`CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`
`761105Orig1s000
`
`
`OTHER REVIEW(S)
`
`

`

`
`Department of Health and Human Services
`Food and Drug Administration
`Center for Drug Evaluation and Research | Office of Surveillance and Epidemiology (OSE)
`
`Epidemiology: ARIA Sufficiency Memo
`Version: 2018-01-24
`
`Date:
`Reviewer:
`
` Team Leader:
`
`Division Deputy Director:
`
`Subject:
`
`
`Drug Name:
`Application Type/Number:
`Applicant/sponsor:
`OSE RCM #:
`
`
`
`
`
`
`
`April 19, 2019
`Michelle R. Iannacone, PhD, MPH
`Division of Epidemiology I
`Patricia Bright, PhD, MSPH
`Division of Epidemiology I
`Sukhminder K. Sandhu, PhD, MPH, MS
`Division of Epidemiology I
`Active Risk Identification and Assessment (ARIA) Sufficiency Memo:
`Theoretical malignancy risk associated with risankizumab treatment
`in psoriasis patients
`Risankizumab
`BLA 761105 / IND
`
`AbbVie
`2019-679
`
`
`
`
`
`Reference ID: 4421562
`
`Page 1 of 11
`
`(b) (4)
`
`

`

`
`
`
`
`Memo type
`
`Source of safety concern
`
`EXECUTIVE SUMMARY (place “X” in appropriate boxes)
`-Initial
`-Interim
`-Final
`-Peri-approval
`-Post-approval
`-Yes
`-No
`-Surveillance or Study Population
`-Exposure
`-Outcome(s) of Interest
`-Covariate(s) of Interest
`-Surveillance Design/Analytic Tools
`
`
`Is ARIA sufficient to help characterize the safety concern?
`
`If “No”, please identify the area(s) of concern.
`
`
`
`
`
`Reference ID: 4421562
`
`
`
`
`
`
`
`
`
`X
`
`
`X
`
`
`Long-term
`Short-term
`All Malignancies
`Lymphoma
`
`
`X
`X
`
`For long-term malignancy:
`
`
`
`
`X
`
`
`
`
`
`
`Page 2 of 11
`
`

`

`
`
`
`
`
`A. General ARIA Sufficiency Template
`
`
`
`1. BACKGROUND INFORMATION
`
`1.1. Medical Product
`Psoriasis is a chronic debilitating immunologic disease characterized by marked inflammation
`and thickening of the epidermis that result in thick, scaly plaques involving the skin. Psoriasis
`may undergo intermittent improvements and relapses in susceptible individuals over the
`course of their lifetime. Although traditional systemic therapies for psoriasis are effective,
`there may be a loss of efficacy during long-term use or patients may experience adverse events
`related to specific treatments.a
` The prevalence of psoriasis in the United States is approximately 2-4%, of which an estimated
`20% have moderate-to-severe disease. Psoriasis can first appear at any age, but more
`commonly appears in adulthood. Two peaks in age of onset have been reported: one at 20-30
`years of age and a second peak at 50-60 years of age.b
` Skyrizi (risankizumab) injection, for subcutaneous use, is indicated for the treatment of adults
`with moderate to severe plaque psoriasis who are candidates for systemic therapy or
`phototherapy. Risankizumab is a humanized immunoglobulin GI (IgG1) monoclonal antibody
`that is specifically directed against IL-23 p19. The framework of the risankizumab antibody
`has been engineered with two mutations in the Fc region to reduce Fcγ receptor and
`complement binding. Binding of risankizumab to IL-23 p19 inhibits the action of IL-23 to
`induce and sustain T helper (Th) 17 type cells, innate lymphoid cells, γδT cells, and natural
`killer (NK) cells responsible for tissue inflammation, destruction, and aberrant tissue repair.c
` The recommended dose of risankizumab is 150mg (two 75mg injections) administered by
`subcutaneous injection at Week 0, Week 4, and every 12 weeks thereafter.
`
`d
`
`Similar to other psoriasis biologics (Table 1), risankisumab poses a theoretical increased risk
`for malignancies based on its immunosuppressive mechanism of action.
`
`
`1.2. Describe the Safety Concern
`
`
`
`
`
`
`a Vide J, Magina S. Moderate to severe psoriasis treatment challenges through the era of biological drugs. An Bras
`Dermatol. 2017; 92(5):668-674.
`b BLA 761105 Multi-disciplinary Review and Evaluation, Skyrizi (risankizumab). Version date: February 1, 2019.
`c Sponsor Original Submission, GlobalSubmit Review: Upload dated April 23, 2018, Risankizumab, Clinical
`Overview.
`d Risankizumab Provider Information Label. DARRTS ID: Pending.
`
`
`
`Reference ID: 4421562
`
`Page 3 of 11
`
`(b) (4)
`
`

`

`Approval date
`for plaque
`psoriasis
`
`
`Table 1. Psoriasis biologics currently marketed in the United States
`
`
`Approved
`Postmarketing
`
`
`for plaque
`requirement for
`Drug
`Class
`psoriasis?
`malignancy?
`
`September 25,
`Yes
`Yes
`Interleukin-12 and -23
`Stelara
`2009
`antagonists
`(ustekinumab)
`January 21, 2015
`Yes
`Yes
`Interleukin-17A
`Cosentyx
`antagonist
`(secukinumab)
`March 22, 2016
`Yes
`Yes
`Interleukin-17A
`Taltz
`antagonist
`(ixekizumab)
`February 15,
`Yes
`Yes
`Interleukin-17 receptor
`Siliq
`2017
`A (IL-17RA) antagonist
`(brodalumab)
`July 13, 2017
`Yes
`Yes
`Interleukin-23 blocker
`Tremfya
`(guselkumab)
`March 20, 2018
`Yes
`Yes
`Interleukin-23 blocker
`Ilumya
`(tildrakizumab)
` For the overall risankizumab drug development program, a total of 21 malignancies (excluding
`non-melanoma skin cancer) were reported in the risankizumab exposed group, which
`corresponds to a rate of 0.62 events/100 person-years. Of these, malignancies reported for
`more than one subject included breast cancer reported in seven subjects, prostate cancer in
`three subjects, and malignant melanoma in two subjects. This observation is consistent with
`the most common cancers seen in the United States (breast cancer is the most common,
`followed by lung and prostate cancers, and the incidence of melanoma of the skin has been
`rising).b
` For the active comparator groups, one case of malignancy for gallbladder cancer was reported
`for adalimumab and one case of malignancy was reported for prostate cancer for ustekinumab.
`Further, the rates of malignant tumors (excluding non-melanoma skin cancer) ranged from
`0.31 – 0.49 events/100 person-years in the clinical development programs for ustekinumab,
`ixekizumab, secukinumab, and guselkumab. Although the event rate for malignancy is slightly
`higher in risankizumab users compared to the malignancy rates observed in the development
`programs for other biologics, there was only one death from malignancy in the risankizumab
`development program.b
` In the risankizumab development program, 25 non-melanoma skin cancer malignancies were
`reported; 10 events of Bowen’s disease/squamous cell carcinoma (SCC) combined and 15
`events of basal cell carcinoma (BCC). The observed ratio of SCC to BCC was 1:1.5. While this
`ratio is narrower than that seen in the immunocompetent general population, it is not inverted
`due to an increase in SCC as is observed in immunosuppressive populations (e.g., organ
`transplant recipients). This suggests that risankizumab has less of an immunosuppressive
`affect than observed in organ transplant patients.b
` The BLA Unireview concluded that the limited duration of observation during the clinical
`development program did not allow for detection of rare events with a long latency period
`such as that required by malignancy events.b Therefore, postmarketing data are needed to
`evaluate the long-term risk of malignancy in patients with psoriasis receiving risankizumab.
` The clinical evaluation of risankizumab had some notable parallels to the clinical evaluation of
`
`
`
`Reference ID: 4421562
`
`Page 4 of 11
`
`

`

`
`
`
`
`
`Assess a known serious risk
`
`Assess signals of serious risk
`Identify unexpected serious risk when available data indicate potential for serious risk X
`
`both guselkumabe and tildrakizumabf, including the following:
` “DDDP Clinical does not consider these clinical data to be a safety signal. The type of risk is
`considered to be a theoretical risk, where biological plausibility exists, yet clinical data are
`limited and not sufficient to support this suspicion of risk. DDDP described the safety concern
`as a variable-onset, where certain cancers may occur short-term, but there may also be a long-
`latency effect after initial exposure. The level of concern is moderate, taking into account that
`malignancy is a very serious adverse event, but the concern is largely theoretical. DDDP was
`also specifically interested in assessing the risk of lymphomas, which may have a shorter
`latency compared to other malignancies. DDDP hypothesized that the risk of lymphoma could
`be related to exposure with risankizumab.”
` No carcinogenicity and mutagenicity studies have been conducted with risankizumab.
` The patient information label does not include any warnings or precautions related to the
`potential malignancy risk. Further, the review team decided that a Risk Evaluation and
`Mitigation Strategy (REMS) is not needed.
`
`1.3. FDAAA Purpose (per Section 505(o)(3)(B))
`
`1.4. Statement of Purpose
`This memo reflects the discussions, recommendations, and determinations between the
`Division of Epidemiology I (DEPI-I), the Division of Dermatology and Dental Products (DDDP),
`and CDER’s Sentinel Team. To better assess malignancy risk, the team considered whether
`ARIA was sufficient or whether to issue a PMR for an observational study to collect additional
`data on long-term safety and evaluate the occurrence of long-latency safety outcomes.
` The purpose of this memo is to describe the determination of whether ARIA could be used to
`assess malignancy risk and lymphoma risk when clinical data could not confirm a safety signal,
`but theoretical concerns indicate the potential for a serious risk. The regulatory goal of ARIA is
`signal detection (i.e. postmarketing surveillance). The anticipated regulatory impact is to
`further characterize malignancy risk to inform labeling decisions. Because the events of
`interest are rare, typically have long-term latency periods (except for lymphoma), and because
`multiple products are available for treatment of the underlying disease (plaque psoriasis), the
`sufficiency determination primarily rests upon the need for a large sample size, the availability
`of long-term follow-up (except for lymphoma), the availability of relevant covariates, and on
`the ensuing market uptake of risankizumab.
`
`The postmarket uptake of risankizumab will in part influence the ARIA approach for the
`malignancy (including lymphoma) assessment. With the availability of comparators, the ARIA
`
`1.5. Effect Size of Interest or Estimated Sample Size Desired
`
`
`e Leishear White, Kira, Division of Epidemiology I, ARIA Sufficiency Memo for Guselkumab, BLA 761061, dated April
`13, 2017, DARRTS Reference ID: 4084180.
`f Bright, Patricia, Division of Epidemiology I, ARIA Sufficiency Memo for Tildrakizumab, BLA 761067, dated March
`16, 2018, DARRTS Reference ID: 4236035.
`
`
`
`Reference ID: 4421562
`
`Page 5 of 11
`
`

`

`
`
`2. SURVEILLANCE OR DESIRED STUDY POPULATION
`2.1 Population
`
`assessment could support an inferential analysis by determining the incidence rate between
`risankizumab exposure and malignancy as compared to the incidence rates following exposure
`to other individual psoriasis biologic medications (Table 1). ARIA could also evaluate a class-
`based effect by comparing the incidence rate of malignancy following exposure to any
`psoriasis biologic medications as compared to the incidence rate of malignancy following
`exposure to non-biologic systemic medications for the indication of psoriasis. Assessing a
`class-based effect would likely yield a higher number of users with events and might increase
`the capacity to detect a difference in effect size.
`Sample size requirements and the corresponding effect estimates will be described in an ARIA
`Planning Concept Brief for any outcomes deemed sufficient to address through ARIA.
`
`Risankizumab is indicated for the treatment of moderate-to-severe plaque psoriasis in adults
`who are candidates for systemic therapy or phototherapy. All patients identified as having
`received a dispensing of risankizumab in Sentinel could be considered in the population for
`postmarket surveillance. A comparator population may include patients that have received a
`dispensing of other psoriasis biologics as listed in Table 1. To evaluate a class-effect, a
`comparator population of patients receiving non-biologic systemic medications for psoriasis
`treatment may be used.
`
`2.2 Is ARIA sufficient to assess the intended population?
`ARIA can be used to identify patients with a risankizumab dispensing in the claims data. If the
`underlying indication of psoriasis is needed to further target this population, the population
`can be screened for the ICD-10 code of L40.XX (psoriasis).
` Few studies have been published that aimed to validate ICD-10 diagnostic codes for estimating
`the prevalence of psoriasis. A Swedish, population-based, validation study demonstrated the
`positive predictive values (PPV) of ICD-10 codes to range from 81% - 100% with a post-
`validation prevalence of 1.23% (95% CI: 1.21 – 1.25) for psoriasis. To date, no studies have
`validated the ICD-10 codes for estimating the prevalence of psoriasis in a U.S. population.
`However, several studies in the United States have aimed to validate ICD-9 diagnostic codes for
`psoriasis. These studies reported PPVs that aligned with the findings from the Swedish
`study.g,h Taken together, findings from these studies suggest that performance of the ICD-10
`codes (L40.XX) to identify psoriasis patients for surveillance purposes in the United States
`would be adequate.
` ARIA is sufficient to identify the indicated population for this analysis and is not a limiting
`factor of concern. However, with several treatment options available to patients (Table 1),
`market uptake of risankizumab will affect whether enough users are available to further
`characterize lymphoma risk given the rarity of these outcomes. The extent of market uptake
`can only be evaluated post-approval.
`
`
`
`g Asgari MM, Wu JJ, Gelfand JM, et al. Validity of diagnostic codes and prevalence of psoriasis and psoriatic
`arthritis in a managed care population, 1996 – 2009. Pharmacoepidemiol Drg Saf. 2013; 22(8): 842 – 849.
`h Icen M, Crowson CS, McEvoy MT, Gabriel SE, Maradit Kremers H. Potential misclassification of patients with
`psoriasis in electronic databases. J Am Acad Dermatol. 2008; 59: 981 – 985.
`
`
`
`Reference ID: 4421562
`
`Page 6 of 11
`
`

`

`
`
`3 EXPOSURES
`3.1 Treatment Exposure
`
`
`3.2 Comparator Exposure
`
`
`3.3 Is ARIA sufficient to identify the exposure of interest?
`
`
`4 OUTCOMES
`4.1 Outcomes of Interest
`
`Patients with pharmacy benefits who receive at least one dispensing of risankizumab can be
`identified in health care claims data.
`As mentioned previously, the regulatory goal of this ARIA assessment is signal detection.
`However, to help interpret the observed incidence rates of lymphoma among psoriasis
`patients treated with risankizumab, two comparator populations may be used: 1) patients
`using other psoriasis biologic medications (Table 1) and 2) patients using non-biologic
`systemic medications (to establish a class-effect). Both of these comparator populations could
`be identified through the Sentinel health care claims data.
`ARIA is sufficient to identify dispensings of both risankizumab and comparator biologics and
`non-biologic medications, and therefore is not a limiting factor.
`The outcomes of interest are: 1) lymphoma and 2) all malignancies.
` A Workgroupi supporting Mini-Sentinel development reviewed the literature to identify
`algorithms that could be used in electronic claims-based data to identify cohorts of vulnerable
`groups, including persons with selected cancers of interest.
` The Workgroup cautioned that:
`
`“Cancers are not typically studied as a homogenous group, given differences in the
` histological type and primary site of lesion – each that often has its distinct risk factors,
`
`screening requirements, pathology, clinical manifestations, diagnostic testing, differential
` diagnoses, staging, treatment and prognosis, as examples. Therefore, studies examining
`
`algorithms for identifying person with any-type of cancer are scant.”i
` Therefore, in the absence of cancer registry data, the Workgroup recommended against
`studying cohorts with an outcome of any cancer, but rather focusing on subcohorts with
`specific cancers. The Workgroup recommended that primary consideration should be given to
`the identification of person with hematopoietic cancers such as leukemias, lymphomas, and
`myelomas.
` As part of the Workgroup’s deliverable, the Workgroup specified an algorithm for lymphoma
`that involved: two or more diagnoses of cancer (ICD-9 codes) within two months (algorithm
`2); this algorithm performed with a PPV of 63% and a sensitivity of 80%.
`Given the findings and recommendations from the Workgroup (as described above), ARIA was
`
`
`4.2 Is ARIA sufficient to assess the outcome of interest?
`
`
`i Leonard C, Freeman C, Razzaghi H, et al. Mini-Sentinel methods: 15 cohorts of interest for surveillance
`preparedness. https://www.sentinelinitiative.org/sites/default/files/Methods/Mini-Sentinel Methods 15-
`Cohorts-of-Interest-for-Surveillance-Preparedness 0.pdf. Accessed March 29, 2019.
`
`
`
`Reference ID: 4421562
`
`Page 7 of 11
`
`

`

`
`
`deemed sufficient to identify lymphoma as an outcome for studying safety in the postmarket
`setting among risankizumab users. However, grouping all malignancies together offers less
`scientific rigor, thus deeming ARIA insufficient.
` The Tremfya (guselkumab) ARIA Sufficiency Memo provided the following additional
`information on validation:
` “Validation of malignancy outcomes has not been assessed in Sentinel. However, there have
`been published validation studies using health care claims data for malignancy. In Medicare, a
`63% positive predictive value was achieved using a complex algorithm.j Different claims-based
`definitions used for specific types of incident cancers all had very high specificity (~99%);
`however, the sensitivity varied between 40 and 90% by type of cancer. Positive predictive
`value (PPV) also varied by type of cancer. Hence, depending on the type of cancer of interest,
`health care claims data may be acceptable. The various definitions used by Setoguchi et al.
`included 1) a combination of diagnosis and procedure codes on the same day or within the
`same hospitalization; 2) two diagnoses of specific cancer within two months; 3) either
`definition 1 or definition 2. For lymphoma, specificity was ≥99.7% for all 3 definitions,
`sensitivity ranged from 55.2% to 83.3%, and PPV ranged from 56.6% to 62.8%, for the 3
`definitions. A study validating ICD-9 codes using Veteran Affairs data, found non-Hodgkin’s
`lymphoma to have the highest PPV (91%) with 100% sensitivity.k The PPV and sensitivity for
`Hodgkin’s lymphoma were not stated in the article. A Mini- Sentinel methods paper states that
`there are multiple types of lymphoma and multiple classifications for categorizing the types of
`lymphoma.l These can be based on etiology (T-cell and B-cell lymphomas) or separated based
`on expected outcomes (e.g., curability). Validation studies for the many specific types of
`lymphoma are not available for claims data, and therefore, it is unknown whether there are
`certain types of lymphoma which may have poor validation.”m
` “In summary, the Medicare validation study of lymphoma in general performed reasonably well
`(i.e., PPV: 57-63%). The VA study showed high PPV (i.e., 91%) for non-Hodgkin’s lymphoma.
`These PPV values are considered to be acceptable for the purpose of surveillance.”
` In addition to the limitation of validating overall malignancy outcomes of any type (i.e. variable
`PPV), there is insufficient long-term follow-up data. As described in the Figure below, roughly
`3.1%, 6.6%, and 9.5% of the Sentinel patient population in age groups 18-30, 31-64, and 65+
`years, respectively, would have at least 8 years of follow-up, as is required for the PMR
`observational study issued for risankizumab (see Section 7).
`
`
`
`j Setoguchi S, Solomon D, Glynn R, Cook E, Levin R, Schneeweiss S. Agreement of diagnosis and its date for
`hematologic malignancies and solid tumors between Medicare claims and cancer registry data. Cancer Causes
`Control. 2007;18 95 0:561-569.
`k Park LS, Tate JP, Rodriquez-Barradas MC, et al. Cancer incidence in HIV-infected versus inunfected veterans:
`Comparison of cancer registry and ICD-9 code diagnoses. JAIDS Clin Res. 2014, 5:7.
`l Schumock GT, Lee TA, Pickard AS, et al. Mini-Sentinel Methods: Alternative methods for health outcomes of
`interest validation. August 31, 2014.
`m Leishear White, Kira, Division of Epidemiology I, ARIA Sufficiency Memo for Guselkumab, BLA 761061, dated April
`13, 2017, DARRTS Reference ID: 4084180.
`
`
`
`Reference ID: 4421562
`
`Page 8 of 11
`
`

`

`Age Group
`(Years)
`
`
`
`Percentage of Patients by Years of Follow-Up Time
`
`
`
`
`
`3+
`4+
`5+
`6+
`7+
`
`
`
`
`
`
`
`8+
`
`
`
`<3
`
`
`
`3.1%
`4.9%
`7.4%
`75.0% 25.0% 16.7% 11.2%
`18-30
`6.6%
`9.7%
`66.1% 33.9% 24.6% 18.1% 13.2%
`31-64
`9.5%
`56.9% 43.1% 32.8% 24.8% 18.6% 13.7%
`65+
` Figure 1 includes data from 16 individual data partners. The start and end dates for data
`collection from these partners range from as early as January 1, 2000 through March 31, 2017.
` Taken together, these limitations deem ARIA insufficient to assess malignancy of any type as
`the outcome of interest.
`
`5 COVARIATES
`The covariates of interest include demographic (e.g., age, sex, calendar year, and geographic
`region), lifestyle (e.g., smoking status, alcohol use), medical history (e.g., family history of
`malignancy), and clinical (e.g., comorbidities and concomitant medications) characteristics.
`Demographic and certain clinical characteristics could be assessed in ARIA. Additional
`characteristics such as smoking or personal or family history of cancer may not be obtained
`reliably. Duration and severity of psoriasis also may not be available in claims data. However,
`covariate information would be important for subsequent study analyses that assess risk
`factors for malignancy outcomes and for assessing bias when comparing incidence rates
`between risankizumab users and other biologic users (Table 1). Therefore, covariate
`information is not critical for the regulatory purpose of signal detection.
`As mentioned previously, the regulatory goal of ARIA is signal detection (i.e. postmarketing
`surveillance). As such, the study design involves identifying the incidence of lymphoma in
`patients exposed to risankizumab (the study would not address the incidence of all
`malignancies due to challenges to ARIA sufficiency described above). However, with the
`availability of comparators, it was also determined that the study design could support an
`inferential analysis that would compare the incidence rates of lymphoma between
`risankizumab users versus cohorts exposed to other psoriasis biologics (Table 1).
`
`
` With a PPV of approximately 63%, nondifferential misclassification could undermine the
`ability of the inferential analysis to detect meaningful differences in lymphoma incidence rates
`between risankizumab users and other biologic users. Table 2 provides information that can
`be used to better understand the impact.
`
`Figure 1. Proportion of Patients with Follow-up Time for Patients Diagnosed with
`Psoriasis in the Sentinel Distributed Databasen
`
`5.1 Covariates of Interest
`
`
`5.2 Is ARIA sufficient to assess the covariates of interest?
`
`
` 6
`
`SURVEILLANCE DESIGN / ANALYTIC TOOLS
`6.1 Surveillance or Study Design
`
`
`
`n Source: Michael D. Nguyen, MD. FDA Sentinel Program Lead. Modular Program Report
`(cder_mpl1p_wp006_nsdp_v01)
`
`
`
`Reference ID: 4421562
`
`Page 9 of 11
`
`

`

`Table 2. Observed Relative Risk (RR) in the case of Non-Differential Misclassification
`
`6.2 Is ARIA sufficient with respect to the design/analytic tools available to assess the
`question of interest?
`
`As described in Table 2, even with a true, modest relative risk of 1.5, a PPV as low as 60%
`would underestimate the relative risk of lymphoma after risankizumab exposure by
`approximately 20%. This would result in an observed relative risk of 1.3, demonstrating that
`the impact of the low PPV would still allow a detectable increase in risk of lymphoma in
`risankizumab users compared to users of other biologics if a risk was in fact present.
`The analytic tools to conduct a surveillance study, an even an inferential assessment in this
`context, are available through ARIA.
`The analytic tools in ARIA are not a major limiting factor to feasibility. ARIA offers the tools
`needed to both describe the incidence of lymphoma and to conduct an inferential assessment
`comparing incidence rates to other psoriasis biologic medications and non-biologic systemic
`medications.
`ARIA was deemed sufficient to identify the risk of lymphoma with risankizumab treatment in
`psoriasis patients. The next step for assessing the lymphoma risk following risankizumab
`exposure is to fill out the ARIA Planning Concept Brief that prompts Sentinel’s routine
`monitoring of market uptake for risankizumab. If market uptake reaches a level sufficient to
`trigger the analysis, FDA investigators can fill in the Analytic Concept Brief and launch the
`assessment.
`ARIA was deemed insufficient for studying the outcome of ‘all malignancies’ among
`risankizumab users on account of the short length of follow-up in Sentinel and variable
`validation characteristics and sensitivity by malignancy. As such, the FDA is issuing a
`postmarket requirement to the Sponsor to evaluate malignancy risk following risankizumab
`exposure. This would be consistent with postmarketing requirements for the other approved
`products in this drug class.
`
`7 NEXT STEPS
`
`Reference ID: 4421562
`
`Page 10 of 11
`
`

`

`
`
`FDA is proposing the use of the guselkumab and tildrakizumab PMR language as a model for
`the risankizumab PMR. The proposed language is as follows:
`“Conduct an observational study to assess the long-term safety of risankizumab compared
`to other therapies used in the treatment of adults with moderate-to-severe plaque psoriasis
`who are candidates for systemic therapy or phototherapy in the course of actual clinical
`care. The study’s primary outcome is long-term malignancy. Secondary outcomes include,
`but are not limited to, serious infections, tuberculosis, opportunistic infections,
`hypersensitivity reactions, autoimmune disease, neurologic or demyelinating disease,
`cardiovascular, gastrointestinal and hematologic adverse events.
` Describe and justify the choice of appropriate comparator population(s) and estimated
`background rate(s) relative to risankizumab-exposed patients; clearly define the primary
`comparator population for the primary objective. Design the study around a testable
`hypothesis to assess, with sufficient sample size and power, a clinically meaningful increase
`in malignancy risk above the comparator background rate(s), with a prespecified statistical
`analysis method. Specify concise case definitions and validation algorithms for both
`primary and secondary outcomes. For the risankizumab-exposed and comparator(s)
`cohorts, clearly define the study drug initiation period and any exclusion and inclusion
`criteria. Enroll patients over an initial 4-year period and follow for a minimum of 8 years
`from the time of enrollment.”b
` The finalized PMR language will be issued upon approval.
`
`
`
`Reference ID: 4421562
`
`Page 11 of 11
`
`

`

`Signature Page 1 of 1
`--------------------------------------------------------------------------------------------
`This is a representation of an electronic record that was signed
`electronically. Following this are manifestations of any and all
`electronic signatures for this electronic record.
`--------------------------------------------------------------------------------------------
`/s/
`------------------------------------------------------------
`
`MICHELLE R IANNACONE
`04/19/2019 09:48:16 AM
`
`PATRICIA L BRIGHT
`04/19/2019 09:50:00 AM
`
`SIMONE P PINHEIRO on behalf of SUKHMINDER K SANDHU
`04/19/2019 10:01:45 AM
`signed as proxy for Dr. Sukhminder Sandhu, who has cleared this review but is currently
`out of office
`
`MICHAEL D BLUM on behalf of JUDITH W ZANDER
`04/19/2019 10:57:34 AM
`
`MICHAEL D NGUYEN
`04/19/2019 11:38:28 AM
`
`ROBERT BALL
`04/19/2019 04:41:01 PM
`
`Reference ID: 4421562
`
`

`

`
`Department of Health and Human Services
`Food and Drug Administration
`Center for Drug Evaluation and Research | Office of Surveillance and Epidemiology (OSE)
`
`Epidemiology: ARIA Sufficiency Memo
`Version: 2018-01-24
`
`April 19, 2019
`Date:
`Michelle R. Iannacone, PhD, MPH
`Reviewer:
`Division of Epidemiology I
`Patricia Bright, PhD, MSPH
`Team Leader:
`Division of Epidemiology I
`Sukhminder K. Sandhu, PhD, MPH, MS
`Deputy Division
`Division of Epidemiology I
`Director:
`Active Risk Identification and Assessment (ARIA) Sufficiency Memo
`Subject:
`for Pregnancy Safety Concerns
`Skyrizi (risankizumab)
`Drug Name:
`Application Type/#: BLA 761105
`Applicant/Sponsor: Abbvie Inc.
`OSE RCM #:
`2019-679
`
`
`
`
`
`Page 1 of 6
`
`
`
`
`
`
`
`
`
`Reference ID: 4421558
`
`

`

`
`Expedited ARIA Sufficiency Template for Pregnancy Safety Concerns
`1. BACKGROUND INFORMATION
`1.1. Medical Product
`
`Psoriasis is a chronic debilitating immunologic disease characterized by marked inflammation
`and thickening of the epidermis that result in thick, scaly plaques involving the skin. Psoriasis
`may undergo intermittent improvements and relapses in susceptible individuals over the
`course of their lifetime. Although traditional systemic therapies for psoriasis are effective,
`there may be a loss of efficacy during long-term use or patients may experience adverse events
`related to specific treatments.a
` Skyrizi (risankizumab) injection, for subcutaneous use, is indicated for the treatment of adults
`with moderate to severe plaque psoriasis who are candidates for systemic therapy or
`phototherapy. Risankizumab is a humanized immunoglobulin GI (IgG1) monoclonal antibody
`that is specifically directed against IL-23 p19. The framework of the risankizumab antibody
`has been engineered with two mutations in the Fc region to reduce Fcγ receptor and
`complement binding. Binding of risankizumab to IL-23 p19 inhibits the action of IL-23 to
`induce and sustain T helper (Th) 17 type cells, innate lymphoid cells, γδT cells, and natural
`killer (NK) cells responsible for tissue inflammation, destruction, and aberrant tissue repair.b
` The recommended dose of risankizumab is 150mg (two 75mg injections) administered by
`subcutaneous injection at Week 0, Week 4, and every 12 weeks thereafter.
`
`c
`The risankizumab BLA Unireviewd includes the following information on Human Reproduction
`and Pregnancy:
`“In cynomolgus monkeys, a dose-dependent increase in fetal/infant loss was noted in the
`risankizumab-treated groups compared to the vehicle control group. The percent
`fetal/infant loss was 19%, 32%, and 43% in the vehicle control, 5 mg/kg, and 50 mg/kg
`groups, respectively.
`In the risankizumab psoriasis clinical development program, male subjects and their female
`partners [who were not study subjects] were not required as per the protocols to use
`contraception. It is not expected that large molecule proteins would interact directly with
`DNA or other chromosomal material and the amount of risankizumb exposure to female
`partners transferred via seminal fluid is likely to be negligible.
`Nine paternal exposure pregnancies were reported in the partner of a male study subject in
`the risankizumab clinical development program, no congenital anomalies were reported as
`outcome.
`
`
`1.2. Describe the Safety Concern – Pregnancy Risk
`
`
`a Vide J, Magina S. Moderate to severe psoriasis treatment challenges through the era of biological drugs. An Bras
`Dermatol. 2017;92(5):668-674.
`b Sponsor Original Submission, Global Submit Review: Upload dated April 23, 2018, Risankizumab, Clinical
`Overview.
`c Risankizumab Provider Information Label. DARRTS ID: Pending.
`d BLA 761105 Multi-disciplinary Review and Evaluation, Skyrizi (risankizumab). Version date: February 1, 2019.
`Referenced, April 4, 2019.
`
`
`
`Reference ID: 4421558
`
`Page 2 of 6
`
`(b) (4)
`
`

`

`
`
`
`Total
`
`0
`
`17
`
`Total
`
`14
`
`3
`
`Table 1. Summary of Pregnancy Outcomes in Risankizu