Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Mar 1.
Published in final edited form as: Women Health. 2019 Jul 8;60(3):241–248. doi: 10.1080/03630242.2019.1635560

“My BMI is too high for Plan B.” Changing trends in body mass index of women seeking ulipristal acetate emergency contraception online from 2011 to 2015

Kelly Cleland a, Brandon Wagner a,b, Nicole K Smith a, James Trussell a,c
PMCID: PMC6946884  NIHMSID: NIHMS1533305  PMID: 31284850

Abstract

Emergency contraceptive (EC) pills may be less effective for women with higher body mass index (BMI), but little is known about public response to the fact that EC may lose efficacy as weight increases. In November 2013, European authorities changed the label for a levonorgestrel EC product to warn of a reduction in effectiveness for women with higher BMI, garnering significant media coverage in the United States. Ulipristal acetate (UPA) EC may be more effective than levonorgestrel for women with BMI levels designated as obese. Among 8,019 women who received UPA from the online pharmacy KwikMed from 2011 to 2015 and self-reported their height, weight and reasons for seeking UPA online, we analyzed changes in the proportion of women in different BMI categories before and after the label change. For the 25 month-period after the label change, the proportion of women in the obese category rose by 26.7 percentage points relative to the 35 months before (B=0.2665, p<0.01). Mean BMI (25.5 versus 29.4, p<0.001) and average weight (148.6 pounds versus 175.5 pounds, p<0.001) of users were higher after the label change. Some women appear to have acted on the information that EC efficacy may be associated with body weight.

Keywords: Emergency contraception, online pharmacy, ulipristal acetate, reproductive health, obesity

Background

Two formulations of dedicated emergency contraceptive (EC) pills are available in the United States: levonorgestrel (LNG) 1.5 mg and ulipristal acetate (UPA) 30 mg. LNG EC is available over-the-counter and is generally easier to obtain than prescription-only UPA. EC pills are more effective the sooner they are taken because they work by preventing or disrupting ovulation (Gemzell-Danielsson, Berger, and Lalitkumar 2013). UPA faces numerous challenges of access for women, including low awareness among providers and inconsistent availability in pharmacies (Shigesato et al. 2018; Bullock et al. 2015; Batur et al. 2016). Online availability of UPA may be an important way for individual women to obtain this time-sensitive medication when they need it.

Access barriers to UPA are important because UPA appears to be more effective than LNG. UPA can disrupt ovulation after the luteinizing hormone surge has begun (Brache et al. 2010; Li et al. 2016), a point at which LNG EC is ineffective (Noe et al. 2011; Novikova et al. 2007). UPA may be particularly more effective for women with higher body mass index (BMI); one study concluded that LNG was no more effective than placebo for women with BMI of 26 kg/m2 or greater, while UPA maintained some efficacy up to a BMI of 35 kg/m2. In 2013, a new analysis of previously published data (Kapp et al. 2015) prompted European regulatory authorities to approve a label change for Norlevo®, a 1.5 LNG EC product, to warn that it was less effective in women weighing 165 pounds or more and ineffective in women weighing more than 176 pounds (Irish Medicines Board 2014). A recent analysis of four trials conducted by the World Health Organization (WHO) also found an increased risk of pregnancy among women who were obese, although the findings appear to be driven by anomalous results from one of the four studies (Festin et al. 2017).

These findings are of potential clinical significance, but have important limitations (Cleland and Wood 2014). In July 2014, the European Medicines Agency (EMA) reversed the 2013 label change (European Medicines Agency 2014), and in May 2016 the US Food and Drug Administration (FDA) determined that a label change was not warranted (US Food and Drug Administration 2016). However, the fact that regulatory agencies did not find the evidence available at the time compelling does not mean that a true relationship between efficacy of EC and body weight does not exist. Two recent pharmacokinetic studies found that serum concentration levels in women who took LNG 1.5 mg were about 50% lower in women who were obese compared with women who were not obese (Edelman et al. 2016; Praditpan et al. 2017).

If LNG EC is less effective in women who weigh more than 165 pounds, millions of women may be at pregnancy risk despite having taken EC, as the average American woman aged 20 years and older weighs 166 pounds (Fryar, Gu, and Ogden 2012). The present study evaluated changes in the proportion of women in different BMI categories seeking UPA EC through the online pharmacy KwikMed before and after the European label change for LNG EC and the media attention that followed.

Methods

Any potential influence of the label change and the research that motivated it on women’s EC usage requires awareness. To estimate the diffusion of information about the European label change, we analyzed the volume of online coverage using Google, which accounted for 63.4% of US internet searches as of October 2013, the month directly preceding the label change (Statista 2016). We searched for articles that appeared in the month before and after the label change (November, 2013) in online sources, including traditional media outlets, magazines, and blogs. To identify relevant articles, we conducted an online search using the term ‘emergency contraception effectiveness weight’. Although this is not necessarily the search term that a woman needing emergency contraception would use, we believe that it indicates online availability of this information.

Study Sample and Data Collection

Our data collection methods are described in detail elsewhere (Smith et al. 2017); in brief, customers purchasing UPA EC from KwikMed answered required medical screening questions in addition to optional survey questions. KwikMed offers next-day shipping and serves US customers who live in states that allow prescription medications to be shipped across state lines (all but North Carolina, Louisiana, Arkansas, and Missouri). Customers self-reported their height and weight, from which we calculated BMI using the following formula: BMI = [(weight in pounds/height in inches2) × 703] (US National Library of Medicine 2016). For our analyses, we divided BMI into three categories, derived from the WHO classification scheme: healthy weight (BMI<25), overweight (BMI≥25 and <30) and obese (BMI ≥ 30) (World Health Organization 2016). Small proportions of the sample were in the “Underweight” (BMI<18.5; 8.2%) and “Extremely obese” (BMI ≥40; 7.7%) categories, so we combined these categories with the next-closest categories for analysis. We also calculated the woman’s age in years using the reported birth date and order date.

Data Analysis

To test whether the BMI composition of the population of women seeking UPA EC from KwikMed changed after the European label change and subsequent publicity, we employed a linear regression of monthly data. We regressed the proportion of monthly orders to women who were obese on whether the month came after the label change in November 2013, using 60 months of data (January 2011-December 2015). Focusing on the proportion of orders to women who were obese, rather than number of orders, reduces the impact of changing monthly volume of orders, including a significant drop in all orders when UPA EC was temporarily unavailable due to a supply chain disruption in April 2014. We also estimated this model controlling for the average age of women ordering UPA in the month. We used t-tests to compare the average BMI of users before and after the label change. As a sensitivity check, we also conducted our analysis using weight, rather than BMI, as the outcome. We present some illustrative open-ended responses about why women were seeking UPA. All quotations from participants are presented in their original form, exactly as entered by participants. Our data include all orders filled from January 2011 through December 2015 from women who provided height and weight information. The Institutional Review Board at Princeton University issued an exemption for this study.

Results

Information about the label change was widely available online. The term ‘emergency contraception effectiveness weight’, searched for a one-month period starting on the day of the label change, returned 5,160 results for this time period. Among these, 100% (N=50) of articles from the first five pages of results directly addressed the possibility that LNG EC was less effective in women with higher body weight, compared with 14% (N=7) for the prior month. All but six of the articles following the European label change specifically named an equivalent US brand and/or used a US product as the cover image; the six that did not mention a US brand were all published outside the US. Nearly half (n=24) of the top 50 articles following the label change mentioned UPA as an alternative for women with higher body weight. Articles about the label change appeared in a broad range of publications, including mainstream media outlets (such as the New York Times, Wall Street Journal, and CNN), online magazines (including Mother Jones and The Atlantic) and websites and blogs aimed at women (such as Mommyish and Our Bodies Ourselves).

Among the 8,019 women who received UPA EC from KwikMed between January 2011 and December 2015, 8,009 (99.9%) provided height and weight information (10 women, 0.1%, were missing height information, and 10 women, 0.1%, were missing weight information). The most frequently occurring BMI category, “Healthy weight”, included 37% of the sample, and after collapsing the “Obese” and “Extremely obese” categories, one-third of the sample was in the highest BMI category (Table 1).

Table 1:

Distribution of UPA EC Orders through KwikMed, by BMI and Year (n=8,009*)

N %
Body Mass Index (kg/m2)
Underweight (<18.5) 654 8.2
Healthy weight (18.5–24.9) 2,988 37.3
Overweight (25–29.9) 1,718 21.5
Obese (30–39.9) 2,034 25.4
Extremely obese (≥40) 615 7.7
Order Year
2011 653 8.2
2012 1,204 15.0
2013 1,921 24.0
2014 1,846 23.1
2015 2,385 29.8
Open in a new tab
*

Sample is comprised of all KwikMed customers approved for prescription UPA who reported their height and weight

The proportion of monthly orders from women in the “obese” BMI category increased substantially after November 2013 (Figure 1). For example, in October 2013, 22% (n=34) of orders were from women with BMI greater than 30, while in December 2013, 39% (n=87) of orders were from women in this category. In a bivariate regression of monthly UPA orders, we found that the proportion of women ordering UPA who were in the obese category was significantly higher for months after the label change than before. On average, 19.5% of monthly orders before the label change were for women who were obese compared to 46.4% in months after the label change (p<0.01). Whether the month preceded or followed the European label change explained 86.4% of the variance in the monthly proportion of orders from women who were obese in this regression. No change was observed in the proportion of users who were in the overweight BMI category after the label change, but the share of monthly orders to underweight/normal weight women decreased from 61.1% to 34.4% (p<0.01). Models that also controlled for average age of women who received a UPA order in a given month produced the same results.

Figure 1.

Figure 1

Open in a new tab

Proportion of UPA EC Orders Fulfilled per Month from KwikMed, by BMI Category

We also tested for differences in average BMI between the months before and after the announced regulatory change, and found that the mean BMI of users was higher after the label change (25.5 before the label change versus 29.4 after, p<0.001). To test the robustness of our findings to measurement specification, we conducted a bivariate regression of average user weight on whether the order month preceded or followed the label change. We found that the average weight of users was 26.8 pounds higher after the label change (p<0.001).

Of the 641 women who provided an open-ended response, 420 (66%) gave additional context for these data by specifically noting their weight or BMI as a motivating factor for seeking UPA EC; 95% of these comments were made after the European label change and media attention in November 2013. For example, users commented: “My BMI is just out of the effective range of plan b type contraception” (age 23, BMI=32, February 2015) and “I am a larger chick and plan b isn’t as effective” (age 36, BMI=35, March 2015). Several participants noted that a health care provider, such as Planned Parenthood, recommended that they use UPA EC because of their weight: “I went to planned parenthood and they told me brands like Plan B are not as effective the heavier you are but ella is!” (age 19, BMI=31, April 2015).

Discussion

News of the European label change cautioning that LNG EC may not be effective for women at higher body weights was readily available online to women in the US. Even though the regulatory change applied only to Europe, nearly all of the top links in our internet search named an equivalent US product and discussed the implications of the underlying research for women in the US. This coverage may have motivated some US women to seek UPA EC as an alternative, as KwikMed’s proportional sales of UPA among women in the “obese” BMI category increased substantially after this change. Some women expressed concern, and even anger, that this information had not been available to them.

A publication in the academic literature raised this issue in 2011 (Glasier et al. 2011), but until regulatory action was taken, this finding was not widely publicized to the general public. The timing of shifts in the body mass index and weight composition of the population of KwikMed users suggests that some of the change found in our data may have been driven by availability of online information about the European label change and its relevance for brands of LNG EC available in the US. However, user comments indicated that some KwikMed customers were seeking UPA because of their weight at the advice of a Planned Parenthood health center or other health care provider who may have changed their protocols and practice independent of the label change.

The results of this study suggest that news of a regulatory change outside the US may have affected some women’s decision-making about EC, driving women in the obese BMI category to seek UPA. The finding that weight may reduce the efficacy of LNG EC is of potential clinical and public health importance, but as the EMA and the FDA concluded after their data reviews, more evidence is needed to clarify this relationship precisely. New pharmacokinetic research helps illuminate the mechanisms through which weight may affect the efficacy of EC, but clinical trials with pregnancy as the endpoint may be the only way to specify this relationship accurately. Because pregnancy following use of EC is a rare outcome, such trials must be large and are therefore expensive; it is unlikely that manufacturers of EC will find it in their economic interests to invest in these trials. Until more data are available, women and their health care providers must make the best decisions that they can in a complex and uncertain landscape. Our study results suggest that women may be self-referring for the care that they understand to be most appropriate for them based on what they read online; it is therefore important to prioritize ensuring that high-quality, accurate information is widely available.

Limitations

Our findings are limited to the population of women in the United States purchasing UPA EC via KwikMed; we are unable to draw any wider conclusions about whether women in the general population sought UPA because of their weight. In addition, our survey did not specifically ask whether women had chosen UPA because of concerns about weight, but many women directly communicated this concern through open-ended responses. Furthermore, because our estimates are based on observational data, we are unable to make a definitive causal statement that the public information available following this label change was solely responsible for the shift in BMI composition in this population of consumers. Further, as in any study of this type, it is also possible that confounders, for which we did not have information (e.g., smoking) and that we thus were unable to control, may have accounted for some of the results. Additionally, because the weight data in our study are self-reported, misclassification of weight is possible and may bias our results.

We also did not conduct a rigorous media analysis of coverage of the European label change, but used internet searches for the months before and after the change to approximate the extent of the coverage. A wide range of sources published information about this change, but we cannot quantify the number of women who read these articles or whether the women seeking UPA EC through KwikMed read this information. However, we believe that the substantial shift in body composition of women seeking UPA through KwikMed directly after the label change strongly suggests that women with higher BMIs received the information that UPA might be a more effective choice for them. Finally, the vast majority of LNG EC is sold over-the-counter in pharmacies by different companies; so we did not have an accurate way to represent LNG EC sales.

Conclusions

Following a label change in Europe and subsequent media attention in the US warning that levonorgestrel EC may be ineffective for women at higher body weights, the proportion of women ordering UPA online categorized as “obese” rose significantly. This suggests that diffusion of this information had some effect on women’s decision-making around their options for EC. Further research is needed to clarify the relationship between the efficacy of EC and weight and to understand how women respond to the availability of this information.

Acknowledgements:

The authors thank Peter Ax, Brian Ackley, and Dan Snyder at KwikMed for their willingness to partner on this research endeavor.

Funding: During the writing of this manuscript, the authors received support provided by Population Research Infrastructure Program (P2CHD047879) awarded to the Office of Population Research at Princeton University by the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

References

  1. Batur P, Cleland K, McNamara M, Wu J, Pickle S, and EC Survey Group. 2016. Emergency contraception: A multispecialty survey of clinician knowledge and practices. Contraception 93 (2): 145–152. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brache V, Cochon L, Jesam C, Maldonado R, Salvatierra AM, Levy DP, Gainer E, and Croxatto HB 2010. Immediate pre-ovulatory administration of 30 mg ulipristal acetate significantly delays follicular rupture. Human Reproduction 25 (9): 2256–2263. [DOI] [PubMed] [Google Scholar]
  3. Bullock H, Steele S, Kurata N, Tschann M, Elia J, Kaneshiro B, and Salcedo J 2015. Pharmacy access to ulipristal acetate in Hawaii: Is a prescription enough? Contraception 92 (4): 388–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cleland K and Wood S 2014. A tale of two label changes. Contraception 90 (1): 1–3. [DOI] [PubMed] [Google Scholar]
  5. Edelman AB, Cherala G, Blue SW, Erikson DW, and Jensen JT 2016. Impact of obesity on the pharmacokinetics of levonorgestrel-based emergency contraception: single and double dosing. Contraception 94 (1): 52–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. European Medicines Agency. Levonorgestrel and ulipristal remain suitable emergency contraceptives for all women, regardless of bodyweight http://www.ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2014/07/news_detail_002145.jsp&mid=WC0b01ac058004d5c1.
  7. Festin MP, Peregoudov A, Seuc A, Kiarie J, and Temmerman M 2017. Effect of BMI and body weight on pregnancy rates with LNG as emergency contraception: Analysis of four WHO HRP studies. Contraception 95 (1): 50–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Fryar CD, Gu Q, and Ogden CL 2012. Anthropometric reference data for children and adults: United States, 2007–2010 National Center for Health Statistics. [PubMed] [Google Scholar]
  9. Gemzell-Danielsson K, Berger C, and Lalitkumar PGL 2013. Emergency contraception - mechanisms of action. Contraception 87 (3): 300–308. [DOI] [PubMed] [Google Scholar]
  10. Glasier A, Cameron S, Blithe D, Scherrer B, Mathe H, Levy D, Gainer E, and Ulmann A 2011. Can we identify women at risk of pregnancy despite using emergency contraception? Data from randomized trials of ulipristal acetate and levonorgestrel. Contraception 84 (4): 363–367. [DOI] [PubMed] [Google Scholar]
  11. Irish Medicines Board. Norlevo 1.5mg, Summary of Product Characteristics Irish Medicines Board, accessed August 9, 2016, http://www.medicines.ie/history/11933/SPC/Norlevo+1.5mg+tablet. [Google Scholar]
  12. Kapp N, Abitbol JL, Mathe H, Scherrer B, Guillard H, Gainer E, and Ulmann A 2015. Effect of body weight and BMI on the efficacy of levonorgestrel emergency contraception. Contraception 91 (2): 97–104. [DOI] [PubMed] [Google Scholar]
  13. Li HW, Lo SS, Ng EH, and Ho PC 2016. Efficacy of ulipristal acetate for emergency contraception and its effect on the subsequent bleeding pattern when administered before or after ovulation. Human Reproduction 31 (6): 1200–1207. [DOI] [PubMed] [Google Scholar]
  14. Noe G, Croxatto H, Salvatierra A, Reyes V, Villarroel C, Munoz C, Morales G, and Retamales A 2011. Contraceptive efficacy of emergency contraception with levonorgestrel given before or after ovulation. Contraception 84 (5): 486–492. [DOI] [PubMed] [Google Scholar]
  15. Novikova N, Weisberg E, Stanczyk F, Croxatto H, and Fraser I 2007. Effectiveness of levonorgestrel emergency contraception given before or after ovulation--a pilot study. Contraception 75 (2): 112–118. [DOI] [PubMed] [Google Scholar]
  16. Praditpan P, Hamouie A, Basaraba CN, Nandakumar R, Cremers S, Davis AR, and Westhoff CL 2017. Pharmacokinetics of levonorgestrel and ulipristal acetate emergency contraception in women with normal and obese body mass index. Contraception 95 (5): 464–469. [DOI] [PubMed] [Google Scholar]
  17. Shigesato M, Ella J, Tschann M, Bullock H, Hurwitz E, Wu Y, Salcedo J 2018. Pharmacy access to ulipristal acetate in major cities throughout the United States. Contraception 97 (3):264–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Smith NK, Cleland K, Wagner B, Trussell J 2017. “I don’t know what I would have done.” Women’s experiences acquiring ulipristal acetate emergency contraception online from 2011 to 2015. Contraception 95 (4): 414–418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Statista. Share of search queries handled by leading U.S. search engine providers as of October 2013 accessed July 2016, https://www.statista.com/statistics/267161/market-share-of-search-engines-in-the-united-states/.
  20. US Food and Drug Administration. FDA communication on levonorgestrel emergency contraceptive effectiveness and weight, accessed May 24, 2017, http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm109775.htm.
  21. US National Library of Medicine. Body mass index, accessed July 7, 2017, https://www.nlm.nih.gov/medlineplus/ency/article/007196.htm.
  22. World Health Organization. Obesity and overweight, accessed October 11, 2017, http://www.who.int/mediacentre/factsheets/fs311/en/.

RESOURCES