Summary Basis of Decision for Saphris ™

3.3.1 Pharmacodynamics

Asenapine has a high affinity for serotonin 5-HT_1A, 5-HT_1B, 5-HT_2A, 5-HT_2B, 5-HT_2C, 5-HT₅, 5-HT₆, and 5-HT₇ receptors (Ki values of 2.5, 2.7, 0.07, 0.18, 0.03, 1.6, 0.25, and 0.11 nM, respectively); dopamine D_2A, D_2B D₃, D₄, and D₁ receptors (Ki values of 1.3, 1.4, 0.42, 1.1, and 1.4 nM, respectively); α_1A, α_2A, α_2B and α_2C-adrenergic receptors (Ki values of 1.2, 1.2, 0.33, and 1.2 nM, respectively); and histamine H₁ receptors (Ki value 1.0 nM); and moderate affinity for H₂ receptors (Ki value of 6.2 nM).

Asenapine had no appreciable affinity for muscarinic cholinergic receptors.

The in vitro assays showed that asenapine acts as an antagonist at most of the 5HT₁ receptors. However, as with other antipsychotic drugs showing appreciable affinity for 5HT_1A receptors, electrophysiological studies showed that asenapine is a partial agonist at this receptor.

3.3.2 Pharmacokinetics

Absorption

Following sublingual administration, asenapine was rapidly absorbed with peak plasma concentrations occurring within 0.5 to 1.5 hours. Up to a dose of 5 mg BID, the systemic exposure of asenapine increased proportionally to the dose. At higher doses [that means (i.e.) 5-10 mg BID] deviation from dose-proportionality was observed.

The absolute bioavailability of sublingual asenapine at 5 mg was estimated to be approximately 35%. Part of the asenapine dose administered via the sublingual route was swallowed, as the AUC of N-desmethylasenapine after sublingual administration of asenapine was considerably lower after treatment with charcoal, than without charcoal pre-treatment, while the AUC of asenapine was essentially unchanged. The absolute bioavailability of asenapine when swallowed was very low (<2% with an oral tablet formulation).

Intake of water 2 or 5 minutes after asenapine sublingual administration resulted in decreased systemic exposure. A high-fat meal immediately before asenapine dosing, as well as food given 4 hours after asenapine administration also reduced asenapine exposure.

Distribution

Asenapine was rapidly distributed with a large volume of distribution. It was highly bound (95%) to plasma proteins, including albumin and α₁-acid glycoprotein.

Metabolism

The primary metabolic pathways for asenapine were direct glucuronidation by the protein UDP-glucuronosyltransferase 1-4 (UGT1A4) and oxidative metabolism by CYP isoenzymes (predominantly CYP1A2).

Metabolite profiling demonstrated at least 5 different peaks in plasma, at least 15 different peaks in urine, and at least 13 different peaks in faeces. The peaks in plasma were associated with asenapine, N-desmethylasenapine, asenapine N⁺-glucuronide, N-desmethylasenapine-N-carbamoylglucuronide and asenapine 11-O-sulfate. None of these metabolites are expected to provide a relevant contribution to the pharmacological activity of asenapine due to lower affinity for the receptors or the inability to cross the blood brain barrier. Asenapine N⁺-glucuronide and unchanged asenapine were the major drug-related entities in plasma samples.

Excretion

After administration of a single dose of radiolabelled asenapine, approximately 90% of the dose was recovered: approximately 50% in the urine; and 40% in the faeces.

Special Populations

Renal Impairment

The pharmacokinetics of asenapine and N-desmethylasenapine following a single dose of 5 mg asenapine were similar among subjects with varying degrees of renal impairment and subjects with normal renal function, indicating that dosage adjustment based upon the degree of renal impairment is not required.

Hepatic impairment

The pharmacokinetics of asenapine and its metabolites N-desmethylasenapine and asenapine N⁺-glucuronide following a single dose of 5 mg asenapine were similar among subjects with mild or moderate hepatic impairment (Child-Pugh Class A and B) and subjects with normal hepatic function, indicating that dosage adjustment is not required for these subjects. In subjects with severe hepatic impairment (Child-Pugh Class C) substantial increases in exposure were observed: 7-fold for asenapine; 3-fold for N-desmethylasenapine; and 2-fold for asenapine N⁺-glucuronide. Based on these observations,asenapine is not recommended for patients with severe hepatic impairment.

Adolescents

At the 5 mg BID dose level, asenapine pharmacokinetics in the adolescent population were similar to those observed in adults. In adolescents, the 10 mg BID dose did not result in higher exposure of asenapine compared to the 5 mg BID dose. This appears to result from subjects partially swallowing the 10 mg dose in the study.

Elderly

In a preliminary data set, exposure to asenapine was on average 30% higher in elderly patients (65-85 years of age) compared to younger adult patients. No dose adjustment is necessary.

Race

A comparative study in Caucasian and Japanese subjects did not detect an effect of race on asenapine pharmacokinetics.

Gender

No clinically significant differences in pharmacokinetics were observed between men and women.

Drug Interaction Studies

Asenapine is cleared primarily through direct glucuronidation and oxidative metabolism (predominantly by CYP1A2). The potential effects of inhibitors and an inducer of several of these enzyme pathways on asenapine pharmacokinetics were studied. With the exception of fluvoxamine (a strong CYP1A2 inhibitor), none of the drugs resulted in clinically relevant alterations in asenapine pharmacokinetics. A summary of the effects of these coadministered drugs is included in the Saphris™ Product Monograph.

3.3.3 Clinical Efficacy

Efficacy in Schizophrenia

The efficacy of Saphris™ in the treatment of schizophrenia was evaluated in five pivotal studies. Four studies (Studies 041004, 041021, 041022, 041023) were short-term (6 weeks), randomized, double-blind, placebo and active-controlled studies of patients who met DSM-IV criteria for schizophrenia and were having an acute exacerbation of their schizophrenic illness. Three of the placebo/active controlled studies were fixed-dose studies: a Phase II study with Saphris™ 5 mg BID (Study 041004); and two Phase III studies with Saphris™ 5 mg BID and 10 mg BID (Studies 041021 and 041023). Study 041022 was a flexible dose study (Saphris™ 5 mg - 10 mg BID). All four studies included an active control group (risperidone, olanzapine, or haloperidol) but were not powered to compare Saphris™ with the active comparator. The fifth pivotal study (Study A7501012) was a standard relapse prevention study with six months of open-label treatment with Saphris™ followed by randomization to placebo or Saphris™ for an additional six months. This study allowed flexible dosing of Saphris™ 5-10 mg BID.

Three out of four of the short-term 6-week studies were Phase III studies that included approximately 100 patients per treatment group. The fourth short-term study was a Phase II study (Study 041004) with approximately 60 patients per group. In the relapse prevention study, there were approximately 200 patients in each of the randomized treatment arms.

The primary efficacy rating scale was the Positive and Negative Syndrome Scale (PANSS). The primary efficacy analysis was a comparison of the least squares (LS) mean changes from baseline to endpoint (LOCF) in the PANSS total score for each Saphris™ treatment group versus (vs.) placebo. The primary and secondary efficacy endpoints included change from baseline on PANSS total score and each of the PANSS subscale scores (PANSS positive, negative and general psychopathology subscales), the PANSS Marder factor scores, and the Clinical Global Impression severity score (CGI-S).

Pivotal Study Results

Study 041004 was the placebo-controlled Phase II study which used Saphris™ at 5 mg BID, and included an active control (risperidone). Saphris™ was significantly more effective than placebo in reducing the symptoms of schizophrenia at Weeks 2 through 6 (endpoint), as measured by total PANSS score. Saphris™ was also significantly more effective than placebo in reducing PANSS subscale scores at Weeks 3 through 6. For Clinical Global Impression (Severity of Illness, and Global Improvement), both Saphris™ and risperidone were significantly better than placebo at endpoint, and showed similar magnitudes of change. Study 041004 was deemed a positive study in support of efficacy of Saphris™ in acute treatment of schizophrenia.

Study 041021 had two fixed-dose treatment arms of Saphris™ (5 mg BID and 10 mg BID), as well as an active control (olanzapine). Statistically significant results were seen only for olanzapine. Therefore, Study 041021 was deemed a negative study.

Study 041022 was a flexible-dose study (initial Saphris™ target dose was 10 mg BID with an option to reduce to 5 mg BID) with an active control (olanzapine). In this study, neither Saphris™ nor olanzapine was statistically superior to placebo. Study 041022 was deemed a failed study.

Study 041023 compared Saphris™ 5 mg BID, Saphris™ 10 mg BID, haloperidol 4 mg BID, and placebo (approximately 110 patients per group). Statistically significant differences were observed in the primary efficacy analysis (baseline to endpoint in the PANSS total score) with patients treated with Saphris™ 5 mg BID and haloperidol 4 mg BID, compared to patients treated with placebo. The results from the Saphris™ 10 mg BID group were not significant. However, in the observed case and repeated measure analyses, the Saphris™ 10 mg BID group also reached statistically significance compared to placebo. This was also seen on some secondary efficacy endpoints (e.g., PANSS positive subscale score, and PANSS response rates). Overall, Study 041023 was deemed positive. However, the study highlighted that Saphris™ 5 mg BID was as, if not more, efficacious than the higher dose.

Based on the four short-term acute treatment studies, there was one negative study, one failed study, and two studies supporting Saphris™ efficacy at the 5 mg BID dose, with equivocal findings at the 10 mg BID dose.

Study A7501012 was a relapse prevention study. While this study design is not considered to support a relapse prevention indication, it is considered of value for the assessment of drug efficacy. In this study, for those patients who were able to tolerate and remain on Saphris™ for a 26-week open-label period prior to randomization to placebo or ongoing Saphris™, Saphris™ was shown to be significantly superior to placebo at controlling symptoms of schizophrenia. While this study allowed flexible dosing (Saphris™ 5-10 mg BID), this study supports the efficacy of Saphris™ 10 mg BID as the majority of patients were receiving this dose compared to the 5 mg  BID dose (80% vs. 20% respectively).

On the basis of all 5 pivotal studies, there is sufficient evidence to support the efficacy of Saphris™ in the treatment of schizophrenia. However, there was no evidence to suggest that a 10 mg BID dose was superior to a 5 mg BID dose. Thus, the recommended dose should be 5 mg BID. (Early Phase II studies failed to demonstrate efficacy at doses lower than 5 mg BID).

Efficacy in Manic or Mixed Episodes Associated with Bipolar I Disorder

The efficacy of Saphris™ in the treatment of manic or mixed episodes associated with bipolar I disorder was assessed in two similarly designed 3-week, Phase III studies (Studies A7501004, A7501005). The evaluation also included longer-term studies; a 9-week extension study (A7501006), and a 12-week co-therapy study (A7501008). Of these four studies, three were considered pivotal. Study A7501006 was considered non-pivotal.

Studies A7501004 and A7501005 were 3-week, randomized, double-blind, placebo-controlled and active-controlled (olanzapine), flexible-dose studies (Saphris™ 10 mg BID with an option to decrease to 5 mg BID) of identical design, in patients with acute manic or mixed episodes associated with bipolar I disorder. Patients were randomized in a 2:2:1 ratio for treatment with Saphris™, olanzapine, or placebo. Each of these two monotherapy studies randomized approximately 100 patients to placebo and 200 patients to Saphris™.

Study A7501006 was a 9-week extension study of patients who completed Studies A7501004 and A7501005 and who were on active treatment. The patients [number (n) = 397] could continue on the same active treatment [Saphris™ 5-10 mg BID or olanzapine 5-20 mg once daily (QD)] for an additional 9 weeks. The primary objective of Study A7501006 was to demonstrate non-inferiority acute efficacy of Saphris™ compared with olanzapine for up to 12 weeks. However, the design was not adequate to assess non-inferiority of two active treatments as the study only continued patients on the same drug that they had a good experience with in the previous studies, keeping them on the same active drug without further randomization. The study was not regarded as a pivotal study. However, the study did demonstrate the continuation of effect of Saphris™ during the additional 9 weeks.

Study A7501008 was a 12-week study conducted in patients who had an acute manic or mixed episode associated with bipolar I disorder, and who had been receiving lithium or valproate for at least 2 weeks prior to randomization. Patients were randomized (1:1) to Saphris™ or placebo, with lithium and/or valproate continued for the duration of the study. Treatment with Saphris™ started at 5 mg BID and thereafter the dose was flexible Saphris™ 5-10 mg BID. This combination study with lithium and/or valproate randomized approximately 106 patients to each group (i.e., Saphris™ or placebo).

The primary rating instrument used for assessing manic symptoms in the efficacy studies was the Young Mania Rating Scale (YMRS). Patients were also assessed on the Clinical Global Impression - Bipolar (CGI-BP) scale. The primary efficacy endpoint was the change from baseline to Day 21 on the YMRS total score based on LOCF analysis. For Study A7501006, the primary efficacy endpoint was as above, but was determined at Day 84 (12 weeks). Study A7501008 used the same primary efficacy endpoint at Day 21, and measurements continued up to Day 84.

Pivotal Study Results

Study A7501004 was conducted in the United States of America, Eastern Europe, India, and a few sites in south-east Asia. The study showed a positive outcome for Saphris™ versus placebo [probability (p) = 0.007], but the study was confounded by a significant treatment-by-centre interaction. Two large sites in India showed differences on YMRS of more than 25 points between Saphris™ and placebo. While these two sites constituted only 3% of the study population, their exclusion not only eliminated the significant treatment-by-site interaction, but also resulted in a non-significant study outcome (p = 0.1412).

Approximately 60% of the dataset was from the United States, 20% from Eastern Europe, and 20% from Asia. There were notable differences between regions in efficacy outcomes. In Asia, there was a marked and significant treatment effect for both active treatment groups (Saphris™ and olanzapine) compared to placebo, while in the much larger United States subgroup, or combined United States/Europe subgroup Saphris™ was no better than placebo, while olanzapine showed superiority over placebo. The treatment-by-centre interaction and the regional differences remain unexplained.

This study was considered equivocal due to the treatment-by-centre interaction on the primary outcome, the lack of any observable treatment superiority of Saphris™ over placebo in the United States/Europe (which contributed 80% of the study population), as well as the consistent superiority of olanzapine compared to placebo across efficacy outcomes, and the lack of such with Saphris™.

Study A7501005, in contrast to Study A7501004, showed a positive outcome for Saphris™ versus placebo (p<0.0001), in the absence of a significant treatment-by-centre interaction. While Study A7501005 did show regional differences similar to Study A7501004, the United States and combined United States/Europe subgroups in Study A7501005 showed significant treatment effects of Saphris™ over placebo, (although the effect was smaller than in the Asian data subset). This study is considered a positive study, in support of the efficacy of Saphris.

Studies A7501004 and A7501005 were identical in study protocols. When the data from these two studies were pooled together, Saphris™ showed statistical superiority over placebo on the primary endpoint not only for the entire study population, but also for the United States/Europe subgroup.

Study A7501008 showed a positive outcome for Saphris™ versus placebo at Week 3 and at Week 12. There was no substantial regional difference in treatment outcome patterns. This study supports the efficacy of Saphris™ in the treatment of mania, in the presence of ongoing treatment with valproate or lithium. However, it does not support a separate augmentation therapy indication for Saphris™. The study population was not clearly defined as “non-responder” or “partial responder” to valproate or lithium.

Given that two out of three of the pivotal studies showed positive results, and that the pooled results of the Europe and United States subpopulation from Studies A7501004 and A7501005 also showed superiority over placebo, it is considered that overall there is sufficient evidence to support efficacy of Saphris™ in the treatment of acute mania and mixed states associated with bipolar disorder, as acute monotherapy or co-therapy with lithium or divalproex sodium.

Saphris™ dosing recommendations with respect to bipolar disorder are somewhat problematic, as no fixed-dose studies were conducted. The monotherapy studies started patients on a 10 mg BID dose regime, while the combination therapy study started patients on a 5 mg BID dose regime of Saphris™. Both studies allowed for up and down titration during the study within the dose range of 5-10 mg BID. In the monotherapy studies, almost all patients remained on the initial 10 mg BID dose, while in the combination study approximately 55% of patients remained on 5 mg BID while approximately 45% had a dosage increase to 10 mg BID at some point during the study. These data are far from optimal with respect to making the most appropriate dosing recommendations for the use of Saphris™ in treatment of bipolar mania. Nevertheless, Saphris™, is reasonably well-tolerated at the 10 mg BID dose, and there is evidence of efficacy at this dose. Therefore, the wording in the Saphris™ Product Monograph recommends a dosage range within 5 mg to 10 mg BID, highlighting the limitation of the current supporting clinical data. Both the European Medicines Agency (EMA) and the United States Food and Drug Administration (FDA) have approved the use of Saphris™ in the treatment of acute mania based on the same clinical study data.

3.3.4 Clinical Safety

In the Phase II or Phase III studies, Saphris™ sublingual tablets were administered at doses of 5-10 mg BID to approximately 3,100 patients with schizophrenia or mania associated with bipolar I disorder. In these studies, approximately 1,700 patients were treated for at least 12 weeks, 1,300 patients for at least 6 months, and almost 800 patients for one year.

As of December 1, 2008, there were 19 deaths (0.5%) in all of the Saphris™ groups: 1 death (0.1%) in the placebo groups; 5 deaths (0.4%) in the olanzapine groups; and 1 death (0.9%) in the haloperidol groups. There were some deaths by suicide, and deaths from a variety of different medical conditions, with no obvious pattern. Furthermore, these clinical study fatalities did not reveal a specific association with Saphris™ treatment.

In the Phase II/III database the following serious adverse events of potential relevance to Saphris™ were observed: syncope (0.2%, n = 7); neuroleptic malignant syndrome (NMS) (0.1%, n = 3); hyponatraemia (0.1%, n = 3); and rhabdomyolysis (0.1%, n = 2). Syncope and NMS are observed with most antipsychotic medications. Hyponatraemia is associated with schizophrenia (i.e., secondary to polydipsia). Rhabdomyolysis was secondary to hyponatraemia and its sequalae or another primary cause.

Common adverse events (at least 2.0% and at least twice the incidence of placebo) associated with Saphris™ (5-10 mg BID) in the short-term placebo-controlled studies (6-week schizophrenia and 3-week bipolar mania) were:

• Sedation (9.1% Saphris™, 4.4% placebo);
• Somnolence (8.4% Saphris™, 2.3% placebo);
• Akathisia (5.4% Saphris™, 2.4% placebo);
• Oral hypoaesthesia (5.0% Saphris™, 0.7% placebo);
• Increased weight (3.5% Saphris™, 0.4% placebo);
• Dystonia (2.5% Saphris™, 0.6% placebo);
• Increased appetite (2.5% Saphris™, 0.7% placebo);
• Parkinsonism (2.4% Saphris™, 1.1% placebo).

The incidence of the common adverse events was similar for both bipolar and schizophrenia studies, except for a somewhat higher incidence of sedation, somnolence and dizziness in the bipolar studies.

In the short-term, schizophrenia studies where fixed doses of Saphris™ were studied, the following adverse events showed a dose-related trend:

• Akathisia (2.4% placebo, 0.7% Saphris™ <5 mg BID, 4.0% Saphris™ 5 mg BID, 10.6% Saphris™ 10 mg BID);
• Tremors (1.4% placebo, 1.7% Saphris™ <5 mg BID, 1.8% Saphris™ 5 mg BID, 4.3% Saphris™ 10 mg BID).

In the Phase II/III studies, the safety profile of Saphris™ was similar to other second-generation antipsychotic drugs that block dopamine 2 and 5HT2 receptors. The only unique adverse event associated with Saphris™ that was identified in the clinical study database was that of hypoesthesia. The pharmacological profile of Saphris™ does demonstrate some local anesthesia activity of the drug. This is only clinically apparent due to the route of administration i.e., sublingual (to avoid a very large first pass effect, which limits oral bioavailability).

In addition to clinical study safety data, there is a post-marketing safety database (as Saphris™ has been marketed for treatment of schizophrenia and bipolar disorder in the United States and Puerto Rico since October 2009, and most recently Germany since December 2010). The post-marketing data (estimated at 1,300 patient years of exposure based on sales) identified two other potential safety signals. As of the most recent periodic safety update (February 2011), there have been four cases of serious hypersensitivity reactions. All of these cases involved throat or tongue swelling. In reviewing the available details on these cases, it was not entirely clear what role numbness of the oropharynx (caused by the anesthetic properties of asenapine) and acute dystonic reactions of tongue and pharynx may have played in the symptoms experienced.

In addition to the above potential hypersensitivity reactions, there were also 4 potential cases of Stevens-Johnson syndrome (SJS). Based upon the available information, a causal relationship between Saphris™ and the onset of SJS cannot be determined. All 4 cases either contained confounding factors or contained information that does not fit well with a diagnosis of SJS (i.e., resolution while continuation of Saphris™ treatment). The Saphris™ Product Monograph contains warnings with respect to the possibility of hypersensitivity reactions associated with Saphris™.