ObjectiveTo compare cumulative live birth rate (LBR) between progestin-primed ovarian stimulation (PPOS) and GnRH antagonist protocols of preimplantation genetic testing (PGT) cycles in different populations.MethodsThis was a retrospective cohort study. A total of 865 patients were enrolled and separate analyses were performed for three populations: 498 patients with predicted normal ovarian response (NOR), 285 patients with PCOS, and 82 patients with predicted poor ovarian response (POR). The primary outcome was cumulative LBR for one oocyte retrieval cycle. The results of response to ovarian stimulation were also investigated, including numbers of oocytes retrieved, MII oocytes, 2PN, blastocysts, good-quality blastocysts, and usable blastocysts after biopsy, as well as rates of oocyte yield, blastocyst formation, good-quality blastocysts, and moderate or severe OHSS. Univariable and multivariable logistic regression analyses were used to identify potential confounders that may be independently associated with cumulative live birth.ResultsIn NOR, the cumulative LBR of PPOS protocol was significantly lower than that of GnRH antagonists (28.4% vs. 40.7%; P=0.004). In multivariable analysis, the PPOS protocol was negatively associated with cumulative LBR (adjusted OR=0.556; 95% CI, 0.377-0.822) compared to GnRH antagonists after adjusting for potential confounders. The number and ratio of good-quality blastocysts were significantly reduced in PPOS protocol compared to GnRH antagonists (2.82 ± 2.83 vs. 3.20 ± 2.79; P=0.032 and 63.9% vs. 68.5%; P=0.021), while numbers of oocytes, MII oocytes and 2PN did not show any significant difference between GnRH antagonist and PPOS protocols. PCOS patients had similar outcomes as NOR. The cumulative LBR of PPOS group appeared to be lower than that of GnRH antagonists (37.4% vs. 46.1%; P=0.151), but not significantly. Meanwhile, the proportion of good-quality blastocysts in PPOS protocol was also lower compared to GnRH antagonists (63.5% vs. 68.9%; P=0.014). In patients with POR, the cumulative LBR of PPOS protocol was comparable to that of GnRH antagonists (19.2% vs. 16.7%; P=0.772). There was no statistical difference in the number and rate of good-quality blastocysts between the two protocols in POR, while the proportion of good-quality blastocysts appeared to be higher in PPOS group compared to GnRH antagonists (66.7% vs. 56.3%; P=0.182). In addition, the number of usable blastocysts after biopsy was comparable between the two protocols in three populations.ConclusionThe cumulative LBR of PPOS protocol in PGT cycles is lower than that of GnRH antagonists in NOR. In patients with PCOS, the cumulative LBR of PPOS protocol appears to be lower than that of GnRH antagonists, albeit lacking statistical difference, whereas in patients with diminished ovarian reserve, the two protocols were comparable. Our findings suggest the need for caution when choosing PPOS protocol to achieve live births, especially for normal and high ovarian responders.