Utility of Follow up Cervical Length Screening in Low Risk Women with Cervical Length 26–29mm

Abstract

Background:

In women with a prior preterm birth, a protocol for serial cervical length screening has been studied and recommended for identification and treatment of short cervix. Cervical length screening has been suggested for low risk singletons in order to treat short cervix vaginal progesterone and reduce preterm birth. However, specific protocols for single versus serial cervical length ultrasound in this population are not established. Cost effectiveness of universal cervical length screening depends on the cost of screening; follow up of borderline measurements can contribute to increased costs with uncertain benefit.

Objective:

To determine the utility of follow up cervical length screening in otherwise low risk singletons with a mid-trimester cervical length measurement of 26–29mm through assessment of the rate of short cervix (≤25mm) on follow up ultrasound and subsequent delivery outcomes.

Study Design:

This is a two year retrospective cohort study at a single urban institution of pregnant singletons, no prior spontaneous preterm birth, and an initial transvaginal ultrasound cervical length measurement of 26–29mm identified during universal cervical length screening at time of anatomy ultrasound (18 0/7 −22 6/7 weeks). The primary outcome is the rate of short cervix (defined as ≤25mm on transvaginal ultrasound) on follow up ultrasound < 24 weeks. Secondary outcomes included the rate of spontaneous preterm birth (<37 weeks and <34 weeks).

Results:

During the study period there were 2,801 singletons 18 0/7 – 22 6/7 weeks with transvaginal cervical length ultrasound at time of anatomy scan, 201 had a CL 26–29mm, and 184 (7%) had no prior spontaneous preterm birth and were included. Of those included 144 (78%) had a follow up cervical length completed prior to 24 weeks. The mean follow up interval was 1.5±0.6 weeks. Overall the percent of short cervix (≤25mm) on follow up was 15% (N=21). Baseline characteristics were similar except initial cervical length measurement was shorter in those who subsequently developed a short cervix (26.7±0.8 vs 27.8±1.0, p<.001). Delivery outcomes were available for 126 patients. The sPTB rate <37wk of patients with an initial CL 26–29mm who subsequently had CL shortening was significantly higher than a historical cohort of low risk women with CL>25mm (16% vs 3%, p=0.03). The rate of sPTB<34 weeks in those with subsequent short cervix was 11% (N=2/19).

Conclusion:

Approximately 15% of low risk singletons with a mid-trimester cervical length measurement of 26–29mm will experience cervical shortening ≤25mm prior to 24weeks. The rate of spontaneous preterm birth (16%) in women with an initial CL 26–29mm and subsequent shortening ≤25mm is significantly higher than would be expected for a cohort of singletons without a history of preterm birth. Performing 111 additional follow up cervical length ultrasounds would prevent 1 early preterm birth <34 weeks.

Condensation:

15% of low risk singletons with a mid-trimester cervical length of 26–29mm will experience cervical shortening ≤25mm prior to 24weeks

Introduction

Preterm birth is a major cause of morbidity and mortality in the United States and all over the world. Over 385,000 infants are born prematurely in the United States each year¹. A number of adverse conditions are associated with preterm birth including respiratory distress syndrome, necrotizing enterocolitis, intraventricular hemorrhage, and cardiovascular abnormalities². Only 10% of women who have preterm births (<34 weeks) also have a history of preterm birth. In fact, approximately half of women who have preterm births have no history of preterm births^3,4. For those without a prior spontaneous preterm birth, the only effective screening tool for preterm birth prevention is transvaginal cervical length ultrasound^2,5. Studies have demonstrated that transvaginal cervical length ultrasound prior to 24 weeks’ with use of vaginal progesterone for prevention of preterm birth in the setting of short cervix is a cost effective measure^6–9. Although interventions are implemented for women with a cervical length measurement ≤25 mm¹⁰, those with a cervical length 26–30mm still have a ~2-fold increased risk of preterm delivery¹¹.

While serial cervical length measurements are suggested for those with prior spontaneous preterm birth^2,5,12,13, the utility of a repeat cervical length screening to assess for shortening <=25mm prior to 24 weeks’ in otherwise low risk women has not been studied. Previous studies on universal cervical length screening in low risk singletons reported on the shortest mid-trimester cervical length measurement obtained, but not on how often multiple cervical length ultrasounds were performed^14–17. Cost effective analysis have demonstrated the cost effectiveness of a single mid-trimester transvaginal ultrasound cervical length measurement with vaginal progesterone treatment for preterm birth prevention in setting of short cervix^6–9. These analyses did not take into consideration variable practices regarding if/when repeat cervical length ultrasound is performed. Ultimately, in deciding the utility of a formal protocol for repeat cervical length ultrasound it is important to first delineate the incidence of short cervix identified. The purpose of this study is to evaluate the utility of repeat cervical length measurement in singletons without prior preterm birth and an initial cervical length measurement of 26–29mm by first reporting the frequency of short cervix identified with follow up cervical length screening.

Materials and Methods

This is a two year retrospective cohort study of pregnant singletons who presented for anatomy ultrasound at Thomas Jefferson University Maternal Fetal Medicine Antenatal Testing Unit (Philadelphia, PA) from June 2017 through May 2019. Our practice is to perform universal transvaginal ultrasound cervical length screening¹², and we recommend vaginal progesterone (200mg micronized vaginal progesterone suppository or 8% gel) for preterm birth prevention in women with no history of spontaneous preterm birth but with a transvaginal cervical length measurement of ≤25mm prior to 24 weeks’ gestation, and continued until 36 weeks. In June 2017 our practice initiated the recommendation that otherwise low risk women (no prior preterm birth) with an initial transvaginal cervical length measurement of 26–29mm undergo a one time follow up cervical length ultrasound for potential diagnosis of short cervix (FIGURE 1). The objective of this study is to evaluate the utility of this additional follow up screening by describing the frequency of short cervix identified.

FIGURE 1:

Protocol for transvaginal ultrasound cervical length screening in singletons without prior spontaneous preterm birth

The cohort was developed through a search of electronic ultrasound database to identify patients from June 2017-May 2019 with an anatomy ultrasound and cervical length ultrasound done. Inclusion criteria were singleton pregnancy with initial cervical length ultrasound at gestational age 18 0/7 – 22 6/7 weeks measuring 26–29mm. Those with a prior spontaneous preterm birth or cerclage in situ, or major congenital or chromosomal anomaly (i.e. gastroschisis, congenital heart defect, chromosomal aneuploidy) were excluded. For those with more than one pregnancy during the study period, only the first pregnancy was included.

The primary outcome is the rate of short cervix (defined as ≤25mm on transvaginal ultrasound) on follow up ultrasound. Secondary outcomes included the rate of spontaneous preterm birth (<37 weeks and <34 weeks) and neonatal outcomes. Additional analyses included comparing the rate of spontaneous preterm birth within this cohort between women with a normal vs short CL on follow up, and comparing preterm birth outcomes between this cohort of women with an initial CL26–29mm to a historical cohort of those with cervical length measurement>25mm using simple chi-square or Mid-P exact analysis as appropriate. We also estimated the cost of additional ultrasounds and number of preterm birth<34 weeks prevented with follow up cervical length ultrasound screening using a relative risk reduction estimate with vaginal progesterone of 0.65¹⁰. Baseline demographic data, obstetric history, antenatal complications, use of vaginal progesterone or cerclage, and delivery outcomes were extracted from the electronic medical record. Patients who delivered at an outside hospital without delivery data available were excluded from analysis of delivery outcomes. Categorical variables were compared with chi-square analysis or Fisher exact as appropriate, continuous variables were compared with non-parametric Mann-Whitney-U test. P<0.05 was considered significant. Statistical analyses were conducted with SPSS v 26.

Results

Cohort Description

From June 2017-May 2019 there were 2,801 singletons 18 0/7 – 22 6/7 weeks with cervical length ultrasound at time of anatomy scan, 201 had a CL 26–29mm, and 184 (7%) had no prior spontaneous preterm birth and were included (FIGURE 2). Of those included 144 (78%) had a follow up cervical length completed prior to 24 weeks. The mean follow up interval was 1.5±0.6 weeks.

FIGURE 2:

Cohort Flow Diagram

TVU-CL: transvaginal ultrasound cervical length; sPTB: spontaneous preterm birth

Primary Outcome

Overall the percent of short cervix (≤25mm) on follow up cervical length ultrasound was 15% (N=21). Baseline differences between those with normal vs short cervical length on follow up are described in Table 1. Baseline characteristics were similar except initial cervical length measurement was shorter in those who subsequently developed a short cervix (26.7±0.8 vs 27.8±1.0, p<.001), FIGURE 3.

Table 1:

Baseline characteristics of patients with initial cervical length 26–29mm comparing those with normal (>25mm) vs short (≤25mm) cervix on follow up ultrasound. P<0.05 considered significant.

Baseline Characteristics	Follow up cervical length measurement ≤25mm (N=21)	Follow up cervical length measurement >25mm (N=123)	P-value
Maternal Age	28.9±6.8	30.1±5.0	.33
BMI (kg/m2)	25.5±6.2	27.6±6.3	.07
Race:			0.69
African American	11 (52)	50 (41)
Caucasian	7 (33)	43 (35)
Asian	3 (14)	15 (12)
Hispanic	0 (0)	6 (5)
Other/Unknown	0 (0)	8 (7)
Tobacco Use	1 (5)	13 (11)	.69
Prior Term Delivery	11 (52)	48 (39)	.25
Gestational Age at Initial CL	20.4±0.6	20.3±0.6	.66
Mean Initial CL	26.7±0.8	27.8±1.0	<.001
Gestational Age at f/u CL	22.1±0.9	21.9±0.7	.26

Figure 3:

Stacked histogram of follow up cervical length findings based on initial cervical length measurement (N=144). Y-axis indicated number (frequency) of short cervix on follow-up, and % indicates #short cervix/total # x-category. P<.001 for difference in rate of short cervix across x-category groups with chi-square analysis.

Preterm Birth Outcomes

Delivery outcomes were available for 127 women. There was one fetal demise at 31 weeks that was excluded from the following analyses, leaving 126 included. The overall rate of preterm birth <37 weeks was 12% (N=15/126), and spontaneous preterm birth was 9% (N=11/126). The overall rate of preterm birth <34 weeks was 3% (N=4/126), all of which were spontaneous. Of the 19 women with short cervix, 15 had accepted vaginal progesterone, 2 received a McDonald cerclage as part of a trial on cerclage in singletons with short cervix (NCT03251729), and 2 declined vaginal progesterone and were managed expectantly. Of those with subsequent short cervix, 3/19 (16%) delivered <37 weeks, and 2/19 (11%) delivered <34 weeks. There was no statistically significant difference in spontaneous preterm birth <37 weeks in those with a follow up cervical length measurement that was normal vs short (8% vs 16%, p=0.37), or in other delivery outcomes (Table 2). A prior study in our institution¹⁸ identified that low risk singletons with a normal cervical length (>25mm) had a spontaneous preterm birth rate of 3.4% (N=53/1548). The sPTB rate overall of patients with an initial CL 26–29mm was significantly higher than the expected sPTB rate in singletons with a CL>25mm (9% vs 3%, p=0.008), as was the sPTB rate of patients who subsequently had CL shortening (16% vs 3%, p=0.03).

Table 2:

Comparison of perinatal outcomes in women with normal vs short cervix on follow up ultrasound. P<0.05 considered significant). f/u: follow up; CL: cervical length; sPTB: spontaneous preterm birth

Outcome	Follow up cervical length measurement ≤25mm (N=19)	Follow up cervical length measurement >25mm (N=107)	P-value
Gestational age at delivery	37.4±2.4	38.2±1.9	.22
sPTB<37 weeks	3 (16)	8 (8)	.37
sPTB<34 weeks	2 (11)	2 (2)	.11
Birthweight (g)	3049±680	3075±563	.93
NICU admission	5 (26)	16 (15)	.32

Cost/Benefit Estimates

Assuming a rate of 15% of short cervix identified on follow up cervical length ultrasound and an early preterm birth rate (<34 weeks) of 11% with vaginal progesterone treatment vs 17% without treatment (relative risk 0.65¹⁰)- 111 repeat transvaginal ultrasound performed in pregnant women with an initial cervical length 26–29mm would be required to prevent a single preterm birth <34 weeks. Based on a previous cost-effectiveness analysis using 2014 estimates, the base rate for a single transvaginal ultrasound is $107 and the cost for neonatal care of a single preterm birth 28–34 weeks is $39,715 and of their early intervention ages 0–3 is an additional $4,490⁷. The cost of 111 additional cervical length ultrasounds ($11,877) is approximately one quarter of the neonatal/early pediatric cost ($44,205) of a single early preterm birth.

Comment

Principal Findings

To our knowledge, this is the first published assessment of the utility of repeat cervical length ultrasound measurement in this otherwise “low risk” (ie no prior spontaneous preterm birth) population. Our objective was to evaluate the utility of repeat cervical length measurement by first reporting the frequency of short cervix identified on follow up. We found approximately 15% of pregnant “low risk” singletons with a cervical length 26–29mm will develop a short cervix prior to 24 weeks’. Furthermore, we found the rate of preterm birth in this cohort (16%) is significantly higher than would be expected for a low risk singleton with a “normal” (>25mm) cervical length, indicating that this is not a spurious diagnosis of short cervix but is in fact clinically relevant. The potential cost savings of a single early preterm birth (<34 weeks) prevented with 111 additional transvaginal ultrasounds is striking. Our findings highlight the need to reassess our management of this subgroup of patients.

Results in Context of What Is Known

We evaluated a protocol for selective follow up cervical length measurement during a time in pregnancy when there is documented risk of spontaneous preterm birth with diagnosis of short cervix (<24 weeks), and proven intervention to reduce that risk (vaginal progesterone)¹⁰. While this specific protocol has not been evaluated in other studies, our results are congruent with related studies. Iams’ landmark study of 3,000 unselected pregnant women identified a two-fold increased risk of spontaneous preterm birth in patients with a cervical length 25–30mm prior to 24 weeks¹¹. Esplin et al performed a large prospective trial of cervical length screening in 9,410 nulliparous women¹⁹, however the gestational age range of the visits was broad (16–22 weeks and 22–30weeks), which is not necessarily when we would expect cervical shortening in a low risk singleton (>18 weeks), or when interventions may be beneficial (<24 weeks). They too identified women who ultimately had a spontaneous preterm birth had a higher incidence and degree of cervical shortening between visits¹⁹.

Clinical Implications

We have identified that approximately 15% of singletons with an intermediate cervical length 26–29mm will shorten to ≤25mm on follow up prior to 24 weeks’, and will be at high risk for preterm birth. These additional cases of short cervix would otherwise have been missed. Based on our data, follow up cervical length measurement in 111 singletons with a mid-trimester cervical length measurement of 26–29mm would prevent one early preterm birth, with significant cost saving. Ultimately, this signifies singletons with a mid-trimester cervical length 26–29mm are at increased risk of spontaneous preterm birth, and individuals warranting intervention to reduce that risk may be identified with a follow up cervical length ultrasound assessment prior to 24 weeks’.

Research Implications

Previous studies have demonstrated the potential utility of individualizing risk assessment^20–22, this can be applied to individualizing the need for repeat cervical length screening as well, allowing us to further limit costs and maximize benefit.

Strengths and Limitations

This study has a number of strengths. As mentioned above, it is a unique evaluation of a formal protocol for selective repeat cervical length screening at a time point in pregnancy where women are routinely receiving an ultrasound and when there is an intervention for preterm birth prevention if short cervix is identified. While cervical length screening patterns in low risk women vary²³, this protocol has not previously been formally evaluated, however often it may be informally used. Adherence to the protocol was high (>75%) with the large majority of patients returning as recommended for follow up transvaginal ultrasound. This reinforces the findings of other studies that patients find transvaginal ultrasound an acceptable diagnostic modality^24,25. We evaluated not just the rate of short cervix, but also what is clinically important, pregnancy outcomes. Finally, although this is a single center study, our patient population is diverse.

There are limitations to our study. It is retrospective and observational, we are unable to conclude clinical benefit from an additional cervical length ultrasound based on this work alone. Although the entire cohort was sizable, the number of women who developed a short cervix on follow up was limited (N=21), thus conclusions regarding delivery outcomes are limited. The study is from a single urban center which may limit external validity.

Conclusion

In summary, we identified that among singletons without a prior sPTB and with an initial transvaginal ultrasound cervical length of 26–29mm, ~15% will develop a short cervix on follow up ultrasound prior to 24 weeks, and the rate of spontaneous preterm birth in this subgroup is high (16%), warranting further examination into the best screening strategy for this population.

AJOG at a glance:

A. Why was this study conducted?

To determine the utility of a single follow up cervical length ultrasound in otherwise low risk singletons with a mid-trimester cervical length measurement 26–29mm through assessment of the rate of short cervix (≤25mm) on follow up ultrasound and subsequent delivery outcomes.

B. What are the key findings?

Approximately 15% (N=21/144) of low risk singletons with a mid-trimester cervical length measurement of 26–29mm will experience cervical shortening ≤25mm prior to 24weeks. The rate of spontaneous preterm birth in those with short cervix on follow up (16%, N=3/19) is significantly higher than would be expected for a cohort of singletons without a history of preterm birth. An additional 111 transvaginal cervical length ultrasounds in women with an initial cervical length measurement of 26–29mm would be required to prevent one early preterm birth <34 weeks.

C. What does this study add to what is already known?

One time follow up cervical length screening in singletons without a history of prior spontaneous preterm birth and a mid-trimester cervical length measurement of 26–29mm may be a useful tool to identify those with short cervix prior to 24 weeks who are at increased risk for spontaneous preterm birth, and who would benefit from vaginal progesterone therapy.