statistical analysis of factors affecting fertilization rates and clinical outcome associated with...

Statistical analysis of factors affectingfertilization rates and clinical outcomeassociated with intracytoplasmic sperminjection

Shehua Shen, M.D.,a Amin Khabani, B.S.,b Nancy Klein, M.D.,b andDavid Battaglia, Ph.D.c

University of California, San Francisco, San Francisco, California

Objective: To identify and evaluate the statistically significant predictors of intracytoplasmic sperm injection(ICSI) fertilization rates and clinical pregnancy in a single population using appropriate statistical techniques.

Design: Retrospective study.

Setting: Fertility and Endocrinology Center, University of Washington Medical Center, Seattle, Washington.

Patient(s): Four hundred forty-one patients undergoing their first attempt at IVF-ICSI from January 1, 1999,to May 21, 2001.

Intervention(s): Each ICSI procedure for an individual patient was performed by a single operator. Spermparameters, oocyte age, culture condition, ICSI technique, and ICSI operator were assessed as variablesinfluencing the fertilization rate. We also assessed the impact of patient age, serum E2 concentration on theday of hCG administration, embryo quality, and number of embryos transferred on the probability of achievinga clinical pregnancy.

Main Outcome Measure(s): Fertilization rate and clinical pregnancy.

Result(s): The 2 pronuclei (2PN) rate was significantly correlated with sperm motility, and there weresignificant differences in the 2PN rates among the ICSI operators. There was no difference in the 2PNrate among different sperm types or among the eight laboratory incubators or whether the eggs werecultured individually or in groups. Patient age, serum E2 concentration on the day of hCG administration,embryo quality, and number of embryos transferred were all statistically significant predictors of clinicalpregnancy.

Conclusion(s): In our program, sperm motility and ICSI operator are the two most important predictors forthe ICSI fertilization rate in vitro. Patient age, serum E2 concentration on the day of hCG administration,embryo quality, and number of embryos transferred were all statistically significant predictors of clinicalpregnancy. (Fertil Steril� 2003;79:355–60. ©2003 by American Society for Reproductive Medicine.)

Key Words: ICSI, fertilization rates, ICSI technique, ICSI operator, serum estradiol, age, embryo quality,number of embryos transferred, clinical pregnancy

Intracytoplasmic sperm injection (ICSI) isa powerful tool to overcome male factor andprior (or risk of) fertilization failure. Thesuccess rate of ICSI is thought to be influ-enced by multiple factors, which include eggquality (1–4), sperm quality (5, 6), ICSI tech-nique (7–19), injection pipette (8), and qualityand number of transferred embryos (20). How-ever, the variation among fully trained opera-tors has not been well assessed, and no previ-ous study has evaluated the simultaneousimpact of all these variables on ICSI success ina single population.

MATERIALS AND METHODS

There were a total of 552 ICSI cases withinthe 3-year interval. We restricted our investi-gation to the first attempt to avoid statisticalproblems with autocorrelation. This yielded441 cycles that met the study criterion. Sincethe study only involved retrospective analysisof laboratory data and no patient identificationwas used, institutional review board approvalwas not necessary.

Indications for ICSI in our clinic are malefactor infertility, unexplained infertility, or pre-

Received February 7, 2002;revised and acceptedAugust 7, 2002.Reprint requests: ShehuaShen, M.D., University ofCalifornia, San Francisco,Center for ReproductiveHealth, 2356 Sutter Street,8th floor, San Francisco,California 94115 (FAX: 415-353-3058; E-mail: [email protected]).a Department of Obstetrics,Gynecology andReproductive Sciences,University of California,San Francisco.b Department of Obstetricsand Gynecology, Universityof Washington, Seattle,Washington.c Department of Obstetricsand Gynecology, OregonHealth and ScienceUniversity, Portland,Oregon.

FERTILITY AND STERILITY�VOL. 79, NO. 2, FEBRUARY 2003Copyright ©2003 American Society for Reproductive MedicinePublished by Elsevier Science Inc.Printed on acid-free paper in U.S.A.

0015-0282/03/$30.00doi:10.1016/S0015-0282(02)04675-7

355

vious fertilization failure (or poor fertilization) with IVF.Patients were treated with the standard IVF protocol. Onlythe oocytes with the presence of the first polar body wereinjected.

After a normal sperm was immobilized, the next step ofthe injection procedure was accomplished by two alternativetechniques. In technique 1, a gentle suction was applied torupture the oolemma. In this method, the membrane wasbroken inside the pipette. In technique 2, the spike on the tipof the needle was used to penetrate the membrane in a lineparallel to the first invagination. The membrane was brokenoutside the pipette.

At 16–18 hours after insemination, 2 pronuclei (2PN)were assessed and then placed in a fresh dish either individ-ually or in a group of 4–6 for culture until transfer.

Embryos were scored on day 3 according to a 1 to 5scoring system (with 5 being the best), which is based onfragmentation and cell symmetry (21). A final score for eachembryo was calculated by multiplying cell number by thescore and dividing by 3. Only cases with observed fetalheartbeats at the 7-week ultrasound examination were con-sidered to have resulted in a clinical pregnancy.

All statistical analyses were performed using SPSS ver-sion 10.0 (SPSS Inc., Chicago, IL). One-way analysis ofvariance with Scheffe’s test and covariance analysis wereutilized to evaluate differences among the groups for con-tinuously distributed variables. Regression and correlationanalyses were used to investigate relationships among con-tinuously distributed variables. Logistic regression was usedto detect effects on the ICSI pregnancy rate as described byWheeler et al. (22) for IVF.

RESULTSAnalysis of variance (Table 1) showed that there were

significant differences in the 2PN rate among the four ICSI

operators, with operators 3 and 4 having 2PN rates that weresignificantly better than operator 1 (P � .019 and P � .008,respectively, using Scheffe’s post hoc test). None of theother pair-wise comparisons among the four operatorsreached statistical significance using Scheffe’s post hoc test.Furthermore, there was no significant difference in spermmotility among the ICSI operators (F � 0.926 with 3 and436 degrees of freedom, P � .428, data not shown), andmultiple linear regression analysis using sperm motility andthree 0/1 “dummy” variables (for ICSI operators 2, 3, and 4)as predictors yielded the equation

Predicted 2PN rate � 62.803 � 7.331 �operator 2�

� 8.691 �operator 3� � 13.562 �operator 4�

� 0.122 �Sperm Motility),

with the 2PN rate expressed as a percentage and R2 � 0.048.

This equation showed that the 2PN rate increased by0.122% for each 1% increase above 0 in sperm motility (P �.008), that operators 2, 3, and 4 had 2PN rates that were7.331%, 8.691%, and 13.562%, respectively, above thatobserved for operator 1 after adjustment for sperm motility,and that the adjusted rate for operator 1 was 62.803%. Therewere no other statistically significant predictors of the 2PNrate. (The number of eggs on which ICSI was performed, thenumber of eggs retrieved, the age of the individual fromwhom the eggs were retrieved, the sperm count, the durationof gonadotropin administration, and the age of patient wereall not statistically significant predictors of the 2PN rate.)There was no difference in the 2PN rate among the fivesperm types (which included partner’s fresh ejaculates, part-ner’s frozen ejaculates, testicular sperm, epididymal sperm,and frozen donor sperm) (P � .087) or among the eightlaboratory incubators (P � .58). There was no difference inthe 2PN rate between eggs cultured individually and ingroups (P � .34).

The logistic regression analyses were restricted to the 246cases that met the following criteria: all embryos incubatedin groups, ET on day 3, and the recipient’s own eggs used.Age of patient, serum E2 concentration on the day of hCGadministration, number of embryos transferred, average em-bryo score, embryo score of the lead embryo, and the totalembryo score were evaluated as predictors of whether aviable pregnancy occurred (as confirmed by detection of abeating heart at the 7-week ultrasound examination). Logis-tic regression analysis showed that before addition of thefirst predictor to the model, patient age (P � .027), E2 (P �.001), embryo quality (average embryo score [P � .003]),lead embryo score (P � .012), and total embryo score(P�.001) were statistically significant predictors, but thenumber of embryos transferred (P � .154) was not a statis-tically significant predictor.

T A B L E 1

Two-pronuclei (2PN) fertilization rates (as percent)among ICSI operators.

ICSIoperators No. of cases 2PN rate (mean � SD)

Total 441 77 � 201 73 69 � 232 75 76 � 203 237 78 � 20a (P � .019)4 46 83 � 20a (P � .008)a Analysis of variance showed there were significant differences in the 2PNrate among the four ICSI operators (F � 4.831 with 3 and 437 d.f., P �.003, R2 � .032) with values for operators 3 and 4 significantly higher thanthose of operator 1 by Scheffe’s test. There was no evidence for heteroge-neity of variance as measured by the Levene statistic (P � .20).

Shen. Factors affecting ICSI results. Fertil Steril 2003.

356 Shen et al. Factors affecting ICSI results Vol. 79, No. 2, February 2003

The two best models that did not include E2 were models1 and 2 (Table 2). The two best models that did include E2

were models 3 and 4 (Table 2). When the categorical vari-ables for the ICSI operators were forced into models 2 or 4,the ICSI operators were not statistically significant predic-tors for the occurrence of viable pregnancy. Attempting toadd additional predictors or the interaction predictors tomodel 4 showed that the added predictors failed to achievestatistical significance.

Pearson correlation analysis showed that log (E2) wassignificantly correlated with the age of the women (R2 ��0.222), the number of eggs retrieved (R2 � 0.732), and theimplantation rate (R2 � 0.254) (P�.001, n � 246 for each).Replacing E2 by egg number in model 4 reduced theNagelkerke R2 from 0.161 to 0.149, clearly showing that E2

was a better predictor than egg number.

To evaluate the relationship between the E2 on the day ofhCG administration and the subsequent probability of

achieving a pregnancy, we rank ordered the 246 cases usedfor the logistic regression analysis based on the E2 concen-tration. We then split the 246 cases into six equal groupsconsisting of 41 (246/6) cases based on the E2 ranking.Because of a tie for E2 rank 41 and 42, group 1 consists of40 cases and group 2 of 42 cases. No other exceptions werenecessary. As shown above by the logistic regression, thepregnancy rate tended to increase with increasing E2 con-centration in the six groups (Table 3).

Likewise, the mean implantation rate also increased withincreasing E2 concentration in the six groups. Analysis ofvariance showed that there were statistically significant dif-ferences among the means for age of patient (P � .028),number of eggs (P�.001), sperm motility (P � .02), numberof eggs injected (P�.001), and implantation rate (P � .005)(Table 3). The pregnancy with the lowest observed E2 con-centration was E2 rank order 12 at 566 pg/mL, and thepregnancy with the highest observed E2 concentration was

T A B L E 2

Four logistic regression models for predicting the probability of conception.

Model Constant AgeTotalscore

Averagescore

No.transferred E2

NagelkerkeR2

1 1.616 �0.104a 0.077b — — — 0.0892 �0.025 �0.110a — 0.203b 0.683b — 0.1373 6.77 �0.089a 0.068b — — 0.000334c 0.1524 �0.727 �0.100a — 0.168a 0.682b 0.000357c 0.161a P�.005.b P�.001.c P�.05.


T A B L E 3

Tabulation of IVF parameters by six E2 groups.

E2 groups 1 2 3 4 5 6 P

E2 value (pg/mL) 196–1,013 1,022–1,267 1,283–1,668 1,675–2,037 2,055–2,765 2,722–5,103No. of cases 40 42 41 41 41 41Age of patient 36.8 � 4.6 36.0 � 4.2 34.6 � 4.3 34.4�4.8 34.2 � 4.3 33.9 � 4.6 �.05Stimulation days 11.3 � 2.0 11.2 � 1.5 11.2 � 1.8 10.7 � 1.7 10.9 � 1.6 11.1 � 1.3 NSNo. of eggs 5.6 � 3.64-6 8.8 � 3.44-6 9.5 � 4.11,4-6 14.1 � 5.21-3,6 17.6 � 5.41-3 20.4 � 6.81-4 �.001Sperm motility (%) 51.2 � 20.5 59.0 � 16.3 56.5 � 20.9 46.0 � 22.8 47.9 � 22.2 56.2 � 18.6 �.05No. of eggs injected 4.5 � 2.84-6 6.8 � 2.54-6 7.4 � 3.34-6 11.1 � 4.81-3,6 12.9 � 4.11-3,6 16.5 � 6.31-5 �.0012PN rate (%) 81.4 � 18.6 73.1 � 19.1 78.0 � 19.5 76.7 � 18.6 78.1 � 17.3 75.7 � 19.3 NSNo. of ETs 2.5 � 1.1 2.8 � 0.8 2.5 � 0.8 2.7 � 1.0 2.4 � 0.6 2.4 � 0.7 NSClinical pregnancy rate

(%)20.0 40.5 39.0 41.5 48.8 58.5

Implantation rate (%) 12.0 19.0 20.0 25.9 26.2 40.61

Note: Only one case had an E2 value over 5,000 pg/mL; P values are for the F test from the analysis of variance. Superscripts show that the value wassignificantly different from the indicated group number at P�.05 (with no further breakdown) after using Scheffe’s test. Lack of superscripts indicates thatno significant differences among the groups were found after application of Scheffe’s test. Analysis of variance was not applied to the clinical pregnancy rate.


FERTILITY & STERILITY� 357

E2 rank order 245 at 4,942 pg/mL. Three of the 4 cases withvalues above 4,000 pg/mL resulted in pregnancy, and 12 ofthe 21 cases in the range 3,006–3,955 pg/mL resulted inpregnancies. The equation for model 4 was used to constructthe table of predicted probability of pregnancy (Table 4).

DISCUSSIONWe examined many factors that might have an impact on

the ICSI fertilization rate to find the crucial predictors forICSI outcome. In our ICSI program, each ICSI procedurewas performed by a single operator, from oocyte preparationto performing injections, which allowed us to study the ICSIoperator as a variable that potentially influences the ICSIoutcome. All four ICSI operators were well trained, and eachof them had at least 5 years of experience in ICSI, whichensured that we were studying the impact of operator ortechnique uninfluenced by the learning curve of trainees.

We found that one of the two important laboratory vari-ables affecting the 2PN rate was the ICSI operator. It waspreviously reported that the rate of development to the blas-tocyct stage is related to the person performing the injection(23), and higher pregnancy rates were observed when theeggs were injected with the polar body at 6 o’clock (whichwas unfortunately confounded with technical differencesamong ICSI operators) (11). In our laboratory, operator 1used both techniques 1 and 2 to break the oolemma. Oper-ators 3 and 4 changed from technique 1 to 2 (no statisticaldifference in 2PN rate was found after the change), andoperator 4 used technique 2 exclusively. The polar body wasalways located at 6 o’clock for all the injections done by

operator 4, and the injection took place in the upper half ofthe oocyte, which might decrease the risk of damaging thespindle. The other three operators located the polar body at6 or 12 o’clock. Our study confirmed that there are signifi-cant differences between ICSI operators in 2PN rates, but wewere unable to show that the differences were due to the twodifferent injection techniques. Monitoring the 2PN rate is animportant and sensitive way to detect factors that will po-tentially affect the success rate of ICSI such as the perfor-mance of the individual ICSI operators.

The other important laboratory variable affecting the 2PNrate was sperm motility. It has been reported that very lowsperm counts resulted in lower 2PN rates and high preg-nancy loss in the cryptozoopermic population (5). Our datadid not show any effect of sperm counts on the 2PN rate. Theeffect of sperm motility is measured by its multiple linearregression coefficient of 0.1223. Analysis of the multiplelinear regression equation showed that an increase of 20% insperm motility (approximately 1 SD) would increase the2PN rate by 2.4%. However, this effect is very small relativeto the observed differences among operators (7.3%, 8.7%,and 13.6%).

We used the outcome variable of viable intrauterine preg-nancy to detect factors affecting the pregnancy rate fromICSI only. Our findings confirm the analysis of Wheeler etal. (22) (conventional IVF data only) that patient age and thetotal embryo score were two predictors when the E2 variablewas not considered. In our ICSI data set, we found thataverage embryo score and number of embryos transferredwere somewhat better as separate predictors than was their

T A B L E 4

Predicted pregnancy rate in different E2 groups.

Patientage

ETno.

For E2 � 1,000 pg/mL For E2 � 2,000 pg/mL For E2 � 3,000 pg/mL

Mean Embryo score Mean embryo score Mean embryo score

3 5 7 9 11 13 3 5 7 9 11 13 3 5 7 9 11 13

28 1 12 16 21 27 34 42 16 21 28 35 43 51 22 28 35 43 52 602 21 27 34 42 51 59 28 35 43 51 60 67 22 28 35 43 52 603 35 43 51 59 67 74 43 52 60 68 74 80 52 60 68 75 81 854 51 60 67 74 80 85 60 68 75 80 85 89 68 75 81 85 89 92

32 1 8 11 15 20 26 33 12 15 20 26 33 41 16 21 27 34 42 502 15 20 26 33 41 49 20 26 33 41 50 58 27 34 42 50 58 663 26 33 41 49 58 66 34 42 50 58 66 73 42 50 59 67 74 804 41 50 58 66 73 79 50 58 66 73 79 84 59 67 74 80 85 89

36 1 6 8 11 14 19 25 8 11 15 19 25 32 11 15 20 25 32 402 11 14 19 25 32 39 15 19 25 32 40 48 20 26 33 40 49 573 19 25 32 39 48 56 25 32 40 48 57 65 33 40 49 57 65 724 32 40 48 56 64 72 40 48 57 65 72 78 49 57 65 72 79 84

40 1 4 5 7 10 14 18 6 8 10 14 18 24 8 10 14 19 24 312 7 10 14 18 24 30 10 14 18 24 31 38 14 19 24 31 39 473 14 18 24 30 38 46 19 24 31 38 47 55 25 31 39 47 55 644 24 31 38 46 55 63 31 39 47 55 63 71 39 47 56 64 71 78



product, the total embryo score. We confirmed that theprobability of achieving a pregnancy increased with increas-ing embryo quality and embryo number and decreased withincreasing age of the patient. Moreover, we also found thatthe probability of achieving a pregnancy increased withincreasing concentrations of serum E2 on the day of hCGadministration.

The implantation rate (fraction of embryos implanting)was significantly higher in the highest E2 group than it wasin the lowest E2 group, and the log of the serum E2 concen-tration on the day of hCG administration is positively cor-related with the implantation rate. These findings providefurther evidence against an adverse effect of high prolifera-tive phase E2 concentrations on the ability of embryos toimplant, a concern that has been a source of controversy(24–29), and actually suggest strongly that high E2 concen-trations increase (rather than decrease) the probability ofpregnancy.

The serum E2 concentration is useful along with the ageof the patient, the number of the embryos transferred, and theaverage embryo score for prospectively predicting the prob-ability of pregnancy since its addition to the logistic regres-sion equation increased the Nagelkerke R2, the statisticalmeasure for logistic regression most closely analogous to R2

in the standard linear regression. Furthermore, both patientage and the E2 variable were predictors of successful im-plantation, but neither was a statistically significant predictorof the 2PN rate, indicating that the effects of age and estro-gen concentration were mediated after the embryos had beentransferred to the patient’s uterus.

Clearly, logistic regression analysis is the best statisticaltool to sort out the intertwined effects of patient age, serumE2 concentration, embryo quality, and embryo number on theprobability of conception. Logistic regression showed thatthe probability of conception increased as the E2 concentra-tion increased after the effects due to the patient age, theembryo quality, and the number of embryos transferred hadbeen adjusted for. To provide reassurance that the resultfrom the logistic regression analysis above was not based onsome statistical fluke, we provided data on the pregnancyoutcome when the cases were divided into six equal groupsbased on the E2 rank order. This showed that both thepregnancy rate (58.5%) and average implantation rate(40.6%) were the highest in the highest E2 group (41 caseswith serum concentration ranging from 2,772 to 5,013 pg/mL), a range that some reports have claimed was associatedwith reduced implantation and/or pregnancy rates (25, 26).Nor was this effect mediated by a corresponding increase inthe number of embryos transferred since the mean numbersof embryos transferred in the groups were 2.45, 2.83, 2.51,2.68, 2.41, and 2.37, respectively.

The highest two pregnancy rates were observed in thegroups with the two lowest means for numbers of embryostransferred. Thus, our data confirm for ICSI the report of

Chenette et al. (27) based on 141 IVF cycles that the im-plantation and pregnancy rates were highest in the highest E2

group. This means that the medication dosages and treatmentprotocols used for producing follicular growth and matura-tion can be adjusted to optimize oocyte number and oocytequality when E2 is in the range covered by our data (196–4,942 pg/mL). Obviously, based on its magnitude, the effectof E2 should be taken into account when attempting topredict the probability of conception. The predicted proba-bilities in Table 4 are program and time specific and may notaccurately predict the actual probability in other programs orin our program in the future. However, the results may be ofuse for providing a general understanding of the complexinterrelationships among the predictor variables and proba-bility estimates until accurate program-specific data becomeavailable.

Acknowledgments: The authors thank Ray Haning, M.D. (retired), forassistance with the comprehensive statistical analyses.

References1. Serhal PF, Ranieri DM, Kinis A, Marchant S, Davis M, Khadum IM.

Oocyte morphology predicts outcome of intracytoplasmic sperm injec-tion. Hum Reprod 1997;12:1267–70.

2. Xia P. Intracytoplasmic sperm injection: correlation of oocyte gradebased on polar body, perivitelline space and cytoplasmic inclusionswith fertilization rate and embryo quality. Hum Reprod1997;12:1750–5.

3. Alikani M, Palermo G, Adler A, Bertoli M, Blake M, Cohen J. Intra-cytoplasmic sperm injection in dysmorphic human oocytes. Zygote1995;3:283–8.

4. Loutradis D, Drakakis P, Kallianidis K, Milingos S, Dendrinos S,Michalas S. Oocyte morphology corralates with embryo quality andpregnancy rate after intracytoplasmic sperm injection. Fertil Steril1999;72:240–4.

5. Strassburger D, Friedler S, Raziel A, Schachter M, Kasterstein E,Ron-El R. Very low sperm count affects the result of intracytoplasmicsperm injection. J Assist Reprod Genet 2000;17:431–6.

6. Nagy ZP, Verheyen G, Tournaye H, Van Steirteghem AC. Specialapplications of intracytosperm injection: the influence of sperm count,motility, morphology, source and sperm antibody on the outcome ofICSI. Hum Reprod 1998;13:143–54.

7. Catt J, O’Neill C. Manipulation of sperm before intracytoplasmic sperminjection improves fertilization rates. Fertil Steril 1995;64:1210–2.

8. Svalander P, Forsberg AS, Jakobsson AH, Wikland M. Factors ofimportance for the establishment of a successful program of intracyto-plasmic sperm injection treatment for male infertility. Fertil Steril1995;63:828–37.

9. Vanserzwalmen P, Bertin G, Lejeune B, Nijs M, Vandamme B, Schoys-man R. Two essential steps for a successful intracytoplasmic sperminjection: injection of immobilized spermatozoa after rupture of theoolema. Hum Reprod 1996;11:540–7.

10. Hlinka D, Herman M, Zesela J, Hredzak R, Horvath S, Pacin J. Amodified method of intracytoplasmic sperm injection without the use ofpolyvinylpyrolidone. Hum Reprod 1998;13:1922–7.

11. Van der Westerlaken LAJ, Helmerhorst FM, Hermans J, NaaktgeborenN. Intracytoplasmic sperm injection: position of the polar body affectspregnancy rate. Hum Reprod 1999;14:2565–9.

12. Blake M, Garrisi J, Tomkin G, Cohen J. Sperm deposition site duringICSI affects fertilization and development. Fertil Steril 2000;15:31–7.

13. Stoddart NR, Fleming SD. Orientation of the first polar body of theoocyte at 6 or 12 o’clock during ICSI does not affect clinical outcome.Hum Reprod 2000;15:1580–5.

14. Carrillo AJ, Atiee SH, Lane B, Pridham, Risch P, Silverman IH, et al.Oolemma rupture inside the intracytoplasmic sperm injection needlesignificantly improves the fertilization rate and reduces oocyte damage.Fertil Steril 1998;70:676–9.

FERTILITY & STERILITY� 359

15. Joris H, Nagy Z, Van de Velde H, De Vos A, Van Steirteghem A.Intracytoplasmic sperm injection: laboratory set-up and injection pro-cedure. Hum Reprod 1998;13:76–85.

16. Payne D. Intracytoplasmic sperm injection: instrumentation and injec-tion technique. Reprod Fertil Dev 1995;7:185–96.

17. Mansour RT, Aboulghar MA, Serour GI, Tawab NA, Amin Y, SattarMA. Successful intracytoplasmic sperm injection without performingcytoplasmic aspiration. Fertil Steril 1996;66:256–9.

18. Nagy ZP, Liu J, Joris H, Bocken G, Desmet B, Van Ranst H, et al. Theinfluence of the site of sperm deposition and mode of oolemma break-age at intracytoplasmc sperm injection on fertilization and embryodevelopment rates. Hum Reprod 1995;10:3171–7.

19. Dozortsev D, Rybouchkin A, De Sutter P, Dhont M. Sperm plasmamembrane damage prior to intracytoplasmic sperm injection: a neces-sary condition for sperm nucleus decondensation. Hum Reprod 1995;10:2960–4.

20. Hsu M, Mayer J, Aronshon M, Lanzendorf S, Muasher S, Kolm P, et al.Embryo implantation in vitro fertilization and intracytoplasmic injec-tion: impact of cleavage status, morphology grade and number ofembryos transferred. Fertil Steril 1999;72:679–85.

21. Veeck L. An atlas of human gemetes and conceptuses. Parthenon, CRCPress LLC, 1998.

22. Wheeler CA, Cole BF, Frishman GN, Seifer DB, Lovegreen SB,Hackett RJ. Predicting probabilities of pregnancy and multiple gesta-tion from in vitro fertilization—a new model. Obstet Gynecol 1998;91:696–700.

23. Dumoulin J, Coonen E, Bras M, Bergers-Janssen JM, Ignoul-Vanvuchelen R, Van Wissen L, et al. Embryo development and chro-mosomal anomalies after ICSI: effect of the injection procedure. HumReprod 2001;16:306–12.

24. Ng E, Yeung W, Lao E, So W, Ho P. High serum oestradiol concen-trations in fresh IVF cycles do not impair implantation and pregnancyrates in subsequent frozen-thawed embryo transfer cycles. Hum Reprod2000;15:250–5.

25. Simon C, Cano F, Valbuena D, Remohi J, Pellicer A. Clinical evidencefor a detrimental effect on uterine receptivity of high and normalresponder patients. Hum Reprod 1995;10:2432–7.

26. Forman R, Fries N, Testart J, Belaisch-Allart J, Hazout A, Frydman R.Evidence of an adverse effect of elevated serum estradiol concentra-tions on embryo implantation. Fertil Steril 1988;49:118–22.

27. Chenette PE, Sauer MV, Paulson RJ. Very high serum estradiol con-centrations are not detrimental to clinical outcome of in vitro fertiliza-tion. Fertil Steril 1990;61:858–63.

28. Pellicer A, Valbuena D, Cano F, Remohi J, Simon C. Lower implan-tation rates in high responders: evidence for an altered endocrine milieuduring the preimplantation period. Fertil Steril 1996;65:1190–5.

29. Simon C, Velasco J, Valbuena D, Peinado J, Moreno C, Remohi J, etal. Increasing uterine receptivity by decreasing estradiol levels duringthe preimplantation period in high responders with the use of a follicle-stimulating hormone step-down regimen. Fertil Steril 1998;70:234–9.


statistical analysis of factors affecting fertilization rates and clinical outcome associated with...

Documents