Professional Documents
Culture Documents
Amanda Roman, MD, Noelia Zork, MD, Sina Haeri, MD, MHSA, Corina N. Schoen,
MD, Gabriele Saccone, MD, Sarah Colihan, MD, Craig Zelig, MD, Alexis C.
Gimovsky, MD, Neil S. Seligman, MD, Fulvio Zullo, MD, Vincenzo Berghella, MD
PII: S0002-9378(20)30672-4
DOI: https://doi.org/10.1016/j.ajog.2020.06.047
Reference: YMOB 13338
Please cite this article as: Roman A, Zork N, Haeri S, Schoen CN, Saccone G, Colihan S, Zelig C,
Gimovsky AC, Seligman NS, Zullo F, Berghella V, Physical Exam Indicated Cerclage in Twin pregnancy:
a Randomized Controlled Trial, American Journal of Obstetrics and Gynecology (2020), doi: https://
doi.org/10.1016/j.ajog.2020.06.047.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
2 Controlled Trial
3 Authors: Amanda ROMAN, MD (1), Noelia ZORK, MD (2), Sina HAERI, MD, MHSA
4 (3), Corina N. SCHOEN, MD (4) Gabriele SACCONE, MD (5), Sarah COLIHAN, MD (6),
5 Craig ZELIG, MD (6), Alexis C. GIMOVSKY, MD (7); Neil S. SELIGMAN, MD (8), Fulvio
7 (1) Maternal Fetal Medicine Division, Obstetrics and Gynecology Department, Sidney
9 States.
10 (2) Maternal Fetal Medicine Division, Obstetrics and Gynecology Department, Columbia
21 DC
2
25 Financial Support: No external financial support was received for this study.
26 Data from this manuscript was presented at the 40th SMFM Meeting, February 3–8,
28
29 Corresponding author
30 Amanda Roman, MD
35 Email: amanda.roman@jefferson.edu
36 Phone: 215-955-9200
37 Fax: 215-955-5041
38
39 Word count
40 Abstract: 447
42
43
3
45 asymptomatic twin pregnancies <24 weeks significantly decreases preterm birth at all
47 Short Title: Physical exam indicated cerclage in twin pregnancy decreases preterm
48 birth
49 AJOG at a Glance:
52 in twin pregnancy
55 birth at all gestational age cutoff studied and decreases perinatal morbidity
56 and mortality
58 • This study is the first randomized controlled trial evaluating physical exam
60
64 https://clinicaltrials.gov/ct2/show/NCT02490384?cond=twin+cerclage&draw=2&rank=5
65
4
66 ABSTRACT
67 Background: Twin pregnancies with dilated cervix in the second trimester are at
68 increased risk of pregnancy loss and early preterm birth; currently there is no proven
71 preterm birth in asymptomatic women with twin gestations and cervical dilation
74 women with twin pregnancy, and asymptomatic cervical dilation from 1-5 cm between
75 16 0/7 to 23 6/7 weeks were enrolled from 7/2015 to 7/2019 in 8 centers. Eligible
79 placenta previa, signs of labor, or clinical chorioamnionitis. The primary outcome was
80 the incidence of preterm birth <34 weeks. Secondary outcomes were preterm birth <32,
81 <28 and <24 weeks, interval from diagnosis to delivery, and perinatal mortality. Data
83 Results: After an interim analysis was performed, the Data Safety Monitoring Board
84 recommended stopping the trial due to significant decrease of perinatal mortality in the
85 cerclage group. We randomized 34 women, four were excluded due to expired informed
86 consent. Seventeen women were randomized to physical exam indicated cerclage and
87 13 women to no cerclage. Four women randomized to cerclage did not receive the
5
89 Maternal demographics were not significantly different. All women in the cerclage group
91 group, the incidence of preterm birth <34 weeks was significantly decreased: 12/17
92 (70%) vs 13/13 (100%) (RR: 0.71 95% CI 0.52-0.96), as well as preterm birth <32
93 weeks 11/17 (64.7%) vs 13/13 (100%) (RR: 0.65 95% CI 0.46-0.92), preterm birth <28
94 weeks 7/17 (41%) vs 11/13 (84%) (RR: 0.49 95% CI 0.26-0.89) and preterm birth <24
95 weeks 5/17 (30%) vs 11/13 (84%) (RR: 0.35 95% CI 0.16-0.75). The mean gestational
96 age at delivery was later: 29.05±1.7 vs. 22.5±3.9 weeks (p<0.01); the mean interval
97 from diagnosis of cervical dilation to delivery was longer: 8.3±5.8 vs. 2.9±3.0 weeks
98 (p=0.02). Perinatal mortality was also significantly reduced in the cerclage group 6/34
102 preterm birth at all evaluated gestational ages. Most importantly, cerclage in this
103 population is associated with a 50% decrease in very early preterm birth <28 weeks and
105 Keywords: Physical exam indicated cerclage, twins, cerclage, cervical dilation, preterm
107
108
109
6
110 INTRODUCTION
111 Twins pregnancies represent a small proportion of annual births in the United States but
112 account for 20% of preterm births. The 2017 twin birth rate in the United States was
113 33.3 twins per 1,000 births but 59.4% of twin pregnancies delivered before 37 weeks,
114 and 9.45% before 32 weeks, compared with singleton pregnancies, of whom 8.13%
115 delivered before 37 weeks and 2.12% before 32 weeks1. Twins are also at increased
116 risk for low birth-weight (LBW), have five times higher risk of early neonatal and infant
118 Singleton pregnancies with cervical dilation of 1 cm or more in the second trimester are
119 associated with a poor prognosis. Greater than 90% will result in pretem birth (PTB),
120 regardless of history of PTB, cervical length, obstetric history, or other PTB risk
121 factors.3-7 Physical exam indicated cerclage, previously called at times rescue,
123 or more cervical dilation detected on physical examination (speculum or digital exam).8,
9
124 Placement of cerclage in patients with singleton gestations who have cervical dilation
125 in the second trimester has been shown to decrease incidences of PTB and perinatal
127 In twins, the data on the efficacy of physical exam indicated cerclage is limited. The only
128 published randomized controlled trial (RCT) evaluating cerclage for cervical dilation
129 compared 13 women who received cerclage, indomethacin and bed rest, and 10
130 women who received bed rest-only; all women received antibiotics. This RCT included 7
131 twin pregnancies, 3 in the cerclage and 4 in the bed rest group. This RCT showed
132 overall a significant decrease in PTB < 34 weeks of gestation and longer latency interval
7
133 from diagnosis to delivery (by 30 days). However, outcomes of twin pregnancies were
135 Multiple case series and cohorts studies on physical exam indicated cerclage in twins or
136 twins compared with singleton pregnancies have been published.12 The three
137 retrospective case-control studies in twins, where physical exam indicated cerclage was
139 <34, and <32 weeks, two of the studies also had significant spontaneous PTB < 28
140 weeks12, 13 and one of them showed a decrease in the incidence of perinatal mortality.12
141 Our aim was to evaluate if physical exam indicated cerclage would reduce the rate of
142 spontaneous PTB before 34 weeks’ gestation and adverse perinatal outcomes in
143 women with twin pregnancy, and asymptomatic cervical dilation from 1-5 cm before 24
145
148 This multicenter, parallel group, open-label, randomized controlled trial was performed
149 by a consortium of 8 clinical centers between July 2015 and July 2019. The Institutional
150 Review Board (IRB) approved this study at each center, and all participants were
151 provided with and signed written informed consent. This trial was registered as a
152 randomized clinical trial (NCT02490384; clinicaltrials.gov). This trial had no external
153 funding. Eligible participants were women between 18 - 50 years of age carrying a
8
154 diamniotic twin gestation with asymptomatic cervical dilation from 1 to 5 cm and/or
155 visible membranes by pelvic exam or speculum exam between 16 0/7 to 23 6/7 weeks.
156 Women were identified either by transvaginal ultrasound cervical length (TVUCL) at the
157 time of the anatomy scan or by pelvic examination during prenatal visits. We excluded
159 twin-twin transfusion syndrome, major fetal malformation or known genetic anomaly,
160 placenta previa, regular painful contractions with cervical changes over 12 hours of
161 observation, active bleeding or clinical chorioamnionitis at the time of diagnosis, cervical
162 dilation 6 cm or more, amniotic membranes prolapsed beyond external os into the
163 vagina, unable to visualize cervical tissue, fetal reduction after 14 weeks from higher-
164 order multiples, ruptured membranes prior to randomization, and cerclage already in
165 place.
166 All pregnancies were dated by crown-rump length during the first trimester, and
167 chorionicity was determined. Before randomization, they underwent anatomy scan for
169 dilation was determined by pelvic exam and/or speculum exam between 16 0/7 and 23
170 6/7 weeks and confirmed to have a dilated cervix from 1-5 cm by the primary
171 investigator at each site. Premature preterm rupture of the membranes (PPROM)
172 defined as rupture of membranes that occurred prior of labor and delivery but not
174 speculum exam, a positive Nitrazine test and/or a ferning pattern on microscopy. Active
175 labor was defined as the presence of regular uterine contractions 3 or more in 10
177 amniocentesis of the presenting twin with: 1) presence of any bacteria on amniotic fluid
178 Gram stain, 2) amniotic fluid leukocyte count ≥ 6 leukocytes per high-power field or >30
179 cells/mm3, 3) amniotic fluid glucose concentration ≤15 mg/dL; or by positive amniotic
180 fluid culture. Clinical chorioamnionitis was defined as: maternal fever ≥38°C (≥100.4° F)
181 plus 1 of the following: maternal tachycardia (>100 beats/min), fetal tachycardia, (>160
182 beats/min), marked leukocytosis (>15,000 cells/mm3), uterine tenderness, or foul odor
183 of the amniotic fluid.15 Amniocentesis was offered to all patients before randomization;
184 however, it was not a prerequisite for enrollment. Patients who declined amniocentesis
185 were observed for approximately 12 hours for signs of labor, bleeding, or infection, and
186 those who remained stable were offered randomization. At the time of pelvic exam,
187 specimens for gonorrhea, chlamydia, group B streptococcus (GBS), bacterial vaginosis
188 and trichomonas’s were collected. Gonorrhea and chlamydia were identified by
189 polymerase chain reaction (PCR). The diagnosis of bacterial vaginosis was done by
190 Amsel criteria,16 (at least three criteria must be present): 1) homogeneous, thin, grayish-
191 white discharge that smoothly coats the vaginal walls, 2) vaginal pH >4.5, 3) positive
192 whiff-amine test, defined as the presence of a fishy odor when a drop of 10% potassium
193 hydroxide (KOH) is added to a sample of vaginal discharge and 4) clue cells on saline
194 wet mount (vaginal epithelial cells studded with adherent coccobacilli at the edge of the
195 cell in at least 20% of the epithelial cells on wet mount). Trichomonas were identified on
196 wet mount. Urinary tract infections (UTI) were diagnosed by positive urine culture. All
197 detected infections were treated. Initiation of indomethacin during the observation
198 period was acceptable. After delivery, the placentas were sent to pathology. Histologic
199 chorioamnionitis was defined by the presence of neutrophils that marginated into the
10
200 placental chorionic plate; funisitis was defined by the presence of neutrophils in the
203 Eligible participants were randomly allocated in a 1:1 ratio by blocks of 4 to either
205 (Research Electronic Data Capture) hosted at Thomas Jefferson University.18, 19 The
206 sequence of randomization was set by a third party. Principal investigators at each site
207 were masked to the allocation of patients; institutions were only able to see their own
208 patients. Due to small sample size of the study, enrollment was not stratified by
210 Interventions
211 Physical exam indicated cerclage was to be placed by a trained physician in the
212 operating room under anesthesia (regional or general at the discretion of the
214 indomethacin, antibiotics before and/or after the cerclage, surgical technique, selection
215 of sutures, additional tocolysis, vaginal progesterone, admission to the hospital for
216 observation, or maternal physical activity after randomization were at the discretion of
219 Women on both arms cerclage and no cerclage were observed in the hospital until they
220 were stable for discharge, and preterm labor precautions were given. Readmission to
11
221 the hospital per physician discretion. Antenatal steroids were offered to all women ≥ 23
222 weeks, and a second course of steroids was offered if delivery was considered
223 imminent ≥ 2 weeks after the first course. Prenatal care continued according to the
224 local guidelines, there were no other study-specific recommendations for pregnancy
225 care.
226 Women were contacted monthly either by phone or in person at the time of their
227 prenatal visit. Consents for release of medical information were signed at the time of
228 randomization to gather information pertinent to the study in case of admissions to other
229 institutions. Fetal growth evaluations were done every 3 to 4 weeks; estimated fetal
230 weight calculations were based on the Hadlock, et al. growth curve.20 Patients
231 diagnosed with intrauterine growth restriction (IUGR) in one or both twins were followed
232 by biweekly non-stress test, weekly umbilical artery Doppler and maximal vertical
233 pocket on each side of the membrane. The primary care provider removed the cerclage
234 during the 36th week of gestation or sooner in the presence of active vaginal bleeding,
235 labor, PPROM, or membranes prolapsed through the suture. Delivery time and mode of
236 delivery were per obstetrical indications. After delivery, all neonates were followed until
239 The primary outcome was the incidence of spontaneous PTB <34 weeks of gestation.
240 Secondary maternal outcomes included: incidence of spontaneous PTB <32, <28, and
241 <24 weeks gestation, PPROM at <34 weeks gestation, GA at delivery, interval between
242 diagnosis and delivery, and birth weight. Secondary neonatal outcomes include: Apgar
12
243 score at 5 minutes, admission to neonatal intensive care unit (NICU), neonatal death,
244 length of stay (LOS) in the NICU until discharge home, and composite neonatal adverse
245 outcome, defined as at least 1 of the following: respiratory distress syndrome (RDS)
246 (ventilator support, intubation, continuous positive airway pressure [CPAP], or use of
249 necrotizing enterocolitis (NEC) grade 3 and 4 requiring surgery,22 proven sepsis
250 (clinically ill infant with suspected infection plus positive blood, cerebrospinal fluid [CSF],
255 rupture.
257 Calculation of sample size was based on a reduction in the incidence of PTB < 34
258 weeks from 80% in the no cerclage group to 40% in the cerclage group, with a power of
259 80%.12 To detect this difference at a significance level of 5%, we calculated 46 subjects
260 with 23 subjects in each arm, plus 10% for loss of follow up. A total sample size of 52
262 It was planned that the Data and Safety Monitoring Board (DSMB) would review data
263 relevant to safety (not efficacy) after approximately 50% of the subjects had delivered
264 and the DSMB would provide a recommendation as to whether the study should
13
266 stopping rules were: significantly different maternal sepsis with admission to the
267 intensive care unit, maternal death, maternal bleeding requiring blood transfusion, or
268 fetal or neonatal mortality associated with the cerclage placement when compared with
269 no cerclage. Adverse events were reported to the Human Research Protection Office
272 Statistical analysis was conducted using Statistical Package for Social Sciences (SPSS)
273 version. 22 (IBM Inc., Armonk, NY, USA). Data are shown as mean ± standard
275 were performed with the use of the chi-square test with continuity correction.
276 Comparisons between groups were performed with the use of the T-test to test group
277 means by assuming equal within-group variances for parametric data, and with the use
278 of Wilcoxon and Mann-Whitney tests for nonparametric data. The primary analysis was
280 incidence of primary and secondary outcomes were quantified by the crude relative ratio
281 (RR) or mean difference with 95% of confidence interval (CI). In addition to standard
282 logistic regression analysis, in which each fetus was treated as an independent unit, we
283 used a generalized model-model approach in which each twin pair was a cluster unit for
284 outcomes of neonates born alive. For these outcomes, intracluster correlation
285 coefficient was also estimated. ICC was calculated using the two-way mixed model with
286 95% CI. Risk of PTB was assessed with the use of Kaplan-Meier analysis, in which
287 gestational age (GA) was time scale, and delivery the event. Log-Rank Hazard Ratio
14
288 was estimated. For outcomes in which cerclage had a statistically significant beneficial
289 or harmful effect number needed to treat (NNT) was calculated. P value <0.05 was
290 considered statistically significant. This trial was reported following the CONSORT
291 guidelines.24
292
293 RESULTS
295 A total of 70 women were assessed for eligibility between July 2015 and July 2019
296 (Figure 1). Of these, 14 women had an exclusion criteria and 22 declined participation,
297 34 women provided informed consent and were randomized to either cerclage (n=18) or
298 from analysis when their consent form was identified to be expired, and a new consent
299 was not available due to lapse in the IRB documents renewal at two enrolling sites (1
300 woman in the cerclage group and 3 women in the no cerclage group). Therefore a total
301 of 30 women were included in the final analysis (17 cerclage and 13 no cerclage). Four
302 women assigned to cerclage did not receive the surgical procedure: 2 of them due to
303 contraindications to the procedure, which occurred soon after randomization (rupture of
304 amniotic membranes, and vaginal bleeding), one due to friable cervix and the cerclage
305 placement was suspended, while one declined cerclage after being randomized. All four
306 patients were included in the intention to treat analysis. At the interim analysis
307 performed with 30 enrolled women (58% of the sample size), the DSMB recommended
308 stopping the trial due to significant decrease of perinatal mortality in the cerclage group.
15
309 Participant demographic characteristics were similar for each group (Table 1).
310 Gestational age at diagnosis and randomization was the same. Most twin gestations
312 11 (36.6%); 6 women (20%) had a history of singleton PTB; 24 (80%) women had one
313 or more TVUCL done as part of the screening for preterm birth; of them, 20 (66.6%) had
314 TVUCL ≤ 25 mm and 7 (23.3%) had TVUCL ≤15mm, while 19 (63.3%) received vaginal
315 progesterone due to the finding of short cervix. All 24 women subsequently were
316 identified as having a dilated cervix. The other 10 women were identified with dilated
317 cervix during a pelvic examination indicated by other symptoms (pain, pressure, or
318 vaginal discharge) but considered not in active labor. Four women (13%) agreed and
319 had an amniocentesis before randomization, and all were negative for intra-amniotic
320 infection. UTI was identified in 4 (13.3%) women; cervico-vaginal evaluation identified
321 bacterial vaginosis in 3 (10%) and chlamydia in 1 (3%), all patients received appropriate
323 with one suture was done on all patients. Manipulation of the amniotic membrane was
324 required on 9/14 (64%) with either Foley balloon or sponge. Mersilene™ tape was used
326 women who underwent cerclage, received antibiotics (12 cephalosporin, 1 clindamycin-
327 gentamycin and 1 azithromycin) and indomethacin (50-100mg loading dose followed by
328 25-50 mg every 6 hours for 48 hours). Intraoperative complications was present in only
329 one case when physical exam indicated cerclage was attempted but the cervix was
330 more effaced and friable than expected, therefore the procedure was stopped and
331 placement was considered unsuccessful. No serious adverse events were identified.
16
333 The primary outcome of spontaneous PTB <34 weeks of gestation was observed in
334 12/17 (70%) of women in the cerclage group vs 13/13 (100%) in the no cerclage group
335 (RR, 0.7, 95% CI, 0.46–0.96; p=0.05). Secondary outcomes of spontaneous PTB <32,
336 <28, and <24 weeks of gestation were also significantly decreased in the cerclage
337 group (Table 2). There were no cases of medically indicated PTB. The latency period
338 from diagnosis to delivery was significantly prolonged in the cerclage group by mean
339 difference of 5.6 (2.0-9.3) weeks (p=0.02). The cumulative percentage of participants
340 who did not give birth preterm was significantly higher in the cerclage group than in the
341 no cerclage group (Log-Rank Hazard Ratio 0.33 [95%CI 0.13-0.80], p<0.001), (Figure
342 2). While more women in the cerclage delivered between 23 0/7 to 27 6/7 weeks: 5/17
343 (29.4%) vs. 1/13 (7.7%), this difference was not significantly different, p=0.2. Overall
344 perinatal mortality was observed in 6/34 (17.6%) of cerclage group versus 20/26 (77%)
345 in the no cerclage group RR 0.23 (0.11-0.49), p< 0.0001. There were no cases of fetal
346 demise. All perinatal mortality cases were associated with delivery < 24 weeks, (Table
347 2). Delivery prior to 23 weeks was seen in 2/17 (11.7%) women in the cerclage group
348 and 10/13 (76.9%) in the no cerclage group, all neonates died secondary to extreme
349 prematurity and received palliative care. Two women (11.7%) in the cerclage group
350 delivered between 23 0/7 and 23 6/7 weeks, they were born alive but one twin of each
351 women died in the first 48 hours (one secondary to sepsis and the other due to
352 respiratory failure). There were no significant differences in other neonatal outcomes,
353 but the trial was not powered for these outcomes (Table 3).
354
17
355 COMMENT
357 In women with twin pregnancies and asymptomatic cervical dilation from 1 to 4 cm
358 before 24 weeks, a combination of physical exam indicated cerclage, indomethacin, and
359 antibiotics significantly decreases the incidence of spontaneous PTB at all evaluated
360 GA providing a longer latency period from diagnosis to delivery by 5.6 weeks. Most
361 importantly, cerclage in these population was associated with a 50% decrease in very
362 early PTB <28 weeks, and a 78% decrease in perinatal mortality.
363 Results
364 Twin pregnancy with asymptomatic cervical dilation before 24 weeks offers a formidable
365 challenge for obstetricians. Physical exam indicated cerclage in twins with dilated cervix
366 before 24 weeks has been evaluated in only three retrospective studies in twins with no
367 cerclage as a control group, all of them favored cerclage placement. Our results are
368 similar to these studies. Roman et al, published in 201612 was the first case control
369 study of twins with cervical dilation ≥ 1 cm before 24 weeks from 6 different institutions
370 in the United States and one in Italy. The study showed that the combination of physical
371 exam indicated cerclage, indomethacin, and antibiotics were associated with a
372 significant prolongation of latency from diagnosis to delivery by 6.7 weeks, decreased
374 mortality by 76% and improved perinatal outcome when compared with no cerclage.
375 Abbasi et al.13 published a retrospective cohort study conducted at a single institution in
376 Toronto. This study included women with twin pregnancies, dilated cervix, and intact
18
377 membranes <25 weeks of gestation, 27 women had cerclage compared to 9 women
378 managed expectantly. Compared with the subjects enrolled in this study, their cohort
379 had more advanced GA and cervical dilation at diagnosis as 16 (44%) of them were ≥ 3
381 in either group. Women in the cerclage group experienced longer latency from
382 presentation to delivery (7.3±5.5 versus 0.5±0.3 weeks, p<.001) and a lower incidence
383 of PTB before 34, 32 and 28 weeks, while all women in the no cerclage group delivered
384 <28 weeks. The overall neonatal survival at discharge was 38/54 (70.4%) in the
385 cerclage group versus 10/18 (55.6%) in the no cerclage group (p=0.2). Han et al.14
386 published a cohort study from two hospitals in California that included twins with cervical
387 dilation > 1cm; the study included 54 women who received physical exam indicated
388 cerclage, and 17 women no cerclage. Maternal demographics were not presented
389 separately limiting comparison of the groups. Nonetheless, the use of cerclage was
390 associated with a significant decrease in the odds of PTB <32, <34, and <36 weeks.
391 Physical exam indicated cerclage in our study had no intraoperative complications; in
392 one case, the surgeon was unable to place the cerclage due to effaced and friable
393 cervix. Larger cohorts in singleton pregnancies have informed a 10-20% risk of
394 intraoperative rupture of the membranes, cervical laceration, and bleeding during the
395 procedure.25
397 The strengths of our study: this is the first randomized study enrolling only twin
398 pregnancies with dilated cervix to cerclage or no cerclage. Expectant management with
19
399 no cerclage is the current standard of care for these women. There are no prior
400 dedicated RCTs in this population. Despite small sample size, we were able to show a
401 significant benefit to physical exam indicated cerclage. In our RCT, physical exam
402 indicated cerclage was associated with decreased perinatal mortality and overall
403 decreased in preterm birth at all gestational ages, however with the potential of more
404 early preterm deliveries. Some of the limitations include: small sample size as the
405 enrollment did not reach the intended 52 cases. The DSMB considered that it was
406 unethical to continue the study due to the considerable perinatal mortality in one of the
407 arms: 76.9% versus 17.6% (p<0.0001) and requested to unmask the arms of the study.
408 The primary and secondary outcomes were also significantly different, favoring the
409 cerclage group. Due to early termination of the study, we were unable to perform
410 additional sub analysis based on the degree of cervical dilation, GA before and after 20
411 weeks, use of progesterone or surgical techniques. The incidence of this challenging
412 obstetrical problem is very low. Only 3% of all deliveries are twins, of which about 5%
413 will have TVUCL ≤15mm,26 and even fewer will be identified with asymptomatic cervical
414 dilation ≥ 1cm before 24 weeks. The challenge of studying this clinical question is
415 highlighted by the fact that enrollment and randomization of 34 women was over a
417 characteristics were not significant different but some imbalance in the participant
418 characteristics were noticed suggesting bias in the intervention allocation as the
419 enrollment was not stratified by institution or any of the maternal characteristics (like
421 acknowledge that four women were excluded from the study due to expired consent.
20
422 We did not renew IRB documentation on time to have a new consent. There is also a
423 discrepancy between groups in withdrawals from the study (1 of 18 in the cerclage
424 group [5.6%] vs 3 of 16 in the no cerclage group [18.8%]). The outcomes of these four
426 The history of PTB in prior pregnancy was present in only 6 (20%) of our study group
427 while TVUCL ≤ 25mm preceded cervical dilation in 20/24 (83.3%) evaluated women. It
428 has been established that cervical length <25 mm is the best predictor of PTB in twin
429 pregnancy,27 and the shorter the TVUCL and the earlier GA at presentation, the higher
430 the risk of PTB.28 Cervical changes are dynamic, in singleton pregnancy TVUCL may
431 present with further shortening as pregnancy progresses; this is the bases of the
432 recommendation of serial TVUCL in women with history of preterm birth.29 In women
433 without risk for preterm birth, 30% of women with TVUCL <11mm had cervical dilation >
434 1cm at pelvic exam.30 Serial cervical evaluation have shown cervical shortening in twins
435 as well,31 however there are no studies assessing how many of them will have cervical
436 dilation. We did not performed serial TVUCL evaluation in twins and we are not able to
437 provide the incidence of cervical dilation. While TVUCL is currently the best tool
438 available to screen for PTB in twin pregnancy, independent of other risk factors,
439 ACOG32 and SMFM still recommends against TVCL in twins as there are no current
440 proven therapies to offer and TVUCL should be reserved for randomized controlled.
443 prior to physical exam indicated cerclage.33, 34 In our trial amniocentesis was not
444 mandatory as IRB found it coercive, only 4/30 (13.3%) participants agreed with the
21
445 procedure. Assessment of amniotic fluid gram stain or amount of glucose in patients
446 prior to physical exam indicated cerclage have shown a positive predictive value of
447 47%, and a negative predictive value of 98% for intrauterine infection.35 Amniotic fluid
448 culture is considered the “gold standard” for diagnosis of infection, but cultures take
449 several days for final result and therefore limit the utility in clinical decision-making.
450 Multiple studies have reported an association between preterm labor/delivery and
453 considering the ascending infection and colonization the most probable pathway.37 In
454 our study only 4/30 (13.3%) had positive vaginal bacterial studies and 4/30 (13.3%)
455 were diagnosed with UTI. Our study had limited financial resources for additional
456 vaginal microbial studies or evaluation of cytokines in amniotic fluid or vagina, as we did
457 not have external funding; those studies included in our study were considered standard
459 While it was not prespecified in our RCT, all women who received physical exam
460 indicated cerclage also received indomethacin and antibiotics. This management is
461 similar to previous studies. The RCT by Althuisius et al. in 2003,3 used a combination of
462 indomethacin and antibiotics in all cases randomized to cerclage. In Roman et al.,12
463 retrospective case-control study of physical exam indicated cerclage in twins, the
465 (94%) at the time of the cerclage placement. In a study by Miller et al.,38 a large
466 retrospective cohort of physical exam indicated cerclage in twins, 59 (57.3%) subjects
467 received indomethacin and 56 (54.3%) received antibiotics. Miller et al39 randomized
22
468 women with singleton pregnancies undergoing physical exam indicated cerclage to
469 indometacin and antibiotics (n=26) or no additional treatment (n=24). The median
470 latency from cerclage placement to delivery, GA at delivery, and neonatal outcomes
471 were no different between groups, but post hoc analysis revealed a greater proportion
472 of pregnancies were prolonged by at least 28 days among women who received
473 indomethacin and perioperative antibiotics (24 [92.3%] vs 15 [62.5%], p=0.01). In our
474 trial, the combination of cerclage, indomethacin, and antibiotics significantly decreased
475 the PTB at different GA cut-offs, increased median latency from cerclage placement to
476 delivery, and decreased perinatal mortality. Finally, the numerous secondary endpoints
477 with no adjustment for multiple comparisons could have led to type 1 error.
479 A prospective registry of twin pregnancies receiving physical exam indicated cerclage
480 would allow us to examine how differences in gestational age, cervical dilation, prolapse
481 of amniotic membranes, surgical technique, and association with other therapies like
482 progesterone supplementation may affect the prognosis and outcome. Prospective
483 studies on the incidence of twin diamniotic pregnancies with TVUCL < 25 mm before 24
484 weeks that presented posteriorly with dilated cervix and lastly prospective studies
485 evaluating therapies in diamniotic twins with short cervix. Our group is enrolling for an
486 RCT in twins with TVUCL <15mm comparing cerclage vs. no cerclage and vaginal
488 Conclusion
23
490 asymptomatic twin pregnancies before 24 weeks significantly decreases preterm birth
492
493 ACKNOWLEDGMENT
494 We thank the participant centers that screened for women with twin pregnancy at risk of
495 preterm birth but were unable to enroll patients: David McKenna, MD at Wright State
496 University, (Dayton, Ohio), Montse Palacio, MD at Hospital Clinic of Barcelona (Spain),
502
503
504
505
506
507
508
509
24
510 REFERENCES
511 1. Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, Drake P. Births: Final Data for 2017. National
512 vital statistics reports : from the Centers for Disease Control and Prevention, National Center for Health
513 Statistics, National Vital Statistics System 2018; 67(8): 1-50.
514 2. Matthews TJ, MacDorman MF, Thoma ME. Infant Mortality Statistics From the 2013 Period
515 Linked Birth/Infant Death Data Set. National vital statistics reports : from the Centers for Disease Control
516 and Prevention, National Center for Health Statistics, National Vital Statistics System 2015; 64(9): 1-30.
517 3. Althuisius SM, Dekker GA, Hummel P, van Geijn HP, Cervical incompetence prevention
518 randomized cerclage t. Cervical incompetence prevention randomized cerclage trial: emergency cerclage
519 with bed rest versus bed rest alone. Am J Obstet Gynecol 2003; 189(4): 907-10.
520 4. Novy MJ, Gupta A, Wothe DD, Gupta S, Kennedy KA, Gravett MG. Cervical cerclage in the second
521 trimester of pregnancy: a historical cohort study. Am J Obstet Gynecol 2001; 184(7): 1447-54; discussion
522 54-6.
523 5. Olatunbosun OA, al-Nuaim L, Turnell RW. Emergency cerclage compared with bed rest for
524 advanced cervical dilatation in pregnancy. International surgery 1995; 80(2): 170-4.
525 6. Daskalakis G, Papantoniou N, Mesogitis S, Antsaklis A. Management of cervical insufficiency and
526 bulging fetal membranes. Obstet Gynecol 2006; 107(2 Pt 1): 221-6.
527 7. Pereira L, Cotter A, Gomez R, et al. Expectant management compared with physical
528 examination-indicated cerclage (EM-PEC) in selected women with a dilated cervix at 14(0/7)-25(6/7)
529 weeks: results from the EM-PEC international cohort study. Am J Obstet Gynecol 2007; 197(5): 483 e1-8.
530 8. Creasy RK, Resnik R, Iams JD. Creasy and Resnik's maternal-fetal medicine : principles and
531 practice. 6th ed. Philadelphia, PA: Saunders/Elsevier; 2009.
532 9. American College of O, Gynecologists. ACOG Practice Bulletin No.142: Cerclage for the
533 management of cervical insufficiency. Obstet Gynecol 2014; 123(2 Pt 1): 372-9.
534 10. Ventolini G, Genrich TJ, Roth J, Neiger R. Pregnancy outcome after placement of 'rescue'
535 Shirodkar cerclage. J Perinatol 2009; 29(4): 276-9.
536 11. Ehsanipoor RM, Seligman NS, Saccone G, et al. Physical Examination–Indicated Cerclage: A
537 Systematic Review and Meta-analysis. Obstetrics & Gynecology 2015; 126(1): 125-35.
538 12. Roman A, Rochelson B, Martinelli P, et al. Cerclage in twin pregnancy with dilated cervix
539 between 16 to 24 weeks of gestation: retrospective cohort study. Am J Obstet Gynecol 2016; 215(1): 98
540 e1- e11.
541 13. Abbasi N, Barrett J, Melamed N. Outcomes following rescue cerclage in twin pregnancies(). J
542 Matern Fetal Neonatal Med 2018; 31(16): 2195-201.
543 14. Han MN, O'Donnell BE, Maykin MM, Gonzalez JM, Tabsh K, Gaw SL. The impact of cerclage in
544 twin pregnancies on preterm birth rate before 32 weeks. J Matern Fetal Neonatal Med 2019; 32(13):
545 2143-51.
546 15. Romero R, Yoon BH, Mazor M, et al. A comparative study of the diagnostic performance of
547 amniotic fluid glucose, white blood cell count, interleukin-6, and gram stain in the detection of microbial
548 invasion in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 1993; 169(4):
549 839-51.
550 16. Workowski KA, Bolan GA, Centers for Disease C, Prevention. Sexually transmitted diseases
551 treatment guidelines, 2015. MMWR Recommendations and reports : Morbidity and mortality weekly
552 report Recommendations and reports 2015; 64(RR-03): 1-137.
553 17. Romero R, Salafia CM, Athanassiadis AP, et al. The relationship between acute inflammatory
554 lesions of the preterm placenta and amniotic fluid microbiology. Am J Obstet Gynecol 1992; 166(5):
555 1382-8.
25
556 18. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: Building an international
557 community of software platform partners. Journal of biomedical informatics 2019; 95: 103208.
558 19. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture
559 (REDCap)--a metadata-driven methodology and workflow process for providing translational research
560 informatics support. Journal of biomedical informatics 2009; 42(2): 377-81.
561 20. Hadlock FP, Harrist RB, Martinez-Poyer J. In utero analysis of fetal growth: a sonographic weight
562 standard. Radiology 1991; 181(1): 129-33.
563 21. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and
564 intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. The Journal of
565 pediatrics 1978; 92(4): 529-34.
566 22. Bell MJ, Ternberg JL, Feigin RD, et al. Neonatal necrotizing enterocolitis. Therapeutic decisions
567 based upon clinical staging. Annals of surgery 1978; 187(1): 1-7.
568 23. International Committee for the Classification of Retinopathy of P. The International
569 Classification of Retinopathy of Prematurity revisited. Archives of ophthalmology 2005; 123(7): 991-9.
570 24. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 explanation and elaboration: updated
571 guidelines for reporting parallel group randomised trials. Bmj 2010; 340: c869.
572 25. Berghella V, Ludmir J, Simonazzi G, Owen J. Transvaginal cervical cerclage: evidence for
573 perioperative management strategies. Am J Obstet Gynecol 2013.
574 26. Souka AP, Heath V, Flint S, Sevastopoulou I, Nicolaides KH. Cervical length at 23 weeks in twins
575 in predicting spontaneous preterm delivery. Obstet Gynecol 1999; 94(3): 450-4.
576 27. Goldenberg RL, Iams JD, Miodovnik M, et al. The preterm prediction study: risk factors in twin
577 gestations. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units
578 Network. Am J Obstet Gynecol 1996; 175(4 Pt 1): 1047-53.
579 28. Kindinger LM, Poon LC, Cacciatore S, et al. The effect of gestational age and cervical length
580 measurements in the prediction of spontaneous preterm birth in twin pregnancies: an individual patient
581 level meta-analysis. BJOG 2016; 123(6): 877-84.
582 29. Berghella V, Roman A, Daskalakis C, Ness A, Baxter JK. Gestational age at cervical length
583 measurement and incidence of preterm birth. Obstet Gynecol 2007; 110(2 Pt 1): 311-7.
584 30. Boelig RC, Dugoff L, Roman A, Berghella V, Ludmir J. Predicting asymptomatic cervical dilation in
585 pregnant patients with short mid-trimester cervical length: A secondary analysis of a randomized
586 controlled trial. Acta obstetricia et gynecologica Scandinavica 2019; 98(6): 761-8.
587 31. Melamed N, Pittini A, Hiersch L, et al. Serial cervical length determination in twin pregnancies
588 reveals 4 distinct patterns with prognostic significance for preterm birth. Am J Obstet Gynecol 2016;
589 215(4): 476 e1- e11.
590 32. American College of O, Gynecologists, Society for Maternal-Fetal M. ACOG Practice Bulletin No.
591 144: Multifetal gestations: twin, triplet, and higher-order multifetal pregnancies. Obstet Gynecol 2014;
592 123(5): 1118-32.
593 33. Airoldi J, Pereira L, Cotter A, et al. Amniocentesis prior to physical exam-indicated cerclage in
594 women with midtrimester cervical dilation: results from the expectant management compared to
595 Physical Exam-indicated Cerclage international cohort study. Am J Perinatol 2009; 26(1): 63-8.
596 34. Mays JK, Figueroa R, Shah J, Khakoo H, Kaminsky S, Tejani N. Amniocentesis for selection before
597 rescue cerclage. Obstet Gynecol 2000; 95(5): 652-5.
598 35. Lisonkova S, Sabr Y, Joseph KS. Diagnosis of subclinical amniotic fluid infection prior to rescue
599 cerclage using gram stain and glucose tests: an individual patient meta-analysis. Journal of obstetrics and
600 gynaecology Canada : JOGC = Journal d'obstetrique et gynecologie du Canada : JOGC 2014; 36(2): 116-
601 22.
602 36. Klein LL, Gibbs RS. Use of microbial cultures and antibiotics in the prevention of infection-
603 associated preterm birth. Am J Obstet Gynecol 2004; 190(6): 1493-502.
26
604 37. Romero R, Gomez-Lopez N, Winters AD, et al. Evidence that intra-amniotic infections are often
605 the result of an ascending invasion - a molecular microbiological study. J Perinat Med 2019; 47(9): 915-
606 31.
607 38. Miller ES, Rajan PV, Grobman WA. Outcomes after physical examination-indicated cerclage in
608 twin gestations. Am J Obstet Gynecol 2014; 211(1): 46 e1-5.
609 39. Miller ES, Grobman WA, Fonseca L, Robinson BK. Indomethacin and antibiotics in examination-
610 indicated cerclage: a randomized controlled trial. Obstet Gynecol 2014; 123(6): 1311-6.
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
27
632 TABLES
644
645
646
647
648
29
649 Table 2. Antepartum and delivery outcomes of twin pregnancies with dilated cervix
Physical
exam No
VARIABLE Indicated Cerclage RR (Cl, 95%) or NNT
p-value
Cerclage MD (Cl, 95%)
n=17 n=13
SPTB < 34 weeks 12 (70.6) 13 (100) 0.71 (0.52-0.96) 0.05 3.4
Antepartum admission
9 (52.9) 1 (7.7) 6.88 (1.43-40) 0.02
ǂ
PPROM <34 weeks 11 (64.7) 5 (38.5) 1.68 (0.83-3.86) 0.26
Antenatal steroids 13 (76.5) 4 (30.8) 2.51 (1.12-5.84) 0.02
Tocolysis 6 (35.3) 3 (23.1) 1.50 (0.47-4.94) 0.7
GA at PPROM (weeks) 28.4±5.6 19.3±0.5 9.0 (3.4-14.6) † 0.02
GA at delivery (weeks) 29.1±1.7 22.5±3.9 6.9 (2.9-10.9) † <0.01
Mode of delivery
• Vaginal delivery
4 (23.5) 10 (76.9) 0.30 (0.12-0.76) 0.008
both
13 (76.5) 3 (23.1) 3.30 (1.19-9.3)
• CS both
Clinical
2 (11.8) 3 (23.1) 0.51 (0.1-2.6) 0.62
chorioamnionitis
Histological
4 (23.5) 8 (61.5) 0.38 (0.14-1.0) 0.06
chorioamnionitis
Abruption 3 (17.6) 1 (7.7) 2.29 (0.26-19.6) 0.6
Birth weight Twin A 1331±850 529±479 801 (1306 to 297) † 0.003
Birth weight Twin B 1349±950 502±432 847 (1384 to 309) † 0.003
Birth weight (g)* 1349±874 544±434 823 (446 to 1200) † <0.0001
24/26 3.3
Birth weight <1500g* 21/34 (61.7) 0.67 (0.50-0.89) 0.007
(92.3)
Apgar <7 at 5 min Twin 4/17 11/13 0.27 (0.11-0.68) 0.003
30
A
Apgar < 7 at 5 min
5/17 11/13 0.34 (0.16-0.75) 0.004
Twin B
22/26 1.7
Apgar < 7 at 5 min* 9/34 (26.5) 0.26 (0.16-0.50) <0.0001
(84.6)
10/13
Both twins died 2/17 (11.7) 0.15 (0.04-0.58) 0.005
(76.9)
One twin died 2/17 (11.7) 0/13 (0) NA 0.49
Both twins alive 13/17 (76.4) 3/17 (17.6) 3.31 (1.18-9.25) 0.009
At least 1 newborn
15/17 (88.2) 3/13 (0.23) 3.8 (1.4-10.4) 0.0005
home
Perinatal mortality Twin
3/17 (17.6) 10/13 (76.9 0.22 (0.1-0.67) 0.002
A
Perinatal mortality Twin 3/17 (17.6) 10/13
B (76.9) 0.22 (0.1-0.67) 0.002
650
651 Variables described as mean ± standard deviation or frequencies (percentage).
652 GA: Gestational age
653 PTB: Preterm birth
654 PPROM: Preterm premature rupture of membranes
655 CS: Cesarean section
656 MD: mean difference
657 RR: relative risk
658 CI: Confidence interval
659 NNT: number needed to treat
660 * Data including both twins A and B
661 † Data correspond to mean difference
662 ǂ Antenatal admission refers to admission for expectant management but not admission
663 for imminent delivery
664
665
666
667
Physical
exam No
VARIABLE Indicated RR (Cl, 95%) or
Cerclage Cerclage p-value
MD (Cl, 95%)
n=30 n=6
669
670 Variables described as mean ± standard deviation or frequencies (percentage).
671 RR: relative ratio
672 MD: Mean difference
673 CI: Confidence interval
674 GA: Gestational age
675 LOS: Length of stay in the hospital
676 NICU: Neonatal intensive care unit
677 RDS: respiratory distress syndrome
678 CPAP: continuous positive airway pressure
679 IVH: Intraventricular hemorrhage
680 NEC: Necrotizing enterocolitis
681 ROP: Retinopathy of prematurity
682 NA: not applicable
683 † Data correspond to mean difference
684 Table 4. Neonatal outcomes treating twin pair as cluster unit
685
32
686
ICC (95% CI)
VARIABLE
OR (CI 95%)* p-value *
687
688 Variables described as mean ± standard deviation or frequencies (percentage).
689 Boldface data, statistically significant
690 OR: odds ratio
691 CI: Confidence interval
692 NICU: Neonatal intensive care unit
693 ICC: intracluster correlation coefficient
694 *Generalized mixed-model analysis, treating twin pair as cluster unit
695 **Outcomes of only neonates born alive
696
697
698
699
700
701
702
703
704
705
706
707
711
712 Figure 2. Survival curves of twin pregnancies that remained undelivered across
713 gestation
714
715 Kaplan-Meier survival curves indicating the proportions of women with twin pregnancy
716 in the physical exam indicated cerclage and control groups. The log-rank test showed
718
719