Article Text
Abstract
Objective: To characterize trends in prescribing carbamazepine (CBZ), sodium valproate (VPA) and lamotrigine (LTG) in adolescent females in the UK and to examine possible reasons for changing trends.
Design: Population-based observational study.
Setting: UK General Practice Research Database between 1 January 1993 and 31 December 2006.
Patients: 12–18-year-old subjects who were issued ⩾1 CBZ, VPA or LTG prescription.
Main outcome measures: Prescribing prevalences stratified by age, gender and antiepileptic drug.
Results: 5417 patients (47.6% females) were prescribed 147 111 prescriptions for CBZ (34.5%), VPA (38.6%) or LTG (26.9%). The prevalence of LTG prescribing in females increased from 0.08 (95% CI 0.04 to 0.12) to 0.80 (95% CI 0.70 to 0.89) per 1000 female population. Conversely, the prevalence in females of CBZ and VPA prescribing significantly decreased from 1.00 (95% CI 0.85 to 1.15) to 0.51 (95% CI 0.44 to 0.58) and from 0.94 (95% CI 0.80 to 1.09) to 0.63 (95% CI 0.55 to 0.72), respectively. This 10-fold rise in LTG prescribing in females is much higher than the fivefold rise in males from 0.09 (95% CI 0.05 to 0.14) to 0.47 (95% CI 0.40 to 0.54) per 1000 male population.
Conclusion: The practice of prescribing antiepileptic drugs in adolescents has changed gradually over the last decade. More females aged 12–18 years are prescribed LTG than CBZ or VPA and the increase is much greater than for males. The increase in LTG prescribing mirrors a corresponding decrease in both VPA and CBZ. Concerns about potential problems to offspring appear to be affecting prescription trends in adolescent females of child-bearing potential.
Statistics from Altmetric.com
The choice of antiepileptic drug (AED) for girls who are likely to continue treatment through their child-bearing years requires careful consideration. Clinicians need to consider the possible adverse effects on the patient, such as increased appetite and weight gain caused by sodium valproate (VPA), as well as the possible adverse effects on the unborn child. AEDs are known to be associated with congenital malformations. In particular, VPA is associated with an increased risk.1 2 Lamotrigine (LTG), a more recently marketed AED, was believed to be a safer choice than VPA in females of child-bearing age, although recent data has questioned this.3
What this study adds
The practice of prescribing antiepileptic drugs in adolescents has changed gradually over the last decade.
Lamotrigine use has increased and sodium valproate and carbamazepine use has decreased specifically in females.
We believe concerns about potential problems to the offspring are affecting prescription trends in adolescent females of child-bearing potential.
A study conducted by the current authors showed a large increase in LTG prescribing and a decrease in the prescribing of conventional AEDs in children and adolescents over the last 13 years in the UK.4 Rates of adolescent pregnancies are high: 7% of babies born in England are to an adolescent mother.5 These pregnancies are often unplanned, implying that the risk of major congenital malformations (MCMs) needs to be considered when choosing AEDs.
The aim of this study was to characterise the prescribing trends, in UK general practice, of the most frequently prescribed AEDs, that is carbamazepine (CBZ), VPA and LTG, in adolescent females of child-bearing potential and to estimate the change in risk of MCMs.
METHODS
The General Practice Research Database (GPRD)6 was used to identify subjects (aged 0–18 years) who were prescribed CBZ, VPA or LTG during the study period 1 January 1993 to 31 December 2006, by a previously described method.4 In summary, the GPRD is a longitudinal clinical database containing anonymised primary care clinical records for approximately 5% of the UK population.6 Validation studies show the quality and completeness of the data are high6 and studies investigating paediatric drug prescribing using the GPRD are becoming well established.7–11 Subjects with less than 12 months’ data or a temporary registration with their general practitioner were excluded. All subjects in the study cohort were screened for diagnoses of epilepsy. The number of subjects prescribed more than one AED, including non-study drugs, was determined. Co-prescribing of oral contraceptive medication and pregnancy records were identified for female subjects treated with AED monotherapy.
Ethics approval
Ethics approval was granted by the GPRD Scientific and Ethical Advisory Group.
Statistical methods
Prevalence was defined as the number of subjects with one or more of the study AED prescriptions per 1000 in the GPRD population (0–18-year olds) in a particular year. The adolescent age category (12–18 years) was defined according to the International Conference on Harmonization Guideline for Clinical Investigation of Medicinal Products in the Paediatric Population. Gender-specific annual prevalences for 12–18-year-old males and females and for 0–11-year-old females of each AED prescribed were calculated. These are presented with confidence intervals in a series of plots. Generalised binomial models were used to quantify the changes in prevalence by age, year and gender for each of the three AEDs. The negative binomial model is a maximum-likelihood regression model of a dependent variable (number of subjects) on independent variables (year, gender and age). The model generates the count variable (number of subjects) by a Poisson-like process but takes into account the greater variation or “over-dispersion” which was present when standard Poisson regression models were used. Age was treated as a continuum within these models. Interactions between pairs of significant variables were investigated to identify differing prescribing patterns over time according to age and gender. Results are presented with 95% confidence intervals (95% CI).
The numbers of adolescent (12–18-year-old) females with epilepsy were combined with established MCM rates3 to determine how prescribing changes over time may have altered the rates of MCM attributable to AED usage. The estimated MCM rate per 100 pregnancies in females taking any of the three study drugs in monotherapy was calculated using the following equation:
(MCM rate for CBZ)×(proportion of females on CBZ monotherapy)+(MCM rate for LTG)×(proportion of females on LTG monotherapy)+(MCM rate for VPA)×(proportion of females on VPA monotherapy).
Statistical analyses were performed using STATA/SE v 9.2 (StataCorp, College Station, TX).
RESULTS
A total of 5417 subjects aged 12–18 years (47.6% females) were prescribed CBZ, VPA or LTG. Of these subjects, 32.3% were being treated with more than one AED (including non-study AEDs). Overall, 87% of subjects had a diagnosis of epilepsy associated with their AED prescriptions. Table 1 shows the number of prescriptions, characteristics of the study cohort (subjects aged 12–18 years), oral contraceptive co-prescribing and pregnancy records by each AED. In the comparison group (females aged 0–11 years), 135 737 prescriptions were issued to 4870 subjects.
Figure 1 shows that the prevalence of LTG prescribing in 12–18-year-old females increased 10-fold between 1993 and 2006, from 0.08 (95% CI 0.04 to 0.12) to 0.80 (95% CI 0.70 to 0.89) per 1000 female population. Conversely, the prevalence in females of CBZ and VPA prescribing significantly decreased from 1.00 (95% CI 0.85 to 1.15) to 0.51 (95% CI 0.44 to 0.58) and from 0.94 (95% CI 0.80 to 1.09) to 0.63 (95% CI 0.55 to 0.72), respectively.
The trends in prescribing prevalence for these drugs in males of the same age are different from those in females (fig 2). For example, the prevalence of LTG prescribing in males only increased fivefold over the study period from 0.09 (95% CI 0.05 to 0.14) to 0.47 (95% CI 0.40 to 0.54) per 1000 male population. The significant increase in the prevalence of VPA prescribing in males from 0.86 (95% CI 0.72 to 0.99) to 1.10 (95% CI 1.00 to 1.27) mirrored the decrease in females. The prevalence of CBZ prescribing fluctuated and decreased slightly over the study period from 0.80 (95% CI 0.67 to 0.93) to 0.70 (95% CI 0.61 to 0.78).
Figure 3 illustrates that in females aged 0–11 years the prevalence of LTG prescribing increased at a lower rate (fivefold) than in females aged 12–18 years (10-fold). CBZ and VPA prescribing prevalence has decreased over time but not to the same degree as in the 12–18-year-old age group.
The generalised binomial models allowed quantification of the annual changes according to gender and age. All three AEDs showed significant differences in trends over time between males and females. Increases in the prescribing of LTG over time were significantly less in males relative to females, with males falling behind by an average of 2.3% (95% CI 0.7% to 3.9%; p = 0.005) per year between 1993 and 2006. In contrast, overall prescribing of CBZ and VPA decreased, with average declines of 2.7% (95% CI 1.1% to 4.3%; p = 0.001) and 1.4% (95% CI 0.2% to 2.6%; p = 0.019) per annum, respectively. The analysis also showed that reductions in prevalence of CBZ and VPA prescribing were significantly less in males than in females (p<0.001 for both AEDs). Hence for all three AEDs the changes over time were less marked for males.
Table 2 shows that the overall estimated risk of MCMs for adolescent females treated with any of the three study drugs in monotherapy decreased by 6% from 4.19 to 3.95 per 100 pregnancies.
DISCUSSION
The three study drugs were chosen for evaluation because of their frequent use as monotherapy exposure during pregnancy, as identified by the UK Epilepsy and Pregnancy Register.3 This study provides detailed analyses of prescribing prevalences in adolescent females, which allows modelling of the potential risk of MCMs for subjects exposed to CBZ, VPA or LTG. A previous study by the authors was a general descriptive analysis of overall AED prescribing in children and adolescents and is a basis for this current study.4
This study shows that the practice of prescribing AEDs in adolescents has changed gradually over the last decade with the introduction of newer drugs to the market. LTG was first licensed in the UK in 1991 and as monotherapy in 1995. The prescribing of LTG has increased 10-fold in adolescent females, with a much smaller (fivefold) rise in males. This indicates that clinicians are preferring to prescribe LTG in females, replacing the more conventional AEDs. In particular, there has been a corresponding marked fall in the prescribing of VPA. Is there substantial research and evidence to support these changes in practice? LTG was marketed as a drug of choice for females of child-bearing age and the increase in prevalence in females aged 12–18 years reflects this. The decline in CBZ and VPA prescribing in these females further suggests that LTG is preferred. However, since LTG was first marketed, studies have highlighted some disadvantages of LTG use in females of child-bearing potential.
Contraception
It is important to consider the choice of contraception in subjects prescribed AEDs because of the potential for interactions between AEDs and the oral contraceptive. The co-prescribing of an oral contraceptive in female subjects prescribed LTG monotherapy was higher (16.4%) than in subjects prescribed CBZ (12.1%) and VPA (14.6%). However, this is likely to be an underestimate as teenagers may obtain oral contraceptives from family planning clinics and these data are not recorded in the GPRD. Subjects could also be using additional contraceptive methods, such as a barrier method. Because of these factors it was not feasible to determine the overall contraceptive rate for subjects prescribed AEDs. Interactions between oral contraceptives and AEDs can occur. CBZ interaction with oral contraceptives is well documented and advice in prescribing guidelines suggests that alternative contraceptive methods should be used.12 A recent study in 20 female healthy volunteers demonstrated that LTG, at a dosage of 300 mg/day, decreases by about 20% the area under the curve of levonorgestrel,13 although the study concluded that none of the subjects showed hormonal evidence of ovulation and consequently no changes in contraception were recommended. Furthermore, oral contraceptives can decrease LTG blood levels and a dosage adjustment of LTG may be necessary, especially following withdrawal of the contraceptive when the female is planning pregnancy, to prevent the fetus from being exposed to unnecessarily high levels of LTG.13–15
The pregnancy rate in adolescent females treated with AED monotherapy was found to be 2% and was similar for each of the three AEDs. This is likely to be an underestimate, as pregnancy tests could be conducted in the home and terminations can be performed at family planning clinics. Additionally, it is not known how many of the reported pregnancies resulted in live births. The hormonal changes during pregnancy are associated with a decrease in levels of LTG which could lead to a loss of seizure control. If the LTG dose is increased to compensate for this, toxicity might occur post-partum, when levels rise.16
Major congenital malformations
Adverse effects of AEDs include intrauterine growth retardation, dysmorphisms, major malformations and postnatal developmental delay.17 The nature of major malformations depends on the AED.18 Neural tube defects have been reported with the use of VPA and less so with CBZ.2 More recently, LTG has been linked with an increased risk of oral clefts.19 However, unlike some other medications, AEDs cannot usually be discontinued in patients with active epilepsy without putting the patient at unacceptable risk. Apart from the risk to the patient, the outcome of the pregnancy may also be affected by seizures.20 21 These factors imply that the choice of AEDs should be considered very carefully. Data from the UK Epilepsy and Pregnancy Registry indicated that VPA monotherapy was associated with the highest frequency of MCMs (6.2%) compared to CBZ (2.2%) and LTG (3.2%).3 The most recent data from the North American Antiepileptic Drug Pregnancy Registry further highlight that VPA monotherapy is associated with a higher risk of major malformations (10.7%)22 compared to CBZ (2.5%)23 and LTG (2.7%).19 However, data from these different registries cannot be directly compared due to different methods of data collection: for example, the North American Registry relies on women to enrol voluntarily whereas the UK Registry is based on reporting from healthcare professionals.
There appears to be an increase in the rate of major malformations with dose for at least some of the antiepileptic drugs, including LTG3 and VPA.1 21 24–26 Our data did not provide adequate information on AED dose to allow analysis. The decline in female VPA prescribing is consistent with the malformation data. However, although the data from the Epilepsy and Pregnancy Registries revealed that CBZ was associated with the lowest frequency of MCMs, our results show that the prescribing of this drug has declined in girls of child-bearing age over the study period. The overall estimated risk of MCMs has declined slightly between 1993 and 2006. Although the decrease in VPA prescribing has led to a fall in this estimated MCM rate, the decrease in CBZ prescribing may have offset this slightly because this drug seems to be associated with the lowest MCM rate, according to the most recent pregnancy register data.3
Limitations and future research
There are several reservations about our estimates of MCM. No account has been taken of the AED dose; furthermore, polytherapy with AEDs has been found to be associated with a significantly higher frequency of major malformations than monotherapy.3 27–30 In this study, 30% of patients were prescribed polytherapy, including non-study AEDs, and consequently the potential risk of major malformations could be increased in these patients. It should also be noted that no account has been taken of the severity of the MCMs for each AED. It would be of considerable value to incorporate data on the effect of factors such as dose, polytherapy and severity of MCMs resulting from different AEDs into future estimates of MCMs, so as to enable the clinician to provide more accurate guidance on AED management.
This study focuses on young females, particularly those aged 12–18, and does not evaluate the use of AEDs in older females of child-bearing potential. Although the risk of malformations increases with age,31 the focus on younger females is in keeping with some very appropriate recent recommendations encouraging clinicians to consider the risk of AED treatment in girls of child-bearing potential and in younger girls who are likely to need continued treatment into adolescence and adulthood.32
It would, in theory, be possible to determine the actual MCM rates in this cohort using the mother–baby link (linkage between mother and baby records using the family ID number). A preliminary study has shown that the GPRD does meet established criteria for evaluating outcomes of pregnancy, but validation studies are required for known teratogenic associations.33 A recently published study has been conducted to validate neural tube defects in the GPRD.34 The study suggests it is feasible to determine a particular MCM. However, an investigation involving all the MCMs would be a considerable undertaking and was beyond the scope of the current study.
General practitioners do not usually record the specific type of epilepsy and the severity (seizure frequency) in the GPRD. The type and severity of the epilepsy will influence treatment preferences35 and the need for polytherapy. Further research is required to assess prescribing trends in females of child-bearing potential, taking these factors into account. Additional newer AEDs are increasingly being prescribed4 and it remains to be seen whether these drugs prove to be less teratogenic, although some preliminary data are emerging from pregnancy registers.36 37 Clinicians should use the best and most up-to-date evidence when choosing treatment for females of child-bearing potential.
Acknowledgments
We wish to thank the general practitioners who contributed data to the GPRD.
REFERENCES
Footnotes
Funding: This study is part of the Network of Excellence TEDDY (Task-force in Europe for Drug Development for the Young) supported by the EC-Sixth Framework Program (Contract no. 0005216 LSHB-CT-2005-005126). IW’s post was funded by a Department of Health Public Health Career Scientist Award to investigate the safety of psychotropic medications in children. The views expressed are those of the authors and not of the Department of Health.
Competing interests: IW has received funding from various pharmaceutical companies including GlaxoSmithKline, Janssen-Cilag and Pfizer (manufacturers of lamotrigine, topiramate and gabapentin) but none was related to the present study. FB has been sponsored to attend conferences and has received research and equipment grants from various pharmaceutical companies. He is the former editor-in-chief of a journal sponsored by GlaxoSmithKline. RA, MM and AW have no competing interests.
Ethics approval: Ethics approval was granted by the GPRD Scientific and Ethical Advisory Group.