Background:
Exposure to secondhand smoke (SHS) during pregnancy is associated with harmful health effects to the foetus and newborn baby. We piloted an intervention targeting pregnant women whose husbands smoked at home, explored intervention acceptability and trial feasibility.

Methods:
The IMPRESS study was administered in Comilla District, Bangladesh. We recruited 48 pregnant women in their first and second trimester, and followed them up at 3 months and within 48 hours after delivery. Outcomes included foetal exposure to SHS measured by pregnant women's salivary cotinine levels, SHS knowledge and smoking behavior of husbands/family members. Interviews explored intervention acceptability. Trial feasibility was assessed by recruitment and retention rates.

Results:
The recruitment target of 48 pregnant women (100%) was achieved in 40 days. Among 96 pregnant women approached, 35 were non-eligible, and 13 declined to participate. Retention was 100% at 3 months, 62% within 48 hours after delivery. Mean cotinine level (intervention arm) declined from 0.43 ng/ml (SD 0.42) at baseline to 0.34 ng/ml (SD 0.6) at first follow up. The difference between intervention and control arms was not statistically significant. Knowledge increased in both arms (+1.1 intervention, +0.6 control, NS). Smoking in the presence of pregnant women reduced by 50% (intervention) versus 17% (control) (p=0.014). Men and women described good engagement with the intervention. Its perceived impact was attributed to new SHS knowledge, women gaining confidence in negotiating a smoke free home and men feeling guilty about smoking close to others.

Conclusions:
Recruitment and retention rates suggest a trial with pregnant women at the community level in a rural setting is feasible in Bangladesh. The intervention appears acceptable and to have potential for change. Its effectiveness should be tested in a fully powered randomised controlled trial.