A device known as a pessary, thought promising for reducing  preterm birth risk due to a short cervix, appears no more effective than usual  medical care, according to a study funded by the National Institutes of Health.  A pessary is a rounded silicone device that fits around a cervix that has  shortened, to keep it from opening and leading to miscarriage or preterm birth.  The device is typically removed before the 37th week of  pregnancy.
This multicenter randomized trial was conducted by Matthew  K. Hoffman, M.D., M.P.H., of the ChristianaCare, Newark, Delaware, and  colleagues in the Eunice Kennedy Shriver National Institute of  Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units  Network. Findings appear in the Journal of the American Medical Association.
Researchers enrolled 544 participants (64%) of a planned  sample of 850 expectant people from 16 through 24 weeks of pregnancy at risk  for preterm delivery because they had a cervical length less than 20  millimeters as measured by ultrasound. At enrollment, attending physicians  could perform cervical cerclage (stitching) or administer the steroid  progesterone at their discretion. Participants were assigned at random to  undergo or forego pessary placement or usual care. The researchers designed the  study to determine if use of a pessary could reduce the risk of a single  primary outcome: delivery or fetal death before 37 weeks.  
Study recruitment was stopped early when the researchers  could see no real difference between the groups in terms of the overall primary  outcome, but the risk of fetal or newborn death appeared unacceptably high in  the pessary group. Among those receiving a pessary, the primary outcome  occurred in 127 participants (45.5%)—a rate which did not differ significantly  from 120 participants (45.6%) in the usual care group. Fetal or infant death  occurred with 13.3% in the pessary group and 6.8% of the usual care group.
The authors noted that the usual care group was more likely  to receive cerclage, which could have influenced the results.
NIH release