Sonographic cervical length assessment to detect shortening has been shown to be an effective screening test for prediction and prevention of spontaneous preterm birth, and variations of the strategy have been widely adopted for clinical practice. Some experts have recommended abandoning transabdominal (TA) ultrasound as a screening modality for short cervix, advocating for exclusive use of transvaginal (TV) ultrasound for this purpose. Drs Khalifeh and Berghella are proponents of exclusive TV ultrasound screening and present the argument for this strategy in their accompanying viewpoint. Their argument is largely based on the fact that, per patient, the accuracy of TV ultrasound in estimating cervical length is superior to that of TA ultrasound. As a counterpoint, we submit the opposing argument in favor of maintaining TA ultrasound as a reasonable initial screening approach. We offer 5 main counterpoints to support the argument for incorporating TA ultrasound in the screening and prevention of spontaneous preterm birth.

First, we think the authors of the accompanying viewpoint summarize the current clinical research data without adhering to an equally objective or rigorous assessment of both techniques of cervical length assessment. The presented summation of the literature seems biased in favor of TV ultrasound, without considering equally the benefits and limitations of each strategy. Several articles that the authors cite actually do not support their argument and firm conclusions. Based on critical review of the same body of literature, we identified several study issues that undermine the argument in favor of exclusive TV ultrasound screening. In the observational study by Stone and colleagues, TA ultrasound underestimates cervical length and therefore overcalls cervical shortening. This leads to concern that TA ultrasound results in an excess of false-positive screens, which would not reduce sensitivity. This error would be eliminated by a 2-tiered screening scheme in which TV ultrasound is used in the minority of patients with a short cervix on TA examination, and the authors of this study conclude that for patients with adequate cervical length detected on TA, TV ultrasound would not provide additional information. Saul et al observed 100% sensitivity for detection of TV cervical length <25 mm using a TA ultrasound cervical length cutoff of 3 cm in a prospective cohort. They also concluded that good TA ultrasound technique for cervical length includes post-void measurements and trained sonographers. In the observational study by Friedman and colleagues, 97% of TV ultrasound short cervices would be detected with a TA ultrasound cervical length cutoff of <34 mm. Using a prospective cohort study design, Rhoades et al observed that a TA cervical length cutoff of 35 mm excludes a short cervix of <30 mm and avoids TV ultrasound in 68% of patients. In the observational study by To and colleagues, TA ultrasound was shown to consistently underestimate cervical length compared to TV ultrasound. In addition, TV measurements were significantly shorter after bladder emptying than prior to bladder emptying. These authors confirmed that high bladder volume erroneously lengthens the cervix, a finding confirmed by Marren et al ; notably, the cervix was well visualized by TA approach in 49% of patients with low urine volume. Marren et al report that in their cohort, the cervix was well visualized in 82.8% of patients with an empty bladder. In this study, using a TA ultrasound cervical length cutoff of 35 mm improved sensitivity for identifying a short cervix, but would result in 77% of women needing a TV scan (inadequate image in 18%; cervix <35 mm in 59%). The study that showed poorest test characteristics for TA cervical length appeared aimed to prove a preconception that TA ultrasound was inferior rather than to objectively compare 2 methods. To quote the authors, “We undertook this study because we were surprised that some investigators continue to propose that transabdominal sonography can be used to screen patients to detect those with a short cervix.” In doing so, the research team optimized the protocol and performance of TV ultrasound, but did not optimize the performance conditions of TA ultrasound. For example, TA ultrasound cervical length assessment was always performed with a maternal full bladder. The study design is a self-fulfilling prophecy for their conclusion that TV is superior to TA ultrasound. In sum, these studies indicate that TA ultrasound can be used effectively to assess for cervical shortening in a considerable proportion of patients (25-80%), when good technique is applied and exam limitations are acknowledged and addressed. In addition, the results of cost-effectiveness studies comparing TV and TA ultrasound are not conclusive based on sensitivity analyses. For example Miller and Grobman concluded from their cost-effectiveness analysis that: “Optimizing TA ultrasound (US) testing characteristics or applying a transabdominal screening strategy in lower risk populations may yield an initial TAUS to be cost-effective.” Other cost-effectiveness analyses do not compare TV and TA ultrasound.

Second, proponents of exclusive TV ultrasound focus on and over-value efficacy (ie, test performance under ideal conditions), which for clinicians and public health policy makers, and likely patients, is less relevant than effectiveness (ie, test performance under real-world conditions). There is no question that TV ultrasound is a more accurate assessment of cervical length, but accuracy and precision are not the only important factors in determining the value or effectiveness of incorporating TA ultrasound cervical length assessment in a screening strategy. For real-world implementation of a screening strategy, one must also take into consideration factors such as level of acceptability of the test for patients and providers, ease of application, and incremental cost and resource utilization. In this screening scenario, using TA ultrasound for cervical length measurement at the time of a routine 18- to 20-week fetal ultrasound represents little or no increase in resource utilization since the examination is already regularly performed for other reasons. Thus, reducing the frequency that TV ultrasound screening is performed via TA ultrasound would likely result in cost and resource savings.

Third, proponents of TA ultrasound cervical length assessment generally do not recommend use of TA ultrasound in place of TV ultrasound but as the first tier in a 2-step screening strategy. TA ultrasound decreases the frequency that TV ultrasound is needed, thereby decreasing the resource burden and patient inconvenience of a TV exam. Using this 2-tier screening scheme has been employed without degrading sensitivity or safety when good technique is applied consistently and when TV ultrasound is used reflexively either when TA visualization is inadequate or when a short cervix is suspected on TA using a higher threshold to identify shortening (eg, 30-35 mm). Thus, the most pertinent comparison is screening with exclusive TV ultrasound vs TA ultrasound that reflexes to a TV ultrasound when not fully informative, not simply TV vs TA. In the absence of clear superiority of TV ultrasound screening effectiveness, other factors should be considered such as time, cost, resource availability, and patient preference.

Fourth, the authors arguing for exclusive TV ultrasound do not consider patient preferences in their discussion. While several studies have shown that TV ultrasound is acceptable to patients, this does not mean that women prefer TV over TA exam if the latter is an option. Therefore, if the frequency of TV ultrasound exams could be reduced with TA ultrasound without degrading screening effectiveness, this could be of significant value to patients and providers alike. Bennett and Richards report that only 24% of patients experience no pain with TV ultrasound, while 26% report that their prior experience with TV ultrasound “hurt a lot.” This is in comparison to abdominal ultrasound, in which 78% of patients experienced “no pain.” Braithwaite and Economides similarly found that almost half of patients experience at least mild discomfort with TV ultrasound. When compared to transperineal or translabial ultrasound, TV ultrasound has also consistently been reported as being associated with a higher discomfort score.

Fifth, as we alluded to above, we believe that standardizing the TA ultrasound technique for cervical length is important to maximize screening effectiveness, just as it is for TV ultrasound or any other screening test. The current observational data suggest that there is likely benefit in optimizing performance and effectiveness of TA ultrasound cervical assessment in the same fashion that has been recommended for TV ultrasound (eg, Cervical Length Education and Review [CLEAR] certification). We propose that, before we abandon TA ultrasound cervical assessment, researchers more accurately evaluate the effectiveness of TA ultrasound when applying the same rigor for TV ultrasound performance to TA ultrasound screening, including technique standardization and training, as well as quality assurance measures. Notably, in the studies referenced above in which TA ultrasound showed adequate screening characteristics, it was done with a standardized protocol, involving experienced sonographers and examination with an empty bladder.

Currently there is little research that compares the screening effectiveness or preventive benefit of TV vs TA ultrasound for spontaneous preterm birth. Although it is true that randomized clinical trials of progesterone therapy for preterm birth prevention relied on TV ultrasound rather than TA ultrasound cervical length screening, the overwhelming majority of patients with a short cervix would receive a TV ultrasound in a 2-tiered screening strategy. Until more conclusive evidence is available directly comparing the effectiveness of a screening scheme that employs exclusive TV ultrasound vs one that incorporates TA ultrasound, we believe that either strategy is reasonable to screen for spontaneous preterm birth.