• ISSN 16748301
  • CN 32-1810/R
Volume 33 Issue 6
Nov.  2019
Article Contents

Citation:

A nomogram predicting clinical pregnancy in the first fresh embryo transfer for women undergoing in vitro fertilization and intracytoplasmic sperm injection (IVF/ICSI) treatments

  • The extent to which factors affect the probability of clinical pregnancy in the first fresh embryo transfer after assisted conception is unknown. In order to examine the predictors of clinical pregnancy, a retrospective cohort study was launched between January 1, 2013 and December 31, 2016 in four infertility clinics including 19 837 in vitro fertilization and intracytoplasmic sperm injection (IVF/ICSI) fresh cycles with known outcomes and relevant records. A multivariable logistic regression was used to select the most significant predictors in the final nomogram for predicting clinical pregnancy. Furthermore, the model was validated by an independent validation set and the performance of the model was evaluated by the receiver operating characteristic (ROC) curves along with the area under the ROC curve (AUC) and calibration plots. In a training set including 17 854 participants, we identified that female age, tubal factor, number of embryos transferred, endometrial thickness and number of good-quality embryos were independent predictors for clinical pregnancy. We developed a nomogram using these five factors and the predictive ability was 0.66 for AUC (95% CI 0.64–0.68), which was independently validated in the validation set (AUC=0.66, 95% CI 0.65–0.68). Our results show that some specific factors can be used to provide infertile couples with an accurate assessment of clinical pregnancy following assisted conception and facilitate to guide couples and clinicians.
  • 加载中
  • [1] Qiao J, Feng HL. Assisted reproductive technology in China: compliance and non-compliance[J]. Transl Pediatr, 2014, 3(2): 91–97.
    [2] Zhou Z, Zheng D, Wu H, et al. Epidemiology of infertility in China: a population-based study[J]. BJOG, 2018, 125(4): 432–441. doi: 10.1111/1471-0528.14966
    [3] Zarinara A, Zeraati H, Kamali K, et al. Models predicting success of infertility treatment: a systematic review[J]. J Reprod Infertil, 2016, 17(2): 68–81.
    [4] Nelson SM, Lawlor DA. Predicting live birth, preterm delivery, and low birth weight in infants born from in vitro fertilisation: a prospective study of 144,018 treatment cycles[J]. PLoS Med, 2011, 8(1): e1000386. doi: 10.1371/journal.pmed.1000386
    [5] De Neubourg D, Gerris J, Mangelschots K, et al. Single top quality embryo transfer as a model for prediction of early pregnancy outcome[J]. Hum Reprod, 2004, 19(6): 1476–1479. doi: 10.1093/humrep/deh283
    [6] De Mouzon J, Goossens V, Bhattacharya S, et al. Assisted reproductive technology in Europe, 2006: results generated from European registers by ESHRE[J]. Hum Reprod, 2010, 25(8): 1851–1862. doi: 10.1093/humrep/deq124
    [7] McLernon DJ, Steyerberg EW, Te Velde ER, et al. Predicting the chances of a live birth after one or more complete cycles of in vitro fertilisation: population based study of linked cycle data from 113873 women[J]. BMJ, 2016, 355: i5735.
    [8] Clarke JF, van Rumste MME, Farquhar CM, et al. Measuring outcomes in fertility trials: can we rely on clinical pregnancy rates?[J]. Fertil Steril, 2010, 94(5): 1647–1651. doi: 10.1016/j.fertnstert.2009.11.018
    [9] Schieve LA, Peterson HB, Meikle SF, et al. Live-birth rates and multiple-birth risk using in vitro fertilization[J]. JAMA, 1999, 282(19): 1832–1888. doi: 10.1001/jama.282.19.1832
    [10] Wang AC, Wang Y, Wu FX, et al. Assessing predictors for the success of GnRH antagonist protocol in reproductive women in IVF/ICSI - in fresh cycles[J]. Biomed Rep, 2017, 7(5): 482–486. doi: 10.3892/br.2017.984
    [11] Cai QF, Wan F, Huang R, et al. Factors predicting the cumulative outcome of IVF/ICSI treatment: a multivariable analysis of 2450 patients[J]. Hum Reprod, 2011, 26(9): 2532–2540. doi: 10.1093/humrep/der228
    [12] Verberg MFG, Eijkemans MJC, Macklon NS, et al. Predictors of ongoing pregnancy after single-embryo transfer following mild ovarian stimulation for IVF[J]. Fertil Steril, 2008, 89(5): 1159–1165. doi: 10.1016/j.fertnstert.2007.05.020
    [13] Blank C, Wildeboer RR, DeCroo I, et al. Prediction of implantation after blastocyst transfer in in vitro fertilization: a machine-learning perspective[J]. Fertil Steril, 2019, 111(2): 318–326. doi: 10.1016/j.fertnstert.2018.10.030
    [14] Lintsen AME, Eijkemans MJC, Hunault CC, et al. Predicting ongoing pregnancy chances after IVF and ICSI: a national prospective study[J]. Hum Reprod, 2007, 22(9): 2455–2462. doi: 10.1093/humrep/dem183
    [15] Borque Á, Esteban LM, Sanz G, et al. Usefulness of clinical nomograms and predictive models for PCA. Predictive clinical factors of tumor agressiveness[J]. Arch Esp Urol (in Spanish), 2015, 68(3): 267–281.
    [16] Ashmita J, Vikas S, Swati G. The impact of progesterone level on day of hCG injection in IVF cycles on clinical pregnancy rate[J]. J Hum Reprod Sci, 2017, 10(4): 265–270. doi: 10.4103/0974-1208.223278
    [17] Dhillon RK, McLernon DJ, Smith PP, et al. Predicting the chance of live birth for women undergoing IVF: a novel pretreatment counselling tool[J]. Hum Reprod, 2016, 31(1): 84–92. doi: 10.1093/humrep/dev268
    [18] Hansen KR, He ALW, Styer AK, et al. Predictors of pregnancy and live-birth in couples with unexplained infertility after ovarian stimulation-intrauterine insemination[J]. Fertil Steril, 2016, 105(6): 1575–1583.e2. doi: 10.1016/j.fertnstert.2016.02.020
    [19] Broekmans FJ, Knauff EAH, te Velde ER, et al. Female reproductive ageing: current knowledge and future trends[J]. Trends Endocrinol Metab, 2007, 18(2): 58–65. doi: 10.1016/j.tem.2007.01.004
    [20] van Loendersloot LL, van Wely M, Repping S, et al. Individualized decision-making in IVF: calculating the chances of pregnancy[J]. Hum Reprod, 2013, 28(11): 2972–2980. doi: 10.1093/humrep/det315
    [21] Ulug U, Ben-Shlomo I, Turan E, et al. Conception rates following assisted reproduction in poor responder patients: a retrospective study in 300 consecutive cycles[J]. Reprod Biomed Online, 2003, 6(4): 439–443. doi: 10.1016/S1472-6483(10)62164-5
    [22] Kawwass JF, Crawford S, Kissin DM, et al. Tubal factor infertility and perinatal risk after assisted reproductive technology[J]. Obstet Gynecol, 2013, 121(6): 1263–1271. doi: 10.1097/AOG.0b013e31829006d9
    [23] Luke B, Brown MB, Stern JE, et al. Using the Society for Assisted Reproductive Technology Clinic Outcome System morphological measures to predict live birth after assisted reproductive technology[J]. Fertil Steril, 2014, 102(5): 1338–1344. doi: 10.1016/j.fertnstert.2014.07.1242
    [24] Rozen S. Defending male fertility[J]. Sci Transl Med, 2011, 3(92): 92ps31. doi: 10.1126/scitranslmed.3002743
    [25] Palermo G, Joris H, Devroey P, et al. Pregnancies after intracytoplasmic injection of single spermatozoon into an oocyte[J]. Lancet, 1992, 340(8810): 17–18. doi: 10.1016/0140-6736(92)92425-F
    [26] Baker VL, Luke B, Brown MB, et al. Multivariate analysis of factors affecting probability of pregnancy and live birth with in vitro fertilization: an analysis of the society for assisted reproductive technology clinic outcomes reporting system[J]. Fertil Steril, 2010, 94(4): 1410–1416. doi: 10.1016/j.fertnstert.2009.07.986
    [27] Huang Y, Li JY, Zhang F, et al. Factors affecting the live-birth rate in women with diminished ovarian reserve undergoing IVF-ET[J]. Arch Gynecol Obstet, 2018, 298(5): 1017–1027. doi: 10.1007/s00404-018-4884-4
    [28] Pandian Z, Marjoribanks J, Ozturk O, et al. Number of embryos for transfer following in vitro fertilisation or intra-cytoplasmic sperm injection[J]. Cochrane Database Syst Rev, 2013, (7): CD003416.
    [29] Li HWR, Lee VCY, Lau EYL, et al. Role of baseline antral follicle count and anti-Mullerian hormone in prediction of cumulative live birth in the first in vitro fertilisation cycle: a retrospective cohort analysis[J]. PLoS One, 2013, 8(4): e61095. doi: 10.1371/journal.pone.0061095
    [30] Roberts SA, Hirst WM, Brison DR, et al. Embryo and uterine influences on IVF outcomes: an analysis of a UK multi-centre cohort[J]. Hum Reprod, 2010, 25(11): 2792–2802. doi: 10.1093/humrep/deq213
    [31] Dechanet C, Brunet C, Anahory T, et al. Effects of cigarette smoking on female reproduction: from oocyte to embryo (Part I)[J]. Gynecol Obstet Fertil (in French), 2011, 39(10): 559–566. doi: 10.1016/j.gyobfe.2011.07.033
    [32] Rockliff HE, Lightman SL, Rhidian E, et al. A systematic review of psychosocial factors associated with emotional adjustment in in vitro fertilization patients[J]. Hum Reprod Update, 2014, 20(4): 594–613. doi: 10.1093/humupd/dmu010
    [33] Yang XK, Li Y, Li CD, et al. Current overview of pregnancy complications and live-birth outcome of assisted reproductive technology in mainland China[J]. Fertil Steril, 2014, 101(2): 385–391.e2. doi: 10.1016/j.fertnstert.2013.10.017
  • 加载中

Figures(3) / Tables(3)

Article Metrics

Article views(79) PDF downloads(4) Cited by()

Related
Proportional views
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

A nomogram predicting clinical pregnancy in the first fresh embryo transfer for women undergoing in vitro fertilization and intracytoplasmic sperm injection (IVF/ICSI) treatments

    Corresponding author: Zhibin Hu, zhibin_hu@njmu.edu.cn
    Corresponding author: Guangfu Jin, guangfujin@njmu.edu.cn
  • 1. State Key Laboratory of Reproductive Medicine, Nanjing Medical University, Nanjing, Jiangsu 211166, China
  • 2. Department of Epidemiology, School of Public Health, Nanjing Medical University, Nanjing, Jiangsu 211166, China
  • 3. Department of Reproduction, Henan Women and Children Health Hospital, Zhengzhou, Henan 450052, China
  • 4. Department of Reproduction, Shengjing Hospital of China Medical University, Shenyang, Liaoning 110011, China
  • 5. Key Laboratory of Modern Toxicology of Ministry of Education, School of Public Health, Nanjing Medical University, Nanjing, Jiangsu 211166, China
  • 6. Department of Reproduction, the Affiliated Obstetrics and Gynecology Hospital of Nanjing Medical University, Nanjing, Jiangsu 210004, China
  • 7. Department of Reproduction, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China

Abstract: The extent to which factors affect the probability of clinical pregnancy in the first fresh embryo transfer after assisted conception is unknown. In order to examine the predictors of clinical pregnancy, a retrospective cohort study was launched between January 1, 2013 and December 31, 2016 in four infertility clinics including 19 837 in vitro fertilization and intracytoplasmic sperm injection (IVF/ICSI) fresh cycles with known outcomes and relevant records. A multivariable logistic regression was used to select the most significant predictors in the final nomogram for predicting clinical pregnancy. Furthermore, the model was validated by an independent validation set and the performance of the model was evaluated by the receiver operating characteristic (ROC) curves along with the area under the ROC curve (AUC) and calibration plots. In a training set including 17 854 participants, we identified that female age, tubal factor, number of embryos transferred, endometrial thickness and number of good-quality embryos were independent predictors for clinical pregnancy. We developed a nomogram using these five factors and the predictive ability was 0.66 for AUC (95% CI 0.64–0.68), which was independently validated in the validation set (AUC=0.66, 95% CI 0.65–0.68). Our results show that some specific factors can be used to provide infertile couples with an accurate assessment of clinical pregnancy following assisted conception and facilitate to guide couples and clinicians.

    HTML

Introduction
  • Infertility is a global issue and the WHO predicts that it will be the third-most serious disease worldwide during the 21st century, after cancer and cardiovascular disease[1]. In China along, data from a recent epidemiology study reported that 15% –25% of couples in China suffer from infertility and the rate is still on the rise[2]. Steadily increasing number of couples experiencing fertility are turning to assisted reproductive technology (ART) for help, such as in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) treatments. Ideally, the goal of ART treatment is to achieve a successful outcome, however, less than 30% of women can conceive in a single cycle of treatment[3]. As ART has potential effects on female emotion and the treatment is expensive, an accurate prediction model for probability of success is useful to facilitate patient counseling and clinical decision making[4].

    In most cases, the success of IVF and ICSI has conventionally been reported as the live birth rate in a single fresh cycle[45], however, some people suggested that clinical pregnancy could also reflect the success of IVF and ICSI treatment[6]. First, the occurrences of clinical pregnancy performed six months earlier than the occurrences of live birth, thus clinicians and patients may know the outcome of transplantation sooner[7]. Second, the pregnancy loss between pregnancies 12 weeks of gestation and live birth occurs in approximately 5% of clinical pregnancy[8], so clinical pregnancy will result in more precise estimate of the success of IVF/ICSI treatment. More specifically, once clinical pregnancy is confirmed, there is almost no difference in live birth rate between pregnancy conceived through ART and spontaneous pregnancy[9]. Consequently, we choose clinical pregnancy as the primary outcome to establish a prediction model.

    Several studies have developed prediction models for clinical pregnancy[1012], and identified key predictors including female age[5], infertility diagnosis[13], and treatment history[14]. However, most of these studies were based on a single center with limited sample size and the predictors were inconsistent. On the other hand, nomograms are considered useful to quantify a risk by combining several prognostic factors to estimation of an individualized probability[15]. Nevertheless, there is no such research using nomogram to predict the probability of clinical pregnancy. To date, prediction model of clinical pregnancy is not well evaluated in couples with fertility.

    Therefore, the objectives of our present study were to examine IVF/ICSI cycle-dependent predictors of clinical pregnancy and develop a nomogram to predict the probability of clinical pregnancy for women undergoing the first fresh IVF/ICSI non-donor procedure.

  • Materials and methods

      Participants of the study

    • The retrospective cohort was comprised by women whose treatment cycles were initiated between January 1, 2013 and December 31, 2016 at the four fertility clinics: the First Affiliated Hospital of Nanjing Medical University (Nanjing, China), the Affiliated Obstetrics and Gynaecology Hospital of Nanjing Medical University (Nanjing, China), Shengjing Hospital Affiliated to China Medical University (Shenyang, China), and Henan Women and Children Health Hospital (Zhengzhou, China). We excluded these women from analysis according to the following criteria: (i) cycles that resulted in frozen thawed embryos; (ii) subjects had a history of IVF/ICSI treatment; (iii) subjects applied preimplantation genetic diagnosis (PGD), donor insemination, egg donation. Thus, 19 837 women with their first fresh embryo transfer cycle were included in this study (Fig. 1). In order to independently validate the prediction model, we randomly divided the subjects into two groups: 90% of all subjects were used for model construction (training set, n=17 854) and 10% of subjects were used for model validation (validation set, n=1 983).

      Figure 1.  Flowchart of analysis cycles.

    • Screening predictive factors

    • From literature review and clinical experiences, we identified the following variables from medical record as potential predictors. Baseline characteristics of subjects included the female age, body mass index (BMI), duration of infertility (years), type of infertility (primary vs. secondary), infertility diagnosis (categorized as diagnosis of male factor, tubal, ovulation disorders, endometriosis, unexplained). Primary infertility refers to a woman who has never established a pregnancy and secondary infertility is defined as a woman who has a history of one or more previous pregnancies[16]. The definition of unexplained diagnosis was that infertile couple still did not find the cause of infertility after standard examination[16]. Treatment information included fertilization methods (IVF vs. ICSI), endometrial thickness, cleavage (2–3 day embryo) vs. blastocyst (5–6 day embryo), number of embryos transferred, and number of good-quality embryos. Good quality embryo was assessed by the number of blastomeres, evenness of blastomere size, and the proportion of mononucleated blastomeres[16]. The outcome for the model was clinical pregnancy, which was defined as an intrauterine pregnancy with fetal cardiac activity confirmed by transvaginal ultrasound at 12 weeks' gestation[16].

      For each candidate prognostic variable, the association with occurrence of clinical pregnancy was assessed in a logistic regression model. Candidate variables were selected as initial predictors with P value <0.05 in univariable logistic regression. Then all initial predictors were introduced to a multivariate logistic regression analysis in a stepwise procedure, using a P value <0.05 remained for the model. A nomogram was then formulated based on these independent predictive factors in a multivariate logistic regression model and delineated by using the R package 'rms'. The predictive ability of the model was assessed on the independent validation set (10% of subjects), and the model was fitted to the validation set using the same parameter estimates derived from the training set. The performance of the model was evaluated by means of discrimination and calibration. Discrimination of the final model was assessed by AUC. Calibration was studied from graphical representations of the relationship between the observed outcome frequencies and the predicted probabilities, which was assessed by means of Hosmer-Lemeshow test. Statistical analyses were carried out using R software (version 3.2.3, The R Foundation for Statistical Computing, http://www.cran.r-project.org/).

    Results
    • A total of 19 837 subjects who underwent their fresh treatment cycles of IVF/ICSI were included in this study (Table 1); of these, 9 816 (49.48%) subjects became clinical pregnancy. The mean age of the subjects was (30.83±4.64) years, and the mean duration of infertility was (3.90±3.15) years. Infertility was primary and secondary in 41.48% and 58.16% of the subjects, respectively. After randomly divided training (n=17 854, 90%) and validation sets (n=1 983, 10%), the characteristics and clinical pregnancies were very similar between two groups (Table 1).

      Variables All (n=19 837) Training set (n=17 854) Validation set (n=1 983)
      Clinical pregnancy rate[n (%)] 9 816 (49.48) 8 798 (49.28) 1 018 (51.34)
      Female age* (year, mean±SD) 30.83±4.64 30.85±4.65 30.65±4.58
      Female BMI* (kg/m2, mean±SD) 22.70±3.36 22.71±3.36 22.64±3.35
      Duration of infertility (year, mean±SD) 3.90±3.15 3.90±3.15 3.93±3.12
      Type of infertility[n (%)]
       Primary 8 299 (41.84) 7 454 (41.75) 845 (42.61)
       Secondary 11 538 (58.16) 10 400 (58.25) 1 138 (57.39)
      Male factor[n (%)] 8 264 (41.66) 7 439 (41.66) 825 (41.60)
      Tubal factor[n (%)] 12 286 (61.93) 11 046 (61.87) 1 240 (62.53)
      Ovulation disorder[n (%)] 2 020 (10.18) 1 817 (10.18) 203 (10.24)
      Endometriosis[n (%)] 1 178 (5.94) 1 059 (6.75) 119 (6.00)
      Unexplained factor[n (%)] 835 (4.21) 750 (5.93) 85 (3.85)
      Infertility diagnosis (>1 type)[n (%)] 6 182 (31.16) 5 570 (28.05) 612 (30.86)
      Insemination method[n (%)]
       IVF 14 458 (72.88) 12 993 (72.77) 1 465 (73.88)
       ICSI 5 379 (27.12) 4 861 (27.23) 518 (26.12)
      Number of embryos transferred[n (%)]
       1 4 078 (20.56) 3 708 (20.77) 370 (18.66)
       2 14 422 (72.70) 12 933 (72.44) 1 489 (75.09)
       ≥3 1 337 (6.74) 1 213 (6.79) 124 (6.25)
      Embryo stage[n (%)]
       Cleavage 17 252 (86.97) 15 545 (87.07) 1 707 (86.08)
       Blastocyst 728 (3.67) 656 (3.67) 72 (3.63)
       Missing 1 857 (9.36) 1 653 (9.26) 204 (10.29)
      Endometrial thickness (mm) 10.33±2.37 10.32±2.37 10.40±2.34
       Missing[n (%)] 786 (3.96) 717 (4.02) 69 (3.48)
      Total number of good-quality embryos 2.61±2.51 2.60±2.50 2.70±2.61
      Clinical pregnancy rate[n (%)] 9 816 (49.48) 8 798 (49.28) 1 018 (51.34)
      BMI: body mass index; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; OR: odds ratio; CI: confidence interval; *Female age indicates the age of the patient undergoing the ART procedure; Female BMI indicates the BMI of the patient undergoing the ART procedure; Variable contains missing data.

      Table 1.  Baseline characteristics of patients undergoing IVF/ICSI treatment

      In the univariate analysis of the training set, we found a series of variables significantly associated with a decreased chance of clinical pregnancy using IVF/ICSI, including older women, longer duration of subfertility, primary subfertility, tubal factor, three or more embryos transferred, lower endometrial thickness and less good-quality embryos (Table 2). However, female BMI, male factor, endometriosis, unexplained factor and embryo stage were not significantly associated with clinical pregnancy.

      VariablesClinical pregnancy [n (%)]OR (95% CI)  P*
      Yes (n=8 798)No (n=9 056)
      Female age
       <31 year 5 048 (57.38) 4 222 (46.62) 1.000 (reference)
       31–36 year 2 794 (31.76) 2 932 (32.38) 0.797 (0.746–0.851) <0.001
       36–41 year 845 (9.60) 1 338 (14.77) 0.528 (0.480–0.581) <0.001
       ≥41 year 111 (1.26) 564 (6.32) 0.165 (0.134–0.203)  <0.001
      Female BMI
       18.5–25.0 kg/m2 575 (6.54) 614 (6.78) 1.000 (reference)
       <18.5 kg/m2 5 555 (63.14) 5 654 (62.44) 1.049 (0.930–1.182) 0.432
       25.0–30.0 kg/m2 2 051 (23.31) 2 124 (23.45) 1.031 (0.906–1.173) 0.641
       ≥30.0 kg/m2 617 (7.01) 664 (7.33) 0.992 (0.847–1.162) 0.923
      Duration of infertility
       <4 year 3 593 (40.84) 3 638 (40.17) 1.000 (reference)
       4–7 year 3 421 (38.88) 3 241 (35.79) 1.069 (1.001–1.142) 0.050
       7–10 year 1 130 (12.84) 1 223 (13.50) 0.936 (0.852–1.027) 0.161
       ≥10 year 654 (7.43) 954 (10.53) 0.694 (0.622–0.775) <0.001
      Type of Infertility
       Primary  3 470 (39.44) 3 984 (43.99) 1.000 (reference)
       Secondary 5 328 (60.56) 5 072 (56.01) 1.206 (1.136–1.280) <0.001
      Male factor 3 718 (43.17) 3 721 (41.09) 1.049 (0.989–1.114) 0.113
      Tubal factor 5 372 (61.06) 5 674 (62.65) 0.935 (0.880–0.993) 0.028
      Ovulation disorder 961 (10.92) 856 (94.52) 1.175 (1.066–1.295) 0.001
      Endometriosis 539 (6.13) 520 (5.74) 1.071 (0.946–1.213) 0.277
      Unexplained factor 358 (4.07) 392 (4.33) 0.945 (0.839–1.064) 0.388
      Infertility diagnosis (>1 type) 2 752 (31.28) 2 818 (31.12) 1.008 (0.946–1.073) 0.815
      Insemination method
       IVF 6 420 (72.97) 6 573 (72.58) 1.000 (reference)
       ICSI 2 378 (27.03) 2 483 (27.42) 0.981 (0.918–1.047) 0.559
      Number of embryos transferred
       1 1 375 (15.63) 2 333 (25.76) 1.000 (reference)
       2 7 037 (79.98) 5 896 (65.11) 2.025 (1.879–2.183) <0.001
       ≥3 386 (4.39) 827 (9.13) 0.792 (0.690–0.909) <0.001
      Embryo stage
       Cleavage 8 267 (95.69) 7 278 (96.23) 1.000 (reference)
       Blastocyst 372 (4.31) 285 (3.77) 1.153 (0.985–1.350) 0.076
      Endometrial thickness
       <7 mm 2 785 (33.06) 3 871 (44.43) 1.000 (reference)
       7–10 mm 4 009 (47.49) 3 572 (41.00) 1.560 (1.460–1.667) <0.001
       ≥10 mm 1 630 (19.35) 1 270 (14.58) 1.784 (1.634–1.948) <0.001
      Number of good-quality embryos
       0 1 269 (14.42) 2 376 (26.24) 1.000 (reference)
       1–4 4 298 (48.85) 4 064 (44.88) 1.980 (1.827–2.146) <0.001
       4–7 1 809 (20.56) 1 435 (15.85) 2.360 (2.142–2.601) <0.001
       ≥7 1 422 (16.16) 1 181 (13.04) 2.254 (2.034–2.499) <0.001
      BMI: body mass index; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; OR: odds ratio; CI: confidence interval; Variable contains missing data; *Adjusted for centers.

      Table 2.  Univariate logistic regression for clinical pregnancy in the training set

      In the final multivariable logistic regression model (Table 3), as compared with women aged less than 31 years, the odds of clinical pregnancy significantly decreased for older women in trend, with ORs being 0.865 (95% CI 0.807–0.928) for 31–36 years, 0.692 (95% CI 0.623–0.769) for 36–41 years, 0.233 (95% CI 0.187–0.289) for 41 years or more. Tubal factor was also significantly associated with a decreased chance of pregnancy (OR=0.885, 95% CI 0.825–0.948). Meanwhile, three other factors were favorable for pregnancy, including transferring 2 embryos (OR=1.805, 95% CI 1.668–1.953, with one embryo as reference), higher endometrial thickness (OR=1.587, 95% CI 1.448–1.738, for ≥10 mm vs. ≤7 mm), and increasing number of good-quality embryos (OR=2.200, 95% CI 1.986–2.437, for ≥7 vs. 0).

      VariablesOR (95% CI)P*
      Female age
       <31 year 1.000 (reference)
       31–36 year 0.865 (0.807–0.928) <0.001
       36–41 year 0.692 (0.623–0.769) <0.001
       ≥41 year 0.233 (0.187–0.289) <0.001
      Tubal factor 0.885 (0.825–0.948) <0.001
      Number of embryos transferred
       1 1.000 (reference)
       2 1.805 (1.668–1.953) <0.001
       ≥3 1.004 (0.861–1.170) 0.959
      Endometrial thickness
       ≤7 mm 1.000 (reference)
       7-10 mm mm 1.405 (1.311–1.506) <0.001
       ≥10 mm 1.587 (1.448–1.738) <0.001
      Total number of good-quality embryos
       0 1.000 (reference)
       1–4 1.962 (1.802–2.137) <0.001
       4–7 2.058 (1.847–2.293) <0.001
       ≥7 2.200 (1.986–2.437) <0.001
      OR: odds ratio; CI: confidence interval; Variable contains missing data; *Adjusted for centers

      Table 3.  Final multivariable logistic regression for clinical pregnancy in the training set

      We developed a nomogram incorporating above five independent factors to predict the probability of clinical pregnancy after IVF/ICSI treatment, by using the effect size from the training set for each variable (Fig. 2). The AUC of the final training data prediction model was 0.66 (95% CI 0.64–0.68), which denotes a good performance. In an independent validation set, the AUC of the validation set was 0.68 (95% CI 0.65–0.70) which was similar to the corresponding ORs in the training set (Fig. 3). Calibration indicated that the predicted probabilities and observed probabilities were well fitted, as assessed by the Hosmer-Lemeshow test (P=0.85).

      Figure 2.  Nomogram [female age, infertility diagnosis (tubal factor), number of embryos transferred, endometrial thickness and number of good-quality embryos] to predict clinical pregnancy following IVF/ICSI treatment.

      Figure 3.  ROC curves for training set and validation set.

    Discussion
    • In current study of 19 837 subjects underwent IVF/ICSI therapy, we identified five independent factors including female age, tubal factor, number of embryos transferred, endometrial thickness and number of good-quality embryos that were shown to have significant effects on the clinical pregnancy probability for couples with IVF/ICSI. The prediction model presented here is the first study reporting on the individual probability of clinical pregnancy over the course of the first fresh embryo cycles in China.

      Female age was the most established predictor included in every prediction model for outcomes of IVF/ICSI[1718]. Only one study did not include the female age into the prediction model of pregnancy due to the age restriction in the inclusion criteria lead to the narrow distribution of age range[12]. Our study observed that progressive decline of clinical pregnancy was associated with increasing age of female. Mounting evidence indicated that the age-related decline on the pregnancy or live birth might be involved to the progressively diminished ovarian with decreases in antral follicle count (AFC), quantity and quality of embryos[1920]. While diminished ovarian reserve may leads to a poor response to gonadotrophin therapy, and ultimately limits the possibility of a successful pregnancy[21].

      Tubal factor is one of the most common diagnosis of infertility, contributes to more than 30%–60% of couples with infertility[22]. Several studies have indicated that infertility couples with tubal factor have significantly reduced live birth probability after IVF/ICSI treatment cycle[23]. However, for clinical pregnancy, there is no significant association with tubal factor reported in previous studies. In our model, tubal factor was firstly reported to be a key predictor for clinical pregnancy probability after IVF/ICSI treatment cycle while a decreased clinical pregnancy chance was observed in couples with tubal factor related infertility. These help to understand the potential relation between tubal related infertility and pregnancy outcomes.

      Male infertility may be caused by testicular and post-testicular deficiencies and account for 40%-50% infertility diagnosis[24]. Although several models for both the live birth and clinical pregnancy have taken male infertility as key predictor, it was not significant in our current predicting model of clinical pregnancy[22]. This may in part be interpreted by the widespread application of ICSI over the past several years, which revolutionized the treatment of couples with male factor infertility, especially for men with nonobstructive azoospermia, or no measurable sperm count[25]. In contrast to conventional IVF, ICSI bypasses the natural barriers to fertilization, thereby increasing the possibility of pregnancy.

      Our model has highlighted the number of embryos transferred as a key predictor for IVF/ICSI success, which has not been used in previous prediction models of clinical pregnancy yet. The number of embryos transferred is a main factor affecting the outcome of ART[2627], and our study showed that double embryo transfer (DET) may help to increase the clinical pregnancy rates. However, compared with single embryo transfer (SET), DET also increases the risk of miscarriage, obstetric complications, and abnormal delivery. Furthermore, triple embryo transfer increases the burden of pregnant women and affects the implantation rate[28]. In recent years, SET has been gradually considered as a better method to reduce multiple birth and avoid perinatal risks, however, it has not been widely applied in fertility clinics of China. Therefore, our model still needs further assessment in external population with different ART treatment strategy.

      The total number of good-quality embryos and endometrial thickness are all established key factors associated with clinical pregnancy and live birth[2930]. Our result also indicated that higher good-quality embryo number and endometrium ≥7 mm exhibited the higher probability of pregnancy in IVF/ICSI. The effect of total number of good-quality embryos on model may partly reflects the accuracy that the currently used morphologic parameters evaluating strategy. While the positive association between higher endometrial thickness and better ART treatment outcome indicated that endometrial thickness as a marker of endometrial receptivity is critical for fertility.

      Although several studies have reported prediction model of pregnancy after IVF/ICSI treatment[12,14,20], no predictive model is available to evaluate the chance of clinical pregnancy in first fresh cycles. In order to make the final model applicable on plain paper, we also developed a nomogram to visualize the association between each predictor and chance of clinical pregnancy. The probability of a successful clinical pregnancy following ART treatment can be obtained by drawing downwards from the total point row to the bottom probability row. The AUC of our model was 0.66, which was similar with previous published models predicting clinical pregnancy or ongoing pregnancy (0.60–0.72)[18,20]. The modest discriminative ability indicated that, there still existed some potentially factors related to clinical pregnancy were not taken in consideration. Actually, a lot of factors have been proved to be associated with pregnancy outcome, such as behavioral factors[31], psychological factors[32], and environmental factors[33]. Therefore, further study with behavioral, psychological, and environmental information may provide better predicting ability.

      This study has some notable strengths. First, our model is the first prediction model of clinical pregnancy for women undergoing their fresh non-donor cycle of treatment. Second, the sample size of our study was relatively large compared to other similar studies. Third, the model was built based on recently collected data from fertility clinics, which may provide better applicability, as the ART treatment strategy is developing rapidly. Fourth, the nomogram analysis provided a visualized information of the model and may help to make the decision for infertility couples and clinicians in clinical practice.

      This study also has some weaknesses. Firstly, this study was limited by some data availability as there is no routine ART surveillance system, therefore we could not have data of other predictors like behavioral factors and psychological factors to resolve existing data gap. Secondly, our model lack of external validation.

      In summary, we have developed a prediction model for clinical pregnancy using multi-center clinical data in China. This novel model provides a useful tool to predict clinical pregnancy after first fresh embryo transfer, which is facilitated to guide couples and clinicians for infertility.

    Acknowledgments
    • The authors thank all of medics in Reproductive Medicine Center of the First Affiliated Hospital of Nanjing Medical University (Nanjing, China), Affiliated Nanjing Maternity and Child Health Hospital of Nanjing Medical University (Nanjing, China), Shengjing Hospital Affiliated to China Medical University (Shenyang, China), and Henan Women and Children Health Hospital (Zhengzhou, China) for recording all the data in this study through the years.

Reference (33)

Catalog

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return