Table 3.
Safety of systemic hormonal contraception initiated after surgical abortion
| Author, year, location, funding | Study design, follow-up, study quality* | Population | Intervention | Comparison | Outcome & results | Strengths | Weaknesses |
|---|---|---|---|---|---|---|---|
| Steinauer, 2014 [23] United States Anonymous donor |
RCT Phone f/u at 2 and 6 months Poor quality |
Women who had a surgical abortion (1st and 2nd trimester) Median weeks GA at time of abortion: Immediate: 15.9 wks delayed: 15.6 wks (n = 298, n = 212 analyzed at 2 month) |
Immediate contraceptive patch: observed initiation in clinic after abortion (n = 154, n = 108 analyzed at 2 month) |
Delayed contraceptive patch: initiated on first Sunday after abortion (n = 144, n = 104 analyzed at 2 month) |
Bleeding Pattern Changes Days of bleeding requiring a pad, self-reported at 2 months Median days of bleeding: Immediate: 5 Delayed: 5p = 0.94 |
Adequate randomization and allocation concealment Blinded outcome assessment by research assistant conducting follow-up Acceptable attrition (immediate: 29.9%, delayed: 27.8%) |
Power calculation rates based on continuation rates Self-reported outcome with potential for recall bias High attrition (29%) at 2 months and at 6 months (47%) Delayed group may have varying times to Sunday start, potentially affecting bleeding outcome |
| Hou, 2017 [30] China Shanghai Population and Family Planning Commission Of PR China |
Prospective cohort Phone f/u at 1, 3, and 6 months Poor quality |
Women who had a vacuum aspiration at 42-70 days GA (n = 705) |
COC (EE/ desogestrel): initiated immediately after abortion (n = 230) |
Non-hormonal: initiated immediately after abortion Cu-IUD (n = 240) Condoms |
Bleeding Pattern Changes Bleeding days after abortion (median, range): COC: 5 (2-7) Cu-IUD: 7 (3-10) Condom: 5.5 (3-8)p = 0.041 |
Outcomes clearly defined | Day of COC initiation was not detailed nor confirmed Non-comparable groups (COC users were younger and had a higher education level; IUD users were older) Self-report of bleeding duration |
| Wang, 2017 [29] China Source of funding NR |
Prospective cohort Clinic f/u on day 21 following abortion Poor quality |
Women who had a surgical abortion, GA was not clearly stated (n = 726) |
COC (30µg EE/3 mg drosperinone): initiated immediately after abortion x 21 days (n = 312) |
Non-hormonal: no COC x 21 days after abortion, recommended condom use (n = 414) |
Bleeding Pattern Changes Bleeding days after abortion (%): <7 days COC: 286/312=91.7% No COC: 294/414=71.0% p < 0.01 >15 days COC: 4/312=1.3% No COC: 27/414=6.5% p < 0.01 Bleeding volume more than menses (%): COC: 13/312=4.2% No COC: 24/414=5.8% p < 0.01 |
No loss to follow-up | GA not specified Day of COC initiation not specified |
| Ortayli, 2001 [21] Turkey Population Council Nobel Co. |
Prospective cohort Clinic f/u at 2 weeks, 6 weeks, 6 and 12 months (Hct at 6 weeks, 6 and 12 months) Menstrual diaries collected at 2 and 6 weeks Poor quality |
Women who had a surgical abortion at ≤ 10 weeks GA by vacuum or electric aspiration (n = 150) Mean GA age Implant: 54.2 days Nonhormonal: 49.4 days |
LNG implant: inserted immediately after abortion (n = 50) |
Non-hormonal/none: withdrawal method or no method (n = 50) |
Bleeding Pattern Changes Hct change from baseline to 6 weeks (mean, SD): Implant: 1.0 ± 3.8 No hormonal/none: 1.7 ± 4.3p = NS Bleeding days, 2 weeks (mean, SD): Implant: 5.0 ± 3.2 No hormonal/none: 6.2 ± 4.6p < 0.05 Bleeding days, 6 weeks (mean, SD): Implant: 6.3 ± 5.7 No hormonal/none: 5.7 ± 6.7p = NS |
Groups similar at baseline by education, pregnancy and abortion history, and use of recent contraception, experience with IUDs or condoms, and BP, Hct and GA Long follow up time Well defined outcomes |
Groups differed by age, time since last pregnancy and experience with OCs, and weight Attrition not reported Some women switched methods during follow-up period No power calculations |
| Querido, 1985 [18] Netherlands Organon Schering |
Multi-site prospective cohort Clinic f/u at 6 weeks Poor quality |
Women who had a surgical abortion, GA was not clearly stated (n = 423) |
OC: initiated after abortion, unclear timing (n = 123) |
Non-hormonal: Cu-IUD inserted immediately after abortion (n = 300) |
Bleeding Pattern Changes Bleeding days after abortion (%): 1-7 days of bleeding: OC: 56.6% Cu-IUD: 29.7% >14 days of bleeding: OC: 9.8% Cu-IUD: 19.9% p < 0.0001 |
OC group was overrepresented with younger, unmarried, nullip, without prior abortion | OC type not specified OC initiation not detailed Groups differed demographically Range of follow-up times (range of less to more than 6 weeks) Unclear randomization of OC group |
| Kurunmaki, 1983 [17] Finland International Development Research Center of Canada, Ford Foundation, Rockefeller Foundation |
Prospective cohort Clinic f/u at 3, 6, 12 months (measured Hgb, BP, weight; daily bleeding records) Poor quality |
Women who had at surgical abortion in the first trimester (n = 68) |
LNG implant: inserted immediately after abortion (n = 38 assigned, n = 36 inserted) |
Non-hormonal: Cu-IUD inserted immediately after abortion (n = 30 assigned, n = 23 inserted) |
Bleeding Pattern Changes Bleeding days after abortion (mean, SD): Implant: 15.1 ± 6.9 Cu-IUD: 14.3 ± 6.9p = NS |
Measured bleeding by both objective and subjective measures Outcomes clearly defined |
Small study size with no power calculations Assignment of contraception type not described Groups only described and compared by assignment not insertions at baseline Attrition not reported No confounding assessment |
| Peterson, 1974 [16] United States Wyeth Labs |
Prospective cohort Clinic f/u at 6 weeks or form filled out by non-study physician at 6 weeks and mailed to clinic Poor quality |
Women who had a surgical abortion, GA was not clearly stated (n = 978, n = 823 analyzed) |
COC: initiated on day of abortion (n = 479 analyzed) |
Non-hormonal/none: diaphragms, condoms, no method (n = 198 analyzed) |
Bleeding Pattern Changes Self-reported at clinic visit or by form at 6 weeks Stopped bleeding by day 7 (%): COC: 58% Control: 59% Stopped bleeding by day 14 (%): COC: 86% Control: 90% Stopped bleeding by day 21 (%): COC: 94% Control: 96% Stopped bleeding by day 28 (%): COC: 97% Control: 98% No significance testing at any time points |
Two control groups using non-hormonal contraceptives Acceptable attrition (overall 15.8%) |
Groups not assessed at baseline or by outcome for potential confounders Outcome ascertainment not described No significance testing to detect outcome differences Unclear what GA included in study Only outcome is a self-reported outcome with potential for recall bias |
BP, blood pressure; COC, combined oral contraception; Cu, copper; EE, ethinyl estradiol; F/U, follow-up; GA, gestational age; Hct, hematocrit; Hgb, hemoglobin; IUD, intrauterine device; LNG, levonorgestrel; LTFU, lost to follow up; NR, not reported; NS, not significant; OC, oral contraception; RCT, randomized controlled trial.