To identify factors predictive of persistent ectopic pregnancy (PEP) in women who have undergone laparoscopic salpingostomy or salpingotomy for tubal pregnancy. Design: Retrospective cohort study (Canadian Task Force classification II-2). Setting: Tertiary referral center. Patients: Ninety-nine women who underwent laparoscopic tubal preservation surgery for ectopic pregnancy. Interventions: Seventy women underwent laparoscopic salpingostomy, and the remaining 29 women underwent laparoscopic salpingotomy. Measurements and Main Results: Factors predicting PEP were evaluated. The change in serum beta human chorionic gonadotropin (b-hCG) levels from baseline observed between postoperative days 5 and 10 (ChCGD5-10) was a predictor of PEP (odds ratio [OR], 0.80; p = .01). Based on receiver operating characteristic (ROC) curve analysis, a cutoff value of 93.1% was determined, with an area under the ROC curve of 0.95 (sensitivity, 85.7%; specificity, 100%). Nonetheless, when considering perioperative variables only, body mass index (BMI) was identified as a predictor of PEP (OR, 0.71; p = .03). Based on the ROC analysis, a BMI cutoff value of ≤ 22 kg/m 2 was determined, with an ROC area of 0.73 (sensitivity, 43.2%; specificity, 100%). In addition, a higher baseline b-hCG level (hazard ratio [HR], 1.0002; p = .009) and left tubal pregnancy (HR, 6.46; p = .03) were predictive of recurrent ectopic pregnancy. There were no differences in the perioperative outcomes, PEP rates, or subsequent intrauterine pregnancy rates between the salpingostomy and salpingotomy groups. In addition, surgical method was not a predictor of recurrent ectopic pregnancy. Conclusions: ChCGD5-10 was identified as a predictor for PEP, suggesting that it might be more clinically useful for the follow-up of PEP. When considering perioperative variables only, BMI was a predictor for PEP. In addition, there was no significant difference in clinical outcomes between the salpingostomy and salpingotomy groups.