FULL-LENGTH ORIGINAL RESEARCH
Generic lamotrigine versus brand-name Lamictal bioequivalence in patients with epilepsy: A field test of the FDA bioequivalence standard *Tricia Y. Ting, †Wenlei Jiang, †Robert Lionberger, ‡Jessica Wong, ‡Jace W. Jones, ‡Maureen A. Kane, *Allan Krumholz, †Robert Temple, and ‡James E. Polli Epilepsia, 56(9):1415–1424, 2015 doi: 10.1111/epi.13095
SUMMARY
Dr. Tricia Y. Ting is an epileptologist and associate professor of neurology at University of Maryland.
Objective: To test the current U.S. Food and Drug Administration (FDA) bioequivalence standard in a comparison of generic and brand-name drug pharmacokinetic (PK) performance in “generic-brittle” patients with epilepsy under clinical use conditions. Methods: This randomized, double-blind, multiple-dose, steady-state, fully replicated bioequivalence study compared generic lamotrigine to brand-name Lamictal in “generic-brittle” patients with epilepsy (n = 34) who were already taking lamotrigine. Patients were repeatedly switched between masked Lamictal and generic lamotrigine. Intensive PK blood sampling at the end of each 2-week treatment period yielded two 12-h PK profiles for brand-name and generic forms for each patient. Steady-state area under the curve (AUC), peak plasma concentration (Cmax), and minimum plasma concentration (Cmin) data were subjected to conventional average bioequivalence (ABE) analysis, reference-scaled ABE analysis, and within-subject variability (WSV) comparisons. In addition, generic-versus-brand comparisons in individual patients were performed. Secondary clinical outcomes included seizure frequency and adverse events. Results: Generic demonstrated bioequivalence to brand. The 90% confidence intervals of the mean for steady-state AUC, Cmax, and Cmin for generic-versus-brand were 97.2– 101.6%, 98.8–104.5%, and 93.4–101.0%, respectively. The WSV of generic and brand were also similar. Individual patient PK ratios for generic-versus-brand were similar but not identical, in part because brand-versus-brand profiles were not identical, even though subjects were rechallenged with the same product. Few subjects had seizure exacerbations or tolerability issues with product switching. One subject, however, reported 267 focal motor seizures, primarily on generic, although his brand and generic PK profiles were practically identical. Significance: Some neurologists question whether bioequivalence in healthy volunteers ensures therapeutic equivalence of brand and generic antiepileptic drugs in patients with epilepsy, who may be at increased risk for problems with brand-to-generic switching. Bioequivalence results in “generic-brittle” patients with epilepsy under clinical conditions support the soundness of the FDA bioequivalence standards. Adverse events on generic were not related to the small, allowable PK differences between generic and brand. KEY WORDS: Bioequivalence, Switchability, Lamotrigine, Generic-brittle, Narrow therapeutic index.
Accepted June 29, 2015; Early View publication July 23, 2015. *Department of Neurology, University of Maryland, Baltimore, Maryland, U.S.A.; †Food and Drug Administration, White Oak, Maryland, U.S.A.; and ‡Department of Pharmaceutical Sciences, University of Maryland, Baltimore, Maryland, U.S.A. Address correspondence to James E. Polli, Department of Pharmaceutical Sciences, University of Maryland, 20 Penn Street, Baltimore, MD 21201, U.S.A. E-mail:
[email protected] Wiley Periodicals, Inc. © 2015 International League Against Epilepsy
1415
1416 T. Y. Ting et al.
Key Points • • • •
A novel bioequivalence study of Lamictal versus generic lamotrigine was performed using “generic-brittle” patients under clinical use conditions. Generic demonstrated bioequivalence to brand in terms of pharmacokinetic AUC, Cmax, and Cmin. The WSV of generic and brand was also similar. Individual patient ratios for generic-versus-brand were similar but not identical, in part because brand-versusbrand profiles were not identical. Results in “generic-brittle” patients with epilepsy under clinical conditions support the soundness of the FDA bioequivalence standards.
The Drug Price Competition and Patent Term Restoration Act, passed in 1984, greatly facilitated the approval of generic drugs. No longer required to directly demonstrate the safety and efficacy of generic products in clinical trials, manufacturers may now rely on pharmaceutical equivalence and bioequivalence (BE) of generic to the original brandname drug for approval.1 Pharmaceutical equivalence requires medicines to have the same exact drug, dose strength, dosage form, and route of administration, and BE requires that medicines have no significant difference in the rate and extent of drug absorption. Today, 86% of prescriptions are filled using generic products.2 For cost savings, patients frequently switch from brand to generic when a generic becomes available. The generic often differs from brand in inactive ingredients, manufacturing process, and appearance. The safety and efficacy of generics have been questioned,3 particularly when there is generic substitution of antiepileptic, immunosuppressant, and psychotropic drugs.4–10 Increased seizures and side effects have been reported after antiepileptic drug (AED) generic substitution, attributed to differences in bioavailability allowed between brand and generic products.11–13 Rare cases of acute rejection in kidney transplant patients have been reported after conversion from reference to generic tacrolimus.14,15 Such observations have led to criticism of U.S. Food and Drug Administration (FDA) approval requirements for generics, considered by some to be less rigorous than for brand-name medications because, unlike most brands, generics are not required to directly demonstrate therapeutic efficacy in patients. Concerns have been raised over whether current BE standards, and the common use of healthy subjects in BE studies, are sufficient to ensure the therapeutic equivalence of generic drug substituted for branded drugs,16–18 particularly for patients who may be at risk for serious, life-threatening therapeutic failure if small differences in drug concentration were to occur. Epilepsia, 56(9):1415–1424, 2015 doi: 10.1111/epi.13095
An important public health standard, BE testing is critical for approval of brand and generic drug products.19 Practically all pharmaceutical companies—both brand-name and generic—employ BE (e.g., pharmacokinetic similarity) to demonstrate equivalence throughout a drug’s life cycle, which can span decades from first-in-human studies to postapproval manufacturing changes. New Drug Applications (NDAs) seeking FDA approval typically include 3–6 BE studies to ensure that the final product is equivalent to prototype formulations.20,21 Even postapproval, some product manufacturing changes (e.g., factory location, new technology, batch size) necessitate in vivo BE testing of the old and new versions of the brand or generic product.22 BE testing may sometimes be the sole source of clinical data for approval of new combination products that contain more than one approved drug and for formulation conversion to extended-release.23 The current BE standard for most drugs is average bioequivalence (ABE).19 ABE assesses average differences between two formulations, just as efficacy studies generally rely on average differences in clinical outcome. Two pharmacokinetic (PK) parameters, peak plasma concentration (Cmax) and area under the curve (AUC), are required to pass FDA standards, that is, the 90% confidence interval (CI) for the test-to-reference ratio of geometric means must fall within the goalposts of 80.00–125.00%. This type of ABE is conventional or unscaled ABE19 and applies to most drugs, but not to highly variable or narrow therapeutic index (NTI) drugs. Unlike clinical trials, which involve patients on chronic dosing, ABE studies typically compare drug PK profiles in healthy volunteers using single dose. ABE largely does not focus on individual profile differences. For this reason, ABE allows for a study to pass (i.e., averages are close) even though individual “outliers” may show testto-reference ratios well beyond 25%. In contrast to ABE, a prior BE standard focused on individual rather than averaged profile differences, allowing greater consideration of “outliers” in the approval process. This 75/75 rule required at least 75% of individual subjects to exhibit no more than a 25% difference between test and reference PK profiles.24 It should be noted also that most ABE studies use only a single exposure of the test and reference products, so that it is not possible to assess whether differences between test and reference are larger than would be seen in a comparison of the reference to itself. The following concerns have been raised about ABE with antiepileptic, immunosuppressant, and psychotropic agents: (1) by relying only on healthy volunteers, ABE testing may miss clinically significant bioavailability differences that could be relevant in patients; (2) by focusing on the average, rather than individual PK profile sameness, ABE testing may not ensure therapeutic equivalence in all individuals, particularly in patients who we refer to as “generic-brittle,” who may be too few to affect the average; and (3) some medications may be NTI drugs “where small differences in
1417 Switchability in “Generic-Brittle” Patients dose or blood concentration may lead to serious therapeutic failures and/or adverse drug reactions that are life-threatening or result in persistent or significant disability or incapacity.”25,26 If so, more stringent BE criteria for NTI drugs using reference-scaled ABE and within-subject variability (WSV) comparisons25,26 are warranted. To date, the FDA has designated several AEDs such as carbamazepine and phenytoin products to be NTI drugs,27,28 which employ reference-scaled ABE and within-subject variability WSV comparisons, rather than conventional/unscaled ABE.19 In this study, we sought to examine whether a generic AED that was approved via healthy volunteer testing would meet the same ABE standard or even tighter BE standards, such as for NTI drugs, when tested in patients who were blinded to formulation and potentially sensitive to problems with generic switching.
Methods Design and “generic-brittle” patients A randomized, double-blind, multiple-dose, steadystate, fully replicated crossover BE study of lamotrigine immediate-release tablets was conducted in patients with epilepsy (ClinicalTrials.gov identifier: NCT01995825). Lamotrigine was studied in this novel, single-center, outpatient BioEquivalence in Epilepsy Patients (BEEP) study because the therapeutic equivalence of generic lamotrigine to brand Lamictal has been frequently questioned in anecdotal reports.29–33 The most commonly dispensed generic immediate-release lamotrigine (manufactured by Teva) was compared to Lamictal (GlaxoSmithKline), employing 100 mg tablets at the patient’s usual dose. For the n = 34 patients who completed, n = 9, 16, and 9 were on 100 mg b.i.d., 200 mg b.i.d., and 300 mg b.i.d. lamotrigine dosing, respectively (mean 200 mg b.i.d.). Study size was selected to allow for 80% power to demonstrate ABE. Subjects were “generic-brittle” adult patients already taking lamotrigine for either focal or generalized epilepsy. We defined “generic-brittle” as having a potential problem with generic switching by virtue of (1) a history of reported prior exacerbation of seizures or side effects following AED formulation changes; (2) intolerable AED side effects within the last year prior to study; or (3) refractory seizures within the last year prior to study, which could reflect clinical sensitivity to slightly higher AED peak plasma concentration or slightly lower drug exposure, respectively. These patients were permitted concomitant AEDs—including inducers and valproic acid (an inhibitor)—and other medications for comorbid disorders. Their medications remained unchanged during the study. Each patient in this four-period, fully replicated crossover design was given identical product (brand and generic) on two periods, allowing PK comparison of generic-to-brand, as well as brand-to-brand and generic-to-generic. Patients
were randomized to one of two sequences (generic, brand, generic, and brand; or the inverse). Because a “nocebo” (negative placebo) effect from anticipation of generic switching had the potential to impact study adherence, retention, and outcomes,34 tablets were over-encapsulated with tamper-resistant capsules to mask administered product. Both over-encapsulated brand and generic lamotrigine tablet demonstrated comparable quality characteristics and conformed to U.S. quality standards. At the end of each 2-week treatment period, there was intensive PK blood sampling over 12 h. Three trough level samples ensured that a steady-state PK lamotrigine level was achieved prior to PK sampling. This steady-state BE study in patients can be considered to be under routine diet conditions, where patients are neither under strict fasting nor fed conditions, but under representative clinical drug use conditions. However, on each 12 h PK sampling day, the patient was fasted overnight for 10 h predose, plus an additional 4 h postdose. The 24 h trough level was also collected after a 10 h overnight fast. Each completed patient yielded two 12-h PK profiles for brand and generic at the end of the 2-week treatment periods. Lamotrigine was quantified using liquid chromatography-tandem mass spectrometry.35 Steady-state AUC, Cmax, and minimum plasma concentration (Cmin) data were subjected to conventional ABE analysis,19 reference-scaled ABE analysis,25–28 and WSV comparisons.25–28 Achievement of steady-state PK via trough measurements, as well as the observed average BE, demonstrated favorable study integrity and patient adherence. Clinical measures of efficacy and tolerability were based on self-reporting, as is routine in phase III epilepsy drug trials. Secondary outcomes included seizure frequency, but it was recognized that the study was not powered to detect small differences. Baseline seizure frequency was not documented prospectively, and a minimum seizure frequency was not an enrollment criterion. Daily seizure frequency was recorded by date but not by hour of the day. Adverse events (AEs) were assessed by staff and recorded by patients. Intention-to-treat (ITT) analysis of tolerability focused on the most common dose-related AEs reported for lamotrigine.36 Prior to data collection, the study was approved by the University of Maryland institutional review board (IRB) and FDA Research in Human Subjects Committee. Study participants gave their written informed consent. Error bars denote standard error of the mean (SEM).
Results Thirty-five generic-brittle patients aged 19–66 years were enrolled, with 34 completing per-protocol (Table 1). Patients had a history of focal and/or generalized seizures. All subjects were considered generic-brittle by virtue of (1) Epilepsia, 56(9):1415–1424, 2015 doi: 10.1111/epi.13095
1418 T. Y. Ting et al. Table 1. Patient demographics Sex Age range (mean years) Epilepsy Focal Generalized AED concomitant Valproic acid (an inhibitor) Inducer Smoking (an inducer) Comorbid conditions None One or more
Male N = 20
Female N = 15
N = 35
19–66 (44)
20–63 (39)
19–66 (42)
17 3
10 5
27 8
3 3 1
0 3 2
3 6 3
9 11
4 11
13 22
a history of prior problems with switching AED formulations, (2) AED adverse events, or (3) refractory seizures. The majority (n = 23) met two or more of the three criteria to define generic-brittle, including nine subjects who had a history of prior problems switching AED formulations. Of these nine, eight subjects had a history of increased seizures and one subject had had adverse drug effects during previous formulation switches. A majority of the subjects, 27 total, had refractory seizures. Eight subjects were older than 50 years of age. Six subjects were on phenytoin or carbamazepine, the only inducers of lamotrigine metabolism in the study. The average number of total comedications (including AEDs and non-AEDs) was 3.4 (median 3; minimum 1; and maximum 9). Fourteen subjects were on four or more co-medications. Nine, 16, and nine subjects took 100, 200, and 300 mg lamotrigine b.i.d., respectively. Trough levels, tablet counts, and patient diaries verified patient adherence and steady-state PK. Comorbid conditions included diabetes, hypertension, hyperlipidemia, headache, and mood disorder.
ABE, reference-scaled ABE, and WSV comparisons Generic was bioequivalent to the brand, with generic-versus-brand geometric mean ratio close to 1 and 90% confidence interval for steady state AUC, Cmax, and Cmin being narrow (Table 2). Figure 1 illustrates the high similarity in the mean profiles (panel A), as well individual profiles for subject 026 (panel B) and subject 024 (panel C). CIs for all three PK parameters fell well within the conventional/unscaled 80–125% limits, meeting FDA criteria for ABE and confirming data from healthy volunteer generic studies, where average differences are about 4–5%.37 AUC, Cmax, and Cmin also fell within the tighter BE goalposts for reference-scaled ABE, which were 93.65–106.78%, 91.85– 108.88%, and 90.80–110.14%, respectively (Table 2). These tighter BE limits were scaled based on brand lamotrigine WSV, which ranged from 6.38% to 9.39% (Table 2). ABE and reference-scaled ABE showed that the means of generic and brand were similar. WSV comparisons showed that generic and brand exhibited similar PK variability (Table 2). WSV of generic was slightly larger than WSV of brand for AUC (1.29–fold), Cmax (1.38-fold), and Cmin (1.44-fold), but each PK metric passed, since the upper limit of the 90% CI for ratio of WSV was ≤2.5.25–28 This upper limit of 2.5 was selected as the variability comparison criterion because it could produce >80% power for similar products and
