Effect of Recombinant Human Granulocyte Colony‐Stimulating Factor on Pregnancy Outcomes in Patients With Recurrent Pregnancy Loss: A Single‐Center Retrospective Study

OBJECTIVE: This study aims to explore how recombinant human granulocyte colony-stimulating factor (rhG-CSF) impacts pregnancy outcomes among patients with recurrent pregnancy loss (RPL), especially the effects of different rhG-CSF administration types on these outcomes. METHODS: This retrospective cohort study analyzed 504 RPL patients treated at the Lanzhou University Second Hospital, including 291 in the treatment group (receiving rhG-CSF) and 213 in the control group. Potential confounding factors were adjusted by Propensity Score Matching (PSM) and Inverse Probability of Treatment Weighting (IPTW). Logistic regression was employed to assess the association between rhG-CSF and pregnancy outcomes. Additionally, exploratory subgroup analysis was performed based on different rhG-CSF administration types (short-acting, long-acting, and sequential therapy). RESULTS: Overall, rhG-CSF treatment did not significantly improve the rates of live births (OR = 1.259, 95% CI: 0.771-2.062, p = 0.357), preterm births (OR = 1.095, 95% CI: 0.304-6.143, p = 0.917), or neonatal unit admissions (OR = 1.123, 95% CI: 0.205-7.511, p = 0.898). However, exploratory subgroup analysis revealed that short-acting rhG-CSF was linked to a lower live birth rate (LBR) (aOR = 0.476, 95% CI: 0.224-0.971), while sequential therapy significantly increased LBRs (aOR = 2.417, 95% CI: 1.384-4.268). Further analysis showed that there were no significant differences in the rates of preterm birth and neonatal unit admissions among different groups. CONCLUSION: Although rhG-CSF showed no significant overall benefit for pregnancy outcomes in RPL patients, exploratory subgroup analyses revealed divergent effects based on administration types. While a potential negative association was observed with short-acting rhG-CSF and a positive impact with sequential therapy, these subgroup-specific findings should be interpreted with caution and require validation in larger, dedicated prospective cohorts.