Summary Review

Decisional Memorandum

New Drug Application 21998, Supplement 5

Levonorgestrel 1.5 mg Tablet Emergency Contraceptive

Labeling Supplement for Update to Mechanism of Action Information

I. Executive Summary

II. Background

III. Applicant’s Rationale for Requested Labeling Change

IV. Summary of Multidisciplinary Review

A. The Typical Events of the Early Reproductive Process

B. Approach to the Review

C. Information on Mechanism of Action that was Available at the Time of the Review

Division’s Analysis in Preparation for the 2003 Nonprescription Drugs Advisory Committee

Meeting

D. Information on Mechanism of Action that Became Available After the 2003 Review

E. Integrated Summary of Current State of Knowledge Regarding Mechanism of Action of

Levonorgestrel Emergency Contraception

V. Decision on Action to be Taken

VI. Labeling

VII. References

I. Executive Summary

Levonorgestrel emergency contraception (LNG-EC) is effective in preventing pregnancy when taken within

3 days after unprotected intercourse. LNG-EC does not terminate an established pregnancy and does not

affect a continuing pregnancy.

In this supplement, the applicant requests an update to the labeling of nonprescription LNG-EC (NDA

21998), in particular to certain parts of the labeling that concern mechanism of action. The

multidisciplinary review team concurs that, based on the totality of both pre-existing and new data,

updated labeling is appropriate. Data are strong for a mechanism of action of delay or prevention of

ovulation, and data are weak to speculative regarding any postovulatory mechanistic effects, such as on

fertilization or implantation.

Key questions answered by the multidisciplinary review include:

A. Does levonorgestrel emergency contraception inhibit ovulation?

Clinical data are strong that LNG-EC, when administered prior to the luteinizing hormone (LH) surge,

inhibits or delays ovulation.

Page 1 of 27

Reference ID: 5100570

B. Does levonorgestrel emergency contraception affect sperm function?

In vitro and human in vivo data do not support a significant effect of LNG on cervical mucus quality, sperm

quantity in the genital tract, sperm motility, or acrosomal reaction, when LNG is administered at doses

similar to that used in EC.

C. After ovulation occurs, does levonorgestrel emergency contraception affect endometrial receptivity

or implantation?

In vitro data and human clinical data show that LNG-EC does not affect exploratory markers of

endometrial receptivity after ovulation occurs. It should be noted that such markers are not validated

biomarkers, and the multidisciplinary team did not find published data that changes in endometrial

receptivity biomarkers translate into a clinical effect on blastocyst implantation. The totality of the

evidence indicates that LNG-EC does not alter the endometrium in a clinically meaningful way to prevent

implantation. Studies showing no evident reduction in the likelihood of pregnancy when LNG is

administered to women postovulation are consistent with histological and biochemical findings and

support the conclusion that LNG-EC is ineffective when administered postovulation.

D. Does levonorgestrel emergency contraception affect pregnancy rate after ovulation?

When LNG-EC is administered after ovulation occurs, the rate of pregnancy is what one would expect if

LNG-EC had not been administered. This is arguably the most important aspect of the evidence.

The data and citations supporting the above answers are included and discussed later in this

memorandum. Given the above findings, it is appropriate to update the Drug Facts label (DFL) and

Consumer Information Leaflet (CIL) for LNG-EC to reflect what has been learned about the LNG-EC

mechanism of action (MoA) in the years since the original nonprescription LNG-EC approval.

The scientific and clinical concept of the time period of pregnancy is generally consistent with the example

found in the Code of Federal Regulations (CFR) under 45 CFR 46.202, which reads: “Pregnancy

encompasses the period of time from implantation until delivery.” LNG-EC prevents pregnancy by acting

on ovulation, which occurs before implantation, and because data do not support that LNG-EC affects

implantation, LNG-EC does not terminate pregnancy.

II. Background

Levonorgestrel is a synthetic progestin. Under the brand name Plan B One-Step® (hereafter referred to as

Plan B One-Step), the levonorgestrel 1.5 mg tablet is approved as a nonprescription emergency

contraceptive, with the Use (nonprescription equivalent to indication) stated on the Drug Facts label (DFL)

as “for women to reduce chance of pregnancy after unprotected sex (if a contraceptive failed or you did

not use birth control).”

Levonorgestrel was first approved as a prescription emergency contraceptive product in 1999 under the

brand name Plan B® (hereafter referred to as Plan B); Plan B was a two-dose product with each tablet 0.75

mg. In 2006, nonprescription Plan B was approved in a partial prescription-to-nonprescription switch. In

2009, Plan B One-Step (the product in this labeling supplement), was approved as a partial prescription-

to-nonprescription switch, with a 1.5 mg tablet requiring only one dose. In 2013, the single dose 1.5 mg

Page 2 of 27

Reference ID: 5100570

product was approved as a full prescription-to-nonprescription switch. The approval history for Plan B and

Plan B One-Step can be found at FDA’s webpage drugs@FDA available at

https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021045

and

https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021998 .

Mechanism of action (MoA) information is rarely included in nonprescription labeling. It is not a required

element in the regulations that govern nonprescription labeling (21 CFR 201.66, Format and Content

Requirements for Over-the-Counter Drug Product Labeling). MoA is often scientifically complex and is

usually determined by nonclinical (animal or in vitro) studies, sometimes augmented by human clinical

pharmacology or human clinical studies. For prescription products, when MoA has been established, it

must be included in the prescription labeling; in that circumstance, the labeling is written for licensed

prescribers who have the scientific training to interpret MoA information correctly. By contrast,

nonprescription labeling is written for consumers of average US health literacy, generally a reading level

of approximately 6

th

to 8

th

grade. The US nonprescription label, called the Drug Facts label, is intended to

provide consumers with the key information they need to understand how to use a drug safely and

effectively without help from a licensed healthcare professional. Mechanism of action information is not

needed for the safe or effective use of nonprescription drugs in general, and not for levonorgestrel

emergency contraception in particular.

However, MoA-related information was included in the labeling of nonprescription levonorgestrel

emergency contraception at the time of its original prescription-to-nonprescription switch in 2006. In

2003, at the public Advisory Committee at which that application (for the LNG-EC 0.75 mg tablet) was

discussed, MoA arose as a topic of discussion. In 2004, the FDA requested that the applicant include

information relevant to MoA in the nonprescription labeling. In 2005, the applicant agreed, and

information relevant to MoA was present in the 2006 labeling for the original approval of the

nonprescription LNG-EC 0.75 mg tablet. When the nonprescription LNG-EC 1.5 mg tablet was approved,

the same MoA-related information was included in labeling.

In the current approved labeling for Plan B One-Step (NDA 21998, LNG- EC 1.5 mg tablet), the following

language relevant to MoA appears:

Drug Facts label:

In the “Warnings” section, under “Do not use”:

“Do not use if you are already pregnant (because it will not work).”

In the “Other information” section:

“• this product works mainly by preventing ovulation (egg release). It may also prevent

fertilization of a released egg (joining of sperm and egg) or attachment of a fertilized egg

to the uterus (implantation).”

Page 3 of 27

Reference ID: 5100570

Consumer Information Leaflet:

Under the heading “What Plan B One-Step® is not.”:

“Plan B One-Step® will not work if you are already pregnant and will not affect an existing

pregnancy.”

Under the heading “When not to use Plan B One-Step®.”:

“Plan B One-Step® should not be used if you are already pregnant, because it will not

work.”

Under the heading “How does Plan B One-Step® work?”:

“Plan B One-Step® is one tablet with levonorgestrel, a hormone that has been used in

many birth control pills for several decades. Plan B One-Step® contains a higher dose of

levonorgestrel than birth control pills, but works in a similar way to prevent pregnancy. It

works mainly by stopping the release of an egg from the ovary. It is possible that Plan B

One-Step® may also work by preventing fertilization of an egg (the uniting of sperm with

the egg) or by preventing attachment (implantation) to the uterus (womb).”

III. Applicant’s Rationale for Requested Labeling Change

In its justification for the requested labeling change, the applicant states that the current labeling of LNG-

EC may be misleading to consumers, because it contains information on a possible contribution of

postovulatory effects to LNG-EC’s mechanism of action, and some consumers are hesitant to use a product

that might affect postovulatory events, in particular implantation of the blastocyst.

The science has evolved since the original approval of LNG-EC, and the applicant submitted publications

from an updated comprehensive literature search that identified a total of 33 publications concerning

LNG-EC MoA. Six of these articles reported on clinical studies. Other types of publications included 4

clinical pharmacology research articles, 2 nonclinical animal research articles, 16 review articles, 1

professional society publication, and 4 opinion/editorial articles. The applicant presented a summary of

the findings of these publications, and concluded that data for an effect on ovulation are strong, but data

for postovulatory effects are weak or speculative. The applicant asserts that updates to the labeling are

needed to make the labeling more accurate, to reduce consumer confusion, and potentially to reduce

barriers to use of the legally marketed approved product.

IV. Summary of Multidisciplinary Review

A. The Typical Events of the Early Reproductive Process

In order to understand the points in the reproductive process at which an emergency contraceptive might

have its effect(s) to prevent pregnancy, it is useful to understand the typical sequence of events that occur

early in the reproductive process. The process is highly complex, but some major events of interest include

ovulation, fertilization, and implantation.

Figure IV.A.1 below illustrates the hormonal and follicular changes that occur in a typical menstrual cycle.

Day 1 of the cycle (the first day of menstruation) starts the follicular phase (referred to in the figure below

as the proliferative phase), in which the pituitary hormone follicle-stimulating hormone (FSH) stimulates

Page 4 of 27

Reference ID: 5100570

the development of follicles in the ovary. As follicles mature, systemic estrogen levels rise and

hypothalamic feedback occurs. In the final stages of follicular maturation, near the middle of the cycle, a

surge in systemic levels of the pituitary hormone luteinizing hormone (LH) occurs. Approximately 24-48

hours later, a dominant follicle ruptures and an ovum is released; this is referred to as ovulation. The

phase after ovulation is called the luteal phase (referred to in the figure below as the secretory phase);

during this time, there is a rise in systemic progesterone and estrogen, with accompanying changes in the

uterine endometrium.

Figure IV.A.1: Hormonal Changes and Ovarian Follicular Development in a Typical 28-Day Menstrual

Cycle

Source: Leung 2010, pg 159.

Abbreviations: FSH = follicle-stimulating hormone; LH = luteinizing hormone

Note: In many sources, the “proliferative phase” is often referred to as the “follicular phase”, and the “secretory phase” is often referred to

as the “luteal phase”.

The days in which unprotected intercourse is most likely to result in fertilization are approximately the

five days prior to ovulation, and the day of ovulation (Wilcox 1995).

Page 5 of 27

Reference ID: 5100570

Figure IV.A.2: Probability of Conception Related to Timing of Intercourse During Menstrual Cycle

*

Source: Adapted from Wilcox 1995, pg 1519

*Day 0 is the day of ovulation

“Probability of Conception” for a given day was calculated by dividing the number of pregnancies observed by the number of cycles with

intercourse only on that specific day (and no intercourse occurring on other days within Days -6 to +1).

After unprotected intercourse, sperm that reach the fallopian tube generally survive for 3-4 days, although

this is variable. The released ovum enters the fallopian tube, and generally must encounter sperm within

12-24 hours in order for fertilization to occur. If fertilization happens, the fertilized egg develops into a

blastocyst, which generally moves through the fallopian tube in about 3 days and enters the uterus on

about Day 17 or 18 of the menstrual cycle. The blastocyst begins implantation about 3 days later. In order

for implantation to occur, the endometrium needs to be in a receptive state. Numerous biochemical

markers of a receptive endometrium have been explored. While none of these is qualified as a validated

biomarker, they include measurements of estrogen and progesterone levels, appropriately timed with

ovulation, followed by assessment of endometrial downregulation of receptors for estrogen and

progesterone; and may also include expression of such factors as glycodelin, prostaglandins, interleukins,

integrins, epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), leukemia inhibitory

factor (LIF), tumor necrosis factor (TNF), and cell adhesion molecules (Achache 2006, Lindhard 2002).

Note that if intercourse occurs prior to ovulation, administration of a hypothetical drug to prevent

pregnancy after ovulation could still prevent pregnancy (based upon the number of days between

intercourse, fertilization, and implantation) if postovulation events (fertilization or implantation) were

Page 6 of 27

Reference ID: 5100570

directly inhibited. However, as discussed in detail below, LNG-EC administered prior to ovulation is

effective to prevent pregnancy, but administered after ovulation is ineffective to prevent pregnancy,

providing evidence that LNG-EC does not directly alter events after ovulation.

B. Approach to the Review

Broadly speaking, publications related to the question of MoA of LNG-EC fall into three categories: original

research articles (meaning publications of original scientific or clinical research studies), review articles,

and commentary/opinion articles. The latter two categories include writings of a wide range of scientific

rigor, varying from careful systematic reviews to commentaries containing hypothetical or speculative

statements. The review for this labeling supplement focuses only on original research articles, both those

submitted by the applicant, and additional articles identified in systematic literature searches conducted

by members of the multidisciplinary review team. Although review and commentary/opinion articles were

not the focus of this review, all articles selected by the applicant (which included both supportive and

dissenting points of view), as well as additional review articles from separate systematic literature

searches by members of the multidisciplinary review team, were reviewed for identification of possible

additional sources of original research, and to consider all viewpoints. The following articles were

identified as original research articles by the applicant: do Nascimento 2007; Lalitkumar 2007; Marions

2002; Meng 2009 and 2010; Muller 2002; Noé 2011; Novikova 2007; Ortiz 2004; Palomino 2010; Ugocsai

2002, and Vargas 2012. Full citations for these and all articles mentioned in this document are in the

References section.

The following additional original research articles were identified by the multidisciplinary review team in

a systematic literature search and were reviewed. (A few of these articles were also submitted by the

applicant, but were not identified as original research articles in the applicant’s Table of Studies on page

10 of the applicant’s justification document.): Durand 2001, 2005, and 2010; Hapangama 2001; Landgren

1989; Ling 1979 and 1983; Marions 2004; Matsuo 2020; Okewole 2007; Tirelli 2008; and Yeung 2002.

Of the aforementioned original research articles, the review team identified the following articles as the

strongest clinical evidence of LNG-EC’s effects on ovulation: Durand 2001; Marions 2002; Croxatto 2004;

Okewole 2007; Novikova 2007; Tirelli 2008; and Noé 2011. The strengths and limitations of each of these

studies are discussed further below.

Other review articles and commentary/opinion articles that were identified by the applicant, and

additional such articles identified by the multidisciplinary review team, are also cited in the References

section. As noted above, the multidisciplinary team reviewed all articles and materials submitted by the

applicant, as well as additional articles from the FDA team’s literature search.

The review below begins with a description of the information (from research studies) that was known at

the time of FDA’s review of the literature in preparation for the 2003 Nonprescription Drugs Advisory

Committee meeting, then proceeds through the data published since then on new original research. The

review then proceeds to an integrated summary of what the current state of knowledge is about the

mechanism of action of LNG-EC.

Page 7 of 27

Reference ID: 5100570

C. Information on Mechanism of Action that was Available at the Time of the Review Division’s

Analysis in Preparation for 2003 Nonprescription Drugs Advisory Committee Meeting

At the time of FDA’s review in preparation for the 2003 public Advisory Committee meeting at which the

application for nonprescription LNG-EC was discussed, limited published data were available from original

research regarding MoA for LNG-EC.

Some work had been published in 1979 and 1983 by Ling and colleagues, but because the treatment given

in these studies was a combination of levonorgestrel and ethinyl estradiol, and not LNG alone, these

studies were not informative as to the individual effect of LNG.

1. Clinical Studies Evaluating Effect of Levonorgestrel on Ovulation and the Endometrium

a. Landgren and colleagues (1989) published a human pharmacodynamic study that showed that

suppression of ovarian function by LNG was seen only when the drug was administered in the late

follicular phase and around the time of ovulation.

On pathological assessment of endometrial biopsy specimens, a decrease in the number and diameter of

endometrial glands was observed only when LNG was administered during the follicular phase, and not

when it was administered in the luteal phase. A limitation of this study was that the total amount of LNG

administered per subject (3 mg, administered as 0.75 mg every 2 days for 4 doses) was higher than that

of the approved LNG-EC product in this application (1.5 mg given as a single dose). However, given this, it

seems unlikely that LNG-EC at the current (lower) approved dose would have exhibited greater effects

than that seen at higher doses in this study.

b. Durand and colleagues (2001) published a clinical study that showed that LNG prevented ovulation

when administered on Day 10 of the menstrual cycle (pre-ovulatory) and did not find impairment of

corpus luteum function or endometrial morphology when LNG was administered in the peri-ovulatory or

postovulatory period.

The study included 45 healthy surgically sterilized women who ranged in age from 29 to 35 years, and

who had regular menstrual cycles. Subjects were randomized into one of three groups. All participants

were studied for two menstrual cycles; a control cycle without administration of LNG, and an active

treatment cycle during which LNG was administered. In the second cycle, each group received two doses

of LNG 0.75 mg, with 12 hours between doses. The difference in regimen between groups was in when in

the menstrual cycle each group received the LNG. Group A received it on Day 10 of the menstrual cycle (a

time expected to be pre-ovulatory); Group B received it immediately after detection of luteinizing

hormone (LH) in the urine (a time expected to be peri-ovulatory); and Group C received it 48 hours after

detection of LH in the urine (a time expected to be postovulatory). A Group D was also formed after it was

noted that there were inconsistencies between serum LH and urinary LH in 12 of the 90 studied cycles.

Because of this observation, the investigators decided to use serum LH, estrogen, and progesterone levels

for cycle dating, rather than using urinary LH. Thus, Group D was formed from four participants from

Group B and four from Group C. Group D received LNG 3 +/- 1 day prior to the serum LH surge.

During both the control cycle and the treatment cycle, from the time of urinary LH detection, the

investigators performed daily transvaginal ultrasound until the investigators observed follicular rupture

(FR). Also, during both the control and treatment cycles, endometrial biopsies were performed on all

participants 9 days after urinary LH detection. This day was chosen for biopsy because it occurs during the

Page 8 of 27

Reference ID: 5100570

time when the endometrium was felt to be optimally receptive to implantation, based on other published

data (Harper 1992, Yaron 1994).

In Group A (the pre-ovulatory group), after receiving LNG, 12 of 15 participants did not ovulate; follicular

rupture was not detected on ultrasound, and urinary LH was not detected. For the 3 of 15 Group A

participants who did ovulate, there was a delay in urinary LH detection, serum LH surge, and follicular

rupture. In Groups B (presumed peri-ovulatory), C (presumed postovulatory), and D (use of serum LH for

ovulation detection, with LNG dosing 3 +/- 1 day prior to LH surge), all participants ovulated, and LNG-EC

did not alter the day of the cycle on which ovulation occurred.

Regarding the endometrial biopsy results, which were all read in blinded fashion, morphology did not

differ between control and treated samples. Of note, there was no difference between groups in the

presence of spiral arteries, which are important for implantation (Pijnenborg 2006, Craven 1998,

Blankenship 1997). Similarly, stromal edema was similar between treated and control groups, also

suggesting a lack of effect on endometrial morphology important for preparation for implantation (Okada

2001).

Table IV.C.1.b: Endometrial Morphology in Control and Treated Groups

Control

Group B

Group C

Group D

(n = 41)

(n = 10)

(n = 11)

(n = 3)

Postovulatory day

8.6

±

1.3

8.7

±

0.6

9.0

±

0.8

9.0

±

0

Total area of tissue (mm

2

)

1,988 ± 55

2,003 ± 45

1,984 ± 62

2,015 ± 26

# of glands/visual field

59 ± 12

58 ± 7

55 ± 8

58 ± 1

# of glands/mm

2

30 ± 6

29 ± 4

28 ± 4

29 ± 0.8

Stromal edema (mm

2

)

1,049 ± 308

1,225 ± 261

1,011 ± 209

1,142 ± 40

% of tissue with stromal edema

53 ± 15

61 ± 14

51 ± 10

57 ± 1.4

Spiral arteries per visual field

6 ± 3

4 ± 1

5 ± 2

4 ± 0.7

Source: Durand 2001, pg 231

Results expressed as mean +/- SD

24/33 biopsies from treated ovulatory cycles included. Reasons for exclusion: Insufficient tissue (4 control, 1 Grp A, 1 Grp D); sampling not

correlated with cycle day (3 Grp A, 4 Grp D)

Overall, the data reported by Durand supported prevention of ovulation by LNG. In addition, peri- and

postovulatory administration did not affect endometrial morphology, making an effect on implantation

unlikely.

c. Marions and colleagues (2002) published a clinical study in healthy fertile women; this study examined

the effect of LNG on ovarian function and endometrial development. With regard to ovarian function, the

authors observed inhibition of the LH surge and ovulation when LNG was given pre-ovulation. However,

the authors did not measure the timing of the LH surge relative to LNG dosing, precluding definitive

conclusions.

The authors also observed no effects on markers of endometrial receptivity, whether LNG was given

before or after ovulation. However, there are no validated markers of endometrial receptivity modulating

blastocyst implantation. Therefore, the biomarkers evaluated in this study would be considered as

exploratory. That being said, the consistent lack of alteration of any biomarker in this study provides

support for the conclusion that there were no meaningful changes in the endometrium with LNG

administration.

Page 9 of 27

Reference ID: 5100570

d. Ugocsai and colleagues (2002) published a report on three women who, “due to enormous fear of

unwanted pregnancy”, had intentionally taken excess LNG (3-3.75 mg as a single dose, followed by at least

one other dose of 0.75 mg). The women then sought care “as anxiety emerged regarding health

consequences and possible sequelae.” Approximately 96 hours after the second dose, each woman

underwent endometrial biopsy. Their specimens were compared to two ovulating women, one in each

half of the menstrual cycle. The investigators performed scanning electron microscopy on the specimens

and reported that the endometria of the women who intentionally overdosed on LNG differed from those

of the control women, notably in a lack of ciliated cells. This study has the limitation that it evaluated

ingestion of doses well above the clinical dose; the relevance of the observations to the approved LNG-EC

product, which is a much lower dose, is uncertain.

e. Hapangama and colleagues (2001) conducted a study in 12 healthy women with regular cycles. This

study intended to examine the effect of LNG at different times during the menstrual cycle. The study

utilized urinary monitoring of estrone-3-glucuronide to detect ovulation with no transvaginal ultrasounds

obtained. The method used in this study does not accurately determine the specific time during the cycle;

that is, the method cannot accurately determine if the women were pre-, peri-, or post-ovulatory. Thus,

the results of this study are confounded, and the study is not clinically interpretable.

2. Effects of Levonorgestrel on Sperm Function

After unprotected intercourse, sperm can be found in the endocervix within 90 seconds. Based on

literature review conducted by the multidisciplinary review team, it is unlikely that LNG, when used for

emergency contraception, directly affects sperm function.

Yeung and colleagues (2002) reported data from an in vitro study that evaluated the effects of LNG on

sperm function. The authors obtained semen samples from men attending infertility clinics and incubated

the samples with either control or LNG in concentrations of 1, 10, or 100 ng/mL. The authors chose their

doses based on prior studies of pharmacokinetic data on serum concentrations of LNG, which reached a

Tmax of 5-10 ng/mL at 2 hours post-dose. The LNG concentrations to which sperm were exposed in the

Yeung study appear to be considerably higher than those that would occur with use of an emergency

contraceptive. The authors noted a decrease in some parameters of sperm motility (curvilinear velocity

and straight-line velocity) at the 10 and 100 ng/mL LNG concentrations, and a decrease in average path

velocity at the 100 ng/mL concentration. As noted above, these observations were at concentrations likely

to be many-fold above pharmacologically relevant concentrations for the LNG-EC. Several other sperm

motility measures (linearity, head beat cross frequency, and amplitude of lateral head displacement) were

the same for control and LNG at all concentrations. Limitations of the study include the high concentration

of LNG used; the unknown correlation between serum and uterine/fallopian LNG concentration; and the

fact that specimens were obtained from men seeking infertility care, who may not have had normal sperm

function at baseline.

To summarize the information on LNG-EC MoA that was available at the time of the 2003 review, the

studies by Landgren, Durand, and Marions, and their colleagues, showed that LNG suppressed the

midcycle LH surge and suppressed ovulation. In addition, all three studies evaluated endometrial changes

in either morphology and/or endometrial biomarkers and reported no meaningful changes. Finally, the

study by Yeung and colleagues showed no evident effect on sperm function. This evidence suggests that

Page 10 of 27

Reference ID: 5100570

the mechanism by which LNG prevents pregnancy is through suppression of ovulation, and not through

changes in endometrial function/receptivity or changes in sperm function leading to impaired fertilization.

However, the limited number of studies available at the time of the 2003 review of the data precluded a

firm conclusion regarding whether other mechanisms were unlikely to contribute.

D. Information on Mechanism of Action that Became Available After the 2003 Review

As noted in the Background (Section II) above, the decision to take the unusual step of inclusion of MoA

in the nonprescription labeling for LNG followed discussion of MoA at the 2003 Advisory Committee (AC)

meeting (note that the labeling for MoA was based upon the assessment of the literature leading up to

the AC meeting). Subsequent to that meeting, additional research (nonclinical, clinical pharmacology, and

clinical) has been published relevant to LNG-EC MoA. Some of these publications became available during

the time period between the 2003 review and the 2006 nonprescription approval of LNG-EC; others were

published later.

1. Nonclinical Studies on Suppression of Ovulation with Levonorgestrel

There are two nonclinical studies, one in rats (Muller 2003) and one in monkeys (Ortiz 2004), that address

LNG effects on ovulation. Full study reports were not available for either publication, but a review was

conducted on peer-reviewed publications of summary information.

a. In the rat study, Muller and colleagues dosed animals to determine if LNG had an effect on

ovulation, fertilization, or implantation. None of the animals were dosed after implantation. First,

rats were dosed subcutaneously (SC) with LNG (50 μg/kg), with Group 1 dosed twice per day, for

2 days early in the menstrual cycle (at diestrus and proestrus); Group 2 dosed once during

diestrus; and Group 3 dosed once at proestrus. These animals were terminated at estrus to

determine the percentage of animals ovulating, and the mean number of released eggs per

ovulating animal. Two SC LNG administrations early in the menstrual cycle (at diestrus and

proestrus) completely prevented ovulation in the rat, but this effect decreased with single doses,

and as dosing approached estrus.

Following the initial rat study, which evaluated Groups 1-3 (as described above), several additional

groups of rats (groups 4-11) were mated and dosed with single doses of LNG during proestrus or

estrus to determine if LNG affected fertilization or implantation. If mating occurred, the day of

estrus was designated Day 1 of pregnancy (P1). The following eight different groups were tested

on P2 to assess the number of fertilized eggs, or P12 to assess the number of implantations:

• Dosing before mating and ovulation: females were dosed during proestrus and then mated

the same day (Group 4), or overnight (Group 5)

• Dosing after mating and before ovulation: females were mated during proestrus, then dosed

with LNG once that day (Group 6), or twice (the day of mating and 12 hours later; Group 7)

• Dosing after ovulation and before mating: females were injected with LNG on the day of

estrus and immediately mated (Group 8)

• Dosing after ovulation and mating: females were mated on the day of estrus, and injected

immediately with LNG (Group 9)

Page 11 of 27

Reference ID: 5100570

• Dosing after fertilization upon implantation: proestrus rats were mated overnight, then

injected with LNG once on P1 and again on P2 (Group 10), or on P3 and again on P4 (Group

11)

All groups had matching vehicle-injected controls, and data from LNG-injected females were

compared to controls. In all groups, there were no notable observed differences in implantation

rates between the control and LNG-treated animals. The primary mode of action of LNG appeared

to be ovulation suppression, based on the absence of ovulation in rats treated with LNG at

proestrus and diestrus (Groups 1-3), and there was no notable effect in the LNG-treated animals

that were mated (but dosed after proestrus), compared to concurrent control animals. LNG does

not appear to exert any postimplantation effects on the rat, nor did it appear to affect

implantation of fertilized eggs. It is not clear how systemic exposure in the rat studies compares

to those expected in humans who take LNG-EC, because systemic exposure was not measured in

rats.

b. The studies reported by Ortiz in monkeys cannot directly inform whether LNG affects implantation

of fertilized eggs or causes loss of fertilized eggs. In the first study, monkeys were mated to

determine if LNG affected pregnancy rates. In the second study, monkeys were dosed with LNG

to determine if LNG suppressed ovulation in the absence of mating. Limitations to both studies

are described below:

i. In a monkey mating study, animals were treated with either one dose of 0.75 mg LNG or

two doses of 0.75 LNG (1.5 mg LNG total), subcutaneously or orally after confirmed

mating. The authors conclude that overall, the pregnancy rate was identical between

LNG-treated and vehicle-treated controls. The authors state, “In 41 [out of] 48 [animals],

mating took place between 2 days before ovulation and the day of ovulation.” In only

seven cycles did mating precede ovulation by greater than 2 days (3 – 5 days). Therefore,

the number of monkeys mated early relative to ovulation is limited making it difficult to

interpret how timing of LNG treatment affects pregnancy rate. Seven animals appear to

have been treated with LNG after fertilization likely occurred; however, the pregnancy

rate in these animals was not different from those treated prior to fertilization.

Pharmacokinetic data were not collected in this study to allow for comparison to human

exposures. There were limited pharmacokinetic data collected for one of the doses (0.75

mg oral or subcutaneous) employed in monkeys in a separate bioavailability assessment

(see item iii below).

ii. A monkey study was also conducted to determine if LNG affected ovulation, but

monkeys in the study were not mated. Because animals were not mated, this study

cannot inform how LNG affects fertilization or implantation. In the monkey ovulation

study, six female monkeys were dosed with two subcutaneous injections of LNG (0.75

mg), 12 hours apart. Follicle development and ovulation were inhibited in monkeys

dosed when the leading follicle was less than 5 mm in size. Ovulation occurred when the

leading follicle was greater than 5 mm at the time of LNG treatment.

Page 12 of 27

Reference ID: 5100570

iii. It is not clear how systemic exposure compares between monkeys and humans in either

the mating or ovulation studies (described directly above). In a bioavailability monkey

study, three monkeys received 0.75 mg LNG (oral or subcutaneous), and the total

systemic exposure (area under the curve) of LNG was not quantified.

In conclusion, the rat data from Muller (2003) show that subcutaneous administration with LNG (50 μg/kg,

twice per day, for 2 days) early in the menstrual cycle inhibits ovulation and does not affect implantation.

The monkey studies from Ortiz (2004) also show LNG inhibits ovulation when administered early in the

cycle but the data did not show that LNG decreased pregnancy rates when administered close to

fertilization. These papers did not address the impact of LNG when dosed after implantation. The data,

however, are hard to extrapolate to consumer use of LNG-EC because it is not clear how the dosing

regimen in the animals compares to clinically relevant exposures of LNG, due to the lack of adequate

pharmacokinetic data in either the rat or monkey study.

2. Clinical Studies Evaluating the Effect of Levonorgestrel Emergency Contraception on Ovulation

a. Marions and colleagues (2004) published another human clinical study further examining the

postovulatory effects of LNG-EC; LNG again delayed or stopped ovulation, and was associated with a delay

in rise of pregnanediol in the time following the expected LH peak, and a lower estrone peak at the time

of expected LH peak. In this study, after a control cycle to determine the expected day of LH peak, 7

women were administered 2 doses of LNG 0.75 mg given 12 hours apart, 2 days prior to their expected

LH peak. Daily transvaginal ultrasounds were performed, as were daily urine assays for LH, pregnanediol

glucuronate, and estrone glucuronate. Follicle rupture did not occur in any participants. Similar to findings

in other studies, the LH peak did not occur or was attenuated and delayed in LNG cycles vs control cycles.

Pregnanediol levels rose more slowly in the LNG-exposed cycle than in the control cycle, reaching peak on

Day +9 (with Day 0 having been the expected LH peak day, 2 days after administration of LNG) as opposed

to Day +7 for the control cycle, but had a similar area under the concentration curve (AUC). Estrone levels

peaked on Day 0 for both the LNG and control cycles, with a lower Cmax for the LNG cycle. The authors

state that the latter two observations suggest that LNG-EC may have some effect in the luteal phase.

However, because ovulation did not occur, this postulation is moot, because pregnancy could not have

occurred anyway. It is unclear whether one can attribute an effect of LNG, or whether the observations

may be the natural findings of reduced luteal activity in an anovulatory cycle. Mean cycle length shortened

by 4 days with LNG vs control. Also, the study examined similar parameters for mifepristone, with each

participant receiving mifepristone in one cycle, then having a washout cycle with no drug given, then

receiving LNG in a third cycle. This presents a small possibility of confounding from mifepristone’s effects

in the prior cycle in the same participant.

b. Tirelli and colleagues (2008) published a study in which women with regular menstrual cycles were

enrolled within 72 hours of unprotected intercourse and received LNG 0.75 mg x 2 doses 12 hours apart.

The study included a subgroup of eight women who received LNG during Days 11-13 of the cycle, and who

underwent detailed evaluation with transvaginal ultrasound every 3 days starting on the day of

recruitment before LNG intake until follicular rupture was documented. This subgroup of eight women

also underwent endocrine blood sampling (LH, FSH, estradiol, and progesterone).

In seven of eight women, an anovulatory cycle was documented with no LH or FSH peak. In addition,

menstrual cycle length was significantly shortened. One of the eight women, who had received LNG on

Day 13 of the cycle, had an ovulatory cycle, and an LH surge. This study demonstrates that follicular phase

Page 13 of 27

Reference ID: 5100570

LNG administration generally (in seven of eight women in this study) results in suppression of ovulation

and consequent shortening of the menstrual cycle. This supports the mechanism of LNG as suppressing

ovulation, which thereby prevents pregnancy.

c. Croxatto and colleagues (2004) published a study evaluating the effect of LNG-EC on ovulation and

ovulatory function. Each subject was studied during a placebo- and a drug-treated (LNG 0.75 mg x 2 doses

at 12-hour interval) cycle. Three groups were studied: one group received drug when the leading follicle

was 12-14 mm (18 women), one group when the leading follicle was 15-17 mm (22 women), and one

group when the leading follicle was ≥ 18 mm (18 women). All women were evaluated with repeated

transvaginal ultrasounds, and venous blood sampling for endocrine measurements (LH, FSH, estradiol,

and progesterone).

In women with the leading follicle 12-14 mm, no rupture of the follicle was documented within 5 days of

dose administration in 83% of women with LNG administration vs 56% of women in the placebo cycle. In

comparison, ovulation was not suppressed relative to placebo in women with the leading follicle ≥ 18 mm

or when the follicle was 15-17 mm. However, with LNG administration, LNG markedly suppressed the LH

peak and/or the rise in progesterone in 94% of women with leading follicle of 12-14 mm (vs 61% with

placebo), 91% of women with their leading follicle 15-17 mm (vs 45% with placebo), and in 47% of women

with leading follicle of ≥18 mm (vs 13% with placebo).

This study shows that LNG suppresses or delays ovulation (beyond 5 days) and/or suppresses ovulatory

function in a high proportion of women with leading follicle sizes under 18 mm, and less consistently in

women with their leading follicle ≥18 mm. In women with a smaller leading follicle size, the primary

mechanism is prevention or delay (beyond 5 days) of follicular rupture, while when the leading follicle is

larger, suppression of ovulatory function (i.e., prevention of the midcycle LH surge) is more prominent as

a mechanism. Both prevention of follicular rupture and suppression of ovulatory function may contribute

to prevention of pregnancy.

d. Okewole and colleagues (2007) studied 14 women with regular menstrual cycles; all were advised to

use condoms during the study to prevent pregnancy. Women were followed from 5 days prior to expected

ovulation, and daily blood samples for estradiol, FSH, and LH were collected. Women were followed

through 2 cycles: a pretreatment cycle and a cycle in which all women received LNG 1.5 mg (8 women

randomly assigned to receive drug 3 days prior to expected ovulation and 6 women to receive drug 1 day

prior to expected ovulation). In the pretreatment cycle, all women had ovulatory cycles as shown by the

LH peak. In women receiving LNG 3 days prior to expected ovulation, mean cycle length was prolonged

and LH and FSH peaks were delayed by 4-5 days relative to the control cycle; in comparison, in 6 women

who received LNG 1 day prior to expected ovulation, mean cycle length was shortened and LH/FSH peak

was only delayed by 1 day relative to the control cycle. The authors concluded that the primary

mechanism is to delay the events triggering ovulation when administered sufficiently in advance of

ovulation.

3. Studies Evaluating the Effectiveness of Levonorgestrel Based Upon Timing in Cycle

a. Novikova and colleagues (2007) published a clinical study showing that LNG 1.5 mg was effective in

preventing pregnancy when taken before ovulation, but when LNG was taken after ovulation, the

pregnancy rate was consistent with expected rates if LNG had not been administered.

Page 14 of 27

Reference ID: 5100570

The study was done in 99 women who presented requesting emergency contraception. While in clinic for

their initial consultation, they had a blood sample taken for LH, estradiol, and progesterone; then they

took 1.5 mg LNG. The serum hormone samples were used to estimate the day of cycle on which the

woman took LNG. Follow-up occurred 4-6 weeks later by phone to determine pregnancy status; those

who became pregnant underwent ultrasound for confirmation.

Of the 99 women studied, 51 had their unprotected intercourse during the “fertile window”, i.e., between

days -5 to 0 of the cycle, with Day 0 being the day of ovulation. Using data from Wilcox (1998), Novikova

and colleagues estimated that eight pregnancies should have occurred among these women if no EC was

used. Three pregnancies occurred: one in a woman who had intercourse on Day -1 of her cycle, and two

in women who had intercourse on Day 0 of their cycle. Notably, these three women took LNG

approximately 2 days after ovulation.

The authors conclude that these results support the conclusion that LNG-EC has little or no effect on

postovulation events but acknowledge the limitation of a study in which only three pregnancies occurred.

It is noted that this study was conducted outside the US, while the Wilcox data upon which expected

pregnancy rates were based came from a US study. The number of expected pregnancies would not,

however, have been expected to differ meaningfully.

b. Noé and colleagues (2011) published a clinical trial on the contraceptive efficacy of LNG given before

or after ovulation; the trial showed that LNG-EC given before ovulation was highly effective, but when

given after ovulation, was ineffective and had no effect on the expected occurrence of pregnancy.

This study enrolled 450 women who presented to a clinic and requested emergency contraception. All

were given 1.5 mg LNG-EC, and had blood drawn for LH, progesterone, and estradiol (in order to

determine the day of the cycle relative to ovulation). Transvaginal ultrasound was also performed to

assess the presence and size of a leading follicle or corpus luteum. Women who had a follicle ≥12 mm or

a recent corpus luteum had follow-up ultrasounds. Women were later contacted to determine if

menstruation occurred; if it had not, women were followed weekly for urinary human chorionic

gonadotropin until pregnancy or menstruation was confirmed. Of most interest were 393 women who

had data sufficient to determine whether they had intercourse on 1 of the 6 fertile days of the cycle (the

day of ovulation and the 5 days prior), as described by Wilcox (1998). Women who had intercourse during

those days (total N = 148) were divided into two groups; those who took LNG before (days -5 to -1, n=103)

or after (day 0 or after, n=45) ovulation.

Among those who received LNG in the fertile window but were preovulatory, there were no observed

pregnancies, while 16 were expected. Among those who took LNG at ovulation or after in the fertile

window, 8.7 pregnancies were expected and 8 were observed. All pregnancies were intrauterine (i.e.,

none were ectopic); five women delivered healthy term babies vaginally. Three pregnancies ended in

abortion (spontaneous versus elective not stated).

The authors conclude that reproductive processes subsequent to ovulation were not affected by LNG-EC.

As with the Novikova study, it is noted that this study was conducted outside the US, while the Wilcox

data upon which expected pregnancy rates were based came from a US study. The number of expected

pregnancies would not, however, have been likely to differ meaningfully.

Page 15 of 27

Reference ID: 5100570

In summary, the studies by Noé and by Novikova both show that LNG is effective when given pre-ovulatory

and not when administered after ovulation. These results support the conclusion that the mechanism of

LNG is to suppress ovulation and that LNG either does not alter postovulation processes (i.e., fertilization

or implantation) or that any effect of LNG on postovulation processes is insufficient to contribute to the

mechanism of drug action to prevent pregnancy.

4. Studies Assessing Levonorgestrel Effect on Endometrial Receptivity or Sperm Function

a. Durand and colleagues (2005) published a clinical study using specimens collected in their previously

described 2001 study; this new study examined one additional exploratory endometrial marker (serum

glycodelin-A). In the study, serum glycodelin-A concentrations rose earlier and endometrial expression of

glycodelin was lower among women who took LNG prior to the LH surge than in control cycles with no

LNG. This effect was not seen when women took LNG on the day of the LH surge. Glycodelin-A is an

endometrial glycoprotein that is postulated to have many possible effects. In an immunologic sense, the

blastocyst and subsequent embryo are partial allografts, and thus at risk of natural immune rejection. One

of glycodelin-A’s roles may be as a suppressor of Natural Killer cells that would otherwise naturally attack

and destroy the blastocyst; another role may be in adhesion of the blastocyst to endometrial epithelial

cells (Uchida 2013). Thus, glycodelin-A may play a role in postovulatory events, although it has not been

validated as a biomarker. Glycodelin-A is not usually expressed in the endometrium until the last week of

the luteal phase (Brown 2000). This timing is potentially important, because in vitro, glycodelin-A may

inhibit binding of human spermatozoa to the zona pellucida of the oocyte (an event relevant to

fertilization), but if glycodelin-A is not normally expressed until later in the luteal phase (after fertilization

would be expected to occur), this potential antifertilization effect would not generally be present

naturally. As noted above, serum glycodelin-A concentrations rose earlier among women who took LNG

prior to the LH surge than among women with control cycles with no LNG. This effect was not seen when

women took LNG on the day of the LH surge. In the endometrial biopsies done on Day +9, staining intensity

for glycodelin-A was less intense for women who took LNG prior to the LH surge than it was in control

cycles and when women took LNG at the time of the LH surge. The observation of low expression of

glycodelin-A in the biopsy specimens while at the same time seeing an early increase in serum glycodelin

concentration is unexplained; the authors postulate that it may be due to variability in glycodelin-A

expression in different parts of the endometrium.

The authors conclude that LNG may affect glycodelin-A in two phases of the cycle: the phase when

fertilization occurs, and the phase of endometrial receptivity. Possible limitations of this study include

that it is unclear how long specimens were stored prior to testing; because these specimens came from a

study published years earlier, storage may have been for years, and the stability of glycodelin-A under the

storage conditions of the study is unknown. Also, as mentioned earlier, glycodelin-A has not been

validated as a biomarker. The practical relevance of a glycodelin-A effect seen only in the absence of

ovulation is unclear, as pregnancy could not have occurred in this circumstance anyway. It is not clear

what the normal expected profile of glycodelin-A would be in an anovulatory cycle with consequent lack

of luteal activity.

b. Lalitkumar and colleagues (2007) published in vitro data showing that levonorgestrel, compared to

control, did not inhibit human blastocyst attachment to an endometrial three-dimensional cell culture

model. However, mifepristone (a drug approved for medical termination of early pregnancy) did inhibit

Page 16 of 27

Reference ID: 5100570

blastocyst attachment. Thus, in an in vitro model with both a positive and negative control, LNG did not

affect blastocyst attachment.

c. do Nascimento and colleagues (2007) published a clinical study conducted in surgically sterilized normal

women; the study showed little to no effect of LNG on sperm function or endometrial glycodelin-A levels.

Note that prior to this work, the same group had published the results of several in vitro studies that

showed no effect of LNG on the fertilizing capacity of spermatozoa (Bahamondes 2003, Brito 2005,

Munuce 2005).

In this clinical study, they recruited fertile, surgically sterilized women (ages 30-41 years). Participants

were divided into four groups:

Group I: LNG 1.5 mg or placebo (pbo) taken 12 hours after coitus; uterine flushing performed 24 hours

after LNG/pbo intake

Group II: LNG 1.5 mg or pbo taken 12 hours after coitus; uterine flushing performed 48 hours after

LNG/pbo intake

Group III: LNG 1.5 mg or pbo taken 36 hours after coitus; uterine flushing performed 24 hours after

LNG/pbo intake

Group IV: LNG 1.5 mg or pbo taken 24 hours after vaginal artificial insemination; uterine flushing

performed 24 hours after LNG/pbo intake

The women underwent daily intravaginal ultrasound and daily serum progesterone measurements.

Couples for groups I-III were asked to abstain from intercourse for 5 days, then have intercourse on the

evening of the day when the greatest ultrasound-detected follicular diameter showed positive correlation

with cervical mucus characteristics (spinnbarkheit >10 cm, crystallization >2+). Group IV participants had

artificial insemination the next morning. In addition to the administration of LNG/pbo and uterine flushing

for sperm retrieval as described above, cervical mucus was also obtained, and endometrial biopsies were

obtained after uterine flushing. Endometrial tissue was assayed for glycodelin-A. Uterine flushings were

examined for quantification of spermatozoa present. Spermatozoa from uterine flushings and cervical

mucus were assessed microscopically for evidence of the acrosomal reaction (the attachment and

dissolution of the acrosome, a structure at the tip of a spermatozoan; this reaction is necessary for

penetration of the ovum and thus fertilization).

Measures of cervical mucus quality, as measured by the World Health Organization Laboratory Manual

(World Health Organization 1999) were not affected by LNG treatment. This observation is of note,

because previous authors had hypothesized that LNG-EC might affect cervical mucus, because progestin-

only routine daily oral contraceptive tablets using a different progestin, such as norethindrone, do appear

to affect cervical mucus (Moghissi 1973).

Number of recovered spermatozoa did not differ between LNG and pbo groups, either at 24 or 48 hours

after coitus. Viable, motile spermatozoa were obtained from the cervical and uterine cavity at Hours 36,

48, and 60 after coitus. There were no differences between LNG and pbo groups in numbers of observed

acrosomal reactions, whether observed in spermatozoa from uterine washings or cervical mucus,

regardless of whether specimens were obtained at 24 or 48 hours.

Page 17 of 27

Reference ID: 5100570

Treatment with LNG was not associated with changes in expression (immunostaining intensity) of

glycodelin-A from endometrial biopsy specimens. This differed from observations in the 2005 Durand

study discussed above.

Potential limitations of the do Nascimento study included the fact that although a total of 33 women

enrolled with a plan for 2 cycles each, 18 women were discontinued due to various problems, resulting in

a total of only 48 experiments being completed. Problems included inadequate cervical mucus (n = 11

cycles), inadequate postcoital mucus test by World Health Organization (WHO 1999) criteria (12 cycles),

premature ovulation (9 cycles), and “problems with the couple” (6 cycles).

The authors of the do Nascimento study concluded that LNG did not affect acrosomal reaction status,

endometrial expression of glycodelin-A, quality of cervical mucus, or penetration of spermatozoa into the

uterine cavity. This points away from postovulatory MoAs.

d. Palomino and colleagues (2010) published a clinical study that showed that LNG 1.5 mg administered

on the day of the LH surge did not alter multiple putative markers of endometrial receptivity in

endometrial biopsies from 12 women, when compared to untreated women. Unchanged markers

included progesterone, L-selectin ligand, integrin ανβ3, and glycodelin-A. In 3 of 12 biopsies from LNG-

treated women, areas of glandular atrophy and stromal decidualization were noted, which are common

after progestin administration (Deligdisch 2000). The authors’ overall conclusion was that the MoA of

LNG-EC given at the time of LH surge does not involve impairment of expression of progesterone receptors

or experimental biomarkers of endometrial receptivity.

e. Meng and colleagues (2009) published in vitro data using a three-dimensional human endometrial cell

culture model similar to that used by Lalitkumar as described above. This in vitro model expressed

numerous putative markers of endometrial receptivity, including alpha and beta receptors for estrogen

and progesterone; VEGF; LIF; interleukin-1β; cyclooxygenase-2 (COX-2); integrin ανβ3; and surface

molecule MUC1. LNG exposure, compared to control, did not affect expression of these markers.

Mifepristone exposure, however, was associated with upregulation of epithelial estrogen receptor-β and

progesterone receptor-β; and with downregulation of stromal VEGF, MUC1, and integrin ανβ3. Thus, an

in vitro experiment with both a positive and negative control did not show an effect of LNG on multiple

putative markers of endometrial receptivity to blastocyst implantation. A limitation of this study is that

these markers of endometrial receptivity, while extensively studied, are not validated biomarkers. The in

vitro model, while robust in that it shows expression of many receptors and markers associated with the

state of endometrial receptivity, may not be fully translatable to the human in vivo state.

f. Vargas and colleagues (2012) studied the direct effects of LNG-EC on the endometrium, focused on

changes in gene expression. Eight women were administered LNG 1.5 mg or placebo after ovulation, so

as to avoid indirect effects of the suppression of ovulation on the endometrium. The results showed, as

expected, no effects on cycle length or bleeding, nor on endometrial morphology or progesterone levels.

A wide range of gene expression was assessed and only small changes were observed in four genes. The

authors concluded that LNG caused minimal changes in transcript levels and that the changes observed

were unlikely to interfere with endometrial receptivity.

g. Brito and colleagues (2005), in an in vitro sperm study, evaluated the acrosome reaction in semen

samples from fertile men. Concentrations relevant to systemic exposure with LNG-EC were used, and no

effects on the acrosome reaction were observed.

Page 18 of 27

Reference ID: 5100570

h. Hermanny and colleagues (2012) comprehensively evaluated sperm function in the fallopian tube after

LNG administration. In this study, fallopian tube segments removed from women undergoing surgical

sterilization were removed and perfused with a suspension containing motile spermatozoa, with or

without LNG. The number of motile spermatozoa recovered was similar in the tubal isthmus and in the

ampulla in the LNG and control groups. In addition, there were no statistically significant differences in

the number of spermatozoa recovered after flushing, the number of sperm adhering to the oviductal

epithelium, or the acrosome reaction rate using a fluorescent probe.

E. Integrated Summary of Current State of Knowledge Regarding Mechanism of Action of Levonorgestrel

Emergency Contraception

The events of the reproductive cycle are highly complex. Since unprotected intercourse may occur at

various stages of the cycle, LNG postcoital dosing may occur either in the follicular phase, right at the time

of the expected midcycle surge in LH, or after ovulation. The totality of the evidence indicates that LNG-

EC is effective when administered during the follicular phase of the menstrual cycle, before the LH surge.

Available clinical evidence suggests that the constellation of events leading to the LH surge and follicular

rupture is interrupted when LNG is administered during the follicular phase, and that this interruption of

pre-ovulatory events is the main mechanism by which LNG prevents pregnancy.

As discussed above, studies available prior to the 2003 review, including clinical studies by Landgren

(1989), Durand (2001), and Marions (2002), showed that LNG suppressed the midcycle LH surge and

suppressed ovulation. Studies subsequently available provide additional support for this mechanism.

Nonclinical studies, by Muller (2003) in rats and by Ortiz (2004) in monkeys, both support the conclusion

that LNG inhibits ovulation, although in monkeys, inhibition was noted predominantly when LNG was

administered before follicle growth exceeded 5 mm in size. Clinical studies published after the 2003

review, by Marions (2004), Okewole (2007), and Tirelli (2008), also demonstrated that LNG suppresses

the LH surge and ovulation.

Studies by Novikova (2007) and Noé (2011) provide the strongest clinical data that suppression of the

midcycle LH surge and interference with ovulation are the main mechanisms by which LNG-EC prevents

pregnancy. These studies showed that after intercourse occurred during the “fertile window”, LNG

administered during the follicular phase, prior to ovulation (through Day -1 prior to ovulation), was

effective in preventing the occurrence of pregnancy, but when administered from Day 0 (ovulation) and

beyond, the expected pregnancy rate was not reduced. In other words, LNG-EC is shown to be effective

only if administered in the follicular phase of the menstrual cycle prior to ovulation when it is possible to

suppress or delay follicle release, impair follicle development, and/or suppress the midcycle LH surge.

These observations also strongly support the conclusion that direct effects on postovulatory processes

are not relevant to the mechanism by which LNG-EC prevents pregnancy since LNG-EC is not effective

when given after ovulation.

Further support for these conclusions comes from the studies that investigated the direct effects of LNG-

EC on the endometrium, implantation, or sperm function/fertilization. As discussed above, the earlier

(before the 2003 review) studies by Durand (2001) and Marions (2002) did not find meaningful changes

in endometrial morphology or endometrial exploratory biomarkers. Landgren (1989) did report a

decreased number of endometrial glands when LNG is administered in the follicular phase, but not when

Page 19 of 27

Reference ID: 5100570

it was administered in the luteal phase, suggesting the possibility of an indirect effect through suppression

of ovulation rather than a direct effect on the endometrium. In studies reported subsequent to the 2003

review, Durand (2005) found reduction in an endometrial biomarker, glycodelin-A, when LNG was

administered during the follicular phase of the cycle, and not when LNG was administered at the time of

ovulation, consistent with the report by Landgren, suggesting that the changes may reflect an indirect

effect of a reduced LH surge. In contrast to the findings of Durand, the study by do Nascimento (2007)

did not find any effects on endometrial glycodelin-A, nor on cervical mucus quality. Similarly, the study

by Palomino (2010) also did not find changes in exploratory endometrial biomarkers (including glycodelin-

A) in women administered LNG, although some morphological changes (e.g., glandular atrophy) were

observed. The study by Meng (2009), using an in vitro system, also found no changes in any of a range of

exploratory endometrial biomarkers assessed. The study by Vargas (2012) found no meaningful changes

in endometrial morphology or gene expression when LNG was administered after ovulation, supporting

the conclusion that there are no direct effects of LNG on the endometrium and hence endometrial

receptivity is unlikely to be reduced. Finally, the study in rats by Muller (2003) showed that LNG did not

alter implantation rates but suppressed ovulation. These studies show that endometrial morphological

changes with LNG administration are variably seen and such changes, if they occur, likely reflect

suppression of ovulation rather than a direct effect on the endometrium. Studies of exploratory

endometrial biomarkers and gene expression generally show no notable effects of LNG, particularly when

administered after ovulation. Overall, these results support the conclusion that changes in the

endometrium are variable, limited, and if seen, may reflect the effects of LNG on ovulation, and are

unlikely to contribute to the mechanism of action of LNG-EC. The observation that postfollicular-period

administration of LNG is not shown to be effective in preventing pregnancy strongly supports the lack of

the drug’s direct effects on postovulation processes involved in conception.

Several studies evaluated sperm function (Yeung 2002, Brito 2005, do Nascimento 2007, Hermanny 2007)

and found no evidence of reduced spermatozoa number recovered from the cervical or uterine cavity;

reduced sperm function; alterations in the acrosomal reaction; or inhibition of fertilization. These studies

support the conclusion that alterations in sperm number or function, or fertilization, are unlikely to be a

relevant mechanism of LNG-EC prevention of pregnancy.

The conclusions from this review are that there is strong evidence that LNG-EC acts by actions on

ovulation, including inhibiting or delaying follicle rupture and suppressing the LH surge; there is no

substantive evidence that other mechanisms such as direct alteration of endometrial receptivity leading

to reduced implantation or alterations in sperm function or fertilization contribute to the mechanism of

action.

It is of note that the evidence prior to the 2003 review already pointed to a predominant mechanism of

suppression of ovulation; however, at that time, the evidence was as yet insufficient to conclude firmly

that other mechanisms were unlikely to be contributory. Evidence subsequent to the 2003 review

provides strong evidence that mechanisms beyond changes in ovulation are noncontributory to the

mechanism of LNG-EC. Novikova (2007) and Noé (2011) provided especially persuasive evidence showing

that dosing after ovulation does not reduce the occurrence of pregnancy. Additional studies provided

evidence that LNG-EC suppresses ovulation and/or ovulatory function, and that there is no substantive

evidence of direct changes in the endometrium or reductions in sperm numbers or function with LNG-EC.

In summary, LNG-EC is shown to act primarily by interfering with ovulation (delay or inhibition of follicle

Page 20 of 27

Reference ID: 5100570

rupture and/or the midcycle LH and related hormonal changes) and not to act through direct effects on

fertilization or on the endometrium to impede implantation.

In the Executive Summary in Section I above, summaries of answers to the following key questions were

provided. These answers are expanded upon in the following paragraphs, with supporting citations:

1. Does levonorgestrel emergency contraception inhibit ovulation?

Clinical data are strong that LNG-EC, when administered prior to the LH surge, inhibits ovulation (Durand

2001; Marions 2002; Croxatto 2004; Okewole 2007; Novikova 2007; Tirelli 2008; and Noé 2011).

2. Does levonorgestrel emergency contraception affect sperm function?

In vitro and human in vivo data do not support a significant effect of LNG on cervical mucus quality; sperm

quantity in the genital tract; sperm motility; or acrosomal reaction when administered at doses similar to

that used in EC (Yeung 2002, Bahamondes 2003, Brito 2005, Munuce 2005, do Nascimento 2007).

3. After ovulation occurs, does levonorgestrel emergency contraception affect endometrial receptivity or

implantation?

In vitro data, and human clinical data, show that LNG-EC does not affect exploratory markers of

endometrial receptivity after ovulation occurs (Lalitkumar 2007; Meng 2009; do Nascimento 2007;

Palomino 2010; Durand 2005). In these research studies, a wide range of endometrial characteristics and

markers was evaluated, and overall, essentially no changes were consistently observed. It should be noted

that such markers are not validated biomarkers, and the multidisciplinary review team did not find

published data showing that changes in endometrial receptivity biomarkers translate into a clinical effect

on blastocyst implantation. The totality of the evidence indicates that LNG-EC does not alter the

endometrium in a clinically meaningful way to prevent implantation. The findings of Noé that there is no

evident reduction in the likelihood of pregnancy when LNG is administered to women postovulation are

consistent with the histological and biochemical findings of the aforementioned studies and support the

conclusion that LNG-EC is ineffective when administered postovulation.

4. Does levonorgestrel emergency contraception affect pregnancy rate after ovulation?

When LNG-EC is administered after ovulation occurs, the rate of pregnancy is what one would expect if

LNG-EC had not been administered (Noé 2011, Novikova 2007). This is arguably the most important aspect

of the evidence.

As noted earlier, the scientific and clinical concept of the time period of pregnancy is generally consistent

with the example found in the Code of Federal Regulations (CFR) under 45 CFR 46.202, which reads:

“Pregnancy encompasses the period of time from implantation until delivery.” LNG-EC prevents

pregnancy by acting on ovulation, which occurs before implantation, and because data do not support

that LNG-EC affects implantation, LNG-EC does not terminate pregnancy.

Page 21 of 27

Reference ID: 5100570

V. Decision on Action to be Taken

In response to the applicant’s submission of this labeling supplement, FDA reviewed currently available

scientific evidence regarding the mechanism of action for LNG-EC as described above. Based on that

review, the multidisciplinary review team agrees that it is appropriate to update nonprescription LNG-EC

labeling information relevant to mechanism of action.

As explained in this review, LNG-EC prevents pregnancy - it does not terminate pregnancy. For people

who are capable of becoming pregnant, LNG-EC is effective in preventing pregnancy when taken within 3

days of unprotected intercourse and before ovulation. LNG-EC is not effective if pregnancy has already

occurred.

The current science supports an effect on inhibiting or delaying ovulation and the midcycle hormonal

changes, and the evidence also supports the conclusion that there is no direct effect on postovulatory

processes, such as fertilization or implantation.

Therefore, based on careful consideration of the applicant’s labeling supplement, as amended, and

additional scientific evidence, the labeling of LNG-EC will be updated. In the Drug Facts label, some (but

not all) information relevant to mechanism of action will be edited. In the Consumer Information Leaflet,

information relevant to mechanism of action will be updated to reflect the current state of knowledge.

The following section details the changes to be made.

VI. Labeling

The multidisciplinary review team proposes the following labeling changes.

In the Drug Facts label for Plan B One-Step, in the “Other Information” section, remove the following

statement:

“ * this product works mainly by preventing ovulation (egg release). It may also prevent

fertilization of a released egg (joining of sperm and egg) or attachment of a fertilized egg to the

uterus (implantation).”

In the Consumer Information Leaflet (CIL), under the heading “How does Plan B One-Step® work?”, make

the following edits; additions are underlined, and removed statements are struck through:

“Plan B One-Step® works before release of an egg from the ovary. As a result, Plan B One-Step® usually

stops or delays release of the egg from the ovary. Plan B One-Step® is one tablet with levonorgestrel, a

hormone that has been used in many birth control pills for several decades. Plan B One Step® that

contains a higher dose of levonorgestrel than birth control pills but and works in a similar way to

prevent pregnancy. It works mainly by stopping the release of an egg from the ovary. It is possible that

Plan B One Step may also work by preventing fertilization of an egg (the uniting of sperm with the egg)

or by preventing attachment (implantation) to the uterus (womb).”

Page 22 of 27

Reference ID: 5100570

Without the above editorial markings, that CIL section will read as follows:

“Plan B One-Step® works before release of an egg from the ovary. As a result, Plan B One-Step®

usually stops or delays the release of an egg from the ovary. Plan B One-Step® is one tablet that

contains a higher dose of levonorgestrel than birth control pills and works in a similar way to

prevent pregnancy.”

Of note, there are several other sections of the DFL and CIL that are relevant to mechanism of action, and

these sections will remain unchanged.

The DFL will retain the following language that is relevant to mechanism of action:

In the “Warnings” section, under “Do not use”, the following statement will remain:

“Do not use if you are already pregnant (because it will not work).”

In the Consumer Information Leaflet, the following statements relevant to mechanism of action will be

retained:

Under the heading “What Plan B One-Step® is not.”, the following statement will remain:

“Plan B One-Step® will not work if you are already pregnant and will not affect an existing

pregnancy.”

Under the heading “When not to use Plan B One-Step®.”, the following statement will remain:

“Plan B One-Step® should not be used if you are already pregnant, because it will not work.”

These retained statements are relevant to mechanism of action because they continue to inform the

consumer that levonorgestrel emergency contraception will only work prior to establishment of a

pregnancy. Thus, when considering where in the reproductive cycle that LNG-EC might exert its

mechanism of action, the consumer can still know that LNG-EC works very early in the reproductive

process, before pregnancy has actually begun.

VII. References

Achache H et al. 2006. Endometrial receptivity markers, the journey to successful embryo implantation.

Hum Reprod Update 12:731-46

Bahamondes L et al. 2003. The in vitro effect of levonorgestrel on the acrosome reaction of human

spermatozoa from fertile men. Contraception 68:55-9

Baird D 2009. Emergency contraception: how does it work? Reprod Biomed Online 18 Suppl 1:32-6

Bastianelli C et al. 2006. Emergency contraception. Minerva Ginecol 58(3):193-204

Blankenship T and Enders A 1997. Trophoblast cell-mediated modifications to uterine spiral arteries

during early gestation in the macaque. Acta Anat 158:227-36

Brache V et al. 2013. Ulipristal acetate prevents ovulation more effectively than levonorgestrel: analysis

of pooled data from three randomized trials of emergency contraception regimens. Contraception 88:611-

8

Page 23 of 27

Reference ID: 5100570

Brito K et al. 2005. The in vitro effect of emergency contraception doses of levonorgestrel on the acrosome

reaction of human spermatozoa. Contraception 72:225-8

Brown S et al. 2000. Endometrial glycodelin-A expression in the luteal phase of stimulated ovarian cycles.

Fertil Steril 74(1):130-3

Craven C et al. 1998. Decidual spiral artery remodeling begins before cellular interaction with

cytotrophoblasts. Placenta 19:241-52

Croxatto H et al. 2001. Mechanism of action of hormonal preparations used for emergency contraception:

a review of the literature. Contraception 63:111-21

Croxatto H et al. 2004. Pituitary-ovarian function following the standard levonorgestrel emergency

contraceptive dose or a single 0.75 mg dose given on the days preceding ovulation. Contraception 70:442-

50

Deligdisch L 2000. Hormonal pathology of the endometrium. Mod Pathol 13(3):285-94

do Nascimento J et al. 2007. In vivo assessment of the human sperm acrosome reaction and the expression

of glycodelin-A in human endometrium after levonorgestrel emergency contraceptive pill administration.

Hum Reprod 22:2190-5

Durand M et al. 2001. On the mechanisms of action of short-term levonorgestrel administration in

emergency contraception. Contraception 64:227-34

Gemzell-Danielsson K and Marions L 2004. Mechanisms of action of mifepristone and levonorgestrel when

used for emergency contraception. Hum Reprod Update 10:341-8

Gemzell-Danielsson K 2010. Mechanism of action of emergency contraception. Contraception 82:401-9

Gemzell-Danielsson K et al. 2013. Emergency contraception – mechanisms of action. Contraception

87:300-8

Gemzell-Danielsson K et al. 2013. Emergency contraception. Gynecol Endocrinol 29 Suppl 1:1-14

Gemzell-Danielsson K et al. 2014. Mechanisms of action of oral emergency contraception. Gynecol

Endocrinol 30:685-7

Hapangama D et al. 2001. The effects of peri-ovulatory administration of levonorgestrel on the menstrual

cycle. Contraception 63:123-9

Harper M 1992. The implantation window. Baillieres Clin Obstet Gyn 6:351-71, 1992.

Hermanny A et al. 2012. In vitro assessment of some sperm function following exposure to levonorgestrel

in human fallopian tubes. Reproductive Biology and Endocrinology 10:8, 2012.

Jin X et al. 2017. Pinopode score around the time of implantation is predictive of successful implantation

following frozen embryo transfer in hormone replacement cycles. Human Reprod 32(12):2394-403

Kahlenborn C et al. 2015. Mechanism of action of levonorgestrel emergency contraception. The Linacre

Quarterly 82(1):18-33

Page 24 of 27

Reference ID: 5100570

Lalitkumar P et al. 2007. Mifepristone, but not levonorgestrel, inhibits human blastocyst attachment to

an in vitro endometrial three-dimensional cell culture model. Hum Reprod 22:3031-7

Lalitkumar P et al. 2013. Emergency contraception. Best Pract Res Clin Endo Metab. 27:91-101

Landgren B et al. 1989. The effect of levonorgestrel administered in large doses at different stages of the

cycle on ovarian function and endometrial morphology. Contraception 39:275-89

Leung V et al. 2010. Mechanisms of action of hormonal emergency contraceptives. Pharmacotherapy

30(2):158-68

Lindhard A et al. 2002. Biochemical evaluation of endometrial function at the time of implantation. Fertil

Steril 78:221-33

Ling W et al. 1979. Mode of action of dl-norgestrel and ethinylestradiol combination in postcoital

contraception. Fertil Steril 32:297-302

Ling W et al. 1983. Mode of action of dl-norgestrel and ethinylestradiol combination in postcoital

contraception. II. Effects of postovulatory administration on ovarian function and endometrium. Fertil

Steril 39:292-7

Ling W et al. 1983. Mode of action of dl-norgestrel and ethinylestradiol combination in postcoital

contraception. III. Effect of preovulatory administration following the luteinizing hormone surge on

ovarian steroidogenesis. Fertil Steril 40:631-6

Marions L et al. 2002. Emergency contraception with mifepristone and levonorgestrel: mechanism of

action. Obstet Gynecol 100(1):65-71

Marions L et al. 2004. Effect of emergency contraception with levonorgestrel or mifepristone on ovarian

function. Contraception 69:373-7

Medard L et al. 2010. Hormonal (levonorgestrel) emergency contraception – effectiveness and

mechanism of action. Ginekol Pol 81(7):532-6

Meng C et al. 2009. Effect of levonorgestrel and mifepristone on endometrial receptivity markers in a

three-dimensional human endometrial cell culture model. Fertil Steril 91:256-64

Meng C et al. 2010. Effects of oral and vaginal administration of levonorgestrel emergency contraception

on markers of endometrial receptivity. Hum Reprod 25:874-83

Moghissi K et al. 1973. Contraceptive mechanisms of microdose norenthindrone. Obstet Gynecol 41:585-

94

Mozzanega B and Cosmi E 2011. How do levonorgestrel-only emergency contraceptive pills prevent

pregnancy? Some considerations. Gynecol Endocrin 27(6):439-42

Muller A et al. 2003. Postcoital treatment with levonorgestrel does not disrupt postfertilization events in

the rat. Contraception 67(5):415-9

Munuce M et al. 2005. In vitro effect of levonorgestrel on sperm fertilizing capacity and mouse embryo

development. Contraception 72:71-6

Page 25 of 27

Reference ID: 5100570

Noé G et al. 2011. Contraceptive efficacy of emergency contraception with levonorgestrel given before or

after ovulation. Contraception 84:486-92

Novikova N et al. 2007. Effectiveness of levonorgestrel emergency contraception given before or after

ovulation- a pilot study. Contraception 75:112-8

Okada Y et al. 2001. Studies on the mechanism of edematous changes at the endometrial stroma for

implantation. Semin Thromb Hemost 27:67-77

Okewole I et al. 2007. Effect of single administration of levonorgestrel on the menstrual cycle.

Contraception 75(5):372-7

Ortiz M et al. 2004. Postcoital administration of levonorgestrel does not interfere with postfertilization

events in the new-world monkey Cebus apella. Hum Reprod 19(6):1352-6

Palomino W et al. 2010. A single midcycle dose of levonorgestrel similar to emergency contraceptive does

not alter the expression of the L-selectin ligand or molecular markers of endometrial receptivity. Fertil

Steril 94:1589-94

Peck R et al. 2016. Does levonorgestrel emergency contraceptive have a postfertilization effect? A review

of its mechanism of action. The Linacre Quarterly 83:35-51

Pijnenborg R et al. 2006. The uterine spiral arteries in pregnancy: facts and controversies. Placenta 27(9-

10):939-58

Pucetti R 2012. Letter to the editor. Postovulatory administration of levonorgestrel: interference with

implantation is not excluded. Contraception 86:770-2

Sarkar N 2003. Levonorgestrel as an emergency contraceptive drug. Int J Clin Pract 57(9):824-8

Sarkar N 2005. The potential of mifepristone as an emergency contraceptive drug. Acta Obstet Gynecol

Scand 84(4):309-16

Suarez V et al. 2010. Effect of levonorgestrel in the ovulation, endometrium, and spermatozoa for

emergency oral contraception. Rev Peru Med Exp Salud Publica 27(2):222-30

Tirelli A et al. 2008. Levonorgestrel administration in emergency contraception: bleeding pattern and

pituitary-ovarian function. Contraception 77(5):328-32

Trussell J 2006. Mechanism of action of emergency contraceptive pills. 74:87-9

Uchida et al. 2013. Glycodelin in reproduction. Reprod Med Biol 12(3):79-84

Ugocsai G et al. 2002. Scanning electron microscope changes of the endometrium in women taking high

doses of levonorgestrel as emergency postcoital contraception. Contraception 66:433-7

Vargas M et al. 2012. Effect of single postovulatory administration of levonorgestrel on gene expression

profile during the receptive period of the human endometrium. Jl of Molec Endocrin 48:25-36

Wilcox A et al. 1995. Timing of sexual intercourse in relation to ovulation: effects on the probability of

conception, survival of the pregnancy, and sex of the baby. NEJM 333:1517-21

Page 26 of 27

Reference ID: 5100570

Wilcox A et al. 1998. Postovulatory aging of the human oocyte and embryo failure. Human Reprod 13:394-

7

World Health Organization 1999. WHO laboratory manual for examination of human semen and semen-

cervical mucus interaction. Cambridge (UK): Cambridge University Press

World Health Organization Task Force on Postovulatory Methods of Fertility Regulation 1998. Randomized

controlled trial of levonorgestrel versus the Yuzpe regimen of combined oral contraceptives for

emergency contraception. Lancet 352:428-38

Yaron Y et al. 1994. Endometrial receptivity in the light of modern assisted reproductive technologies.

Fertil Steril 62:225-32

Yeung W et al. 2002. The effects of levonorgestrel on various sperm functions. Contraception 66:453-7

Page 27 of 27

Reference ID: 5100570

--------------------------------------------------------------------------------------------

------------------------------------------------------------

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/

(b) (6)

12/23/2022 02:16:21 PM

Reference ID: 5100570