Strategic implementation of ICSI leads to higher success rates than routine ICSI


      To evaluate if there are differing outcomes in fresh autologous cycles in clinics that use an algorithm to determine whether to use in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) versus clinics that use ICSI exclusively.


      Retrospective cohort study

      Materials and Methods

      All women with fresh, frozen, donor, and non-donor embryo transfers from the Society for Assisted Reproductive Technology (SART) database between the years 2004 to 2013 were analyzed. Clinics with greater than or equal to 95% of ICSI performed were termed routine ICSI clinics, whereas those with less than 95% were designated as algorithm-based clinics. The main outcome measures were pregnancy, live birth, multiple gestation, birth weight, neonatal death, and gender rates. Independent variables included demographic and cycle characteristics. SAS 9.4 (Cary, NC) was used to calculate frequencies and perform bivariate analyses.


      A total of 159,005 patients from 465 clinics in 3,055 distinct clinic-years were reviewed. Of routine ICSI clinics, 24,826 patients underwent fresh and 1,528 patients underwent frozen embryo transfers from 557 distinct clinic-years. Among algorithm-based clinics, 132,379 patients underwent fresh and 5,490 underwent frozen embryo transfers from 2,498 distinct clinic-years. Patients in algorithm-based clinics undergoing fresh, non-donor cycles have significantly higher pregnancy and live birth rates, while having lower neonatal death rates among singleton and multiple gestations. There is no difference in rates of spontaneous abortion, multiple gestation, live birth weight, or length of gestation. Patients in algorithm-based clinics had more male infertility, were older and had higher BMIs than those in routine ICSI clinics.


      Performing routine ICSI for the purpose of improving fertilization rates leads to an overall decrease in birth rate. This decrease in success rates was seen in all age strata (data not shown).
      Tabled 1Fresh Non-Donors with routine ICSI versus Algorithm-based ICSI
      Fresh Non-DonorRoutine ICSI (N = 22833)Algorithm-based ICSI (N = 122219)P-value
      Age (years) at cycle start, mean (# cycles)34.70 (24774)34.80 (129289)0.0052
      Patient Weight (lbs), mean (# cycles)143.04 (15913)143.40 (72986)0.3945
      BMI (kg/m2), mean (# cycles)25.53 (14970)25.76 (68363)<0.0001
      Male Infertility, % (N)35.04 (8680/24774)35.72 (46182/129289)0.0397
      Pregnancy Rate Per Cycle, % (N)41.38 (10250/24774)44.15 (57080/129289)<0.0001
      Live birth rate per cycle, % (N)29.33 (7265/24774)31.44 (40650/129289)<0.0001
      Miscarriage rate, % (N)13.54 (1388/10250)13.29 (7588/57080)0.4967
      Neonatal death per live birth, % (N) Singleton0.65 (33/5089)0.57 (161/28155)<0.0001
      Neonatal death per live birth, % (N) Multiple2.34 (51/2176)1.44 (180/12495)<0.0001
      Multiple gestation rate per CP, % (N)31.97 (2753/8777)31.96 (15597/48806)0.2739
      Length of gestation (days), mean (N)227.94 (8679)228.84(48380)0.2932