SEARCH

SEARCH BY CITATION

Keywords:

  • Valproic acid;
  • Contraceptive steroids;
  • Pharmacokinetics;
  • Drug interaction;
  • Epilepsy

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES

Summary: Purpose: To determine potential changes in total and unbound serum valproic acid (VPA) concentrations at steady-state during a cycle of intake of combined hormonal contraceptive (HC) steroids.

Methods: Blood samples were collected from nine women stabilized on VPA monotherapy on two separate randomized occasions: (i) at the end of the 4- to 7-day HC-free interval, and (ii) on the last day of the HC intake period. Trough concentrations of VPA in serum and serum ultrafiltrates were determined by fluorescence polarization immunoassay.

Results: In all women, total and unbound VPA concentrations were higher during the HC-free interval than during HC intake (means ± SD: 425 ± 184 vs. 350 ± 145 μmol/L, respectively, for total VPA, p = 0.002, and 55 ± 37 vs. 39 ± 25 μmol/L, respectively, for unbound VPA, p = 0.005). Compared with the HC-free interval, HC intake was associated with a mean 21.5% increase in VPA total apparent oral clearance (from 8.0 ± 5.2 to 9.7 ± 6.4 ml/h/kg, p = 0.01) and a 45.2 % increase in VPA unbound apparent oral clearance (from 79 ± 81 to 115 ± 121 ml/h/kg, p = 0.029).

Conclusions: The apparent oral clearance of total and unbound VPA increases during the HC intake period compared with the HC-free interval, probably due to induction of glucuronosyltransferase by ethinylestradiol. The magnitude of the change varies across individuals, being potentially clinically relevant in some cases. Serum VPA concentrations should be monitored when adding or discontinuing HC steroids, and possibly during the on–off intervals of a HC cycle.

Despite the fact that combined hormonal contraceptive (HC) steroids are frequently used in women with epilepsy, interactions between these agents and antiepileptic drugs (AEDs) have not been thoroughly investigated (1). Specifically, while many studies assessed the effects of AEDs on the disposition of the contraceptive pill (2), little consideration has been given to the possibility of HCs interfering with the pharmacokinetics of AEDs. In fact, HCs can have important effects on drug metabolism, which vary depending on the enzyme system(s) involved. While HCs may inhibit cytochrome P450 enzymes and increase by this mechanism the plasma concentration and the effect of many tricyclic antidepressants, hydroxylated benzodiazepines, some β-blockers, methylxanthines, prednisolone, and cyclosporin A (3), they can also induce the activity of some glucuronosyltransferases, most notably UGT1A4, and enhance the metabolism of drugs which are cleared by glucuronide conjugation, including lamotrigine (4,5), certain benzodiazepines, clofibric acid, paracetamol, and possibly, morphine (3). The interaction with lamotrigine is especially prominent because it leads to >50% reduction in serum lamotrigine concentration (4,5), and alterations in response (4).

Recent studies have shown that the interaction of combined HCs with lamotrigine is dependent on the phase of the contraceptive cycle, with a decrease in serum lamotrigine concentration during the period of HC intake and a rebound increase (up to 100%, or even more) during the HC-free interval (6,7). Whether a similar interaction occurs with other AEDs cleared by glucuronosyltransferases is unclear. Valproic acid (VPA) is one of such AEDs, and a case report suggested that its serum concentration may fluctuate widely during a HC cycle (8). These observations led us to assess prospectively serum VPA concentrations during HC intake and in the HC-free interval in women with epilepsy.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES

Patients and procedures

Women were enrolled according to the following main criteria: (i) age 18–50 years; (ii) treatment with VPA at a stable dosage for ≥3 months; (iii) intake for ≥3 months of a combined HC; (iv) steady-state conditions for any associated comedication; (v) absence of metabolic, hepatic, renal, or progressive illness; (vi) no history of poor compliance; and (vii) written informed consent.

According to the protocol, which received Ethics Committee approval, blood samples had to be collected during an HC cycle on two separate randomized occasions: (i) on the last day of the HC-free interval (reference was made to a 7-day oral HC-free interval, but women using 24-day oral HC cycles separated by four HC-free days, and women using 21-day transdermal HC patches were not excluded), and (ii) on the last day (±1 day) of the HC intake period. All samples were collected in the morning, approximately 12 h after the last dose of VPA. No changes in dosages of medications occurred between sampling sessions. The serum was separated within 4 h and frozen (−20°C) until assay.

Drugs assays, pharmacokinetic, and statistical analysis

The concentration of VPA in serum (total concentration) and serum ultrafiltrates (unbound concentration) was quantified by fluorescence polarization immunoassay (TDx, Abbott, Rome). Ultrafiltrates were prepared by centrifuging 800 μl serum at 1,500g for 20 min at 25°C through the Centrifree Micropartition System (Millipore, Milan). Limits of quantitation were 5.0 and 0.7 μmol/L for total and unbound VPA respectively, and precision was better than 5%.

Apparent oral clearance (CL/F, ml/h/kg) was calculated as Molar daily dosage (μmol/kg)/[Ct (μmol/ml) × 24 h], and unbound oral clearance (CLu/F, ml/h/kg) as Molar daily dosage (μmol/kg)/[Cu (μmol/ml) × 24 h], where Ct and Cu represent the total and unbound concentration respectively at steady-state, and F is the oral bioavailability. Because trough rather than mean concentrations were used, reported CL/F and CLu/F represent overestimates of actual values.

Statistical comparisons between the two sampling sessions were made by Student's t-test for paired data (two tailed). Reported values are means ± SD.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES

Characteristics of patients

Ten subjects were enrolled, and nine completed the study without major protocol deviations. A 24-year-old woman assessed on day 21 failed to return for the second sample 1 week later, and reported back only after 6 months. This long interval was considered a major violation and the subject was excluded from analysis.

Details of the women included in the pharmacokinetic assessment are given in Table 1. All used a HC containing ethinylestradiol (15–40 μg) as estrogen component, with HC-free intervals of 7 days in seven cases and 4 days in two. Transdermal patches were used in two women. All women had been seizure-free for >12 months on monotherapy with VPA, given as a sustained-release formulation (except patient 8) at dosages of 500–1,500 mg/day, and took no additional medications apart from HCs.

Table 1. Details of the patients included in the assessment
SubjectAge (yr)Weight (kg)Weight Epilepsy syndromeVPA daily dosing schedule (mg)HC (μg/day)Duration of HC-free interval (days)
  1. GTCS, generalized tonic-clonic seizures.

  2. *Transdermal formulation.

14454Juvenile absence epilepsy with GTCS150 + 600Gestodene 604
 Ethinylestradiol 15 
22250Juvenile myoclonic epilepsy1,000Gestodene 757
 Ethinylestradiol 20 
33453Childhood absence epilepsy with GTCS750 + 750Norgestromine 150*7
 Ethinylestradiol 20* 
41864Childhood absence epilepsy with GTCS500 + 500Gestodene 757
 Ethinylestradiol 20 
52380Juvenile myoclonic epilepsy750 + 500Drospirenone 3,0007
 Ethinylestradiol 30 
63160Idiopathic generalized epilepsy with GTCS on awakening750 + 750Gestodene 50,70,1007
 Ethinylestradiol 30,40 
74564Idiopathic generalized epilepsy with GTCS500Gestodene 604
 Ethinylestradiol 15 
82065Idiopathic generalized epilepsy with GTCS on awakening500Drospirenone 3,0007
 Ethinylestradiol 30 
92760Idiopathic generalized epilepsy with GTCS on awakening500 + 800Norgestromine 150*7
 Ethinylestradiol 20* 

Pharmacokinetic data

In all women included in the pharmacokinetic assessment, samples were obtained on the last day of the HC-free interval (except for subjects 5 and 9 in whom samples were collected on day 6 of the HC-free interval) and on the last day of HC-intake. In every patient, total and unbound serum VPA concentrations were higher in the HC-free interval than during HC intake (Fig. 1). Mean total serum VPA levels on the two occasions were 425 ± 184 μmol/L and 350 ± 145μmol/L, respectively (p = 0.002). Corresponding values for unbound levels were 55 ± 37 μmol/L and 39 ± 25 μmol/L, respectively (p = 0.005).

image

Figure 1. Individual changes in total (left panel) and unbound (right panel) serum VPA concentrations in nine women assessed at the end of the hormonal contraceptive (HC)-free interval and on the last day of HC intake.

Download figure to PowerPoint

Individual VPA CL/F and CLu/F values are illustrated in Fig. 2. On an average, CL/F was 21.5% higher during HC intake than in the HC-free interval (9.7 ± 6.4 vs. 8.0 ± 5.2 ml/h/kg, p = 0.01). The increase in CLu/F between the two sessions was 45.2%, from 79 ± 81 to 115 ± 121 ml/h/kg (p = 0.029).

image

Figure 2. Individual changes in total (left panel) and unbound (right panel) VPA CL/F in nine women assessed at the end of the hormonal contraceptive (HC)-free interval and on the last day of HC intake.

Download figure to PowerPoint

In the subject excluded for protocol violation, total and unbound serum VPA concentrations were 247 and 22 μmol/L, respectively, during HC intake and 198 and 15 μmol/L, respectively, in a HC-free interval 6 months later. Changes in total and unbound VPA levels between the two sessions remained statistically significant (p < 0.02) when the protocol violator was included in the analysis (data not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES

This study shows that, in agreement with observations previously made in a single case (8), serum VPA levels at steady-state vary in relation to the HC intake cycle. Mean total VPA CL/F was 21.5% higher on the last day of HC intake compared with the end of the HC-free interval. The increase in clearance of unbound, pharmacologically active drug (CLu/F) between the two sessions was even greater (45.2%), which may be explained by the known concentration-dependency of VPA binding to serum proteins (9).

These changes are likely to reflect a modification in metabolic drug clearance. In fact, combined HCs are known to induce the microsomal enzymes involved in glucuronide conjugation (3), which contribute to an important extent to VPA metabolism (10). The reduction in VPA CL/F and CLu/F values during the HC-free interval compared with the period of HC intake can thus be ascribed to deinduction of glucuronosyltransferases after interruption of HC exposure.

Similar changes in serum drug concentration during a HC monthly cycle has been reported recently for lamotrigine. In a first study in 16 healthy women assessed during lamotrigine treatment at steady-state, trough lamotrigine concentrations were found to be 27%, 63%, and 116% higher on the 3rd, 5th, and 7th day, respectively, of the first week after withdrawal of an oral contraceptive, compared with those found during HC intake (6). In a second study that used a protocol similar to ours, lamotrigine concentrations in eight women with epilepsy were 27% higher (range 0–100%) on day 7 of the HC-free interval compared with days 18–21 of HC intake (7). At least for lamotrigine, the HC component responsible for the interaction appears to be ethinylestradiol (11).

As described with lamotrigine, the degree of interaction in our study varied markedly across subjects. The increase in serum VPA concentration during the HC-free interval ranged from 2% to 38% for total drug and from 15% to 69% for unbound drug, while in a previous case report (8) the increase in total serum VPA levels in a single woman assessed in two different cycles was 159% and 57%, respectively. Such variability may reflect differences in HC doses, formulations and duration of the HC-free period, differences in functional polymorphisms of glucuronosyltransferases (12) and associated susceptibility to enzyme induction, variation in enzyme deinduction rate during the HC-free interval, differences in VPA dosage and VPA fraction cleared by conjugation (10), and possibly, HC influences on other pathways involved in VPA metabolism (3). At least in some subjects, HC-induced changes in VPA concentration may be clinically relevant, as in the woman reported by Herzog et al. (8) who had far more frequent seizures during HC intake than during HC-free periods.

Our study was limited to a comparison between the last day of HC intake and, in most cases, day 6 or 7 after HC interruption, and therefore did not assess the time course of interaction throughout the cycle. It is probable that a 4–7-day period is insufficient for complete deinduction, and therefore the changes in VPA concentration observed in our study are likely to underestimate the magnitude of the increase in VPA clearance caused by HCs. This will have to be investigated in women starting HCs de novo (or discontinuing HCs permanently), or by between-group comparisons of women on and off HCs. It also remains to be determined whether a similar degree of interaction occurs in women receiving combination therapy with enzyme inducing AEDs, which have an independent stimulating effect on VPA metabolism (13). In any case, these observations suggest that serum VPA concentrations and clinical response should be monitored closely when adding or discontinuing HC steroids, and possibly during the on–off intervals of a HC cycle. Further studies are also indicated to determine whether a similar interaction occurs with other antiepileptic agents, which are extensively cleared by glucuronide conjugation, the most notable example of which is the active monohydroxy-derivative of oxcarbazepine.

Acknowledgments

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES

Acknowledgment:  Mr. Roberto Marchiselli and Mrs. Cristina Valisi are acknowledged for technical assistance. This study was supported by grants from the Italian Ministry for University and Scientific Research (FAR, University of Pavia) and the Italian Ministry of Health (RC 2005 I.R.C.C.S. C. Mondino Foundation).

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgments
  7. REFERENCES