Академический Документы
Профессиональный Документы
Культура Документы
org
OBJECTIVE: The purpose of this study was to assess funnel type and els that controlled for randomization cervical length and cerclage,
pregnancy duration in women with previous spontaneous preterm birth women with U-shaped funnel delivered earlier than women with either
and cervical length ⬍25 mm. V-shaped funnel or no funnel. Interaction between cerclage and
STUDY DESIGN: We performed a secondary analysis of a multicenter U-shaped funnel was observed, and analyses that were stratified by
randomized trial of cerclage. At the randomization scan that docu- cerclage showed that women with a U-shaped funnel and cerclage de-
mented short cervix, the presence and type of funnel (U or V) were livered at a mean of 33.8 ⫾ 6.6 weeks of gestation, compared with
recorded. women who did not receive cerclage (28.9 ⫾ 6.9 weeks of gestation).
RESULTS: One hundred forty-seven of 301 women (49%) had funnel- CONCLUSION: U-shaped funnels in high-risk women with a short cervix
ing: V-shaped funnel, 99 women; U-shaped funnel, 48 women. are associated with earlier birth.
U-shaped funnel was associated significantly with preterm birth at
⬍24, ⬍28, ⬍35, and ⬍37 weeks of gestation. In multivariable mod- Key words: preterm birth, U-shaped funnel, V-shaped funnel
Cite this article as: Mancuso MS, Szychowski JM, Owen J, et al. Cervical funneling: effect on gestational length and ultrasound-indicated cerclage in high-risk
women. Am J Obstet Gynecol 2010;203:259.e1-5.
differently to cerclage intervention. The cerclage after random assignment for tests and logistic regression. Time to de-
aim of this study was to assess the rela- the clinical diagnosis of acute cervical livery was estimated with the Kaplan-
tionship between the type of cervical insufficiency. Meier method, and group differences
funneling and pregnancy duration in Cervical length was measured by the were evaluated with the log-rank statis-
women with previous spontaneous PTB standard technique as described by Iams tic. Descriptive statistics for these 3
and cervical length of ⬍25 mm who were et al.4 Trial sonologists underwent a uni- groups of patients were compared using
enrolled in a randomized intervention form certification process by a single in- analysis of variance for continuous mea-
trial of ultrasound-indicated cerclage. vestigator (J.O.) to ensure uniformity in sures and chi-square tests for categoric
sonographic equipment, measurement measures. Multivariable linear regres-
technique, completion of study forms, sion, logistic regression, and Cox pro-
M ATERIALS AND M ETHODS and adherence to protocol. The cervical portional hazards models were then con-
This was a planned, secondary analysis of length at each visit was measured along a sidered for gestational age, rates of PTB,
the National Institute of Child Health closed endocervical canal. Minimal de- and time to delivery, respectively. An al-
and Human Development–sponsored grees of apparent dilation ⬍5 mm were pha level of .05 was selected to represent
randomized trial of cerclage for PTB pre- considered closed. After a baseline cervi- statistical significance for main effects
vention that was performed by a consor- cal length was measured, fundal pressure and .10 for interactions. All analyses
tium of 15 U. Clinical Centers between was applied for 30 seconds as a provoca- were performed using SAS software (ver-
January 2003 and November 2007.11 tive maneuver; each scan included an sion 9.2; SAS Institute Inc, Cary, NC).
Healthy, multiparous women who en- evaluation period of at least 5 minutes to The protocol and data forms were re-
rolled for prenatal care before 22 weeks detect spontaneous cervical shortening. viewed and approved by the human-use
of gestation were screened to identify the The shortest cervical length at each ex- committees at all participating centers.
women with at least 1 previous sponta- amination was recorded as the cervical
neous PTB between 170/7 and 336/7 length, regardless of whether the mea-
weeks of gestation. surement was obtained with pressure or R ESULTS
Exclusion criteria were fetal anomaly, was the result of spontaneous dynamic Of the 1044 women who had a qualifying
planned history-indicated cerclage for a shortening. previous PTB, 1014 women (99%) gave
clinical diagnosis of cervical insuffi- During the ultrasound examination, informed consent and underwent their
ciency, acute cervical insufficiency (de- the presence and type of funnel (U- or initial sonographic assessment of cervi-
fined as 2-cm dilation and visible mem- V-shaped) were recorded. Cervical fun- cal length. From this cohort, we observed
branes at the external os), and clinically neling was defined as the protrusion of 318 women who experienced cervical
significant maternal-fetal complications. the amniotic membranes of ⱖ5 mm into length shortening of ⬍25 mm. Sixteen
Eligible women were invited to consent the internal os, as measured along the patients were excluded (13 women did
for the ultrasound screening phase of the lateral border of the funnel. Care was not consent to randomization; 2 women
trial. Other details of the study protocol taken to differentiate between a true fun- were ruled ineligible at the randomiza-
are described elsewhere.11 nel and a pseudo-funnel. A pseudo-fun- tion visit, and 1 woman withdrew from
Consenting women underwent serial nel may occur when the lower uterine the trial), which left 302 (95%) who were
transvaginal sonographic evaluations, segment forms what appears to be a fun- assigned randomly to the no-cerclage (n
the first of which was scheduled in the nel above an otherwise normal-length ⫽ 153) or cerclage groups (n ⫽ 149). Pri-
temporal window of 160/7 to 216/7 weeks cervix.12 mary outcome information was avail-
of gestation. Subsequent scans were Study outcomes included gestational able for all 153 women in the no-cerclage
scheduled every 2 weeks, unless the cer- age at birth, rates of PTB at several gesta- group and for 148 of 149 women in the
vical length was observed to be 25-29 tional age cutoffs, and time to birth that cerclage group, which left a total of 301
mm, after which scans were scheduled was assessed by survival analysis. Gesta- women in the analysis.11 As depicted in
on a weekly basis. Women with a cervical tional age at delivery was modeled as a Figure 1, of the 153 women who were
length that remained at least 25 mm by function of funnel type in a simple linear assigned to the no-cerclage group, 14
the final sonographic evaluation, which regression model. The cervical length women underwent cerclage placement; 4
was scheduled to be no later than 226/7 and funnel type at the qualifying evalua- cerclage placements were at the discre-
weeks, were ineligible for random as- tion for randomization were recorded. tion of their treating physicians (off-pro-
signment and resumed their obstetric Women were classified as having U-fun- tocol treatment crossover) , and 10 cer-
care. If on any evaluation the cervical nel, V-funnel, or no cervical funneling. clage placements were undertaken for a
length was ⬍25 mm, the woman became For this study, we considered actual cer- diagnosis of acute cervical insufficiency
eligible for the random assignment to ei- clage placement, not assigned random- (protocol-sanctioned treatment cross-
ther receive a McDonald cerclage or to ization group. over). Similarly, of the 149 women who
enter a no cerclage group. Women who Multiple and pairwise comparisons were assigned to receive cerclage, 11
were assigned to no cerclage could re- for the rates of PTB in these funnel clas- women did not undergo surgery; 8
ceive a physical examination–indicated sification groups were evaluated with 2 women declined to undergo surgery,
TABLE 1
Demographic characteristics
Funnel type
Variable None (n ⴝ 154) V (n ⴝ 99) U (n ⴝ 48) P value
Actual cerclage placement, n (%) 67 (44) 53 (54) 32 (67) .015
................................................................................................................................................................................................................................................................................................................................................................................
a
Race/ethnicity, n (%) .14
.......................................................................................................................................................................................................................................................................................................................................................................
Black (non-Hispanic) 89 (58) 62 (62) 22 (46)
.......................................................................................................................................................................................................................................................................................................................................................................
White (non-Hispanic) 31 (20) 15 (15) 7 (15)
.......................................................................................................................................................................................................................................................................................................................................................................
Hispanic 23 (15) 11 (11) 10 (21)
.......................................................................................................................................................................................................................................................................................................................................................................
Other 11 (7) 11 (11) 9 (10)
................................................................................................................................................................................................................................................................................................................................................................................
Maternal age, y b
25.9 ⫾ 5.4 26.5 ⫾ 5.1 28.2 ⫾ 4.8 .03
................................................................................................................................................................................................................................................................................................................................................................................
c
Previous births, n 2 (1–4) 2 (1–4) 1 (1–3) .15
................................................................................................................................................................................................................................................................................................................................................................................
Earliest previous preterm birth, wk b
25.5 ⫾ 4.6 23.9 ⫾ 4.7 21.5 ⫾ 3.9 ⬍ .0001
................................................................................................................................................................................................................................................................................................................................................................................
First vaginal sonogram, wk b
17.5 ⫾ 1.3 17.2 ⫾ 1.2 17.4 ⫾ 1.5 .22
................................................................................................................................................................................................................................................................................................................................................................................
Randomization, wk b
19.4 ⫾ 2.0 19.5 ⫾ 2.0 19.3 ⫾ 1.9 .91
................................................................................................................................................................................................................................................................................................................................................................................
Baseline cervical length at randomization visit, mm b
23.6 ⫾ 5.3 20.3 ⫾ 5.6 14.2 ⫾ 7.2 ⬍ .0001
................................................................................................................................................................................................................................................................................................................................................................................
Shortest cervical length at randomization visit, mm b
21.1 ⫾ 4.3 18.8 ⫾ 5.2 13.0 ⫾ 6.7 ⬍ .0001
................................................................................................................................................................................................................................................................................................................................................................................
a
Race and ethnic group were self-reported; Values are given as means ⫾ 1 SD; Values are given as median (interdecile range).
b c
Mancuso. Cervical funneling: effect on gestational length and cerclage. Am J Obstet Gynecol 2010.
ear regression model, women with a U- U-funnel and no-funnel cohorts, beta REFERENCES
shaped funnel and cerclage delivered a error should not be a concern and should 1. Surbek DV, Hoeslie IM, Holzgreve W. Mor-
mean 4.9 weeks later in gestation than not affect our results and conclusions. phology assessed by transvaginal ultrasonog-
raphy differs in patients in preterm labor with vs
those women with a U-funnel who did U-shaped funnels may have a differ-
without bacterial vaginosis. Ultrasound Obstet
not receive cerclage. ent pathophysiologic mechanism than Gynecol 2000;15:242-5.
We acknowledge that funnel identifi- V-shaped funnels because they appear to 2. Spong C. Prediction and prevention of recur-
cation and characterization of shape is be associated with significantly earlier rent spontaneous preterm birth. Obstet Gy-
somewhat subjective and that the recog- birth when compared with no funnel or necol 2007;110:415.
V-shaped funnels. Women with a 3. Tucker JM, Goldenberg RL, Davis RO, Cop-
nition and characterization of a funnel
per RL, Winkler CL, Hauth JC. Etiologies of pre-
may not be highly reproducible. In addi- U-funnel and short cervix also appear to term birth in an indigent population: is preven-
tion to the fact that previous reports on have enhanced benefit from ultrasound- tion a logical expectation? Obstet Gynecol
cervical funneling did not systematically indicated cerclage, when compared with 1991;77:343-7.
distinguish between funnel types, this V-shaped funnel or no funnel. The de- 4. Iams JD, Goldenberg RL, Meis PJ, et al. The
might explain the inconsistent effects of velopment of a U-funnel may be in the length of the cervix and the risk of spontaneous
premature delivery. N Engl J Med 1996; 334:
funneling in the literature. However, for evolution pathway of acute cervical in-
567-72.
this trial, sonologists were trained specif- sufficiency, which might explain the dis- 5. Owen J, Yost N, Berghella V, et al. Mid-tri-
ically in the recognition and character- proportionate benefit from cerclage in- mester endovaginal sonography in women at
ization of funnels. Cervical funneling was tervention when compared with women high risk for spontaneous preterm birth. JAMA
clearly defined in the study protocol as with no funnel or a V-funnel.14 Ten 2001;286:1340-8.
women who were not assigned randomly 6. Burger M, Weber-Rossler T, Willmann M.
the protrusion of the amniotic mem- Measurement of the pregnant cervix by trans-
branes of ⬎5 mm into the internal os to cerclage intervention later received a vaginal sonography: an interobserver study and
that was measured along the lateral bor- physical examination–indicated cerclage new standards to improve the interobserver
der of the funnel. All sonologists were after they were found to have clinical ev- variability. Ultrasound Obstet Gynecol 1997;
trained and certified by a single investi- idence of acute cervical insufficiency. 9:188-93.
Seven of those women (70%) had a U- 7. Sonek JD, Iams JD, Blumenfeld M, Johnson
gator; the training included those aspects
F, Landon M, Gabbe S. Measurement of cervi-
of funnel recognition that are necessary shaped funnel on their randomization
cal length in pregnancy: Comparison between
to differentiate a V- from a U-shape. In scan. This further supports our pos- vaginal ultrasonography and digital examina-
particular U-shaped funnel assessment tulation that a U-shaped funnel may tion. Obstet Gynecol 1990;76:172-5.
required both the funnel depth and be in the pathway of acute cervical 8. Yost NP, for the NICHD MFM Units Network.
insufficiency. Midtrimester endovaginal sonography: ancillary
width to be measured, whereas V-shaped
These findings have important clin- markers for spontaneous preterm birth. Obstet
funnels required only measurement of Gynecol 2004;103:457-62.
depth. Such training should help to re- ical implications. High-risk women with 9. Rust OA, Atlas RO, Kimmel S, Roberts WE,
duce the subjectivity in clinical practice; a previous spontaneous PTB and short Hess LW. Does the presence of a funnel in-
however, sonologist interpretation of cervix benefit from cerclage; this ben- crease the risk of adverse perinatal outcome in
funneling was not reviewed by the pri- efit increases as the cervical length de- a patient with a short cervix? Am J Obstet Gy-