Objective
To evaluate the percentage change in total βeta‐unit human chorionic gonadotropin (βhCG) levels (%ΔβhCG) in the prediction of treatment outcomes following intravaginal misoprostol for missed miscarriage before 13 weeks.

Methods
A secondary analysis of a randomised controlled study of medical management of miscarriage was performed. Total βhCG levels were collected before misoprostol (baseline) and after a planned seven day interval (follow‐up), when a transvaginal ultrasound (TVUS) reported a gestational sac as present or not. If no sac at TVUS, surgery was indicated on clinical criteria. %ΔβhCG ((baseline βhCG – follow‐up βhCG)/baseline βhCG × 100) was evaluated in the prediction of a sac at TVUS and surgery on clinical criteria.

Results
%ΔβhCG was calculated for cases with βhCG levels within two days of misoprostol and TVUS; calculation interval determined case number. The median %ΔβhCG for 24 cases with a persistent sac (6–9 day interval) was significantly lower than for 145 with no sac (58.75% (interquartile range (IQR): 37.59–76.69; maximum 86.54) vs 97.65% (IQR: 95.44–98.43); P < 0.0001). The median %ΔβhCG for eight cases needing surgery on clinical criteria (5–9 day interval) was significantly lower than for 140 cases with no sac not needing surgery (79.68% (IQR: 64.63–91.15; maximum 94.06) vs 97.68% (IQR: 95.61–98.50); P < 0.0001). The area under the receiver‐operator curve was 0.975 for prediction of a persistent sac and 0.944 for prediction of surgery on clinical criteria, respectively. %ΔβhCG > 87% predicted no sac at TVUS. %ΔβhCG > 94.5% predicted no surgery on clinical criteria.

Conclusion
%ΔβhCG calculation over one week reliably predicted treatment outcomes after medical management of missed miscarriage.