May 8, 2008

Petition to Ban Ethinyl Estradiol/Norelgestromin (Ortho-Evra)

Andrew von Eschenbach, M.D., Commissioner
U.S. Food and Drug Administration
Food and Drug Administration
5600 Fishers Lane 
Rockville, MD 20857 

Dear Dr. von Eschenbach:

Public Citizen, representing more than 80,000 consumers nationwide, hereby petitions the Food and Drug Administration (FDA), pursuant to the Federal Food, Drug and Cosmetic Act 21 U.S.C. Section 355(e)(3) and 21 C.F.R. 10.30, to ban the transdermal contraceptive patch, Ortho-Evra (ethinyl estradiol/norelgestromin, Johnson & Johnson), with market removal of the device to occur within 6 months.

Compared to standard 35 microgram estrogen/progestin oral contraceptives, Ortho-Evra results in:

• 60% more estrogen on average exposure;
• greater variability in estrogen levels;
• a possible two-fold increased risk of venous thrombosis (typically, painful blood clots of the leg which can travel to the lungs and cause death);
• increased risk of side effects such as breast discomfort, severe menstrual pain, nausea, and vomiting;
• a 50% increased likelihood of discontinuation;
• no improvement in contraceptive outcomes.

In the past several years, millions of women and their healthcare providers have discovered that Ortho-Evra is a less attractive product than advertised.  Demand for the patch has dropped dramatically, from over 9.9 million prescriptions filled in 2004 to 2.7 million prescriptions filled in 2007 (a decline of 73%), yet Ortho-Evra was still among the top 200 brand-name drugs by sales and prescriptions in the United States in 2007.[1]

Ongoing litigation on behalf of women injured or killed after using Ortho-Evra has led to the release of two unpublished studies that show Johnson & Johnson knew about potentially higher levels of estrogen in the patch as well as the greater variability in estrogen delivery before FDA approval in November 2001.[2]  The label has also never mentioned the increased odds of side effects on the patch.  Women deserve a level of risk from their contraceptive comparable to standard combined oral contraceptives (the “pill”), especially when contraceptive efficacy is no greater with the patch. 

Background

Ortho-Evra is a patch that contains 0.75 milligrams of ethinyl estradiol (EE, an estrogen) and 6 mg of norelgestromin (NGMN, a progestin).  Each patch is designed to be worn on the skin for 7 consecutive days before removal.  Each cycle consists of three patches worn in a row followed by a patch-free week.  According to the label, the patch may be worn on the abdomen, buttocks, upper outer arms, or upper torso.  Johnson & Johnson expected two primary advantages compared to existing oral contraceptives: (1) stable blood levels of hormones due to weekly dosing and avoidance of initial hormone metabolism through the gastrointestinal tract and/or liver, and (2) improvements in compliance because of skin application and the weekly instead of daily dosing regimen. 

The analysis in this petition focuses on the safety and efficacy of elevated estrogen bioavailability in Ortho-Evra compared to 35 mcg EE pills.  However, comparison with pills is complicated by different pharmacokinetic (PK) profiles.  For example, ingestion of a contraceptive pill quickly produces a pronounced spike in blood estrogen before it is metabolized over the course of 24 hours and the next pill is taken.  In contrast, patch application results in a gradual increase in estrogen that plateaus within three days and remains relatively constant while the patches are worn.  In addition, steady-state, or the state where blood levels of a hormone are stable because the rate of excretion matches the rate of intake, is attained at different times; around the third day for pill use and the second week for patch application. 

Bioavailability describes how much of an administered drug is absorbed into systemic circulation and is thus available at the site of action.  It is usually determined with a combination of PK parameters, including the local maximum of chemical concentration in the blood (Cmax), the steady-state concentration of chemical in the blood (Css), and, most importantly, the area under the drug concentration-time curve (AUC).  AUC provides a well-validated measure of exposure to EE because it measures the total amount of EE that reaches the circulation over a period of time (unlike Cmax or Css which provide only snapshots of blood estrogen concentration).[3]  Focusing on AUC thus helps to account for the differences between patches and pills in daily estrogen delivery.  Thus, this petition restricts itself to studies that compare the AUC levels over time (AUC0-t) at steady state of the patch to pills unless otherwise noted.

The original Ortho-Evra label approved in 2001 cited a single-dose PK study that claimed Ortho-Evra delivered only the equivalent of 20 mcg of EE every day,[4] even though the company already had evidence from another pre-approval study that the patch delivered more estrogen than 30 mcg pills.[5]  In 2005, Johnson & Johnson changed the Ortho-Evra label to cite a post-market pharmacokinetic (PK) study that found that overall exposure to EE as measured by both AUC and Css was 55-60% higher from the patch than from pills with 35 micrograms of EE.[6]

Unfortunately, high levels of EE exposure can have devastating consequences for health, particularly higher rates of venous thromboembolism (VTE).   An FDA advisory committee had concluded in 1988 that “efficacy for the prevention of pregnancy is the same for all oral contraceptives equal to or greater than 30 mcg of estrogen,” and that high-dose oral contraceptives were not of “sufficiently unique value in clinical practice to warrant their continued availability for contraception.” [7]  Acting on this recommendation, manufacturers of oral contraceptives voluntarily withdrew all high-dose (>50 mcg of estrogen) products.[8]

Currently, among the more than 60 pills on the market, only two generic drugs have 50 mcg EE doses; the rest have only 35 mcg of EE or less.  Indeed, according to a standard pharmacological textbook, hormonal contraception should consist of the “minimum dosage of steroids that provides effective contraceptive coverage.  This is often a pill with 30 to 35 mcg of estrogen, but preparations with 20 mcg may be adequate.”[9]

Given the unnecessary dangers of increased estrogen levels in contraception, it is not surprising that the Ortho-Evra label had to be changed again in both 2006 and 2008 to include findings from epidemiological studies on the up to 2-fold increased risk of VTE in women using the patch compared to standard pills. 

Three Studies Document that Ortho-Evra Delivers a High Dose of Ethinyl Estradiol

PHI-017, the first steady-state study to directly compare the pharmacokinetics of Ortho-Evra to pills, was completed two years before FDA approval of the product.[10]  PHI-017 compared the patch to a triphasic pill (30 mcg EE days 1 to 6, 40 mcg EE days 7 to 11, and 30 mcg EE days 12 to 21/levonorgestrel) and found a 2-fold increase in exposure to EE from the patch, equating the patch to a 47 mcg EE pill (see Appendix A and B).  Finally released as a result of ongoing litigation, these results have never been published, mentioned in the label, or publicly acknowledged.

After FDA approval, two additional comparative PK studies were done.  The first study, known as NED-1, was conducted in 2003 as a post-marketing requirement of the European Medicines Evaluation Agency (EMEA).[11],[12]  This study showed a 60% greater exposure to EE on the patch compared to a standard 35 mcg EE pill, and became the basis for the 2005 label change.   The label failed to mention that exposure to EE (in AUC) from use of the patch equated to a daily 56 mcg pill (95% C.I.=18.9-94.8 mcg) (see Appendix B).  This exceeds the amount determined by the 1988 FDA advisory committee to be necessary for contraception, thereby raising the risk of adverse effects without any increase in efficacy.

The final PK study was conducted by van den Heuvel et al., researchers unaffiliated with Johnson & Johnson.[13]  They found a 60% greater AUC level in Ortho-Evra compared to a 30 mcg EE pill.

Patch Application Sites Deliver Different Amounts of Estrogen

The only study to compare levels of hormone exposure from the abdomen, buttock, arm, and torso found that the abdominal site delivered approximately 20% less EE than the other three sites.[14]  Presumably, the FDA deemed this difference acceptable after Johnson & Johnson argued that the abdominal site still delivered a “therapeutically equivalent,” or just as efficacious, dose of hormones.[15]  Of course, an equally efficacious dose of estrogen is even lower than the abdominal site dose.  Notably, both study PHI-017 and van den Heuvel et al. measured estrogen exposure from abdominal application sites only.  It is possible that estrogen levels in these studies would have been even higher if other sites had been included.

Ortho-Evra Delivers a Greater Variability in Exposure to Ethinyl Estradiol than Pills

A primary theoretical advantage of Ortho-Evra over contraceptive pills was that Ortho-Evra would deliver a relatively constant amount of hormone through transdermal absorption without the peaks and troughs seen in daily EE pill dosing.  In fact, evidence from comparison studies of the two contraceptives shows that overall variability in estrogen exposure from the patch is actually greater than EE exposure from pills (see Figure 1).  For example, the 68% confidence interval (1 standard deviation)[16] for estrogen exposure from the oral contraceptive Triphasil in study PHI-017 was 534-902 picogram-hr/mL while the same interval for estrogen exposure from Ortho-Evra was 1025-2050 pg-hr/mL.  A comparison of standard deviations across three studies finds that, adjusted for the fact that measurements with greater means tend to have higher standard deviations, Ortho-Evra has 1.2-3.5 times more variability in estrogen exposure than pills.[17]

Figure 1: Ethinyl Estradiol Exposure in Multiple-Dose Comparison Studies, 68% CI[18](patch application on abdomen site only)

The variability in steady-state EE concentrations (Css) recorded from PK studies of Ortho-Evra is also alarming.  During the development of Ortho-Evra, a target range of Css for the patch was determined to ensure delivery of efficacious levels of hormones.  This target range was based upon average steady-state concentrations of estrogen in women using a 35 mcg pill, or between 25 and 75 pg/mL.[19]   In a controlled study of 10 healthy Ortho-Evra users, White et al. found mean Css levels of 111±63 and 114±59 pg/mL using two different blood assay methods.[20] The current label notes that the mean Css found in patch users has been found to range from 11.2-137 pg/mL.[21]

Higher Risk of Blood Clots

Given the higher average dose of EE resulting from Ortho-Evra and long-standing evidence that increased estrogen exposure increases risk for VTEs,[22] it is no surprise that both pharmacodynamic (PD) and epidemiological studies suggest an increased risk for VTEs among women who use the patch compared to 20-35 mcg EE pills.

Pharmacodynamic Studies: Markers for Increased Blood Clot Risk

PD studies measure chemical changes in the body from a drug.  These studies are useful for tracking surrogate markers of potential adverse events when the events themselves are rare or hard to study conclusively.  The best surrogate marker for assessing the risk of thrombosis in users of hormonal contraceptives is sex hormone-binding globulin (SHBG), because it combines the effects of both the estrogen and progestin components of the contraceptive to measure the total estrogenic effect of the pill, which in turn has been linked to a higher risk of thrombosis.[23],[24]

The three known pharmacodynamic studies of Ortho-Evra were conducted as randomized, controlled trials that compared, among other biochemical responses, the SHBG of women who used the patch compared to pills.  Blood samples were drawn at baseline and at the conclusion of the last cycle (ie. at steady-state) for analysis.  A ratio of the increase from baseline is used to compare the differences in SHBG levels.

The first study was conducted as a component of study NED-1 mentioned above.  The ratio of SHBG change from baseline in Ortho-Evra users compared to 35 mcg EE/norgestimate pill users was 41% upon reaching steady-state (p=0.0001).[25]

White et al. analyzed SHBG levels among 10 women who wore Ortho-Evra compared to nine control women who took a 35 mcg EE/norgestimate pill for three cycles.  Ortho-Evra users experienced a 73% greater change from baseline SHBG (p<0.05).[26]

The most recently published study found a 33% difference in change from baseline between 32 women on Ortho-Evra compared to 33 women using a 20 mcg EE/desogestrel pill after six cycles of contraceptive use (p=0.005).[27]  Thus, all three studies documented statistically significant increases of risk-associated SHBG in Ortho-Evra users compared to pill users. 

Epidemiological Studies

In response to concerns about increased VTEs experienced by women using Ortho-Evra, Johnson & Johnson also sponsored three case-control studies comparing VTE risk in Ortho-Evra to the risk in pills.  Case-control studies are intended to detect the association between a rare adverse event (such as idiopathic, or no-known-cause, VTE in healthy young women) and a specific intervention (such as Ortho-Evra use).  This retrospective method compares “cases” of women who have had an adverse event to “control” women without adverse events and, in this instance, determines whether Ortho-Evra use is more common among the cases than controls.  To mitigate the pitfalls of the non-randomized study design, the three studies matched cases and controls by relevant characteristics such as age and risk factors for VTE and analyzed the data with conditional logistic regressions to control for other possible confounders.

Cole et al. were the first investigators to completely publish their Ortho-Evra case-control study.  They used a database of pharmacy and medical claims from a healthcare insurer to identify cases of idiopathic VTE among all women who had recently filled prescriptions for either Ortho-Evra or a 35 mcg EE/ norgestimate pill.  Cases of VTE were confirmed with abstracted medical records.  Four control women without VTE were then randomly selected from the same database and matched to each case by birth year, VTE risk factors, and particular history of exposure to the relevant contraceptive.  This study reported that, compared to women without idiopathic VTEs, women with idiopathic VTEs were 2.4 times more likely to have used Ortho-Evra than a 35 mcg EE/norgestimate pill (95% C.I.=1.1-5.5).[28]

The second study, by Jick et al., used slightly different methods and found no increased association between Ortho-Evra and non-fatal, idiopathic VTE compared to a 35 mcg EE/norgestimate pill (Odds Ratio=1.0; 95% C.I.=0.7-1.5).[29]   This study drew its population from a different insurance database, and unlike the Cole study, included only women who were new users of one of the study drugs, excluded cases of fatal VTE, and did not confirm cases of VTE with medical records.  The FDA used these two case-control studies as the basis for its 2006 label revision.

Some information about a third, as yet unpublished, case-control study was released as the basis for the 2008 Ortho-Evra label revision.[30]  This study was also conducted by the Jick group and used similar methods as their previous study.  This time the investigators compared Ortho-Evra to a 30 mcg EE/levonorgestrel pill and found an increased, but not quite statistically significant, risk of non-fatal idiopathic VTE for the Ortho-Evra users (OR=2.0; 95% C.I.= [0.9-4.1). 

The 2008 label concluded that, “it is not known whether there are changes in the risk of serious adverse events based on the differences in pharmacokinetic profiles of EE in women using ORTHO EVRA® compared with women using oral contraceptives containing 35 mcg of EE.”  While these epidemiological studies are not conclusive on their own, in addition to the PK and PD data, they do suggest a higher risk of VTEs than the label admits.  

More Adverse Effects and Higher Overall Discontinuation Rates

In addition to VTEs, increased rates of other adverse effects have been shown from two large clinical efficacy trials of Ortho-Evra compared to pills.  Audet et al. randomized 1495 women to receive either the patch or a triphasic pill (30 mcg EE days 1 to 6, 40 mcg EE days 7 to 11, and 30 mcg EE days 12 to 21/levonorgestrel).  One-third of participants were enrolled for 13 cycles and the remaining women were enrolled for 6 cycles.[31]  Urdl et al. similarly randomized 1517 women to receive either the patch or a monophasic pill (20 mcg EE/desogestrel for 21 days) for either 6 or 13 cycles of use.[32]  Both trials found that Ortho-Evra users were statistically significantly more likely than pill users to experience painful side effects.  For example, Audet and Urdl found 19 and 25% of Ortho-Evra users in both trials experienced breast discomfort, three times more than pill users (OR=3.09; 95% C.I.=2.26-4.22], OR=2.98; 95% C.I.=2.29-3.90], respectively).  Audet et al. found that 13% of Ortho-Evra users had severe menstrual pain (OR=1.43; 95% C.I.=1.03-1.99).  Urdl et al. also found increased odds of nausea (OR=2.08; 95% CI=1.46,2.95) and vomiting (OR=1.88; 95% CI=1.12-3.16) (see Appendix C.  In cases where side effects were not reported, the results were either not collected or not significant).

Side effects have the further effect of inducing patients to stop using a particular product.  Women randomized to receive the patch were consistently more likely to drop out of the aforementioned efficacy trials than those assigned to pills.  For example, Audet et al. found that 12% of all women randomized to Ortho-Evra dropped out due to adverse events compared to only 5% of women who used the triphasic pill (OR=2.27; 95% C.I.=1.59-3.25).  Urdl et al. found a similar rates of discontinuation due to adverse events (OR=2.01; 95% C.I.=1.38-2.95).  In these studies, discontinuation from Ortho-Evra for any reason (including adverse effects) was 1.45 to 1.58 (p<0.05) times more likely compared to a pill.  Other non-randomized or controlled studies conducted to observe typical use of contraceptives in high-risk populations suggest low user acceptability and higher pregnancy rates on the patch compared to the pill.[33],[34]  Studies on pills have shown that any discontinuation frequently leads to a period of no contraceptive use at all, and a greater subsequent risk of pregnancy.[35]

Same Efficacy as Pills Despite Higher Compliance

Audet et al. and Urdl et al. also measured compliance, defined for pills as the proportion of completed cycles with 21 consecutive days of drug taking and for the patch as no patch worn more than 7 or 8 days, respectively.  Compliance results did not include the potential cycles of women who decided to discontinue in the study.  Both studies found that compliance was better in patch users compared to pill users (OR=2.05; 95% C.I.=1.83-2.29] and OR=2.76; 95% C.I.=2.35-3.24]).  Yet this advantage failed to translate into statistically significantly fewer pregnancies (OR=0.57; 95% C.I.=0.18-1.77] and OR=1.49; 95% C.I.=0.30-7.53) in studies that were statistically underpowered to detect a difference.  While the increased compliance of Ortho-Evra users compared to pill users is much touted by Johnson & Johnson, the similar pregnancy outcomes among users who continued in the trial question the relevance of that outcome.

Rationale for Phased Withdrawal from the Market

Without any additional efficacy and considerable additional risk beyond the accepted levels from pills, Ortho-Evra is a poor choice for women.  Yet the patch is still vastly superior to no contraception at all.  Withdrawal of any contraceptive from the market carries the risk that some proportion of users will not immediately replace their contraception with a method that is as effective as the banned product.  Ideally, Johnson & Johnson could immediately release a safer reformulation of the patch that delivered lower and less variable levels of estrogen, although this has not happened in the seven years Ortho-Evra has been available.  Public Citizen therefore requests a six-month transition period in which Ortho-Evra will be available for refill prescriptions to allow women time to meet with their healthcare provider and seek an alternative contraceptive method. 

Conclusion

The theoretical benefits of Ortho-Evra over alternative contraceptive options included less variability in estrogen delivery and better compliance.  Unfortunately, Ortho-Evra has more than theoretical risks.  Ortho-Evra users are exposed to levels of estrogen that are on average comparable to a 56 mcg EE pill, but could range widely for the individual, from 18 to 95 mcg of EE.  Documented adverse events include an up to two-fold increase in venous thromboembolism and an increased variety of painful side effects.  Finally, despite better compliance, there is no difference in pregnancy outcomes over pills.  If Ortho-Evra had been designed as a pill, it is unlikely to have been approved because of its increased estrogen content.  The considerable safety concern of high-dose, variable estrogen exposure tips the balance of risks and benefits against the availability of Ortho-Evra as a contraceptive. 

Environmental Impact

Removal of Ortho-Evra from the market will be beneficial for the environment, as discarded patches still contain considerable amounts of estrogen (80% of the original estrogen dose) that pose a risk of leakage and contamination greater than from pills.[36]

Sincerely,

Eunice Yu
Staff Researcher

Sidney M. Wolfe, M.D.
Director
Public Citizen’s Health Research Group

Appendix A: Summary of Multiple-Dose Ortho-Evra Pharmacokinetic Studies

Premarket Studies

PHI-005[37]
This open-label study of Ortho-Evra was conducted on 12 women to determine the PK profile of two consecutive patches.  The first patch was applied for 7 days and the second patch was applied for 10 days to assess the PK effect of improperly wearing a patch for three days beyond the prescribed 7-day dosage interval.  All patches were worn on the abdomen and blood samples were taken throughout the study.

AUC results were recorded on the 7th day of the second patch, providing multiple-dose AUC measurements that are comparable to other studies (see Table A1).

PHI-013[38]
Study PHI-013 followed 24 women for three cycles of patch wear, and is the longest PK study ever done on Ortho-Evra.  Women were randomized to wear the patch according to package directions on either the buttock or abdomen patch site during measurement periods (Week 1 of Cycle 1, and Weeks 1,2,3 of Cycle 3), but were free to wear the patch on any of the four anatomical sites during the non-measurement weeks (see Table A1).

Table A1: Ethinyl Estradiol Exposure from Multiple-Dose PK Studies of Ortho-Evra

Study	Application Site	Duration of Patch Application	Sample Size	Avg AUC0-168h(pg*h/mL)	SD
PHI-005	Abdomen	7+10 days	12	8353	3098
PHI-013	Abdomen	3 x 21 days	12	12139	3241
PHI-013	Buttock	3 x 21 days	12	8840	5176
PHI-017*	Abdomen	21 days	12	10761	3589
NED-1	Abdomen and Buttock	2 x 21 days	31	12974	4295
Van den Heuvel et al.	Abdomen	21 days	8	11933	1833

*Triphasil arm only

PHI-017[39]
Study PH1-017 compared Ortho-Evra worn on the abdomen to three oral contraceptives; Triphasil (a triphasic regimen of levonorgestrel and 30 mcg EE for six days, 40 mcg EE for five days, and 30 mcg EE for ten days), Alesse (levonorgestrel and 20 mcg EE), and Mercilon (desogestrel and 20 mcg EE).  Thirty-six women were randomized to receive one of the three pills in a cross-over study in which subjects first received either their assigned pill or the patch, followed by a minimum 28-day drug-free period and the alternative treatment.  The Evra-Triphasil group received the full 21-day cycle of pills and Ortho-Evra for 21 days while the Evra-Alesse and Evra-Mercilon groups received only 7 days (1 pill/day or 1 patch/week) of each treatment.  Blood samples were taken only during the last week for the Evra-Triphasil group and throughout the week for Evra-Alesse and Evra-Mercilon groups (Figure 1 in text).

The results showed that women had steady-state AUC levels of EE that were twice as high when using Ortho-Evra compared to 30-40 mcg Triphasil.  EE exposure was three times as high on Ortho-Evra compared to the 20 mcg EE Alesse and Mercilon.  The study’s conclusion suggested that these comparisons were not valid because the EE exposure measured from the pills in this study were unusually low compared to their reference values.  At the same time, the Ortho-Evra results from this study were similar to past studies.  Thus, “the lower-than-expected exposure to ethinyl estradiol after administration of the oral contraceptives led to a greater relative exposure to ethinyl estradiol after dosing with EVRA.”  Based on this reasoning, the conclusions submitted to the FDA read, “data from this and other studies indicate that EVRA is comparable in exposure and estrogenicity to 35 mcg oral contraceptives.”

This statement attempts to make a spurious comparison between AUC results from the patch in this study to past studies of pills conducted on different populations, while ignoring that a concurrent, randomized comparison group was designed into the study.  Indeed, the cross-over study design puts the burden on the company to show why the same women had substantially different results using the patch compared to the pill.  Although this study suggested a higher level of EE exposure on the patch compared to the pill well before approval (the study was completed in June 1999, nearly two years before FDA approval), the FDA-approved labels for Ortho-Evra have never mentioned study PH1-017 or its results.  In addition, the study is not mentioned by the FDA in its clinical pharmacology reviews nor acknowledged by Johnson & Johnson in their later submissions to the FDA.  Not surprisingly, it has also never been published.

Post-market studies

NED-1[40],[41]
In the first post-market PK study, 36 women were randomized into an open-label cross-over study that compared Ortho-Evra to Cilest (norgestimate and 35 mcg EE).  Women either wore the patch or took daily pills for two cycles followed by a wash-out period of 28 days and the alternative treatment for two cycles.  Each cycle consisted of 21 consecutive days of treatment followed by 7 days of no treatment.  Study subjects were further randomized to wear the patch on the abdomen or buttock.  Blood samples were taken during Week 1 of Cycle 1 and Week 3 of Cycle 2. 

As noted above, this study showed that the average exposure to EE at steady-state on the patch was 60% higher than the average AUC of the 35 mcg EE pill, the highest typical contraceptive dose on the market (see Figure 1 in text).

Van den Heuvel et al.[42]
In this independently conducted post-market study, 24 subjects first took a pill, Microgynon (levonorgestrel and 30 mcg EE) for 2-8 weeks before a 7-day pill free period and randomization to continue with 21 days of either the patch, the NuvaRing, or Microgynon.  All patch users wore the patch on their abdomen.  Blood samples were taken throughout treatment (Figure 1 in text).

Appendix B: Linear AUC-Dose Relationship

The claim that Ortho-Evra contains 56 mcg of ethinyl estradiol comes from study NED-1, in which blood levels of EE in women on the patch were compared to their EE levels while taking a 35 mcg pill.  The study found that Ortho-Evra exposed women to 60% more EE, thus 35 mcg EE times 60% equals a 56 mcg equivalent dose of estrogen.  The linear relationship between estrogen AUC and equivalent pill dose has been confirmed by Johnson & Johnson’s experts[43] as well as their own studies.  Study PHI-017 explains that “the EE estimated dose (compared to OC’s) for EVRA in a 24 hour period can be calculated from the regression equation,”[44] derived from plotting the AUCs of various doses of pills (see Figure 1B).  Study NED-1 again uses this linear regression equation to calculate the equivalent EE dose for all PK studies done on Ortho-Evra.[45]  Table 1B shows that the equivalent dose of EE delivered from commercial lots of Ortho-Evra patches at steady-state averages 56 mcg of EE (95% CI: 18.9-94.8) using this linear regression equation. 

Figure 1B: EE Dose vs. AUC0-24 for Six Oral Contraceptive Pills[46]

It is not clear why Ortho-Evra’s label change indicating the true level of exposure to an active ingredient in the product came in 2005, nearly four years after its initial approval.  It is possible that the smaller lots of patches developed initially for clinical trials actually did deliver less EE on average than those made later on a larger scale for commercial release.  A comparison of AUC results from multiple-dose studies of Ortho-Evra show that commercial patches exposed women to an average of 12 mcg more EE per day than the clinical trial patches (not significant, p>0.05).  There is limited publicly-available evidence that Johnson & Johnson has had other troubles with manufacturing Ortho-Evra.  For example, the European and Canadian versions of the patch contain only 0.6 mg of EE, yet claim bioequivalence to the 0.75 mg EE patch sold in the U.S.[47],[48]

Table 1B: Equivalent Ethinyl Estradiol Dose from Multiple-Dose PK Studies of Ortho-Evra

Study[49]	Manufacturing Lot	Application Site	EE Dose[50] (mcg/day)	95% CI		Avg. EE Dose[51] (mcg/day)
				Lower	Higher
PHI-005	Phase 2 clinical	Abdomen	36.4	9.0	63.8
PHI-013	Phase 3 clinical	Abdomen	53.2	24.5	81.8
PHI-013	Phase 3 clinical	Buttock	38.6	0.0	84.3
PHI-017*		Abdomen	47.1	15.3	78.8	43.8
NED-1	Commercial	Abdomen and Buttock	56.8	18.9	94.8
Van den Heuvel et al.	Commercial	Abdomen	52.2	36.0	68.5	55.9

*Triphasil arm only

Appendix C: Discontinuation and adverse event rates on the patch compared to an oral contraceptive[52]

	Study	Audet 2001	Urdl 2005
	Reference Pill	30/40/30 mcg EE/ 50/75/125 mcg LNG	20 mcg EE/ 150 mcg DSG
	Sample Size (Ortho Evra / Pill)	856 / 639	846 / 643
Discontinuation- Overall	% O-E	30%	20%
	% Pill	21%	14%
	Odds Ratio	1.58 [1.25, 1.99]	1.45 [1.11, 1.90]
Discontinuation- Due to adverse event	% O-E	12%	10%
	% Pill	5%	5%
	Odds Ratio	2.27 [1.59, 3.25]	2.01 [1.38, 2.95]
Breast discomfort	% O-E	19%	25%
	% Pill	6%	9%
	Odds Ratio	3.09 [2.26, 4.22]	2.98 [2.29, 3.90]
Severe menstrual pain	% O-E	13%	5%
	% Pill	10%	5%
	Odds Ratio	1.43 [1.03, 1.99]	1.15 [0.72, 1.83]
Nausea	% O-E	20%	12%
	% Pill	18%	6%
	Odds Ratio	1.14 [0.88, 1.49]	2.08 [1.46, 2.95]
Vomiting	% O-E	Not reported	5%
	% Pill		3%
	Odds Ratio		1.88 [1.12, 3.16]

 

---

[1] Drug Topics “Top 200 brand drugs by retail dollars in 2007” Accessed March 20, 2008. . Now available at: http://drugtopics.modernmedicine.com/drugtopics/data/articlestandard//dr….

[2] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-014; Phase 1.” May 4, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.583. and Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

[3] Shargel L and Yu A. Applied Biopharmaceutics and Pharmacokinetics.   1985, Appleton-Century-Crofts/Prentice-Hall. East Norwalk, Connecticut. P.132.

[4] Ortho-Evra label.  November 20, 2001.  Available at: http://www.fda.gov/cder/foi/label/2001/21180lbl.pdf. Now available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021180s026lbl.pdf

[5] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

[6] Ortho-Evra label.  November 10, 2005. Available at: http://www.fda.gov/cder/foi/label/2001/21180lbl.pdf. Now available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021180s026lbl.pdf

[7] Food and Drug Administration, Transcript of the Fertility and Maternal Health Drugs Advisory Committee. Volume II.  January 15, 1988, P.166.

[8] Ortho Pharmaceutical Corporation., “Dear Doctor letter.” March 30, 1988.  In re: Ortho Evra Products Liability Litigation,  MDL 1742, N.D. Ohio, Case No., 1:06-40000, Deposition Ex. 674.

[9] Goodman and Gilman’s The Pharmacological Basis of Therapeutics. 10th Ed. 2001.  McGraw Hill. 

[10] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

[11] Johnson & Johnson Pharmaceutical Research and Development LLC. ”Clinical Study Report Protocol PRI/EDN-NED-1; Phase 1.” March 10, 2003.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.629.

[12] Devineni D, Skee D, Vaccaro N, Massarella J, Janssens L, LaGuardia KD, Leung AT. J Clin Pharmacol. 2007; 47:497-509.

[13] van den Heuvel MW, van Bragt AJM, Alnabawy AKM, Kaptein MCJ. Contraception 2005; 72:168-174.

[14] Abrams LS, Skee DM, Natarajan J, Wong FA, Anderson GD. Br J Clin Pharmacol, 2002; 53: 141-146.

[15] Abrams LS, Skee DM, Natarajan J, Wong FA, Anderson GD. Br J Clin Pharmacol, 2002; 53: 141-146.

[16] A 68% confidence interval means there is a 68% chance of the true parameter (in this case the mean) falling within the interval.  A wider interval gives greater confidence but in this case is used to indicate the large range of AUC levels in Ortho-Evra compared to pill users.

[17] Calculated as a ratio of the percent coefficient of variation (%CV) of Ortho-Evra over the pill.  %CV is the standard deviation divided by the mean, expressed as a percentage.

[18] Dotted lines show variation in AUC data from the patch worn on the abdomen only (other sites are approximately equivalent, or higher).  Solid lines show variation in AUC from the same study of the pill.  Sample sizes in each group are approximately equal.

[19] Ortho-Evra label.  November 20, 2001.  Available at: http://www.fda.gov/cder/foi/label/2001/21180lbl.pdf. Now available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021180s026lbl.pdf

[20] White T, Ozel B, Jain JK, Stanczyk FZ. Contraception. 2006; 74: 293-296.

[21] Ortho-Evra label.  January 18, 2008.  Available at: http://www.fda.gov/cder/foi/label/2001/21180lbl.pdf. Now available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021180s026lbl.pdf

[22] van Vliet HAAM, Frolich M, Christella M, Thomassen LGD, Doggen CJM, Rosendaal FR, Rosing J, Helmerhorst FM. Hum Reprod. 2005; 20: 563-568.

[23] van Vliet HAAM, Frolich M, Christella M, Thomassen LGD, Doggen CJM, Rosendaal FR, Rosing J, Helmerhorst FM. Hum Reprod. 2005; 20: 563-568.

[24] Van Rooijen M, Silveira A, Hamsten A, Bremme K. Am J Obstet Gynecol. 2004; 190:332-337.

[25] From study NED-1, p-values not published in Devenini et al. (2007)

[26] White T, Ozel B, Jain JK, Stanczyk FZ. Contraception. 2006; 74: 293-296.

[27] Kluft C, Meijer P, LaGuardia KD, Fisher AC. Contraception. 2008; 77: 77-83.

[28] Cole JA, Norman H, Doherty M, Walker AM. Obstet Gynecol. 2007; 109: 339-346.

[29] Jick S, Kaye J, Li L, Jick H. Contraception 2007; 76:4-7.

[30] Johnson & Johnson Pharmaceutical Research & Development, L.L.C. ClinicalTrials.gov ID: NCT00511784. Accessed on March 11, 2008. https://clinicaltrials.gov/ct2/show/NCT00511784?term=ortho+evra&rank=10

[31] Audet M, Moreau M, Koltun WD, Waldbaum AS, Shangold G, Fisher A, Creasy GW. JAMA. 2001; 285:2347-2354

[32]Urdl W, Apter D, Alperstein, Koll P, Schonian S, Bringer J, Fisher A, Preik M. Obstet Gynecol. 2005; 121: 202-210.

[33] Thurman AR, Hammond N, Brown HE, Roddy ME. J Pediatr Adolesc Gynecol. 2007; 20: 61-65.

[34] Bakhru A, Stanwood N. Obstet Gynecol. 2006; 108: 378-86.

[35] Rosenberg MJ, Waugh MS, Long S. J Reprod Med. 1995; 40: 355-60.

[36] Prescrire Editorial Staff. “Do Evra contraceptive patches represent an advance?” Prescrire Int   2004; 13: 123-135.

[37] Abrams LS, Skee DM, Wong FA, Anderson NJ, Leese PT. J Clin Pharmacol, 2001; 41:1232-1237.

[38] Abrams LS, Skee DM, Natarajan J, Wong FA, Lasseter KC. Contraception. 2001; 64:287-294.

[39] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

[40] Johnson & Johnson Pharmaceutical Research and Development LLC. ”Clinical Study Report Protocol PRI/EDN-NED-1; Phase 1.” March 10, 2003.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.629.

[41] Devineni D, Skee D, Vaccaro N, Massarella J, Janssens L, LaGuardia KD, Leung AT. J Clin Pharmacol. 2007; 47:497-509.

[42] van den Heuvel MW, van Bragt AJM, Alnabawy AKM, Kaptein MCJ. Contraception 2005; 72:168-174.

[43] Deposition of Frank Z. Stanczyk PhD. in re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex. H.

[44] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591. pp 32.

[45] See study NED-1 Table 24.  But this table uses only single-dose data, even when multiple-dose data are available

[46] Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

[47] Prescribe Editorial Staff “Evra dose change: an air of secrecy” Prescribe Int 2007; 16 (89):107.

[48] Health Canada “New Safety Information about Evra Transdermal System” November 21, 2006. Accessed on March 31, 2008. http://www.hc-sc.gc.ca/dhp-mps/medeff/advisories-avis/prof/2006/evra_hpc-cps_e.html

[49] Data on manufacturing lots comes from study NED-1 Table 24 pp.87.  Data used to calculated EE dose and CI taken from the following published and unpublished sources.

PHI-005: Abrams LS, Skee DM, Wong FA, Anderson NJ, Leese PT. J Clin Pharmacol, 2001; 41: 1232-1237.

PHI-013: Abrams LS, Skee DM, Natarajan J, Wong FA, Lasseter KC. Contraception. 2001; 64: 287-294.

PHI-017: Johnson & Johnson Pharmaceutical Research Institute. ”Clinical Study Report Protocol NRGEEP-PHI-017; Phase 1.” June 11, 1999.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.591.

NED-1: Johnson & Johnson Pharmaceutical Research and Development LLC. ”Clinical Study Report Protocol PRI/EDN-NED-1; Phase 1.” March 10, 2003.  In re: Ortho Evra Products Liability Litigation, MDL 1742, N.D. Ohio, Case No. 1:06-cv-40000, Ex.629.

[50] Equivalent EE OC Dose = (AUC24h-17.3)/32.3; equation from PHI-017 and NED-1 submissions to the FDA

[51] Average is weighted by sample size of studies. 

[52] Data from: Audet M, Moreau M, Koltun WD, Waldbaum AS, Shangold G, Fisher A, Creasy GW. JAMA. 2001; 285:2347-2354 and Urdl W, Apter D, Alperstein, Koll P, Schonian S, Bringer J, Fisher A, Preik M. Obstet Gynecol. 2005; 121: 202-210.  Odds ratio data from: Lopez LM, Grimes DA, Gallo MF, Schulz KF. Cochrane Database Syst Rev. 2008; 1: CD003552.