- Email: [email protected]
Fertility awareness-based methods for contraception: systematic review of randomized controlled trials
Contraception 72 (2005) 85 – 90
Review article
Fertility awareness-based methods for contraception: systematic review of randomized controlled trials David A. GrimesT,1, Maria F. Gallo, Vera Grigorieva, Kavita Nanda, Kenneth F. Schulz Family Health International, Research Triangle Park, NC 27709, USA Received 3 March 2005; revised 30 March 2005; accepted 30 March 2005
Abstract Context: Fertility awareness-based (FAB) methods of family planning estimate the fertile days of the menstrual cycle whether by observing fertility signs such as cervical secretions and basal body temperature or by monitoring cycle days. The efficacy and acceptability of these methods are unclear. Methods: We retrieved and analyzed all randomized controlled trials that examined any FAB method used for contraception. Results: Because of poor methods and reporting, pregnancy rates could not be determined. A trial in Colombia found similar numbers of pregnancies among women randomized to the ovulation and symptothermal methods, but a trial in Los Angeles observed more pregnancies in the group assigned to the ovulation method. In the two American trials, recruitment of participants was unexpectedly difficult. Continuation rates were poor. Conclusion: Despite intensive training and ongoing support, most participants in these trials discontinued prematurely. The comparative efficacy of these methods remains unknown. However, with the ovulation and symptothermal methods, pregnancies appear to be common; method continuation rates are low. D 2005 Elsevier Inc. All rights reserved. Keywords: Contraception; Female; Periodic abstinence; Rhythm; Systematic review; Meta-analysis
1. Introduction Fertility awareness-based (FAB) methods of family planning binvolve identification of the fertile days of the menstrual cycle, whether by observing fertility signs such as cervical secretions and basal body temperature, or by monitoring cycle days. Fertility awareness-based methods can be used in combination with abstinence or barrier methods during the fertile timeQ [1]. Several names have been used to describe this approach to contraception—rhythm, natural family planning and periodic abstinence [2]. FAB methods can be used with abstinence from sexual intercourse. Alternatively, these methods can be used with barrier contraceptives or withdrawal during presumed fertile times.
T Corresponding author. Tel.: +1 919 544 7040x541; fax: +1 919 544 7261. E-mail address: [email protected] (D.A. Grimes). 1 Overnight mail address: Family Health International, 2224 E. NC Highway 54, Durham, NC 27713. 0010-7824/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.contraception.2005.03.010
FAB methods include several approaches to timing coitus. The earliest was the calendar (or rhythm) method. This method calculates the estimated fertile days of the menstrual cycle based on past cycle lengths. In part because of biologic variation in ovulation, this method reportedly is the least effective type of FAB method [2]. The basal body temperature method was another early approach, stemming from the observation that basal body temperature rises after ovulation [2]. Basal body temperature is the temperature measured upon waking and before any activity. The rise in serum progesterone after ovulation causes a small increase in this baseline temperature, which can be detected with a digital or mercury thermometer with fine gradations. With this method, a couple refrains from coitus from the first day of a menstrual period until the third consecutive day of temperature elevation above baseline. In the 1970s, attempts began to predict ovulation based on characteristics of the cervical secretions and sensations at the vaginal introitus (known as the ovulation, cervical mucus or the Billings method) [3]. During the menstrual cycle, women may detect changes in the characteristics of the cervical
86
D.A. Grimes et al. / Contraception 72 (2005) 85–90
mucus and different sensations at the opening of the vagina [2]. These changes are used to estimate fertile days during the cycle. With this method, couples avoid coitus during the presumed fertile days, indicated by the presence of fertile mucus (estrogenic). The final day of having transparent, slippery mucus (termed bpeak dayQ) coincides with ovulation. The fertile interval ends on the fourth morning after the peak day [4]. Moreover, in the pre-ovulatory phase, couples should avoid coitus on consecutive days to enable detection of cervical mucus changes. A potential advantage of this method over earlier ones is that coitus is allowed both before and after presumed ovulation [2]. Another variation in the abovementioned themes is the symptothermal method. It combines two or more of the abovementioned approaches such as calendar calculations, temperature shift and mucus changes. For example, abstinence begins when the calendar or cervical mucus suggests the onset of the fertile period; coitus can resume when another indication (basal body temperature or cervical mucus) suggests the end of the fertile phase. Other approaches to FAB methods include hormonal surveillance to predict the timing of ovulation. For example, a personal hormone monitor has been used to identify fertile days in the cycle during which coitus should be avoided [5]. Although numerous cycle monitors have been developed, not much is known about their potential clinical efficacy [6]. FAB methods have both advantages and disadvantages. In most methods, there are no or minimal costs, no drugs or physical devices are required and medical personnel are not required. No medical contraindications exist. On the other hand, signs and symptoms of the fertile phase may be subtle [2,7]. Irregular cycles can make these methods difficult. The frequency of vaginal intercourse may be decreased [8]. In the 1980s, the proportion of all women of reproductive age who relied on FAB methods was reported to be 0–11% in developing countries and 1–13% in developed countries [9]. By 2002, approximately 3% of women of reproductive age worldwide were using FAB methods [10]. Hence, determining the comparative efficacy and acceptability of available methods is important for current and potential future users of these methods. 2. Materials and methods We retrieved and analyzed all randomized controlled trials that examined any FAB method used for contraception (Table 1). We included all randomized controlled trials in any language that compare any FAB method for contraception with a placebo; another method, including an alternative FAB methods; or a FAB method used in conjunction with another contraceptive. All couples included in the eligible trials were incorporated in the review. Most trials included only highly motivated volunteers who received intensive training and reinforcement in the proper use of the methods. For example, in a Colombia trial [12], participants were visited in their
Table 1 Characteristics of included studies Study Methods Participants
Interventions
Outcomes Notes
Allocation concealment Study Methods Participants
Interventions
Outcomes
Notes
Allocation concealment
Kass-Annese et al. [11] Randomized controlled trial Fifty-three women aged 18 – 40 years in Los Angeles who were healthy, sexually active, with a history of regular menstrual cycles, at least one menses since last pregnancy, no hormonal contraception in 3 months, normal pelvic examination and Pap smear results and negative urine pregnancy test results. Recruitment was from October 1984 to November 1985. Today vaginal sponge used in usual way versus Today vaginal sponge used only during presumed fertile days as judged by calendar calculation and basal body temperature. Pregnancy over 12 months of observation; reasons for premature termination from the trial. Method of randomization and allocation concealment not specified. Sample size calculation not provided, although recruitment fell short of planned number. Two participants dropped after randomization for discovery of ineligibility. Large premature discontinuation rates: 75% and 78%, respectively, failed to complete 12 months of observation. Two pregnancies in the sponge-only group and four in the periodic abstinence plus sponge group. Unclear Medina et al. [12] Randomized controlled trial Five hundred sixty-six women aged 18–39 years in Bogota and Palmira, Colombia, who were healthy, not desiring pregnancy in the next 2 years, in a stable union and willing to use the assigned method for 12–24 months. Recruitment was from August 1976 to December 1978; trial closed in June 1979. Ovulation method based on an atlas [13] versus symptothermal method based on double-check method [14] that included basal body temperature, changes in cervical mucus, symptoms such as breast tenderness and length of six prior menstrual cycles. Use-effectiveness as judged by pregnancy rates; characteristics of volunteers for the trial; characteristics of those who did and did not drop out of the trial. Method of randomization and allocation concealment not specified. Sample size and power calculation not specified. All participants entered a training program lasting 3–5 months. Thereafter, all participants were visited monthly by study personnel for follow-up and counseling. Large premature discontinuation rates primarily due to accidental pregnancy or problems with the methods. Overall, 57% of participants quit during the training period and another 30% dropped out during the active phase of the trial. Investigators excluded participants for noncompliance with assigned regimen. Unclear
(continued on next page)
D.A. Grimes et al. / Contraception 72 (2005) 85–90 Table 1 (continued) Study Methods Participants Interventions Outcomes Notes Allocation concealment Study Methods Participants
Interventions
Outcomes
Notes
Allocation concealment
Wade et al. [15]
Interim report from Wade et al. [16] Unclear Wade et al. [16] Randomized controlled trial A total of 1247 women aged 20 – 39 years in Los Angeles County and Orange County, California, who were healthy, with regular menstrual cycles, in a stable relationship, expressed a desire to avoid pregnancy for 2 years and agreed to randomization and participation for at least 12 months. Recruitment was from January 1976 to August 1978. Ovulation method based on an atlas [13] versus symptothermal method based on double-check method [14] that included basal body temperature, changes in cervical mucus, symptoms such as breast tenderness and length of six prior menstrual cycles. Use-effectiveness as judged by pregnancy rates; acceptability as judged by dropout rates; difficulties in recruitment of volunteers. Method of randomization and allocation concealment not specified. No sample size or power calculation provided. Noncompliant participants were dropped from the study by investigators or encouraged to drop out. High rate of discontinuation in both groups. Unclear
homes monthly by study personnel b. . .to discuss problems in using the method, and to offer counseling as needed.Q Similarly, in a larger Los Angeles trial [16], participants interacted with study staff monthly. We included any FAB method used to prevent pregnancy. These included but were not limited to the calendar method, the basal body temperature method, the ovulation or the Billings method, the symptothermal method and ovulation prediction devices that rely on hormone assays. Interventions could include FAB methods used with or without a barrier contraceptive or withdrawal. Pregnancy rates were the primary outcome measure. Secondary outcome measures included continuation rates and acceptability, if reported. We searched the computerized databases Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, POPLINE, EMBASE and LILACS (each from its inception to January 2004) for randomized controlled trials of FAB methods. We examined the reference list of each trial as well as that of review articles. We searched the websites of relevant organizations, including the World Health Organization and the International Planned Parenthood Federation, for citations. We wrote to the corresponding author of each
87
trial identified to solicit other published or unpublished trials that we may have missed. Finally, we wrote to organizations in the United States, United Kingdom, Canada, Australia and New Zealand that teach these methods. Our search strategy for CENTRAL included the search term bcontraception.Q Our search strategy for MEDLINE (PubMed) used the following approach: [(natural family planning OR NFP OR periodic abstinence OR rhythm method* OR calendar method* or symptothermal method*) AND contracept*] AND (clinical trial OR clinical trials OR randomized controlled trial* OR clinical trial[pt] OR randomized controlled trial[pt]). Our search strategy for POPLINE included the following: (natural family planning/nfp/rhythm method calendar/symptothermal method/bperiodic abstinenceQ/bcalendar methodQ/ cervical mucus method/sexual abstinence) and (clinical trial*/controlled clinical trial*/randomized controlled trial*). We also searched POPLINE with the following terms: fertility awareness/fertility awareness method*/fertility knowledge/FAM/periodic abstinence/natural family planning/nfp/cervical secretion*/cervical mucus/Billings method/Creighton method/basal body temperature/double-check method/mucothermic/symptothermic/calendar method/ rhythm method/simple rule method/cycle beads/two-day method/personal hormone monitoring/abstinence and clinical trials (kw). Our search in EMBASE included the following: [periodic(w)abstinence or periodic(w)sexual(w)abstinence or natural family planning or rhythm(w) method or calendar(w) method or sympto(w)thermal(w)method]. Our search in LILACS used the following terms: natural family planning or nfp or periodic abstinence or calendar method or metodo del calendario or symptothermal method or rhythm method or metodo de ritmo or method do ritmo or birth intervals or intervalo entre nacimientos. We assessed all titles and abstracts found for inclusion. Because of high losses after randomization, we did not enter the data into RevMan or calculate odds ratios. We evaluated the methodological quality of the trials for potential biases by qualitatively assessing the study design, randomization method, allocation concealment, blinding, premature discontinuation rates and loss to follow-up rates. The trial conducted in Los Angeles by Kass-Annese et al. [11] compared use of the Today vaginal contraceptive sponge only during the presumed fertile period (as estimated by calendar calculations and basal body temperature) versus customary use of the sponge (with each act of coitus). The FAB method used the calendar rhythm method for judging the onset of the fertile days and the shift in basal body temperature for estimating its end; mucus was not evaluated. Those assigned to fertility awareness plus sponges were given sponges, a basal body temperature thermometer, a diary for recording events and instructions on use of the sponge. They returned within 2 weeks for instructions on fertility awareness. Thereafter, they returned for a maximum of three
88
D.A. Grimes et al. / Contraception 72 (2005) 85–90
follow-up training visits. Those assigned to the spongeonly received sponges, diaries and instructions on use of the sponge. The trial was designed to run for 12 months. bSince the drop-out rate was high, recruitment was ended and all women followed-up as long as possible, for up to 12 month. . .About one-quarter of the volunteers completed 12 months of follow-up. . ..Q [11]. Medina et al. [12] enrolled 566 couples in two cities in Colombia. The trial was designed to compare the efficacy of the ovulation method versus the symptothermal method. For the former, the atlas of Billings [13] was the reference, and, for the latter, it was the text of Thyma [14]. The symptothermal method included information about basal body temperature, cervical mucus changes, symptoms such as breast tenderness and length of the preceding six menstrual cycles. All couples enrolled entered a training period lasting 3 –5 months. Thereafter, couples were visited monthly by research staff who provided ongoing counseling about the assigned method. The observation period of the trial was intended to be 12 –24 months. Wade et al. [16] conducted a companion trial to the Colombian trial [12] in Southern California; the treatments and analytic methods were similar. An interim report [15] from the full trial [16] will not be considered separately. The methods compared were the ovulation method [13] and the symptothermal method [14]. Couples enrolled entered a training period that included both individual and group instructions. Randomized participants were allowed to drop out for any reason. Those who completed the training phase were invited to enter the follow-up phase; those who were noncompliant or who failed to learn the method were involuntarily dropped by the investigators. During the follow-up phase, noncompliant couples were encouraged to drop out. During the entire study, each female participant met monthly with her teacher and had a pregnancy test done. The planned length of follow-up was 12 months. The conduct and reporting of the three trials were poor. Because the smaller Los Angeles trial was abandoned, we will focus on the two larger trials [12,16]. Although each reported to be a randomized controlled trial, no information was provided concerning the method used to generate the random sequence. Similarly, neither mentioned allocation concealment. Sample size calculations were not provided. Blinding was apparently not used for either trial, but this would be unlikely to bias determination of the principal outcome (pregnancy). A more serious violation of research methods was the endorsement and, indeed, encouragement of discontinuation after randomization. Investigators in both the Colombia and the larger Los Angeles trials unilaterally excluded participants after randomization if they were having difficulty during the training phase. Those noncompliant with the method in the main follow-up phase were encouraged to discontinue. In contrast, the investigators should have attempted to keep all randomized participants in the trial. Loss of participants after randomization introduces bias
[17,18], and the losses here were large. In general, if more than 20% of participants are lost after randomization, trial validity is suspect [17]. 3. Results The three trials revealed difficult recruitment and high dropout rates. Because of unanticipated difficulty in recruiting participants and high discontinuation rates in the second Los Angeles trial [11], the trial was abandoned. The investigators cautioned that they could not assess the efficacy of using the vaginal sponge with fertility awareness versus using the sponge alone. However, four pregnancies occurred among the 24 women assigned to the FAB method plus the sponge in contrast to two among the 27 women in the spongeonly group. Similarly, pregnancy rates cannot be determined from the Colombia trial [12]. Most randomized participants dropped out before beginning the observation period: 149 of 279 couples (53%) assigned to the ovulation method discontinued during training in contrast to 176 of 287 assigned to the symptothermal method (61%). Eleven women assigned to the ovulation method and 32 assigned to the symptothermal method were excluded from analysis because of noncompliance during the training phase, and 1 more in each group was excluded during the active observation phase. Only a minority of participants entered the follow-up phase: 130 assigned to the ovulation method and 111 to the symptothermal method. Of these participants, 86 (31%) and 82 (30%), respectively, dropped out during the follow-up phase. With the training and follow-up phases combined, 72 women assigned to the ovulation method became pregnant as compared with 71 assigned to the alternative method. The corresponding number of participants who discontinued because of lack of interest or dissatisfaction with the method was 63 and 69, respectively. Because of inadequate methods, pregnancy rates cannot be determined from the Southern California trial [16]. The total number of participants randomized to the ovulation method and symptothermal method was 619 and 628, respectively. Of these participants, 265 (43%) and 238 (39%) dropped out during the training period. Another 314 women (163 and 151, respectively) were still in training at the end of the study and apparently contributed no outcome data. Further losses seem to have occurred since the report indicates that only 191 and 239 participants entered the formal follow-up phase. Stated alternatively, 428 participants (69%) assigned to the ovulation method and 389 (62%) assigned to the symptothermal method were dropped after randomization and before entering the planned observation period. During the main observation period, additional 123 (20%) and 100 (16%) participants, respectively, dropped out. Pregnancies were reported more often among those assigned to the ovulation method [16]. During the training phase, 92 pregnancies occurred in the group assigned to the ovulation method in contrast to 55 in the group assigned to
D.A. Grimes et al. / Contraception 72 (2005) 85–90
the symptothermal method. During the follow-up phase, the number of pregnancies was 42 and 20, respectively. The total number of reported pregnancies was 134 and 75, respectively. Discontinuations because of lack of interest or dissatisfaction with the method were similar for the two groups. A total of 85 and 90 participants, respectively, discontinued for these reasons. 4. Discussion The comparative efficacy of FAB methods remains unresolved. The conduct and reporting of randomized controlled trials of FAB methods were of such poor quality that valid pregnancy rates cannot be determined [18]. One small trial [11] was abandoned, and the other trials [12,16] lost most participants after randomization. No large difference in efficacy was likely in the Colombia trial [12], although the symptothermal method may have been more effective than the ovulation method in the Los Angeles trial [16]. A re-analysis of the World Health Organization case series report of the ovulation method [19] by Trussell [20] found that 23% of women became accidentally pregnant in the first year. Two trial reports addressed recruitment problems. In the larger Los Angeles trial, the investigators noted that recruitment bdifficulties were quite unexpected since this study featured the most promising methods of NFP [natural family planning], had the wholehearted support of the Catholic hierarchy in the Los Angeles area, received wide media coverage, and utilized two public relations firms to aid recruitmentQ [16]. The smaller Los Angeles trial was stopped because of poor recruitment. In it, the problems with the FAB method were boverwhelmingly related to the inconvenience of using BBT [basal body temperature]Q [11]. Continuation rates were low. Despite enrolling motivated volunteers, having thorough instruction and providing ongoing education and counseling of participants, most couples discontinued their assigned method before entering the planned observation period [12,16]. These techniques can be difficult for many couples since sexual abstinence for 7–17 days of the cycle, excluding menses, can be required [7,8,21]. An international study over 13 menstrual cycles reported that 5–62% of couples experienced difficulties with the mandated abstinence [22]. Focus groups conducted by the World Health Organization in developing countries revealed that couples commonly were sexually active during the presumed fertile days. Focus group respondents reported that drawbacks to FAB methods included trouble learning the method, difficulty using it and the challenge of abstinence. On the other hand, respondents liked the absence of side effects, perceived greater safety than with hormonal or intrauterine methods and acknowledged low cost [23]. Properly done randomized controlled trials are required to determine the comparative effectiveness of contraceptive methods [20,24,25]. Such evidence is not available for FAB methods.
89
The comparative efficacy of FAB methods for contraception remains unknown. However, with the ovulation and symptothermal methods, pregnancies appear to be common; method continuation rates are low.
Acknowledgments A version of this review has been published in the Cochrane Library (FORMAT: Grimes DA, Gallo MF, Grigorieva V, Nanda K, Schulz KF. Fertility awarenessbased methods for contraception. The Cochrane Database of Systematic Reviews 2004, Issue 4. Art. No.: CD004860.pub2. DOI: 10.1002/14651858.CD004860.pub2.) Cochrane reviews are regularly updated to include new research and in response to comments and criticism from readers. If you wish to comment on this review, please contact Dr. David Grimes at Family Health International (FHI), P.O. Box 13950, Research Triangle Park, NC 27709, USA, or email him at [email protected]. The results of a Cochrane review can be interpreted differently, depending on a person’s perspective and circumstance. Please consider the conclusions presented carefully. They are the opinions of the review authors and are not necessarily shared by the Cochrane Collaboration. Carol Manion of FHI assisted with the literature searches. Support for this study was provided by the FHI with funds from the U.S. Agency for International Development and the National Institute of Child Health and Human Development although the views expressed in this article do not necessarily reflect those of the FHI or the agencies funding the research.
References [1] World Health Organization. Medical eligibility criteria for contraceptive use. 3rd ed. http://www.who.int/reproductive-health/publications/ MEC_3/index.htm [accessed February 11, 2005]. [2] International Planned Parenthood Federation. Statement on periodic abstinence for family planning. London (England)7 International Planned Parenthood Federation; 1995. [3] Billings JJ. Natural family planning. Med J Aust 1978;2:436. [4] Pyper CMM, Knight J. Fertility awareness methods of family planning: the physiological background, methodology and effectiveness of fertility awareness methods. J Fam Plann Reprod Health Care 2001;27:103 – 10. [5] Bonnar J, Flynn A, Freundl G, Kirkman R, Royston R, Snowden R. Personal hormone monitoring for contraception. Br J Fam Plann 1999; 24:128 – 34. [6] Freundl G, Godehardt E, Kern PA, Frank-Herrmann P, Koubenec HJ, Gnoth Ch. Estimated maximum failure rates of cycle monitors using daily conception probabilities in the menstrual cycle. Hum Reprod 2003;18:2628 – 33. [7] Hughes J. Natural family planning: very risky business. Conscience 1988;9:7 – 12. [8] Neumann GH. How natural is bnatural family planningQ? Critical remarks from the biological behavior viewpoint. in German. Gynakologe 1989;22:1 – 6. [9] Gray RH, Kambic RT. Epidemiological studies of natural family planning. Hum Reprod 1988;3:693 – 8.
90
D.A. Grimes et al. / Contraception 72 (2005) 85–90
[10] Population Reference Bureau. Family planning worldwide. 2002 Data sheet. http://www.prb.org/Template.cfm?Section = PRB&template = / Content/ContentGroups/Datasheets/Family_Planning_Worldwide_ 2002_Data_Sheet.htm [accessed February 11, 2005]. [11] Kass-Annese B, Kennedy KI, Forrest K, Danzer H, Reading A, Hughes H. A study of the vaginal contraceptive sponge used with and without the fertility awareness method. Contraception 1989;40: 701 – 14. [12] Medina JE, Cifuentes A, Abernathy JR, Spieler JM, Wade ME. Comparative evaluation of two methods of natural family planning in Columbia. Am J Obstet Gynecol 1980;138:1142 – 7. [13] Billings J. Natural family planning: the ovulation method. 2nd American edition. Collegeville (Minn)7 The Liturgical Press; 1973. [14] Thyma P. Fertile and infertile days in marriage life. Author’s edition. Fall River (Mass)7 Raritan Printing Company; 1973. [15] Wade ME, McCarthy P, Abernathy JR, et al. Randomized prospective study of the use-effectiveness of two methods of natural family planning: an interim report. Am J Obstet Gynecol 1979;134: 628 – 31. [16] Wade ME, McCarthy P, Braunstein GD, et al. A randomized prospective study of the use-effectiveness of two methods of natural family planning. Am J Obstet Gynecol 1981;141:368 – 76.
[17] Schulz KF, Grimes DA. Sample size slippages in randomised trials: exclusions and the lost and wayward. Lancet 2002;359:781 – 5. [18] Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallelgroup randomised trials. Lancet 2001;357:1191 – 4. [19] World Health Organization. A prospective multicentre trial of the ovulation method of natural family planning: II. The effectiveness phase. Fertil Steril 1981;36:591 – 8. [20] Trussell J. Methodological pitfalls in the analysis of contraceptive failure. Stat Med 1991;10:201 – 20. [21] Ryder B, Campbell H. Natural family planning in the 1990s. Lancet 1995;346:233 – 4. [22] World Health Organization. A prospective multicenter trial of the ovulation method of natural family planning: V. Psychosexual aspects. Fertil Steril 1987;47:765 – 72. [23] d’Arcangues C, Kennedy KI. Periodic abstinence and calendar method use in Hungary, Peru, the Philippines, and Sri Lanka. Contraception 2001;64:209 – 15. [24] Trussell J, Grummer-Strawn L. Further analysis of contraceptive failure of the ovulation method. Am J Obstet Gynecol 1991;165:2054 – 9. [25] Glasziou P, Vandenbroucke JP, Chalmers I. Assessing the quality of research. BMJ 2004;328:39 – 41.
Review article
Fertility awareness-based methods for contraception: systematic review of randomized controlled trials David A. GrimesT,1, Maria F. Gallo, Vera Grigorieva, Kavita Nanda, Kenneth F. Schulz Family Health International, Research Triangle Park, NC 27709, USA Received 3 March 2005; revised 30 March 2005; accepted 30 March 2005
Abstract Context: Fertility awareness-based (FAB) methods of family planning estimate the fertile days of the menstrual cycle whether by observing fertility signs such as cervical secretions and basal body temperature or by monitoring cycle days. The efficacy and acceptability of these methods are unclear. Methods: We retrieved and analyzed all randomized controlled trials that examined any FAB method used for contraception. Results: Because of poor methods and reporting, pregnancy rates could not be determined. A trial in Colombia found similar numbers of pregnancies among women randomized to the ovulation and symptothermal methods, but a trial in Los Angeles observed more pregnancies in the group assigned to the ovulation method. In the two American trials, recruitment of participants was unexpectedly difficult. Continuation rates were poor. Conclusion: Despite intensive training and ongoing support, most participants in these trials discontinued prematurely. The comparative efficacy of these methods remains unknown. However, with the ovulation and symptothermal methods, pregnancies appear to be common; method continuation rates are low. D 2005 Elsevier Inc. All rights reserved. Keywords: Contraception; Female; Periodic abstinence; Rhythm; Systematic review; Meta-analysis
1. Introduction Fertility awareness-based (FAB) methods of family planning binvolve identification of the fertile days of the menstrual cycle, whether by observing fertility signs such as cervical secretions and basal body temperature, or by monitoring cycle days. Fertility awareness-based methods can be used in combination with abstinence or barrier methods during the fertile timeQ [1]. Several names have been used to describe this approach to contraception—rhythm, natural family planning and periodic abstinence [2]. FAB methods can be used with abstinence from sexual intercourse. Alternatively, these methods can be used with barrier contraceptives or withdrawal during presumed fertile times.
T Corresponding author. Tel.: +1 919 544 7040x541; fax: +1 919 544 7261. E-mail address: [email protected] (D.A. Grimes). 1 Overnight mail address: Family Health International, 2224 E. NC Highway 54, Durham, NC 27713. 0010-7824/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.contraception.2005.03.010
FAB methods include several approaches to timing coitus. The earliest was the calendar (or rhythm) method. This method calculates the estimated fertile days of the menstrual cycle based on past cycle lengths. In part because of biologic variation in ovulation, this method reportedly is the least effective type of FAB method [2]. The basal body temperature method was another early approach, stemming from the observation that basal body temperature rises after ovulation [2]. Basal body temperature is the temperature measured upon waking and before any activity. The rise in serum progesterone after ovulation causes a small increase in this baseline temperature, which can be detected with a digital or mercury thermometer with fine gradations. With this method, a couple refrains from coitus from the first day of a menstrual period until the third consecutive day of temperature elevation above baseline. In the 1970s, attempts began to predict ovulation based on characteristics of the cervical secretions and sensations at the vaginal introitus (known as the ovulation, cervical mucus or the Billings method) [3]. During the menstrual cycle, women may detect changes in the characteristics of the cervical
86
D.A. Grimes et al. / Contraception 72 (2005) 85–90
mucus and different sensations at the opening of the vagina [2]. These changes are used to estimate fertile days during the cycle. With this method, couples avoid coitus during the presumed fertile days, indicated by the presence of fertile mucus (estrogenic). The final day of having transparent, slippery mucus (termed bpeak dayQ) coincides with ovulation. The fertile interval ends on the fourth morning after the peak day [4]. Moreover, in the pre-ovulatory phase, couples should avoid coitus on consecutive days to enable detection of cervical mucus changes. A potential advantage of this method over earlier ones is that coitus is allowed both before and after presumed ovulation [2]. Another variation in the abovementioned themes is the symptothermal method. It combines two or more of the abovementioned approaches such as calendar calculations, temperature shift and mucus changes. For example, abstinence begins when the calendar or cervical mucus suggests the onset of the fertile period; coitus can resume when another indication (basal body temperature or cervical mucus) suggests the end of the fertile phase. Other approaches to FAB methods include hormonal surveillance to predict the timing of ovulation. For example, a personal hormone monitor has been used to identify fertile days in the cycle during which coitus should be avoided [5]. Although numerous cycle monitors have been developed, not much is known about their potential clinical efficacy [6]. FAB methods have both advantages and disadvantages. In most methods, there are no or minimal costs, no drugs or physical devices are required and medical personnel are not required. No medical contraindications exist. On the other hand, signs and symptoms of the fertile phase may be subtle [2,7]. Irregular cycles can make these methods difficult. The frequency of vaginal intercourse may be decreased [8]. In the 1980s, the proportion of all women of reproductive age who relied on FAB methods was reported to be 0–11% in developing countries and 1–13% in developed countries [9]. By 2002, approximately 3% of women of reproductive age worldwide were using FAB methods [10]. Hence, determining the comparative efficacy and acceptability of available methods is important for current and potential future users of these methods. 2. Materials and methods We retrieved and analyzed all randomized controlled trials that examined any FAB method used for contraception (Table 1). We included all randomized controlled trials in any language that compare any FAB method for contraception with a placebo; another method, including an alternative FAB methods; or a FAB method used in conjunction with another contraceptive. All couples included in the eligible trials were incorporated in the review. Most trials included only highly motivated volunteers who received intensive training and reinforcement in the proper use of the methods. For example, in a Colombia trial [12], participants were visited in their
Table 1 Characteristics of included studies Study Methods Participants
Interventions
Outcomes Notes
Allocation concealment Study Methods Participants
Interventions
Outcomes
Notes
Allocation concealment
Kass-Annese et al. [11] Randomized controlled trial Fifty-three women aged 18 – 40 years in Los Angeles who were healthy, sexually active, with a history of regular menstrual cycles, at least one menses since last pregnancy, no hormonal contraception in 3 months, normal pelvic examination and Pap smear results and negative urine pregnancy test results. Recruitment was from October 1984 to November 1985. Today vaginal sponge used in usual way versus Today vaginal sponge used only during presumed fertile days as judged by calendar calculation and basal body temperature. Pregnancy over 12 months of observation; reasons for premature termination from the trial. Method of randomization and allocation concealment not specified. Sample size calculation not provided, although recruitment fell short of planned number. Two participants dropped after randomization for discovery of ineligibility. Large premature discontinuation rates: 75% and 78%, respectively, failed to complete 12 months of observation. Two pregnancies in the sponge-only group and four in the periodic abstinence plus sponge group. Unclear Medina et al. [12] Randomized controlled trial Five hundred sixty-six women aged 18–39 years in Bogota and Palmira, Colombia, who were healthy, not desiring pregnancy in the next 2 years, in a stable union and willing to use the assigned method for 12–24 months. Recruitment was from August 1976 to December 1978; trial closed in June 1979. Ovulation method based on an atlas [13] versus symptothermal method based on double-check method [14] that included basal body temperature, changes in cervical mucus, symptoms such as breast tenderness and length of six prior menstrual cycles. Use-effectiveness as judged by pregnancy rates; characteristics of volunteers for the trial; characteristics of those who did and did not drop out of the trial. Method of randomization and allocation concealment not specified. Sample size and power calculation not specified. All participants entered a training program lasting 3–5 months. Thereafter, all participants were visited monthly by study personnel for follow-up and counseling. Large premature discontinuation rates primarily due to accidental pregnancy or problems with the methods. Overall, 57% of participants quit during the training period and another 30% dropped out during the active phase of the trial. Investigators excluded participants for noncompliance with assigned regimen. Unclear
(continued on next page)
D.A. Grimes et al. / Contraception 72 (2005) 85–90 Table 1 (continued) Study Methods Participants Interventions Outcomes Notes Allocation concealment Study Methods Participants
Interventions
Outcomes
Notes
Allocation concealment
Wade et al. [15]
Interim report from Wade et al. [16] Unclear Wade et al. [16] Randomized controlled trial A total of 1247 women aged 20 – 39 years in Los Angeles County and Orange County, California, who were healthy, with regular menstrual cycles, in a stable relationship, expressed a desire to avoid pregnancy for 2 years and agreed to randomization and participation for at least 12 months. Recruitment was from January 1976 to August 1978. Ovulation method based on an atlas [13] versus symptothermal method based on double-check method [14] that included basal body temperature, changes in cervical mucus, symptoms such as breast tenderness and length of six prior menstrual cycles. Use-effectiveness as judged by pregnancy rates; acceptability as judged by dropout rates; difficulties in recruitment of volunteers. Method of randomization and allocation concealment not specified. No sample size or power calculation provided. Noncompliant participants were dropped from the study by investigators or encouraged to drop out. High rate of discontinuation in both groups. Unclear
homes monthly by study personnel b. . .to discuss problems in using the method, and to offer counseling as needed.Q Similarly, in a larger Los Angeles trial [16], participants interacted with study staff monthly. We included any FAB method used to prevent pregnancy. These included but were not limited to the calendar method, the basal body temperature method, the ovulation or the Billings method, the symptothermal method and ovulation prediction devices that rely on hormone assays. Interventions could include FAB methods used with or without a barrier contraceptive or withdrawal. Pregnancy rates were the primary outcome measure. Secondary outcome measures included continuation rates and acceptability, if reported. We searched the computerized databases Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, POPLINE, EMBASE and LILACS (each from its inception to January 2004) for randomized controlled trials of FAB methods. We examined the reference list of each trial as well as that of review articles. We searched the websites of relevant organizations, including the World Health Organization and the International Planned Parenthood Federation, for citations. We wrote to the corresponding author of each
87
trial identified to solicit other published or unpublished trials that we may have missed. Finally, we wrote to organizations in the United States, United Kingdom, Canada, Australia and New Zealand that teach these methods. Our search strategy for CENTRAL included the search term bcontraception.Q Our search strategy for MEDLINE (PubMed) used the following approach: [(natural family planning OR NFP OR periodic abstinence OR rhythm method* OR calendar method* or symptothermal method*) AND contracept*] AND (clinical trial OR clinical trials OR randomized controlled trial* OR clinical trial[pt] OR randomized controlled trial[pt]). Our search strategy for POPLINE included the following: (natural family planning/nfp/rhythm method calendar/symptothermal method/bperiodic abstinenceQ/bcalendar methodQ/ cervical mucus method/sexual abstinence) and (clinical trial*/controlled clinical trial*/randomized controlled trial*). We also searched POPLINE with the following terms: fertility awareness/fertility awareness method*/fertility knowledge/FAM/periodic abstinence/natural family planning/nfp/cervical secretion*/cervical mucus/Billings method/Creighton method/basal body temperature/double-check method/mucothermic/symptothermic/calendar method/ rhythm method/simple rule method/cycle beads/two-day method/personal hormone monitoring/abstinence and clinical trials (kw). Our search in EMBASE included the following: [periodic(w)abstinence or periodic(w)sexual(w)abstinence or natural family planning or rhythm(w) method or calendar(w) method or sympto(w)thermal(w)method]. Our search in LILACS used the following terms: natural family planning or nfp or periodic abstinence or calendar method or metodo del calendario or symptothermal method or rhythm method or metodo de ritmo or method do ritmo or birth intervals or intervalo entre nacimientos. We assessed all titles and abstracts found for inclusion. Because of high losses after randomization, we did not enter the data into RevMan or calculate odds ratios. We evaluated the methodological quality of the trials for potential biases by qualitatively assessing the study design, randomization method, allocation concealment, blinding, premature discontinuation rates and loss to follow-up rates. The trial conducted in Los Angeles by Kass-Annese et al. [11] compared use of the Today vaginal contraceptive sponge only during the presumed fertile period (as estimated by calendar calculations and basal body temperature) versus customary use of the sponge (with each act of coitus). The FAB method used the calendar rhythm method for judging the onset of the fertile days and the shift in basal body temperature for estimating its end; mucus was not evaluated. Those assigned to fertility awareness plus sponges were given sponges, a basal body temperature thermometer, a diary for recording events and instructions on use of the sponge. They returned within 2 weeks for instructions on fertility awareness. Thereafter, they returned for a maximum of three
88
D.A. Grimes et al. / Contraception 72 (2005) 85–90
follow-up training visits. Those assigned to the spongeonly received sponges, diaries and instructions on use of the sponge. The trial was designed to run for 12 months. bSince the drop-out rate was high, recruitment was ended and all women followed-up as long as possible, for up to 12 month. . .About one-quarter of the volunteers completed 12 months of follow-up. . ..Q [11]. Medina et al. [12] enrolled 566 couples in two cities in Colombia. The trial was designed to compare the efficacy of the ovulation method versus the symptothermal method. For the former, the atlas of Billings [13] was the reference, and, for the latter, it was the text of Thyma [14]. The symptothermal method included information about basal body temperature, cervical mucus changes, symptoms such as breast tenderness and length of the preceding six menstrual cycles. All couples enrolled entered a training period lasting 3 –5 months. Thereafter, couples were visited monthly by research staff who provided ongoing counseling about the assigned method. The observation period of the trial was intended to be 12 –24 months. Wade et al. [16] conducted a companion trial to the Colombian trial [12] in Southern California; the treatments and analytic methods were similar. An interim report [15] from the full trial [16] will not be considered separately. The methods compared were the ovulation method [13] and the symptothermal method [14]. Couples enrolled entered a training period that included both individual and group instructions. Randomized participants were allowed to drop out for any reason. Those who completed the training phase were invited to enter the follow-up phase; those who were noncompliant or who failed to learn the method were involuntarily dropped by the investigators. During the follow-up phase, noncompliant couples were encouraged to drop out. During the entire study, each female participant met monthly with her teacher and had a pregnancy test done. The planned length of follow-up was 12 months. The conduct and reporting of the three trials were poor. Because the smaller Los Angeles trial was abandoned, we will focus on the two larger trials [12,16]. Although each reported to be a randomized controlled trial, no information was provided concerning the method used to generate the random sequence. Similarly, neither mentioned allocation concealment. Sample size calculations were not provided. Blinding was apparently not used for either trial, but this would be unlikely to bias determination of the principal outcome (pregnancy). A more serious violation of research methods was the endorsement and, indeed, encouragement of discontinuation after randomization. Investigators in both the Colombia and the larger Los Angeles trials unilaterally excluded participants after randomization if they were having difficulty during the training phase. Those noncompliant with the method in the main follow-up phase were encouraged to discontinue. In contrast, the investigators should have attempted to keep all randomized participants in the trial. Loss of participants after randomization introduces bias
[17,18], and the losses here were large. In general, if more than 20% of participants are lost after randomization, trial validity is suspect [17]. 3. Results The three trials revealed difficult recruitment and high dropout rates. Because of unanticipated difficulty in recruiting participants and high discontinuation rates in the second Los Angeles trial [11], the trial was abandoned. The investigators cautioned that they could not assess the efficacy of using the vaginal sponge with fertility awareness versus using the sponge alone. However, four pregnancies occurred among the 24 women assigned to the FAB method plus the sponge in contrast to two among the 27 women in the spongeonly group. Similarly, pregnancy rates cannot be determined from the Colombia trial [12]. Most randomized participants dropped out before beginning the observation period: 149 of 279 couples (53%) assigned to the ovulation method discontinued during training in contrast to 176 of 287 assigned to the symptothermal method (61%). Eleven women assigned to the ovulation method and 32 assigned to the symptothermal method were excluded from analysis because of noncompliance during the training phase, and 1 more in each group was excluded during the active observation phase. Only a minority of participants entered the follow-up phase: 130 assigned to the ovulation method and 111 to the symptothermal method. Of these participants, 86 (31%) and 82 (30%), respectively, dropped out during the follow-up phase. With the training and follow-up phases combined, 72 women assigned to the ovulation method became pregnant as compared with 71 assigned to the alternative method. The corresponding number of participants who discontinued because of lack of interest or dissatisfaction with the method was 63 and 69, respectively. Because of inadequate methods, pregnancy rates cannot be determined from the Southern California trial [16]. The total number of participants randomized to the ovulation method and symptothermal method was 619 and 628, respectively. Of these participants, 265 (43%) and 238 (39%) dropped out during the training period. Another 314 women (163 and 151, respectively) were still in training at the end of the study and apparently contributed no outcome data. Further losses seem to have occurred since the report indicates that only 191 and 239 participants entered the formal follow-up phase. Stated alternatively, 428 participants (69%) assigned to the ovulation method and 389 (62%) assigned to the symptothermal method were dropped after randomization and before entering the planned observation period. During the main observation period, additional 123 (20%) and 100 (16%) participants, respectively, dropped out. Pregnancies were reported more often among those assigned to the ovulation method [16]. During the training phase, 92 pregnancies occurred in the group assigned to the ovulation method in contrast to 55 in the group assigned to
D.A. Grimes et al. / Contraception 72 (2005) 85–90
the symptothermal method. During the follow-up phase, the number of pregnancies was 42 and 20, respectively. The total number of reported pregnancies was 134 and 75, respectively. Discontinuations because of lack of interest or dissatisfaction with the method were similar for the two groups. A total of 85 and 90 participants, respectively, discontinued for these reasons. 4. Discussion The comparative efficacy of FAB methods remains unresolved. The conduct and reporting of randomized controlled trials of FAB methods were of such poor quality that valid pregnancy rates cannot be determined [18]. One small trial [11] was abandoned, and the other trials [12,16] lost most participants after randomization. No large difference in efficacy was likely in the Colombia trial [12], although the symptothermal method may have been more effective than the ovulation method in the Los Angeles trial [16]. A re-analysis of the World Health Organization case series report of the ovulation method [19] by Trussell [20] found that 23% of women became accidentally pregnant in the first year. Two trial reports addressed recruitment problems. In the larger Los Angeles trial, the investigators noted that recruitment bdifficulties were quite unexpected since this study featured the most promising methods of NFP [natural family planning], had the wholehearted support of the Catholic hierarchy in the Los Angeles area, received wide media coverage, and utilized two public relations firms to aid recruitmentQ [16]. The smaller Los Angeles trial was stopped because of poor recruitment. In it, the problems with the FAB method were boverwhelmingly related to the inconvenience of using BBT [basal body temperature]Q [11]. Continuation rates were low. Despite enrolling motivated volunteers, having thorough instruction and providing ongoing education and counseling of participants, most couples discontinued their assigned method before entering the planned observation period [12,16]. These techniques can be difficult for many couples since sexual abstinence for 7–17 days of the cycle, excluding menses, can be required [7,8,21]. An international study over 13 menstrual cycles reported that 5–62% of couples experienced difficulties with the mandated abstinence [22]. Focus groups conducted by the World Health Organization in developing countries revealed that couples commonly were sexually active during the presumed fertile days. Focus group respondents reported that drawbacks to FAB methods included trouble learning the method, difficulty using it and the challenge of abstinence. On the other hand, respondents liked the absence of side effects, perceived greater safety than with hormonal or intrauterine methods and acknowledged low cost [23]. Properly done randomized controlled trials are required to determine the comparative effectiveness of contraceptive methods [20,24,25]. Such evidence is not available for FAB methods.
89
The comparative efficacy of FAB methods for contraception remains unknown. However, with the ovulation and symptothermal methods, pregnancies appear to be common; method continuation rates are low.
Acknowledgments A version of this review has been published in the Cochrane Library (FORMAT: Grimes DA, Gallo MF, Grigorieva V, Nanda K, Schulz KF. Fertility awarenessbased methods for contraception. The Cochrane Database of Systematic Reviews 2004, Issue 4. Art. No.: CD004860.pub2. DOI: 10.1002/14651858.CD004860.pub2.) Cochrane reviews are regularly updated to include new research and in response to comments and criticism from readers. If you wish to comment on this review, please contact Dr. David Grimes at Family Health International (FHI), P.O. Box 13950, Research Triangle Park, NC 27709, USA, or email him at [email protected]. The results of a Cochrane review can be interpreted differently, depending on a person’s perspective and circumstance. Please consider the conclusions presented carefully. They are the opinions of the review authors and are not necessarily shared by the Cochrane Collaboration. Carol Manion of FHI assisted with the literature searches. Support for this study was provided by the FHI with funds from the U.S. Agency for International Development and the National Institute of Child Health and Human Development although the views expressed in this article do not necessarily reflect those of the FHI or the agencies funding the research.
References [1] World Health Organization. Medical eligibility criteria for contraceptive use. 3rd ed. http://www.who.int/reproductive-health/publications/ MEC_3/index.htm [accessed February 11, 2005]. [2] International Planned Parenthood Federation. Statement on periodic abstinence for family planning. London (England)7 International Planned Parenthood Federation; 1995. [3] Billings JJ. Natural family planning. Med J Aust 1978;2:436. [4] Pyper CMM, Knight J. Fertility awareness methods of family planning: the physiological background, methodology and effectiveness of fertility awareness methods. J Fam Plann Reprod Health Care 2001;27:103 – 10. [5] Bonnar J, Flynn A, Freundl G, Kirkman R, Royston R, Snowden R. Personal hormone monitoring for contraception. Br J Fam Plann 1999; 24:128 – 34. [6] Freundl G, Godehardt E, Kern PA, Frank-Herrmann P, Koubenec HJ, Gnoth Ch. Estimated maximum failure rates of cycle monitors using daily conception probabilities in the menstrual cycle. Hum Reprod 2003;18:2628 – 33. [7] Hughes J. Natural family planning: very risky business. Conscience 1988;9:7 – 12. [8] Neumann GH. How natural is bnatural family planningQ? Critical remarks from the biological behavior viewpoint. in German. Gynakologe 1989;22:1 – 6. [9] Gray RH, Kambic RT. Epidemiological studies of natural family planning. Hum Reprod 1988;3:693 – 8.
90
D.A. Grimes et al. / Contraception 72 (2005) 85–90
[10] Population Reference Bureau. Family planning worldwide. 2002 Data sheet. http://www.prb.org/Template.cfm?Section = PRB&template = / Content/ContentGroups/Datasheets/Family_Planning_Worldwide_ 2002_Data_Sheet.htm [accessed February 11, 2005]. [11] Kass-Annese B, Kennedy KI, Forrest K, Danzer H, Reading A, Hughes H. A study of the vaginal contraceptive sponge used with and without the fertility awareness method. Contraception 1989;40: 701 – 14. [12] Medina JE, Cifuentes A, Abernathy JR, Spieler JM, Wade ME. Comparative evaluation of two methods of natural family planning in Columbia. Am J Obstet Gynecol 1980;138:1142 – 7. [13] Billings J. Natural family planning: the ovulation method. 2nd American edition. Collegeville (Minn)7 The Liturgical Press; 1973. [14] Thyma P. Fertile and infertile days in marriage life. Author’s edition. Fall River (Mass)7 Raritan Printing Company; 1973. [15] Wade ME, McCarthy P, Abernathy JR, et al. Randomized prospective study of the use-effectiveness of two methods of natural family planning: an interim report. Am J Obstet Gynecol 1979;134: 628 – 31. [16] Wade ME, McCarthy P, Braunstein GD, et al. A randomized prospective study of the use-effectiveness of two methods of natural family planning. Am J Obstet Gynecol 1981;141:368 – 76.
[17] Schulz KF, Grimes DA. Sample size slippages in randomised trials: exclusions and the lost and wayward. Lancet 2002;359:781 – 5. [18] Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallelgroup randomised trials. Lancet 2001;357:1191 – 4. [19] World Health Organization. A prospective multicentre trial of the ovulation method of natural family planning: II. The effectiveness phase. Fertil Steril 1981;36:591 – 8. [20] Trussell J. Methodological pitfalls in the analysis of contraceptive failure. Stat Med 1991;10:201 – 20. [21] Ryder B, Campbell H. Natural family planning in the 1990s. Lancet 1995;346:233 – 4. [22] World Health Organization. A prospective multicenter trial of the ovulation method of natural family planning: V. Psychosexual aspects. Fertil Steril 1987;47:765 – 72. [23] d’Arcangues C, Kennedy KI. Periodic abstinence and calendar method use in Hungary, Peru, the Philippines, and Sri Lanka. Contraception 2001;64:209 – 15. [24] Trussell J, Grummer-Strawn L. Further analysis of contraceptive failure of the ovulation method. Am J Obstet Gynecol 1991;165:2054 – 9. [25] Glasziou P, Vandenbroucke JP, Chalmers I. Assessing the quality of research. BMJ 2004;328:39 – 41.