Contraception 75 (2007) 299 – 304
Original research article
Oral contraceptives: side effects and depression in adolescent girlsB,BB Katharine O’Connell4, Anne R. Davis, Jennifer Kerns Department of Obstetrics and Gynecology, Columbia University, New York, NY 10032, USA Received 3 August 2006; revised 20 August 2006; accepted 6 September 2006
Abstract Objective: The study was undertaken to determine the side effects, including depression, of oral contraceptives (OCs) in adolescent girls. Methods: We conducted a randomized trial of OCs for dysmenorrhea, which assessed side effects and depression. Seventy-six adolescents received an OC (20 Ag of ethinyl estradiol/100 mg of levonorgestrel) or a placebo in a double-blind fashion for 3 months. We ascertained OC side effects using open-ended and closed question formats. Participants self-administered the Center for Epidemiologic Studies Depression Scale (CES-D) to assess depressive symptoms. Results: Fifty-seven participants (77%) reported at least one side effect (median = 2, range = 0–8, interquartile range = 1.0–3.25). The number and the type of side effects reported in the OC group and in the placebo group were similar. Mean exit CES-D scores were comparable between groups [OC group, 14.0 (SD = 9.2); placebo group, 14.4 (SD = 8.1); p = .86]. Conclusion: Adolescents treated with an OC or a placebo experienced similar numbers and types of OC side effects, as well as depressive symptoms. D 2007 Elsevier Inc. All rights reserved. Keywords: Oral contraceptives; Depression; Side effects; Adolescents
1. Introduction Oral contraceptives (OCs) are the most often used form of reversible contraception in the United States and throughout much of the world. More than 18 million American women use OCs; about 30% of contracepting women [1] and 80% of women use OCs at some point in their lives [2]. While use is prevalent, discontinuation rates of OCs are high, especially among adolescents. Multiple studies using convenience sample studies of diverse populations of adolescents demonstrate discontinuation rates of z 50% during the first 6 months of use [3–5]. Beliefs and experiences related to the side effects of hormonal contraception influence OC initiation and continuation. A survey of 345 adolescents showed that teens who believed that OC use would likely affect menses, physical appearance or health were much less likely to initiate OCs
B Funding support for this study was provided by the National Institute of Child Health and Human Development (grant R03-HD39776-RCT). BB The views in this manuscript do not necessarily reflect those of the National Institute of Child Health and Human Development. 4 Corresponding author. Tel.: +1 212 305 4805. E-mail address:
[email protected] (K. O’Connell).
0010-7824/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.contraception.2006.09.008
or to use OCs consistently [6]. Among adolescents beginning OC use postpartum or postabortion, 46% of users discontinued use by the end of the 6- to 18-month follow-up period, primarily due to side effects [7]. In one study of 209 adolescents initiating OCs, more than 45% of participants reported worrying about weight gain, and those with perceived weight gain discontinued more often (37%) than those who did not perceive weight change (14%), even though the mean weight gain of discontinuers (0.46F2.4 kg) was similar to that of continuers (0.99F3.3 kg) [4]. In addition to these physical effects, physicians and women believe that mood changes and depression are expected side effects of OCs [8]. Data from controlled studies in adult women fail to support the assertion that OCs commonly cause side effects. In a placebo-controlled trial of 507 adult women, there was no difference in the reported side effects between OC users and placebo users [9]. Another placebo-controlled trial found no difference in weight change between the OC group and the placebo group [10]. A subsequent pooled analysis of two placebo-controlled trials reported similar findings of weight gain [11]. In addition, other side effects, such as nausea, breast pain and headache, were reported equally between the OC group and the placebo group. Among
300
K. O’Connell et al. / Contraception 75 (2007) 299–304
50 women who completed a mood questionnaire daily during their menstrual cycle, users of OC experienced a higher level of positive affect during the cycle than did nonusers [12]. Among 122 women who kept a daily diary for 3 months, there were no differences in mood changes throughout the menstrual cycle between OC users and nonusers [13]. No such controlled trials have been conducted in adolescents; other studies, however, demonstrate similar findings. In a prospective study of 46 adolescents, subjects anticipated more side effects than they actually experienced at 6 months of use [14]. A retrospective chart review of 130 adolescents using OCs and DepoProvera showed that the majority of users of hormonal contraception actually lost weight or gained b 5% of their baseline weight at 1 year [15]. We used data from a placebo-controlled trial to examine OC side effects, including depression, in adolescent girls.
2. Materials and methods This study was conducted in an academic medical center from August 2001 to November 2003 with approval of the Medical Center Institutional Review Board and the Pediatric Institutional Review Board. We collected baseline data via interviews with adolescent girls upon their enrollment in a clinical trial comparing the use of an OC to the use of a placebo for dysmenorrhea treatment. The main results, as well as a detailed description of study methods, are published elsewhere [16]. Healthy adolescent girls aged V 19 years experiencing moderate or severe dysmenorrhea were eligible for enrollment. We used the Robinson modification of the Andersch scale to classify the severity of dysmenorrhea [17]. Using this scale, moderate dysmenorrhea indicates sometimes or always experiencing very painful menstrual cramps. Severe dysmenorrhea indicates sometimes or always cutting back on activities in addition to experiencing painful menstrual cramps. Other inclusion criteria were: parental consent if aged b18 years, English speaking, regular menstrual cycles for at least 1 year and menstrual cycle length from 21 to 35 days. Exclusion criteria included: current or recent use of hormonal contraception; history of term pregnancy, recent abortion or miscarriage, followed by fewer than three menstrual cycles; or history of pelvic pathology related to dysmenorrhea. After initial telephone screening, the first visit was scheduled within 1 week of the last menstrual period. We required parental attendance and consent for those adolescents aged b18 years; these girls also provided assent. Adolescents aged N 18 years provided their own informed consent. After enrollment, the investigator administered a demographic and menstrual questionnaire. Participants completed the pain subscale of the Moos Menstrual Distress Questionnaire (MMDQ), a validated pain scale for previous
menses [18]. To measure depressive symptoms, we used the Center for Epidemiologic Studies Depression Scale (CES-D). Subjects responded to 20 statements (e.g., bI thought my life had been a failureQ) by describing how often they felt that way during the past week. Each item was rated from 0 (rarely or none of the time) to 3 (most or all of the time), for a total highest score of 60. Scores of z 16 are commonly interpreted as indicative of depression [19]. The CES-D has been validated cross-culturally and across age groups using clinical ratings of depression and other selfreport measures [20,21]. The test–retest reliability for this scale has been reported as ranging from .45 to .70, and the scale has high internal consistency (a = .85–.90) [22,23]. We conducted a brief physical examination, which included height, body weight and blood pressure measurements. We measured the weight of participants wearing only one layer of clothing. Participants were then randomly assigned to receive either an OC (20 Ag of ethinyl estradiol/100 mg of levonorgestrel) or a matching placebo. At the end of three menstrual cycles, participants returned for an exit visit and self-administered the MMDQ and the CES-D. Body weight was measured with the same scale. We ascertained OC side effects in two ways. First, we asked participants, through an open-ended question, to list any side effects or changes they experienced during the study. Second, we asked if they experienced any of the 12 specific side effects commonly attributed to OCs: headache, nausea, acne, abdominal pain, back pain, vomiting, breast tenderness, breast enlargement, mood swings, weight gain, premenstrual syndrome and irregular bleeding. We used SPSS version 10.0 (SPSS, Inc., Chicago IL) to perform data analysis. We used descriptive statistics to describe demographic and menstrual pain characteristics, as well as side effects. We performed exploratory analyses of possible associations between the occurrence of side effects, age and dysmenorrhea severity using chi-square test and independent-sample Student’s t test, where appropriate. We analyzed weight changes using paired-sample Student’s t test. Sample size was calculated for power to detect differences in main outcome (dysmenorrhea) between the OC group and the placebo group. This analysis presents post hoc analyses of secondary outcomes only.
3. Results Screening, randomization, discontinuation and loss to follow-up are summarized in Fig. 1. In order to complete enrollment, we screened 300 adolescents. Of those screened, 173 (58%) were eligible. Of those ineligible, reasons included: age of N19 years, pain inconsistent with menses, irregular menses and requiring OC for contraception. Of those eligible, 97 (56%) did not enroll. Logistical issues (63%) were the most common reason for nonenrollment. Additionally, 20% of those eligible did not enroll due to
K. O’Connell et al. / Contraception 75 (2007) 299–304
301
Table 1 Demographic and baseline characteristics of enrolled subjects All subjects (n = 76) Age in years 16.8 (2.2) [mean (SD)] Race [n (%)] White 20 (26) Black 8 (11) Hispanic 36 (47) Other 12 (16) Education in years [n (%)] 0–8 15 (20) 9–12 48 (63) N 12 13 (17) Body mass index 24.1 (4.5) [mean (SD)] Severity of dysmenorrhea [n (%)] Moderate 32 (42) Severe 44 (58) Baseline 11.4 (5.2) MMDQ score [mean (SD)] Baseline 16.8 (10.9) CES-D score [mean (SD)]
OC group (n = 38)
Placebo group (n = 38)
16.7 (2.3)
16.9 (2.0)
8 (21) 13 (34) 11 (29) 6 (16)
.75
(32) (26) (26) (16)
.74
(21.1) (57.8) (21.1) (4.6)
.71
16 (42.1) 22 (57.9) 11.0 (5.1)
16 (42.1) 22 (57.9) 11.8 (5.3)
1.00
16.3 (10.0)
17.3 (11.8)
.70
7 26 5 23.7
(18.4) (68.4) (13.2) (4.4)
12 10 10 6
p
8 22 8 24.6
.46
.53
effect, 10 participants reported two side effects, 4 participants reported three side effects and 1 participant reported four side effects. Participants reported experiencing positive effects (less nausea) and negative effects (more nausea). The Table 2 Side effects reported during the trial (open-ended question)
Fig. 1. Flow diagram of participants’ progress throughout the trial.
adolescents’ or parents’ concerns with OC side effects or safety. Of the 76 participants enrolled, we obtained 3-month exit information for 74 participants (one participant became pregnant and discontinued participation, and one was lost to follow-up). The demographic characteristics of the study population are presented in Table 1. Randomization was successful; the groups were similar in terms of demographic characteristics, menstrual severity and baseline questionnaire scores. The mean CES-D score was 16.8, suggesting high prevalence of depressive symptoms in this group of urban adolescents. 3.1. Side effects Participants reported many OC side effects in response to the open-ended question asked on exit visit. Forty-five participants (61%) reported at least one side effect, as detailed in Table 2. Thirty participants reported one side
Menstrual effects Less pain or no pain with periods Lighter periods Amenorrhea Shorter periods bPeriods easierQ More pain with periods Heavier bleeding Prolonged bleeding Irregular bleeding Change in the timing of cycle Nonmenstrual effects More nausea Less nausea More acne Less acne More headaches Fewer headaches Bloating Weight gain Increased appetite Dizziness Fatigue Less constipation bMore hyperactiveQ Moodier Depression
OC group (n = 38)
Placebo group (n = 38)
15 2 1 1 1 1 0 1 2 0
10 7 0 2 0 0 1 1 0 1
2 1 1 0 1 1 2 1 1 1 0 0 0 1 0
1 0 0 1 1 0 0 1 0 0 1 1 1 0 1
302
K. O’Connell et al. / Contraception 75 (2007) 299–304
to an 8-lb gain. No participant discontinued use due to weight gain.
Table 3 Side effects reported during the trial (from a list)
Headache Nausea Acne Abdominal pain Back pain Vomiting Breast tenderness Breast enlargement Mood swings Weight gain Premenstrual syndrome Irregular bleed
OC group (n = 38)
Placebo group (n = 38)
p
8 6 6 4 7 1 9 9 10 11 12 11
8 7 5 6 6 1 7 7 13 6 13 5
1.0 .95 .95 .79 .95 1.0 .85 .85 .75 .38 .97 .23
median number of reported side effects in both the OC group and the placebo group was 1 [with an interquartile range (IQR) =0–1]. Using a list, participants were asked if they had experienced any of the 12 specific side effects listed during the trial (Table 3). Using this question format, the number of participants reporting side effects increased to 57 (77%). The median number of reported side effects increased to 2 (IQR =1.0–3.25) and was essentially the same in the OC group and in the placebo group (OC group: median = 2, IQR =1.0–3.5; placebo: median =2, IQR = 0.5–3.5). Younger girls (aged V 17 years) were no more likely to report any side effects than older girls, and girls with severe dysmenorrhea were no more likely to report any side effects than girls with moderate dysmenorrhea (data not shown). Three participants discontinued study participation due to side effects: acne (OC group), nausea (OC group) and moodiness (placebo group). We measured baseline weight in 66 participants and exit weight in 56 participants; we had both measurements for 51 participants (OC group, n = 29; placebo group, n = 22). The remaining subjects had their enrollment or exit interview at a location other than the research office (where no scale was available). There was no significant weight change for the group as a whole (baseline mean weight =145.5 lb, exit mean weight = 145.2 lb, p= .62), within the OC group (baseline mean weight = 142.4 lb, exit mean weight = 143.5 lb, p = .22) or within the placebo group (baseline mean weight = 149.7 lb, exit mean weight =147.5 lb, p =.09). Seventeen participants reported weight gain when prompted (as compared to two participants who spontaneously reported weight gain): 11 in the OC group and 6 in the placebo group. We had baseline and exit weight measurements for 12 of these participants. The other five subjects had exit interviews away from the medical center. The nine participants in the OC group with reported weight gain had a weight change ranging from a 4-lb loss to an 8-lb gain. The three participants in the placebo group who reported weight gain with weight information had a weight change ranging from no gain
3.2. CES-D Because of typographical error, 19 of 76 participants (25%) received CES-D questionnaires that had one item (bI could not get goingQ) missing. In an additional 13 cases, the participants skipped one item in the exit questionnaire; there was no pattern on skipped items. In all of these cases, the CES-D score was calculated by dividing the summed score by the number of answered items and multiplying the result by 20, as recommended by McDowell and Newell [19]. Overall, participants’ mean CES-D scores decreased slightly during the study, from 16.8 to 14.2 (SD = 8.6, paired t = 2.193, p= .03), indicating less depressive symptomatology over time. This improvement was similar within the OC group and within the placebo group. In the OC group (n = 37), the mean CES-D score decreased from 16.0 (SD =9.9) to 14.0 (SD = 9.2) (paired t test, p =.26). In the placebo group (n =37), the mean CES-D score decreased from 17.8 (SD =11.6) to 14.4 (SD = 8.1) (paired t test, p = .06). There was no significant difference in mean CES-D scores on exit between the OC group and the placebo group (independent-sample t test, t = 0.184, p = .86). To determine the effect of OCs on a potentially vulnerable subgroup, we separately examined participants (n =11) who had high scores at baseline (CES-D z 27, 1 SD above the mean), indicating an elevated risk of depression. These participants also showed improvement over time [baseline mean = 35.7 (SD =8.5); exit mean = 19.1 (SD = 9.2); t =3.908, p= .003], with similar improvements in the OC group and in the placebo group (data not shown). There were 20 participants who crossed the threshold of a positive screen for depression during the study, with similar numbers of participants changing in each group. Eleven participants had scores that decreased from N 16 to b 16 (OC group, six participants; placebo group, five participants), and nine participants had scores that increased from b 16 to N 16 (OC group, four participants; placebo group, five participants). 4. Discussion In this randomized placebo-controlled trial, we found that adolescents treated with placebos reported types and quantities of side effects similar to those of adolescents treated with OCs. Our results are consistent with a small number of randomized placebo-controlled trials in adult women, which found similar experiences of side effects in those treated with placebo and in those treated with OC [23]. Taken together, these controlled studies suggest that OCs have little causative role in the experience of common bhormonalQ side effects. Perceived side effects have an important public health impact, as they are a leading cause of OC discontinuation.
K. O’Connell et al. / Contraception 75 (2007) 299–304
Premature discontinuation of OCs leads to many unintended pregnancies and abortions because discontinuers often remain at risk for pregnancy but do not use another highly effective method of contraception. The influence of beliefs on health is observed throughout medicine; in studies of healthy nonpatients, those instructed to expect negative physical experiences prior to receiving placebos consistently reported to have experienced more negative effects than those not instructed [24]. Beliefs about medications are powerful predictors of compliance. In one clinical trial, subjects who were told that aspirin had gastrointestinal side effects were more likely to report subjective minor complaints and were six times more likely to discontinue participation in the study than those who were not bwarnedQ about this side effect [25]. Participants in our study experienced no significant changes in weight, either in the group as a whole or within the OC and placebo subgroups. However, participants in both groups complained of weight gain, even in several instances when they lost weight, reflecting the common concern that OCs cause weight gain. While our findings are based on a small sample, they reflect a recent Cochrane review that found no major effects of OCs on weight [26]. Using a standardized measure of depression, our study also demonstrated no difference in depressive symptoms between those treated with OC and those treated with placebo after 3 months of use. Early studies implicating OCs in depression were conducted using high-dose pills and were case–control and case–series designs that often relied on self-reported history of OC use among women with mood disorders [27–29]. More recent observational studies show inconsistent results and rely on nonstandardized measures to assess depressive symptoms [30–32]. We identified two studies that used a measurement tool to assess depression. Maes et al. [33] compared scores on the Beck Depression Scale between OC users and nonusers and found no difference. Similarly, Rasgon et al. [34] conducted a cross-sectional study of OCs and depression among women with polycystic ovary syndrome and reported that OC users were less depressed than nonusers, as measured by the CES-D. However, cross-sectional studies were unable to assess changes in mood with OC use. We identified one randomized placebo-controlled trial that examined the effect of OCs on depression. Graham and Sherwin [35] randomized 82 women with complaints of moderate to severe premenstrual symptoms to receive either OCs or placebos for 3 months, and they used a visual analogue scale to measure mood before and during the treatment phase. Both groups reported significant improvement in mood over the study period, consistent with the results of our study, although their study was limited by a high discontinuation rate. The modest decrease in depression scores seen across both the OC group and the placebo group may represent regression to the mean. People tend to enroll in clinical studies when symptoms are at their worst, which soon
303
improve with the passage of time. Another explanation for our results may be that this study enrolled adolescents with moderate to severe dysmenorrhea; as their pain improved during the study, their mood may have improved as well. Similarly, participants’ mood may have improved simply as a result of the increased medical attention that comes from being part of a clinical trial. Our findings demonstrate that OC use neither caused depression nor worsened depressive symptoms in this group of adolescents, regardless of the reason for improvement in mood. Young women [36,37] and minority inner-city populations [38,39] have higher scores on the CES-D. Our results are consistent with these findings; the mean CES-D score at baseline was 16.8, indicating a high rate of positive screens for depression. However, the CES-D is not a diagnostic tool; when the scale is applied in general population settings, it generates a high rate of false positives [19] and, therefore, some researchers have argued for a higher cutoff. By using the standard screen-positive of 16 in our study, our sample yielded a high rate of depressive symptoms, but this may not indicate an equivalent rate of clinical depression. 5. Conclusions Our study has several limitations. The small sample size and the relatively short duration of follow-up limited our ability to detect small differences between groups and to determine how the experience of side effects and depressive symptoms changes over time. Our study only included adolescents with moderate or severe dysmenorrhea; our findings may not apply to other populations of adolescents. We did not have sufficient power to detect small effects, but we did have power to find moderate effects; we had 81% power to detect depressive symptom differences (a six-point difference in mean CES-D between groups). Effects smaller than these may not be clinically meaningful. Overall, we found no differences in the number or in the type of side effects, and we found no large effects on mood, caused by a commonly used low-dose OC in this adolescent population. Physicians counseling patients should consider that OCs are unlikely to cause clinically important side effects. Acknowledgments OC and placebo pills were provided by Wyeth Pharmaceuticals (Collegeville, PA). References [1] Rosenberg MJ, Waugh MS. Oral contraceptive discontinuation: a prospective evaluation of frequency and reasons. Am J Obstet Gynecol 1998;179:577 – 82. [2] Centers for Disease Control and Prevention, National Center for Health Statistics. Fertility, family planning, and women’s health: new data from the 1995 National Survey of Family Growth 1997 [Series 23, No. 19, PHS No. 97-1995].
304
K. O’Connell et al. / Contraception 75 (2007) 299–304
[3] Oakley D, Sereika S, Bogue EL. Oral contraceptive pill use after an initial visit to a family planning clinic. Fam Plann Perspect 1991;23: 150 – 4. [4] Emans SJ, Grace E, Woods ER, Smith DE, Klein K, Merola J. Adolescents’ compliance with the use of oral contraceptives. JAMA 1987;257:3377 – 81. [5] Berenson AB, Wiemann CM. Contraceptive use among adolescent mothers at 6 months postpartum. Obstet Gynecol 1997;89:999 – 1005. [6] Moore PJ, Adler NE, Kegeles SM. Adolescents and the contraceptive pill: the impact of beliefs on intentions and use. Obstet Gynecol 1996; 88:48S – 56S. [7] Blumenthal PD, Wilson LE, Remsburg RE, Cullins VE, Huggins GR. Contraceptive outcomes among postpartum and post-abortal adolescents. Contraception 1994;50:451 – 60. [8] Stenchever MA, Herbst AL, Droegemueller W, Mishell D, editors. Comprehensive gynecology. St. Louis (MO)7 Elsevier Health Sciences; 2001. p. 309 [chapter 12]. [9] Redmond G, Godwin AJ, Olson W, Lippman JS. Use of placebo controls in an oral contraceptive trial: methodological issues and adverse event incidence. Contraception 1999;60:81 – 5. [10] Goldzieher JW, Moses LE, Averkin E, Scheel C, Taber BZ. A placebocontrolled double-blind crossover investigation of the side effects attributed to oral contraceptives. Fertil Steril 1971;22:609 – 23. [11] Coney P, Washenik K, Langley RGB, DiGiovanna JJ, Harrison DD. Weight change and adverse event incidence with a low-dose oral contraceptive: two randomized, placebo-controlled trials. Contraception 2001;63:297 – 302. [12] Almagor M, Ben-Porath MS. Mood changes during the menstrual cycle and their relation to the use of oral contraceptives. J Psychosom Res 1991;35:721 – 8. [13] Walker A, Bancroft J. Relationship between premenstrual symptoms and oral contraceptive use: a controlled study. Psychosom Med 1990; 52:86 – 96. [14] Rosenthal SL, Cotton S, Ready JN, Potter LS, Succop PA. Adolescents’ attitudes and experiences regarding levonorgestrel 100 mcg/ethinyl estradiol 20 mcg. J Pediatr Adolesc Gynecol 2002;15: 301 – 5. [15] Risser WL, Gefter LR, Barratt MS, et al. Weight change in adolescents who used hormonal contraception. J Adolesc Health 1999;24:433. [16] Davis AR, Westhoff CL, O’Connell KJ, et al. Oral contraceptives for dysmenorrhea in adolescent girls: a placebo-controlled randomized trial. Obstet Gynecol 2005;106:97 – 104. [17] Hillen TI, Grbavac SL, Johnston PJ, et al. Primary dysmenorrhea in young Western Australian women: prevalence, impact, and knowledge of treatment. J Adolesc Health 1999;25:40 – 5. [18] Moos RH. Menstrual Distress Questionnaire manual. Los Angeles (Calif)7 Western Psychological Services; 1991. [19] McDowell I, Newell C. Measuring health: a guide to rating scales and questionnaires. New York7 Oxford University Press; 1996. [20] Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977;1: 385 – 401. [21] Blascovich J, Tomaka J. Measures of self-esteem. In: Robinson JP, Shaver JR, Wrightsman LS, editors. Measures of personality and
[22] [23] [24] [25] [26]
[27]
[28] [29] [30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
social psychological attitudes. San Diego (CA)7 Academic Press; 1991. p. 115 – 60. Buckelew SP, Parker JC, Keefe FJ, et al. Self-efficacy and pain behavior among subjects with fibromyalgia. Pain 1994;59:377 – 84. Davis AR, Teal SB. Controversies in adolescent hormonal contraception. Infertil Reprod Med Clin North Am 2003;14:141 – 56. Barsky AJ, Saintfort R, Rogers MP, Borus JF. Nonspecific medication side effects and the Nocebo phenomenon. JAMA 2002;287:622 – 7. Myers MG, Cairns JA, Singer J. The consent form as a possible cause of side effects. Clin Pharmacol Ther 1987;42:250 – 3. Gallo MF, Grimes DA, Schultz KF, Helmerhorst FM. Combination estrogen–progestin contraceptives and body weight: systematic review of randomized controlled trials. Obstet Gynecol 2004;103: 359 – 73. Nilsson A, Almgren PE. Psychiatric symptoms during the postpartum period as related to use of oral contraceptives. Br Med J 1968; 2:453 – 5. Herzberg BN, Johnson AL, Brown S. Depressive symptoms and oral contraceptives. Br Med J 1970;4:142 – 5. Worsley A, Chang A. Oral contraceptives and emotional state. J Psychosom Res 1978;22:13 – 6. Joffe H, Cohen L, Harlow B. Impact of oral contraceptive pill use on premenstrual mood: predictors of improvement and deterioration. Am J Obstet Gynecol 2003;189:1523 – 30. Rosenberg M, Waugh M. Oral contraceptive discontinuation: a prospective evaluation of frequency and reasons. Am J Obstet Gynecol 1998;179:577 – 82. Bancroft J, Rennie D. The impact of oral contraceptives on the experience of perimenstrual mood, clumsiness, food craving and other symptoms. J Psychosom Res 1993;37:195 – 202. Maes M, Claes M, Schotte C, et al. Disturbances in dexamethasone suppression test and lower availability of l-tryptophan and tyrosine in early puerperium and in women under contraceptive therapy. J Psychosom Res 1992;36:191 – 7. Rasgon NL, Rao RC, Hwang S, et al. Depression in women with polycystic ovary syndrome: clinical and biochemical correlates. J Affect Disord 2003;74:299 – 304. Graham CA, Sherwin BB. The relationship between mood and sexuality in women using an oral contraceptive as a treatment for premenstrual symptoms. Psychoneuroendocrinology 1993;18: 273 – 81. Garrison CZ, Addy CL, Jacksosn KL, et al. The CES-D as a screen for depression and other psychiatric disorders in adolescents. J Am Acad Child Adolesc Psychiatry 1991;30:636 – 41. Pumariega AJ, Johnson NP, Sheridan D, et al. The influence of race and gender on depressive and substance abuse symptoms in high-risk adolescents. Cult Divers Ment Health 1996;2:115 – 23. Prescott CA, McArdle JJ, Hishinuma ES, et al. Prediction of major depression and dysthymia from CES-D scores among ethnic minority adolescents. J Am Acad Child Adolesc Psychiatry 1998;37: 495 – 503. Thomas JL, Nones GN, Scarinci IC, et al. The utility of the CES-D as a depression screening measure among low-income women attending primary care clinics. Int J Psychiatry Med 2001;31:25 – 40.
