Hohmann 2012.
| Methods | Randomised, open label, safety study | |
| Participants | 88 women, 18 years or older were randomised. Inclusion criteria: gestational age 15 weeks to 23 weeks 6 days, already consented to an induced abortion, desires to use the LNG‐IUS for contraception for 12 months or more; lives in specified counties of Pennsylvania Exclusion criteria: allergy to either polyethylene or levonorgestrel, urgent need for abortion (active bleeding or infection), exposure to or treatment for gonorrhoea or Chlamydia in past 90 days, PID in past year, leiomyomata > 3 cm diameter, uterine anomaly (other than repaired septate uterus) Post‐enrollment pre‐randomisation exclusion criteria (assessed at D&E completion): uterine perforation; haemorrhage as defined by (1) need for transfusion, (2) estimated blood loss > 500 cc, (3) intrauterine placement of a Foley catheter, or (4) use of >= 3 doses of uterotonic medications; infection at time of D&E, including fever (temperature >= 38°C) or pus at the cervical os; subject no longer desires a LNG‐IUS |
|
| Interventions | Insertion of LNG‐IUS immediately following dilation & evacuation compared to delayed insertion (3 to 6 weeks post‐abortion) | |
| Outcomes | Primary: use at 6 months Secondary: uptake, expulsion, continuation, acceptability, utility of ultrasound in predicting expulsion | |
| Notes | This study was completed and data in this update was obtained from the published study | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Subjects were stratified into two strata by parity (parous or nulliparous) with random block sizes of 2,4 and 6 using computer generated sequence |
| Allocation concealment (selection bias) | Low risk | Eligible subjects were randomised by opening the next sequentially numbered sealed opaque envelope in the operating room |
| Blinding of participants and personnel (performance bias) All outcomes | Low risk | A statistician not affiliated to the study prepared the envelopes |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Study staff for these evaluations were blinded to subjects' randomisation assignments |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | Study staff for these evaluations were blinded to subjects' randomisation assignments |
| Incomplete outcome data (attrition bias) All outcomes | High risk | Not all randomised participants were included in the final analysis. Participants lost to follow‐up were not included in the analysis of important outcomes Number of participants lost to follow‐up was high. 17 participants were lost to follow‐up in each group. |
| Selective reporting (reporting bias) | Unclear risk | unclear |
| Other bias | Unclear risk | unclear |