bone health, no significant group effects on BMD or on laboratory parameters of bone metabolism were found in 26 patients enrolled at the two participating study centres. Intra-individually, however, two patients passed the threshold for osteopenia measured at the femoral neck during the 1-year observation period.

ESL is a third generation ASM, which, like OXC, is a weak enzyme inducer of CYP3A4 and UDP-glucuronyltransferase, with a smaller effect size and effects on a smaller range of enzymes than carbamazepine. The overall (S)-licarbazepine to (R)-licarbazepine ratio of ESL is 20:1, which is considered to be advantageous towards OXC (ratio 4:1), since (S)-licarbazepine is thought to be more effective, less toxic and more efficient at crossing the blood–brain barrier than (R)-licarbazepine [23].

We hypothesized that ESL may exert similar effects on bone metabolism as reported in some of the studies assessing OXC, despite differences related to its molecular structure and partially distinct metabolic properties. In our study, we aimed to describe the effect of ESL on bone health for the first time, while at the same time adding information to the already existing body of evidence of weak enzyme inducers on BMD and on laboratory parameters of bone metabolism. Most studies evaluating ASM effects on bone mineral metabolism are cross-sectional. Longitudinal studies such as ours focusing on the effects of single agents provide better evidence for causal relationships.

Measurement of BMD is the most commonly used predictor of fracture risk in postmenopausal women and older men, thus it is a suitable primary outcome parameter when studying the effects of ASM on bone health.

Data on potentially detrimental effects on bone density of some enzyme inducing ASMs like CBZ and OXC are heterogeneous, yet undisputed for PHT. In a longitudinal study of premenopausal women with epilepsy on ASM monotherapy, Pack et al. [7] found that those patients receiving PHT sustained significant femoral neck bone loss after 1 year of treatment but did not have bone loss at other sites. In contrast, patients treated with carbamazepine, valproate, or lamotrigine had no loss in bone mass at any site. The femoral neck bone loss in young, premenopausal women treated with PHT monotherapy for 1 year was 2.6%, which was eight times greater than described in a larger cohort of women without ASM treatment [7]. In these women, lower serum 25-hydroxyvitamin D concentrations were associated with higher parathyroid hormone, bone alkaline phosphatase, and urine N-telopeptide levels, a biochemical pattern consistent with secondary hyperparathyroidism and increased bone remodeling. Not all studies found significant changes in BMD or bone metabolism during therapy with CBZ [24,25,26], yet several investigations showed increased bone turnover in epilepsy patients [3, 27].

In our study, the proportion of postmenopausal women was 19.2% (n = 5). Among the six patients with osteopenia at the baseline screening with DEXA scans there was only one postmenopausal woman aged 59, the other five patients were male, mean age 40.2 ± 9.7 years. This finding makes it seem likely that a relevant proportion of epilepsy patients outside the special risk group of postmenopausal women have an increased susceptibility to bone mass loss due to a variety of possible predisposing factors, including treatment with ASM.

For OXC, the evidence on catabolic effects on bone structure is unequivocal. One study in 14 children treated with OXC showed that serum ALP was elevated and BMD values were lower in patients after more than 1 year of treatment compared with healthy controls [5]. Since data collection in this cross-sectional study was performed only at one point of time, the causality of these findings with OXC treatment remains unclear. Notably, in our study, 24 of 26 patients had at least one metabolic parameter outside the standard range at baseline prior to introduction of ESL, pointing to the complex attribution of the effects of individual ASMs on bone health in cross-sectional studies. Another study in 34 children with newly diagnosed epilepsy reported a significant decrease in 25-OHD levels after 18 months of monotherapy with OXC, without a correlation between 25-OHD levels and BMD [11]. Comparing 44 children treated with OXC with a healthy control group, Babacan et al. [28] showed that serum levels of calcium, osteocalcin, 25-hydroxyvitamin D, and BMD did not differ significantly between the study and control group. In contrast, serum levels of parathyroid hormone, alkaline phosphatase and phosphorus were higher within the group on OXC treatment, and calcitonin levels were lower [28]. Mintzer et al. [6] reported 24 adult patients on monotherapy with OXC and 21 on CBZ to have reduced 25-OHD-levels and elevated parathyroid hormone and bone specific alkaline phosphatase after treatment periods longer than 1 year, as compared with healthy controls.

On the other hand, the study of Koo et al. [13] did not find negative effects on BMD and on biochemical markers in 41 drug naïve epilepsy patients who were treated with OXC for about 1 year, similar to findings of Cetinkaya et al. [12] in 28 adult epilepsy patients.

We here for the first time report the results from an observational real world longitudinal study focusing on the effects of ESL on BMD and on bone metabolism. Since effects on bone health of new ASMs are not evaluated during controlled trials, real world data with sufficiently long exposure periods are crucial to estimate the risk new agents may pose to bone health. The mean number of concomitantly administered ASMs in our patient population was 1.19 (± 0.98); maximum 3 ASMs. No patient took ASMs known to have enzyme-inducing effects. Nine patients received ESL as monotherapy. None of the patients received comedication with other drug classes that increase the risk of osteoporosis, such as selective serotonin reuptake inhibitors (SSRIs) or proton pump inhibitors. Thus, the observational conditions in our study allow us to draw conclusions on pharmacological effects of ESL on bone health.

It needs to be considered that besides the potentially detrimental effects of ASMs on bone health, ASM use may be associated with poor health, other medical conditions, lifestyle behaviour and neurological impairments that can also lead to greater rates of bone loss. Since immobilization was one of the exclusion criteria for our study, we aimed to rule out this potentially important factor as a confounding variable. In other studies on bone health in epilepsy patients an apparent association between ASM use and bone loss might be due, at least to some extent, to these confounding factors.

As significant effects were found neither on bone density nor on bone metabolism at the group level in our study, and as confounding variables in the form of anabolic effects are not likely to have influenced the results significantly, this prospective study design provides the first evidence that ESL has no detrimental effects on bone health after exposure to the drug for 1 year.

4.1 Limitations

Key limitations of the study are the small patient cohort and a relatively short follow-up period. Larger cohorts would provide more robust statistical power, and longer follow-up periods can reveal trends or consequences that may not manifest within a year. Furthermore, for the 2/26 patients in whom the osteopenia threshold was crossed during the treatment period, it cannot be automatically assumed that ESL treatment was the (only) cause; notably these patients were an otherwise healthy young man aged 31 years and a postmenopausal woman aged 51 years. Another limitation is the potential lack of diversity in the patient population. The patients included in the study may not adequately represent the broader population of epilepsy patients who might be prescribed ESL. This could limit the generalizability of the findings. Further information on larger patient cohorts with longer follow-up periods are needed to see if weak enzyme-inducing ASMs as ESL may exert adverse effects on bone health in susceptible subgroups.

5 Conclusion

Neither osteodensitometry nor bone metabolism parameters showed significant group effects after 1 year of treatment with ESL. Individual fluctuations were observed, however, which may warrant monitoring for longer follow-up periods. Bone health aspects should be part of pharmacovigilance for both old and new generation ASMs. Ideally, the effects of ASMs on bone metabolism should be included in the initial clinical trials and followed up in the long-term extension phase of the trials.