Comparison of tamoxifen and clomiphene citrate for ovulation induction: a meta-analysis

Anne Z. Steiner1,3, Mishka Terplan2 and Richard J. Paulson1

1 Department of Obstetrics and Gynecology, University of Southern California, Los Angeles, California and 2 Center for Women's Health Research, University of North Carolina, Chapel Hill, North Carolina, USA

3 To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Women's and Children's Hospital, 1240 North Mission Road, Room 8K9, Los Angeles, CA 90033, USA. Email: asteiner{at}usc.edu


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Both selective estrogen receptor modulators, tamoxifen and clomiphene have been used for ovulation induction for patients with anovulatory infertility. This meta-analysis sought to compare the effectiveness of tamoxifen to clomiphene for the induction of ovulation and achievement of pregnancy. METHODS: We searched MEDLINE, BIOSIS, PreMEDLINE, CINAHL, International Pharmaceutical Abstracts, DDSR, ACP Journal Club, DARE and CCTR, along with reference lists and national experts. Inclusion criteria were prospective clinical trials, which compared tamoxifen and clomiphene for ovulation induction in infertile couples with isolated anovulatory infertility. Main outcome measures were ovulation rate and clinical pregnancy rate. Pooled odds ratios were obtained using random effects meta-analysis. RESULTS: Four trials were included. After pooling all the trials, the use of tamoxifen or clomiphene citrate resulted in similar ovulation rates [odds ratio (OR) 0.755, 95% confidence interval (CI) 0.513–1.111]. There was no benefit of tamoxifen over clomiphene citrate in achievement of pregnancy per cycle (OR 1.056, 95% CI 0.583–1.912) or per ovulatory cycle (OR 1.162, 95% CI 0.632–2.134). CONCLUSIONS: Clomiphene citrate and tamoxifen are equally effective in inducing ovulation. Although data regarding pregnancy rates and outcome are limited, there does not appear to be a significant benefit of one medication over the other.

Key words: clomiphene citrate/ovulation induction/pregnancy/tamoxifen


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Non-steroidal selective estrogen receptor modulators (SERM), such as clomiphene citrate and tamoxifen, are commonly used to induce ovulation. In the clinical setting of anovulation, SERM are thought to act primarily by binding with estrogen receptors at the hypothalamus. This competitive inhibition results in a perceived drop in endogenous estrogen levels, eventually leading to increased gonadotrophin secretion and subsequent induction of ovulation.

Clomiphene citrate has been the first-line method of ovulation induction in couples with anovulatory infertility since its introduction in 1956 (Greenblatt et al., 1961Go; Wolf, 2000Go). Approximately 80% of women ovulate while using clomiphene (Gorlitsky et al., 1978Go); however, only 40% of women will achieve pregnancy (Gysler et al., 1982Go). Some authors have proposed that this discrepancy is due to the antiestrogenic effects of clomiphene on the uterus, cervix and vagina, resulting in a thin endometrial lining (Eden et al., 1989Go) and poor cervical mucus (Gysler et al., 1982Go).

Another non-steroidal SERM, tamoxifen, has also been used to induce ovulation. Although commonly used today as an adjuvant therapy in the treatment of breast cancer, its use as an ovulatory agent was first reported by Williamson and Ellis (1973)Go. Unlike clomiphene, tamoxifen acts as an agonist on the estrogen receptors of the vaginal mucosa and endometrium. Studies on the effects of tamoxifen on cervical mucus have been contradictory (Roumen et al., 1984Go; Acharya et al., 1993Go; Annapurna et al., 1997Go).

A randomized controlled trial found that tamoxifen was as effective as clomiphene in inducing ovulation. Despite a trend toward improved pregnancy rates with tamoxifen, the study was underpowered to confirm this finding (Boostanfar et al., 2001Go). The purpose of this meta-analysis was to review the literature comparing tamoxifen and clomiphene for ovulation induction and subsequent pregnancy and to obtain a higher-powered estimate of the difference in efficacy between tamoxifen and clomiphene.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Search
We searched MEDLINE between 1966 and Week 4, January 2005, and PreMEDLINE, CINAHL, International Pharmaceutical Abstracts, CDSR, ACP Journal club, DARE, CCTR, and BIOSIS through January 2005 with the assistance of a reference librarian. The following key words and subject headings were used when applicable, ‘tamoxifen, nolvadex, or zitazonium’ and ‘ovulation induction, ovarian stimulation, artificial induction, anovulation, pregnancy, or infertility’, limiting the search to humans and removing studies referencing breast neoplasms. In addition, expert informants provided articles from personal files. References from identified articles and review articles were also checked. English and foreign language articles were included. Abstracts were included only if additional information could be obtained from the primary authors.

Selection and data extraction
We included prospective clinical trials which compared tamoxifen and clomiphene for ovulation induction in infertile couples with isolated anovulatory infertility. Only studies which enrolled oligo-ovulatory or anovulatory women with no known tubal disease and partners with normal semen analysis were included. Primary outcomes included ovulation and pregnancy per cycle.

Three reviewers (A.S., M.T., A.B.) independently screened all titles and abstracts, when necessary. All articles selected by the authors for further review were independently evaluated by two reviewers (A.S., M.T.) using the predefined inclusion and exclusion criteria. After article selection was completed, both reviewers (A.S., M.T.) independently abstracted data from the articles on subject characteristics, interventions, outcomes and study design. Results were compared and differences resolved by consensus agreement. Three authors were successfully contacted via e-mail and additional information on methodology and outcomes was obtained in one case.

Quantitative data synthesis
For both outcomes, results for each trial were expressed as numbers of events per cycle of SERM. All cycles were included in cross-over and parallel trials. STATA statistical software (USA) was used to perform the meta-analysis. Heterogeneity between the results was examined using the Cochrane's Q statistic (assuming statistical significance at P<0.1). As summary statistics, pooled odds ratio (OR) and 95% confidence intervals (CI) were calculated using the, random effects method of DerSimonian and Laird, 1986Go. Publication bias was assessed using Begg's test for publication bias (Begg and Muzumdar, 1994Go) (assuming significant publication bias at P<0.1).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The initial searches identified 247 potentially relevant articles. After review of 248 titles and abstracts and 37 articles, we identified five studies that met our inclusion criteria (Figure 1). One study (Buvat et al., 1987Go) was excluded from the meta-analysis because the number of cycles in each treatment group could not be determined, and the primary author was unable to provide additional information. The four remaining trials were included in the meta-analysis. The analysis thus included a total of 273 women and a total of 921 cycles.



View larger version (24K):
[in this window]
[in a new window]
 
Figure 1. Trial flow diagram.

 
Information on trials
Table I displays the details of the five trials. A funnel plot showed no suggestion of asymmetry due to publication bias (Figure 2). However, larger trials are pictorially absent. The included trials were all published in English. All trials stated that randomization had occurred. Method of allocation was described in three of the five trials, although concealment of allocation was only described and adequate in one. All the trials were open-label. Power calculations were preformed in only one of the trials. One trial was published in abstract form only but complete information was obtained directly from the authors. The results of all meta-analyses are detailed in Table II.


View this table:
[in this window]
[in a new window]
 
Table I. Trials comparing tamoxifen and clomiphene

 


View larger version (8K):
[in this window]
[in a new window]
 
Figure 2. Begg funnel plot with pseudo 95% confidence limits and P-value for Begg's test for publication bias.

 

View this table:
[in this window]
[in a new window]
 
Table II. Summary data of outcomes

 
Ovulation induction
All four trials included in the meta-analysis used ovulation induction as a reported endpoint. There was no significant difference between the tamoxifen and the clomiphene groups in the odds of ovulation (OR 0.755, 95% CI 0.513–1.111). There was no significant statistical heterogeneity detected between the trials (P=0.141).

Pregnancy rate
Three studies evaluated pregnancy as an outcome for a total of 743 cycles of which 504 were ovulatory (Elstein and Fawcett, 1984Go; Vegetti et al., 1999Go; Boostanfar et al., 2001Go). There was no significant difference between the tamoxifen and control groups in the odds of pregnancy per cycle (OR 1.056, 95% CI 0.583–1.912) or per ovulatory cycle (OR 1.162, 95% CI 0.632–2.134). No statistical heterogeneity was detected (P=0.815).

Live-birth rate
Only one trial, Elstein and Fawcett (1984)Go, reported all pregnancy outcomes, including miscarriages and live births. In this one trial, there was no significant difference between the groups in the odds of live birth per cycle (OR 0.261, 95% CI 0.005–2.711).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
In approaching this review, we theorized that tamoxifen's estrogenic effects on the endometrium and cervical mucus would result in higher pregnancy rates. In this meta-analysis, both SERM were equally effective in inducing ovulation. The ovulation rates were high in both groups; however, pregnancy rates were much lower. Despite the theoretical benefits of tamoxifen, this meta-analysis failed to find a significant benefit of tamoxifen over clomiphene for inducing pregnancy. Although the odds of pregnancy were higher after ovulation with tamoxifen, this finding was not statistically significant.

This meta-analysis is limited by the quality of the included papers. Others have expressed concerns with the use of the cycle as a unit of analysis in subfertility trials (Vail and Gardener, 2003Go), cross-over design use in subfertility trials (Daya, 2003Go), and combining parallel and cross-over trials (Elbourne et al., 2002Go). However, these methods continue to be commonly used in the fertility literature (Elbourne et al., 2002Go; Hughes et al., 2004Go).

In general, this meta-analysis of randomized controlled trials provides a fairly precise estimate of the effect of SERM on ovulation induction. This analysis was powered to detect an 8% absolute difference in ovulation rates, assuming 80% power, {alpha}=0.05, and a 70% ovulation rate in the clomiphene group. The similarity in ovulation rates differs from the conclusions by Borenstein et al. (1989)Go who found in a retrospective study of 43 clomiphene-resistant patients that tamoxifen was a superior ovulatory agent. Other retrospective studies have concluded that the two agents act similarly (Gerhard and Runnebaum, 1979Go; Weseley and Melnick, 1987Go). However, Gerhard and Runnebaum (1979)Go concluded that ovulation rates with tamoxifen did tend to be higher than those with clomiphene in women with oligomenorrhoea.

This analysis failed to demonstrate any difference in ovulation rates between clomiphene and tamoxifen. However, the familiarity of clinicians with the dosing regimens and side-effects of clomiphene may make this SERM more favourable for clinical use. No studies have addressed the use of adjuvant medications, such as metformin or hCG, with tamoxifen. In addition, patients may have concerns about taking a medication commonly used as a chemotherapeutic agent for breast cancer. In contrast, tamoxifen may be more affordable and some studies have shown that tamoxifen is better tolerated with fewer side-effects (Gerhard and Runnebaum, 1979Go; Buvat et al., 1987Go; Borenstein et al., 1989Go), although none of the trials was powered or designed to detect differences in side-effects. Therefore, in women suffering intolerable side-effects from clomiphene, tamoxifen may be an option.

Despite the clinical heterogeneity between the studies, various dosing regimens and patient characteristics, there was no statistically significant heterogeneity. This implies that these findings can be generalized to a variety of dosing regimens and patient populations commonly seeking ovulation induction. Unfortunately the number of published studies and amount of patient information is too limited to permit subgroup analysis. Further studies evaluating the effects of either regimen for women with polycystic ovarian syndrome or obesity are warranted. In addition, this analysis does not evaluate the efficacy of either of these agents to induce super-ovulation. Further studies are needed to compare these two regimens for this clinical scenario.

Our findings on the relative effects of clomiphene and tamoxifen on pregnancy rates or outcome were inconclusive. Pregnancy rates were less frequently described in the literature, and the overall pregnancy rate was low. Unlike the pooled odds ratio for ovulation induction, these limitations resulted in a fairly imprecise estimate of the odds of pregnancy with tamoxifen versus clomiphene. Thus more cycles are needed to quantify this association.

This meta-analysis would imply however, that if there is a difference in pregnancy rates, the difference is probably not large and therefore may not be clinically significant. This analysis was powered to detect a 6% absolute difference in cycle pregnancy rates, assuming 80% power, {alpha}=0.05, and a 0.06% cycle pregnancy rate in the clomiphene group. Other retrospective studies have reached similar conclusions (Gerhard and Runnebaum, 1979Go; Ruiz-Velasco et al., 1979Go; Weseley and Melnick, 1987Go). However, Buvat et al. (1987)Go conducted a prospective randomized controlled trial and concluded that tamoxifen was superior in the face of luteal phase defects.

Only one study included live birth as an outcome. This one study with eight pregnancies found a higher miscarriage rate and lower live-birth rate among those women on tamoxifen. Similarly, Ruiz-Velasco et al. (1979)Go and Buvat et al. (1987)Go found that miscarriage rates among those women prescribed tamoxifen was ~35% compared to miscarriage rates between 11 and 14% in the clomiphene group. Elstein and Fawcett (1984)Go found a similar tendency toward higher miscarriage rates with tamoxifen in their prospective trial. However, in a retrospective study of 87 patients, Wu (1997)Go found that miscarriage rates were significantly lower for those women receiving tamoxifen for treatment of luteal phase defect.

In conclusion, there are no appreciable differences in ovulation or pregnancy rates after treatment with tamoxifen or clomiphene for isolated anovulatory infertility. Clinicians may choose to use the regimen with which they are most familiar for ovulation induction in women with anovulatory infertility. Future studies may help to elucidate clinical settings in which one agent may be superior to the other.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The authors thank Janet Nelson, our research librarian, and Aykut Bayrak, MD for assistance in screening abstracts.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Acharya U, Irvine DS, Hamilton MPR and Templeton AA (1993) The effect of three anti-oestrogen drugs on cervical mucus quality and in-vitro sperm-cervical mucus interaction in ovulatory women. Hum Reprod 8, 437–441.[Abstract]

Annapurna V, Dhaliwal LK and Gopalan S (1997) Effect of two anti-estrogens, clomiphene citrate and tamoxifen, on cervical mucus and sperm-cervical mucus interaction. Int J Fertil Women's Med 42, 215–218.[ISI][Medline]

Begg CB and Mazumdar M (1994) Operating characteristics of a rank correlation test for publication bias. Biometrics 50, 1088–1101.[ISI][Medline]

Boostanfar R, Jain JK, Mishell DR, Jr and Paulson RJ (2001) A prospective randomized trial comparing clomiphene citrate with tamoxifen citrate for ovulation induction. Fertil Steril 75, 1024–1026.[CrossRef][ISI][Medline]

Borenstein R, Shoham Z, Yemini M, Barash A, Fienstein M and Rozenman D (1989) Tamoxifen treatment in women with failure of clomiphene citrate therapy. Aust, N.Z. J Obstet Gynaecol 29, 173–175.[ISI][Medline]

Buvat J, Buvat-Herbaut M, Marcolin G and Ardaens-Boulier K (1987) Antiestrogens as treatment of female and male infertilities. Horm Res 28, 219–229.[ISI][Medline]

Daya S (2003) Pitfalls in the design and analysis of efficacy trials in subfertility. Hum Reprod 18, 1005–1009.[Free Full Text]

Dersimonian R and Laird NM (1986) Meta-analysis in clinical trials. Control Clin Trials 7, 177–188.[CrossRef][ISI][Medline]

Eden JA, Place J, Carter GD, Jones J, Alaghband-Zadeh J and Pawson ME (1989) The effect of clomiphene citrate on follicular phase increase in endometrial thickness and uterine volume. Fertil Steril 73, 187–190.

Elbourne DR, Altman DG, Higgins JPT, Curtin F, Worthington HV and Vail A (2002) Meta-analyses involving cross-over trials: methodological issues. Int J Epidemiol 31, 140–149.[Abstract/Free Full Text]

Elstein M and Fawcett GM (1984) Effects of the anti-oestrogens, clomiphene and tamoxifen, on the cervical factor in female infertility. Ciba Foundn Symp 109, 173–179.

Gerhard I and Runnebaum B (1979) Comparison between tamoxifen and clomiphene therapy in women with anovulation. Arch Gynecol 227, 279–288.[ISI][Medline]

Gorlitsky GA, Kase NG and Speroff L (1978) Ovulation and pregnancy rates with clomiphene citrate. Obstet Gynecol 51, 265–269.[Abstract]

Greenblatt RB, Barfield WE, Jungck EC and Ray AW (1961) Induction of ovulation with MRL/41, preliminary report. J Am Med Assoc 178, 101–104.[ISI][Medline]

Gysler M, March CM, Mishell DR, Jr and Bailey EJ (1982) A decade's experience with an individualized clomiphene treatment regimen including its effect on the postcoital test. Fertil Steril 37, 161–167.[ISI][Medline]

Hughes E, Collins J and Vanderkerckhove P (2004) Clomiphene citrate for unexplained subfertility in women (Cochrane Library). In The Cochrane Library, Volume 2.

Messinis IE and Nillius SJ (1982) Comparison between tamoxifen and clomiphene for induction of ovulation. Acta Ostet Gynecol Scand 61, 377–379.

Roumen FJ, Doesburg WH and Rolland R (1984) Treatment of infertile women with a deficient postcoital test with two antiestrogens: clomiphene and tamoxifen. Fertil Steril 41, 237–243.[ISI][Medline]

Ruiz-Velasco V, Rosas-Arceo J and Matute MM (1979) Chemical inducers of ovulation: comparative results. Int J Fertil 24, 61–64.[ISI][Medline]

Vail A and Gardener E (2003) Common statistical errors in the design and analysis of subfertility trials. Hum Reprod 18, 1000–1004.[Abstract/Free Full Text]

Vegetti W, Riccaboni A, Colombo M, Baroni E, Diaferia D, Ragni G and Crosignani PG (1999) Randomized study of induction of ovulation by two different molecules with antiestrogenic effects, in patients with chronic anovulation disorders. Fertil Steril 72 (Suppl), S234–S235.

Weseley AC and Melnick H (1987) Tamoxifen in clomiphene-resistant hypothalamic anovulation. Int J Fertil 32, 226–228.[ISI][Medline]

Williamson JG and Ellis JD (1973) The induction of ovulation by tamoxifen. J Obstet Gynaecol Br Cmnwlth 80, 844–847.

Wolf LJ (2000) Ovulation induction. Clin Obstet Gynecol 43, 902–915.[CrossRef][ISI][Medline]

Wu CH (1997) Less miscarriage in pregnancy following tamoxifen treatment of infertile patients with luteal phase dysfunction as compared to clomiphene treatment. Early Pregn 3, 301–305.

Submitted on January 12, 2005; resubmitted on February 10, 2005; accepted on February 11, 2005.





This Article
Abstract
Full Text (PDF )
All Versions of this Article:
20/6/1511    most recent
deh840v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Request Permissions
Google Scholar
Articles by Steiner, A. Z.
Articles by Paulson, R. J.
PubMed
PubMed Citation
Articles by Steiner, A. Z.
Articles by Paulson, R. J.