Human Reproduction Vol.19, No.3 pp.

611615, 2004 Advance Access publication 29 January, 2004

DOI: 10.1093/humrep/deh127

Late, but not early, paternal effect on human embryo development is related to sperm DNA fragmentation
Jan Tesarik1,2,5, Ermanno Greco3 and Carmen Mendoza2,4
MAR&Gen, Molecular Assisted Reproduction and Genetics, Gracia 36, 18002 Granada and 4University of Granada, Campus Fuentenueva, 18004 Granada, Spain, 2Laboratoire d'Eylau, 55 rue Saint-Didier, 75116 Paris, France and 3European Hospital, Via Portuense 700, Rome, Italy
5 1

To whom correspondence should be addressed at: MAR&Gen, Gracia 36, 18002 Granada, Spain. E-mail: cmendoza@ugr.es

BACKGROUND: It is known that repeated failure of assisted reproduction treatment (ART) can be due to a paternal effect. This study was undertaken to analyse the possible relationship between ART failure and sperm DNA fragmentation. METHODS: Zygote morphology and the percentage of spermatozoa with fragmented DNA (assessed by TUNEL) were compared in two groups using donor oocytes for ICSI attempts. The experimental group consisted of 18 infertile couples who had each undergone three previous failed ART attempts. The control group included 18 randomly selected infertile couples undergoing their rst ICSI attempt. Both groups used sibling oocytes from the same donors. RESULTS: In 10 couples of the experimental group, the adverse paternal effect was evident as early as the zygote stage. This early paternal effect was not associated with sperm DNA fragmentation. In eight couples of the experimental group, the adverse paternal effect did not produce any perceptible deterioration of zygote morphology. However, this late paternal effect was associated with an increased percentage of spermatozoa with fragmented DNA. CONCLUSIONS: Early paternal effect can compromise ART outcomes in the absence of increased sperm DNA fragmentation. Evaluation of sperm DNA integrity is useful to detect late paternal effect, which is not associated with morphological abnormalities at the zygote and early cleavage stages.
Key words: embryo/implantation/paternal effect/preimplantation development/sperm DNA fragmentation

Introduction Previous studies have shown that a paternal effect can be responsible for repeated failure of assisted reproduction treatment (ART) attempts (Vanderzwalmen et al., 1991; Parinaud et al., 1993; Janny and Menezo, 1994; Hammadeh et al., 1996; Sanchez et al., 1996; Shoukir et al., 1998). Adverse paternal effects are sometimes apparent as early as the pronuclear zygote stage (Tesarik et al., 2002). However, we have also observed cases of repeated implantation failure after ICSI with donor oocytes, contrasting with high pregnancy, implantation and birth rates achieved with the use of sibling oocytes from the same donors shared by other recipient couples, without any apparent impairment of zygote and cleaving embryo morphology. In this study, these two conditions are referred to as early paternal effect and late paternal effect, respectively. The early paternal effect has been suggested to be mediated by centrosome dysfunction or deciency of oocyte-activating factors (Tesarik et al., 2002). The implication of sperm chromatin packaging or DNA damage remains a controversial issue. In fact, the early paternal effect was observed before the major activation of embryonic genome expression which begins at the 4-cell stage in humans (Tesarik et al., 1986,

1988; Braude et al., 1988). However, limited RNA synthesis can be detected in human pronuclei (Tesarik and Kopecny, 1989), and failure of this early zygotic transcription is associated with abnormal pronuclear development (Tesarik and Kopecny, 1990). On the other hand, the late paternal effect may involve sperm aneuploidy, DNA damage or abnormal chromatin packaging, which can inuence the orderly activation of paternal gene expression. Several studies have evaluated DNA structure and integrity in human ejaculated sperm samples and reported a negative effect of high percentages of spermatozoa with damaged DNA on pregnancy rates in ART (Filatov et al., 1999; Host et al., 2000; Larson et al., 2000; Morris et al., 2002; Tomsu et al., 2002; Benchaib et al., 2003). These studies have suggested that DNA damage may serve as a marker for the diagnosis of an adverse paternal effect on human preimplantation development. However, no study dealing with the possible relationship between sperm DNA damage and the early paternal effect on zygote quality has yet been performed. In this study, we tested potential paternal effects by comparing two groups. The experimental group consisted of infertile couples using donor oocytes who had failed three previous ICSI attempts. The control group included randomly 611

Human Reproduction vol. 19 no. 3 European Society of Human Reproduction and Embryology 2004; all rights reserved

J.Tesarik, E.Greco and C.Mendoza

selected infertile couples also using donor oocytes in their rst ICSI attempt. Both groups used sibling oocytes from the same donors. The results of this analysis were related to the percentage of spermatozoa with fragmented DNA in the corresponding patient groups. Materials and methods
Patients This study involved 18 infertile couples who had each undergone three unsuccessful ART attempts with donor oocytes at our clinic (experimental group) compared with 18 randomly chosen infertile couples in their rst ICSI attempt (control group) who shared sibling oocytes from the same donors as the experimental group. All oocyte recipients had a normal uterus without any pathological formations, and their endometrial thickness on the day of embryo transfer was b8 mm. All of them had normal serum concentrations of glucose, insulin thyroid-stimulating hormone (TSH) prolactin, and antiphospholipid, anti-cardiolipin and anti-nuclear antibodies. Women suffering from endometriosis were not included. Based on the evaluation of early embryo morphology in the previous unsuccessful attempts, the 18 couples with a history of three previous ART failures were divided into two groups referred to as early paternal effect and late paternal effect, respectively. Early paternal effect This group was formed by couples in whom b5 oocytes (b50%) were fertilized in each of the three previous ART attempts, but <25% of the resulting zygotes were classied as having good morphology according to the previously described criteria (Tesarik and Greco, 1999; Tesarik et al., 2002). Late paternal effect This group was formed by couples in whom b5 oocytes were fertilized in each of the three previous ART attempts that resulted in implantation failure in spite of no apparent impairment of zygote and cleaving embryo morphology (b50% of zygotes were classied as having good morphology). Design In our oocyte donation programme, oocytes from each donor are shared by two infertile couples. Each of the two couples received half of the metaphase II oocytes recovered from the corresponding donor. If an odd number of oocytes was recovered, the couple receiving the extra oocyte was determined randomly. The analyses described in this study concern the outcomes of a fourth ART attempt with donor oocytes performed for the couples showing either an early or a late paternal effect (experimental group). The data obtained were compared with those concerning randomly chosen infertile couples undergoing their rst ICSI attempt (control group) who shared sibling oocytes with the couples involved in the experimental group. In parallel, sperm samples used for ART in each of these attempts were analysed for basic sperm parameters and for DNA fragmentation (see below). Oocyte donor preparation Oocyte donors were healthy volunteers aged between 20 and 28 years, without any apparent reproductive or other pathology and with a good ovarian reserve (basal serum FSH and inhibin B concentrations of <7 IU/l and >100 pg/ml, respectively). The general health status of each candidate for oocyte donation was assessed by the examinations dened by the Spanish Law of Assisted Reproduction, including

karyotype and tests for human immunodeciency virus (HIV), hepatitis B and C, cytomegalovirus, herpes virus, rubella, clamydia, toxoplasmosis and syphilis. Ovarian stimulation of oocyte donors was performed with the use of a combination of recombinant and urinary human gonadotrophins after pituitary downregulation started in the mid-luteal phase as described (Tesarik and Mendoza, 2002). Ovulation was induced with 10 000 IU of HCG (Profasi; Serono, Rome, Italy) when there were at least ve follicles of b18 mm in diameter in both ovaries. Ultrasoundguided transvaginal follicular aspiration was performed 36 h later. Oocyte recipient preparation Oocyte recipients were ovulating women in whom the recourse to oocyte donation was indicated because of a poor response to ovarian stimulation or poor oocyte morphology. The ovarian activity of the oocyte recipients was suppressed by pituitary desensitization with triptorelin (Decapeptyl; Ipsen Pharma, Madrid, Spain) started in the midluteal phase. When the serum estradiol concentration was <45 pg/ ml and there was no cystic structure of >10 mm detectable by ultrasonography in the patient's ovaries, the desensitization was considered complete, and the patient was ready for the beginning of endometrial growth stimulation. This condition was maintained until the respective oocyte donor was ready for ovarian stimulation. Accordingly, the period between the start of triptorelin treatment and embryo transfer varied between 32 and 45 days. Triptorelin was rst applied in a single i.m. injection of a long-acting preparation (Decapeptyl LP, 3.75 mg) followed, when necessary, by daily s.c. injections of a short-acting preparation of triptorelin (Decapeptyl, 0.1 mg). Stimulation of endometrial growth was started 1 day before the beginning of ovarian stimulation of the respective donor. It was performed by oral estradiol valerate (Progynova; Schering, Madrid, Spain) at daily doses of 2 mg on days 14, 4 mg on days 58 and 6 mg from day 9 until 1 day after the injection of HCG in the respective oocyte donor when the daily dose of estradiol valerate was reduced to 2 mg. On the same day, intravaginal application of natural micronized progesterone (Utrogestan; Laboratoires Besins-Iscovesco, Paris, France) was started, beginning with a daily dose of 200 mg which was augmented on the two following days to 400 and 600 mg, respectively. The daily dose of 600 mg micronized progesterone was maintained until the pregnancy test, and in the case of pregnancy, this medication was prolonged through the 3 months following embryo transfer. The treatment with 2 mg of estradiol valerate was maintained until the negative pregnancy test or, in the case of pregnancy, until the serum estradiol concentration had reached 1000 pg/ml. Assisted reproduction techniques Oocytes were prepared and subjected to ICSI using previously described techniques and equipment (Tesarik and Sousa, 1995). Fertilization was assessed 1416 h after ICSI. Oocytes were considered fertilized when they contained two pronuclei. Abnormally fertilized (three or more pronuclei) and parthenogenetically activated oocytes (one pronucleus) were discarded at this stage and are not taken into account. Zygotes were classied as having good morphology or poor morphology according to previously described criteria based on the number and distribution of nucleolar precursor bodies (Tesarik and Greco, 1999; Tesarik et al., 2000). Embryo in vitro culture was performed in IVF medium (Vitrolife; Goteborg, Sweden) equilibrated with 5% CO2 in air up to day 3 after ICSI when three embryos were transferred to the patient's uterus. Embryos that showed the best scores at the cleavage stages (see the section on ygote and cleaving embryo evaluation) were selected for transfer. Where possible, these embryos were selected from those that

612

Paternal effect and sperm DNA

developed from good morphology pronuclear zygotes. If supernumerary embryos of adequate quality were available, they were cryopreserved for eventual later transfer. However, only treatment attempts with fresh embryos are involved in this study. Only clinical pregnancies, characterized by the presence of an embryonic heartbeat, are taken into account in this study. Sperm evaluation Basic evaluation. Basic evaluation of sperm samples used in this study was performed according to World Health Organization instructions (World Health Organization, 1999). Evaluation of sperm DNA fragmentation. Sperm DNA fragmentation was evaluated in the swim-up sperm preparation used for ICSI by the terminal deoxyribonucleotidyl transferase-mediated dUTP nick-end labelling (TUNEL) assay using the in situ Cell Death Detection Kit with uorescein isothiocyanate (FITC)-labelled dUTP (Roche, Monza, Italy). Samples of each sperm preparation were smeared on at least four microscope slides, xed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate. The rest of the procedure was carried out according to the manufacturer's instruction. Briey, the specimens were incubated in TUNEL reaction mixture in the dark at 37C for 1 h followed by evaluation in a uorescence microscope. Four hundred spermatozoa were randomly analysed in two slides for each sample (200 spermatozoa per slide). The remaining two slides were used for the negative (omitting the enzyme terminal transferase) and positive (using DNase I, 1 mg/ml for 20 min at room temperature) controls. Zygote and cleaving embryo evaluation Zygotes were evaluated with the use of a previously described scoring system based on the examination of pronuclear morphology (Tesarik and Greco, 1999) in its simplied version (Tesarik et al., 2000). Accordingly, zygotes with normal pronuclear patterns are classied as having good morphology, and the ve originally described abnormal pronuclear patterns (Tesarik and Greco, 1999) are grouped together as poor morphology zygotes. The quality of cleaving embryos was scored using criteria based on the cleavage speed and morphology (blastomere regularity and the volume occupied by anucleate cell fragments). The latter was quantied with the use of previously described embryo morphologygrading criteria (Bolton et al., 1989). Statistics Paired Student's t-test and Wilcoxon matched pairs test were used as applicable. Continuous outcomes and average percentage differences between matched groups are represented as mean T SD. All analyses were performed using the Statistica 5.0 software (Statsoft Version 5.1, Hamburg, Germany).

The values of basic sperm parameters, including sperm count (12.6 Q 106 T 10.1 Q 106 versus 14.1 Q 106 T 10.9 Q 106), motility (35.2 T 12.4% versus 32.0 T 12.5%) and the percentage of normal forms (24.8 T 10.5% versus 25.3 T 10.9%), were not different between the experimental and the control group. Moreover, when sperm DNA fragmentation was evaluated in samples used for the ART attempts under study, there was no difference between the two groups (Table I). One singleton pregnancy was established in the experimental group. This contrasted with seven pregnancies, of which four were twin, in the control group. The implantation rate was signicantly lower in the former group as compared with the latter (Table I). Endometrial thickness on the day of embryo transfer was similar in both groups (Table I). Late paternal effect Similarly to the early paternal effect, there were b50% normally fertilized oocytes in this group of patients as well as in the group of randomly chosen patients who shared sibling oocytes from the same donors. However, in contrast to the early paternal effect, no difference in the percentage of good morphology zygotes was observed between these two groups of patients (Table II). The examination of sperm samples used for the ART attempt described in this study did not reveal any signicant differences in sperm count (14.0 Q 106 T 11.2 Q 106 versus 13.6 Q 106 T 10.8 Q 106), motility (32.4 T 12.8% versus 31.9 T 13.3%) and percentage of normal forms (22.5 T 10.9% versus 25.1 T 11.8%) between the experimental and the control group. However, the percentage of spermatozoa showing DNA fragmentation by TUNEL assay was signicantly higher in the experimental group as compared with the control group (Table II). There was no pregnancy in the experimental group. This contrasted with six clinical pregnancies, of which three were twin, obtained in the control group. The difference in the implantation rate between the two groups was signicant (Table II). Endometrial thickness on the day of embryo transfer was similar in both groups (Table II). Discussion Previously published data have shown that adverse paternal effects on embryo viability can be detected as early as the rst cell cycle after fertilization, at the pronuclear stage (Tesarik et al., 2002). It is well known that major gene expression in preimplantation human embryos starts two cell cycles later, between the 4-cell and the 8-cell stage (Tesarik et al., 1986, 1988; Braude et al., 1988). Thus, we did not expect DNA fragmentation to be directly involved in the events leading to the development of the morphological abnormalities at the pronuclear stage. However, it could not be excluded that the pathological process which ultimately leads to DNA fragmentation also affects other sperm functions, some of which may be related to the early paternal effect. If this were the case, DNA fragmentation could serve as a marker for sperm insufciency leading to the early adverse paternal effect on embryo 613

Results Early paternal effect Similar to the three previous ART attempts with donor oocytes, the fertilization rate was b50%, but <25% of the fertilized oocytes developed into good morphology zygotes in all treatment attempts involving the patients of the experimental group. This contrasted with the prevalence of good morphology zygotes developing from sibling oocytes from the same donors that were shared by the patients of the control group (Table I).

J.Tesarik, E.Greco and C.Mendoza Table I. Early paternal effect: zygote morphology, endometrial thickness, implantation rate and the percentage of spermatozoa showing DNA fragmentation by TUNEL in treatment attempts performed in 10 couples with a history of repeated failure of ART with donor oocytes (experimental group) and in 10 randomly chosen couples who shared oocytes from the same donors (control group)a Patient group Experimental group Control group %Difference in matched pairsc
aData bSignicantly

Good morphology zygotes 8/76 (10.5%)b 49/74 (66.2%) 55.9 T 4.8

Endometrial thickness on embryo transfer day 10.2 T 0.8 10.4 T 0.9 0.2 T 0.1

%TUNEL-positive sperm nuclei 8.9 T 3.7 8.7 T 3.6 0.2 T 3.5

Implantation rate 1/30 (3.3%)b 11/30 (36.7%) 33.3 T 3.1

are mean T SD. different from control group (P < 0.01). cAverage of differences obtained by subtracting the value for each data point of the experimental group from the corresponding (same donor) value of the control group data point.

Table II. Late paternal effect: zygote morphology, endometrial thickness, implantation rate and the percentage of spermatozoa showing DNA fragmentation by TUNEL in treatment attempts performed in eight couples with a history of repeated failure of ART with donor oocytes (experimental group) and in eight randomly chosen couples who shared oocytes from the same donors (control group)a Patient group Experimental group Control group %Diference in matched pairs
aData bSignicantly

Good morphology zygotes 36/59 (61.0%) 35/59 (59.3%) 1.7 T 4.6

Endometrial thickness on embryo transfer day 10.5 T 0.9 10.4 T 0.8 0.2 T 0.1

%TUNEL-positive sperm nuclei 27.6 T 8.3b 8.5 T 3.7 19.3 T 5.7

Implantation rate 0/24 (0%)b 9/24 (37.5%) 37.5 T 3.4

are mean T SD. different from the sibling oocyte sharer group (P < 0.001). cAverage of differences obtained by subtracting the value for each data point of the experimental group from the corresponding (same donor) value of the control group data point.

development, and its evaluation could thus be used to predict the chance of an eventual ART attempt. However, the data obtained in this study fail to demonstrate an association between sperm DNA fragmentation and the early adverse paternal effect on embryo development. On the other hand, the results of this study show that high percentages of spermatozoa with fragmented DNA are often found in cases of repeated ART failures without any apparent impairment of zygote and cleaving embryo morphology, a condition referred to as late paternal effect in this study. The absence of a relationship between sperm DNA fragmentation and cleaving embryo morphology has also been reported by another recent study (Benchaib et al., 2003). Because embryos were transferred on day 3 after ICSI in this study, it is not clear whether this late paternal effect could inuence blastocyst development if in vitro culture were prolonged to days 56 after ICSI or whether this effect becomes manifest later during implantation and post-implantation development. No differences in endometrial thickness on the day of embryo transfer were found between the experimental and the control group. Moreover, all oocyte recipients had normal serum concentration of insulin, TSH and prolactin, none of them showed elevated serum concentrations of anti-cardiolipin, anti-phospholipid and anti-nuclear antibodies, and none suffered from endometriosis. Even though the existence of a hidden female factor, related to uterine receptivity, cannot be completely excluded, the fact that a sperm-related anomaly (sperm-derived impairment of zygote morphology or sperm DNA fragmentation) was detected in all cases included in the experimental group strongly suggests that implantation failure 614

in the experimental group was caused by a paternal rather than a maternal effect. The principal clinical message of this study can be derived from the observation that the results of ART using ICSI for fertilization can be compromised by paternal effects acting at different times of embryo development and detectable with the use of different diagnostic means. Analysis of sperm DNA fragmentation by the TUNEL assay is a valuable diagnostic method to reveal the paternal origin of unexplained repeated ICSI failures. However, this study also shows that normal values of the TUNEL assay do not exclude the existence of an adverse paternal effect. If the percentage of TUNEL-positive spermatozoa is not increased compared with values found in fertile individuals, the nding of high frequencies of zygote and cleaving embryo morphological abnormalities is sufcient to suspect a paternal origin of ART failure, especially if this condition occurs repeatedly or with oocytes of supposedly good quality, such as in an oocyte donation programme. Moreover, this study also shows that a (late) adverse paternal effect on embryo development can be present even if no morphological abnormalities are apparent at the zygote stage. It is important to note that it is in just such cases that increased percentages of spermatozoa with fragmented DNA can often be detected by the TUNEL assay. In conclusion, sperm DNA fragmentation should be evaluated in cases of unexplained ICSI failure without any apparent impairment of embryo morphology. If embryo morphology is impaired, a paternal factor is to be suspected even if the analysis of sperm DNA fragmentation shows normal values. The validity of these ndings for conventional IVF remains to

Paternal effect and sperm DNA

be evaluated. With regard to the development of future treatment strategies for sperm deciencies responsible for sperm-borne embryo demise, the possibility of the existence of an early and a late paternal factor as two different pathological entities should be taken into consideration. References
Benchaib M, Braun V, Lornage J, Hadj S, Salle B, Lejeune H and Guerin JF (2003) Sperm DNA fragmentation decreases the pregnancy rate in an assisted repreoductive technique. Hum Reprod 18,10231028. Bolton VN, Hawes M, Taylor CT and Parsons JH (1989) Development of spare human preimplantation embryos in vitro: an analysis of the correlations among gross morphology, cleavage rates and development to the blastocyst. J In Vitro Fertil.Embryo Transfer 6,3035. Braude P, Bolton V and Moore S (1988) Human gene expression rst occurs between the four and eight-cell stages of preimplantation development. Nature 332,459461. Filatov MV, Semenova,EV, Vorob'eva OA, Leont'eva OA and Drobchenko EA (1999) Relationship between abnormal sperm chromatin packing and IVF results. Mol Hum Reprod 5,825830. Hammadeh ME, al-Hasani S, Stieber M, Rosenbaum P, Kupker D, Diedrich K and Schmidt W (1996) The effect of chromatin condensation (aniline blue staining) and morphology (strict criteria) of human spermatozoa on fertilization, cleavage and pregnancy rates in an intracytoplasmic sperm injection programme. Hum Reprod 11,24682471. Host E, Lindenberg S and Smidt-Jensen S (2000) The role of DNA strand breaks in human spermatozoa used for IVF and ICSI. Acta Obstet Gynecol Scand 79,559563. Janny L and Menezo YJR (1994) Evidence for a strong paternal effect on human preimplantation embryo development and blastocyst formation. Mol Reprod Dev 38,3642. Larson KL, DeJonge CJ, Barnes AM, Jost LK and Evenson DP (2000) Sperm chromatin structure assay parameters as predictors of failed pregnancy following assisted reproductive techniques. Hum Reprod 15,17171722. Morris ID, Ilott S, Dixon L and Brison DR (2002) The spectrum of DNA damage in human sperm assessed by single cell gel electrophoresis (Comet assay) and its relationship to fertilization and embryo development. Hum Reprod 17,990998. Parinaud J, Mieusset R, Vieitez G, Labal B and Richoilley G (1993) Inuence of sperm parameters on embryo quality. Fertil Steril 60,888892. Sanchez R, Stalf T, Khanaga O, Turley H, Gips H and Schill WB (1996)

Sperm selection methods for intracytoplasmic sperm injection (ICSI) and andrological patients. J Assist Reprod Genet 13,228233. Shoukir Y, Chardonnens D, Campana A and Sakkas D (1998) Blastocyst development from supernumerary embryos after intracytoplasmic sperm injection: a paternal inuence? Hum Reprod 13,16321637. Tesarik J and Greco E (1999) The probability of abnormal preimplantation development can be predicted by a single static observation on pronuclear stage morphology. Hum Reprod 14, 13181323. Tesarik J and Kopecny V (1989) Nucleic acid synthesis and development of human male pronucleus. J Reprod Fertil 86,549558. Tesarik J and Kopecny V (1990) Assembly of the nucleolar precursor bodies in human male pronuclei is correlated with an early RNA synthetic activity. Exp Cell Res 191,153156. Tesarik J and Sousa M (1995) Key elements of a highly efcient intracytoplasmic sperm injection technique: Ca2+ uxes and oocyte cytoplasmic dislocation. Fertil Steril 64,770776. Tesarik J, Junca AM, Hazout A, Aubriot FX, Nathan C, Cohen-Bacrie P, and Dumont-Hassan M (2000) Embryos with high implantation potential after intracytoplasmic sperm injection can be recognized by a simple, noninvasive examination of pronuclear morphology. Hum Reprod 15,1396 1399. Tesarik J, Kopecny V, Plachot M and Mandelbaum J (1986) Activation of nucleolar and extranucleolar RNA synthesis and changes in the ribosomal content of human embryos developing in vitro. J Reprod Fertil 78,463470. Tesarik J, Kopecny V, Plachot M and Mandelbaum J (1988) Early morphological signs of embryonic genome expression in human preimplantation development as revealed by quantitative electron microscopy. Dev Biol 128,1520. Tesarik J, Mendoza C and Greco E (2002) Paternal effects acting during the rst cell cycle of human preimplantation development after ICSI. Hum Reprod 17,184189. Tomsu M, Sharma V and Miller D (2002) Embryo quality and IVF treatment outcomes may correlate with different sperm comet assay parameters. Hum Reprod 17,18561862. Vanderzwalmen P, Bertin-Segal G, Geerts L, Debauche C and Schoysman R (1991) Sperm morphology and IVF pregnancy rate: comparison between Percoll gradient centrifugation and swim-up procedures. Hum Reprod 6,581588. WorldHealthOrganization (1999) WHO Laboratory Manual for the Examination of Human Semen and SpermCervical Mucus Interaction, 4th edn. Cambridge University Press, Cambridge, UK. Submitted on: August 22, 2003; resubmitted on October 20, 2003; accepted on November 19, 2003

615