Over 1 million babies die in the world every year of consequences that are related to being born too soon. Every 30 seconds 1 baby dies of preterm birth (PTB). The key to prevention of PTB is early detection of changes that precede eventual late signs and symptoms of preterm labor or premature rupture of membranes.

When I started getting interested in transvaginal ultrasound (TVU) evaluation of cervical length (CL) in 1994 and 1995, this was just a novel sonographic approach with no clinical application. Since then, >1000 studies have been reported on the predictive accuracy of TVU evaluation of CL for PTB in several different populations, and several randomized trials have been reported on interventions that were based on the finding of a short CL.

The interesting article by Vaisbuch et al in this issue of the Journal relates to predictive accuracy of TVU evaluation of CL, specifically for women whose CL is 0 mm. Although they do not report the incidence rate of this finding in their population, this is indeed a rare finding in the second trimester. A CL of <25 mm and ≤15 mm only occurs in approximately 10% and 1-2% of pregnant women, respectively. Even in high-risk women whose cases have been followed in our institution over the last 15 years, a CL of 0 mm occurred in approximately 65/2500 women (2.6%) with serial TVU evaluation of CLs at 16-23 weeks of gestation. In low-risk women, a CL of 0 mm happens in <1% of women. The population that was studied by Vaisbuch et al was high-risk, since 44% of the women had significant risk factors for PTB, mostly because of previous spontaneous PTB. Although approximately 75% of women with a short CL of <25 mm have a cervix that feels “closed and long” on physical examination, women with 0 mm cervix on TVU evaluation of CL mostly have a cervix that appears to have membrane prolapse on speculum examination and “feels” dilated on manual examination. It would be interesting to know whether this is what was found in their study as well, because digital assessment of the cervix was performed in all women.

The main findings of the study of Vaisbuch et al are several. First, asymptomatic women with a CL of 0 mm at 14-28 weeks of gestation have a median diagnosis-to-delivery interval of 3 weeks, and 75% delivered at <32 weeks of gestation. Second, the earlier the CL of 0 mm was detected (eg, <24 weeks of gestation), the sooner the women deliver. A CL of 0 mm was associated with a 38% chance of delivering within 7 days, if detected at <24 weeks of gestation. These findings (as stated by Vaisbuch et al ) confirm previous data and have clinical implications. Steroids for fetal maturation should be considered at ≥24 weeks of gestation in women with a CL of 0 mm (especially if detected at <24 weeks of gestation), given this very high chance of PTB. Counseling regarding management of a possible periviable birth is paramount in these circumstances. From our previous data, high-risk women with a CL of ≤5 mm <21 weeks of gestation or a CL of ≤10 mm before 18 weeks of gestation have a >50% risk of delivering at <32 weeks of gestation and may also be candidates for steroid therapy. It is important to note that women with a CL of 0 mm that is noted later, at 24-28 weeks of gestation, gain a median of 41 days from diagnosis to delivery and are less likely therefore to benefit from steroid therapy at the diagnosis of CL of 0 mm.

Third, approximately 11% of women with a CL of 0 mm at 14-28 weeks of gestation have intraamniotic infection (aerobic/anaerobic bacteria or genital mycoplasmas), and 63% of the women have intraamniotic inflammation (interleukin-6 ≥2.6 ng/mL). An astonishing 59% of these women have sludge on ultrasound examination, which is a clinical sign of these invasive findings. These data have research and clinical implications. Research studies should evaluate whether antibiotic therapy may benefit these pregnancies. Clinically, amniocentesis to evaluate for infection might be discussed with patients with these uncommon findings. Fourth, in women with a CL of 0 mm, the risk of poor outcome is so high that previous historic factors do not further contribute to prediction of PTB. This clinical scenario is an extreme example. We agree with Vaisbuch et al that individualized integrated risk assessment (with the use of historic and current pregnancy risk factors) is the way of the future for the prediction of PTB, with CL currently playing an important role.

Apart for prediction and preparation for PTB, management of a short CL should include discussion of possible measures to avoid the PTB. Several interventions have been studied to prevent PTB once a TVU for CL is detected in the second trimester. Interventions would have a higher chance of showing significant results in women at high-risk for PTB, such as those with previous PTB, because the sensitivity of a short CL of <25 mm at <24 weeks of gestation for PTB in this population is approximately 70%. In the general population with singleton gestations, the sensitivity is only approximately 30%, because most of the few low-risk women who deliver preterm do not do so through a pathophysiologic pathway that involves early cervical shortening. The 2 interventions that have been studied so far by randomized trials for women with short CL are cerclage and progesterone.

In women with singleton gestations, previous spontaneous PTB, and a TVU finding of a CL of <25 mm at <24 weeks of gestation, cerclage is associated with a significant 30% decrease in PTB at <35 weeks of gestation (odds ratio, 0.70; 95% confidence interval [CI], 0.55–0.89) by analysis of data from the 5 randomized studies that included this very selected population. Only 10-15% of women have previous PTB, and only approximately 30-40% of them have a short CL of <25 mm. Therefore, the screening of women with previous spontaneous PTB between 16-34 weeks of gestation with TVU evaluation of CL every 2 weeks from 16-23 weeks of gestation and the performance of cerclage in the minority who have a CL of <25 mm potentially would prevent at least 20,000 PTBs in the United States alone. Cerclage in women with twins and a short CL has been shown to be possibly detrimental. TVU screening for CL therefore cannot be recommended in multiple gestations, because there is no proven intervention to prevent PTB if a short CL is detected.

Vaginal progesterone (200 mg daily) has been shown to be associated with a significant 44% decrease (relative risk, 0.56; 95% CI, 0.36–0.86) in PTB at <34 weeks of gestation in women with a short CL of ≤15 mm at 22-24 weeks of gestation. Only 1-2% of women in the general population experience this finding at <24 weeks of gestation. TVU screening for CL cannot be recommended yet in all singleton gestations, because this finding should be replicated first in at least another large placebo-controlled randomized trial. Several trials on progesterone for the prevention of PTB are underway and include women with a short CL.

Other interventions, not reported in the article by Vaisbuch et al, may hold promise in the future for the prevention of PTB in women with a short CL. These include tocolytics (eg, indomethacin), antibiotics, activity restrictions, and others. At present, given the lack of data for benefit, these interventions should be restricted to research trials.