{"paper_id":"93340d84-5dd9-4792-9d3d-a630db753a03","body_text":"Intracytoplasmic Sperm injection (ICSI) brings\nan operative technology in the fields of assisted\nreproduction technology (ART) ( 1 ). Nowadays,\nthis way becomes one of the most important and\nefficient treatment approaches which is used by\ninfertility clinics. One of problems in infertility\ncenters is low rate of pregnancy in ICSI cycles\n( 2 ). There is still an ongoing debate among reproductive\nembryologists and endocrinologists\nabout using ICSI for the treatment of the infertile\ncouples. There is a general agreement that\nICSI has become a gold standard technique for\nthe treatment of male factor infertility ( 3 ,  4 ),\nbut many physicians recommend ICSI or invitro\nfertilization (IVF) to patients with tubal\nfactor infertility ( 5 ). However, there is controversy\nabout the replacement of ICSI with IVF in\nwomen with sever tubal factor infertility ( 5 ,  6 ).\nIt could be helpful in planning strategies for the\ntreatment of infertile couples with unexplained\ninfertility ( 7 ,  8 ). Although ICSI has been applied\nfor the male factor infertility more than\ntubal factor infertility and endometriosis, it may\nbe a useful technique to overcome these types\nof fertilization defect in women ( 9 ,  10 ).\nMultiple studies have also showed a higher\nrisk of congenital abnormalities, cardiovascular,\nmusculoskeletal defect, low birth weight, preterm\ndelivery and increase perinatal mortality\nin IVF/ ICSI offspring ( 11 - 15 ). Nevertheless,\nthere is an imperfect image of the risk factors\nassociated with the application of ICSI about\noffspring. Although this technique is accepted\nas a standard routine for couples with different\ncauses of infertility ( 11 ). There are no data to\nsuggest that ICSI is a proper method in order to\napply for all ART cases. Since ICSI has a great\ncost to both couple and the health care system,\nit is necessary to assess its efficacy ( 3 ).\nDifferent success rate of ICSI has also been observed\nin various causes of infertility. Our prediction\nmodel has been developed in reproductive\nmedicine to help gynecologists in assessing the\nchances of pregnancy following ICSI. With these\nmodels, gynecologists can calculate the probability\nof a treatment pregnancy as well as the probability\nof pregnancy success with ICSI. In this study,\nwe evaluated the relation between ICSI outcome and special cause of infertility. We also aimed to\nexamine parameters that might predict pregnancy\nsuccess rate following ICSI.\n\nThis cross sectional study included 1492 infertile\nwomen referred to Infertility Center of Royan\nInstitute between 2010 and 2011.We assigned\ntwo groups including pregnant (n=504) and nonpregnant\n(n=988), while all participants underwent\nICSI cycles.\nInformation about age, menstrual duration,\nnumber of previous cycle, endometrial thickness,\nnumber of embryo transfer, embryo quality,\nduration of infertility, type of infertility, cause of\ninfertility (ovulatory factor, unexplained factor,\nmale factor...) were obtained from patient’s file,\nthen the collected data were analyzed and compared\nbetween the two groups. Inclusion criteria\nwere as follows: male infertility, ovarian infertility\n(including polycystic ovary syndrome (PCOS)\nand diminished ovarian reserve), tubal infertility,\nunexplained infertility, recurrent abortion, and endometriosis.\nThe couples with testicular atrophy, anatomical\nabnormalities, infection, uterine fibroids,\nsystemic disease and history of ICSI/IVF failure\nmore than three times were excluded from the\nstudy. In all participants, serum follicle-stimulating\nhormone (FSH; Pishtaz-Tab, Tehran,\nIran) and luteinizing hormone (LH; Pishtaz-\nTab, Tehran, Iran) were measured on day 3 of\nthe cycle preceding ovarian stimulation. The\novarian stimulation protocol for all patients\nwas performed according to the standard long\nprotocol. Both groups started with Bucereline\nacetate (Superfact; Aventis Pharma Deutshlan,\nFrankfurt, Germany) 500 μg on day 21 of the\nprevious cycle and continued daily until the day\nof hCG administration. After ovarian stimulation,\ninjection of 10000 IU of human chorionic\ngonadotropin (hCG; Choriomon; IBSA, Switzerland)\nwas given when at least two follicles\n≥18 mm were detected. Transvaginal follicular\naspiration was performed under ultrasound\nguidance, 34-36 hours after the administration\nof hCG. Afterwards, ICSI cycle was completed.\nAll patients received luteal phase support\nthrough daily administration of 100 mg natural\nprogesterone (sterop Laboratories, Brussls, Belgium) and progesterone up to 8 week of gestation.\nA clinical pregnancy was confirmed by\nthe observation of gestational sac in ultrasonography.\nThis study was approved by the Royan\nEthics Committee.\nIn addition, all subjects agreed to participate in\nthe study were required to sign a consent form approved\nby the Royan Ethics Committee.\nIn order to build a prediction model, we used\nbackward logistic regression analysis, in which\na p-value of 0.15 was used as an entry criterion,\nwhereas a p-value of 0.10 was the threshold for\na variable to stay in the model. The performance\nof the model was calculated as the area under the\nreceiver operating characteristic (ROC) area under\nroc curve (AUC). An AUC of 0.5 indicates no discriminative\nperformance, whereas an AUC of 1.0\nindicates perfect discrimination.\nCalibration of the model was assessed by\ncomparing the predicted probability of pregnancy\nin a category of patients with the observed\npercentage of pregnant woman in the same category.\nWe first categorized the predicted probabilities\nof pregnancy in 10 groups, and then\nwe compared the mean predicted probability\nof pregnancy within a group with the observed\nprobability of the same group.\nAll statistical analysis was performed by SPSS\nprogram (Version 18; USA). Chi-square and t test\nwere used for analysis. Also, In order to predict\nthe result of ICSI, we used logistic regression. The\ndata were expressed as means ± standard deviation\n(SD). Odds ratio (OR) and 95% confidence interval\n(95% CI) were also calculated for each factor.\nThe value of p<0.05 was considered to be statistically\nsignificant.\n\nIn this study, 1492 women were enrolled. The\nmean maternal age was 32.3 ± 5.3 years, while the\nmean duration of infertility was 7.2 ± 5.07 years. Of\n1492 women, 1172 (78.5%) individuals were with\nprimary infertility, while 320 (21.5%) individuals\nwere with secondary infertility. Overall, 59.8% of\npatients had previous treatment for infertility, and\nwe also found a significant reduction of pregnancy\nrate in the group with previously failed ICSI attempts.\nThe general characteristics of all participants\nundergoing ICSI, divided into pregnant and\nnon-pregnant groups,are provided in table 1.\nOur findings confirmed that the pregnancy\nrate reduced when the woman’s age increased\n(OR=0.93, 95% CI=0.91-0.95). In addition, our\nresult revealed that pregnancy rate were lower in\nprimary infertility than secondary infertility, but it\nis n’t significant. Our results also showed no significant\neffect of body mass index (BMI) on pregnancy\nrate ( Table 1 ).\nCharacteristics of women undergoing ICSI in two groups of pregnant and non-pregnant\n*; OR=Odds ratio, CI=Confidence interval, **; Values are mean ± SD, ***; Reference category, a; Independent sample t test\nand b; Chi-square test.\nTable 2 indicates the different causes of infertility\nand their likelihood of occurrence, like\novulatory factor (7.4%), tuboperitoneal factor\n(5.2%), unexplained factor (10%), male factor\n(59.1%), recurrent abortion (2.1%), uterine factor\n(0.4%), Mix (14%), and others (impotency,\nvaginismus, genetic disorder, etc) (1.7%). Furthermore,\ntable 2 shows outcome of ICSI cycles\nin different causes of infertility, where as\nthere was no statistically significant difference\nbetween groups.\nCycle characteristics are depicted in table 3.\nWe found total dose of gonadotropin of nonpregnant\ngroup to be significantly higher than\nthat of pregnant group (p<0.0001). Table 3 reveals\nthat there was statistically significant difference\nbetween the pregnant and non-pregnant\ngroups in endometrial thickness (p<0.05). The\nnumber of retrieved metaphase II (MII) oocytes\nwas significantly higher in pregnant group than\nthat in non-pregnant group. There was no significant\ndifference between two groups in the\nmean serum concentrations on day 3 after application\nof the following hormones: FSH (7.04 ±\n3.43 IU/ml), LH (5.51 ± 4.06 IU/ml), TSH (2.35\n± 1.81 IU/ml), metoclopramide-stimulated prolactin\n(PRL) (163.40 ± 264.82 IU/ml) ( Table 3 ).\nSuccess rate of ICSI outcome in infertile couples with different cause of infertility\n*; Impotency, vaginismus, genetic disorder and a; Chi-square test.\nTable 3: Cycle parameters of the patients undergo ICSI\n* ; OR=Odds ratio, CI=Confidence interval, **; Values are mean ± SD and a; Chi-square test.\nHowever, there was statistically significant\ndifference in the mean serum level on day 3 after\napplication of LH between the pregnant and\nthe non-pregnant groups. There were no statistically\nsignificant differences between groups\nin number of MII oocytes, embryo transfer and\nnumber of good embryo (grade A, B, AB) ( Table 3 ). No significant difference was also observed\nbetween the pregnant and non-pregnant\ngroups in the number of metaphase I (MI) (OR=\n0.97, CI=0.85-1.1; p=0.68). Fertilization rate\nin the pregnant and non-pregnant groups was\n71 and 68%, respectively (OR=1.31, CI=0.92-\n1.89; p=0.13) ( Table 3 ).\nFinally, in order to build a prediction model\nand find the most important factors that affect\npregnancy rate, we used a logistics regression\nmodel in a backward manner. Table 4 shows\nthe result of fitting logistic regression model\nto the data.\nAge, menstrual duration, number of previous\ncycle, endometrial thickness, number of embryo\ntransfer, and embryo quality in the logistic regression\nmodel were significantly associated with\npregnancy outcome. Age and number of previous\ncycle were negatively associated with pregnancy\noutcome, while menstrual duration, endometrial\nthickness, embryo quality and number of embryos\ntransferred were positively associated with pregnancy\noutcome ( Table 4 ).\nTable 4: Result of logistic regression analysis\n* OR; Odds ratio and CI; Confidence interval.\nAUC shows the discriminative performance of\nthe logistic model. The AUC of 0.5 shows no discriminative\nperformance, while AUC of 1.0 indicates\nperfect discrimination. The AUC for the fitted\nlogistic model was 0.681 (95% CI=0.653-0.709)\nthat shows good predictive performance ( Fig 1 ).\nFigure 2 indicates the calibration of the prediction\nmodel for pregnancy after ICSI. The predictive\nperformance of the model is acceptable because the\n95% confidence intervals of the observed pregnancy\nrates overlap with the predicted pregnancy rates.\nROC curve for assessment discrimimative preformance\nof logistic regression.\nCalibration plot, showing the relationship between\npredicted and observed rate of pregnancy after ICSI.\n\nAlthough ICSI is originally developed to treat\nmale infertility, it has been used for infertile couples\nwith different causes of infertility ( 1 ). We\nevaluated the relation between ICSI outcome and\ndifferent causes of infertility, while our obtained\ndata illustrated the different success rate of ICSI\nin various causes of infertility ( Table 2 ). Our data\nalso supported maternal age as an important predictor\nfor having a successful outcome in ICSI,\nwhich is found to bein agreement with results of\nsome studies ( 16 - 20 ), while contradicts to a study\nof Spandorfer et al. in which they have found no\nimpact of paternal age on ICSI success rates ( 21 ).\nIn the present study, endometrial thickness was\nregarded as prognostic parameters for successful\npregnancy in ICSI, which was in agreement with\nseveral studies showing endometrial thickness was\ngreater in cycles resulting in pregnancy than in cycles\nnot resulting in pregnancy ( 16 ,  19 ,  20 ,  22 ,  23 ).\nIn addition, other studies have showed increased\nendometrial thickness is not associated with decreased\npregnancy rates in ART treatment ( 24 ,  25 ).\nIn our study, no significant difference was\nfound in BMI between pregnant and not-pregnant\ngroups. Usoniene et al. reported the same results\nof pregnancy rate for different BMI groups ( 19 ).\nThe other studies indicated that when BMI was\nhigher than 25 kg/m 2 , the pregnancy rate was significantly\nlower ( 26 ,  27 ). Non-PCOS patients undergoing\nIVF/ICSI showed an increase in BMI independently\nof age, LH, FSH, as well as duration\nand type of infertility, whereas these factors affect\npregnancy rate, significantly ( 28 ).\nIn our study, mean LH serum concentration in\npregnant group were significantly higher than\nthat in non-pregnant group (OR=1.02, CI=1-1.05;\np=0.04). At the beginning of stimulation, high LH\nconcentration can lead to increased endometrial\nmaturation at oocyte pick-up ( 29 ). Balasch et al.\nfound LH is not necessary for follicular growth, but\nexternally administered LH possibly shows a primary\nrole in complete maturity of the oocyte and follicle\n( 30 ). LH is required for normal folliculogenesis,\nwhile low LH concentrations on day 3 could lead to\na poor ovarian response ( 31 ). However, basal LH\nand E2 levels are not considered as proper factors in\norder to distinguish the infertile patients responding\ndifferently to ovarian stimulation ( 32 ).\nHistory of menstrual cycle length (MCL) will be\nused as a simple sign of ovarian reserve, which is\nprimarily determined by the growth rates and quality\nof ovarian follicles. Our result showed women\nwith menstrual duration more than 6 days had more\nchance of pregnancy than those with cycles less\nthan 6 days (OR=1.12, CI=1.04-1.21; p=0.002).\nShortening of MCL causes an abbreviated follicular\nphase, but in general, luteal phase length is preserved\n( 33 ). The shorter follicular phase is associated\nwith a decrease in inhibin B and an increase\nin FSH, which could be due to smaller number of\nantral follicles ( 33 ).\nTomas et al. found the influence of embryo\nquality in pregnancy prediction ( 34 ). Our result\nshowed higher pregnancy rate in the best quality\nof embryo transferred. Different reports have also\nrevealed that ICSI may be performed successfully\nincases with a history of fertilization failure ( 35 ).\nWe also found that women with repeated fertilization\nfailure were at higher risk of pregnancy loss\n(Tables  2 ,  4 ).\nEsinler et al. showed that larger doses of gonadotropin\nare required for overweight and obese\nwomen ( 36 ). Our findings also revealed the significant\neffects of total dose of gonadotropin and\nendometrial thickness on outcome of pregnancies\nconceived by ICSI.\nIt is evident that various factors may influence\nthe outcome of ICSI. Kovacs et al. showed, women\nwho became pregnant after ART showed thicker\nendometrium, better quality of embryo, as well\nas more follicles, oocytes and embryos ( 25 ) as we\nobserved in our study. In this study, we examined\nvariable parameters in patients with different causes\nof infertility and predicted pregnancy success\nrates following ICSI. In addition, patient characteristics,\ntotal dose of gonadotropin, endometrial\nthickness, number of previous cycle and quality\nof embryo transferred were evaluated as predictors\nof success rates following ICSI. Our findings\nconfirmed that larger doses of gonadotropin, a decrease\nin endometrial thickness and in appropriate\nembryo quality reduced significantly pregnancy\nsuccess rate. Our results also indicated that ICSI\nin an effective option in couples with different\ncauses of infertility. We did not find an association\nbetween cause of infertility and clinical outcomes.\nThe overall pregnancy rate in our study was 33.9%\n(n=1492).\nUndoubtedly, this database was not large enough\nto allow definite conclusion, and needed further\nsupports to continue the follow-up of pregnancy\noutcome after ICSI.\nOur models can be used reliably as a guide for\nmaking decisions for fertility management in infertile\npatients. The effects of using these models\nin patient care need further experimental investigation.\n\nICSI is an important treatment option for various\nindications of infertility. The present study\nshowed that pregnancy rate affected by the number\nof previous cycle, total dose gonadotropin,\nendometrial thickness, number of previous cycle,\nquality of embryo transferred and menstrual duration.\nIt is required that each infertility center gather\nenough information about the causes of infertility\nin order to provide more information and better\nassistance to patients. Therefore, we suggest that\nphysician sprepare adequate training and required\ninformation regarding these procedures for infertile\ncouples in order to improve their knowledge.","source_license":"CC-BY-4.0","license_restricted":false}