Treatments for short cervix: cerclage, progesterone

It does no good for individuals to know that they are at risk if there is no effective treatment. Instead, it may actually do harm, triggering anxiety and a desperate search for unproven, costly, or hazardous treatments. For those at risk for PTB, well-meaning clinicians are tempted to deploy a host of unproven or even discredited interventions, including activity restrictions, bedrest, long-term tocolysis, or antibiotic treatment. Cervical cerclage and progesterone are treatments proposed specifically for women with a short cervix. But as recently as 2001, Owen and colleagues cautioned, “Until properly designed trials of cerclage or other interventions prove a benefit from the finding of a ‘short’ cervix in the midtrimester, we recommend that cervical length measurement … remain investigational.”

Since then, evidence has accumulated that cerclage can indeed reduce PTB and neonatal morbidity in a specific subset of women with short cervix: those with a singleton pregnancy and a history of a spontaneous PTB. Because there is an effective treatment, experts now recommend serial midtrimester screening of TVCL in women with a history of spontaneous PTB. On the other hand, cerclage has not been shown to reduce PTB in women without a history of PTB and may actually be harmful to those with twins.

In this issue of the journal, Romero and colleagues present evidence that vaginal micronized progesterone treatment reduces PTB and neonatal morbidity in asymptomatic women with a short cervix, whether or not there is a history of PTB. Using a sophisticated individual patient data metaanalysis based on 5 placebo-controlled trials, they report that vaginal progesterone reduced the rate of early PTB (<33 weeks) among those with TVCL ≤25 mm (12.4% with progesterone vs 22.0% with placebo) and reduced neonatal complications. The 2 largest trials in the metaanalysis, comprising 91% of the subjects, had remarkably consistent results despite differences in the TVCL (10-20 mm at enrollment vs <15 mm at enrollment ) and progesterone dosage (90 mg daily vs 200 mg daily ): progesterone reduced early PTB by 45% and 42% compared to placebo.

The analysis of Romero and colleagues clearly establishes that treatment with vaginal progesterone is beneficial for asymptomatic women with singleton pregnancies in whom a short cervix is detected. The metaanalysis of the combined patient-level data yields key insights that could not have been gleaned from the individual trials:

• •
  

  The beneficial effects of progesterone on early PTB and neonatal morbidity was similar in various subgroups of TVCL (21-25 mm vs 10-20 mm vs <10 mm), but only the subgroup with TVCL 10-20 mm had sufficient statistical power to reach significance.

  
  
• •
  

  The benefits of progesterone were similar whether or not the women had a history of PTB.

  
  
• •
  

  The benefits were similar whether the dose of progesterone was 90-100 mg daily or 200 mg daily.

  
  
• •
  

  The benefits did not appear to depend on maternal age, body mass index, or race/ethnicity.

  
  
• •
  

  Although the subgroup of twin pregnancies with short cervix in the 5 trials was small, progesterone treatment was associated with a significant reduction of neonatal morbidity but not a significant reduction in early PTB.

Should we screen for short cervix?

Even though the metaanalysis clearly establishes that treatment with vaginal progesterone is effective for women in whom a short cervix has been detected, it would be a large leap to conclude that it would necessarily be worthwhile to embark on a large-scale, costly program to screen all pregnant women with transvaginal sonography to detect those with a short cervix. Only 1.7% of asymptomatic women have a TVCL ≤15 mm and 2.3% have TVCL of 10-20 mm in the late second trimester. Even among those few with a short cervix, progesterone reduces but does not entirely eliminate PTB and neonatal morbidity. Do these reductions in such a small subset of women justify the costs of a screening program?

The World Health Organization has outlined 10 general principles for a good screening test. Let us examine how well TVCL screening satisfies these:

• 1
  

  The condition being sought should be an important health problem. The importance of PTB as a health problem is well known.

  
  
• 2
  

  There should be an accepted treatment for patients with recognized disease. The metaanalysis provides compelling evidence that vaginal progesterone is beneficial for those with singleton pregnancy and TVCL 10-20 mm, whether or not there is a history of PTB. For women with a singleton pregnancy and history of PTB, the efficacy of vaginal progesterone appears comparable to the efficacy of cerclage.

  
  
• 3
  

  Facilities for diagnosis and treatment should be available. Endovaginal sonography is standard for any center that performs first-trimester or gynecologic ultrasound examinations. Valid TVCL measurement requires a standardized protocol (empty bladder, standardized placement of transducer and calipers, and observations over several minutes at rest and after application of fundal pressure or Valsalva). Training for physicians and sonographers requires a small investment of time and should be available through a World Wide Web–based training program similar to the program used by the Nuchal Translucency Quality Review Program. Many centers have already been routinely measuring TVCL in women with history of PTB and other risk factors.

  
  
• 4
  

  There should be a recognizable latent stage. Progressive cervical shortening is often present on serial TVCL measurements in women who later deliver preterm, with a rate of shortening almost identical among those who later present with preterm labor and those with preterm prelabor rupture of membranes. These observations suggest that cervical shortening is part of a long latent phase of the spontaneous PTB “phenotype.”

  
  
• 5
  

  There should be a suitable test or examination. TVCL measurement is such a test. Transabdominal cervical length measurement is not a reliable screen because the cervix often cannot be imaged well transabdominally with an empty bladder, but bladder filling tends to make the cervix appear longer than it actually is.

  
  
• 6
  

  The test should be acceptable to the population. Most women readily accept a vaginal sonogram once they understand its value.

  
  
• 7
  

  The natural history of the disease, including development from latent to declared disease, should be adequately understood. Although we do not understand all the mechanisms involved, there is substantial observational experience about the natural history of the short cervix. Many women remain asymptomatic for weeks or months before presenting with preterm labor or rupture of the membranes. Others deliver shortly after diagnosis. Still others deliver at or near term.

  
  
• 8
  

  There should be an agreed policy on whom to treat as patients. Professional organizations have yet to issue consensus statements regarding treatment of women with a short cervix. The most compelling evidence favors treatment for women with singleton pregnancy and TVCL ≤20 mm.

  
  
• 9
  

  The cost of case-finding (including diagnosis and treatment of patients diagnosed) should be economically balanced in relation to possible expenditure on medical care as a whole. Two independent cost-effectiveness analyses concluded that there would be a net cost savings due to a reduction in neonatal complications with a program involving universal TVCL screening followed by vaginal progesterone treatment if a short cervix is detected. Despite differences in assumptions regarding costs for screening and treatment, the 2 reports agreed that universal TVCL screening would result in a net savings of at least $19 million dollars for every 100,000 women screened, averting hundreds of cases of early PTB and dozens of cases of perinatal death and severe neonatal morbidity.

  
  
• 10
  

  Case-finding should be a continuing process and not a “once and for all” project. This principle requires that a screening program enroll new patients over time. It does not imply that we need to do serial TVCL measurements in each woman. Indeed, the studies in the metaanalysis of Romero et al were mainly based on a single measurement at 19-25 weeks of gestation. The 2 cost-effectiveness analyses also assumed a single TVCL measurement. Cost-effectiveness certainly decreases if screening becomes more expensive, as would happen with serial measurements.

Universal TVCL screening in midpregnancy thus appears to satisfy all the World Health Organization principles for screening.

Public and private health care payers frequently balk at paying for new or expensive screening tests or treatments. Payers should take notice: universal TVCL screening followed with progesterone treatment for those with a short cervix should result in substantial net cost savings and reductions in perinatal morbidity and mortality.