Superior Live Birth Rates, Reducing Sperm DNA Fragmentation (SDF), and Lowering Miscarriage Rates by Using Testicular Sperm Versus Ejaculates in Intracytoplasmic Sperm Injection (ICSI) Cycles from Couples with High SDF: A Systematic Review and Meta-Analysis.

This study aimed to compare sperm DNA fragmentation (SDF) levels between ejaculate and testicular sperm and evaluate clinical outcomes of intracytoplasmic sperm injection (ICSI) cycles using testicular sperm (T-ICSI) versus ejaculate sperm (E-ICSI) in males with high ejaculate SDF, prior ICSI failures, or severe male infertility. A systematic review of major databases and a subsequent meta-analysis were performed to compare clinical outcomes in men with high SDF, oligozoospermia, or prior ICSI failures undergoing T-ICSI or E-ICSI. Thirteen studies met the inclusion criteria. Outcomes analyzed included SDF levels, fertilization rate (FR), clinical pregnancy rate (CPR), live birth rate (LBR) per embryo transfer (ET), and miscarriage rate (MR) per pregnancy. The mean difference (MD) and odds ratio (OR) were calculated for each outcome. Paired assessments of SDF showed significantly lower levels in testicular sperm compared to ejaculated sperm (MD = -25.42 [-31.47, -17.30], p < 0.00001). While no significant difference in FR was observed in T-ICSI cycles overall (OR = 0.94 [0.74, 1.20]), a subgroup analysis revealed significantly higher FR with E-ICSI in men with oligozoospermia and no prior ICSI failures (OR = 0.61 [0.52, 0.71], p < 0.00001). CPR was significantly higher in T-ICSI cycles (OR = 2.13 [1.35, 3.36], p < 0.001; n = 540 ET), along with a significantly lower MR (OR = 0.31 [0.14, 0.70], p = 0.004; n = 35) and increased LBR (OR = 2.40 [1.32, 4.36], p = 0.004; n = 446 ET). In conclusion, using testicular sperm in cases of elevated ejaculate SDF, oligozoospermia, or prior failed ICSI cycles enhances the selection of sperm with lower DNA damage, leading to improved pregnancy rates, reduced miscarriage rates, and higher live birth rates. However, the studies included were rated as having a moderate to serious risk of bias. Further well-designed randomized controlled trials are necessary to confirm these findings with stronger evidence.