Folliculogenesis and oogenesis are coupled processes that

Transcription

1 Follicular Fluid Levels of the Long Pentraxin PTX3 Alessio Paffoni, BS, Guido Ragni, MD, Andrea Doni, BS, Edgardo Somigliana, MD, Fabio Pasqualini, BS, Liliana Restelli, DSc, Giorgio Pardi, MD, Alberto Mantovani, MD, and Cecilia Garlanda, VetScD OBJECTIVE: Pentraxin-3 (PTX3) is a long pentraxin that plays a key role in female fertility as a structural and essential constituent of the cumulus oophorus extracellular matrix. Despite considerable evidence supporting this role of PTX3 in mice, data in humans are scanty. The aim of the present study was (1) to evaluate follicular fluid concentrations of PTX3; (2) to test the hypothesis that levels of the molecule correlate with oocyte characteristics (corona radiata, aspect of the cumulus, nuclear maturity, and fertilization); and (3) to evaluate the possibility that peripheral concentration of PTX3 may be of clinical help in monitoring ovarian hyperstimulation. METHODS: ELISA was used to determine PTX3 concentration. Levels of PTX3 were tested in 96 follicles. RESULTS: The mean SD and the median (interquartile range) were and 12.1 ( ) ng/ml, respectively. Levels of the molecule did not appear to be normally distributed. At the day of ovum pick-up, levels of PTX3 were 6.3-fold higher in follicular fluid than in peripheral blood (95% CI, ). No statistically significant difference emerged linking follicular fluid concentration of PTX3 and oocyte quality. In a series of ten women, plasma concentration of PTX3 did not vary during ovarian hyperstimulation, resulting in levels of at the 3rd day of the menstrual cycle and ng/ml at the day of oocyte retrieval. CONCLUSIONS: Results from the present study support the following conclusions: (1) elevated levels of soluble PTX3 can be found in follicular fluid; (2) follicular fluid concentration of PTX3 cannot by used as a marker of oocyte quality; and (3) plasma concentration of the molecule is not influenced by ovarian hyperstimulation. (J Soc Gynecol Investig 2006;13:226 31) Copyright 2006 by the Society for Gynecologic Investigation. KEY WORDS: Pentraxin-3, follicular fluid, oocyte quality, ovarian hyperstimulation. Folliculogenesis and oogenesis are coupled processes that are coordinated through a precise communication resulting in mature oocytes competent for fertilization and successive embryonic development. No follicle is formed without an oocyte and surrounding granulosa cells are essential to ensure an adequate oocyte development. 1 In recent years, the establishment of mouse models in which specific genes have been disrupted has contributed to clarifying the molecular mechanisms that control oogenesis and folliculogenesis. Despite progress, many of the molecules defining the circuits that control these fascinating processes remain unknown. In this context, two independent groups have recently established that pentraxin-3 (PTX3), a molecule known to be involved in From the Infertility Unit, Department of Obstetrics, Gynecology, and Neonatology, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, Milan, Italy; Department of Immunology and Cell Biology, Mario Negri Institute for Pharmacological Research, Milan, Italy; and Centro IDET, Institute of General Pathology, Università degli Studi di Milano, Milan, Italy. Supported by the European Commission. Address correspondence and reprint requests to: Alessio Paffoni, BS, Infertility Unit, Regina Elena Hospital, Via M. Fanti, 6, Milan, Italy. alessio.paffoni@alice.it innate immunity against selected pathogens, 2,3 also plays a crucial role in oogenesis and folliculogenesis in mice. 4,5 PTX3 appears to be a structural and essential constituent of the cumulus oophorus extracellular matrix that allows fecundation to occur. Specifically, Varani et al, through gene chip technology, identified PTX3 in mouse as a novel protein downstream of the growth differentiation factor 9 (GDF-9) signal transduction pathway in the periovulatory stage of folliculogenesis. They found PTX3 mrna transcription in granulosa cells after the luteinizing hormone (LH) surge and postulated a role as an integral factor in cumulus cell-oocyte complex. Moreover, using a gene knock out approach, they showed that PTX3 null ( / ) female mice develop normally but are highly subfertile due to defects in cumulus cell-oocyte integrity. 4 More recently, Salustri et al confirmed PTX3 as a key molecule in cumulus formation and expansion. 5 They showed that, in PTX3-deficient mice, corona radiata cells fail to polarize and randomly surround the oocyte in the preovulatory follicle and cumulus cell-oocyte complexes are instable and no longer identifiable after ovulation. Moreover, these authors have reported that PTX3 displays its matrix-stabilizing activity Copyright 2006 by the Society for Gynecologic Investigation /06/$32.00 Published by Elsevier Inc. doi: /j.jsgi

2 Follicular Fluid J Soc Gynecol Investig Vol. 13, No. 3, April through interaction with another crucial component of the cumulus matrix, tumor necrosis factor -induced protein 6 (TNFAIP6). TNFAIP6 is a multifunctional protein usually associated with inflammation, which has the ability to specifically bind hyaluronan acid (HA) and, as PTX3, colocalizes with HA throughout the matrix, from the periphery of the cumulus to the zona pellucida. 6,7 As PTX3 is unable to bind directly HA, TNFAIP6 would be the component involved in mediating such interaction. More specifically, PTX3, which is predominantly assembled as a large multimer complex consisting of two decamers that interact with TNFAIP6, would serve as an anchoring site for multiple HA molecules. 5 Despite a considerable body of evidence supporting the non-redundant role of PTX3 in the mechanisms surrounding folliculogenesis in mice, data in humans are scanty. 5,8 Salustri et al qualitatively observed that both mrna and the protein of PTX3 are present in the cumulus oophorus of oocytes retrieved during in vitro fertilization (IVF) cycles. The molecule was also found to be present in a soluble form in the follicular fluid obtained from women performing an IVF cycle. The concentration reported in the five samples analyzed was found to vary considerably. 5 Moreover, preliminary results from Zhang et al have shown that the amount of mrna for PTX3 contained in the cumulus significantly correlates with the rate of fecundation, the quality of the embryos at 48 to 72 hours, and the pregnancy rate. 8 The aim of the present study was to evaluate follicular fluid concentrations of PTX3 in a large series of follicular fluid samples and to test the hypothesis that levels of the molecule correlate with oocyte quality. Moreover, the possibility that peripheral concentration of PTX3 may be of clinical help in monitoring ovarian hyperstimulation was also evaluated. MATERIALS AND METHODS Subjects All subjects enrolled in this study were less than 40 years of age and had regular menstrual cycles. Two groups of patients were recruited. One hundred thirty-four infertile women who performed an intracytoplasmic sperm injection (ICSI) procedure for severe male infertility and eight women who underwent a natural cycle ICSI due to both a male factor and a previous poor responder cycle (fewer than three oocytes retrieved despite the administration of 450 IU recombinant folliclestimulating hormone [rfsh] per day). Women were enrolled only if the number of retrieved oocytes was at least three for women undergoing ovarian hyperstimulation and one or more for those performing natural cycle ICSI. Two different types of pharmacologic regimens were used in women performing hyperstimulation cycles: the long protocol with daily 0.1 mg gonadotropin-releasing hormone (GnRH) agonist (triptoreline; Decapeptyl, Ipsen Pharma, Pavia, Italy) (n 70) and a protocol using daily 0.25 mg GnRH antagonist (cetrorelix 0.25 mg daily; Cetrotide, Serono Laboratories, Inc, Rome, Italy) (n 64). In the long protocol, GnRH agonist was administered daily starting in the mid-luteal phase of the preceding cycle and continued until ovulation was triggered. Seventeen to 18 days after GnRH agonist initiation, rfsh (Gonal-F, Serono Laboratories, Inc) was started. Conversely, in cycles using GnRH antagonist, rfsh was initiated the third day of the menstrual cycle. The antagonist was started from the day in which at least one follicle was 14 mm. In both regimens, concentration of rfsh prescribed was decided based on age, hormonal tests, ultrasound characteristics of the ovaries, and results from previous pharmacologic ovarian hyperstimulation cycles. In all cases, follicular growth was monitored by serial transvaginal ultrasound and serum levels of 17 estradiol. Regardless of the type of ovarian stimulation, ovulation was triggered administering 5,000 IU of human chorionic gonadotropin (hcg) (Profasi, Serono Laboratories, Inc) and transvaginal oocyte retrieval was performed 36 hours later. Approval for the study was granted by the local ethical committee. All patients gave a written consent prior to participate to the study. Blood Samples In 10 women who underwent ovarian hyperstimulation using a protocol of rfsh and GnRH antagonist, samples of blood were collected at both the third day of the cycle (prior to the initiation of the stimulation) and the day of ovum pick-up. Levels of PTX3 were also determined at the time of oocyte retrieval in 31 women undergoing ovarian hyperstimulation cycles (long protocol in 16 and protocol with GnRH antagonist in 15 cases). All blood samples were collected in EDTA and successively centrifugated at 800 g for 10 minutes at room temperature to separate plasma from cells. Plasma was then transferred into cryovials and stored 20C until PTX3 assay. Follicular Fluid Follicular fluids were collected in two ways. In one group of patients (n 10), a pool of all aspirated follicular fluids was prepared. Levels of PTX3 in these pools were then correlated to characteristics of the stimulated cycle. In another group of patients (n 96), follicular fluid and the corresponding oocyte were collected separately in order to correlate levels of PTX3 to morphologic and functional characteristics of the oocyte. No more than one follicle per patient was studied. Immediately after recovery of oocytes, the volume of follicular fluids belonging to single follicles was measured by aspiration into a graduated pipette and recorded. All follicular fluid samples collected were successively centrifuged at 350 g for 10 minutes at room temperature to separate them from cells. Samples were transferred into cryovials and stored 20C until PTX3 assay. Oocyte and Embryo Evaluation Oocytes were evaluated using a Nikon inverted T200 microscope (Nikon Instruments S.P.A., Sesto Fiorentino, Italy). Oocytes coming from follicular fluids recruited for the study were cultured and treated alone in every step of the assisted reproduction procedure in order to evaluate characteristics of

3 228 J Soc Gynecol Investig Vol. 13, No. 3, April 2006 Paffoni et al morphology, fertilization, and development and correlate them with follicular PTX3 concentration. Morphologic evaluation was performed as previously reported All oocytes/ embryos were evaluated separately by two of the authors (A.P. and L.R.) who were blinded to the concentration of PTX3. If discordant evaluations were given, the result was not considered. The following dicotomic variables were evaluated: corona radiata (normal versus anomalous), aspect of the cumulus (normal versus anomalous), nuclear maturity (competent versus not competent), and fertilization (normal versus anomalous). More specifically, these dicotomic classifications were determined as follows. The corona layer was graded according to the degree of expansion, considering normal a corona with good or excellent expansion and a clear delimitation of the oocyte outline. Corona layers appearing dense and with no clear delimitation of the oocyte outline (immature) or dark and clumped (postmature) were considered anomalous. The cumulus was classified on the basis of the degree of expansion, defined by the spacing between the cells, and dimension. Cumuli appearing well expanded with fluffy cells were considered normal, while absent to sparse cumulus cells, dense and tightly packed cumulus cells or scanty, dark, or clumped cumulus cells where classified as anomalous. After cumulus removal by conventional use of 40 IU/mL hyaluronidase and hand-pulled pipettes, nuclear competence was established at 200: only oocytes showing a single polar body in the perivitelline space were considered competent, while oocytes showing a germinal vesicle, no polar body, or degeneration were considered not competent. Oocyte cytoplasm morphology was assessed at 200 magnification. Clear cytoplasm with uniform texture and fine granularity was considered normal. Conversely, dark oocytes with granularity affecting the whole cytoplasm or the central portion of the cytoplasm, oocytes showing cytoplasmic inclusions, vacuoles, or refractile bodies, were considered anomalous. Only oocytes classified as competent, even if considered abnormal on the basis of cytoplasmic appearance, were used for IVF. Fourteen to 18 hours after ICSI and 18 to 20 hours after conventional IVF, oocytes were examined for the presence of pronuclei. The term normal fertilization was assigned to oocytes with two pronuclei and two extruded polar bodies. Oocytes showing no pronuclei or a single or more than two pronuclei, as oocytes showing signs of degeneration, were considered anomalous. Enzyme-Linked Immunosorbent Assay for PTX3 Plasma and follicular fluid sample collection, handling, and storage were carefully standardized. The sandwich enzymelinked immunosorbent assay (ELISA) for PTX3 was performed as previously described. 12,13 Briefly, ELISA plates (96-well; Nunc Immuno Plate, MaxiSorp; Nunc A/S, Roskilde, Denmark) were coated with 100 ng/well of rat monoclonal anti- PTX3 antibody (mab) MNB4 diluted in coating buffer (15 mm carbonate, Na 2 CO 3 NaHCO 3, buffer ph 9.6) and were incubated overnight at 4C. The plates were washed with washing buffer (Dulbecco s phosphate-buffered saline containing 0.05% Tween 20) and 300 L of 5% dry milk were added to block nonspecific binding sites. A 50- L quantity of recombinant human PTX3 standards (100 pg/ml to 2 ng/ml) and unknown plasma samples were added in duplicate, and incubated for 2 hours at 37C. After three washes with the washing buffer, 25 ng/well of biotin conjugated PTX3 affinity-purified rabbit IgG was added for 1 hour at 37C. Wells were extensively washed and incubated with 100 L of streptavidin-peroxidase conjugated to dextran backbone (Am- Dex, Copenhagen, Denmark) diluted 1:4,000 for 1 hour at room temperature. After incubation, the plates were washed four times and 100 L of TMB chromogen (Pharmingen BD, San Diego, CA) was added. Absorbance values were read at 405 nm in an automatic ELISA reader. PTX3 levels in follicular fluids were analyzed after addition of EDTA 0.35% final concentration to sample dilution buffer in order to eliminate interference in the assay by calciumdependent factors. PTX3 levels in follicular fluids were also normalized to total protein content, determined by using a Comassie Plus protein assay reagent in micro plates (Pierce, Rockford, IL). Statistical Analysis The Social Package for Statistical Science (SPSS 11.5 for Windows, Chicago, IL) was employed. Data were reported as the mean SD if normally distributed and as the median (interquartile range [IQR]) if not. Wilcoxon test for unpaired data was used to compare quantitative variables not normally distributed. Paired or unpaired Student t test was used to compare normally distributed variables. Results from previous studies suggested that plasma concentration of PTX3 can be considered normally distributed. 12,13 P.05 was considered statistically significant. RESULTS Distribution of PTX3 concentration in the follicular fluid obtained from 96 follicles is illustrated in Figure 1. The mean SD and the median (IQR) were and 12.1 ( ) ng/ml, respectively. As shown in Figure 1, levels of the molecule did not appear to be normally distributed. The skewness index ( SE) and the kurtosis index ( SE) were and , respectively. Distribution of total amount of PTX3 per follicle (calculated as C V, where C concentration and V volume of follicular fluid) was similar (data not shown). Since levels were positively skewed, nonparametric statistics were used for analysis on data regarding follicular concentrations of PTX3. Median (IQR) levels of PTX3 in eight follicular fluids obtained from women undergoing natural cycle IVF were 10.1 ( ) ng/ ml. These levels resulted not significantly different from those observed in 88 follicular fluids obtained from women undergoing ovarian hyperstimulation (median 12.1, IQR ng/ml) (P.93).

4 Follicular Fluid J Soc Gynecol Investig Vol. 13, No. 3, April Number of observations Follicular fluid concentration of PTX3 (ng/ml) Figure 1. Distribution of follicular fluid concentrations of PTX3. Data did not appear to be normally distributed. The skewed index ( SE) and the kurtosis index ( SE) were and , respectively. Table 1. Follicular Levels of PTX-3 According to Oocyte Quality Characteristics No. PTX-3 (ng/ml) P Corona radiata Normal ( ).42 Anomalous ( ) Cumulus Normal ( ).73 Anomalous ( ) Nuclear competence Competent ( ).17 Not competent ( ) Cytoplasm Normal ( ).43 Anomalous ( ) Fertilization* Normal ( ).61 Anomalous ( ) Data are expressed as median (interquartile range). P values were calculated using nonparametric Wilcoxon test for unpaired data. * Only competent oocytes (n 82) were fertilized. Concentrations of PTX3 in the follicular fluids were correlated to morphologic characteristics of the corresponding oocyte. The hypothesis that fertilization rate may be associated to follicular fluid levels of PTX3 was also tested. Results from these analyses are illustrated in Table 1. No statistically significant difference emerged linking follicular fluid concentration of PTX3 and oocyte quality. Similar results were observed when total amount of follicular PTX3 rather than concentration of the molecule was used and when concentration of Figure 2. Plasma and follicular fluid concentrations of PTX3. A pool of all aspirated follicular fluids and a sample of peripheral blood were concomitantly obtained at the time of oocyte retrieval in 10 women. Levels of PTX3 were and ng/ml in pools of follicular fluid and in plasma, respectively (P.001). Levels of the molecule were 6.3-fold higher in follicular fluid (95% CI, ). PTX3 (ng/ml) Plasma 2 Follicular fluid Flli l flid

5 230 J Soc Gynecol Investig Vol. 13, No. 3, April 2006 Paffoni et al PTX3 was normalized to the total amount of proteins (data not shown). In 10 patients, a pool of all aspirated follicular fluids and a sample of peripheral blood were concomitantly obtained at the time of oocyte retrieval. Data were compared using Student t test for paired data since, according to the theorem of the central limit tendency for averages, PTX3 levels in pools of follicular fluids are expected to be normally distributed. As shown in Figure 2, levels of PTX3 were and ng/ml in pools of follicular fluid and in plasma, respectively (P.001). Levels of the molecule were 6.3-fold higher in follicular fluid (95% CI, ). The hypothesis that plasma concentration of PTX3 could be modified during ovarian hyperstimulation was also tested. Ten women who were prescribed a regimen of recombinant FSH and GnRH antagonist gave a blood sample on the third day of the menstrual cycle (prior to initiating ovarian hyperstimulation) and at the time of oocyte retrieval. Levels of PTX3 at these two periods were and ng/ml, respectively (paired Student t test, P 1.00). To rule out a confounding role of the regimen used, peripheral levels of PTX3 at the time of oocyte retrieval were evaluated in women who were administered the long protocol (n 16) and in those receiving the protocol with GnRH antagonist (n 15). Concentrations of PTX3 were and ng/ml, respectively (unpaired Student t test, P.86). DISCUSSION In this study, we have documented that PTX3 is a component of follicular fluid at the time of oocyte retrieval in women. In this milieu, concentrations of the molecule were sixfold higher than in plasma. From a statistical point of view, levels of PTX3 were characterized by a relevant variability; distribution of concentrations appeared to be positively skewed and the kurtosis index was remarkably high. Based on previous studies supporting a crucial non-redundant role of PTX3 in the fine mechanisms surrounding ovulation, 4,5 we have hypothesized that the quality of the oocyte may significantly influence follicular levels of molecule. However, we have failed to detect any significant association between morphologic characteristics of the oocyte and concentration of PTX3 in the corresponding follicular fluid. Furthermore, fertilization rate did not appear to be associated with follicular concentration of PTX3. Overall, these results do not support the idea that levels of PTX3 in the follicular fluid may mirror the quality of the corresponding oocyte. As a consequence, the use of this concentration cannot be proposed to predict oocyte quality. Since quality of the oocyte did not appear to be significantly associated with follicular concentrations of PTX3, it cannot be speculated that variability of this concentration can be explained by variable oocyte quality. Alternative explanations for the variable distribution have thus to be hypothesized. In this regard, it has to be noted that Northern blot analysis of whole ovary revealed that PTX3 mrna expression is finely regulated in mice. Specifically, its expression starts 2 hours after injection of an ovulatory dose of hcg, peaks at 6 hours, and declines thereafter. 5 In stimulated cycles in women, ovulations usually begin 36 hours after hcg administration and are sequential over several hours. 14 For this reason, oocyte retrieval is typically performed at 36 hours after hcg injection: this strategy recruits a high rate of mature oocytes while minimizing the risk that ovulation has occurred prior to initiating retrieval. It might thus be hypothesized that the extremely temporary limited expression of PTX3 on one hand and variability in time between hcg administration and ovulation on the other hand may at least in part explain the considerable variability in follicular concentrations of PTX3. Zhang et al have recently reported that, in women, PTX3 mrna expression in the cumulus oophorus of retrieved oocyte is associated with oocyte quality. In particular, higher expression of the molecule was associated with higher probability of fertilization, improved embryo quality, and higher pregnancy rate. 8 Our results showed that similar conclusions cannot be drawn considering follicular concentration of soluble PTX3 at the time of oocyte retrieval. These observations are not conflicting. Indeed, several biologic steps have to be considered between mrna expression in the cumulus and release of PTX3 in follicular fluid. Of note, PTX3 has been shown to play a crucial non-redundant role in allowing an optimal architectural structure of the cumulus in mice. 4,5 Conversely, a role for PTX3 shedding in the follicular fluid has not been demonstrated. Thus, it cannot be ruled out that the presence of PTX3 in this milieu should be viewed as an epiphenomenon without biologic relevance. If true, this shedding would not be finely regulated and this would explain the extreme variability of follicular fluid concentrations of the molecule and the failure to detect significant correlations with oocyte quality. However, this hypothesis needs to be confirmed. Finally, results from the present study do not support the use of plasma concentration of PTX3 to monitor ovarian response during pharmacologic stimulation. Indeed, levels of the molecule were not significantly modified during therapy. At the time of oocyte retrieval, peripheral concentrations were extremely similar to those measured prior to the initiation of ovarian hyperstimulation. This conclusion is further supported by the observation that plasma levels of PTX3 reported in our study at the time of oocyte retrieval were similar to those detected in the control healthy groups of two recent large studies on the use of peripheral levels of PTX3 in coronary heart disease and sepsis. 12,13 This observation supports the idea that PTX3 is mostly confined to the follicle, at least until ovulation occurs. Although release of the molecule into the bloodstream cannot be excluded, this phenomenon should be limited and biologically not relevant. In summary, results from the present study support the following conclusions: (1) elevated levels of soluble PTX3 can be found in follicular fluid; (2) follicular fluid concentration of PTX3 cannot be used as a marker of oocyte quality; and (3) plasma concentration of the molecule is not influenced by ovarian hyperstimulation.

6 Follicular Fluid J Soc Gynecol Investig Vol. 13, No. 3, April REFERENCES 1. Amleh A, Dean J. Mouse genetics provides insight into folliculogenesis, fertilization and early embryonic development. Hum Reprod Update 2002;8: Garlanda C, Hirsch E, Bozza S, et al. Non-redundant role of the long pentraxin PTX3 in anti-fungal innate immune response. Nature 2002;420: Garlanda C, Bottazzi B, Bastone A, Mantovani A. Pentraxins at the crossroads between innate immunity, inflammation, matrix deposition and female fertility. Annu Rev Immunol 2005; 23: Varani S, Elvin JA, Yan C, et al. Knockout of pentraxin 3, a downstream target of growth differentiation factor-9, causes female subfertility. Mol Endocrinol 2002;16: Salustri A, Garlanda C, Hirsch E, et al. PTX3 plays a key role in the organization of the cumulus oophorus extracellular matrix and in in vivo fertilization. Development 2004;131: Lee TH, Wisniewski HG, Vilcek J. A novel secretory tumor necrosis factor-inducible protein (TSG-6) is a member of the family of hyaluronate binding proteins, closely related to the adhesion receptor CD44. J Cell Biol 1992;116: Milner CM, Day AJ. TSG-6: A multifunctional protein associated with inflammation. J Cell Sci 2003;116: Zhang X, Jafari N, Barnes RB, Confino E, Milad M, Kazer RR. Studies of gene expression in human cumulus cells indicate pentraxin 3 as a possible marker for oocyte quality. Fertil Steril 2005;83 Suppl 1: Staessen C, Camus M, Bolle N, Devroey P, Van Steirteghem AC. The relationship between embryo quality and the occurrence of multiple pregnancies. Fertil Steril 1992;57: Rattanachaiyanont M, Leader A, Léveilé MC. Lack of correlation between oocyte-corona-cumuls complex morphology and nuclear maturity of oocytes collected in stimulated cycles for intracytoplasmic sperm injection. Fertil Steril 1999;71: Ng ST, Chang TH, Wu J. Prediction of the rates of fertilization, cleavage, and pregnancy success by cumulus-coronal morphology in an in vitro fertilization program. Fertil Steril 1999;72: Peri G, Introna M, Corradi D, et al. PTX3, A prototypical long pentraxin, is an early indicator of acute myocardial infarction in humans. Circulation. 2000;102: Muller B, Peri G, Doni A, et al. Circulating levels of the long pentraxin PTX3 correlate with severity of infection in critically ill patients. Crit Care Med 2001;29: Templeton AA, Van Look P, Angell RE, Aitken RJ, Lumsden MA, Baird DT. Oocyte recovery and fertilization rates in women at various times after administration of hcg. J Reprod Fertil 1996;76:771-8.