1 Center for Reproductive Medicine, Department of Obstetrics and Gynecology (H4-205) and 2 Department of Clinical Epidemiology and Biostatistics (H4-205), Academic Medical Center, University of Amsterdam, PO Box 22700, 1100 DE Amsterdam, The Netherlands
3 To whom correspondence should be addressed. Email: m.vanwely{at}amc.uva.nl
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: clomiphene citrate/ovulation induction/PCOS/pregnancy/rFSH
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Clomiphene citrate is still the first line treatment in women with PCOS, yet a significant proportion of women do not respond to treatment with clomiphene citrate. In these clomiphene citrate-resistant women, laparoscopic electrocautery or ovulation induction with gonadotrophins are indicated.
For ovulation induction with gonadotrophins, recombinant FSH (rFSH) is widely used. Pregnancy rates between 35 and 70% after ovulation induction with rFSH have been reported in the literature (Hull, 1987; White et al., 1996
; Christin-Maitre and Hugues, 2003
; Van Wely et al., 2003
). Ovulation induction with FSH is, especially in PCOS, also associated with a high risk of multiple pregnancies caused by multiple follicular development (Hull, 1987
; Van Wely et al., 2003
). For this reason, adjusted dose regimens and intense monitoring of the ovarian response to prevent multiple follicular development have been implemented.
An alternative treatment option for clomiphene citrate-resistant women is laparoscopic electrocauterization. If not successful, electrocautery can be followed by clomiphene citrate, as some clomiphene citrate-resistant women will become responsive after the cauterization. Multiple pregnancies can largely be prevented by treating women with electrocautery and clomiphene citrate, and, if not successful, by subsequent rFSH treatment (Bayram et al., 2004; Farquhar et al., 2004
). Laparoscopic cauterization followed by clomiphene citrate has been put forward as the treatment of first choice in women with clomiphene citrate-resistant PCOS (Dutch Health Council guideline, 2003
; National Institute of Clinical Excellence (NICE) of the British National Health Service, 2004
). A limitation of the laparoscopic procedure, however, is that it is a surgical procedure under general anaesthesia for which patients and/or gynaecologists might not opt. The counselling of couples would benefit from knowledge about the chances of success with rFSH in specific patient groups.
Therefore, we performed a study to establish whether clinical, ultrasonographic and endocrine parameters during initial screening of clomiphene citrate-resistant PCOS patients can predict treatment success after ovulation induction with rFSH.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Transvaginal ultrasonographic evaluation included assessment of the ovarian volume. Ovarian volume was determined by the prolate ellipsoid formula, volume = 4/3abc, with a, b and c representing the mean radius of the length, width and thickness of the ovaries. The mean radius was determined by taking the mean diameter of length, width and thickness of both ovaries, divided by 2.
Endocrine evaluation included serum assays for FSH, LH, estradiol, testosterone, androstenedione, sex hormone-binding globulin (SHBG), fasting insulin, glucose and leptin. The free androgen index (FAI) was calculated as total testosterone x100/SHBG. Blood samples were centrifuged within 2 h after withdrawal and stored at 20°C until assayed. All hormonal assays were performed at one laboratory (Endocrinology Department, Academic Medical Center Amsterdam) on cryopreserved blood.
All male partners had two semen analyses after a minimal sexual abstinence of 2 days and analysis of the semen was performed within 1 h of ejaculation. After liquefaction, the volume, concentration and motility were determined.
rFSH (follitropin alpha, Gonal-F; Serono Benelux BV, The Hague, The Netherlands), was administered in a chronic low-dose step-up protocol. Treatment started on cycle day (CD) 3 of a progesterone-induced bleeding with daily 75 IU s.c. injections. Follicle development was assessed by transvaginal ultrasonography at weekly intervals or more frequently if indicated by follicle growth. If the diameter of the follicles remained <10 mm, the dose was increased with half an ampoule on CDs 16 and 23. If no follicle development was seen on CD 30, i.e. no follicle with a diameter of >10 mm, the cycle was cancelled because of poor response. A cycle was also cancelled if more than six follicles were present with a diameter of 14 mm or more than three follicles were present with a diameter of
16 mm in order to prevent hyperstimulation or multiple pregnancy. A dose of 10 000 IU of HCG (Pregnyl, NV Organon, Oss, The Netherlands) was administered by i.m. or s.c. injection if one follicle >18 mm in diameter or no more than two follicles >15 mm diameter were present.
Women completed the study when they achieved an ongoing pregnancy within 12 months.
Data analysis
Ongoing pregnancy, defined as a vital pregnancy at 12 weeks of gestation, was the primary end-point for this study. Associations between clinical, ultrasonographic and endocrine parameters and response to rFSH treatment were evaluated using multivariable logistic regression analysis, adjusting for treatment cycle number. First, correlations between parameters were studied to prevent the selection of highly correlated parameters in the logistic model. We subsequently checked the linearity of the association between continuous variables and ongoing pregnancy, using visual inspection and spline functions. As the continuous variables approached linearity, it was not necessary to transform these variables. For both dichotomous and continuous variables, odds ratios (ORs) and 95% confidence intervals (CIs), as well as P-values were calculated.
Subsequently, multivariable logistic regression analysis was used to construct a prediction model for the occurrence of ongoing pregnancy. The initial model included clinical data only. Subsequently, ultrasonographic and endocrine variables were added using a stepwise procedure. To minimize erroneous exclusion of some factors that are of prognostic relevance, a significance level of P0.2 was used for selection in our multivariable analysis (Steyerberg et al., 1999
; Mol et al., 2000
).
To adjust for overfitting, an internal validation was performed with bootstrapping, and a shrinkage factor was calculated. Bootstrapping is a technique to create comparable virtual populations. We bootstrapped 200 times. In each of the 200 new data sets, the same multivariable logistic regression was assessed.
To express the discriminative performance of the final logistic model, the area under its receiver operating characteristic (ROC) curve was calculated.
To measure the agreement between predicted and observed outcomes, we studied the reliability of the model. The predicted fraction and the observed fraction of ongoing pregnancies were compared in five groups of women, by plotting the respective observed proportions of women with an ongoing pregnancy against the predicted proportions, as calculated from the model. The five groups were constructed by using the quintiles of the model-based calculated probabilities of success. We evaluated the goodness-of-fit of the model with the Hosmer and Lemeshow test statistic.
Data were analysed using SPSS 11.0.1 (SPSS Inc., Chicago, IL) and S-PLUS 2000 (MathSoft Inc.).
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
When including clinical and ultrasonographic parameters in the multivariable analysis, only cycle history and duration of infertility were found to be significantly associated with ongoing pregnancy rate. Women with oligomenorrhoea had a higher probability of attaining an ongoing pregnancy, while the chances of becoming pregnant decreased with a longer duration of infertility [area under the curve (AUC) 0.64, 95% CI 0.560.71].
When considering the additional value of endocrine parameters to the clinical and ultrasonographic parameters, cycle history, duration of infertility and FAI were identified (Table II). Internal validation by bootstrapping showed a 6% (95% CI 0.0490.070) overfitting of the model. Therefore, a shrinkage of 6% was performed to reduce the extremeness of predictions, and to make predictions more reliable. The multivariable-adjusted ORs after shrinkage were 2.34 for cycle history (95% CI 1.115.38), 0.87 for duration of infertility (95% CI 0.731.04) and 0.93 for FAI (95% CI 0.890.98).
For the FAI, the largest difference was found at a cut-off level of 9, based on an ROC curve analysis.
The reliability of the model was tested by comparing model-based cumulative probabilities with the observed cumulative proportions of women achieving an ongoing pregnancy (Figure 2). The group with the lowest probabilities of achieving an ongoing pregnancy had a mean cumulative probability of <5% and could be easily distinguished from women with a cumulative probability of >25%. The group of 14 women with the lowest cumulative probability of achieving an ongoing pregnancy included those who were amenorrhoeic, had an infertility duration >3 years and an FAI >9. In contrast, the group of 34 women that had a mean cumulative probability of achieving an ongoing pregnancy of >25% included those who were oligomenorrhoeic (only one woman had amenorrhoea). Furthermore, all women in this high cumulative probability group had an infertility duration of 2 years and a FAI of <9.
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In this study, we have used ongoing pregnancy as the measure of success. Not all ongoing pregnancies resulted in a delivery. Using live birth as an outcome measure is likely to be more meaningful in counselling patients, but it would have reduced the power of this study.
To our knowledge, this is the first study on prognostic factors for success of rFSH treatment in a chronic low-dose step-up scheme to use a multivariable prediction model based on structured, complete and carefully monitored patient-based data collection in clomiphene citrate-resistant women with PCOS. The strength of the study is that we included only women with clomiphene citrate-resistant PCOS, without other causes of infertility including male partner subfertility. Furthermore, we considered all parameters known to be associated with PCOS, including female age, type of infertility, age at menarche, cycle history, BMI, size of the ovary, LH/FSH ratio, FAI, dehydroepiandrosterone sulphate, androstenedione, glucose, insulin and leptin. A potential weakness of our study is that it is based on 85 women only; therefore, additional studies will be needed to confirm out results.
An alternative analysis used to predict the chance of becoming pregnant over time is with Cox proportional hazard regression. Logistic regression, as applied in this study, will give results that are identical to a Cox regression using cycles as the unit of time when the conditional probability of success remains constant over cycles. Logistic regression has the advantage over Cox regression that it provides both a baseline probability and ORs rather than hazard ratios, making it easier to estimate the probability of becoming pregnant.
One other study examined variables prognostic for success of rFSH treatment in women with normogonadotrophic anovulatory infertility (Mulders et al., 2003). Most of these women probably had PCOS. There are four major differences between our study and the study of Mulders et al. First, we used the chronic low-dose step-up regimen while the other study used a step-down regimen. Secondly, we included only clomiphene citrate-resistant women while the other study also included women that had failed to conceive in six ovulatory cycles. Thirdly, we induced ovulation with rFSH only while the other study used urinary FSH or rFSH. Finally, our population of women was more hyperandrogenaemic, i.e. women had higher LH, testosterone and FAI levels than the women in the other study [mean (SD) LH of 11 (5.1) versus 8.0 (4.6) and FAI 13 (10.5) versus 7.8 (9.2)].
In contrast to the study of Mulders et al. (2003), we did not find a negative effect of age on the predictive model. In the study of Mulders et al., the association between age and ongoing pregnancy was relatively weak (P=0.1). Therefore, the strength of the evidence that age is a clinically relevant predictor for ongoing pregnancy is rather limited. Just like in our study, Mulders et al. found that the chances of achieving an ongoing pregnancy decreased with increasing testosterone level. However, we found the FAI to be a better predictor than testosterone alone. The FAI level expresses the available bioactive testosterone serum level. There is evidence that local high androgen concentrations have a detrimental effect on follicle growth and oocyte quality. In atretic follicles, high androgen/estrogen levels have been found in mice (Gore-Langton and Armstrong, 1988
). Furthermore, it has been observed that the follicular steroid environment is correlated with oocyte quality in humans (Teissier et al., 1999
). In a recent study in mouse follicle culture under a constant gonadotrophin tonus, local androgen increase was not found to be atretogenic but did adversely affect the fertilization of in vitro grown oocytes (Hu et al., 2002
). Unfortunately, we do not know whether the FAI levels in blood serum reflect the levels in the follicular environment.
In our study, 94% of the women ovulated with rFSH treatment. As follicle growth was observed in almost all women, it is unlikely that FAI serum levels are associated with impaired follicle growth. However, a detrimental effect of high androgen levels expressed as a negative association between FAI serum levels and oocyte quality or fertilization rate is possible. Such an association may explain why the FAI level was found to be a prognostic parameter for pregnancy.
In our study group, the FAI was positively correlated with several other clinical and endocrine parameters. The strongest correlation was found for androstenedione and the WHR, implying that women with a high FAI were likely to have a high androstenedione level and a high WHR. Furthermore, we found a modest correlation beween FAI and leptin independent of BMI and WHR. Previous studies on leptin in women with PCOS did not find such an association (Mantzoros et al., 2000; Spritzer et al., 2001
). The women in those studies, however, were not clomiphene citrate resistant.
Our model allowed us to distinguish between women with a probability of <5% of achieving an ongoing pregnancy and those with a probability of >25%. The group of women with a poor prognosis consisted of those who had amenorrhoea, an infertility duration >3 years and an FAI >9. In contrast, most women with a probability of >25% of achieving an ongoing pregnancy had oligomenorrhoea, an infertility duration of <2 years and an FAI <9. After external validation on data from women with clomiphene citrate-resistant PCOS treated with rFSH in a chronic low-dose step-up regimen, our model can be used in the counselling of women with clomiphene citrate-resistant PCOS.
![]() |
Acknowledgements |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Bayram N, Van Wely M, Kaaijk EM, Bossuyt PM and Van der Veen F (2004) Using an electrocautery strategy or recombinant follicle stimulating hormone to induce ovulation in polycystic ovary syndrome: randomised controlled trial. Br Med J 328, 192195.
Christin-Maitre S and Hugues JN (2003) A comparative randomized multicentric study comparing the step-up versus step-down protocol in polycystic ovary syndrome. Hum Reprod 18, 16261631.
Diamanti-Kandaralis E, Kouli CR, Bergiele AT, Filandra FA, Tsiannateli TC and Spina GG (1999) A survey of the polycystic ovary syndrome in the Greek island of Lesbos: hormonal and metabolic profile. J Clin Endocrinol Metab 84, 40064011.
Dutch Health Council guideline. http://www.cvz.nl/resources/rSBA97-007_tcm13-2650.pdf.
Farquhar C, Vandekerckhove P and Lilford R (2004) Laparoscopic drilling by diathermy or laser for ovulation induction in anovulatory polycystic ovary syndrome (Cochrane Review). In The Cochrane Library, Issue 1. John Wiley & Sons, Chichester, UK.
Gore-Langton RE and Armstrong DT (1988) Follicular steroidogeneis and its control. In Knobil E and Neill J (eds) The Physiology of Reproduction. Raven Press, New York, pp. 331385.
Hu Y, Cortvrindt R and Smitz J (2002) Effects of aromatase inhibition on in vitro follicle and oocyte development analyzed by early preantral mouse follicle culture. Mol Reprod Dev 61, 549559.[CrossRef][ISI][Medline]
Hull MG (1987) Epidemiology of infertility and polycystic ovarian disease: endocrinological and demographic studies. Gynecol Endocrinol 1, 235245.[Medline]
Knochenhauer ES, Key TJ, Kahsar-Miller M, Waggoner W, Boots LR and Azziz (1998) Prevalence of the polycystic ovary syndrome in unselected black and white women of the southeastern United States: a prospective study. J Clin Endocrinol Metab 83, 30783082.
Mantzoros CS, Cramer DW, Liberman RF and Barbieri RL (2000) Predictive value of serum and follicular fluid leptin concentrations during assisted reproductive cycles in normal women and in women with the polycystic ovarian syndrome. Hum Reprod 15, 539544.
Mol BW, Van Wely M and Steyerberg EW (2000) Using prognostic models in clinical infertility. Hum Fertil 3, 199202.
Mulders AGM, Eijkemans MJ, Imani B and Fauser BC (2003) Prediction of chances for success or complications in gonadotrophin ovulation induction outcome in normogonadotrophic anovulatory infertility. Reprod Biomed Online 7, 170178.[Medline]
National Institute of Clinical Excellence (NICE) of the British National Health Service. http://www.nice.org.uk/pdf/CG011niceguideline.pdf. Guideline 1.6.3.1.
Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group (2004) Revised 2003 consensus on diagniostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 19, 17.[CrossRef][ISI][Medline]
Spritzer PM, Poy M, Wiltgen D, Mylius LS and Capp E (2001) Leptin concentrations in hirsute women with polycystic ovary syndrome or idiopathic hirsutism: influence on LH and relationship with hormonal, metabolic, and anthropometric measurements. Hum Reprod 16, 13401346.
Steyerberg EW, Eijkemans MJ and Habbema JD (1999) Stepwise selection in small data sets: a simulation study of bias in logistic regression analysis. J Clin Epidemiol 52, 935942.[CrossRef][ISI][Medline]
Teissier MP, Chable H, Paulhac S and Aubard Y (1999) Recombinant human follicle stimulating hormone versus human menopausal gonadotrophin induction: effects in mature follicle endocrinology. Hum Reprod 14, 22362241.
Van Wely M, Bayram N and Van der Veen F (2003) Recombinant FSH in alternative doses or versus urinary gonadotrophins for ovulation induction in subfertility associated with polycystic ovary syndrome: a systematic review based on a Cochrane review. Hum Reprod 18, 11431149.
White DM, Polson DW, Kiddy D, Sagle P, Watson H, Gilling-Smith C, Hamilton-Fairley D and Franks S (1996) Induction of ovulation with low-dose gonadotropins in polycystic ovary syndrome: an analysis of 109 pregnancies in 225 women. J Clin Endocrinol Metab 81, 38213824.[CrossRef][ISI][Medline]
Submitted on May 21, 2004; resubmitted on November 8, 2004; accepted on February 28, 2005.
|