CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`
`APPLICATION NUMBER:
`022036Orig1s000
`
`
`
`MEDICAL REVIEW(S)
`
`
`
`
`
`
`
`
`
`
`

`

`December 2, 2009
`
`Division Director
`Division of Neurology Products/HFD-120
`
`MEMORANDUM
`
`DATE:
`
`FROM:
`
`
`
`TO:
`
`SUBJECT: Action Memo for NDA 22-036, for the use of Silenor (doxepin HCl)
`in the treatment of insomnia
`
`NDA 22-036, for the use of Silenor (doxepin HCl) in the treatment of insomnia,
`was submitted by Somaxon Pharmaceuticals on January 7, 2008. The
`application was submitted as a 505(b)(2) application, relying on the approved
`applications for Sinequan (doxepin) capsules and Oral Concentrate, as well as
`Zonalon (doxepin) Cream. Sinequan is approved and has been marketed since
`1969 as an anti-depressant and anxiolytic at doses up to 300 mg/day (usual daily
`dose of 75-150 mg/day). Zonalon Cream is a topical preparation and is indicated
`in the treatment of pruritis.
`
`The initial application contained the results of 6 controlled trials. The Agency
`issued a Complete Response (CR) letter on 2/25/09; the primary reasons for this
`action were as follows:
`
`Effectiveness
`
`The division primarily considered the evidence purporting to establish substantial
`evidence of effectiveness for Silenor as a treatment for insomnia characterized
`by difficulty in maintaining sleep (there were no consistent positive findings on
`measures of sleep latency). However, we had concluded that there was
`inadequate subjective evidence of sleep maintenance (as assessed by the
`subjective [sWASO]) in non-elderly adults at the 6 mg dose. Specifically,
`although there was objective evidence of an effect (as measured by objective
`Wake Time After Sleep Onset [oWASO]) on sleep maintenance at days 15 and
`29 in non-elderly adults, there was no evidence of a beneficial effect on those
`nights on a subjective measure of sleep maintenance (sWASO) in this
`population, the protocol-specified primary nights at which a subjective response
`was to be measured. There were statistically significant drug-placebo
`differences on nights 16 and 30 on sWASO in this population at this dose, and on
`the mean of Nights 15 and 16 and 29 and 30. There were also significant
`findings on sWASO out to 2 months in elderly adults (in a separate study) at 6
`mg, but, as noted in the CR letter, we could not be certain that the effects seen
`on subjective measures at 6 mg in the elderly were applicable to non-elderly
`adults (possibly because of the higher plasma levels achieved in the elderly
`
`
`
`File, NDA 22-036
`
`
`
`1
`
`

`

`compared to the non-elderly at this dose, or perhaps related to increased
`sensitivity to drug effect in the elderly).
`
`Further, we noted that there were significant subjective findings on oWASO in the
`non-elderly population at 3 mg out to one month, but no significant findings on
`sWASO in this population after Night 1 (and no robust effect on sWASO in the
`elderly at this dose). Taken together, the division concluded that there was no
`clear effect on subjective measures of sleep maintenance at any dose in the non-
`elderly population.
`
`Safety
`
`The division concluded that there was evidence that Silenor might have been
`associated with a prolongation of the QT interval of between 5-10 msec. We
`were aware at the time we issued the CR letter that the sponsor had performed,
`or was in the process of performing, a thorough QT study, and in the letter we
`asked the sponsor to submit the results of this study.
`
`
`The sponsor responded to the CR letter with a complete response on 6/4/09.
`The response primarily consisted of additional statistical analyses performed in
`an effort to provide evidence that there were robust effects on subjective
`measures of sleep maintenance at a 6 mg dose in the non-elderly population.
`This submission has been reviewed by Dr. June Cai, medical officer, Dr. Abiola
`Olagundoye, SEALD, the Interdisciplinary Review Team for QT Studies, Dr.
`Tristan Massie, statistician, Jessica Diaz and Melissa Hulett, Division of Risk
`Management, and Dr. Ronald Farkas, neurology team leader. In this memo, I
`will very briefly review the relevant issues, and offer the rationale for the
`division’s action.
`
`As noted above, the sponsor has submitted the results of additional statistical
`analyses that they believe establish a reliable effect of Silenor 6 mg on sWASO.
`
`Specifically, as discussed by Dr. Massie, the sponsor asserts that the treatment
`by time interaction is not statistically significant for the 6 mg dose based on a
`Mixed Model Repeated Measures (MMRM) analysis, on the basis of which they
`conclude that the average treatment difference over the double-blind period can
`stand for the difference at the end of the study. On the basis of this new
`analysis, the sponsor obtains a significant drug-placebo difference. Based on the
`MMRM, differences between 6 mg and placebo at days 15 and 16 did not reach
`statistical significance nor did the 6 mg-placebo difference reach significance at
`Night 29 (see Dr. Massie’s Table 6), though the between-treatment contrasts for
`the average of each two night pair does reach nominal significance (see Dr.
`Massie’s Table 8).
`
`
`
`
`2
`
`

`

`However, according to Dr. Massie, the power of this test to detect an interaction
`is quite low (43%). For this reason, we cannot with confidence reject the
`hypothesis that there is no treatment by time interaction.
`
`For example, Dr. Massie notes that the p-value for the interaction test based on
`the first night of each visit is 0.14. Including all nights for each visit, the p-value
`for the test of the interaction between time and treatment is 0.27. However, for a
`test of 90% at the 0.05 significance level, the null hypothesis of no interaction
`would be rejected if the p-value for the interaction test was <0.54. For a test with
`80%, we would reject the hypothesis of no interaction with p<0.33.
`
`In addition, a simple inspection of the data suggests that the treatment effect is
`not constant over time. In this regard, see Dr. Massie’s Figure 1, which depicts
`the mean sWASO over time (at Nights 1 and 2, 15 and 16, and 29 and 30), and
`clearly documents the inconstant pattern of responses, especially at the end of
`the study. In fact, the difference in treatment effect between Nights 29 and 30 is
`statistically significant. This makes it difficult to reliably estimate the true
`treatment effect at the end of the study, making comparisons between this
`(unknown) treatment effect and estimates of the treatment effects at earlier
`timepoints unreliable.
`
`
`Further, as Dr. Massie notes, there were likely not sufficient assessments during
`the 30 days of the study to conclude that the treatment difference was constant
`at times between assessments.
`
`For these reasons, then, in his view, for an assessment of the drug effect at the
`end of the study, we must continue to rely on the data at that time point (that is,
`at Nights 29 and 30; again, the assessment at Night 29 was specified in the
`protocol as the primary assessment).
`
`In addition, the sponsor also applied an MMRM approach to subjective Total
`Sleep Time (sTST), their preferred subjective measure of sleep maintenance.
`Using this analysis, statistical significance was not achieved for either Night 29 or
`Night 30.
`
`The sponsor asserts that a pre-specified plan for performing the MMRM analysis
`was followed, though they acknowledge that this plan was proposed after the
`submission of the NDA (that is, after the data and results of the previous
`analyses were obviously known).
`
`Finally, Dr. Massie performed calculations to determine the potential size of the
`interaction that could not be excluded, with an eye to examining whether or not
`the difference in the size of any treatment effect among timepoints might be
`sufficiently small to be considered unimportant. As he notes, the findings on the
`MMRM performed by the sponsor are consistent with a treatment difference on
`Nights 15 or 29 of about 10 minutes less than on Night 1. This difference is
`
`
`
`3
`
`

`

`about 50% of the estimate of the treatment difference at Night 1, a difference that
`seems non-dismissible.
`
`Safety
`
`The sponsor has submitted the results of a thorough QT study examining
`doxepin doses of 6 and 50 mg. The QT Review Team has concluded that
`neither dose is associated with a meaningful increase in the QT interval.
`
`Conclusions
`
`The sponsor has submitted numerous additional analyses that purport to
`establish a consistent effect of a 6 mg dose of doxepin on subjective measures of
`sleep maintenance in the non-elderly population out to one month. The
`statistically significant between-treatment differences that the sponsor presents,
`however, are as the result of MMRM analyses performed after the original data
`were known and analyzed. Further, and importantly, the results are based on the
`presumption that there is a constant treatment effect over time, and that there is
`no treatment by time interaction. Although the sponsor’s formal test for such an
`interaction did not reach significance, Dr. Massie points out that the power to
`detect such a difference was very small (43%). Inspection of the data also
`suggests that the effect may not have been constant over time (and that there
`were likely not sufficient assessments over the 30 days of the study to permit a
`conclusion that the effects were constant over time). For these reasons, we
`cannot accept the sponsor’s assertions that the MMRM analyses are appropriate.
`As a result, I believe that we should rely on the original analyses on which we
`based our original decision.
`
` I
`
` note that Dr. Farkas continued to recommend that the application be approved.
`He bases this conclusion on his original reasoning, and he acknowledges that
`the sponsor has presented no new statistical arguments that persuasively
`counter the reasons for the initial CR action. In short, in his view, no meaningful
`change in the data package has occurred, and so his original conclusion still
`applies. I agree that the sponsor has provided no new arguments that
`adequately address our concerns, as articulated in the original CR letter, and
`which transmitted my decision to not approve the drug at that time. Although I
`note Dr. Farkas’s recommendation, I have not changed my original views, and,
`for this reason, will issue the attached CR letter.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Russell Katz, M.D.
`
`
`
`4
`
`

`

`Application
`Type/Number
`--------------------
`NDA-22036
`
`Submission
`Type/Number
`--------------------
`ORIG-1
`
`Submitter Name
`
`Product Name
`
`--------------------
`SOMAXON
`PHARMACEUTICA
`LS INC
`
`------------------------------------------
`SILENOR (DOXEPIN HCL)
`
`---------------------------------------------------------------------------------------------------------
`This is a representation of an electronic record that was signed
`electronically and this page is the manifestation of the electronic
`signature.
`---------------------------------------------------------------------------------------------------------
`/s/
`----------------------------------------------------
`
`RUSSELL G KATZ
`12/04/2009
`
`

`

`February 23, 2009
`
`Director
`Division of Neurology Products/HFD-120
`
`
`
`File, NDA 22-036
`
`MEMORANDUM
`
`DATE:
`
`FROM:
`
`
`
`TO:
`
`SUBJECT: Action Memo for NDA 22-036, for the use of Silenor (Doxepin HCl)
`in the treatment of insomnia
`
`NDA 22-036, for the use of Silenor (Doxepin HCl) in the treatment of insomnia,
`was submitted by Somaxon Pharmaceuticals on January 7, 2008. The
`application was submitted as a 505(b)(2) application, relying on the approved
`applications for Sinequan (doxepin) Capsules and Oral Concentrate, as well as
`Zonalon (doxepin) Cream. Sinequan is approved and has been marketed since
`1969 as an anti-depressant and anxiolytic at doses up to 300 mg/day (usual daily
`dose of 75-150 mg/day). Zonalon Cream is a topical preparation and is indicated
`in the treatment of pruritis.
`
`This application contains the results of 6 controlled trials, in which doses of 1, 3,
`and 6 mg/night were evaluated in various models (transient and chronic) of, and
`in several populations of patients (non-elderly and elderly adults) with, insomnia.
`In addition, safety data from these studies are presented.
`
`The application has been reviewed by Dr. June Cai, medical officer, Dr. Tristan
`Massie, statistician, Dr. Ju-Ping Lai, Office of Clinical Pharmacology, Drs. Houda
`Mahayni and Sherita McLamore, Office of New Drug Quality Assessment, Dr.
`Jinhee J. Lee, Division of Medication Error Prevention and Analysis, Dr.
`Katherine Bonson, Controlled Substance Staff, Dr. Antoine El-Hage, Division of
`Scientific Investigations, and Dr. Ron Farkas, Acting Neurology Team Leader. In
`this memo, I will briefly review the relevant effectiveness and safety data, and
`offer the rationale for the division’s action.
`
`Effectiveness
`
`As noted, the application contains the results of 6 controlled trials, as follows:
`
`Study 401
`
` A
`
` 4 period cross-over study in non-elderly adults evaluating placebo, doxepin 1,
`3, and 6 mg, each dose given for 2 nights. Efficacy was to be determined by the
`average Wake Time During Sleep (WTDS), measured by polysomnography
`(PSG) for the 2 nights.
`
`
`
`
`1
`
`

`

`Study 402
`
` A
`
` similar design as Study 401 in elderly adults.
`
`
`Study 501
`
` A
`
` parallel group study in non-elderly adults evaluating placebo, doxepin 3 and 6
`mg/night for 35 nights. The primary outcome was Waketime After Sleep Onset
`(WASO) as assessed by PSG.
`
`Study 502
`
` A
`
` one night, parallel group study utilizing an advance phase model of transient
`insomnia, evaluating placebo and doxepin 6 mg. The primary outcome was
`Latency to Persistent Sleep (LPS) as assessed by PSG.
`
`Study 503
`
` A
`
` parallel group study in elderly adults evaluating placebo, 1 and 3 mg/night for 3
`months. The primary outcome was WASO, assessed by PSG.
`
`Study 509
`
` A
`
` parallel group study in elderly adults evaluating placebo and 6 mg/night for 1
`month. The primary outcome was Total Sleep Time (TST), assessed
`subjectively by patients.
`
`
`All of the studies, except for Study 509, assessed various sleep-related
`parameters by both objective and subjective measures. The reviewers describe
`the times at which the primary measures in each study were to be assessed. In
`many of the trials, the primary time of assessment was to be at Night 1. This
`reflects the traditional view that hypnotics must be effective immediately, on Night
`1 or 2. However, current standards require that hypnotics be shown to be
`effective over time (for at least one month). Therefore, despite the protocol
`specification of Night 1 as the primary time of assessment of drug effect, in my
`view, the appropriate way to analyze these trials is to examine first the high dose
`in each study at the nominal study endpoint, and then to examine the drug’s
`effects at earlier time points (there were not significant numbers of
`discontinuations in these studies; this makes an assessment of drug effect out in
`time by the last observation carried forward (LOCF) method reasonable
`(although observed cases [OC] analyses were done as well). Although the
`reviewers discuss in some detail the protocol specified (and resultant post hoc)
`analyses, and possible (or perhaps traditionally considered necessary)
`adjustments for multiple comparisons, I believe it is appropriate to analyze the
`studies as I have described, without adjustments for multiple comparisons
`
`
`
`2
`
`

`

`(because even though the analyses I suggest should be done were mostly post
`hoc, they are the way all such studies are currently analyzed).
`
`Further, several of the studies evaluated LPS, a measure of the drug’s effects on
`patients’ difficulty in falling asleep, in addition to WASO or WTDS, measures of
`the drug’s effects on patients’ difficulties staying asleep. The results were also
`examined.
`
`The analyses demonstrate that there are no consistent effects on sleep latency
`beyond Night 1, and the effect on Night 1 is not entirely consistent across all
`studies (see, for example, Dr. Massie’s Table 52, page 72 of his review).
`
`With regard to doxepin’s effects on sleep maintenance, there was a consistent
`beneficial effect on Night 1 across studies. The following results were seen in
`Studies 501, 503, and 509, the studies that examined sleep maintenance beyond
`one or two nights.
`
`Study 501 (non-elderly adults).
`
`In this study, there were statistically significant differences favoring doxepin 6
`and 3 mg over placebo on WASO on Nights 29 (end) and 15, the nights the
`protocol specified as the nights on which the assessments were to be made.
`However, there were no statistically significant between-treatment differences for
`either night for either dose compared to placebo on subjective measures of sleep
`maintenance (subjective WASO). It should be noted that sleep assessments
`were done on two nights at each evaluation: Nights 15 and 16, and Nights 29
`and 30. Statistically significant differences between doxepin 6 mg and placebo
`were seen on sWASO on Nights 16 and 30, and for the average of Nights 15 and
`16, and for the average of Nights 29 and 30 (the protocol specified assessing
`WASO on the first night of each assessment).
`
`Study 503 (elderly adults)
`
`There were statistically significant differences between doxepin 3 mg and
`placebo in WASO on Nights 85, 57, 29, and 15. There was a statistically
`significant difference on WASO between 1 mg and placebo on Night 85, but not
`on Nights 57, 29, or 15. There were inconsistent statistically significant
`differences between doxepin 3 mg and placebo on sWASO (Nights 85 and Night
`29) and between doxepin 1 mg and placebo (Night 85 only).
`
`Study 509 (elderly adults)
`
`There were clear statistically significant differences favoring doxepin 6 mg over
`placebo at all time points on sWASO (Nights 57, 29, 15, and 1). There were no
`objective measures assessed.
`
`
`
`
`3
`
`

`

`Safety
`
`There were 966 unique subjects exposed to doxepin in this application. Drs. Cai
`and Farkas describe the adverse events seen with the use of doxepin at the
`doses proposed. As noted by the reviewers, there is some evidence that Silenor
`does cause some next day residual effects (and when taken with meals, the
`Tmax increases from about 3-4 hours to about 6-8 hours), and is associated with
`some other, not unexpected, adverse events. There are no adverse events,
`however, that would preclude approval of Silenor (recall that doxepin, in the form
`of Sinequan, is marketed as an anti-depressant and anxiolytic at doses up to 300
`mg/day).
`
`However, as described in some detail by Dr. Farkas, what data we do have is
`suggestive of a capacity of doxepin to prolong the QT interval at the doses
`evaluated in these studies.
`
`Specifically, the QT interval was prolonged in several studies submitted. The
`data displayed below are taken from Dr. Farkas’s review, pages 22-23.
`
`In Study 501, EKGs were evaluated at Baseline and in the morning of Day 38,
`2½ days after the last dose. The change from baseline in msec QT varied as
`follows:
`
`
`
`QTcF
`QTcB
`
`In Study 505, in which doxepin was administered with cimetidine, a non-specific
`CYP inhibitor that induces an approximately 2-fold increase in doxepin levels,
`EKGs were performed at Baseline and 96 hours after a single dose of doxepin 6
`mg. The results are given below in msec:
`
`
`
`QTcF
`QTcB
`
`In Study 503, EKGs were evaluated at Baseline and in the morning of the final
`study day, about 9 hours post-dose:
`
`
`
`QTcF
`QTcB
`
`
`Placebo
`
`3 mg 6 mg
`
`.9
`.1
`
`
`
`
`3.9
`4.2
`
`5.1
`6.6
`
`Baseline
`
`6 mg
`
`396
`406
`
`
`
`
`405
`416
`
`
`
`
`
`
`
`Placebo
`
`3 mg 6 mg
`
`1.4
`3.0
`
`
`
`
`4.9
`6.0
`
`6.3
`5.8
`
`
`
`4
`
`

`

`
`
`Placebo
`
`6 mg
`
`In Study 509, EKGs were evaluated at Baseline and on the Final Study Day :
`
`
`
`QTcF
`QTcB
`
`
`As Dr. Farkas also notes, in Study 506, which examined the interaction between
`doxepin 6 mg and sertraline, a moderate CYP 2D6 inhibitor, EKGs were obtained
`at baseline and on the final study day. There was no placebo, but the change
`from baseline in QT interval was about 8-9 msec (the Cmax of doxepin increased
`about 30% in the presence of sertraline).
`
`As further described by Dr. Farkas, analyses of outliers did not yield a consistent
`picture. There was an increase in the incidence of outliers with absolute QT
`intervals of >480 msec on drug compared to placebo (9/720 doxepin-treated
`subjects vs 3/560 placebo-treated patients), but there was a greater incidence of
`placebo-treated subjects compared to doxepin-treated patients who met outlier
`criteria as defined by an increase in QT interval of >60 msec.
`
`
`COMMENTS
`
`The sponsor has presented the results of 6 randomized controlled trials, 5 in
`patients with chronic insomnia, and one in healthy volunteers in a model of
`transient insomnia.
`
`The standard requirements for a demonstration of effectiveness for hypnotics is
`that the drug in question demonstrate an effect on both objective and subjective
`measures of some aspect of sleep disturbance (e.g., difficulty in falling asleep,
`difficulty in staying asleep). Typically, this is required to be shown in the same
`study, and, also typically, these effects are required to be shown in both non-
`elderly and elderly adults; these populations are typically evaluated in separate
`studies. Further, and importantly, hypnotics are generally expected to be
`effective on the first or second night of administration, but also in extended use,
`at least out to one month of dosing.
`
`The sponsor has performed 3 studies of at least one month in duration, 1 in non-
`elderly (one month), and 2 in elderly adults (2 months and 3 months).
`
`There has been considerable discussion in the various reviews about the
`appropriate statistical analyses of these studies, given that for most of them, the
`primary outcome was to be assessed at Night 1, and measures of sleep latency
`were to be assessed prior to measures of sleep maintenance at times after Night
`1. Given the usual statistical rules, if a particular outcome does not reach
`
`-6.7
`-5.5
`
`
`
`
`-2.5
`0.9
`
`
`
`5
`
`

`

`statistical significance at a given time point (for example, sleep latency),
`subsequent outcomes cannot be analyzed (for example, measures of sleep
`maintenance after Night 1). My view, however, is that the primary outcomes in
`essentially all of these studies were measures of sleep maintenance, and
`patients were required to have sleep maintenance difficulties to enroll in the
`studies. For this reason, I believe it is reasonable to inspect the results of
`analyses of the primary maintenance outcomes, independent of the results on
`the sleep latency measures (almost all of which do not reach statistical
`significance at any dose). Further, as I noted earlier, I believe it is reasonable to
`examine the results first at the latest time points, and then “work backwards” in
`time in evaluating the effects of doxepin on sleep maintenance. In addition, the
`effects of the highest dose in any study should be examined first, at all time
`points, and then the same should be done for the lower doses in any given study.
`The fact that this approach is largely post hoc in these studies is no bar to
`proceeding in this way; it is the way all modern studies of hypnotics are analyzed,
`and the fact that this is our choice, not the sponsor’s makes it, in my view,
`acceptable. Finally, when approached in this manner, I do not believe that
`corrections to the alpha are necessary.
`
`Given this position, the results can be briefly summarized.
`
`There is clear evidence of an effect of doxepin 6mg on objective measures of
`sleep maintenance out to one month in Study 501. However, the evidence of an
`effect of 6 mg doxepin nightly on subjective measures of sleep maintenance is
`somewhat less clear. We do not have clear evidence of such an effect in Study
`501, the only study that examined both objective and subjective effects of 6 mg
`nightly. Specifically, there were no statistically significant differences between 6
`mg and placebo in Study 501 on nights 15 and 29. However, significant
`differences were seen on Nights 16 and 30, and on the average of Nights 15 and
`16 and Nights 29 and 30. Further, there were significant differences between 6
`mg and placebo at all time points (out to 2 months) on subjective measures in the
`elderly in Study 509. I agree with Dr. Farkas that these results, taken as a whole,
`suggest that doxepin 6 mg given nightly, is effective in the treatment of sleep
`maintenance difficulties, but I also note that this seems to be somewhat less
`compelling data than we would typically have for most hypnotics. In this regard, I
`note that the clear subjective findings in the elderly could possibly be the result of
`the slightly higher plasma levels of doxepin seen in the elderly (although we do
`not have completely adequate data to establish this difference) and/or an
`increased sensitivity of elderly patients to a given plasma level/dose of doxepin.
`
`There is clear evidence of an effect of doxepin 3 mg on objective measures of
`sleep maintenance out to 1 month in non-elderly adults (Study 501) and out to 3
`months in the elderly (Study 503). However, there is no evidence of a subjective
`benefit of 3 mg in non-elderly adults, and a very inconsistent effect out to 3
`months in the elderly. For these reasons, I do not believe that the sponsor has
`demonstrated an adequate effect of a dose of 3 mg dose of doxepin.
`
`
`
`6
`
`

`

`
`Finally, although there are nominally statistically significant treatment differences
`between 1 mg and placebo at 3 months on both objective and subjective
`measures of sleep maintenance in the elderly, these effects are inconsistent and
`do not establish 1 mg as an effective dose.
`
`For these reasons, I do not believe that the sponsor has established substantial
`evidence of effectiveness for doxepin as a hypnotic for patients with sleep
`maintenance difficulties, at any dose. Of course, the data are suggestive at 6 mg
`nightly, and the sponsor should be asked to make the case that this, or any other
`dose, is effective.
`
` do not believe that the statistically significant findings at Night 1 on both
`objective and subjective measures of sleep maintenance for both 3 and 6 mgs is
`adequate evidence to support approval either as an initial dose, or as a dose for
`one night of treatment. As I noted above, hypnotics are required to be effective
`for at least one month, even though patients may not take a hypnotic every night
`(that is, even though chronic intermittent use may “mimic” a series of repeated
`single night uses).
`
`With regard to safety, as I noted earlier, no adverse effects were noted that
`would preclude approval. However, the suggestion that doxepin may prolong the
`QT interval is of concern.
`
`As noted earlier, the estimate of the degree of QT prolongation seen in several
`studies varied, but ranged from a difference between drug and placebo in the
`change from baseline from 3-4 msec to up to 10 msec. What makes these
`changes of potential concern is that EKGs were obtained long (sometimes days)
`after Tmax. Whether or not the changes are, in fact, drug-related, is certainly
`open to question, but an increase of some degree does appear to be consistent,
`and dose related. It is unclear why such an effect should occur in some cases
`several days after a last dose of doxepin; one possible explanation might be that
`the changes are due to a metabolite (the nordoxepine metabolite has a T1/2 of
`about 30 hours). However, at this time, the explanation for the finding is obscure.
`Nonetheless, the finding appears to exist (even though, again, the finding could
`be spurious, given the vagaries of the way the data were collected, and
`especially that some studies did not employ placebo), and cannot, in view my, be
`dismissed without further explanation. Doxepin is known to be associated with
`cases of torsades de pointes, but, as noted earlier, there is a long marketing
`history of doxepin at much higher doses than those studied here without an
`overwhelming signal attributable to QT prolongation and its consequences.
`Nonetheless, data on the QT prolonging effects of any dose of doxepin have not
`been submitted. We have, however, commented on a protocol for a thorough QT
`study for doxepin fairly recently. However, we do not know if that study has been
`done, and certainly no data from this study have been submitted to this
`application.
`
` I
`
`
`
`7
`
`

`

`
`In my view, before this application can be approved, the sponsor must
`adequately address our concerns about the potential for doxepin to prolong the
`QT interval to a clinically meaningful degree.
`
`For the reasons stated above, therefore, I will issue a Complete Response letter
`in which we will ask the sponsor to address our concerns about both
`effectiveness and safety, as described above.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Russell Katz, M.D.
`
`
`
`8
`
`

`

`---------------------------------------------------------------------------------------------------------------------
`This is a representation of an electronic record that was signed electronically and
`this page is the manifestation of the electronic signature.
`---------------------------------------------------------------------------------------------------------------------
` /s/
`---------------------
`Russell Katz
`2/25/2009 08:42:51 AM
`MEDICAL OFFICER
`
`

`

`Ronald Farkas MD, PhD
`NDA 22036, CDTL Review
`
`
`Cross Discipline Team Leader Review
`
`
` Date
` From
` Subject
` NDA
` Applicant
` Date of Submission
` PDUFA Goal Date
`Proprietary Name /
`Established (USAN) names
`Dosase forms / Strength
`Proposed Indication(s)
`Recomended:
`
` Feb 12, 2009
` Ronald Farkas, MD, PhD
` Cross-Discipline Team Leader Review
` 22-036
`Samaxon Pharmaceuticals
` January 7, 2008
`February 27, 2009
`Silenor / Doxepin HCl
`
` 1 mg, 3 mg, 6 mg tablets
` Treatment of Insomnia
`Complete response
`
`Introduction
`
`
`
`
`1.
`
`Silenor (doxepin HCl) is being developed by Somaxon Pharmaceuticals under section 505(b)(2) for the
`treatment of insomnia in adult (18-64 years old) and elderly (65 years of age or older) patients. To
`support the application, the sponsor is referencing safety and efficacy information FDA relied on for
`approval of NDA 016-798 (Sinequan® Capsules), NDA 017-516 (Sinequan® Oral Concentrate), NDA
`020-126 (Zonalon® 5% Cream), published literature, and data generated by the sponsor. Doxepin has
`been marketed in the U.S. by Pfizer since 1969. Oral doxepin as Sinequan®
` is indicated for depression
`and anxiety. Topical doxepin as Zolalon®
` is indicated for treatment of pruritis.
`
`
`2. Background
`
`Doxepin is a tricyclic antidepressant with sedating effects. While doxepin binds to a number of CNS
`targets at the doses used for anxiety and depression, at low doses the sponsor believes that doxepin
`mainly antagonizes histaminergic H1 receptors, thereby inducing drowsiness and sleep, similar to the
`mechanism of currently approved over-the-counter antihistamine sleep aids.
`
`Sinequan labeling indicates that the usual dose range of doxepin for depression or anxiety is 75- to 150
`mg/day, up to 300 mg/day. These doses are roughly 10- to 100-fold higher than doses of doxepin
`studied in the present application for sleep: 1 mg and 3 mg in elderly subjects, and 3 mg and 6 mg in
`adults.
`
`The sponsor’s development program attempted to demonstrate the efficacy and safety of Doxepin for
`both sleep onset and sleep maintenance endpoints. Doxepin seemed particularly promising for sleep
`maintenance, an aspect of insomnia in which new treatment options are needed. The sponsor notes that
`
`
`
`1
`
`

`

`Ronald Farkas MD, PhD
`NDA 22036, CDTL Review
`
`many drugs currently approved or used off-label for treatment of insomnia are not effective in
`promoting sleep maintenance, and that some drugs that are effective in sleep maintenance (e.g. longer-
`acting benzodiazepines) are associated with undesirable effects including next-day sedation and the risk
`of tolerance and dependence.
`
`3. CMC
`
`Dr. Sherita McLamore was the primary reviewer, and Dr. Ramjesh Sood was the secondary reviewer. Both
`recommend the approval of Silenor under the conditions specified in the package insert, with no
`recommendation for phase 4 commitments, agreements, or risk management steps.
`
`
`4.