Impact of Combined Hormonal Contraceptive Use on Weight Loss: A Secondary Analysis of a Behavioral Weight Loss Trial

Abstract

Objective:

To perform a preliminary investigation of the impact of combined hormonal contraceptive (CHC) use on weight loss during an 18-month behavioral weight loss trial.

Methods:

Adults (n=170; 18-55 yrs; BMI 27-42 kg/m²) received a weight loss intervention that included a reduced-calorie diet, a progressive exercise prescription, and group-based behavioral support. Premenopausal women (n=110) were classified as CHC users (CHC, n=17) or non-CHC users (non-CHC, n=93). Changes in weight were examined within groups using a linear mixed model, adjusted for age and randomized group assignment.

Results:

At 6 months, weight was reduced from baseline in both CHC (mean, −6.7kg; 95% CI, −9.8 to −3.7kg) and non-CHC (−9.1kg; −9.1 to −6.4kg). Between 6 and 18 months, CHC regained weight (4.9kg; 0.9 to 8.9kg), while weight remained relatively unchanged in non-CHC (−0.1kg; −1.8 to 1.6kg). At 18 months, weight was relatively unchanged from baseline in CHC (−1.8kg; −7.3 to 3.6kg), and was reduced from baseline in non-CHC (−7.9kg; −10.2 to −5.5kg).

Conclusion:

In this secondary data analysis, CHC use was associated with weight regain after initial weight loss. Prospective studies are needed to further understand the extent to which CHC use influences weight loss and maintenance.

Introduction

Obesity rates have stabilized in men over the last decade, but continue to increase in women (1, 2). The prevalence of obesity (BMI ≥30 mg/kg²) has more than tripled between 1960 and 2014 among women of reproductive age (20-39 years) in the United States (1, 3). Comprehensive behavioral weight loss programs are the cornerstone of obesity treatment. Despite the short-term effectiveness of behavioral weight loss interventions, weight regain is a common challenge that undermines current obesity treatment (4, 5). Notably, behavioral weight loss interventions are generally less successful in women than in men (6).

The health risks of obesity may be greater for women compared to men (7, 8). In addition, obesity in women results in distinct health risks during pregnancy, delivery, and the postpartum period (9). Pregnant women with obesity face increased risk for gestational diabetes, hypertensive disorders, thromboembolic disorders, higher cesarean delivery rates, pre-term delivery, and wound infections (10). Infants of mothers with obesity are at increased risk of perinatal death, macrosomia, congenital abnormalities, and developing obesity (10, 11). Given these additional health risks, and the higher prevalence of obesity among women, there is a critical need to understand factors underlying sex differences in obesity treatment response.

A significant proportion of reproductive-age women use combined estrogen and progestin hormonal contraceptives (CHCs). The most commonly used CHC is the oral contraceptive pill (12.6%) followed by the transdermal patch and vaginal ring (2%) (12, 13). Commonly used CHCs contain synthetic versions of estradiol (E2, the primary estrogen in humans) and progesterone, which inhibit production of follicle stimulating hormone and luteinizing hormone, thereby preventing ovulation (14).

Most data suggest CHC use does not cause weight gain (15–17). To date, there are no known studies that have examined the impact of CHCs on women with overweight/obesity who are actively trying to lose weight and/or maintain weight loss. Data from a recently completed 18-month behavioral weight loss trial provided an opportunity to compare weight loss and weight loss maintenance between premenopausal women using CHCs to premenopausal women not using CHCs as a preliminary investigation of the potential impact of CHC use on weight loss and maintenance. Changes in fat mass (FM), lean mass (LM), waist circumference (WC), self-reported dietary energy intake (EI) and macronutrient intake, and device-measured moderate-to-vigorous physical activity (MVPA) and steps were also compared between groups.

Methods

Participants

This secondary data analysis included premenopausal women (n=110) who participated in an 18-month behavioral weight loss trial (NCT01985568) conducted at the University of Colorado Anschutz Health and Wellness Center. The study was approved by the Colorado Multiple Institutional Review Board. The randomized weight loss trial was designed to compare the efficacy of delivering diet and exercise interventions simultaneously versus sequentially. The primary outcome of the trial was weight change at 18 months. A detailed description of the methods for the interventional trial, as well as results for the primary and secondary outcomes has been reported previously (18). Participants in the interventional trial were healthy men and women with overweight/obesity (18-55 yrs, BMI 27-42 kg/m²) who were free of diabetes or cardiovascular disease. Potential participants were excluded if they were pregnant or lactating, or had lost >5% body weight the 6 months prior to enrolling in the study. Participants provided written informed consent and received compensation for participating. Women were excluded from the sample if they were postmenopausal (n=19) or had unknown menopausal status (n=4). One hundred nineteen premenopausal women were identified at baseline. Women were excluded from the secondary analysis if they went through menopause during the study (n=1), stopped or started CHC use during the trial (n=4), used non-contraceptive oral or transdermal estrogens and/or progestins (n=2), used progestin-only oral contraceptives (n=1), or were using an unknown type of oral contraceptive pill (n=1) (Figure 1).

Figure 1.

Flowchart of participants included in the current secondary analysis

Abbreviations: IUD – intrauterine device; OCP – oral contraceptive pill; CHC – combined hormonal contraceptives

^aNon-contraceptive oral medications included Activella (Estradiol 0.5-1mg/norethindrone 0.1-0.5mg) ashormone replacement therapy for perimenopause, and Micronized Progesterone 150mg for sleep.

^bOne participant switched from an OCP to the vaginal ring without interruption in CHC exposure during the study, thus the total from CHC users below is >17

^cParticipants were able to report more than one type of contraceptive method, so the totals from the non-CHC Users is >93.

^dOf 21 IUD users, 13 women were using Mirena^® IUD (levonorgestrel-releasing intrauterine system, 52mg), 5 women were using Paragard^® IUD (intrauterine copper contraceptive), and 3 women who did not know which kind of IUD she had in place.

Randomization and Blinding

Justification for sample size (n=170), randomization, and the consort diagram for the interventional trial have been published previously (18). Given that weight change was not different between randomized groups at 18 months (18), we combined randomized groups for this analysis to increase sample size and preserve statistical power. Randomization was used as a covariate in statistical models.

Diet and exercise interventions

As previously reported, both randomized groups in the interventional trial received a reduced-calorie diet intervention along with group-based behavioral support. The standard group simultaneously began a supervised exercise program progressing to 300 minutes/week of moderate intensity aerobic exercise (65-75% maximum heart rate) during months 0-6. The sequential group was asked not to begin exercise during months 0-6, and received an identical supervised exercise program during months 7-12. Upon completion of the 6-month supervised exercise program, both groups were asked to continue to exercise 300 minutes/week at moderate-intensity for the duration of the 18-month study.

Behavioral support was delivered in a group setting by a registered dietician and focused on dietary behavior modification using cognitive-behavioral strategies. Group meetings lasted approximately 60 minutes and were held weekly during weeks 0-20, every other week during weeks 21-26, and every month during weeks 27-78. The targeted EI for women was between 1200-1600 kcal/day and the targeted macronutrient content was 20-30% fat, 50-55% carbohydrate, and 20-25% protein.

Combined Hormonal Contraceptive Use and Menopause Status

Menopausal status was confirmed by the study physician based on the date of last menstruation self-reported by participants at baseline and study conclusion. Post-menopausal status was defined as ≥ 1 year without menstruation. Participants provided information regarding contraceptive use at baseline (Table 2). Changes in medications (including contraceptives) were assessed at months 3, 6, 9, 12, 15, and 18. An endocrinologist blinded to weight loss outcomes reviewed medication usage at each time point to identify any changes in CHC use, and use of any non-contraceptive oral or transdermal estrogens and/or progestins during the study. Participants using hormonal IUDs containing levonorgestrel were included in the non-CHC group because the hormone exposure is primarily localized. Published studies indicate that maximum systemic levonorgestrel levels are at least 10 times lower than progesterone levels among women using combined oral contraceptive pills (19). In addition, endogenous estradiol and progesterone levels among women using hormonal IUDs are comparable to those of normally cycling women (20, 21).

Table 2.

Combined hormonal contraceptive types and formulations used by participants.

CHC Name	# on CHC	Estrogen Component	Day	Progesterone Component	Day	Other Components	Day	Route
Yaz	1	Ethinyl Estradiol 20mcg	1-24	Drosperinone 3mg	1-24	Inert	25-28	Oral
Necon	3	Ethinyl Estradiol 35mcg	1-21	Norethindrone 0.5mg	1-10	Inert	22-28	Oral
				Norethindrone 1mg	11-21
Generess	1	Ethinyl Estradiol 25mcg	1-24	Norethindrone 0.8mg	1-24	Inert	25-28	Oral
Junel Fe	1	Ethinyl Estradiol 20mcg	1-24	Norethindrone Acetate 1mg	1-24	Ferrous Fumarate 75mg	25-28	Oral
Sprintec	1	Ethinyl Estradiol 35mcg	1-21	Norgestimate 0.25mg	1-21	Inert	22-28	Oral
Ortho Tri-Cyclen	1	Ethinyl Estradiol 35mcg	1-21	Norgestimate 0.18mg	1-7	Inert	22-28	Oral
				Norgestimate 0.215mg	8-14
				Norgestimate 0.25mg	15-21
Aviane	1	Ethinyl Estradiol 20mcg	1-21	Levonorgestrel 0.1mg	1-21	Folic Acid 1mg	22-28	Oral
Orsythia	1	Ethinyl Estradiol 20mcg	1-21	Levonorgestrel 0.1mg	1-21	Folic Acid 1mg	22-28	Oral
Levora	1	Ethinyl Estradiol 30mcg	1-21	Levonorgestrel 0.15mg	1-21	Inert	22-28	Oral
Quasense	1	Ethinyl Estradiol 30mcg	1-84	Levonorgestrel 0.15mg	1-84	Inert	85-90	Oral
Seasonale	2	Ethinyl Estradiol 30mcg	1-84	Levonorgestrel 0.15mg	1-84	Inert	85-90	Oral
NuvaRing	3	Ethinyl Estradiol 15mcg	1-21	Etonorgestrel 0.12mg	1-21	n/a	n/a	Vaginal

One unknown CHC formulation; some people took more than one formulation during the study.

Abbreviations – CHC: Combined hormonal contraceptives.

Outcome Measures

All outcomes were measured at baseline and months 6, 12, and 18. Due to the cohort design of the trial, it was not possible to time outcome measure study visits to the menstrual cycle.

Body weight and body composition.

Body weight and body composition were measured with a calibrated digital scale (to the nearest 0.2 lbs.). WC (cm) was measured just over the iliac crests using a tape measure. FM and LM were measured with the dual-energy X-ray absorptiometry (DXA, Hologic Discovery QDR Series, Bedford, MA).

Dietary outcomes.

Self-reported dietary EI and macronutrient intake were assessed with three-day diet records at baseline and months 6, 12, and 18. Participants were asked to log all food and beverages over 3 days at each time point. Diet records were analyzed using Nutrition Data System for Research software (version 2016, Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN).

Physical Activity.

Free-living physical activity (PA) was assessed with a multi-sensor monitor (SenseWear Mini, BodyMedia Inc., Pittsburgh, PA, version 7.0). Participants were asked to wear the device for 7 consecutive days at each time point. Minutes of MVPA (≥3 METs) were quantified using a proprietary algorithm (22). A day was considered valid if the participant wore the SenseWear armband for ≥22.8 hours/day (95% wear time). To be included in the analysis, ≥4 valid days, including ≥1 valid weekend day were required. Bout MVPA (minutes accumulated in bouts lasting ≥10 minutes where ≥80% of the bout was MVPA), and steps (count/day) served as the outcome measures of PA (23).

Statistical Analysis

Baseline demographic and clinical characteristics were summarized using descriptive statistics. Baseline differences between CHC and non-CHC were analyzed using two sample t-tests for continuous variables, and chi-square or Fisher’s exact test where appropriate for categorical variables. Our analysis plan focused primarily on descriptively reporting within- and between-group changes over time; as a secondary analysis, this study was not powered a priori to detect changes between the CHC and non-CHC groups. Intent-to-treat analysis was used to assess the effect of CHC use on outcomes over time. Linear mixed models with an unstructured covariance matrix were used to delineate the longitudinal profile of repeated outcome variables in each group (CHC, non-CHC). Outcome measures at baseline and any subsequent time points were the dependent variable in the mixed effects modeling. Fixed effects included time (months), group (CHC, non-CHC), and their interaction term. The models were adjusted for age and randomized group assignment (standard, sequential). Under this saturated model, we examined within- and between-group changes in the outcomes between months 0-6 (i.e., weight loss phase), 6-18 (i.e. weight loss maintenance phase), and 0-18 (overall weight loss and maintenance). All analyses were performed with SAS version 9.4 (SAS System for Microsoft, SAS Institute Inc., Cary, NC, USA). We performed post-hoc sensitivity analyses excluding women using levonorgestrel-releasing intrauterine devices and controlling for race and parity but none changed the pattern of results. Thus, results are presented with the full sample not including these covariates. Results are presented as mean; 95% CI unless otherwise stated.

Results

Participant Characteristics

Of the 110 premenopausal women included in our analytical sample, 17 (15%) were classified as CHC users (CHC) and 93 (85%) were classified as non-CHC users (non-CHC), which is representative of CHC use among women of this age range in the United States (12). Of this initial sample, 11/17 CHC users (65%) and 60/93 non-CHC (65%) completed the 18-month intervention (Figure 1). Participants who were lost to follow-up were similar to completers in all baseline characteristics, except they were younger (mean±SD; 35±8 years versus 38±8 years; P<0.05). Baseline characteristics are displayed in Table 1. There were no significant differences between CHC and non-CHC for outcome variables at baseline except for % fat intake: CHC self-reported a significantly higher percent fat intake compared to non-CHC.

Table 1.

Baseline characteristics of study population^a

Characteristic	CHC users (n=17)	Non-CHC users (n=93)
Age (y) (mean ± SD)	34 ± 10	37 ± 8
Anthropometric Measures (mean ± SD)
Weight (kg)	95.5 ± 14.9	94.2 ± 14.8
BMI (kg/m2)	34.6 ± 4.1	35.1 ± 4.1
Waist Circumference (cm)	108.7 ± 10.7	106.4 ± 11.3
Total Body Fat Mass (kg)	41.1 ± 7.7	39.6 ± 8.4
Intervention Group (count (%))
Standard	9 (52.9%)	44 (46.8%)
Sequential	8 (47.1%)	49 (52.7%)
Intervention Participation
Attrition (count (%))	6 (35.3%)	33 (35.1%)
Weeks to attrition (mean± SD)	45.9 ± 4.7	60.0 ± 2.9
Group-based class attendance (mean± SD)	71% ± 47%	70% ± 46%
Supervised exercise session attendance (mean± SD)	44.7% ± 30.8%	51.8% ± 30.5%
Race/Ethnicity (count (%))
Caucasian	16 (94.1%)	67 (72.0%)
Asian		4 (4.3%)
Black or African American	1 (5.9%)	20 (21.5%)
Not reported		2 (2.2%)
Ethnicity (count (%))
Hispanic or Latino	2 (11.8%)	23 (24.7%)
Not Hispanic or Latino	15 (88.2%)	70 (75.3%)
Education
High school or less	1 (5.9%)	9 (9.7%)
Some College	6 (35.3%)	20 (21.5%)
4 years of college or graduate degree	10 (58.5%)	63 (67.7%)
Reproductive Health and History
History of PCOS	1 (5.9%)	4 (4.3%)
Parity: Ever pregnant (%)	6 (35.3%)	58 (62.4%)
Dietary Intake (mean ± SD) b
Energy Intake (kcal/d)	1,836 ± 487	1,852 ± 533
Fat Intake (%)	40.3% ± 6.4%	34.8% ± 6.5%*
Carbohydrate Intake (%)	41.4% ± 8.1%	45.4% ± 7.6%
Protein Intake	18.0% ± 3.4%	17.6% ± 4.2%
Physical Activity Measures (mean ± SD)
Bout MVPA (min/d) c, d	17 ± 15	19 ± 17
Steps (count/d) c	5,398 ± 1,910	6,337 ± 2,432

^aContinuous variables analyzed using two sample t-tests, and categorical variables analyzed using chi-square test or Fisher’s exact test.

^bn = 14 for CHC users and n = 73 for non-CHC users

^cn = 15 for CHC users, n = 88 for non-CHC users.

^dMVPA accumulated in bouts ≥ 10 minutes.

^*Significant difference between groups (p < 0.01).

Bold denotes (p < 0.05); Abbreviations - CHC: Combined hormonal contraceptives, MVPA: moderate-to-vigorous physical activity.

Body Weight and Body Composition

There were consistent differences between CHC and non-CHC in changes over time for weight, FM, and WC (Table 3, Figure 2a–c). Between months 0 to 6 both CHC and non-CHC exhibited decreases in weight (CHC: −6.7kg; −9.8 to −3.7kg; non-CHC: −7.7kg; −9.1 to −6.4kg), FM (CHC: −5.1kg; −7.3 to −2.9kg; non-CHC: −5.2kg; −6.2 to −4.3kg) and WC (CHC: −5.5cm; −8.8 to −2.2cm; non-CHC: −6.8cm; −8.2 to −5.4cm). However, between months 6 to 18, CHC users gained an average of 4.9 kg (0.9 to 8.9 kg), while body weight in non-CHC remained relatively unchanged (−0.2kg; −1.8 to 1.6kg). Increases in FM were also observed between 6 to 18 months in CHC (4.4kg; 1.3 to 7.5kg); while FM was relatively unchanged in non-CHC (−0.2kg; −1.5 to 1.1 kg). Thus, at 18 months, weight loss in CHC was minimal (−1.8kg; −7.3 to 3.6kg), but remained clinically meaningful in non-CHC (−7.9kg; −10.2 to −5.5kg). Similarly, FM at 18 months was relatively unchanged from baseline at 18 months among CHC (−0.7kg; −4.8 to 3.4kg), but was reduced from baseline in non-CHC (−5.4kg; −7.1 to −3.6kg). Among completers of the 18-month intervention, 3/11 (27%) of CHC exhibited ≥5% weight loss at 18 months, compared to 33/60 (55%) of non-CHC. Changes in LM over time followed a similar pattern in both groups (Table 3, Figure 2d).

Table 3.

Adjusted means and change in all outcome variables

		Assessment Period, Mean (SEM)				Change
Outcome		0M	6M	12M	18M	0M–6M	95% CI	ES	6M-18M	95% CI	ES	0M-18M	95% CI	ES
Weightb (kg)	CHC	95.6 (3.7)	88.9 (3.6)	90.3 (3.8)	93.8 (4.0)	−6.7	(−9.8, −3.7)		4.9	(0.9, 8.9)		−1.8	(−7.3, 3.6)
	Non-CHC	94.1 (1.6)	86.4 (1.5)	85.4 (1.6)	86.2 (1.7)	−7.7	(−9.1, −6.4)		−0.1	(−1.8, 1.6)		−7.9	(−10.2, −5.5)
	Difference	1.5 (4.0)	2.5 (3.9)	4.9 (4.2)	7.5 (4.3)	1.0	(−2.4, 4.4)	0.12	5.0	(0.6, 9.4)	0.44	6.0	(0.1, 12.0)	0.39
WCc (cm)	CHC	109.4 (2.7)	104.0 (2.7)	104.5 (3.0)	108.2 (3.2)	−5.5	(−8.8, −2.2)		4.4	(0.0, 8.7)		−1.1	(−6.3, 4.1)
	Non-CHC	106.4 (1.2)	99.6 (1.1)	97.3 (1.3)	98.7 (1.3)	−6.8	(−8.2, −5.4)		−0.9	(−2.7, 1.0)		−7.7	(−9.9, −5.5)
	Difference	3.0 (3.0)	4.4 (3.0)	7.2 (3.3)	9.6 (3.4)	1.3	(−2.2, 4.9)	0.14	5.2	(0.5, 9.9)	0.43	6.6	(0.9, 12.2)	0.45
Fat Massc (kg)	CHC	41.1 (2.1)	36.1 (2.0)	37.7 (2.4)	40.2 (2.6)	−5.1	(−7.3, −2.9)		4.4	(1.3, 7.5)		−0.7	(−4.8, 3.4)
	Non-CHC	39.5 (0.9)	34.3 (0.8)	33.1 (1.0)	34.2 (1.1)	−5.2	(−6.2, −4.3)		−0.2	(−1.5, 1.1)		−5.4	(−7.1, −3.6)
	Difference	1.6 (2.2)	1.7 (2.2)	4.5 (2.6)	6.3 (2.8)	0.1	(−2.3, 2.5)	0.02	4.6	(1.2, 7.9)	0.53	4.7	(0.3, 9.2)	0.41
Lean Massc (kg)	CHC	50.0 (1.7)	48.7 (1.7)	48.4 (1.7)	49.0 (1.7)	−1.2	(−2.5, 0.0)		0.3	(−1.2, 1.8)		−0.9	(−2.7, 0.8)
	Non-CHC	50.1 (0.7)	47.9 (0.7)	48.0 (0.7)	47.8 (0.7)	−2.1	(−2.7, −1.4)		−0.1	(−0.7, 0.5)		−2.3	(−3.1, −1.6)
	Difference	−0.2 (1.9)	0.8 (1.9)	0.4 (1.8)	1.2 (1.8)	1.0	(−0.4, 2.4)	0.27	0.4	(−1.2, 2.0)	0.10	1.4	(−0.5, 3.3)	0.29
Dietary Intaked (kcal/day)	CHC	1828 (141)	1530 (95)	1449 (131)	1696 (122)	−298	(−662, 66)		166	(−169, 502)		−132	(−481, 217)
	Non-CHC	1859 (62)	1714 (37)	1390 (57)	1446 (49)	−146	(−301, 10)		−268	(−403, –132)		−413	(−560, –266)
	Difference	−32 (154)	−184 (102)	59 (143)	250 (131)	−152	(−548, 244)	−0.15	434	(72, 795)	0.48	282	(−97, 660)	0.30
% Fat Intaked	CHC	40.2 (1.8)	33.9 (2.3)	30.7 (1.8)	39.5 (2.6)	−6.2	(−12.0, −0.5)		5.6	(−1.3, 12.4)		−0.7	(−6.0, 4.7)
	Non-CHC	34.7 (0.8)	33.6 (0.9)	30.3 (0.8)	35.3 (1.1)	−1.1	(−3.5, 1.2)		1.8	(−1.0, 4.5)		0.6	(−1.5, 2.8)
	Difference	5.5 (1.9)	0.4 (2.5)	0.4 (19)	4.2 (2.8)	−5.1	(−11.3, 1.1)	−0.33	3.8	(−3.2, 10.8)	0.21	−1.3	(−7.1, 4.5)	−0.09
% Carb Intaked	CHC	41.3 (2.1)	42.4 (2.7)	47.7 (2.4)	42.0 (3.0)	1.2	(−5.4, 7.7)		−0.4	(−8.5, 7.6)		0.8	(−6.1, 7.6)
	Non-CHC	45.6 (0.9)	47.2 (1.1)	46.6 (1.0)	43.6 (1.2)	1.6	(−1.0, 4.3)		−3.7	(−6.9, −0.5)		−2.0	(−4.9, 0.8)
	Difference	−4.3 (2.3)	−4.8 (2.9)	1.1 (2.6)	−1.5 (3.2)	−0.5	(−7.5, 6.6)	−0.03	3.3	(−5.4, 11.9)	0.15	2.8	(−4.6, 10.2)	0.15
% Protein Intaked	CHC	18.1 (1.1)	20.7 (1.3)	21.0 (1.6)	17.5 (1.6)	2.7	(−0.8, 6.1)		−3.2	(−7.2, 0.7)		−0.6	(−4.0, 2.9)
	Non-CHC	17.7 (0.5)	17.7 (0.5)	20.9 (0.7)	20.4 (0.7)	−0.0	(−1.4, 1.4)		2.7	(1.2, 4.3)		2.7	(1.3, 4.1)
	Difference	0.4 (1.2)	3.1 (1.4)	0.1 (1.7)	−2.9 (1.8)	2.7	(−1.0, 6.4)	0.29	−6.0	(−10.2, −1.7)	−0.56	−3.3	(−7.0, 0.4)	−0.36
Bout MVPAe (min/d)	CHC	17 (5)	35 (8)	34 (9)	31 (8)	18	(3, 33)		−4	(−20, 12)		14	(−2, 30)
	Non-CHC	20 (2)	32 (3)	38 (4)	29 (3)	12	(6.0, 18)		−3	(−10, 4)		9	(3, 15)
	Difference	−3 (5)	3 (8)	−4 (10)	3 (9)	6	(−10, 22)	0.15	−1	(−18, 16)	−0.02	6	(−12, 23)	0.12
Stepse (count/d)	CHC	5486 (620)	7252 (742)	6999 (744)	6725 (809)	1766	(337, 3195)		−528	(−2093, 1038)		1238	(−237, 2713)
	Non-CHC	6407 (258)	7872 (306)	8341 (320)	6994 (338)	1434	(862, 2007)		−848	(−1503, –193)		587	(−7, 1180)
	Difference	−921 (671)	−589 (803)	−1341 (811)	−269 (877)	332	(−1207, 1871)	0.08	320	(−1377, 2017)	0.07	652	(−938, 2241)	0.16

^aResults are from linear mixed effect model with unstructured covariance using an intent-to-treat analysis; Effect size (ES) is calculated as 2 × t value)/√DF; WC: Waist Circumference; Carb: Carbohydrate; MVPA: Moderate-to-Vigorous Physical Activity.

^bSample sizes for weight are: CHC: n=17 at 0M, n=13 at 6M, n=12 at 12M, n=11 at 18M; non-CHC: n=93 at 0M, n=73 at 6M, n=65 at 12M, n=60 at 18M.

^cSample sizes for waist circumference, fat mass, and lean mass are: CHC: n=17 at 0M, n=13 at 6M, n=12 at 12M, n=10 at 18M; Controls: n=93 at 0M, n=73 at 6M, n=65 at 12M, n=60 at 18M.

^dSample size for dietary intake, % fat intake, % carb intake, and % protein intake: CHC: n=14 at 0M, n=10 at 6M, n=11 at 12M, n=8 at 18M; non-CHC: n=73 at 0M, n=66 at 6M, n=59 at 12M, n=49 at 18M.

^eSample size for bout MVPA and steps: CHC: n=15 at 0M, n=12 at 6M, n=12 at 12M, n=10 at 18M; non-CHC: n=88 at 0M, n=71 at 6M, n=63 at 12M, n=58 at 18M.

Figure 2.

Adjusted means of body weight, waist circumference, and body composition over time

^a Error bars represent standard error of the mean; CHC: Combined hormonal contraceptives.

^b Sample sizes for weight are: CHC: n=17 at 0M, n=13 at 6M, n=12 at 12M, n=11 at 18M; non-CHC: n=93 at 0M, n=73 at 6M, n=65 at 12M, n=60 at 18M.

^c Sample sizes for waist circumference are: CHC: n=17 at 0M, n=13 at 6M, n=12 at 12M, n=10 at 18M; non-CHC: n=93 at 0M, n=73 at 6M, n=65 at 12M, n=60 at 18M

^d Sample sizes for fat mass and lean mass are: CHC: n=17 at 0M, n=13 at 6M, n=12 at 12M, n=10 at 18M; non-CHC: n=93 at 0M, n=71 at 6M, n=65 at 12M, n=60 at 18M

Diet-Related Outcomes

Self-reported dietary intake are shown in Table 3 and Figure 3. Change in EI between baseline and month 6 followed a similar pattern within both groups. However, between months 6 to 18, CHC exhibited a trend for an increase in EI (166 kcal/day; −169 to 502 kcal/day), while non-CHC exhibited decreases in EI (−268 kcal/day; −268 to −132 kcal/day). In addition, CHC tended to report decreased % protein intake (−3.2%; −7.2 to 0.7%), while non-CHC reported increased % protein intake between months 6 to18 (2.7%; 1.2 to 4.3%). Between months 6 to 18, the between-group difference in change in EI was 434 kcal/day (72 to 795 kcal/day), and change in % protein intake was −6% (−10.2 to −1.7%), with CHC serving as the reference group. Changes in % fat and % carbohydrate intake over time were similar across groups.

Figure 3.

Adjusted means of dietary intake outcomes over time

^a Error bars represent standard error of the mean; CHC: Combined hormonal contraceptives, kcal: kilocalories.

^b Sample sizes for energy intake and macronutrient composition are: CHC: n=14 at 0M, n=10 at 6M, n=11 at 12M, n=8 at 18M; non-CHC: n=73 at 0M, n=66 at 6M, n=59 at 12M, n=49 at 18M.

Exercise-Related Outcomes

Changes in bout MVPA minutes and steps over time followed a similar pattern in both groups (Table 3 and Figure 4).

Figure 4.

Adjusted means of physical activity outcomes over time

^a Error bars represent standard error of the mean; CHC: Combined hormonal contraceptives.

^b Sample sizes for MVPA Bouts and Steps are: CHC: n=15 at 0M, n=12 at 6M, n=12 at 12M, n=10 at 18M; non-CHC: n=88 at 0M, n=71 at 6M, n=63 at 12M, n=58 at 18M.

Discussion

To our knowledge, this is the first study to evaluate the impact of CHC use on weight loss among premenopausal women with overweight/obesity who were actively trying to lose weight. While our results are preliminary, we observed that both CHC and non-CHC achieved clinically meaningful (>5%) weight loss at 6 months within a comprehensive behavioral weight loss intervention. However, CHC exhibited weight regain of almost all initial weight loss between months 6 to 18, while weight remained essentially unchanged in non-CHC over the same time frame. Ultimately, only the non-CHC group exhibited clinically meaningful weight loss on average at 18 months. Changes in FM and WC followed a similar pattern.

A recent systematic review (17) determined that a large effect of CHCs on body weight in free-living conditions was not evident, but there was insufficient evidence to determine the effects of CHCs on weight gain as the quality of reporting in these studies was generally poor. Notably, of the 49 studies included in the systematic review, only one had weight as the primary outcome. Existing placebo-controlled randomized trials to date do not suggest a causal association between CHCs and weight gain over 4-9 months following initiation (15, 24–28). A longer study duration may be needed to observe significant changes in weight, or it is possible that CHC use does not impact weight in women who are not specifically seeking to lose weight or maintain weight loss. However, the differences observed in the current study between CHC and non-CHC in weight loss at 18 months within a behavioral weight loss intervention suggests that CHC use may have clinically meaningful effects on weight loss and weight loss maintenance. Because this was a secondary analysis, unobserved confounders may impact results as participants were not randomized to CHC or non-CHC. Thus, our findings are preliminary, but suggest this is an important topic for future research. Replication in existing larger datasets and prospective, randomized trials are needed to elucidate the extent to which CHC use may influence weight loss and/or maintenance within a behavioral weight loss program.

Several mechanisms through which CHCs may influence body weight have been hypothesized, including increased EI (29). In the current study, CHC users differed from non-CHC users in change in EI from months 6 to 18, consistent with the time-frame for weight regain in CHC. It is possible that an effect of CHCs on hunger and/or satiety made it more difficult for CHC users to sustain adherence to reduced calorie diet over 18 months. CHC use may affect appetite and eating behavior because of the distinctly different hormonal profile of women using CHCs. Commonly used CHCs lead to lower levels of endogenous estrogens and progesterone across the cycle, without cyclic fluctuation (30). Moreover, a recent study by Lovett et al. (31) measured serum concentrations of endogenous and exogenous (synthetic) estradiol and progesterone in women receiving seven typical CHCs after adjusting for their relative binding affinity to their receptors. This study demonstrated that the median exogenous E2 exposure across a 28-day cycle in CHC users was similar to median endogenous E2 exposure; however, median synthetic progesterone exposure in CHC users was 4-fold higher on average than median circulating endogenous progesterone (31). These higher progesterone levels in particular have the potential to influence EI.

Progesterone has been shown to increase appetite and trigger binge- or emotional-eating (32). Several human and animal studies have examined the role of progesterone on EI, macronutrient intake, and hormones regulating energy metabolism in normally cycling women (33–35). These studies have found that women may consume up to 500 kcal/day more, with a propensity for foods higher in carbohydrate or fat during the luteal phase, when levels of progesterone are higher (34). The baseline differences in fat intake between CHC and non-CHC observed in the current study are consistent with these studies. In addition, a pilot study demonstrated that pictures of high-calorie foods led to significantly greater activation in areas of the brain associated with food motivation relative to pictures of low-calorie foods or non-food objects 8 weeks after injections of Depo Medroxyprogesterone Acetate (progestin) compared to baseline (33). These data suggest that progestins used for contraceptives may affect central appetite pathways, leading to increases in appetite as well as binge and hedonic eating, which then could lead to increases in overall EI and/or a change in macronutrient preference.

Endogenous estrogens, on the other hand, are demonstrated to have an anorexigenic effects through their action on central estrogen receptors (34, 36). In both humans and rhesus macaques, EI is lowest in the periovulatory phase, when estradiol is the highest (35). However, EI is also high during the luteal phase, when both estradiol and progesterone are high, suggesting that progesterone may counter these anorexigenic effects of estradiol (35). In addition, ovariectomized rats exhibit increases in EI, rapid weight gain, and increased adiposity that reverses with administration of physiological doses of estradiol (37). It is not currently known if E2 exerts a similar anorexigenic effect as endogenous estrogens, or if suppressing endogenous estrogens may increase appetite and EI. Thus, the increases in EI and weight gain observed in CHC users between months 6 to 18 may be associated with the higher levels of progestins, and lower levels of endogenous estrogens.

CHC use could also influence weight regain through effects on PA and/or energy expenditure (EE). We did not observe differences in device–measured MVPA or steps between CHC users and non-CHC users at baseline; and patterns of changes in these PA parameters over the 18-month intervention were similar in both groups. We are not aware of any other data on the impact of CHC use on free-living PA behavior within or outside the context of a weight loss trial. We did not measure daily EE in this study. However, oral contraceptive pill users have been shown to have a resting energy expenditure that is ~5% higher than non-users (38) which would theoretically enhance weight management. To our knowledge, there is no other published data on the impact of CHCs on total daily EE or its components. However, experimentally suppressing ovarian hormones with gonadotropin-releasing hormone receptor agonists in pre-menopausal women leads to reductions in total and exercise EE measured under the same conditions by whole room indirect calorimetry, which were only partially attenuated with synthetic estradiol add-back (39). This could suggest that even if MVPA was not less in CHC users, the energy they expended per unit of activity could be lower, although future studies are needed to directly address this question. Given this literature, and the differences in 18 month weight loss we observed, the impact of CHCs on PA and EE warrants further investigation.

It is important to note that there are adaptations that occur in response to weight loss that predispose individuals to weight regain (40–42). These include changes in appetite and appetite related hormones, EE including resting EE and EE expended in PA, and changes in metabolism including insulin sensitivity. It is possible that CHCs have only modest effects on these variables when women are in energy balance, but have clinically meaningful effects in the weight reduced state. Future controlled studies are needed to further explore the effect of CHCs on these variables in the weight reduced state.

Given the current lack of published data on the potential effects of CHC use on weight loss and weight loss maintenance, this study provides important preliminary findings. However, our results should be interpreted with caution and careful consideration of the limitations of this secondary data analysis comprised of a relatively small sample of CHC users (n=17). Because of the small sample size, we were unable to evaluate the effect of different sub-categories of CHCs (i.e., triphasic versus monophasic, different preparations of E2 and/or progestins, oral versus vaginal/transdermal) on weight loss/regain. Different formulations of CHCs used by participants (Table 2) may have contributed to the variability in weight loss maintenance observed among CHC users. Missing data is also a limitation of the ITT approach. Although we used a mixed-level model to minimize the bias of our estimates, if data were not missing at random, the reason for missing data could have introduced bias in our estimates. It is possible that each group differed in ways that we did not measure (e.g., socioeconomic status), or include in the statistical models reported (e.g., race, parity; though post-hoc sensitivity analyses did not change the pattern of results). Moreover, EI was assessed based on self-report, which suffers from inaccuracies and bias (43). While PA was measured with the SenseWear device, we did not have measures of EE. Finally, the outcome measures in the parent trial were not timed to phase of the menstrual cycle, and EI (32–35) has been shown to change throughout the cycle. Despite these limitations, this study had several strengths, particularly in light of the fundamental gaps in knowledge that exist on the effects of CHCs on body weight and composition, EI, and PA. CHC users and non-CHC users were not different in any baseline characteristics (except % fat intake) and both groups had similar rates of attrition and attendance for the group-based weight loss classes and supervised exercise. We obtained detailed information on CHC use, as well as menopausal status throughout the study. Given these limitations, it is important not to overestimate the results of this preliminary study. This study is not meant to discourage the use of CHCs, which offer many benefits for women’s health, economic empowerment, educational opportunities, and quality of life (44, 45). Rather, these results highlight a need for replicating this secondary analysis in other large, interventional weight loss trials and exploration in prospective randomized trials, including studies of the effects of different CHC formulations. These studies should include objective measures of EI, EE, macronutrient preference and PA timed to the menstrual cycle. If future studies yield similar findings, this could suggest that CHC use is an important factor to consider in obesity treatment and that patient’s goals for weight loss and maintenance may need to be considered when counseling patients on the best form of contraception.

Conclusion

Within a comprehensive behavioral weight loss trial, both CHC users and non-CHC users achieved clinically meaningful (>5%) weight loss at 6 months. However, premenopausal women using CHCs regained almost all of initial weight loss between months 6 to 18, while women not using CHCs maintained initial weight loss at 18 months. During months 6 to 18, changes in EI followed a pattern consistent with weight regain, with increases in EI in CHC relative to non-CHC. These results suggest a potential effect of CHCs on appetite, eating behavior, and/or macronutrient preference in the weight reduced state. However, there are significant limitations to this secondary data analysis and future research is urgently needed to explore this preliminary signal and determine if CHC use impacts weight loss and/or weight loss maintenance.

STUDY IMPORTANCE QUESTIONS.

What is already known about this subject?

• Approximately 60% of women of reproductive age (20-39 yrs) are overweight or obese.

• A significant proportion of reproductive-age women use combined (estrogen and progestin) hormonal contraceptives (CHCs), including oral contraceptive pills, transdermal patches, and vaginal rings.

• Although most studies do not indicate that CHCs cause weight gain, the impact of CHC use on weight loss or weight loss maintenance has never been studied.

What does your study add?

• In a secondary analysis of data from an 18-month behavioral weight loss trial, both women using CHCs and women not using CHCs achieved clinically meaningful weight loss on average at 6 months.

• However, between months 6 and 18, women using CHCs regained almost all of their initial weight loss, while women not using CHCs maintained their initial weight loss.

• As a result, women using CHCs exhibited lower weight loss than women not using CHCs at 18 months.

How might your results change the direction of research or the focus of clinical practice? Please remember to also include between the title page and structured abstract in your paper.

Combined hormonal contraceptives (CHCs) are a commonly used method of contraception in women. Our secondary data analysis suggests that CHC use was associated with lower mean weight loss at 18-months within a comprehensive behavioral weight loss intervention. However, there are important limitations to this secondary data analysis and future studies are needed to further explore this preliminary signal.