SEARCH

SEARCH BY CITATION

Keywords:

  • AMPA ;
  • antiepileptic drugs;
  • efficacy;
  • secondary generalized seizures

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

To evaluate the efficacy of adjunctive, once-daily perampanel against secondarily generalized (SG) seizures in three Phase III trials (studies 304, 305, and 306) and their extension (study 307). The Phase III studies enrolled patients (≥12 years) with uncontrolled partial-onset seizures despite treatment with 1–3 concomitant antiepileptic drugs. Patients completing the core Phase III studies were eligible for the extension study. Endpoints included median percent change in SG seizure frequency, 50% responder (proportion of patients achieving a ≥50% reduction in SG seizure frequency), 75% response, and seizure-freedom rates. In total, 1480 patients were randomized and treated in the three perampanel Phase III trials. At baseline, 71.9% of placebo-treated and 68.4% of perampanel-treated patients had a history of SG seizures. In the individual core Phase III studies, perampanel (4–12 mg) reduced seizure frequency and improved responder rates. Consistent with this, in pooled analyses of the Phase III data, the median percent change in SG seizure frequency was −48.6%, −62.9%, and −53.3% with perampanel 4, 8, and 12 mg, respectively, vs −19.4% with placebo; 50% responder rates were 49.3%, 60.5%, and 53.7% vs 37.0% with placebo. More perampanel-treated patients had ≥75% reductions in SG seizure frequency, and seizure-freedom rates improved, compared with placebo. Improvements in seizure frequency and responder rate were maintained during the extension study. Perampanel consistently demonstrated efficacy against SG seizures when assessed using various endpoints. Furthermore, reductions in seizure frequency and improvements in responder rate were sustained with long-term perampanel treatment.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

Patients with partial-onset seizures that secondarily generalize are often inadequately treated with existing antiepileptic drugs (AEDs). Almost a third of patients with this clinically severe seizure type are resistant to treatment [1]. Uncontrolled chronic epilepsy is associated with an increased risk of mortality and physical injury, as well as significant comorbidities [1-3].

There are very few reports from randomized controlled trials of the efficacy of currently available AEDs against secondarily generalized (SG) seizures. The most extensive evidence is more than 15 years old, and demonstrated that topiramate reduced SG seizure frequency and improved SG responder rate (the proportion of patients achieving a ≥50% reduction in SG seizure frequency) compared with placebo [4-7]. In two reports from randomized, double-blind, placebo-controlled Phase III trials with either a small number of patients with SG seizures in each treatment group (n = 21–28) [8] or where the numbers of patients with SG seizures were not reported [9], levetiracetam significantly reduced SG seizure frequency compared with placebo. In another small randomized, double-blind, placebo-controlled, crossover trial in 41 patients, lamotrigine improved responder rates for SG seizures compared with placebo [10]. Moreover, in the last decade, there has only been a single Phase III AED study reporting evidence for efficacy against SG seizures. Chung et al. [11] reported improvements in both SG tonic-clonic seizure frequency and responder rate in patients with uncontrolled partial-onset seizures treated with lacosamide. A systematic review of studies in patients with partial-onset generalized tonic-clonic seizures demonstrated that the proportion of patients who achieved seizure freedom was comparable in patients treated with carbamazepine (48%), phenytoin (49%), and valproate (52%) [12]. To date, no comprehensive assessments of the efficacy of newer AEDs against SG seizures have been published.

In three randomized, double-blind, placebo-controlled Phase III trials (study 304, NCT00699972; study 305, NCT00699582; and study 306, NCT00700310), perampanel, a selective, non-competitive α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type receptor antagonist has been shown to be effective as an adjunctive therapy in patients with treatment-resistant partial-onset seizures [13-16]. Once-daily perampanel (4–12 mg) significantly reduced overall seizure frequency and improved responder rates compared with placebo in these patients. Furthermore, at effective doses (4–12 mg), perampanel had a favorable tolerability profile [13-15]. Here, we present analyses from the Phase III clinical development program of perampanel efficacy against SG seizures. Results from the three pivotal trials and their extension study (study 307, NCT00735397) are presented.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

Patients

Patients (≥12 years) with a diagnosis of simple or complex partial-onset seizures, with or without secondary generalization according to the 1981 International League Against Epilepsy Classification of Epileptic Seizures [17], were enrolled in the Phase III trials. Detailed inclusion and exclusion criteria have been reported elsewhere [13-15]. At baseline, patients were receiving treatment with stable doses of 1–3 approved AEDs.

All trials were performed in accordance with the Declaration of Helsinki, ICH-E6 Guideline CPMP/ICH/135/95, European Directive 2001/83/EC, and the US Code of Federal Regulations Part 21. The trial protocols, amendments, and informed consents were reviewed by national regulatory authorities in each country and Independent Ethics Committees or Institutional Review Boards for each site. Before trial participation, all patients gave written informed consent statements.

Phase III trial design

Core Phase III trials

The designs of the randomized, double-blind, placebo-controlled Phase III trials have been reported previously [13-15]. Briefly, following a 6-week baseline period, patients were randomized to once-daily treatment with: placebo, perampanel 8 mg, or perampanel 12 mg (1:1:1; studies 304 and 305) or placebo, perampanel 2 mg, perampanel 4 mg, or perampanel 8 mg (1:1:1:1; study 306) over a 19-week double-blind treatment phase (6-week titration period; 13-week maintenance period). Patients who completed the maintenance period of the core Phase III studies were offered entry into a long-term extension study (study 307) [18].

Phase III extension trial

The design of the extension study has been reported previously [18]. The study consisted of three phases: a 16-week blinded conversion period, a planned 256-week open-label maintenance period, and a 4-week follow-up phase [18]. The first patient entered the study in October 2008, and the interim data cutoff date for analysis was December 1, 2010.

Efficacy assessments

Seizure frequency and type were captured daily in a diary by the patient or the patient's caregiver. All simple partial seizures (with or without motor signs), complex partial seizures, and partial seizures with secondary generalization (SG seizures) were recorded. Percentage change in the frequency of SG seizures per 28 days (double-blind treatment phase vs baseline) and 50% responder rates for SG seizures (maintenance period relative to baseline) were a priori defined exploratory endpoints in the Phase III studies. The 75% response rate based on SG seizures, and SG seizure-freedom rates, were post hoc analyses.

Statistical analysis

In the three core Phase III studies, efficacy endpoints were analyzed based on randomized dose in the intent-to-treat (ITT) analysis set (all randomized and treated patients who provided any seizure frequency data during the double-blind treatment phase). The study 307 ITT analysis set included all patients who received at least one dose of perampanel during the extension study and had valid seizure data during the extension study.

All analyses are descriptive. Multiple, non-independent, unadjusted P values are reported, to highlight effects of interest.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

Patients

Overall, 1480 patients were randomized and treated during the Phase III perampanel clinical development program: 388, 386, and 706 from studies 304, 305, and 306, respectively. Patient seizure history was comparable across the Phase III studies (Table 1). The mean time since epilepsy diagnosis ranged from 19.1 to 23.7 years and complex partial seizures were the most common seizure type at baseline in all three Phase III studies (Table 1). Overall, the proportion of patients with a history of SG seizures at baseline in the core trials was 69.5%; the relative distribution of patients with a history of SG seizures was comparable across the individual studies (Table 1).

Table 1. Epilepsy-specific past medical history of patients randomized and treated in perampanel Phase III studies
 Study 304 (n = 388)Study 305 (n = 386)Study 306 (n = 706)Pooled (n = 1480)
  1. N/A, data not available; SD, standard deviation.

Mean time since diagnosis, years (SD)23.7 (13.6)22.0 (13.2)19.1 (11.7)21.3 (12.9)
Seizure type, n (%)
Simple partial without motor signs143 (36.9)133 (34.5)210 (29.7)N/A
Simple partial with motor signs128 (33.0)107 (27.7)213 (30.2)N/A
Complex partial345 (88.9)328 (85.0)593 (84.0)1266 (85.5)
Complex partial with secondary generalization279 (71.9)262 (67.9)487 (69.0)1028 (69.5)

The overall ITT analysis set included 1478 patients from studies 304 (n = 387), 305 (n = 386), and 306 (n = 705). Although 70% of patients overall (71.9% placebo-treated and 68.4% perampanel-treated) had a history of SG seizures, fewer than half of patients randomized to each dose actually had an SG seizure during the baseline period. The proportion of patients randomized to each dose who had an SG seizure during the baseline period was: 39.1%, 37.8%, 41.3%, 36.4%, and 37.3% for placebo, perampanel 2, 4, 8, and 12 mg, respectively.

Efficacy

Individual core Phase III trials

Median baseline SG seizure frequency per 28 days was comparable across the three core Phase III studies and ranged from 3.4 to 4.1 in study 304, 3.4 to 3.8 in study 305, and 2.7 to 3.7 in study 306. In all three Phase III studies, perampanel 4–12 mg reduced the frequency of SG seizures compared with placebo (Table 2). Similarly, the proportion of patients achieving a ≥50% reduction in SG seizure frequency in each of the Phase III studies also improved with perampanel 4–12 mg compared with placebo (Table 2).

Table 2. Key efficacy data for SG seizures in individual perampanel Phase III studies (ITT analysis set)
 Study 304Study 305Study 306
Placebo (n = 56)PerampanelPlacebo (n = 48)PerampanelPlacebo (n = 69)Perampanel
8 mg (n = 51)12 mg (n = 52)8 mg (n = 44)12 mg (n = 43)2 mg (n = 68)4 mg (n = 71)8 mg (n = 62)
  1. ITT, intent-to-treat; SG, secondarily generalized.

  2. a

    Seizure freedom during maintenance period in patients who completed the Phase III study maintenance period.

  3. *P < 0.05; **P < 0.01; ***P < 0.001 vs placebo. Multiple, non-independent, unadjusted P values are reported, to highlight effects of interest.

Median baseline seizure frequency per 28 days (min, max)4.1 (0.6, 79.9)3.4 (0.7, 74.2)4.1 (0.6, 138.6)3.5 (0.6, 169.4)3.4 (0.6, 158.7)3.8 (0.7, 112.0)3.4 (0.7, 37.1)3.4 (0.7, 31.4)3.7 (0.7, 104.5)2.7 (0.6, 16.1)
Median percent change in seizure frequency per 28 days (double-blind phase vs baseline), %−14.2−61.1***−75.4***−6.7−52.0***−47.4*−35.8−28.0−48.6−69.2
50% responder rate (maintenance period vs baseline), %37.566.7**59.6*25.050.0**46.544.944.149.362.9
75% responder rate (maintenance period vs baseline), %21.449.0**51.9**14.638.6**23.333.332.431.050.0
Seizure-freedom rates, %a10.033.3**35.7**4.825.7*15.623.815.322.727.3

Among those patients who completed the maintenance period of the core studies, there was generally a dose-dependent increase in the proportions of perampanel-treated patients achieving SG seizure freedom during the maintenance period (Table 2). Furthermore, more patients randomized to perampanel 8 and 12 mg in studies 304 and 305 achieved greater reductions (≥75%) in SG seizure frequency compared with placebo (Table 2).

Pooled Phase III trials

Results from the pooled Phase III ITT analysis set by randomized dose were consistent with those from the individual studies in demonstrating improvements in the median percentage change from baseline in SG seizure frequency per 28 days compared with placebo (Fig. 1A; Table 3). Similarly, 50% responder rates for SG seizures also improved in patients randomized to perampanel 2–12 mg (Fig. 1B; Table 3).

image

Figure 1. Median percentage change in SG seizure frequency per 28 days (double-blind phase vs baseline) and 50% responder rate for SG seizures (maintenance period vs baseline) in the pooled Phase III ITT analysis set. Multiple, non-independent, unadjusted P values are reported to highlight effects of interest. **P < 0.01; ***P < 0.001 vs placebo. ITT, intent-to-treat; SG, secondarily generalized.

Download figure to PowerPoint

Table 3. Key efficacy data for SG seizures by randomized dose in the pooled Phase III ITT analysis set
 PlaceboPerampanel
2 mg4 mg8 mg12 mg
  1. ITT, intent-to-treat; SG, secondarily generalized.

  2. a

    Seizure freedom during maintenance period in patients who completed the Phase III study maintenance period.

  3. *P < 0.05; **P < 0.01; ***P < 0.001 vs placebo. Multiple, non-independent, unadjusted P values are reported, to highlight effects of interest.

Randomized dosen = 173n = 68n = 71n = 157n = 95
Median percent change in seizure frequency per 28 days (double-blind phase vs baseline), %−19.4−28.0−48.6**−62.9***−53.3***
50% responder rate (maintenance period vs baseline), %37.044.149.360.5***53.7**
75% responder rate (maintenance period vs baseline), %24.332.431.046.5***38.9*
Seizure-freedom rates, %a14.215.322.728.9**27.0*

The pooled Phase III analyses were also consistent in demonstrating that, overall, more perampanel-treated patients achieved large (≥75%) reductions in SG seizure frequency compared with placebo-treated patients across the Phase III studies: 32.4% with perampanel 2 mg, 31.0% with perampanel 4 mg, 46.5% with perampanel 8 mg, and 38.9% with perampanel 12 mg vs 24.3% with placebo (P = 0.279 for 4 mg; P < 0.001 for 8 mg, and P = 0.012 for 12 mg vs placebo). Overall, within the effective dose range of perampanel, SG seizure-freedom rates were also higher compared with placebo: 22.7% with 4 mg, 28.9% with 8 mg, and 27.0% with 12 mg vs 14.2% with placebo (P = 0.167 for 4 mg, P = 0.002 for 8 mg, and P = 0.028 for 12 mg vs placebo).

Impact of common concomitant AEDs on efficacy

Patients in the Phase III studies had uncontrolled partial-onset seizures, despite the large majority receiving ongoing treatment with ≥2 AEDs at baseline (1 AED, n = 206, 13.9%; 2 AEDs, n = 751, 50.7%; 3 AEDs, n = 522, 35.3%). The mean number of AEDs at baseline was 2.2. The most common concomitant AEDs at baseline were carbamazepine (n = 491, 33.2%), valproic acid (n = 478, 32.3%), lamotrigine (n = 458, 30.9%), and levetiracetam (n = 435, 29.4%).

In subgroups of patients who were receiving concomitant treatment with at least 1 of the 4 most common AEDs across the Phase III studies, the overall median reduction in SG seizure frequency increased dose dependently with perampanel 4–12 mg compared with placebo, irrespective of the mechanism of action of these 4 AEDs (Table 4). Furthermore, across these subgroups, perampanel 8 mg and 12 mg consistently improved 50% responder rates for SG seizures compared with placebo.

Table 4. Median percent change in SG seizure frequency per 28 days (double-blind phase vs baseline) and 50% responder rate for SG seizures (maintenance period vs baseline) by randomized dose in the pooled Phase III ITT analysis set, according to the concomitant AEDs at baseline
 PlaceboPerampanel
2 mg4 mg8 mg12 mg
  1. ITT, intent-to-treat; SG, secondarily generalized.

  2. No inferential analyses were performed on pooled data.

Median percent change in seizure frequency per 28 days
Concomitant carbamazepine, % (number of patients evaluated)−12.7 (51)−30.1 (21)−33.2 (20)−53.8 (43)−52.8 (33)
Concomitant valproic acid, % (number of patients evaluated)−34.4 (64)−19.7 (29)−48.3 (30)−65.4 (46)−75.4 (24)
Concomitant lamotrigine, % (number of patients evaluated)−22.9 (45)−65.4 (22)−60.5 (27)−61.1 (54)−75.5 (22)
Concomitant levetiracetam, % (number of patients evaluated)−19.1 (43)−26.9 (20)−63.6 (15)−61.1 (41)−70.7 (27)
50% responder rate
Concomitant carbamazepine, % (number of patients evaluated)37.3 (51)38.1 (21)35.0 (20)51.2 (43)51.5 (33)
Concomitant valproic acid, % (number of patients evaluated)42.2 (64)27.6 (29)43.3 (30)67.4 (46)62.5 (24)
Concomitant lamotrigine, % (number of patients evaluated)35.6 (45)54.5 (22)63.0 (27)55.6 (54)68.2 (22)
Concomitant levetiracetam, % (number of patients evaluated)39.5 (43)45.0 (20)66.7 (15)61.0 (41)66.7 (27)

Phase III extension study

Of 1264 patients who completed the double-blind phase of the three core Phase III trials, 1218 (96.4%) patients continued into study 307. As in the core studies, approximately 70% (n = 854) of patients had a history of SG seizures.

The greater reduction in SG seizure frequency per 28 days observed in perampanel-treated patients at the end of the core Phase III double-blind phase was sustained throughout the extension study conversion and maintenance periods. At interim analysis, the overall median percent changes in SG seizure frequency were as follows: −65.3% for the extension conversion period (n = 465); −77.5% for weeks 1–13 of the open-label maintenance period (n = 333); −88.2% for weeks 14–26 (n = 233); −86.1% for weeks 27–39 (n = 168); and −87.1% for weeks 40–52 (n = 104). During the 16-week conversion period, there was only a small difference in the median percentage change in SG seizure frequency between patients previously randomized to placebo in the double-blind studies (−59.9%, n = 148) and those randomized to perampanel (−69.9%, n = 317). This reduction was then maintained in both groups of patients during the extension study maintenance period (Fig. 2A).

image

Figure 2. Median percent change in SG seizure frequency per 28 days (relative to pre-perampanel baseline) and 50% responder rate for SG seizures at the end of the double-blind maintenance period, the end of the extension study blinded conversion period, and by each 13-week interval during the open-label maintenance period (ITT analysis set). ITT, intent-to-treat; SG, secondarily generalized.

Download figure to PowerPoint

Similarly, the improvement in 50% responder rate for SG seizures with perampanel during the core Phase III studies was also maintained throughout the extension study. The overall responder rates for SG seizures were as follows: 59.1% for the extension study conversion period (n = 465); 68.2% for weeks 1–13 of the open-label maintenance period (n = 333); 69.1% for weeks 14–26 (n = 233); 67.3% for weeks 27–39 (n = 168); and 68.3% for weeks 40–52 (n = 104). At the end of the 16-week conversion period, there was little difference in responder rates for SG seizures between patients previously randomized to placebo in the double-blind studies (55.4%, n = 148) and those randomized to perampanel (60.9%, n = 317). The initial improvement in responder rate at the end of the conversion period was again maintained in both groups of patients during the extension study maintenance period (Fig. 2B).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

Convulsive seizures are the most clinically severe and disruptive seizure type in adults. Novel therapies efficacious against SG seizures have the potential to provide substantial clinical benefit and improved quality of life to patients who are refractory to existing AEDs. There is a paucity of data demonstrating the efficacy of second- and third-generation AEDs against SG seizures. While there are a number of reports demonstrating the efficacy of topiramate [4-7], these were published more than 15 years ago at a time when fewer AEDs were available for treatment of refractory epilepsy. Consequently, in these studies, between 36 and 45% of all patients were receiving concomitant treatment with only one AED [4-6]. Furthermore, there are only limited results from one or two studies each with levetiracetam, lamotrigine, and lacosamide [8-11]. The results reported here demonstrate the efficacy of perampanel against SG seizures, based on multiple efficacy endpoints, from three randomized, double-blind, placebo-controlled Phase III trials and their long-term extension study. These results are borne out in analyses that demonstrate that perampanel continued to show impressive efficacy in over 100 patients with at least 1 year of data in the extension trial.

At baseline in the perampanel Phase III clinical development program, 35% of patients had uncontrolled partial-onset seizures despite ongoing treatment with the maximum of 3 concomitant AEDs. Furthermore, patients generally had higher baseline seizure frequencies and more often had a history of SG seizures compared with registration trials for previous second- and third-generation AEDs [19]. In each of three large Phase III studies, once-daily perampanel demonstrated efficacy in terms of reducing seizure frequency and increasing responder rate compared with placebo in this hard-to-treat patient population. In studies 304 and 305, which included the highest dose of perampanel investigated (12 mg), there were large reductions in SG seizure frequency compared with placebo at this dose (percentage changes in seizure frequency of −75.4% vs −14.2% and −47.4% vs −6.7% in studies 304 and 305, respectively). Furthermore, the proportion of patients achieving a ≥50% reduction in SG seizure frequency almost doubled with perampanel 12 mg vs placebo in both studies. At higher perampanel doses, in general, more patients had large reductions in seizure frequency, and more patients were free from SG seizures with perampanel 12 mg. Seizure-free rates ranged from 15.3% to 35.7% in the perampanel-treated groups. Consistent with results from the individual studies, pooled analyses of the Phase III trials also demonstrated efficacy of perampanel against SG seizures. Furthermore, concomitant treatment with at least 1 of the 4 most common AEDs at baseline did not diminish the efficacy of perampanel against SG seizures in the pooled Phase III analysis, irrespective of the mechanism of action of the AED. The large improvements in seizure frequency and responder rate with perampanel were sustained with ongoing treatment during the extension study.

While topiramate has been shown to reduce SG seizure frequency and improve responder rate in a number of studies [4-7], only two of these studies assessed the efficacy of topiramate at doses recommended for treatment of patients with partial-onset seizures (maximum 400 mg/day) [5, 7]. Both of these reported large median reductions in SG seizure frequency and high responder rates with topiramate 400 mg/day; however, the number of patients investigated was low (n = 14–15), so these results should be interpreted with caution. Only the study by Pritivera et al. [6], which assessed high doses of topiramate (600–1000 mg/day), included patient numbers comparable to those reported here for perampanel. In this study, the change in SG seizure frequency and improvements in responder rate observed with topiramate 1000 mg/day were more modest compared with placebo (−78.0% vs −40.3% and 55% vs 35%, respectively). In a Phase III study, the magnitude of change in SG seizure frequency with levetiracetam (−37.4% and −28.2% with levetiracetam 1000 mg/day and 2000 mg/day, respectively, vs +16.8% with placebo) was comparable to that observed with the highest dose of perampanel in the individual studies reported here [8]. More recently, lacosamide was shown to reduce SG tonic-clonic seizures dose dependently compared with placebo in a Phase III study. At the highest dose approved for maintenance treatment of seizures (400 mg/day), the change in seizure frequency with lacosamide compared with placebo (−59.4% vs −14.3%, respectively) was comparable to that seen with the highest dose of perampanel in studies 304 and 305 [11]. Furthermore, similar to the improvement observed with perampanel 12 mg in studies 304 and 305, responder rates for SG seizures also improved approximately twofold with lacosamide 400 mg/day vs placebo [11].

In summary, perampanel has consistently demonstrated efficacy against SG seizures across the Phase III clinical development program. Reductions in seizure frequency and improvements in responder rate were maintained with ongoing perampanel treatment during the long-term extension study. Perampanel has the potential to be a valuable addition to the armament of AEDs available for the treatment of uncontrolled SG seizures.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

David Squillacote, MD, of Eisai Inc. assisted in the preparation of the article by providing up-to-date data access and editorial support. Editorial support was provided by Deborah McGregor, PhD, of Complete Medical Communications and was funded by Eisai Inc.

Conflicts of interest and sources of funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References

D. Ko has participated in the speakers’ bureau for GlaxoSmithKline, Lundbeck, and UCB. He has also served as a consultant to Eisai, Supernus, and Westward. E. Ramsay has participated in the speakers’ bureau for Lundbeck and UCB. He has also served as a consultant to Supernus, Eisai, UCB, and GlaxoSmithKline and has had research funded by Lundbeck and Eisai. Studies were funded by Eisai Inc.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest and sources of funding
  9. References