- Top of page
Summary: Purpose: To determine whether long-term treatment with valproate (VPA) and/or lamotrigine (LTG) in children with epilepsy is associated with altered growth and/or bone metabolism.
Methods: Twenty-seven boys and 26 girls, aged 3 to 17 years (9.2 ± 3.9, mean ± SD), with epilepsy treated with VPA and/or LTG for ≥2 years were evaluated for growth, nutrient intakes, physical activity, bone mineral density (BMD), and blood biochemical indices of mineral and bone metabolism.
Results: Twenty-three (43.4%) of the children had a body height below the 10th percentile. Z-scores for BMD below –1.5 occurred in 24.4% of the children. When patients were divided into two groups according to daily activity score, a significantly lower Z-score for total body BMD (p = 0.007), percentile for body height (p = 0.05), and plasma parathyroid hormone (PTH; p = 0.04), osteocalcin (p = 0.04) and 25-hydroxyvitamin D (25OHD) (p = 0.01) were found in the inactive compared with the active group. Z-score for total body BMD was correlated with daily activity score (r = 0.43, p = 0.008). Plasma intact osteocalcin and intact PTH values correlated significantly (r = 0.36, p = 0.02). Plasma 1,25-dihydroxyvitamin D was within normal range for all subjects. When patients were divided into LTG-alone, VPA-alone, and LTG-plus-VPA treatment groups, significantly lower (p < 0.05) plasma osteocalcin and percentile for body height were found in the VPA-plus-LTG treatment group.
Conclusions: Long-term VPA and LTG therapy, particularly when combined, is associated with short stature, low BMD, and reduced bone formation. These alterations may be mediated primarily through reduced physical activity rather than through a direct link to the VPA and/or LTG therapy.
Antiepileptic drugs (AEDs) such as phenytoin (PHT), carbamazepine (CBZ), and phenobarbital (PB) are associated with bone abnormalities when used in children (1–7). Children treated with these AEDs compared with healthy children may have reduced circulating calcium (Ca) (2–5), 25-hydroxyvitamin D (25OHD) (2,4), and low bone mineral density (BMD) (1,7). Current evidence suggests that there are drug-induced increases in hepatic microsomal catabolism of vitamin D and its biologically active metabolites (8–10). However, previous studies reported that treatment with vitamin D increased the low serum 25OHD concentration but failed to correct the hypocalcemia (11–13). These findings suggest that other factors, rather than vitamin D alone, may play an important role in the abnormal bone metabolism induced by AEDs. There is evidence that AEDs might directly affect bone cells and inhibit intestinal transport of Ca (14).
Alternative AEDs such as valproate (VPA) and lamotrigine (LTG) have been widely used as effective seizure control agents in children. In previous descriptive studies that measured bone mass, reduced BMD was reported in children with epilepsy treated with VPA in some (15,16,17) but not all (18) studies. It is generally agreed that VPA is not an inducer of liver enzymes, nor does it reduce circulating calcium, 25OHD, or alkaline phosphatase concentration (19,20). A study in pubertal girls with epilepsy showed that VPA alone did not affect statural growth (21). Furthermore, it is unclear whether the reduced BMD is mediated by VPA itself or whether the reduced BMD is associated with other conditions such as insufficient physical activity or inadequate calcium and vitamin D intake. The purpose of this study was to determine whether the long-term treatment of epilepsy with VPA and/or LTG is associated with altered growth and bone metabolism in children and adolescents, and if so, to elucidate a possible mechanism by which interference with these growth processes occurs.
- Top of page
The clinical characteristics of the subjects are summarized in Table 1. Gender was balanced within the treatment and activity groups. The mean ages and gender did not differ between the three treatment groups. A mean daily vitamin D intake in excess of the recommended adequate intake (AI) (31) occurred in 51.9% of the children (Table 2). Mean calcium intake above the recommended AI occurred in 46.4% in the children aged 3–8 years, whereas this occurred in only 8.3% of the children aged 9–18 years (Table 2). Vitamin D and Ca supplements were taken by 30.8% and 25.0% of patients in the respective age groups.
Table 1. Clinical characteristics
|Treatment||Gender||Age (y)||Tanner stage|
Table 2. Daily calcium and vitamin D intake
|Nutrient||Age (y)||No.||Daily intake||AI||% Subject >AI|
|Calcium (mg/d)||3–8||29||803||539||800||46.4 (13/28)|
| ||9–18||24||696||416||1,300||8.3 (2/24)|
|Vitamin D (μg/d)||3–18||53||7.9||5.7||5||51.9 (27/52)|
Twenty-three (43.3%) patients had a body height less than the 10th percentile (Fig. 1). A Z-score for BMD below –1.5 was found in nine (24.3%) patients (Fig. 1). Forty-three (81.1%) patients had plasma intact osteocalcin less than the 25th percentile, and none of 53 patients had plasma intact osteocalcin higher than the 75th percentile (Fig. 2). Compared with reference values, plasma intact PTH was in the lower limit, whereas observed values for plasma 25OHD and 1,25(OH)2D were not below reference ranges (Fig. 3). An activity score of –25% was set as a “cut-off” point to divide patients into active (n = 24, 13 boys and 11 girls) and inactive (n = 29, 14 boys and 15 girls) groups. The cut-off point is based on the lower limit of activity score observed in a small group of healthy children with ages similar to those of the study subjects. The median activity score was –12.5% (range, +20% to –25%) and –58.3% (–33% to –100%) in the active and inactive groups, respectively. Mean age was not different between the active (8.0 ± 4.1 years, mean ± SD) and inactive (8.8 ± 3.8) groups. Patients in the inactive group had a lower body height percentile (p = 0.05; Fig. 1), lower Z-score for total body BMD (p = 0.007; Fig. 1), and lower plasma intact PTH (p = 0.03; Fig. 3) and osteocalcin (p = 0.03). A Z-score for BMD below –1.5 occurred in seven (36.8%) of 19 in the inactive group but only in two (11.1%) of 18 in the active group (Fig. 1). Plasma CrossLaps was similar in the active and inactive groups.
Figure 1. Comparison in body height (top), weight (middle), and bone mineral density (BMD; bottom) between children with epilepsy in inactive (low activity score) and active (high activity score) groups. The horizontal lines represent medians for body height and weight percentile, and represent medians for BMD Z-scores.
Download figure to PowerPoint
Figure 2. Plasma intact osteocalcin in boys (top) and girls (middle), and the correlation between activity score and Z-score for bone mineral density (BMD) (r = 0.43, p = 0.008) (bottom). The solid lines represent 50th percentile values, and the areas between dotted lines represent 25–75th percentile in plasma intact osteocalcin (27).
Download figure to PowerPoint
Figure 3. Comparison in plasma intact parathyroid hormone (PTH) (top), 1,25(OH)2D (middle), and 25OHD (bottom) between children with epilepsy in inactive (low activity score) and active (high activity score) groups. The horizontal lines represent means. The boxes represent reference ranges (29,30).
Download figure to PowerPoint
When patients were divided into treatment groups of VPA alone, LTG alone, or a combined therapy of VPA and LTG, lower body height percentile and plasma intact osteocalcin (p < 0.05; Fig. 4), but not plasma CrossLaps and activity score, were found in patients treated with the combined therapy of VPA and LTG. The mean physical activity score was 25% and 32% lower in the combined-therapy group compared with the LTG-alone and VPA-alone groups, respectively, but statistical analysis did not show a significant difference between the three treatment groups (Fig. 4). Only six patients completed a DXA scan in the combined-therapy group. The limited number did not allow us to compare the difference in BMD between the three treatment groups. Z-score for total body BMD was significantly correlated with daily activity score (r = 0.43; p = 0.008; Fig. 2). Plasma intact osteocalcin and CrossLaps were correlated with plasma intact PTH (r = 0.36, p = 0.02, and r = 0.29, p = 0.05, respectively).
Figure 4. Comparison in body height (upper left), activity score (upper right), plasma osteocalcin (lower left), and CrossLaps (lower right) between children with epilepsy treated with valproate (VPA) alone, lamotrigine (LTG) alone, or a combined therapy of VPA and LTG. The bars represent means, and the whiskers represent standard errors. Significant differences (p < 0.05) were found between groups with different letters.
Download figure to PowerPoint
- Top of page
In the children with epilepsy studied, treatment with VPA and/or LTG therapy for >2 years was associated with short stature, low bone mass, and reduced bone formation. These outcome measures were particularly compromised in those children whose physical activity was low and who took a combined therapy of VPA and LTG.
The observation of low BMD in children treated with VPA is consistent with some previous studies (15,16,17) but not others (18). Physical activity was not analyzed in the previous studies. It is well known that physical activity influences bone-mass acquisition in growing children (32,33), whereas high turnover offers a greater potential for skeletal hypertrophy. Physical exercise also was associated with an increase in circulating intact PTH and osteocalcin (32–36). We propose that the lower physical activity in children with epilepsy in the present study led to an uncoupled bone remodeling, as indicated by the reduced plasma intact osteocalcin in the presence of unchanged plasma CrossLaps. The reduced Z-score for BMD in the inactive patients likely results from the reduced bone formation, which in turn is a function of the reduced physical activity.
In previous studies of children and adolescents receiving AEDs, limitation of mobility was significantly associated with low bone mass and delayed bone development (2,37,38). However, in children receiving VPA alone, low bone mass was observed in the absence of restricted mobility. Both Sheth et al. (18) and Kafali et al. (17) observed a low bone mass in the lumbar spine and middle or distal radius in children without physical handicaps who were treated for ≥6 (17) or 18 (18) months. In contrast, a normal bone mass of the lumbar spine was observed in a similar population with normal physical activity who were treated for ≥12 months (18). Although the reason for the inconsistent results observed among the previous studies and our present study is not clear, the following hypotheses may provide a partial explanation. First, whole-body bone mass was measured in our study, whereas regional skeletal mass was measured in other studies. Measurements of the regional skeleton such as forearm, lumbar spine, or proximal femur are often done clinically because of their importance for predicting fracture risk in the elderly. However, measurement of the whole-body bone mass is critical in evaluating bone acquisition in children. Measurement of the whole-body bone mass is particularly important if low physical activity is suspected, because the regional skeletal site for weight bearing is different in different life styles. Second, to date, all studies in children and adolescents receiving VPA are cross-sectional in design. To monitor bone growth and development, well-controlled longitudinal designs are needed in future studies. Third, the inconsistency on the findings suggests that there may be other factors that have not yet been identified, such as growth failure, that contribute to the development of low bone mass in children and adolescents treated with VPA.
Although the circulating vitamin D metabolites were normal in the present study, which is consistent with previous reports (19,20), we cannot entirely rule out the possibility that high vitamin D intake in this study population masked an altered vitamin D homeostasis. This hypothesis is based on the following facts: (a) inactive patients had a lower plasma 25OHD than did active patients, which is possibly mediated by the insufficient exposure to sunshine; and (b), plasma intact PTH was lower in the inactive patients. Vitamin D and Ca supplements were taken by many of the participating children with epilepsy, particularly by young children. It is likely that parents are aware of the side effects of traditional AEDs on vitamin D and Ca metabolism. We believe that for children who do not consume adequate dietary sources of calcium and vitamin D (or who are not habitually exposed to sunlight), supplementation is appropriate. There is no evidence that children requiring therapy with VPA and LTG require intakes of these nutrients in excess of the Dietary Reference Intake recommendations (31).
Statural growth was not affected by a single drug therapy such as VPA, as also observed by Rattya et al. (21). However, the combined therapy of VPA and LTG was associated with lower height percentiles. Patients treated with more than one drug usually have more severe seizures and, in turn, may have lower physical activity. Because reduced body height is associated with a significantly reduced circulating intact osteocalcin and a 25–32% reduction in a physical activity score in the combined-therapy group, we speculate that the alteration in growth is primarily mediated through reduced physical activity. The lack of statistical significance in physical activity between the three treatment groups in our study may simply reflect the small number of study subjects.
In conclusion, long-term VPA and LTG therapy, particularly when combined, is associated with short stature, reduced bone formation, and low bone mass, especially in those children whose physical activity is low. Future studies in children with epilepsy should focus on reduced physical activity, rather than a direct link to the VPA and/or LTG, as a component for intervention to prevent clinically significant reductions in growth and bone-mass accretion. A limitation of the present study is that we could not establish a geographically similar age- and sex-matched healthy control group for comparison with our population with epilepsy. This, however, does not diminish the clinical relevance of the observed differences in outcomes of growth and bone between the children treated with single and combined therapy, and in relation to activity level.