Effect of switching hepatic enzyme-inducer antiepileptic drug to levetiracetam on bone mineral density, 25 hydroxyvitamin D, and parathyroid hormone in young adult patients with epilepsy
We sought to determine the effect of changing phenytoin therapy on bone mineral density (BMD) and 25-hydroxyvitamin D in patients with epilepsy. Of the 90 patients, 54 patients had switched to levetiracetam, 19 patients had stopped, and 17 patients continued taking phenytoin. We proposed a 2-year period to examine 25-hydroxyvitamin D, parathyroid hormone, and BMD. The patients who switched or stopped phenytoin showed a significant increase in BMD of the lumbar spine and left femur, and in 25-hydroxyvitamin D. In contrast, those who continued phenytoin had a significant decrease in BMD at both sites and in 25-hydroxyvitamin D. Patients who were taken off phenytoin and those switching to levetiracetam did not show a significant difference in BMD, 25-hydroxyvitamin D, parathyroid, or calcium at follow-up. Compared with those who continued phenytoin, the BMD was significantly higher in patients switching to levetiracetam and those who stopped using phenytoin. Switching medications may be necessary in some cases to avoid low BMD.
Research over the last decades has added to the considerable evidence that older antiepileptic drugs (AEDs) may have diffuse and profound effects on bone mineral density (BMD; Farhat et al., 2002; El-Hajj Fuleihan et al., 2008; Phabphal et al., 2009). Patients prescribed newer, non–enzyme-inducing anticonvulsants are less likely to have a diagnosis of osteoporosis (Lee et al., 2012). Evidence pertaining to the effect of levetiracetam on BMD is limited. This study was conducted with two objectives using a repeat-measure within-patient design: to determine the change in BMD; and to determine the changes in parathyroid hormone (PTH), 25-hydroxyvitamin D, calcium, and phosphate in young adult patients with epilepsy who switched from phenytoin to levetiracetam or discontinued phenytoin.
Subject and Methods
We recruited young adults from Songklanagarind Hospital between April 2004 and January 2012. Inclusion criteria were a Thai national with epilepsy, age 20–40 years, having used phenytoin for at least 1 year and whose physician had decided, for clinical reasons other than bone disorders, to switch them over to monotherapy with levetiracetam, regular menstruation, stable weight (over the previous 6 months), stable dosage of phenytoin (in the previous 12 months), no chronic medical illness other than epilepsy, taking no medication except antiepileptic(s), with an active daily life, and not using alcohol or smoking. Exclusion criteria included pregnancy; having significant disability; previous fracture in the past year; taking medications known to affect bone turnover, for example, glucocorticoids, bisphosphonates, thiazides, GnRH analogues, calcium, hormonal replacement, and steroids. None of the patients received vitamin D, as there is no evidence of low vitamin D being a risk factors for low BMD in patients with epilepsy in the South of Thailand (Phabphal et al., 2009).
Patients who were taking phenytoin and physicians deciding on withdrawal from AED for at least 1 year because of seizure remissions were recruited as one control group, and patients who continued taking phenytoin were recruited as a second control group.
Demographic variables were recorded. Body mass index (BMI) was calculated (see Supporting Information, Data S1). At baseline before switching phenytoin, weight and height were measured, and blood samples for calcium, phosphate, 25-hydroxyvitamin D level, and PTH level determination. The BMD of the lumbar spine and left femur were measured. During follow-up, each subject underwent repeat measurements (see Data S1).
Basic clinical characteristics, chemical data, and bone mineral density parameters were compared between patients who switched to levetiracetam and those stopping AED and those continuing phenytoin. Categorical variables were shown as number and percent and compared across groups using chi-square test, whereas continuous variables were shown as median and interquartile range and compared using the Kruskal-Wallis test followed, if significant, by rank-sum (Mann-Whitney) tests of each pair of treatment groups. Because of the variable interval between first and second measurements, the changes in chemicals and BMD value were standardized by dividing the change by the time interval. Within each group, the rate of change of each parameter was tested for significant difference from zero using the sign test, and the rate of change of each parameter compared across groups using the Kruskal-Wallis test followed, if significant, by rank-sum tests of each pair of treatment groups. To investigate the possible association in the switching group between rate of change in BMD measurements and that of 25-hydroxyvitamin D and of calcium, Spearman correlation coefficients were examined. In all analyses, statistical significance was set at p < 0.05. Statistical analyses were performed in STATA version 10.0 (Stata Statistical Software College Station, TX, U.S.A.).
Basic demographic and clinical characteristics of the patients are shown in Table 1. There were no differences among the three groups in age, sex, duration between first and second BMD measurement, BMI, duration of treatment with phenytoin, and duration of epilepsy. The results of first measurement and second measurement of serum PTH, 25-hydroxyvitamin D, calcium, and serum phosphate showed no statistically significant difference among the three groups (Table 1).
Table 1. Basic clinical characteristics, chemical data, and bone mineral density
|Gender|| || || || |
|Male, n (%)||26 (48.2%)||6 (31.6%)||7 (41.2%)||0.44|
|Female, n (%)||28 (51.8%)||13 (68.4)||10 (58.8%)|
|Age at inclusion (years)|| || || || |
|Median (IQR)||33 (28,39)||32 (27,39)||33 (29,37)||0.89|
|Body mass index|| || || || |
|Duration of second BMD measurement (years)|| || || || |
|Median (IQR)||3 (2,3)||2 (2,3)||2.5 (2,3)||0.12|
|Duration of epilepsy (years)|| || || || |
|Median (IQR)||14 (4,23)||9 (6,18)||10 (9,12)||0.28|
|Duration of phenytoin treatment (years)|| || || || |
|Median (IQR)||2 (1,2)||2 (1,2)||2 (1.5,2.5)||0.70|
|25-hydroxyvitamin D at first measurement (ng/ml)|| || || || |
|Median (IQR)||38.16 (32.11,42.21)||34.72 (29.77,42.81)||34.70 (31.7,44.3)||0.88|
|25-hydroxyvitamin D at second measurement (ng/ml)|| || || || |
|Median (IQR)||38.14 (32.87,42.48)||35.9 (30.59,42.16)||32.89 (28.13,40.21)||0.24|
|Parathyroid hormone at first measurement (pg/ml)|| || || || |
|Median (IQR)||38.06 (23.4,41.19)||40.21 (28.11,57.79)||39.50 (38.02,40.3)||0.25|
|Parathyroid hormone at second measurement (pg/ml)|| || || || |
|Median (IQR)||38.41 (23.12,41.91)||40.22 (28.21,56.71)||40.00 (37.3,47.6)||0.14|
|Phosphate (mg/ml) at first measurement|| || || || |
|Median (IQR)||3.5 (2.9,3.9)||3.7 (3.3,3.9)||3.8 (3.5,3.9)||0.9|
|Phosphate (mg/ml) at second measurement|| || || || |
|Median (IQR)||3.5 (3,3.8)||3.7 (3.3,3.9)||3.6 (3.4,3.8)||0.11|
|Calcium (mg/ml) at first measurement|| || || || |
|Median (IQR)||9 (8.9,9.3)||9.2 (9,9.3)||9.4 (9,9.5)||0.57|
|Calcium (mg/ml) at second measurement|| || || || |
|Median (IQR)||9.3 (9,9.6)||9.2 (8.8,9.5)||9.4 (9,9.5)||0.71|
|Left femur at first measurement|| || || || |
|Z-score Median (IQR)||−0.7 (−1.1,−0.3)||−0.5 (−1.2,0.2)||−0.65 (0.99,−0.01)||0.41|
|T-score Median (IQR)||−0.85 (−1.2,−0.3)||−0.6 (−1.4,0.2)||−0.6 (−0.9,−0.02)||0.27|
|Lumbar spine at first measurement|| || || || |
|Z-score Median (IQR)||−1.2 (−1.8,−0.5)||−0.8 (−1.4,−0.3)||−.98 (−1.1,−0.9)||0.41|
|T-score Median (IQR)||−1.1 (−1.6,−0.5)||−0.4 (−1.7,−0.2)||−0.95 (−1.1,−0.86)||0.14|
|Left femur at second measurement|| || || || |
|Z-score Median (IQR)||0.15 (−0.1,0.7)a||0 (−0.1,1)a||−0.9 (−1.1,−1.2)b||<0.01|
|T-score Median (IQR)||0.5 (0,1.1)a||0.2 (−0.1,1.1)a||−1.0 (−1.3,−0.7)b||<0.01|
|Lumbar spine at second measurement|| || || || |
|Z-score Median (IQR)||−0.1 (−1.1,0.5)a||0 (−0.6,0.4)a||−1.2 (−1.9,−1.1)b||<0.01|
|T-score Median (IQR)||0 (−0.4,1.1)a||0.4 (0,1.1)a||−1.7 (−1.9,−1.3)b||<0.01|
Bone mineral density
Summary results for distribution of BMD, together with the other primary outcome measures are shown in Table 1. At baseline, the median values of BMD at lumbar spine and left femur did not show a statistical difference among the three groups.
The median values of BMD at each site showed no statistically significant difference between patients switching to levetiracetam and patients who stopped phenytoin. Of interest, patients who switched to levetiracetam and patients who stopped phenytoin showed median values of BMD at both sites that were higher than those of patients who continued taking phenytoin.
Rate of change per year of BMD, PTH, 25-hydroxyvitamin D, calcium, and phosphate are shown in Table 2.
Table 2. Rate change per year in each outcome measure between first and second measure
|Left femur|| || || || || |
|Lumbar spine L1–L4|| || || || || |
|25-hydroxyvitamin D (ng/ml)||Switched group||0.07a||0,0.24||<0.01||<0.01|
|Parathyroid hormone (pg/ml)||Switched group||0.03||−0.15,0.21||0.49||0.53|
|Phosphate (mg/ml)||Switched group||−0.03||−0.08,0.04||0.40||0.14|
|Calcium (mg/ml)||Switched group||−0.03||−0.05,0.02||0.05||0.42|
All patients who switched to levetiracetam and those who stopped phenytoin had significant improvement of the BMD at both sites over those who continued on phenytoin, but did not differ significantly themselves. In contrast, patients who continued taking phenytoin had significant worsening of the BMD at both sites. 25-hydroxyvitamin D significantly increased in the switching phenytoin groups. In contrast, 25-hydroxyvitamin D showed a significant decrease in those continuing to take phenytoin. The change in 25-hydroxyvitamin D was not significantly correlated with change in T-score or Z-score at either site in any of the three groups.
The results of this study of patients who switched from phenytoin to levetiracetam revealed a significant improvement of BMD as well as an increase in 25-hydroxyvitamin D. By contrast, patients who continued taking phenytoin had significant deterioration of BMD as well as lowered 25-hydroxyvitamin D.
Previous studies of phenytoin-induced bone loss in patients were conducted mostly in western and middle western countries (Farhat et al., 2002; El-Hajj Fuleihan et al., 2008) and included patients with confounding factors related to BMD (Stephen et al., 1999; Andress et al., 2002). In this prospective study in a tropical Asian country, we followed 90 patients who had been taking phenytoin for at least 1 year and met strict inclusion criteria before enrollment. At the first measurement of BMD, osteopenia or osteoporosis was detected in 31 (57%) of 54 patients in groups that had switched from phenytoin to levetiracetam and 6 (32%) of 19 patients in the group that stopped phenytoin. At the second measurement, patients who switched to levetiracetam had significant increases in both BMD and 25-hydroxyvitamin D, whereas patients who continued taking phenytoin showed significant decreases in BMD and in 25-hydroxyvitamin D. This finding supports the positive effect on BMD of switching from phenytoin to levetiracetam.
An animal model study demonstrated a biphasic dose-dependent effect of levetiracetam on biomechanical bone strength at the femoral neck, rather than on bone mass, which may be related to microstructural changes in the bone matrix. The weakening effect of low-dose levetiracetam on the femoral neck, despite a constant BMD, suggests a primary effect on bone quality (Nissen-Meyer et al., 2007). Recently, two retrospective studies have been conducted to investigate the effects on BMD of newer AEDs (Beniczky et al., 2012; Lee et al., 2012). Beniczky et al. (2012) conducted a retrospective study of 49 patients taking various monotherapy and found that a reduced BMD occurred significantly more often in patients treated with levetiracetam and in those treated with oxcarbazepine. In contrast, a study by Lee et al. (2012), investigating patients prescribed newer AEDs, including 152 patients on levetiracetam compared to patients prescribed traditional AEDs, found that newer AEDs were not associated with lower BMD. However, the study had some limitations such as being retrospective and having numerous confounding factors and including heterogeneous AEDs. To date, there has been only one prospective study, by Koo et al. (2013). They evaluated BMD and biochemical markers of bone and mineral metabolism of 61 new onset patients with epilepsy and suggested that levetiracetam monotherapy may have no harmful effect on bone strength and metabolism for 1 year. To our knowledge, the present study is the first to report the effects of levetiracetam on BMD in humans using follow-up data and a repeated-measure within-patient design. Our study showed that patients who switched to levetiracetam had significant increases in BMD and 25-hydroxyvitamin D. This finding supports the proposition that levetiracetam has no adverse effect on bone health or 25-hydroxyvitamin D.
This investigation has a number of important limitations due to a lack of data on calcium intake. However, dietary intake estimated from a recall technique over a short period of time may not be a good representation of usual or long-term calcium intake. Another limitation of our study is the rather modest control sample size and variable length of follow-up among the three study groups. However, the differences are not statistically significant, and their effect was minimized in the analysis by using the rate of change of each parameter. The strength of this study includes using a large population of young adult patients, who switched from phenytoin to levetiracetam, allowing the dissection of effects of phenytoin. The second strength is the analysis of drug-induced changes in BMD using a repeated-measures within-patient design. This effectively removed potential confounding due to genetic variation among subjects.
In conclusion, our study suggests that either switching from phenytoin to levetiracetam or stopping phenytoin significantly improved BMD and 25-hydroxyvitamin D levels.
This study was support by grant from Faculty of Medicine, Prince of Songkla University. We thank Ms Walailuk Jitpiboon and Anongtip Sae for data analysis. Also we wish to thank David Leather for the editing and suggestions on English writing of the manuscript.
None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.