Oral contraceptive effectiveness according to body mass index, weight, age, and other factors

Article Outline

• Abstract
• Materials and Methods
  • Study design
  • Study participants
  • Baseline questionnaire and follow-up
  • Analysis of contraceptive failure
• Results
  • Contraceptive failure
  • BMI
  • Weight
  • Age
  • Other factors
• Comment
• Acknowledgments
• References
• Copyright

Objective

The purpose of this study was to assess the use-effectiveness of oral contraceptives (OCs) in Europe according to body mass index (BMI), weight, age, and other factors.

Study Design

In a planned secondary analysis, we used data from the European Active Surveillance Study on Oral Contraceptives, which was a prospective active cohort surveillance study of 59,510 OC users, to assess the effectiveness of OCs overall and by BMI, weight, age, duration of use, ethinylestradiol dose, regimen type, starting/switching status, and parity. Self-reported unplanned pregnancies during OC use were confirmed by interview.

Results

An analysis of OC effectiveness (112,659 women-years of exposure and 545 unplanned pregnancies) found little variation in effectiveness by BMI/weight. Failure rates decreased after 30 years of age and with an increasing duration of use.

Conclusion

OC users in Europe reported high contraceptive effectiveness with “typical use.” Failure rates decreased with age and duration of use. BMI and weight had little, if any, influence on effectiveness.

Key words: body mass index, contraceptive effectiveness, oral contraceptives, weight

Oral contraceptives (OCs) are a popular method of reversible birth control that provide women with convenience and flexibility in family planning. OCs are extremely effective at preventing unplanned pregnancies, and their consistent and correct use is associated with an unintended rate of pregnancy as low as 0.3% during the first year of use.¹

Typically, preapproval trials that have investigated the efficacy of OCs have been of insufficient size to allow subgroup comparisons regarding unplanned pregnancies. In addition, preapproval trials often have somewhat narrow entry criteria; for example, women who are overweight or obese generally have been excluded from trials of OCs,² despite the fact that data on the relationship between contraceptive effectiveness and body composition are conflicting. Several typical use studies that have assessed contraceptive effectiveness according to weight and body mass index (BMI) have suggested that overweight and obese women may be at an increased risk of contraceptive failure.³, ⁴, ⁵ Other studies, however, show no significant association between increasing weight/BMI and contraceptive failure.⁶, ⁷, ⁸, ⁹

The European Active Surveillance Study on Oral Contraceptives (EURAS-OC) was a large-scale, prospective active cohort surveillance study of OC users in 7 European countries.¹⁰ The study was carried out as a Bayer Schering Pharma–sponsored phase IV commitment to European drug authorities, and followed OC users for up to 5 years. The main outcome of interest was the occurrence of cardiovascular events.

Using the large EURAS-OC dataset, we assessed the contraceptive effectiveness of OCs overall and according to BMI, weight, age, duration of use, dose of ethinylestradiol, type of regimen, starting/switching status, and parity.

Back to Article Outline

Materials and Methods 

Study design 

EURAS-OC was a large-scale, prospective, active cohort surveillance study conducted between 2000 and 2005 that characterized and compared the short- and long-term risks of both drospirenone-containing and other progestin-containing OCs in a large cohort of OC users in Europe. The main clinical outcome of interest in EURAS-OC was the occurrence of cardiovascular events (ie, the occurrence of venous thromboembolism, myocardial infarction, stroke, arrhythmia, and sudden death). Secondary a priori-defined outcomes included contraceptive failure and return to fertility after OC cessation. The study was designed to test the noninferiority of drospirenone-containing OCs vs levonorgestrel-containing OCs and other progestin-containing OCs in terms of venous thromboembolism risk. To this end, a sample size of 50,000 women with a total OC exposure of at least 100,000 women-years (WY) was considered sufficient.

Study participants 

Recruitment into the study was conducted through a network of 1113 study centers in 7 European countries (Austria, Belgium, Denmark, France, Germany, The Netherlands, and the United Kingdom) between November 2000 and June 2004. All women who received a prescription for a new OC at the participating centers were asked by the physician if they were willing to participate in the study. Study participation was discussed after a prescribing decision had been made to avoid physicians' prescribing behavior being affected by study participation. All women who were enrolled had to be either first-ever users of OCs or existing OC users who were switching to a new OC to avoid enrolling long-term users into the study and to ensure comparability of cohorts.¹⁰ More specific inclusion and exclusion criteria were not applied because of the noninterference nature of the study.

The study was approved by the Berlin Medical Board Ethics Commission. All women provided written informed consent before entry into the study.

Baseline questionnaire and follow-up 

After informed consent was given, participants provided baseline data with a self-completed questionnaire that addressed their state of health and potential risk factors and their height, weight, and prescribed OC. Each participant's medical history, which included medication history and history of OC use, and contact details, which included contact details of relatives, friends, and gynecologists/primary care physicians who could act as back-up contacts, were documented. In most cases, women completed the questionnaire immediately (ie, while at the physician's office). The physicians and their office staff checked the questionnaires for completeness and accuracy (eg, name of the prescribed OC).

Follow-up assessments were completed by the OC users on a 6-month basis for a period of up to 5 years. Follow-up questionnaires addressed the occurrence of adverse events and, if applicable, reasons for OC cessation or a switch to another OC. All questionnaires were reviewed for completeness and plausibility/consistency. In the case of missing and/or inconsistent information, women were contacted directly by telephone. Where necessary, information that was provided by participants regarding adverse events was clarified and validated with the diagnosing and/or treating physician.

Among other items, the 6 monthly follow-up questionnaires were designed to collect information on self-reported height and weight and on regular and continued OC use. The questionnaires did not collect information on the frequency of intercourse. The questionnaires also addressed the occurrence of pregnancy during OC use. All women who self-reported an unplanned pregnancy while using their OC were interviewed by a study physician to analyze in more detail potential reasons for OC failure. The interviews established whether any pills might have been missed or whether the OC might have been used incorrectly.

To ensure a low loss-to-follow-up rate, a multifaceted 4-level follow-up procedure was adopted. In the first instance, questionnaires were mailed to the participating women to be completed and returned, as were 2 reminder letters in the case of no response (level 1). In the absence of a completed questionnaire, multiple attempts were made to contact the participants and their relatives, friends, and gynecologists/primary care physicians (level 2), and searches in national and international telephone and address directories were initiated (level 3). If these measures were unsuccessful, an official address search through governmental administrations was conducted (level 4).

Analysis of contraceptive failure 

Calculations of contraceptive failure were based on all reported unplanned pregnancies that occurred during OC use. Study physicians contacted all study participants who, in a follow-up questionnaire, self-reported an unplanned pregnancy, despite OC use; the interviewers ascertained the onset of the pregnancy and whether the subject had used the OC during the month that the pregnancy began. The numerator for the rate of contraceptive failure was all confirmed unplanned pregnancies during the relevant time period. The denominator was the total OC exposure during the same time period. Study participants who continued using the OCs during episodes of abstinence were included in the denominator because of a lack of information about these episodes. Contraceptive failure rates were calculated with the use of both the Pearl Index (PI; total number of all unintended pregnancies per 100 WYs of use) and life-table estimate methods (rate of contraceptive failure for each time interval [eg, months]) in agreement with current guidelines on the clinical investigation of steroid contraceptives in women that were published by the European Medicines Agency.¹¹

Contraceptive failure was assessed overall and according to BMI, weight, age, duration of use, dose of ethinylestradiol, type of regimen, starting/switching status, and parity. Stratified analyses were performed for these individual factors according to the following categories: BMI, <20.0, 20.0-24.9, 25.0-29.9, and ≥30.0 kg/m²; weight, <55.0, 55.0-59.9, 60.0-64.9, 65.0-69.9, 70.0-74.9, and ≥75.0 kg; age, <18, 18-24, 25-29, 30-34, 35-39, and ≥40 years; duration of OC use, 1, 2, 3, and 4 years of use; dose of ethinylestradiol, <30 and ≥30 μg; regimen type, monophasic and multiphasic regimens; starting/switching status, first-ever users of an OC and existing OC users switching to a new OC; and parity, nulliparous, and primi-/multiparous women. To allow for a robust analysis of up to 6 subcategories for each individual factor, 10,000 WYs of exposure (≥1500 WYs per subgroup) were needed to ensure that the estimate was precise and the bounds of the 95% confidence interval (CI) were sufficiently narrow. As such, separate subgroup analyses by progestin type were performed only for the 5 OCs that contained progestins with an exposure of ≥10,000 WY.

Ninety-five percent confidence limits for PI and life-table estimates were calculated according to the methods of Gerlinger et al.¹² Similar to the analyses of other time-to-event outcomes of the EURAS-OC study,¹⁰ inferential statistics for contraceptive failure were based on Cox regression analysis. Each of the factors mentioned earlier was included in the Cox regression model.

Back to Article Outline

Results 

A total of 59,510 women were enrolled in EURAS-OC. Of these, 836 women were excluded because of protocol violations (failure to sign an informed consent form, enrolled ≥2 times by ≥1 study center, continued use of an existing OC, or failure to initiate treatment). Overall, 58,674 women were observed for 142,475 WYs. The maximum individual duration of follow-up evaluation was up to 5 years (mean, 2.4 years). Overall, 1401 women (2.4%) were lost to follow-up during the study (Figure 1).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 1. 

  Flow chart of women in the study

• OC, oral contraceptive; WY, women-years.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

The baseline demographic characteristics of all women who were included in EURAS-OC have been reported previously.¹⁰ In brief, the mean age was 25.2 ± 8.2 years; the mean weight and BMI were 63.1 ± 12.1 kg and 22.1 ± 4.1 kg/m², respectively. Overall, 11,997 women (20.4%) were OC starters (ie, first-ever users of OCs).

Five progestins fulfilled the criterion of an exposure of >10,000 WYs for the current analysis: drospirenone, levonorgestrel, chlormadinone acetate (CMA), desogestrel, and dienogest. The baseline characteristics of women who were included in the current analysis are shown in Table 1. At baseline, users of OCs that contained drospirenone tended to have a slightly higher weight and BMI, compared with women who used OCs that contained other progestins.

TABLE 1. Baseline characteristics

Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

Contraceptive failure 

The overall analysis of contraceptive failure was based on a total OC exposure of 112,659 WYs and 545 reported unplanned pregnancies, which resulted in a PI of 0.48 (95% CI, 0.44–0.53). The OC exposures for the subgroup analyses are given in Table 2. The life-table estimate of the rate of contraceptive failure ranged from 0.75% (95% CI, 0.68–0.82%) after the first year of OC use to 1.67% (95% CI, 1.61–1.85%) after the fourth year of OC use (Table 3). A comprehensive analysis of potential reasons for contraceptive failure showed that 230 (42.2%), 100 (18.3%), and 85 (15.6%) unplanned pregnancies were reported to be associated with irregular OC intake, vomiting and/or diarrhea, and use of antibiotics, respectively. In 99 cases of contraceptive failure (18.2%), the analysis suggested perfect OC use. The reasons for unplanned pregnancy could not be analyzed in 31 women (5.7%) because of missing information.

TABLE 2. OC exposure in selected subgroups

Subgroup	Exposure (women-years)
Regimen type
Monophasic OCs	99,798
≥30 μg ethinylestradiol	80,306
<30 μg ethinylestradiol	19,492
Multiphasic OCs	12,211
Duration of OC use
First year	48,988
Second year	34,490
Third year	20,161
Fourth year	8,928
Age, y
<18	21,225
18-24	43,580
25-29	16,605
30-34	12,881
35-39	10,005
≥40	8,363
Parity
Nulliparous	64,750
Primi-/multiparous	47,832

OC, oral contraceptive.

Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

TABLE 3. Life-table estimates of the rate of contraceptive failure after 1, 2, 3, and 4 years of OC use

Variable	Year
	1	2	3	4
Rate of contraceptive failure (%)	0.75(0.68-0.82)	1.33(1.24-1.42)	1.53(1.41-1.66)	1.67(1.61-1.85)

Data are shown as point estimate (95% confidence limit).

OC, oral contraceptive.

Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

Separate subgroup analyses were performed for those OCs that contain progestins with an exposure of ≥10,000 WYs of observation, which included CMA, desogestrel, dienogest, drospirenone and levonorgestrel. The PI for each progestin was as follows: CMA, 0.48 (95% CI, 0.35–0.63); desogestrel, 0.50 (95% CI, 0.40–0.63); dienogest, 0.40 (95% CI, 0.30–0.53); drospirenone, 0.44 (95% CI, 0.37–0.53); and levonorgestrel, 0.53 (95% CI, 0.46–0.62). Life-table estimates of the rate of contraceptive failure after the first year of OC use for OCs that contain these progestins were 0.78% (95% CI, 0.57–1.06%), 0.88% (95% CI, 0.69–1.12%), 0.57% (95% CI, 0.41–0.79%), 0.62% (95% CI, 0.50–0.76%), and 0.78% (95% CI, 0.57–1.06%), respectively (Figure 2). There was no significant difference in contraceptive effectiveness according to the type of progestin that was received.

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 2. 

  Life-table estimates of contraceptive failure

• Life-table estimates of the rate of contraceptive failure for oral contraceptives (OCs) that contain chlormadinone acetate (CMA), desogestrel (DSG), dienogest (DNG), drospirenone (DRSP), and levonorgestrel (LNG). The 95% CIs are not shown to avoid overlapping graphic elements. Typically, confidence limits were clearly smaller than ± 0.002 and ± 0.005 in years 1 and 2 and in years 3 and 4.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

BMI 

An analysis of all 5 progestins combined showed little variation in contraceptive failure according to BMI (hazard ratio [HR], 1.00; 95% CI, 0.98–1.03 in the Cox regression analysis; Figure 3). An analysis of the individual progestins indicated little variation in contraceptive failure according to BMI with desogestrel, dienogest, drospirenone, and levonorgestrel. In contrast, the contraceptive failure of CMA-containing OCs showed a statistically significant correlation with BMI (P = .028); an increased BMI (≥30 kg/m²) being associated with an increased rate of contraceptive failure (Figure 4).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 3. 

  Oral contraceptive failure vs overall body mass index

• Data are shown as contraceptive failure rate per 100 women-years of exposure.

• BMI, body mass index.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 4. 

  Oral contraceptive failure vs body mass index by progestin type

• Data are shown as contraceptive failure rate per 100 women-years of exposure to oral contraceptives that contain chlormadinone acetate (CMA), desogestrel (DSG), dienogest (DNG), drospirenone (DRSP), and levonorgestrel (LNG).

• BMI, body mass index.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

Weight 

The results of the weight analyses were similar to the BMI analyses. An analysis of all 5 progestins combined showed that contraceptive failure was affected only slightly by weight. When each of the progestins was analyzed separately, it was shown (with the exception of CMA; P = .024) that weight had little or no effect on the contraceptive failure of desogestrel, dienogest, drospirenone, and levonorgestrel.

Age 

Contraceptive failure vs age followed a biphasic pattern; the highest rate of contraceptive failure was observed in the 18- to 24-year-old group, and a marked decline was observed in women >29 years old (Figure 5). The HR for contraceptive failure in women ≥30 years old vs <30 years old was 0.20 (95% CI, 0.15–0.27; P < .001).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 5. 

  OC failure vs age

• Data are shown as contraceptive failure rate per 100 women-years of exposure.

• OC, oral contraceptive.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

Other factors 

Contraceptive failure was also assessed according to duration of use, dose of ethinylestradiol, type of regimen, starter/switching status, and parity. The rate of contraceptive failure decreased with increasing duration of use (Figure 6). No clinically relevant difference in the rate of contraceptive failure was observed with OC preparations that contained an ethinylestradiol dose of <30 or ≥30 μg or with OC preparations that were monophasic or multiphasic in design. The overall PI in women who received a dose of ethinylestradiol of <30 μg was 0.57 (95% CI, 0.46–0.69), compared with 0.47 (95% CI, 0.42–0.52) for women who received a dose of ethinylestradiol of ≥30 μg. Life-table estimates of contraceptive failure at year 1 were 0.84% (95% CI, 0.64–1.08%) and 0.73% (95% CI, 0.63–0.84%) for ethinylestradiol doses of <30 μg and ≥30 μg, respectively. The overall PI in women who received multiphasic regimens (0.51; 95% CI, 0.39–0.65) was similar to that in women who received monophasic regimens (0.48; 95% CI, 0.44–0.53). Life-table estimates of contraceptive failure at year 1 were 0.74% (95% CI, 0.65–0.83%) and 0.78% (95% CI, 0.56–1.06%) for multiphasic regimens and monophasic regimens, respectively. Whether a woman was a first-ever user of an OC (starter) or was switching from another OC had no impact on the rate of contraceptive failure, after adjustment for age. Finally, contraceptive failure rates were higher in primi-/multiparous women than in nulliparous women (adjusted HR, 1.55; 95% CI, 1.40–1.70; P < .001).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 6. 

  OC failure vs duration of use

• Data are shown as life-table estimates at years 1, 2, 3, and 4, with 95% CI.

• CI, confidence interval; OC, oral contraceptive.

• Dinger. Oral contraceptive effectiveness. Am J Obstet Gynecol 2009.

Back to Article Outline

Comment 

The results of the current analysis, which was conducted in a large cohort of women from several European countries and based on a total OC exposure of 112,659 WYs, showed that OCs are associated with high contraceptive effectiveness during routine use in women in Europe. Both increasing age (for age groups of >24 years) and increasing duration of use were associated with lower contraceptive failure rates. Although BMI and absolute weight had little, if any, influence on the contraceptive effectiveness of OCs that contained dienogest, drospirenone, desogestrel, and levonorgestrel, an increased BMI (≥30.0 kg/m²) and an increased weight (≥75.0 kg) appeared to reduce the contraceptive effectiveness of CMA-containing OCs. The contraceptive effectiveness of CMA-containing OCs was still acceptable nonetheless; the contraceptive failure rate per 100 WYs of exposure was 1.49 in women with a BMI of ≥30.0 kg/m² and 1.17 in women with a bodyweight of ≥75 kg.

In contrast with the outcomes of some studies,³, ⁴, ⁵ we found that overweight and obese women who used OCs were not at an increased risk of contraceptive failure, which is in agreement with studies that have shown no significant association between increasing weight/BMI and contraceptive failure.⁶, ⁷, ⁸, ⁹ In the current analysis, BMI and weight had little or no impact on the contraceptive effectiveness of OCs, at least for those preparations that contained dienogest, drospirenone, desogestrel, and levonorgestrel as the progestin component.

The very low OC failure rates that were observed in this cohort suggest a high level of correct OC use in a “typical use” setting in Europe. It is not possible, however, to assess whether the low failure rates reported here for overweight and obese women would also be seen in populations with a substantial amount of incorrect OC use or in populations with a higher proportion of women with World Health Organization class II or III obesity (≥35 kg/m²). It is probable that most OCs are dosed adequately for OC users who are overweight or who are World Health Organization class I obese (BMI, ≥30.0-34.9 kg/m²); however, it is conceivable that many OCs are not dosed adequately for OC users who are World Health Organization class II or III obese. In short, although the current study shows little or no impact of body composition on contraceptive effectiveness in a compliant “typical use” population in Europe, the same may not be true in other populations, such as in the United States, where the rate of obesity is high. For this reason, studies in different populations with different characteristics are warranted.

Our analysis showed that there was little variation in the contraceptive effectiveness of OCs that contained drospirenone, levonorgestrel, desogestrel, and dienogest across various weight and BMI categories. In contrast, there was a significant correlation between body composition and contraceptive effectiveness for OCs that contained CMA, which is a progestin that is not marketed in the United States; increased BMI (≥30 kg/m²) and weight (≥75 kg) were associated with a higher rate of failure. The reduced contraceptive effectiveness of CMA-containing OCs in heavier women may reflect the fact that CMA is highly lipophilic and has been shown to accumulate in adipose tissue.¹³ In overweight and obese women in whom adipose tissue mass is increased, this could result in an altered volume of distribution and potentially lead to (temporary) subtherapeutic levels of systemic hormone.

The current analysis also showed that contraceptive failure by age followed a biphasic pattern, which is in line with findings that female fecundity peaks between age 20 and 30 years.¹⁴, ¹⁵, ¹⁶ A lower rate of contraceptive failure was found in women who were >30 years old and in women who had used OCs for a longer duration of time. The decrease in contraceptive failure rates with increasing age likely reflects the decline in the overall rate of fecundity with advancing age. Indeed, lower rates of planned pregnancy were observed in older women who stopped OC use during the EURAS-OC study with the specific intention of becoming pregnant.¹⁷ The reduced failure rate with an increasing duration of use may be related to the fact that those women who are most likely to experience contraceptive failure do so early, which results in a group of women that is increasingly dominated by the most conscientious users and/or the least fertile users.¹⁸

No clinically relevant difference in the rate of contraceptive failure was observed with OC preparations that contain an ethinylestradiol dose of <30 or ≥30 μg; both doses were associated with high effectiveness.

Parous women were more likely to experience a pregnancy during OC use than nulliparous women. This outcome was not unexpected; parity is associated with many factors, which include established fecundity in the past.¹⁸ Although parous women have “proved” their fertility, nulliparous women have not; thus, among any group of nulliparous women will be some women who are infertile (or have partners who are infertile) whose inclusion in the group will reduce the overall rate of fertility.¹⁹ This was illustrated by Howe et al,¹⁹ who showed that parity was associated strongly with fecundity. The finding that fertility is higher in parous women, compared with nulliparous women, was also observed in those women who discontinued using OCs in EURAS-OC with the intention of becoming pregnant.²⁰

The current study has limitations. Both weight and height were self-reported by the study participants, and it is known that the prevalence of obesity in women is underestimated when based on self-reported data. Indeed, studies suggest that the prevalence of obesity may be underestimated by as much as 30%,²¹, ²² and the most important determinant of underreporting BMI is a high BMI.²² Having noted this, an analysis of self-reported and measured data that were collected at gynecologic practices in women who use OCs has shown that the mean of self-reported weight is only 7% lower than the mean of objectively measured values (data not shown). Although it is possible that self-reporting of weight and height in the current analysis might have led to differential misclassification (with more women who were overweight or obese being misclassified as normal weight), this would not impact the findings of the current analysis because (1) any increase of failure rates would be overestimated by differential misclassification and (2) in general, no effect of body composition on the rate of contraceptive failure was observed. Weight changes over time; however, this was not an issue in EURAS-OC because self-reported weight was recorded at study entry and at up to 10 follow-up visits during the study. Only the last available weight value before the occurrence of an unplanned pregnancy was used for our analysis. Furthermore, sensitivity analyses that use baseline weights showed almost identical results.

A key limitation of the study is that the low rate of contraceptive failure that was observed in this European analysis may not be generalizable elsewhere. Specifically, this study identified OC failure rates that were substantially lower than those that were found in the 2002 cycle of the US National Survey of Family Growth (NSFG), in which the probability of failure of OCs during the first 12 months of use was 9%.²³ There are several possible explanations for the differences in contraceptive failure rates in this study vs the NSFG. First, failure rates with OCs and other hormonal contraceptives in Europe are known to be substantially lower than those in North America.²⁴, ²⁵, ²⁶ Second, it is possible that the women who chose to participate in EURAS-OC (and to undergo several years of follow-up) were more conscientious OC users, compared with women who refused to participate in the study. In contrast, the NSFG is a representative study of women in the United States that includes subgroups of women who are known to experience higher contraceptive failure rates. Also, the participants are not self-selected volunteers in a multiyear study. Importantly, the NSFG methods differ from the EURAS-OC methods with regard to the frequency of data collection, the definition of an unplanned pregnancy, and the definition of person-time in the denominator. In particular, the fact that the NSFG eliminated exposure months in which no sexual activity occurred may have contributed greatly to the large difference in failure rates in the 2 studies. As a final point, the NSFG reported failure rates for the first year of use only, and it is well known that first-year failure rates are much higher than long-term user rates (as shown in the EURAS-OC study).

In populations with higher contraceptive failure rates, such as women in the United States, the associations that are reported here might be different, depending on other factors that contribute to higher rates of contraceptive failure. Although there is essentially no association between obesity and OC failure in this population, it is at least plausible that such an association could emerge in women with higher failure rates, if obesity is associated with higher levels of incorrect OC use, which is something that has been shown with prescription medications other than OCs.²⁷

Nevertheless, these potential limitations should not detract from the strengths of the current analysis. First, EURAS-OC was a very large, multinational, prospective, active cohort surveillance study. Because the study was noninterventional in nature and specific inclusion and exclusion criteria were not applied, participants were representative of “typical” European users of OCs. Second, contraceptive failure was a prospectively defined endpoint of EURAS-OC, and the study was powered sufficiently to assess this outcome. Indeed, even in the smallest BMI category (≥30 kg/m²), data for >6000 WYs of exposure were available. Finally, the intensive follow-up procedure of EURAS-OC ensured an extremely low loss-to-follow-up rate of just 2.4% (<0.5% of participants lost at each of the semiannual follow-up time points). As a result of this low loss-to-follow-up rate, it may be assumed that the study results were not biased by a substantial underreporting of outcomes of interest and that the study is, in general, methodologically valid.¹⁰

In conclusion, the findings of this analysis show that OCs were associated with a high level of contraceptive effectiveness in a “typical use” setting (ie, outside of a clinical trial setting) in Europe. Contraceptive failure rates were found to decrease substantially from 30 years of age and as the duration of OC use increased. BMI and absolute weight had little, if any, influence on the contraceptive effectiveness of OCs that contained drospirenone, dienogest, desogestrel, and levonorgestrel.

Back to Article Outline

Acknowledgments 

We thank Lyndal McMillan and Danielle Turner from Wolters Kluwer Health for writing assistance during the preparation of this article.

Back to Article Outline

References 

1. Trussell J. Contraceptive efficacy. In:  Hatcher RA,  Trussell J,  Nelson AL,  Cates W,  Stewart FH,  Kowal D editor. Contraceptive technology. 18th ed.. New York: Ardent Media; 2007;p. 792
   • View In Article
2. Food and Drug Administration. Transcript: Advisory Committee for Reproductive Health Drugs, January 24, 2007. http://www.fda.gov/OHRMS/DOCKETS/AC/07/transcripts/2007-4274t2-part2.pdfAccessed Jan. 7, 2009
   • View In Article
3. Brunner Huber LR, Hogue CJ. The association between body weight, unintended pregnancy resulting in a livebirth, and contraception at the time of conception. Matern Child Health J. 2005;9:413–420
   • View In Article
   • MEDLINE
   • CrossRef
4. Holt VL, Cushing-Haugen KL, Daling JR. Body weight and risk of oral contraceptive failure. Obstet Gynecol. 2002;99:820–827
   • View In Article
   • MEDLINE
   • CrossRef
5. Holt VL, Scholes D, Wicklund KG, Cushing-Haugen KL, Daling JR. Body mass index, weight, and oral contraceptive failure risk. Obstet Gynecol. 2005;105:46–52
   • View In Article
   • MEDLINE
   • CrossRef
6. Brunner Huber LR, Hogue CJ, Stein AD, Drews C, Zieman M. Body mass index and risk for oral contraceptive failure: a case-cohort study in South Carolina. Ann Epidemiol. 2006;16:637–643
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (106 KB)
   • CrossRef
7. Brunner Huber LR, Toth JL. Obesity and oral contraceptive failure: findings from the 2002 National Survey of Family Growth. Am J Epidemiol. 2007;166:1306–1311
   • View In Article
   • CrossRef
8. Brunner LR, Hogue CJ. The role of body weight in oral contraceptive failure: results from the 1995 national survey of family growth. Ann Epidemiol. 2005;15:492–499
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (112 KB)
   • CrossRef
9. Vessey M. Oral contraceptive failures and body weight: findings in a large cohort study. J Fam Plann Reprod Health Care. 2001;27:90–91
   • View In Article
   • MEDLINE
   • CrossRef
10. Dinger JC, Heinemann LAJ, Kühl-Habich D. The safety of a drospirenone-containing oral contraceptive: final results from the European Active Surveillance study on Oral Contraceptives based on 142,475 women-years of observation. Contraception. 2007;75:344–354
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (267 KB)
    • CrossRef
11. European Medicines Agency Committee for Medicinal Products for Human Use. Guideline on clinical investigation of steroid contraceptives in women, July 27, 2005. http://www.emea.europa.eu/pdfs/human/ewp/051998en.pdfAccessed Jan. 6, 2009
    • View In Article
12. Gerlinger C, Endrikat J, van der Meulen EA, Dieben TO, Dusterberg B. Recommendation for confidence interval and sample size calculation for the Pearl Index. Eur J Contracept Reprod Health Care. 2003;8:87–92
    • View In Article
    • MEDLINE
    • CrossRef
13. Gallegos AJ, Gonzáles-Diddi M, Merino G, Martínez-Manautou J. Tissue localization of radioactive chlormadinone acetate and progesterone in the human. Contraception. 1970;1:151–161
    • View In Article
    • Abstract
    • Full-Text PDF (629 KB)
    • CrossRef
14. Hamilton BE, Sutton PD, Ventura SJ. Revised birth and fertility rates for the 1990s and new rates for Hispanic populations, 2000 and 2001: United States. Natl Vital Stat Rep. 2003;51:1–94
    • View In Article
    • MEDLINE
15. Larsen U, Yan S. The age pattern of fecundability: an analysis of French Canadian and Hutterite birth histories. Soc Biol. 2000;47:34–50
    • View In Article
    • MEDLINE
16. Wood JW. Fecundity and natural fertility in humans. Oxf Rev Reprod Biol. 1989;11:61–109
    • View In Article
    • MEDLINE
17. Cronin M, Möhner S, Minh Thai D, Westhoff C. Past oral contraception use does not negatively affect time to conception. Obstet Gynecol. 2007;109:3S
    • View In Article
18. Trussell J. Methodological pitfalls in the analysis of contraceptive failure. Stat Med. 1991;10:201–220
    • View In Article
    • MEDLINE
    • CrossRef
19. Howe G, Westhoff C, Vessey M, Yeates D. Effects of age, cigarette smoking, and other factors on fertility: findings in a large prospective study. Br Med J (Clin Res Ed). 1985;290:1697–1700
    • View In Article
    • MEDLINE
20. Cronin M, Möhner S, Dinger J, Schellschmidt I. Return to fertility after use of oral contraceptives: results from the EURAS study. In: Proceedings of the 9th World Congress on Controversies in Obstetrics, Gynecology & Infertility (COGI). Barcelona, Spain, March 22-25, 2007. 2007;A-68
    • View In Article
21. Nyholm M, Gullberg B, Merlo J, Lundqvist-Persson C, Rastam L, Lindblad U. The validity of obesity based on self-reported weight and height: Implications for population studies. Obesity (Silver Spring). 2007;15:197–208
    • View In Article
    • MEDLINE
    • CrossRef
22. Visscher TL, Viet AL, Kroesbergen IH, Seidell JC. Underreporting of BMI in adults and its effect on obesity prevalence estimations in the period 1998 to 2001. Obesity (Silver Spring). 2006;14:2054–2063
    • View In Article
    • MEDLINE
    • CrossRef
23. Kost K, Singh S, Vaughan B, Trussell J, Bankole A. Estimates of contraceptive failure from the 2002 National Survey of Family Growth. Contraception. 2008;77:10–21
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (172 KB)
    • CrossRef
24. Darney P. Safety and efficacy of a triphasic oral contraceptive containing desogestrel: results of three multicenter trials. Contraception. 1993;48:323–337
    • View In Article
    • Abstract
    • Full-Text PDF (1092 KB)
    • CrossRef
25. Ferguson H, Vree ML, Wilpshaar J, Eskes TK. Multicenter study of the efficacy, cycle control and tolerability of a phasic desogestrel-containing oral contraceptive. Eur J Contracept Reprod Health Care. 2000;5:35–45
    • View In Article
    • MEDLINE
    • CrossRef
26. Grubb GS, Archer DF, Constantine GD. Differences between the United States and Europe in clinical trials of hormonal contraceptive efficacy. Obstet Gynecol. 2008;111:63S
    • View In Article
27. Salas M, Kiefe CI, Schreiner PJ, et al. Obesity modifies the association of race/ethnicity with medication adherence in the CARDIA study. The Patient. 2008;1:41–54
    • View In Article