Changes in cervical effacement and dilatation occur well before preterm birth1–4 and must precede all spontaneous deliveries. Screening strategies to predict preterm birth based on early identification of cervical change in asymptomatic gravidas have intuitive appeal and have been incorporated into preterm birth prevention protocols.5,6 However, basic issues remain unresolved, such as whether existing research accurately quantifies the usefulness of clinical cervical examination for routine screening; how cervical length, and its interaction with dilatation, are related to risk; and whether baseline cervical dimensions are associated with spontaneous preterm birth regardless of pathway, or are differentially linked to the risk of preterm labor or preterm premature rupture of membranes (PROM).
The relation between clinical assessment of cervical dimensions and prematurity is not firmly established for general obstetric populations in the United States. To accrue a sufficient number of preterm births, many investigators have restricted their studies to high-risk subjects such as those with histories of preterm births, multifetal gestations, or other combinations of known risk factors.7–12 Some have included subjects with symptoms such as uterine contractions at the time of evaluation. The largest studies of general obstetric patients have been conducted in Europe, in populations in which prevailing risk factors are divergent from those in the United States and rates of preterm birth are quite low (3–5%).13,14
Dilatation and effacement and their interdependence may contribute to the usefulness of cervical examination as a screening tool. Dilatation early in gestation, when screening must have utility, is rare. Variations in length are more common and potentially more informative. Yet measures of cervical length by digital vaginal examination have been underemphasized. General categories are most often used: complete effacement compared with incomplete, or a cutoff point such as 1 cm.3,13,14 Multiple categories of length and dilatation, with potential clinical significance, should be considered. Composite measurements of cervical status such as cervical score8,10 (length of the cervix minus cervical dilatation in centimeters) should be explored to summarize the contributions of length and dilatation and to accommodate differences in prominence of dilatation versus effacement between nulliparous and multiparous subjects.
Preterm labor and preterm PROM are hypothesized to have distinct causal mechanisms, with preterm labor more closely related to uterine activity and PROM related to ascending infections or inflammatory processes.15,16 Although these pathways to prematurity likely share some etiologic elements, we wanted to know whether cervical change is more closely related to the former than the latter because previous work combined births resulting from both pathways.
Our objective was to examine prospectively the relation between cervical dilatation and length in centimeters and the risk of spontaneous preterm birth, including its subtypes preterm labor and preterm PROM, in a general obstetric population. We used survival analysis to describe how cervical examination findings relate to degree of prematurity.
The subjects were participants in an ongoing prospective cohort study. The Pregnancy, Infection and Nutrition Study17 enrolled patients from four prenatal care sites: the resident physician clinics and faculty clinics at the University of North Carolina Women's Hospital, a tertiary-care academic center; the resident-staffed clinics at Wake Medical Center, an urban community hospital; and Wake County Health Department prenatal clinics. Eligibility requirements included singleton gestation; maternal age of 16 years or older; fluent spoken English; access to a telephone; at least one routine prenatal care visit before the study enrollment window of 24–29 weeks; and willingness to complete all phases of the study, including a pelvic examination, blood and urine specimens, telephone interview, and dietary history.
During the study period, 1702 potentially eligible women entered prenatal care, 92% were invited to participate, and 970 (57%) were enrolled. The rate of preterm birth among those not enrolled was similar to that of the participants.17 At the time of this analysis, pregnancy outcomes were not known for seven subjects who were at least 3 weeks past their estimated dates of delivery, and they were excluded. Clinical cervical measurements were not initially included in the study protocol and for a short time included only dilatation and not length. As a result, 139 early enrollees had incomplete data for cervical examinations (92 did not have cervical examinations done, 40 had dilatation recorded but not length, and seven had length recorded but not dilatation).
A total of 871 subjects with prospectively recorded cervical examination data were included in these analyses. Dilatation of the internal cervical os and cervical length, measured in centimeters from the plane of the external cervical os to the lateral vaginal fornix, were assessed by routine clinical examination at 24–29 completed weeks. These examinations were conducted by the patients' usual care providers, including attending physicians, resident physicians, or health department nurses, after completion of speculum examinations. Providers were asked to report findings in centimeters without restriction of increment. Thirty-five reports of “fingertip” dilatation of the internal cervical os, intended to indicate less than 1 cm, were converted to 0.5 cm for analysis.
Gestational age was assigned using the last menstrual period if these dates were within 14 days of the estimated dates by earliest ultrasound before enrollment; otherwise ultrasound dating was used. Thirty-seven completed weeks was considered term. Preterm births were assigned to three categories by two of the obstetricians (JMT, TLM) after review of the hospital records and discharge summaries. These physicians were unaware of results of the baseline cervical examinations. The following standardized definitions were used: pre-term labor—spontaneous labor and delivery before 37 completed weeks' gestation; preterm PROM—rupture of membranes at least 4 hours before the onset of regular uterine contractions; and other causes or indications. Births in the last category, such as abruption or induction for maternal or fetal indications, were not considered in this report.
We analyzed each cervical dimension as both continuous and categoric variables. To create a composite measurement of cervical change, we calculated cervical scores10 (length minus dilatation in centimeters). Differences between groups were tested using Student t test for continuous variables. The association between pre-term birth and cervical-dimension categories was initially investigated by examining relative risk (RRs) and 95% confidence intervals (CIs), based on Pearson χ2, in both crude and stratified analyses. We selected maternal age, race, parity, previous live births, previous preterm births, previous terminations, previous miscarriages, previous stillbirths, marital status, education, and employment as potential a priori predictors of preterm birth. All were examined as potential modifiers of the relation between cervical examination findings, other covariates, and the risk of preterm delivery. Point estimates of RR by strata and Mantel-Haenszel tests for heterogeneity were evaluated. Each covariate that was related to the risk of preterm birth in the cohort was investigated as a potential confounder by assessing the relation between the covariate and cervical dimensions among those who delivered at term. Logistic regression was used to calculate a multiple-adjusted odds ratio (OR) as an estimate of the risk ratio. Time to delivery was evaluated using the Kaplan-Meier product-limit estimate of the survival curve.
Excluding medically indicated deliveries, 73 subjects (8.3%) gave birth before 37 completed weeks. The mean gestational age at delivery among the preterm group was 33.9 ± 2.4 weeks, with a range of 26–36. Forty-six of these births were attributed to preterm labor and 27 to preterm PROM. Table 1 presents the characteristics of the cohort summarized by pregnancy outcome.
Table 2 summarizes the relation between cervical dimensions and RRs of preterm birth and its subtypes preterm labor and preterm PROM. All three measures (dilatation, length, and cervical score) were associated with the risk of spontaneous preterm delivery. The mean cervical length was 2.7 ± 0.8 cm, with a range of 0–4.5 and an interquartile range of 2–3. Cervical dilatation ranged from 0 to 4 cm; 94% of the subjects had no dilatation of the internal os. Among those with dilatation (n = 55), the mean was 0.9 ± 0.7 cm, with an interquartile range of 0.5–1. Cervical scores (length minus dilatation) had a mean of 2.6 ± 0.8 with a range of −2.0 to 4.5.
The risk of preterm delivery was not associated with marital status, education, employment, history of miscarriage, or previous stillbirth. Black women had an RR of 1.5 (95% CI 0.9, 2.3) compared with whites. The number of subjects from other minority groups (n = 51, including four preterm deliveries) was too small to derive a stable estimate of risk. The RR of preterm birth was 2.5 (95% CI 1.6, 3.8) for those who had had previous preterm deliveries. Any previous live births or induced abortions were associated with an increased risk of preterm birth of 1.5 (95% CI 0.9, 2.4) and 1.6 (95% CI 1.0, 2.6), respectively. Excluding nulliparas, the number of previous pregnancies and number of induced abortions were significantly related to risk (P = .005 and P = .002).
Although related to the risk of preterm birth, race and history of induced abortions were not associated with abnormal cervical examination findings in dilatation, length, or score, and thus did not introduce confounding. Parity was only minimally related to abnormal cervical examination findings, whereas age and previous live births were modestly associated. The risk of having a cervical score less than 2.0 cm among women aged 30 and older was 1.3 times (95% CI 0.8, 2.1) that of younger subjects; among women who had had previous births, the risk was also 1.3 times (95% CI 1.1, 1.6) that of nulligravid subjects. Among women who had previously given birth prematurely but carried to term in the study pregnancy, the prevalence of cervical scores less than 2.0 cm was increased 1.9-fold; the prevalence of dilatation at least 0.5 cm was increased 2.7-fold; and the prevalence of length up to 1.5 cm was 1.8-fold that of the prevalence among women who had never had a preterm birth. Although the probability of abnormal cervical examination findings was higher among those with histories of preterm birth, it did not explain the association of parity or previous live birth with abnormal findings. When women with histories of preterm birth were excluded, the point estimates for the association between parity and live birth were stable.
Cervical scores were related to the time from assessments of cervical dimensions to delivery (Figure 1). The mean time to delivery for scores greater than 3 (n = 457) was 12 ± 2.4 weeks; for scores of 2–3 (n = 342), 11.8 ± 2.4 weeks; for scores of 1–2 (n = 66), 10.8 ± 3.2 weeks; and for scores less than 1 (n = 13), 10.3 ± 4.2 weeks. Compared with women who had cervical scores greater than 3, those with scores of 2–3 gave birth an average of 2.2 days earlier (95% CI −0.1, 4.6); those with scores of 1–2, an average of 9.5 days earlier (95% CI 5.1, 13.8); and those with scores less than 1, 11.9 days earlier (95% CI 2.2, 21.7). No deliveries occurred within 48 hours of cervical examinations.
Table 3 presents the performance characteristics of selected cutoff points, including dilatation up to 1.0 cm, length less than 2.0 cm, and cervical score less than 2, for use as screening tests to predict the risk of preterm birth. The sensitivity of all three was poor, at 8–20%, whereas the specificity was good, at 93–99%.
Our findings confirmed that cervical change is associated with and precedes preterm delivery in a general obstetric population of black and white women, independent of traditional markers of risk such as race and previous preterm birth.3,6,7,13,14 Analysis of preterm delivery by etiologic subtypes suggests that cervical effacement and changes in cervical score are associated more strongly with preterm PROM than preterm labor. To our knowledge, this difference has not been shown by others and has implications for further study of the etiologic pathways of these conditions. Given that secondary preventive measures believed to ameliorate the risk of preterm delivery (tocolysis and bed rest) are directed specifically at preterm labor, our observation that cervical change is strongly related to preterm birth after preterm PROM offers challenges for screening and preventive interventions if we are to reduce preterm births.
Our study has several limitations. The chief limitation is the small number of preterm events in each subtype, resulting in imprecise estimates of risk. Measurement by multiple examiners introduces variability that cannot be controlled. Nonetheless, such measurements reflect the expected clinical usefulness of cervical examination in similar populations and appear to be sufficient to rank the degree of abnormality among subjects. Although providers were unaware of our study hypothesis, we cannot assess the extent to which their findings modified subject management. If abnormal findings prompted intervention that was more likely to prevent preterm labor than preterm PROM, our results may be related in part to differential probabilities of prolonging gestation. The association between dilatation and effacement and the risk of preterm labor would thus be weakened by success in preventing or treating preterm labor.
Despite these weaknesses, the probability is higher that women who have preterm births will have had abnormal cervical examination findings than those who give birth at term. However, the benefit of early detection of abnormal findings is marginal. Each of the cervical measures evaluated had low sensitivity. Most women who deliver before term do not have cervical change at 24–29 weeks. Few women with effacement or dilatation go on to deliver before term, representing false positives. This confirms the work of Buekens et al18 who, in a large randomized, controlled trial, could not show improvement in perinatal outcomes from routine cervical examinations.
Nevertheless, assessment of the cervix may be important. We found that cervical examinations are quite specific, ie, the absence of cervical change correctly identifies women at low risk of preterm delivery. This finding suggests that a two-step evaluation of preterm delivery risk might be fruitful. Sequential testing amplifies the specificity of screening and identifies a smaller group for heightened surveillance. In the first step, women would have cervical examinations, and those without cervical change, and thus at low risk for preterm delivery, could be given routine precautions and excluded from further immediate evaluation. In the second step, women with cervical change would undergo additional testing, such as salivary estriol,19 fetal fibronectin,20 and endovaginal sonography.21 Such stepped evaluation would, in theory, focus the use of more costly tools and increase the precision of risk assessment. If the association of cervical change with preterm PROM is supported by future investigations and in other populations, closer surveillance and more aggressive treatment of infections would also be indicated for those with cervical change. However, obstetricians must be wary of intuitively appealing prevention programs that lack supportive evidence. We must first test screening systems in prospective studies and then prove that perinatal outcomes are enhanced by foreknowledge of impending preterm delivery.
1. Wood C, Bannerman RHO, Booth RT, Pinkerton JHM. The prediction of premature labor by observation of the cervix and external tocography. Am J Obstet Gynecol 1965;91:396–402.
2. Anderson AB, Turnbull AC. Relationship between the length of gestation and cervical dilatation, uterine contractility, and other factors during pregnancy. Am J Obstet Gynecol 1969;105:1207–14.
3. Papiernik E, Bouyer J, Collin D, Winisdoerffer G, Dreyfus J. Precocious cervical ripening and preterm labor. Obstet Gynecol 1986;67:238–42.
4. Blondel B, Le Coutour X, Kaminski M, Buekens P. Routine vaginal examinations. Lancet 1988;ii:846–7.
5. Alexander GR, Weiss J, Hulsey TC, Papiernik E. Preterm birth prevention: An evaluation of programs in the United States. Birth 1991;18:160–9.
6. Blondel B, Le Coutour X, Kaminski M, Chavigny C, Breart G, Sureau C. Prediction of preterm delivery: Is it substantially improved by routine vaginal examination? Am J Obstet Gynecol 1990;162:1042–8.
7. Copper RL, Goldenberg RL, Davis RO, Cutter GR, DuBard MB, Corliss DK, et al. Warning symptoms, uterine contractions, and cervical examination findings in women at risk of preterm delivery. Am J Obstet Gynecol 1990;162:748–54.
8. Neilson JP, Verkuyl DAA, Crowther CA, Bannerman C. Preterm labor in twin pregnancies: Prediction by cervical assessment. Obstet Gynecol 1988;72:719–23.
9. Newman RB, Godsey RK, Ellings JM, Campbell BA, Eller DP, Miller MC 3d. Quantification of cervical change: Relationship to preterm delivery in multifetal gestation. Am J Obstet Gynecol 1991;165:264–71.
10. Houlton MCC, Marivate M, Philpott RH. Factors associated with preterm labour and changes in the cervix before labour in twin pregnancy. Br J Obstet Gynaecol 1982;89:190–4.
11. Stubbs TM, Van Dorsten JP, Miller MC. The preterm cervix and preterm labor: Relative risks, predictive values, and change over time. Am J Obstet Gynecol 1986;155:829–34.
12. Leveno KJ, Cox K, Roark ML. Cervical dilatation and prematurity revisited. Obstet Gynecol 1986;68:434–5.
13. Bouyer J, Papiernik E, Dreyfus J, Collin D, Winisdoerffer B, Gueguen S. Maturation signs of the cervix and prediction of preterm birth. Obstet Gynecol 1986;68:209–14.
14. Mortensen OA, Franklin J, Lofstrand T, Svanberg B. Prediction of preterm birth. Acta Obstet Gynecol Scand 1987;66:507–12.
15. Klebanoff MA. Conceptualizing categories of preterm birth. Prenat Neonat Med 1998;3:13–5.
16. Savitz DA, Blackmore CA, Thorp JM. Epidemiologic characteristics of preterm delivery: Etiologic heterogeneity. Am J Obstet Gynecol 1991;164:467–71.
17. Savitz DA, Dole N, Williams J, Thorp JM, McDonald T, Carter AC, et al. Determinants of participation in an epidemiologic study of preterm delivery. Paediatr Perinat Epidemiol 1999;13:114–25.
18. Buekens P, Alexander S, Boutsen M, Blondel B, Kaminski M, Reid M. Randomised controlled trial of routine cervical examinations in pregnancy. European Community Collaborative Study Group on Prenatal Screening. Lancet 1994;344:841–4.
19. McGregor JA. Salivary estriol as risk assessment for preterm labor: A prospective trial. Am J Obstet Gynecol 1995;173:1337–42.
20. Peaceman AM, Andrews WW, Thorp JM, Cliver SP, Lukes A, Iams JD, et al. Fetal fibronectin as a predictor of preterm birth in patients with symptoms: A multicenter trial. Am J Obstet Gynecol 1997;177:13–8.
21. Iams JD, Goldenberg RL, Meis PJ, Mercer BM, Moawad A, Das A, et al. The length of the cervix and the risk of spontaneous premature delivery. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. N Engl J Med 1996;334:567–72.