Interpretation
Universal CL screening meets World Health Organization criteria for an effective screening test 24. CL is inversely related to PTB risk, but its predictive value is variable as short cervix (CL <2.5cm) is rare and not ubiquitous to all PTB cases6, 25-28. At our institution, CL screening is performed at the time of the second-trimester anatomy scan. Our study suggests that qualitative CGA evaluation may be useful in building upon a pre-existing, well-accepted screening modality with negligible impact on patient or provider performance burden.
Wide variation in CGA detection rates in the literature is likely due to differences in study design, CGA density, and subjectivity of TVUS interpretation. Our study demonstrates good reproducibility for CGA detection when prevalence-adjusted for our low-risk population. Several studies have investigated using CGA to triage high-risk patients, such as those with threatened PTL or with prior spontaneous PTB, quoting rates of sonographic CGA absence as high as 56-69% 19, 29. One longitudinal study used serial TVUS and digital cervical exams to chart the natural course of CL, CGA, and Bishop score with advancing GA in low-risk patients delivered at term; they found increasing rates of CGA absence from 31 weeks to term, correlating with a rising cervical maturation index 13. Conversely, persistent visibility of CGA at term has been associated with increased need for late-term labor induction and increased rates of failed induction30, 31. Loss of a sonographically detectable CGA consistently correlates with cervical maturation, supporting the biologic plausibility of a connection between this ultrasound finding and increased PTB risk from premature cervical ripening.
The biologically active endocervical glands play a crucial role in the balance of collagen and other proteins that contribute to the biomechanical strength of the cervix, as well as the cervical mucous required to maintain a barrier between the vagina and an intrauterine gestation 15, 17, 32, 33. Sonographic CGA evaluation may lend itself to an improved functional appraisal of PTB risk as the loss of visualizable CGA coincides with underlying cervical ripening13, 19, 29-31. This connection may be applicable across several disciplines of obstetrics and gynecology, such as for birth outcomes after infertility or oncologic care. For instance, cervical mucus collected from patients following intrauterine insemination or in vitro fertilization features unique proteome profiles34. The composition of cervical mucus has proven predictive of PTB when it features abnormal proportions of inflammatory mediators, such as in the presence of dysbiotic vaginal flora35. The known role of progesterone in PTB prevention is related to maintenance of cervical collagen organization and inflammatory mediation; thus, it is not surprising that progesterone supplementation after embryo transfer results in higher live birth rates17, 33, 36, 37. Further, history of excisional treatment for cervical intraepithelial neoplasia is a known risk factor for second-trimester pregnancy loss and PTB 38. A study found that cumulative excision length significantly predicted PTB, even when second-trimester CL did not vary by gestational age at delivery 39. Higher grade cervical lesions are more likely to involve endocervical glands, requiring deeper excisions and imparting greater PTB risk, perhaps due to more substantial loss of glandular activity 40. Incorporation of CGA evaluation at CL screening, may better assess for dysfunction in the mechanical and biological activities of the glands, both of which are integral to cervical sufficiency.
Such activity was the rationale behind our study’s aim to quantify the CGA visualized on TVUS and to determine if variations in CGA size might reflect underlying physiologic changes that translate to differences in birth outcomes. However, our study found no appreciable association between birth timing and quantitative CGA measurements – though this may reflect insufficient power given our cohort’s low PTB rate. Larger studies with higher PTB prevalence may be necessary to adequately explore the use of CGA measurements in PTB prediction.