25th Anniversary! Congress on Women’s Health April 27-30, 2017

An Update on Hormonal Contraception and The Changing U.S. Population 1 Anita Nelson, MD Professor and Chair Department of Obstetrics and Gynecology Western University of Health Sciences Women’s Health Congress, 2017

3 FDA Meta-Analysis on the Effect of Obesity on HC Effectiveness Individual participant data meta-analysis of 6 pivotal combination oral contraceptive (COC) trials and 1 contraceptive patch trial Analysis suggests 44% increased risk of pregnancy during COC use in obese compared to non-obese women FDA authors called for more data in obese women from Phase 3 clinical trials after an FDA meta-analysis showed an effect of obesity on hormonal contraceptive effectiveness A candidate patch AG200-15 has just completed Phase 3 studies and a potential effect of increased BMI was observed, consistent with the FDA findings We undertook a review of contraceptive trials outcomes over time, with a focus on the role of obesity Yamazaki M et al, Contraception 2015; 92: 445-52

4 Measures of Contraceptive Efficacy in Clinical Trials The primary measure of efficacy used by FDA is the Pearl Index (PI): On-drug pregnancies have an estimated date of conception between start date of study drug and 7-14 days after date of last study drug use Cycles in which study subjects use an alternate method of contraception are not counted in the PI denominator For some trials, cycles in which subjects report no sexual activity are also excluded from the PI calculation Yamazaki M et al, Contraception 2015; 92: 445-52 Number of On-drug pregnancies Number of On-drug cycles (13) (100)

Impact of Study Populations in Contraceptive Trial Outcomes Enrollment of All-Comers BMI/weight Racial/ethnic minorities Range of experience with contraception Stringent Trial Design Cycle exclusion rules In addition to study design changes, variety of other factors shown to impact study outcomes Socioeconomics Race/ethnicity Prior pregnancy Prior hormonal contraceptive use Some factors may be surrogates for non-compliance: Difference between probability of contraceptive failure during typical (incorrect/inconsistent) and perfect use (correct/consistent) shows impact of imperfect use 5 Trussell et al, 2011 Westhoff et al, Contraception 2012 Gerlinger et al, Contraception 2014

Creeping Pearl: Changing Study Design and Populations Over Time Pearl Indices from early contraceptive trials were typically <1 Trussell describes steadily increasing Pearl Index results from pivotal hormonal contraceptive trials over past 30+ years Reasons not entirely clear: Better detection of pregnancies More frequent testing/home testing More sensitive tests Less adherent study populations “Because study populations appear to be increasingly representative of the likely actual use once the product is marketed, we can expect to see even higher failure rates in ongoing and future studies.” 6 Trussell and Portman Contraception 2013

Creeping Pearl-Trussell 7 Trussell and Portman Contraception 2013

8 Examples of Oral Contraceptives With Higher Pearl Indices When Used As Comparators in Later Studies Product Trial Year Mean Weight/ BMI Pearl Index UB 95% CI Loestrin Fe 1/20 Original U.S. Registration 1973 Not available 0.75 Not available Ortho Tri-Cyclen Lo Phase 3 2002 23.6 kg/m2* 3.80 Loestrin 24 Fe U.S. Phase 3 2006 68.2 kg 3.67 13.20 Levlite Original German Registration 1998 62.7 kg 0.29 0.91† Original U.S. Registration 1998 63.0 kg 1.08 2.34† Seasonale Phase 3 2003 69.7 kg 3.75 8.60 Nordette Original U.S. Registration 1982 Not available 0.48 1.04† Seasonale Phase 3 2003 71.0 kg 2.22 6.38 Seasonique Phase 3 2006 71.8 kg 4.40 Not available Sourced from publicly available NDA Reviews *Mean weight not available †Calculated based on cycle and pregnancy data in NDA review

2002 Publication Draws Attention to a Potential Relationship Between Obesity and Hormonal Contraceptive Effectiveness “Body Weight and Risk of Oral Contraceptive Failure” Retrospective case-control analysis comprising 755 group health cooperative enrollees Relative risk of 1.6 of OC failure for women in highest body weight quartile (> 70.5 kg) compared to women of lower weight Follow-up study titled “Body Mass Index, Weight, and Oral Contraceptive Failure Risk” Risk of pregnancy 60% higher in women with BMI > 27.3 kg/m2 and over 70% higher in women with BMI > 32.2 kg/m2 Risk of pregnancy was more than doubled in women with BMI > 27.3 kg/m2, over 70% higher in women weighing > 74.8 kg, and nearly doubled in women weighing > 86.2 kg among consistent users of oral contraceptives 9 Holt et al, 2002 Holt et al, 2005

Reproductive Health Considerations Related to BMI BMI is a significant risk factor for multiple outcomes related to reproductive health Higher morbidity and mortality with pregnancy and childbirth Higher risk of VTE with hormonal contraception Growing evidence of greater risk of hormonal contraceptive failure Although risk of morbidity and mortality on hormonal contraception higher in women of high BMI, pregnancy-related risk of bad outcomes due to VTE and other events still higher Women of high BMI need contraceptive options--not all will want IUD or other LARC Understanding the role of weight in contraceptive effectiveness is important for women and prescribers 10 ACOG Practice Bulletin No. 156, Dec 2015. Ovesen P et al. Effect of Prepregnancy Maternal Overweight and Obesity on Pregnancy Outcome. Obstet Gynecol 2011; 118: 305-312.

United States Weight Trends 11 Source: CDC/NCHS. National Health and Nutrition Examination Survey BMI Trends Among Adults Aged 20 to 74 years: United States, 1960-1962 Through 2009-2010

Contraceptive Trials Historically Enrolled Only Women of Normal Weight/BMI Product Year Mean Weight/BMI‡ U.S. Pearl Index Nordette 1982 Not available 0.48 Ortho-Cyclen 1989 61 kg 0.96 Ortho Tri-Cyclen 1992 60 kg 1.21 Ortho-Cept 1992 Not available 1.12 Alesse 1997 66 kg 0.84 Levlite 1998 63 kg 1.08 Mircette 1998 23.5 kg/m2 1.11 Yasmin 2001 63 kg 0.40 12 Mean baseline weight and Pearl Index results from pivotal phase 3 trials of contraceptives conducted from 1980-2000: ‡All weights standardized to kilograms for comparison Yasmin’s clinical trials contained an inclusion criterion requiring subjects to be “within 25% of ideal body weight” Sourced from publicly available NDA Reviews and product labels

Impact of Body Weight on Hormonal Contraceptive Effectiveness-Ortho Evra FDA concluded during review that Ortho Evra effectiveness was reduced for women weighing > 90 kg. < 3% of the Ortho Evra study population had a baseline weight> 198 pounds (90 kg), but accounted for 33% of the pregnancies 17% of subjects had baseline weight 74 to 94 kg, but accounted for 27% of the pregnancies Data published by M Zieman et al, 2002 Ortho Evra (*Xulane) label advises the patch “may be less effective in preventing pregnancy in women who weigh 198 lbs. (90 kg) or more.” 13 Zieman M et al, Fertil Steril 2002 Sourced from publicly available NDA Reviews and product labels *Xulane is the generic equivalent of the Ortho Evra patch

14 Quartette Trial Results Weight Pearl Index UB 95% CI 28-Day Cycle-Equivalents All 3.19 4.03 < 70 kg 2.59 3.67 > 70 kg to < 90 kg 3.38 5.17 > 90 kg 4.82 7.60 91-Day Cycles All 3.52 4.44 < 90 kg 3.14 4.12 > 90 kg 5.37 8.45 The Phase 3 study for Quartette enrolled a racially and ethnically diverse population of subjects 18 to 40 years of age 28.3% were obese (BMI > 30 kg/m2 ) 44% weighed < 70kg, 28% weighed > 70 to < 90kg, and 18% weighed > 90kg The Quartette label does not contain information regarding BMI or weight impact on the product’s effectiveness Quartette Sub-Analysis by Weight: Sourced from publicly available Quartette package insert and NDA review

AG200-15 Candidate Hormonal Contraceptive Patch AG200-15 is a once-weekly contraceptive patch Comprises Levonorgestrel (LNG) and ethinyl estradiol (EE) Intended to improve convenience and compliance1 Designed to deliver approximately 30 mcg of EE per day and approximately 120 mcg of LNG per day2 Currently available patch delivers approximately 56mcg of EE per day3 Alternatives to daily birth control pills are important Women report fitting daily pills into their busy lifestyles can be challenging4 Women frequently forget to take their pills (approximately 1-4x per month)4; up to 15% of birth control users have difficulty taking pills consistently and correctly5 Inconsistent use of oral contraceptives is a major contributor to unplanned pregnancy6 15 1Archer et al, Contraception 2004; 2Kaunitz et al, Contraception 2014; 3Devineni et al, J Clin Pharmacol 2007; 4Qualitative consumer and HCP market research, RG&A 2012; ACE Survey 2013; 5Rosenberg et al, Contraception 1995; 6Finer et al, Perspect Sex Reprod Health 2006

SUN MON TUE WED THU FRI SAT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 The AG200-15 Regimen is Based on an Established 21/7 Schedule Patch applied once weekly for three weeks followed by a 4th week with no patch Can be applied to abdomen, buttock, or upper torso Buttock and abdomen were the most common patch placement choices in clinical trials 16 Source: Data on File, Agile Therapeutics 7 days no patch Patch 1 Patch 2 Patch 3 Patch 1 Upper Torso 12% Abdomen 40% Buttock 48%

The AG200-15 Development Program Comprehensive clinical program conducted Over 4,100 women enrolled; over 3,500 have received the AG200-15 patch Phase 1 and Phase 2 trials support hormone delivery consistent with current 30mcg OCs In Phase 3 studies, over 1,000 women have received AG200-15 for 12 months Approval was originally sought for AG200-15 in 2012 based on two original Phase 3 trials (CL12 and CL13); FDA denied approval and requested a third Phase 3 trial (CL23-The SECURE Trial) Based on data from the SECURE trial, resubmission for the AG200-15 patch is expected by the end of June, 2017, with potential for FDA approval by the end of 2017 17 Source: Data on File, Agile Therapeutics

The SECURE Trial Was Designed to Assess the Efficacy and Safety of AG200-15 in a Real-World Population Multicenter, single-arm, open-label 13-cycle trial at 102 experienced U.S. clinical sites ~ 2,000 healthy subjects aged > 18 treated with laser-etched patches Representative sample of women seeking hormonal contraception No exclusions for BMI/weight Stringent Trial Design Subject daily entries in electronic diaries Frequent pregnancy testing, including provision of home pregnancy tests Exclusion of cycles for BOTH use of back-up contraception and lack of sexual activity Analysis Efficacy measure was Pearl Index in an ITT population of subjects 35 years of age and under Prespecified analyses related to BMI and body weight 18 CRL = Complete Response Letter; ITT = Intent to Treat

Impact of Study Population and Study Design SECURE Trial Design vs Historical Trials 19 Factor Study Design Historical SECURE Trial Study Design Enrollment Criteria BMI Restricted All comers-representative of US population seeking HC Race/ethnicity Majority Caucasian Representative of US population seeking HC New users Experienced populations Enroll substantial number of new users Ex-US subjects Substantial proportion of European subjects supported US approvals Approvability based on US-only data Study Conduct Pregnancy testing If pregnancy suspected Every visit Provide home pregnancy tests/encourage testing Serum hCG at final visit Primary Analysis BMI Exclude subjects with BMI above 35/30 Include all BMI categories Cycle exclusion Use of back-up contraception Use of back-up contraception Lack of sexual intercourse

20 SECURE Study Population Demographics Study SECURE Ortho Evra Trials Quartette Trial Age Mean age 28 years 28 years 27 years < 35 years 90% 83% 90% > 35 10% 17% 10% Body Mass Index Mean BMI* 28.3 kg/m2 23.6 kg/m2 27.4 kg/m2 < 25 (normal) 39% Not available 47% 25 - < 30 (overweight) 25% 25% > 30 (obese) 35% 28% Race White 67% 91% 64% Black 24% 5% 19% Asian 3% 2% 2% Other 6% 2% 14% Ethnicity Hispanic 20% Not available 11% Non-Hispanic 80% 89% Hormonal Contraception Use Current user 35% Not available 44% Recent user 13% Former user 43% 39% *Based on CDC BMI categories Information is based on currently marketed Ortho Evra and Quartette product labels and publicly available information. We have not performed a head-to-head comparison of AG200-15 to Ortho Evra or Quartette. Percentages in table are rounded to nearest integer; may not add up to 100%

Pearl Indices for AG200-15 Patch--SECURE Trial 21 BMI Category BMI (kg/m2) % of Study Population Pearl Index UB 95% CI Normal* < 25 39% 3.03 4.62 Overweight > 25 - < 30 25% 5.36 7.98 Obese > 30 35% 6.42 8.88 Non-Obese* < 30 65% 3.94 5.35 Obese > 30 35% 6.42 8.88 ITT = Intent to Treat; all results shown are based on ITT subjects < 35 years of age UB 95% CI = upper bound of the 95% confidence interval Population (ITT) Pearl Index UB 95% CI < 35 years of age 4.80 6.06 *Reflective of Historical CHC Trial Populations Overall results: An effect of obesity was observed:

Observable Trend in Pearl Indices for Approved Combined Hormonal Contraceptives (CHCs) 22 Product (Year Approved) Pearl Index Ortho Evra (2001) 1.07 Nuvaring (2001) 2.02 Yasmin (2001) 0.4 Seasonale (2003) 1.98 Loestrin 24 FE (2006) 1.82 Seasonique (2006) 1.34 Yaz (2007) 1.41 Lybrel (2007) 2.38 LoSeasonique (2008) 2.74 Natazia (2010) 1.64 LoLoestrin (2010) 2.92 Generess (2010) 2.01 Quartette (2013) 3.19 Quartette >90kg 4.82 Quartette >70 to <90kg 3.38 Quartette <70kg 2.59 Sources: Trussell, et al., The Creeping Pearl (2013), currently marketed product labels, and publicly available information Quartette Overall AG200 - 15 Normal (BMI <25) AG200 - 15 Non - Obese (BMI <30) AG200 - 15 Overall (mean BMI 28.3) AG200 - 15 Overweight (BMI 25 to <30) AG200 - 15 Obese (BMI ?30) Quartette ?90kg Quartette ?70kg to <90kg Quartette <70kg 0 1 2 3 4 5 6 7 2000 2005 2010 2015 2020 Pearl Index Historical Pearl Indices for CHCs Approved Since 2000 and the Pearl Indices Observed in the SECURE Trial Approved Combined Hormonal Contraceptives, CHCs AG200-15 Quartette Weight Groups Pearl Index Trend for Approved Products Since 2000

Conclusions Women of all weight categories need contraceptive choices There is increasing evidence that obesity may affect hormonal contraceptive effectiveness; the SECURE trial has contributed to this evidence The weight of women of reproductive age in the U.S. is increasing, making this effect more important to understand and quantify to inform contraceptive choices More research is needed to understand the role of other factors, including compliance, race/ethnicity, pharmacokinetics, etc. 23

DISCUSSION 24