Summary Basis of Decision for Cymbalta ™

3.3.1 Pharmacodynamics

Duloxetine hydrochloride was shown to be a selective serotonin and norepinephrine re-uptake inhibitor in various models. It increased extracellular concentrations of both norepinephrine and serotonin within the CNS, thereby contributing to the descending pain inhibitory pathway and alleviating persistent pain. In healthy subjects, duloxetine hydrochloride significantly interfered with the serotonin re-uptake process as demonstrated by marked decreases in blood serotonin concentration, but showed no clear effect on the norepinephrine re-uptake, as measured by the tyramine stressor test. In additional clinical studies, duloxetine hydrochloride decreased 5-hydroxytryptamine concentrations in the blood and decreased the urinary excretion of norepinephrine and its metabolites, consistent with the blockage of serotonin and norepinephrine uptake processes, respectively. Duloxetine hydrochloride did not appear to have significant affinity for a variety of other receptors such as dopaminergic, cholinergic, adrenergic, or histaminergic type.

3.3.2 Pharmacokinetics

Absorption

In humans, duloxetine hydrochloride is well absorbed, with its maximal plasma concentrations (C_max) occurring 6 hours post-dose. Although food does not affect the C_max of duloxetine, it delays the time to reach peak concentration from 6 to 10 hours. Duloxetine plasma exposure increases in proportion to dose for doses up to 60 mg twice a day. Steady-state plasma concentrations are typically achieved after 3 days of dosing. There is no clinically important difference in the pharmacokinetic parameters of morning and evening doses.

Distribution

Duloxetine hydrochloride was highly bound (>90%) to plasma proteins. The mean volume of distribution was 1640 L. Plasma protein binding of duloxetine is not affected by renal or hepatic impairment.

Metabolism

Duloxetine hydrochloride is extensively metabolized in the liver by two cytochrome P450 enzymes, primarily by CYP1A2 and to a lesser extent CYP2D6. The two major circulating metabolites are not pharmacologically active.

Excretion

The elimination half-life of duloxetine hydrochloride was 8-17 hours (mean: 12 hours). Apparent plasma clearance of duloxetine does not differ between once daily and twice daily dosing. The majority of metabolites were excreted in the urine (72%) and feces (~19%).

Special Populations

Gender and Age
Mean duloxetine exposure were higher in female patients compared to males.

Compared to younger adults, elderly patients had increased levels of drug exposure and the half-life of the drug was longer in elderly females (vs middle-aged females). These differences in drug exposure did not appear to be clinically significant. Therefore, dosage adjustment based on gender or age is not necessary. Greater sensitivity may occur in older individuals as the elderly routinely take other concomitant medications and the likelihood of drug interactions would be greater. Safety and efficacy of duloxetine in pediatric patients have not been established.

Hepatic Impairment
Cymbalta is contraindicated in patients with any liver disease resulting in hepatic impairment. A pharmacokinetic study showed that single non-therapeutic dose of 20 mg duloxetine in six cirrhotic patients with moderate liver impairment (Child-Pugh Class B) had a mean plasma duloxetine clearance approximately 15% that of age- and gender-matched healthy subjects, with a 5-fold increase in mean exposure, and a half-life approximately 3-times longer.

Renal Impairment
Compared to subjects with normal renal function, duloxetine Cmax and drug exposure (AUC) values were approximately 2-fold higher in patients with end stage renal disease receiving chronic intermittent hemodialysis. Thus, Cymbalta is not recommended for patients with severe renal impairment.

Patients with Substantial Alcohol Use
Use of duloxetine in patients who consume substantial amounts of alcohol may be associated with severe liver injury. Cymbalta should only be used in exceptional circumstances and with extreme caution in these patients.

Drug Interactions

Duloxetine hydrochloride is predominantly metabolized by CYP1A2 and to a lesser extent by CYP2D6. Duloxetine should not be used concomitantly with potent inhibitors of CYP1A2 enzyme.

Duloxetine is a moderate inhibitor of the enzyme CYP2D6 and caution should be used when it is co-administered with drugs that are metabolized by CYP2D6.

Potential drug interactions are included in the Cymbalta Product Monograph.

3.3.3 Clinical Efficacy

Clinical studies were submitted for the evaluation of efficacy in major depressive disorder (MDD) and for pain associated with diabetic peripheral neuropathy (DPN).

MDD Studies

The efficacy of Cymbalta in MDD was studied in six randomized, double-blind, placebo-controlled, fixed-dose studies in adult outpatients (range:18 to 83 years, mean: 41 years) meeting the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria for MDD, with a 17-Item Hamilton Depression Rating Scale (HAMD-17) ≥15 and a Clinical Global Impression Severity Score ≥4 at study entry. The studies were 8 or 9 weeks in duration; two of the studies had a 26-week continuation phase. Doses tested were 20 mg, 40 mg and 60 mg, all twice daily (BID) and 60 mg once daily (QD). The primary efficacy endpoint was the HAMD-17 total score. The secondary efficacy endpoints included: HAMD-17 Depressed Mood Item, the HAMD-17 Core and Anxiety subscales; Global Impression Scales (Patient Global Impressions (PGI) Improvement and CGI-Severity); and the Quality of Life in Depression scale.

In four of the six studies, Cymbalta demonstrated statistical superiority over placebo as measured by improvement in the HAMD-17 total score. The secondary efficacy endpoints were supportive of the primary efficacy endpoints. In comparison with placebo, Cymbalta at doses of 60 mg QD, and both 40 mg and 60 mg BID, produced a significantly higher rate of response and remission as defined respectively by ≥50% decrease in the HAMD-17 total score and a total endpoint HAMD-17 score of ≤7. There was no evidence that doses greater than 60 mg QD of Cymbalta conferred any additional benefit.

DPN Studies

The efficacy of Cymbalta in the treatment of neuropathic pain associated with Diabetic Peripheral Neuropathy (DPN) was studied in three placebo-controlled, fixed-dose trials of 12 weeks in duration in adult outpatients (age range: 20-84 years, mean: 61 years), having neuropathic pain associated with DPN for at least 6 months. Patients had to have a baseline pain score of ≥4 on the 11-point Likert scale at study entry. Patients meeting diagnostic criteria for major depression were excluded. Patients treated with Cymbalta were also permitted to have most other therapies, including non-medicinal treatments but antidepressants, anticonvulsants, and antipsychotics were excluded. The primary efficacy endpoint was the weekly mean of the 24-hour average pain intensity/severity scores based on the 11-point Likert scale. Secondary endpoints included response/responder rates (% of patients who achieved ≥30% reduction in mean pain score at the end of the study vs baseline), 24-hour worst pain scores, night pain scores, and Patient's Global Impression of Improvement scores.

In all three studies, Cymbalta 60 mg QD and 60 mg BID were statistically significantly superior in the primary efficacy endpoint when compared to placebo. The effects were apparent after one week of treatment and lasted throughout the 12-week period of each study. In all three studies, Cymbalta 60 mg QD and 60 mg BID demonstrated statistically significantly better responder rates compared to placebo (63-70% vs. 42-47%). Most secondary efficacy endpoints were also positive. The Cymbalta dose of 20 mg QD is non-therapeutic.

3.3.4 Clinical Safety

The clinical safety database included data from the short-term controlled studies referred to in Section 3.3.3 Clinical Efficacy, as well as data from longer-term or extension studies (two MDD 6-month extension studies, a one-year open-label MDD study, and five DPN extension studies ranging from 24-52 weeks in duration) and the most recent Periodic Safety Update Report (PSUR), dated August 2007, which included all clinical trial and post-market safety data to date.

MDD Studies

Approximately 10% of the 1139 MDD patients who received Cymbalta in the short-term placebo-controlled studies discontinued treatment due to an adverse event (AE), compared with 4% of the 777 patients receiving placebo. Nausea (Cymbalta 1.4%, placebo 0.1%) was the only common AE reported as reason for discontinuation and was considered to be drug-related (i.e., discontinuation in at least 1% of the Cymbalta-treated patients and at a rate of at least twice that of placebo). The most commonly observed AEs (an incidence 5% or greater and at least twice the incidence in placebo patients) in Cymbalta-treated MDD patients were: nausea, dry mouth, constipation, decreased appetite, fatigue, somnolence, and increased sweating.

Of a total of 2418 patients in the MDD clinical studies, 993 Cymbalta-treated patients were exposed for at least 180 days and 445 Cymbalta-treated patients were exposed for at least one year. The safety profile of Cymbalta in the longer-term open-label studies appeared similar to that seen in the controlled studies.

DPN Studies

Approximately 12% of the 800 DPN patients who received Cymbalta in the DPN placebo-controlled studies discontinued treatment due to an adverse event, compared with 5% of the 339 patients that received placebo. Nausea (Cymbalta 3.0%, placebo 0.3%), dizziness (Cymbalta 1.1%, placebo 0.3%), and somnolence (Cymbalta 1.2%, placebo 0%) were the common AEs reported as reasons for discontinuation and mostly considered to be drug-related. The most commonly observed AEs in Cymbalta-treated DPN patients were: nausea, constipation, dry mouth, vomiting, fatigue, decreased appetite, somnolence, erectile dysfunction, and sweating. The 60 mg BID dose was not as well-tolerated and was associated with a higher frequency of non-serious AEs.

The safety profile of Cymbalta in the longer-term open-label studies appeared to be similar to that seen in the controlled studies, with one exception: in the one-year study, three patients treated with 60 mg BID experienced hyperglycemia vs. none in the group that received routine care. Of a total of 1429 patients in the DPN clinical studies, 881 had ≥ 6 months of exposure to Cymbalta, and 515 had >12 months of exposure.

MDD and DPN Studies

Discontinuation symptoms have been systematically evaluated in all Cymbalta studies. Following abrupt or tapered discontinuation in MDD and DPN placebo-controlled clinical trials, the following symptoms occurred at a rate ≥ 1% and at a significantly higher rate in Cymbalta-treated patients compared with those discontinuing from placebo: dizziness, nausea, headache, paresthesia, vomiting, irritability, nightmare, fatigue, insomnia, diarrhea, anxiety, sweating, and vertigo. These are consistent with AEs following discontinuation of other serotonin-norepinephrine re-uptake inhibitors (SNRIs) and selective serotonin reuptake inhibitors (SSRIs).

Safety Topics of Interest

Hepatic Toxicity
Cymbalta increases the risk of elevation of serum aminotransferase levels, which led to trial discontinuation in 0.3% of duloxetine-treated patients. The median time to detection of such elevations in patients was approximately 2 months. In these patients, these were usually transient and self-limiting with continued use, or resolved upon discontinuation of Cymbalta. In the full cohort of placebo-controlled studies in any indication, elevation of alanine aminotransferase (ALT) >3 times the upper limit of normal occurred in 1.1% (75/6871) of Cymbalta-treated patients compared to 0.3% (13/5036) of placebo-treated patients. In placebo-controlled studies using a fixed-dose design, there was evidence of a dose-response relationship for ALT and aspartate aminotransferase (AST) elevation of >3 times the upper limit of normal and >5 times the upper limit of normal, respectively.

Post-marketing reports indicate that severe elevations of liver enzymes (>10 times the upper limit of normal) or liver injury with a cholestatic or mixed pattern have been rarely reported, in some cases, associated with excessive alcohol use or pre-existing liver disease. Because it is possible that duloxetine and alcohol may interact to cause liver injury or that duloxetine may aggrevate pre-existing liver disease, Cymbalta should ordinarily not be prescribed to patients with substantial alcohol use and should not be used in patients with any liver disease resulting in hepatic impairment. Furthermore, this drug should be used with caution in patients treated with other drugs associated with hepatic injury. Warnings regarding the possible liver-related effects of Cymbalta have been included in the Cymbalta Product Monograph.

Glucose Regulation
Data from DPN trials indicate that duloxetine treatment worsens glycemic control in some diabetic patients. In the 12-week acute treatment phase of these studies, Cymbalta was associated with a small increase in mean fasting blood glucose as compared to placebo. In the extension phase of these studies (up to 52 weeks), mean fasting blood glucose increased by 0.67 mmol/L (12 mg/dL) in the Cymbalta group and decreased by 0.64 mmol/L (11.5 mg/dL) in the routine care group. This difference was statistically significant. HbA1c increased by 0.5% in the Cymbalta group and by 0.2% in the routine care groups.

Bleeding
Of a total of 619 bleeding cases in the clinical studies (all indications), 33 cases were considered serious. Of these, 7 were coded as gastrointestinal bleeding. In the MDD clinical studies, duloxetine-treated patients had a significantly higher incidence of adverse events "possibly" indicative of bleeding vs. patients that received placebo (most common AE: contusion). Patients in DPN trials reported the highest percentage of cases "possibly" indicative of bleeding (5%) but the majority of the cases appeared to be due to injection site bruising or contusion and similar to placebo rates.

The post-marketing database (up to August 2007) contains a total of 246 adverse reactions possibly related to bleeding with at least 39 considered serious. The cumulative number of cases categorized under the cluster of Gastrointestinal Bleeding was 114 with 17 cases listed in the most recent PSUR (dated August 2007). Where the indication was stated, over 50% of the patients were taking Cymbalta for MDD and about 3% for pain associated with DPN.

While a causal relationship to Cymbalta is unclear, impaired platelet aggregation may result from platelet serotonin depletion and contribute to bleeding occurrences. Thus, caution is advised in patients taking anticoagulants and/or medicinal products known to affect platelet function (e.g. non-steroidal anti-inflammatorr drugs and ASA), and in patients with known tendency for bleeding or those with predisposing conditions. The Product Monograph for Cymbalta contains warnings with regards to abnormal bleeding including references to gastrointestinal bleeding and hematochezia.