The accuracy of multivariate models predicting ovarian reserve and pregnancy after in vitro fertilization: a meta-analysis
1 Department of Obstetrics and Gynecology, Máxima Medical Center Veldhoven, De Run 4600, PO Box 7777, 5500 MB Veldhoven NL, The Netherlands 2 Division of Perinatology and Gynecology, Department of Reproductive Medicine, University Medical Center Utrecht, The Netherlands 3 Department of Obstetrics and Gynecology, Rijnstate Hopsital, Arnhem, The Netherlands 4 Department of Obstetrics and Gynecology, Academic Medical Centre, Amsterdam, The Netherlands
5 Correspondence address. Tel: +31-40-8888385; Fax: +31-40-8888387; E-mail: t.verhagen{at}mmc.nl
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BACKGROUND: To review the accuracy of multivariate models for the prediction of ovarian reserve and pregnancy in women undergoing IVF compared with the antral follicle count (AFC) as single test.
METHODS: We performed a computerized MEDLINE and EMBASE search to identify articles published on multivariate models for ovarian reserve testing in patients undergoing IVF. In order to be selected, articles had to contain data on the outcome of IVF in terms of either pregnancy and/or poor response and on the prediction of these events based on a multivariate model. For the selected studies, sensitivity and specificity of the test in the prediction of poor ovarian response and non-pregnancy were calculated. Overall performance was assessed by estimating a summary receiver operating characteristic (ROC) curve, which was compared with the ROC curve for the AFC as the current best single test.
RESULTS: We identified 11 studies reporting on the predictive capacity of multivariate models in ovarian reserve testing. All studies reported on the prediction of poor ovarian response, whereas none reported on the occurrence of pregnancy. The sensitivity for prediction of poor ovarian response varied between 39% and 97% and the specificity between 50% and 96%. Logistic regression analysis indicated that cohort studies provided a significantly better discriminative performance than case–control studies. As cohort studies are superior to case–control studies, further analysis was limited to the cohort studies. For the cohort studies, a summary ROC curve could be estimated, which had a shape similar to that previously made for the AFC.
CONCLUSIONS: The accuracy of multivariate models for the prediction of ovarian response in women undergoing IVF is similar to the accuracy of AFC. No data are available on the capacity of these models to predict pregnancy, let alone live birth. On the basis of these findings, the use of more than one single test for the assessment of ovarian reserve cannot currently be supported.
Key words: meta-analysis / IVF / ovarian reserve / multivariate models / AFC
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In recent decades, societal changes in family planning have caused a significant increase in the incidence of unwanted infertility due to female reproductive ageing (Weinstein et al., 1993
; Abma et al., 1997). Ageing of women leads to female fertility problems because of a reduction in primordial follicle pool and loss of oocyte quality (Te Velde and Pearson, 2002). An age-related effect on female fecundity has also been shown in reports on the results of IVF treatment (Templeton et al., 1996; Quban et al., 2005). In an IVF programme, ovarian ageing is characterized by decreased ovarian responsiveness to gonadotrophin administration and lowered pregnancy rates. Identification of women at increased risk for poor ovarian response prior to IVF could be useful, as this either could prevent unnecessary continuation of treatment (Klinkert et al., 2004) or result in adjustment of the gonadotrophin dose in order to maximize ovarian response (Popovic-Todorovic et al., 2003; Klinkert et al., 2005).
In a recent systematic review of tests predicting ovarian reserve and IVF outcome, we concluded that none of the individual ovarian reserve tests (ORT) can routinely be used for outcome prediction (Broekmans et al., 2006). ORT evaluated were early follicular phase blood values of FSH, estradiol, inhibin B and anti-Mullerian Hormone (AMH), antral follicle count (AFC), ovarian volume (OV) and ovarian blood flow. Furthermore, the clomiphene citrate challenge test (CCCT), the exogenous FSH ovarian reserve test and the gonadotrophin agonist stimulation test (GAST) were evaluated. The capacity of each of the evaluated tests to predict ovarian reserve was modest to poor.
When patient and doctor valuation of mismatches between test results and treatment outcome are taken into account, testing for ovarian reserve seems not useful for current IVF programmes (Mol et al., 2006). In view of these results, it is currently recommended to start IVF treatment without any prior testing.
Combining tests that predict the ovarian reserve might lead to a better estimation of the ovarian reserve capacity. If this were to be the case, testing for ovarian reserve prior to IVF might become more useful. Our group has previously reported on the performance of tests for ovarian reserve prior to IVF in a multivariate setting (Bancsi et al., 2002; Van Rooij et al., 2002; Hendriks et al., 2005b). These individual studies showed that the combination of tests leads to a better prediction than individual tests. In a previously published meta-analysis, AFC and basal FSH were compared as predictors of ovarian reserve and this meta-analysis showed a significantly better predictive performance of AFC (Hendriks et al., 2005b). An estimated summary ROC curve was constructed for AFC in that study. A preliminary assessment of the literature on the subject has been performed (Broekmans et al., 2006). In this article, we report on a systematic review and meta-analysis on the subject. The aim of this meta-analysis was to assess the accuracy of a multivariate approach of ovarian reserve testing. We compared the multivariate prediction models of ovarian reserve testing with a previously study on univariate prediction, where we concluded that AFC might be considered the test of first choice in the assessment of ovarian reserve testing prior to IVF (Hendriks et al., 2005b).
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Sources
ORT that were considered as potential part of a predictive model were basal FSH, estradiol, inhibin B and AMH, as well as sonographic markers such as the AFC, total OV and ovarian blood flow. Moreover, female age, being an indicator for ovarian reserve, was also considered as potential prognosticator. Challenge tests like the CCCT or GAST were also included in the search.
For every ORT, a computerized MEDLINE search was performed to identify articles on the subject published from 1978 until April 2006. Keywords used for the various searches were ‘in vitro fertilisation’or ‘in vitro fertilization’ or ‘assisted’ or ‘intracytoplasmic’ or ‘intracytoplasmatic’, in combination with ‘test-specific’ keywords as published elsewhere (Broekmans et al., 2006). To be selected for further reading abstracts that were recognized by the search had to report clearly on one or two outcomes of IVF or ICSI treatment: poor response and/or pregnancy.
Selected studies were carefully read and scored by two of the authors (D.J.H and F.J.M.B) and judged upon the presence of multivariate prediction models that enabled the construction of 2 x 2 tables to calculate test characteristics. The authors of studies containing multivariate models in which it was not possible to construct 2 x 2 tables were asked to supply us with data needed for the construction of 2 x 2 tables or otherwise these studies were excluded. In addition, cross-references of the selected studies were checked for other articles meeting the inclusion criteria. All articles, selected according to the previously mentioned search strategy, were checked using multivariate prediction models.
Two-by-two tables comparing results of the multivariate prediction models and the occurrence of poor ovarian response and/or pregnancy were constructed independently by two of the authors (D.J.H. and F.J.M.B.) and in the event of disagreement, the judgement of a third author (B.W.J.M) was decisive. For each study, we calculated the prevalence of poor ovarian response and/or pregnancy. There was no uniform criterion for the definition of poor ovarian response. In this analysis poor response encompassed cycle cancellation as well as insufficient follicular growth or oocyte yields according to the standards of each study. Also, data on clinical and ongoing pregnancies were not analysed separately.
The following characteristics of each study were registered: (1) sampling (consecutive versus other), (2) data collection (prospective versus retrospective), (3) study design (cohort study versus case-control study), (4) blinding (present or absent), (5) selection bias and (6) verification bias (Lijmer et al., 1999). The definition of poor response or pregnancy was documented as well as whether an included study reported on only one cycle per couple or on multiple cycles.
The analysis was conducted according to a methodology that has been described in detail elsewhere (Bancsi et al., 2003). First, we assessed the predictive performance of each model in detail. In brief, for each study, sensitivity and specificity were calculated from the published data. Subsequently, sensitivity–specificity points for each study were plotted in ROC space. Homogeneity of the studies was tested by means of the chi-square test. A summary point estimate of sensitivity and specificity was calculated if homogeneity could not be rejected. In case of heterogeneity, logistic regression was used to evaluate whether the study characteristics influenced the discriminative capacity of the models. A P-value of <0.05 was considered to indicate statistical significance. If one of the study characteristics was found to have a statistically significant impact on the performance of the model, further analysis was performed in subgroups of patients.
A Spearman correlation coefficient was calculated for the association between sensitivity and specificity to explore possible heterogeneity due to a shift in cut-off levels between the models. If there was a negative correlation between sensitivity and specificity as defined by a correlation coefficient of –0.5 or less, a summary ROC curve was estimated (Littenberg and Moses, 1993; Midgette et al., 1993; Moses et al., 1993), using a random-effects regression model (SAS Institute Inc., 1993).
To compare the predictive capacity of the multivariate models with that of AFC, the constructed summary ROC curves for the multivariate models and a previously constructed curve of AFC were tested for statistically significant differences using linear regression (Hendriks et al., 2005b).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Search strategy
A total of 11 studies reporting on the predictive capacity of several multivariate models were identified and considered suitable for data extraction and meta-analysis (Balasch et al., 1996; Ranieri et al., 1998; Creus et al., 2000; Fábregues et al., 2000; Bancsi et al., 2002; Van Rooij et al., 2002; Durmusoglu et al., 2004; Erdhem et al., 2004; Muttukrishna et al., 2004; Hendriks et al., 2005a; Penarrubia et al., 2005). All studies reported on poor ovarian response as outcome, but only one of the studies reported on pregnancy as outcome (Creus et al., 2000).
Characteristics of the included studies are listed in Table I. Sampling of patients was consecutive in eight studies and data collection was prospective in five studies. Seven studies were designed as cohort studies. None of the studies was blinded. Selection bias was present in 6 of 11 included studies, whereas verification bias was present in two studies. All studies reported on the outcome of one cycle per couple.
|
As with most studies on ORT, definitions of poor response varied considerably. Each of the 11 studies used a different model. All studies reported on models that were constructed by the authors themselves in a so-called internal validation, whereas studies validating an existing prediction model were not present. The number of variables combined in the different models varied from two to four. The prediction variables used were age, inhibin B, FSH, AFC, AMH, GAST, OV and the CCCT. In the models, FSH was used seven times, inhibin B six times, age five times, AFC four times, CCCT and AMH two times and GAST or OV each only one time. The combinations of prediction variables that were most frequent were FSH+inhibin B (Balasch et al., 1996
; Creus et al., 2000; Fábregues et al., 2000; Bancsi et al., 2002) and age+FSH (Balasch et al., 1996; Creus et al., 2000; Durmusoglu et al., 2004; Penarrubia et al., 2005). The sensitivities and specificities for the prediction of poor ovarian response, as calculated from each model, are summarized in Table II. The sensitivity varied between 39% and 97%, and the specificity between 50% and 96%. A plot of sensitivity–specificity points in a ROC space is shown in Fig. 1.
|
|
Logistic regression analysis indicated that cohort studies reported a significantly better discriminative performance than case–control studies, whereas none of the other study characteristics had a statistically significant impact on the reported performance of the different models. As a consequence, further analysis was performed separately in the cohort studies and case–control studies.
In the seven cohort studies, homogeneity was rejected for both sensitivity and specificity (both P-values <0.001), and therefore a summary point estimate was not given. The Spearman correlation coefficient between sensitivity and specificity was –0.77 (P-value 0.006). This indicated that an increase in sensitivity occurred when specificity dropped, or vice versa. Thus, the combination of sensitivity–specificity points could be thought of as originating from an ROC curve, in which variations in sensitivity–specificity were explained by different cut-off values. Therefore, we summarized the test-performance of the multivariate models by estimating a summary ROC curve (Fig. 1). This summary ROC curve had a shape almost similar to the summary ROC curve for the AFC as published elsewhere (Hendriks et al., 2005b).
In the four case–control studies, homogeneity was also rejected for both sensitivity and specificity (both P-values <0.001). The Spearman correlation coefficient for sensitivity and specificity was –0.50 (P-value 0.006). As cohort studies are methodologically superior to case–control studies, a separate analysis of the case–control studies was not performed.
Figure 1 also shows the summary ROC curves of the AFC that was constructed in a previous meta-analysis (Hendriks et al., 2005b). Comparison of this summary ROC curve of AFC with the summary ROC curve of the multivariate models as estimated from cohort studies did not indicate a better performance of multivariate models than AFC (P-value 0.45).
Only one study reported results of the prediction of non-pregnancy (Creus et al., 2000). The variables analysed were age, basal FSH and basal inhibin B. The likelihood of pregnancy was analysed and age had the highest area under curve in the ROC curve (0.72), which was better than basal FSH or basal inhibin B alone (both 0.60) or any combination of the three variables.
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The results of this meta-analysis show that multivariate models as a test for the occurrence of poor ovarian response after IVF have a similar potential compared with AFC alone. However, no data are available on the predictive capacity of multivariate models for pregnancy, let alone live birth. According to these findings, there is at present no place for multivariable models in the prediction of ovarian reserve prior to IVF. Future models, should take into account the AFC, as this test is at present the best univariate predictor of diminished ovarian reserve.
A concern in the meta-analysis of diagnostic studies is heterogeneity among the included studies with respect to the definitions used, study design and clinical characteristics of the population studied (Irwig et al., 1994; Deeks, 2001). We found a clear difference between cohort studies and case–control studies. Surprisingly, cohort studies reported a better performance than case–control studies. Other sources of heterogeneity might be caused by differences in patient population, different treatment protocols used and variability in the definition of poor ovarian response. Despite this heterogeneity, we were able to provide an estimate summary ROC curve, which we feel is at present the best available evidence on the subject. Especially our finding that multivariate models did not perform better than AFC alone is in our opinion of clinical importance.
This meta-analysis indicates that AFC performed equally to multivariate models in assessing quantitative ovarian reserve. However, we are aware of the fact that we used different populations in the comparison of both tests in the present study. This could hamper comparability of the performance of these tests. It would be possible to overcome this problem by limiting the analysis to studies reporting on the predictive capacity of both test approaches in the same group of women. Only three studies in this meta-analysis incorporated AFC in their models (Bancsi et al., 2002; Durmusoglu et al., 2004; Hendriks et al., 2005a). Durmusoglu et al. (2004) found an area under the ROC curve for basal AFC of 0.82 with an optimum cut-off for the prediction of poor response of six primordial follicles. Sensitivity and specificity for the AFC were 85% and 74%, respectively. This performance was not improved by adding female age together with the AFC in a predictive model. Bancsi et al. (2002) evaluated inhibin B, basal FSH and AFC. AFC as single variable was the best predictor of poor response with a sensitivity of 61% and specificity of 88%. The area under the ROC curve for AFC was 0.87, better than for inhibin B (0.77) and FSH (0.84). All the multivariate models constructed in this study had a better overall ability to discriminate between normal and poor response, with a maximum increase of the ROC-AUC to a value of 0.92 for a model containing AFC, basal FSH and inhibin B. With this model, a sensitivity of 0.75 and specificity of 0.95 was obtained (Table II). In the study from Hendriks et al., the best single variable predictor of poor response was basal FSH with an area under the ROC curve of 0.85 followed by AFC with the area under the ROC curve of 0.83. Combining these two variables yielded the best predictor model for ovarian reserve (AUC-ROC 0.89) (Hendriks et al., 2005b), but with a modest improvement of sensitivity and specificity. All this indicates that multifactor models may have some benefit over using the AFC as a single test, although the added clinical value has not been consistently shown.
Recent years have shown several studies on the role of AMH as univariate predictor of outcome after IVF. AMH is likely to represent the size of the primordial follicle pool as accurately as the AFC, as AMH serum levels are produced by small antral follicles and AFCs relate closely to primordial follicle numbers (Kevenaar et al., 2006). Also, antral follicle numbers measured on Day 3 are highly correlated to the serum levels of AMH (Fanchin et al., 2003). From several studies, it has been shown that AMH levels do not fluctuate across and between cycles making it a cycle independent marker (Hehenkamp et al., 2006; La Marca et al., 2006; Tsepelidis et al., 2007). In contrast, the AFC is more likely to be a marker that is more prone to observer bias and at least varies more clearly between cycles in patients (Fanchin et al., 2003; van Rooij et al., 2005). All this may allow AMH to become an easy and powerful tool for quantitative ovarian reserve testing. Several clinical studies suggest AMH to have the same accuracy compared with the AFC in predicting the poor responder after IVF, although pregnancy prediction continues to be troublesome (van Rooij et al., 2005; Elgindy et al., 2007; Kwee et al., 2007; La Marca et al., 2007; McIlveen et al., 2007). Meta-analysis of the accumulating studies on AMH may further exemplify its merits in this field compared with the AFC.
The performance of a prognostic model can be assessed in the sample or population in which it has been developed (apparent or internal validity) or in other populations, i.e. patients seen in another period or at another place (external validity or generalizability). All studies included in the present meta-analysis reported on models that were assessed in the sample or population in which they had been developed. External validation is needed to assess the true accuracy of the models that we identified (Mol et al., 2003). Such validation studies are lacking for ORT. However, most studies on prediction models in reproductive medicine have shown that the internal model often overestimates the true predictive capacity (Stolwijk et al., 1996).
Even if a test has good predictive accuracy, this does not mean that the test is good enough to have clinical implications. For assessment of the clinical applicability of a test, the relative weight of both false positive and false negative predictions should be considered. When patients are interviewed on issues of incorrect withholding IVF when compared with incorrect performing IVF, they value the first much worse than the second, thereby implicating that the present tests for ovarian reserve have insufficient accuracy to withhold IVF (Mol et al., 2006). As a consequence of this point of view, testing for ovarian reserve is not needed. Whether incorporation of the AFC in future models improves the predictive capacity should be answered in future studies.
Apart from the prediction of pregnancy or live birth after IVF, the prediction of ovarian reserve is also potentially important for individual adjustment of the dose of gonadotrophins prior to IVF. Patients with an expected poor response might benefit from a higher starting dose than 150 IU/day. A randomized trial on the subject showed that an individual dose regimen in a well-defined ‘standard’ patient population increased the proportion of appropriate ovarian responses and decreased the need for dose adjustments during controlled ovarian stimulation (Popovic-Todorovic et al., 2003). A higher ongoing pregnancy rate was observed in the individual dose group. In contrast, no beneficial effect was observed in a randomized dose study of predicted poor responders (Klinkert et al., 2005). This all implies that additional studies are needed to provide clear evidence for the beneficial effects of dose adaptation in poor responder patients, as many previous studies have failed to show such effect (Tarlatzis et al., 2006). In our opinion, such studies provide more information, when the randomization is limited to women in whom the prediction model indicates a starting dose different than the standard dose of 150 IU/day (Bossuyt et al., 2000). Furthermore, the end-point of such a study should be live birth after IVF and not only the ovarian response to stimulation (Farquhar, 2000).
In conclusion, this study shows that the performance of multivariate models in the prediction of poor ovarian response after IVF is comparable with that of AFC. Therefore, the AFC may be considered as the test of first choice in the assessment of diminished ovarian reserve. Future models for the prediction of poor response and pregnancy should incorporate the AFC.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Abma JC, Chandra A, Mosher WD, Peterson CS, Piccinino CJ. Fertility, family planning and women's health: new data from the 1995 national survey of family growth. Vital Health Stat (1997) 23:1–114.
Balasch J, Creus M, Fabregues F, Carmona F, Casamitjana R, Ascaso C, Vanrell JA. Inhibin, follicle-stimulating hormone, and age as predictors of ovarian response in in vitro fertilization cycles stimulated with gonadotropin-releasing hormone agonist-gonadotropin treatment. Am J Obstet Gynecol (1996) 175:1226–1230.[CrossRef][Web of Science][Medline]
Bancsi LF, Broekmans FJ, Eijkemans MJ, de Jong FH, Habbema JD, te Velde ER. Predictors of poor ovarian response in in vitro fertilization: a prospective study comparing basal markers of ovarian reserve. Fertil Steril (2002) 77:328–336.[CrossRef][Web of Science][Medline]
Bancsi LF, Broekmans FJ, Mol BW, Habbema JD, te Velde ER. Performance of basal follicle-stimulating hormone in the prediction of poor ovarian response and failure to become pregnant after in vitro fertilization: a meta-analysis. Fertil Steril (2003) 79:1091–1100.[CrossRef][Web of Science][Medline]
Bossuyt PM, Lijmer JG, Mol BW. Randomised comparisons of medical tests: sometimes invalid, not always efficient. Lancet (2000) 356:1844–1847.[CrossRef][Web of Science][Medline]
Broekmans FJ, Kwee J, Hendriks DJ, Mol BW, Lambalk CB. A systematic review of tests predicting ovarian reserve and IVF outcome. Hum Reprod Update (2006) 12:685–718.
Creus M, Penarrubia J, Fábregues F, Vidal E, Carmona F, Casamitjana R, Vanrell JA, Balasch J. Day 3 serum inhibin B and FSH and age as predictors of assisted reproduction treatment outcome. Hum Reprod (2000) 15:2341–2346.
Deeks JJ. Systematic reviews in health care: systematic reviews of evaluations of diagnostic and screening tests. BMJ (2001) 323:157–162.
Durmusoglu F, Elter K, Yoruk P, Erenus M. Combining day 7 follicle count with the basal antral follicle count improves the prediction of ovarian response. Fertil Steril (2004) 81:1073–1078.[CrossRef][Web of Science][Medline]
Elgindy EA, El-Haieg DO, El-Sebaey A. Anti-Müllerian hormone: correlation of early follicular, ovulatory and midluteal levels with ovarian response and cycle outcome in intracytoplasmic sperm injection patients. Fertil Steril (2007) [Epub ahead of print].
Erdhem M, Erdhem E, Gursoy R, Biberoglu K. Comparison of basal and clomiphene citrate induced FSH and inhibin B, ovarian volume and antral follicle counts as ovarian reserve tests and predictors of poor ovarian response in IVF. J Assist Reprod Genet (2004) 21:37–45.[CrossRef][Web of Science][Medline]
Fábregues F, Balasch J, Creus M, Carmona F, Puerto B, Quinto L, Casamitjana R, Vanrell JA. Ovarian reserve test with human menopausal gonadotropin as a predictor of in vitro fertilization outcome. J Assist Reprod Genet (2000) 17:13–19.[CrossRef][Web of Science][Medline]
Fanchin R, Schonäuer LM, Righini C, Frydman N, Frydman R, Taieb J. Serum anti-Müllerian hormone dynamics during controlled ovarian hyperstimulation. Hum Reprod. (2003) 18:328–332.
Farquhar CM. Extracts from the ‘clinical evidence’. Endometriosis. BMJ (2000) 321:1077–1078.
Hehenkamp WJ, Looman CW, Themmen AP, de Jong FH, Te Velde ER, Broekmans FJ. Anti-Müllerian hormone levels in the spontaneous menstrual cycle do not show substantial fluctuation. J Clin Endocrinol Metab (2006) 91:4057–4063.
Hendriks DJ, Broekmans FJ, Bancsi LF, de Jong FH, Looman CW, te Velde ER. Repeated clomiphene citrate challenge testing in the prediction of outcome in IVF: a comparison with basal markers for ovarian reserve. Hum Reprod (2005) a 20:163–169.
Hendriks DJ, Mol BW, Bancsi LF, Te Velde ER, Broekmans FJ. Antral follicle count in the prediction of poor ovarian response and pregnancy after in vitro fertilization: a meta-analysis and comparison with basal follicle-stimulating hormone level. Fertil Steril (2005) b 83:291–301.[CrossRef][Web of Science][Medline]
Irwig L, Tosteson AN, Gatsonis C, Lau J, Colditz G, Chalmers TC, Mosteller F. Guidelines for meta-analyses evaluating diagnostic tests. Ann Intern Med (1994) 120:667–676.
Kevenaar ME, Meerasahib MF, Kramer P, van de Lang-Born BM, de Jong FH, Groome NP, Themmen AP, Visser JA. Serum anti-Mullerian hormone levels reflect the size of the primordial follicle pool in mice. Endocrinology (2006) 147:3228–3234.
Klinkert ER, Broekmans FJ, Looman CW, Te Velde ER. A poor response in the first in vitro fertilization cycle is not necessarily related to a poor prognosis in subsequent cycles. Fertil Steril (2004) 81:1247–1253.[CrossRef][Web of Science][Medline]
Klinkert ER, Broekmans FJ, Looman CW, Habbema JD, te Velde ER. Expected poor responders on the basis of an antral follicle count do not benefit from a higher starting dose of gonadotrophins in IVF treatment: a randomized controlled trial. Hum Reprod (2005) 20:611–615.
Kwee J, Schats R, McDonnell J, Themmen A, de Jong F, Lambalk C. Evaluation of anti-Müllerian hormone as a test for the prediction of ovarian reserve. Fertil Steril (2007) [Epub ahead of print].
La Marca A, Stabile G, Artenisio AC, Volpe A. Serum anti-Mullerian hormone throughout the human menstrual cycle. Hum Reprod (2006) 21:3103–3107.
La Marca A, Giulini S, Tirelli A, Bertucci E, Marsella T, Xella S, Volpe A. Anti-Müllerian hormone measurement on any day of the menstrual cycle strongly predicts ovarian response in assisted reproductive technology. Hum Reprod (2007) 22:766–771.
Littenberg B, Moses LE. Estimating diagnostic accuracy from multiple conflicting reports: a new meta-analytic method. Med Decis Making (1993) 13:313–321.
Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, Prins MH, van der Meulen JH, Bossuyt PM. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA (1999) 282:1061–1066.
McIlveen M, Skull JD, Ledger WL. Evaluation of the utility of multiple endocrine and ultrasound measures of ovarian reserve in the prediction of cycle cancellation in a high-risk IVF population. Hum Reprod (2007) 22:778–785.
Midgette AS, Stukel TA, Littenberg B. A meta-analytic method for summarizing diagnostic test performances: receiver-operating-characteristic-summary point estimates. Med Decis Making (1993) 13:253–257.
Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med (1993) 12:1293–1316.[Web of Science][Medline]
Mol BW, van Wely M, Steyerberg EW. Using prognostic models in clinical infertility. Hum Fertil (Camb) (2003) 3:199–202.
Mol BW, Verhagen TE, Hendriks DJ, Collins JA, Coomarasamy A, Opmeer BC, Broekmans FJ. Value of ovarian reserve testing before IVF: a clinical decision analysis. Hum Reprod (2006) 21:816–823.
Muttukrishna S, Suharjono H, McCarrigle H, Sathanandan M. Inhibin B and anti-Mullerian hormone: markers of ovarian response in IVF/ICSI patients? BJOG (2004) 111:1248–1253.[Web of Science][Medline]
Penarrubia J, Fabregues F, Manau D, Creus M, Carmona F, Casamitjana R, Vanrell J, Balasch J. Previous cycle cancellation due to poor follicular development as a predictor of ovarian response in cycles stimulated with gonadotropin-releasing hormone agonist-gonadotropin treatment. Hum Reprod (2005) 20:622–628.
Popovic-Todorovic B, Loft A, Bredkjaeer HE, Bangsboll S, Nielsen IK, Andersen AN. A prospective randomized clinical trial comparing an individual dose of recombinant FSH based on predictive factors versus a ‘standard’ dose of 150 IU/day in ‘standard’ patients undergoing IVF/ICSI treatment. Hum Reprod (2003) 18:2275–2282.
Quban HS, Malkawi HY, Tahat YA, Areidah S, Nusair B, Khreisat BM, Al-Quraan G, Abu-Assaf A, Hadaddein MF, Abu-Jassar H. In-Vitro fertilisation treatment: factors affecting its results and outcome. J Obstet Gynaecol (2005) 25:689–693.[CrossRef][Medline]
Ranieri DM, Quinn F, Makhlouf A, Khadum I, Ghutmi W, McGarrigle H, Davies M, Serhal P. Simultaneous evaluation of basal follicle-stimulating hormone and 17β-estradiol response to gonadotropin-releasing hormone analogue stimulation: an improved predictor of ovarian reserve. Fertil Steril (1998) 70:227–233.[CrossRef][Web of Science][Medline]
SAS Institute Inc. SAS Technical Report P-242, SASR: Changes and Enhancements, Release 6.08 (1993) Cary, NC: SAS Institute INC. 113–116.
Stolwijk AM, Hamilton CJCM, Hollanders JMG, Bastiaans LA, Zielhuis GA. A more realistic approach to the cumulative pregnancy rate after in-vitro fertilization. Hum Reprod (1996) 11:660–663.
Tarlatzis BC, Fauser BC, Kolibianakis EM, Diedrich K, Rombauts L, Devroey P. GnRH antagonists in ovarian stimulation for IVF. Hum Reprod Update (2006) 12:333–340.
Templeton A, Morris JK, Parslow W. Factors that affect outcome of in-vitro fertilisation treatment [see comments]. Lancet (1996) 348:1402–1406.[CrossRef][Web of Science][Medline]
Tsepelidis S, Devreker F, Demeestere I, Flahaut A, Gervy Ch, Englert Y. Stable serum levels of anti-Müllerian hormone during the menstrual cycle: a prospective study in normo-ovulatory women. Hum Reprod. (2007) 22:1837–1840.
Te Velde ER, Pearson PL. The variability of female reproductive aging. Hum Reprod Update (2002) 8:141–154.
Van Rooij IA, Broekmans FJ, te Velde ER, Fauser BC, Bancsi LF, de Jong FH, Themmen AP. Serum anti-Mullerian hormone levels: a novel measure of ovarian reserve. Hum Reprod (2002) 17:3065–3071.
Van Rooij IA, Broekmans FJ, Scheffer GJ, Looman CW, Habbema JD, de Jong FH, Fauser BJ, Themmen AP, te Velde ER. Serum antimullerian hormone levels best reflect the reproductive decline with age in normal women with proven fertility: a longitudinal study. Fertil Steril (2005) 83:979–987.[CrossRef][Web of Science][Medline]
Weinstein M, Wood JW, Chang MC. Age patterns in fecundability. In: Biochemical and Demographic Determinants of Reproduction.—Gray R, Leridon H, Spina A, eds. (1993) Oxford: Clarendon Press. 209–220.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. van Disseldorp, C.B. Lambalk, J. Kwee, C.W.N. Looman, M.J.C. Eijkemans, B.C. Fauser, and F.J. Broekmans Comparison of inter- and intra-cycle variability of anti-Mullerian hormone and antral follicle counts Hum. Reprod., October 19, 2009; (2009) dep366v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. J. Broekmans, M. R. Soules, and B. C. Fauser Ovarian Aging: Mechanisms and Clinical Consequences Endocr. Rev., August 1, 2009; 30(5): 465 - 493. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||