The CHIPS Trial

The CHIPS Trial: The Control of Hypertension In Pregnancy Study

Formal Publication Title: Less-Tight versus Tight Control of Hypertension in Pregnancy

https://www.nejm.org/doi/full/10.1056/nejmoa1404595

Some general background information

• Who did the study and who published it?

  • An open, multicenter, international, randomized controlled trial. 

    • Coordinating center: University of British Columbia (go Canada again!)

  • Where was it published? The New England Journal of Medicine in 2015 

• Why was the study done? 

  • Hypertension is common – at the time of this publication, it was estimated to affect 10% of pregnancies, with 1% being cHTN, 5-6% being gHTN, and 2-3% being preeclampsia.

  • Treatment of blood pressure at specific thresholds had not really been well defined.

    • On one hand – using antihypertensives liberally and early might help prevent maternal / fetal complications related to uncontrolled HTN.

    • On the other hand – antihypertensives might have their own consequences, as shown in other smaller studies (i.e., FGR).

• What was the research question?

  • To compare tight versus less-tight control of non-proteinuric, non-severe hypertension in pregnancy

Methods

• Who participated and when?

  • Subjects were recruited from March 2009 to August 2012 - 95 sites in 16 countries enrolled at least one patient.

  • Eligibility: 

    • Had non-severe, non-proteinuric preexisting hypertension or gestational hypertension

      • That’s right – they treated gHTN too! More on that later

      • Preexisting HTN defined as diagnosis pre-20 wks, gestational HTN defined as diagnosis after 20wks

    • A DBP of 90-105 if not receiving therapy, or 85-105 if already on treatment

      • BP were obtained at least 4 hours apart or at two consecutive outpatient visits, with the second measurement taken within 1 week prior to randomization.

      • Both BPs needed to be elevated to be included.

    • Live singleton fetus between 14w0d and 33w6d

  • Exclusion criteria

    • SBP of 160 or higher (but could be included later if they were treated and met all other eligibility criteria)

    • Proteinuria > 0.3mg/day on 24h, or a P:C >0.263, or a dipstic of 2+ or more

    • Used an ACE-I at or after 14 weeks

    • Had a contraindication to either trial group because of preexisting disease

      • Examples provided included pregestational diabetes or renal disease 

    • Multiple gestations, anomalies, or plans for TOP

    • Previous participation in the trial 

• How was the study done?

  • Randomized in blocks of 2 or 4 patients using a telephone line and pager system

  • 1:1 ratio of less-tight control (defined as target DBP 100 or lower) versus tight control (target DBP 85 or lower)

    • Control of BP was expected to the target level until delivery, with a goal of between-group difference of DBP of 5mmHg (goal based on a pilot trial of the protocol).

  • Recommendation for labetalol as drug of first choice.

    • ACE-I, ARBs, renin inhibitors, and atenolol were not permitted prior to delivery.

    • No drugs were provided by the study – this was left to physician discretion. 

  • BP at subsequent prenatal visits were obtained 3x per visit. The average of the 2nd and 3rd DBPs obtained were considered to be the DBP for the visit and used for med targeting.

    • Participants also kept a diary to record this info as well as medications and co-interventions (i.e., ultrasound, clinic visit info)

  • Adherence to protocol based on a “clinically reasonable standard” was assessed within 4 weeks of randomization.

    • This isn’t totally elaborated on, but did follow to some degree blood pressure measurements in the patient’s diary and the interventions listed.

    • Thereafter, patients were seen on a schedule dictated by their doctor/midwife. 

  • A standardized questionnaire was then given to patients at 6 weeks postpartum to identify post discharge complications.

• What outcomes were they looking for?

  • Primary outcome

    • Composite of pregnancy loss (miscarriage, ectopic, termination, stillbirth, or neonatal death) or high-level neonatal care (“greater than normal” newborn care) for more than48 hours until 28 days of life or discharge home, whichever was later.

  •  Secondary outcome

    • Maternal outcomes and complications up to six weeks postpartum, including:

      • Stroke, death, eclampsia, blindness, uncontrolled HTN, use of inotropic agents, pulmonary edema, respiratory failure, myocardial ischemia/infarction, hepatic dysfunction, hepatic hematoma or rupture, renal failure, and transfusion. 

  • Outcomes were adjudicated by a committee who were not aware of group assignments and not involved in patient’s care.

  • Additional outcomes analyzed included fetal growth and newborn complications, and incidence of severe hypertension (> 160/110) in the mother

• Some statistics interestingness:

  • This trial had some interesting analyses that we don’t frequently see in RCTs:

    • There were multiple levels of comparisons planned, and for this reason, the alpha level for significance (i.e., p value to look for) was 0.046.

    • Similarly, for secondary outcome, p<0.01 was needed, and for the additional subsequent outcomes, p<0.001 was needed.

      • We would love to have you super statistics-minded brains email us about why these adjustments are made – it has to do with the number of comparisons made and the two interim analyses that were performed to assure safety during the trial.

Results

• Who did they recruit? 

  • 1030 eligible women were recruited - 519 for less-tight, and 511 for tight control

    • Ultimately, one site needed to be excluded due to concerns about data integrity – so 497 patients were assigned to less-tight, and 490 to tight control.

    • Six patients were lost to follow up or withdrew so no data was available

    • 24 patients discontinued BP treatment prior to delivery, but their data was included as part of an intention-to-treat analysis

    • 10 patients (five in each group) had incomplete data after they were lost to follow up for the postpartum survey. 

    • 21 patients were found to have been ineligible after data analysis.

  • “Clinically reasonable adherence” to assigned treatment protocol was slightly worse in the less-tight group (76.6%) versus the tight group (82%).

  • Baseline characteristics were overall very similar:

    • Similar BMI, nulliparity, gestational age at randomization, gestational DM rate, smoking rate.

    • 25% in each group had gestational hypertension, whereas 75% had chronic HTN

      • 16% in the less-tight and 12% in the tight group had a severe-range BP at some point prior to enrollment (only statistical difference at p=0.049)

      • About 57% in each group were on antihypertensive meds at enrollment

  • Blood pressure was higher in the less-tight control group by average of 5.8 mmHg systolic, and 4.6 mmHg diastolic.

    • SBP: 138.8 vs 133.1 mmHg, p<0.001

    • DBP: 89.9 vs 85.3 mmHg, p<0.001

  • Antihypertensive meds were taken by fewer patients in the less-tight control group after randomization (73.4% vs 92.6%) and this continued after delivery (65.5% vs 78.3%). 

  • Labetalol was most commonly used agent (68.9% vs 68.8% between groups)

    • Four protocol violations for use of atenolol prior to delivery.

• Outcomes

  • Primary: neonatal composite – no difference. 

    • No significant differences with respect to other perinatal outcomes for newborns, including SGA <10% or <3%, or rates of respiratory complications.

• Secondary outcomes

• Maternal outcomes – no difference overall but rare serious events.

  • No maternal deaths.

  • In less severe events:

    • Frequency of severe hypertension was higher in the less-tight control group than tight control group (40.6% vs 27.5%, p<0.01)

    • Higher rates of abnormal labs consistent with severe preeclampsia in less-tight group (more frequent rates of thrombocytopenia, liver enzyme elevations) – however, these did not meet prespecified limit for statistical significance (0.001 for these other outcomes)

Conclusions and What We Do Now

• The authors conclude from this study that: 

  • Infant: “tight versus less tight control of maternal hypertension resulted in no significant difference in risk of adverse perinatal outcomes”

  • Maternal: “Less-tight control did not significantly increase risk of overall serious maternal complications.” 

    • While there was a more significant rate of severe hypertension and markers of severe preeclampsia, they didn’t meet the study’s threshold for significance (admittedly very challenging at p<0.001).

• CHIPS is interesting in that it has dictated how we treat hypertension and allowed for “less-tight control” as the dominant paradigm in US practice:

  • In most places, treatment of hypertension prior to more significant values consistent with severe BP is not performed.

  • Gestational hypertension is not typically treated unless severe-range pressures result

    • And nowadays, that’s classified as severe preeclampsia!

• It is interesting to think about this and the challenges with preeclampsia management – maybe we would prevent some severe preeclampsia with more aggressive treatment?

  • Those numbers are so small though, it’s hard to know.

  • But severe BP control in preeclampsia we know is very important to prevent stroke, seizures, and other complications…

• CHIPS did provide some reassuring data in that tighter and less-tight control paradigms didn’t seem to adversely affect birth weight.

• Given some of the limitations of CHIPS and some of these open questions, the CHAP trial was performed to better evaluate the strategy for treatment of specifically mild chronic hypertension in pregnancy. 

  • We’ll review this in a future podcast – but as a preview, it seems to favor more tight control! So perhaps a new strategy is already being employed at your institution or is incoming!

The ARRIVE Trial

Here’s the RoshReview Question of the Week!

A 32-year-old G1P0 woman at 39 weeks gestation is admitted for induction of labor. Her induction is started with vaginal misoprostol. Cervical ripening by this method is caused by the action of which of the following substances?

• Phospholipase A2

• Prostaglandin E1

• Prostaglandin E2

• Prostaglandin F2 alpha

Check out the links above to see if you got this week’s question correct!

---

Actual Title: Labor Induction vs. Expectant Management in Low-Risk Nulliparous Women 

ARRIVE = A Randomized tRial of Induction Verses Expectant Management

Background: 

• Who did the study? 

  • As many big studies in Ob/Gyn, it was done by the MFMU and the Eunice Kennedy Shriver National Institute of Child Health and Human Development 

  • First author was Dr. William Grobman from Northwestern University (now at Ohio State!)

  • The study was done in 41 hospitals participating in the MFMU 

• Where was the study published? 

  • Published in NEJM in August 2018 

• Why was the study done? 

  • In previous observational trials, there was worse perinatal outcomes when delivery before 39w0d was done without medical indication than when there was delivery at full term 

  • However, there was also thought that delivery after 41 weeks can lead to increased perinatal risks (ie. increasing risk of stillbirth) 

  • There was also thought that induction of labor should be avoided if there was no reason to induce  (ie. elective induction) because induction led to increased cesarean section and possible adverse maternal outcomes 

  • There was one previous study in the UK of 619 women 35 years and older that showed increased risk of c-section 

• What was the objective? 

  • Test if elective induction at 39 weeks would result in lower risk of composite outcome of perinatal death or severe neonatal complication than expectant management among low-risk nulliparous women 

  • So key things: 

    • Purpose of the trial was to look at NEONATAL outcomes (not mom!) 

    • Second: the population they studied was low-risk nulliparous patients, not everyone! 

Methods

• Who was in the study? 

  • As we said above, the study was done at 41 centers in the United States that were part of the MFMU (ie. most were large academic centers) 

  • Low-risk nulliparous patients were included in the study 

    • Low risk = no maternal or fetal indication to be delivered before 40w5d (ie. hypertensive disorders, fetal growth restriction etc) 

  • They had to be 34w0d-38w6d at the time of enrollment 

    • Patient had to be certain of LMP or if dating was done with ultrasound before 21w0d 

  • Had to have live, singleton fetus in vertex presentation with no contraindication to vaginal delivery and no C/S planned 

  • Patients who were consented to participate were assessed again between 38w0d and 38w6d to ensure they did not have new indication for delivery that would make them ineligible 

    • Patients who were in labor or who had PROM or bleeding were not eligible 

• How was the study done?

  • Patients who were eligible were randomized in 1:1 ratio to either:

    • Labor induction - assigned to undergo induction of labor at 39w0d-39w4d 

    • Expectant management - had to forego elective induction before 40w5d, and had to have delivery initiated no later than 42w2d 

    • Of note: no specific induction protocol for either group 

  • Randomization was stratified to site 

  • Participants were then followed and data was abstracted from their chart

  • Patients also had interview to rate their labor pain on Likert Scale and also rate their experience with Labor Agentry Scale 

• What outcomes did they look for?

  • Primary outcome: composite of perinatal death or severe neonatal complications

    • Consisted of one or more of many things (don’t have to list all): perinatal death, need for respiratory support within 72 hours after birth, Apgar of 3 or less at 5 min, HIE, seizure, infection, meconium aspiration syndrome, birth trauma, intracranial or subgaleal hemorrhage, or hypotension requiring vasopressor support 

  • Main secondary outcome: cesarean section 

    • Lots of other neonatal secondary outcomes and maternal secondary outcomes that we don’t need to list 

  • Other prespecified subgroups: race, age >/= 35 or <35, BMI, modified Bishop score at time of randomization of <5 vs. 5 or more 

Results 

• Participants 

  • Recruited from March 2014 - August 2017

  • Out of 22,533 eligible women, 6106 (27%) consented and were randomized

    • 3062 in induction and 3044 to expectant management 

    • 63% had unfavorable Bishops (<5) 

    • Both groups were similar  

  • 3 in induction group and 7 in expectant management group were lost to follow-up 

  • 6% of induction group and 4.6% of expectant management group had protocol violation 

• Outcomes 

  • Those in induction group had shorter median time from randomization to delivery than in expectant management group (7 vs. 12 days) 

  • Women in induction group underwent delivery at a significantly earlier median gestational age (39.3 wks IQR 39.1-39.6 vs. 40.0 IQR 39.3-40.7 weeks) 

  • Primary outcome 

    • Occured in 4.3% of neonates in induction group vs. 5.4% in the expectant management group (RR 0.8, 95% CI 0.64-1.00) 

    • Did not change after adjustment for previous pregnancy loss 

    • Neonates in induction group had shorter duration of respiratory support and total hospital stay 

      • Other secondary outcomes for neonates were the same 

  • Secondary outcomes for mom 

    • Cesarean delivery was 18.6% in IOL group vs. 22.2% in expectant management group (RR 0.84, 95% CI 0.76-0.93), p<0.001 

    • HTN disorders of pregnancy was 9.1% in IOL group vs 14.1% in expectant management group (RR 0.64, 95% CI 0.56-0.74), p<0.001 

    • Of note, interestingly there were also higher scores on LAS both immediately after and 4-8 weeks after delivery in the IOL group 

    • Median labor pain was also reported as less (8 vs. 9) in IOL group vs. expectant management group 

      • Of note though, for LAS and median labor pain score, the scores were statistically significant but overall difference was small 

    • Women in the IOL group spent more time on the labor and delivery unity but their postpartum stay was shorter 

    • Subgroup analyses showed no significant difference between group differences 

Discussion 

• What happened after the study? 

  • Don’t know about your hospitals, but we have begun offering 39 week inductions to all nulliparous patients if they desire them 

    • ACOG also made a statement that reasonable to offer 39 week induction as long as we also take patient preference into consideration 

  • It is not “recommended” but offered 

  • Why? 

    • Some people take this to mean that by having IOL at 39 weeks that we are not only not changing neonatal outcomes, we are also decreasing CS rates and HTN disorders of pregnancy per this study 

    • They also would argue that there is no difference in LAS or pain overall (and if anything, patients feel more agentry and less pain) with IOL

  • What’s the other side of the story? 

    • In 2018, the ACNM responded to the ARRIVE Trial study results 

    • Discussed that potentially by increasing IOLs, we are also increasing the use of hospital resources (ie. staff, capacity of hospital beds, etc) 

    • Also stated that the study criteria were very strict (low risk, nulliparous), and discussed that we should be careful of broadening the outcomes and applying IOL to all patients 

    • Basically: concern that we will be offering IOL to everyone without knowing the actual implications 

  • More to the story 

    • There was a lot of concern about cost and hospital resources , so in 2020 Einerson et al came out with a study looking at cost 

    • Reviewed health-system cost of elective IOL at 39 weeks vs. expectant management in Utah hospitals 

      • No cost difference between expectant vs. IOL 

        • Maternal outpatient antenatal cost were 47% lower in the induction arm, and intrapartum and delivery costs were 16.9% higher 

• How do we practice now? 

  • One study did look at rates of IOL pre and post ARRIVE 

    • Gilroy et al looked at rates of IOL in the country in patients who were nulliparous who started prenatal care by 12 weeks and delivered at 39 weeks or later  

    • There was a significant increase in IOL after ARRIVE 

      • 36.1% vs 30.2%, OR 1.36

      • Also more likely to deliver by 39w6d (42.8% vs. 39.9%) 

      • Less likely to have a CS (27.3% vs. 27.9%) ← but that is a much higher rate than 18% in ARRIVE 

The Twin Birth Study

Here’s the RoshReview Question of the Week:

Which of the following needs to be met to undergo vaginal delivery with a monochorionic-diamniotic twin pregnancy with vertex twin A?

• Estimated fetal weight of twin B < 1,500 g

• Growth discordance < 20%

• Minimum 24 weeks gestation

• Vertex-vertex presentation

Check your answer and get a special RoshReview deal for listeners at the links above!

---

Actual title: A Randomized Trial of Planned Cesarean or Vaginal Delivery for Twin Pregnancy

https://www.nejm.org/doi/full/10.1056/nejmoa1214939 

Background:

• Where was the study published?

  • NEJM, October 3, 2013

• Why was the study done?

  • Tthrough the 1990s and 2000s there was a significant rise in twin births in the USA, likely attributed to advancing maternal age (when twinning is more common spontaneously) and the use of reproductive technology – ovulation induction and IVF.

  • In the wake of the Term Breech trial, as well as some observational studies looking at twins specifically, there was concern that breech birth risks could be extended to twins – and practice was changing!

    • In 1995, 53.9% of twin births were by CS. By 2008, this number was 75%. 

  • Not all observational studies were in agreement about the risk of “breech extraction” of a second twin, specifically – so a new study was planned and performed.

• Who performed the study?

  • The “Twin Birth Study Collaborative Group” – a large multinational collaborative, but with the main site at the University of Toronto and funded by the Canadian Institutes of Health Research – the same funders that brought you the Term Breech Trial!

    • You’ll note a lot of similarities (but also some important differences!) between this study and the Term Breech Trial. We definitely recommend a compare-contrast session!

• What was the research objective?

  • To compare the risk of fetal/neonatal death or serious morbidity between planned cesarean or planned vaginal delivery for twin pregnancies between 32w0d and 38w6d, if the presenting twin was in cephalic presentation. 

Methods:

• Who participated and when?

  • Recruitment between December 13, 2003 and April 4, 2011 at 106 centers in 25 countries.

  • Enrolled 1392 patients in the planned cesarean group and 1392 patients in the planned vaginal delivery group.

• Eligibility:

  • Needed to have:

    • Twin pregnancy between 32w and 38w6d

    • First twin in cephalic presentation

    • Both fetuses alive with EFW between 1500g and 4000g, confirmed by ultrasound within 7 days before randomization

  • Exclusions:

    • Monoamniotic twins

    • Lethal fetal anomalies

    • Other contraindication to labor or vaginal delivery (including 2nd twin being “substantially larger” than the first)

    • Prior cesarean with vertical incision or more than one LTCS

• Management:

  • Delivery by cesarean or by labor induction was planned between 37w5d and 38w6d

  • If in the CD group, if the first twin delivered vaginally, then a c-section was attempted for the second twin if logistically possible.

  • In the VD group:

    • Continuous EFM was “recommended” during active labor

    • Use of oxytocin and epidural analgesia were left to OB provider discretion

    • After delivery of first twin, use of US was “encouraged” to check second twin presentation

      • If cephalic, amniotomy was delayed until head was engaged and SVD anticipated, unless for other OB indication

      • If non-cephalic, OB decided on best delivery option – spontaneous or assisted breech delivery, total breech extraction +/- internal podalic version, ECV and vaginal cephalic delivery, or intrapartum CD

    • Deliveries were attended by qualified OB experienced in twin delivery, defined as a OB who judged themselves to be experienced at twin delivery and whose department head agreed with this judgment (similarly to Term Breech Trial).

• Outcomes:

  • Primary: fetal/neonatal mortality or serious neonatal morbidity, assessed up to 28 days after birth.

    • Morbidities included many of the same things in the Term Breech Trial, and were serious neonatal morbidities (for the sake of brevity, we won’t list them out).

  • Secondary: maternal death or serious maternal morbidity, assessed up to 28 days after delivery.

    • Again, this was very similar to the Term Breech Trial. 

  • A number of subgroup analyses were planned for the primary outcome, including by nulliparity; gestational age at randomization; maternal age; presentation of the second twin; chorionicity; and the perinatal mortality rate in the mother’s country of residence. 

Results

• Who was recruited?

  • Outcome data was available for 1392 women (2783 fetuses/infants) in the cesarean group and 1392 women (2782 fetuses/infants) in the vaginal delivery group. 

  • Baseline characteristics were overall similar, and most patients (82.4%) underwent randomization between 32w0d and 36w6d. 

    • More than half of the infants in each group were born at 37w0d or later. 

      • Around 5-6% in each group were between 32w and 33w6d, and another 42% between 34w0d to 36w6d. 

    • The time from randomization to delivery was similar but slightly different between groups (12.4 vs 13.3 days).

• In the planned CD group: 

  • 90% had CD

  • 1% had a combined vaginal-cesarean delivery, and 

  • 9% had both twins vaginally.

    • Almost 60% of the CDs were performed before the onset of labor.

• In the planned VD group:

  • 56% delivered both twins vaginally, 

  • 4% had a combined vaginal-cesarean delivery, and 

  • 40% had a cesarean for both twins.

    • Of those in the VD group who had a CD, 67.5% of them were performed during labor (or another way to look at it, 32.5% had a CD prior to labor in the planned VD group).

  • 95% had an experienced OB present, according to the study definition

• Primary Outcome:

  • The frequency of composite primary outcome did not differ between planned CD (60, or 2.2%) and planned VD (52, or 1.9%) groups.

    • The only variable that appeared to modify the risk of the primary outcome was earlier gestational age at randomization. 

    • The number of deaths in each group was 24 (0.9%) in CD group and 17 (0.6%) in VD group. 

      • 11 of these deaths in the CD group and 8 in the VD group were before labor onset.

  • In subgroup analyses, there was no significant interaction with the primary outcome with respect to parity, gestational age at randomization, presentation of the second twin, chorionicity, or national perinatal mortality rate. 

  • The second twin was more likely than the first to have the primary outcome, but this was not different between the groups. 

• Secondary outcome:

  • There were no differences in primary maternal composite outcome rates (7.3% CD, 8.5% VD). 

Impact

• What is the impact of all of this, and what are we doing now?

  • This paper certainly helped to encourage the training and planning of vaginal delivery of the second twin, including by breech delivery by stating that no increased risk was seen with a policy of planned vaginal delivery. 

    • In ACOG PB 231 on multifetal gestation, it notes that vaginal delivery of a non-cephalic second twin is reasonable, provided an OB with experience is present.

    • That’s key – it’s apparent in this paper that, compared with the Term Breech Trial, there was more emphasis on patient counseling / selection (i.e., 13 day median from randomization to delivery, protocolized assessment of EFW by US within 7 days, 95% presence of “experienced OB”). 

      • And this is heavily noted in the conclusions of the paper – stating “only centers that can provide OB management as specified by the protocol, including ability to perform a CD within 30 minutes if necessary” should undertake this.

• Methodologically, this group responded to many criticisms of the Term Breech Trial:

  • An improved randomization scheme that was block-based, stratified by gestational age and parity.

  • Improved use of ultrasound and CTG in labor, as well as higher standard of care at all sites to prevent misappropriation of primary outcome.

  • More explicit counseling – happening weeks before delivery on average, rather than in labor!

• And finally - and most importantly - this represents a well-selected, high-resource, best-case scenario work.

  • For our US listeners who mostly practice in centers where there is ability to perform cesarean within 30 minutes, the Twin Birth Study included:

    • Twins delivering between 32w0d and 38w6d

    • With EFW estimated by US within 7 days of delivery, ranging from 1500g - 4000g

      • Second twin not significantly larger (with expert opinion putting this around a max of 15% discordance)

    • Ability to perform CD within 30 minutes, and use CTG and intrapartum US

    • With someone with experience and ability to perform breech extraction and internal podalic version available 

The CREST Study

Here’s the RoshReview Question of the Week!

Which of the following methods of sterilization has the highest relative risk for ectopic pregnancy?

• Bipolar coagulation

• Postpartum partial salpingectomy

• Silicone band methods

• Spring clips

Check your answer and get a special deal on RoshReview at the link above!

---

Background: 

• Who did the study? 

  • Study was done by the US Collaborative Review of Sterilization Working Group 

  • CREST was part of the CDC and conducted with the NICHD 

  • Conducted with 10 year follow up and was done at multiple medical centers (Baltimore, MD, Buffalo, NY, Chapel Hill, NC, Honolulu, HI, Houston, TX, Memphis, TN, Sacramento CA, St. Louis, MO, San Francisco, CA) 

• Where was the study published? 

  • AJOG in 1996 

  • Presented at the Annual Meeting of the American Gynecological and Obstetrical Society in Napa, CA in 1995 

• Why was the study done?

  • Tubal sterilization is the most prevalent form of contraception among married women and formerly married women in the US 

  • While sterilization was common, there was not widespread data about their efficacy, especially over time 

• Objective: To assess the effectiveness of various methods of tubal occlusion 

Methods: 

• Who was included? 

  • Prospective study of women undergoing tubal sterilization at the above mentioned medical centers from 1978 -1986 

  • Ages 15-44 years 

  • Patients were approached before their sterilization procedure 

• How was it done? 

  • If the patient agreed, information about her history was obtained 

    • Characteristics of the surgical procedure, including complications during the surgery and afterward, were recorded 

    • Contacted at 1 month for brief follow-up 

    • Annual follow-up planned for 5 years for all patients 

    • If they were enrolled early enough, patients also had annual followup for 8-14 years after sterilization 

    • If the patient could not be contacted for the follow up then the last completed interview was used in the analysis 

  • At the follow up, all patients were asked: “Since your tubal sterilization, have you had a positive pregnancy test or been told by a physician that you were pregnant?” 

    • If yes, the interviewer then had a separate form with additional info about the pregnancy 

    • Excluded from further follow up if they became pregnant, had a repeat sterilization, a tubal anastomosis, or hysterectomy 

  • Type of tubal occlusion included: (don’t need to say all of these) 

    • Laparoscopic unipolar coagulation - don’t do these anymore! 

    • Laparoscopic bipolar coagulation - I have never seen this 

    • Laparoscopic silicone rubber band application - I think I did a few of these 

    • Laparoscopic spring clip application - Filshie clips? 

    • Partial salpingectomy (including Pomeroy, other types of partial, and total salpingectomy) 

  • If a pregnancy was identified, they were classified into: 

    • True failure (pregnancy conceived after sterilization) 

    • Luteal phase pregnancy (pregnancies conceived before sterilization but ID’ed after) 

    • Pregnancy resulting from tubal anastomosis or IVF 

    • Or pregnancy of unknown status (didn’t get the information) 

Results 

• Who: 

  • 10,863 women enrolled → 178 were excluded from analysis

    • Some were due to loss to follow up, refusal to be interviewed at 1 month follow up, or refusal for prolonged follow up 

    • Others excluded because of hysterectomy, repeat tubal ligation, or death  

  • Demographics 

    • Median age: 30 (so pretty young!) 

    • Most women were non-Hispanic White (52.7%) and had had at least 2 pregnancies 

    • Most common procedure: silicone band (31.2%), followed by bipolar coagulation (21.2%), postpartum partial salpingectomy (15.3%)

      • For us, that is super different! Since I think what i have done the most is postpartum or interval total salpingectomies 

      • Though for a bit, we also did Pomeroys and Parklands  

• Follow-up 

  • 89.2% were interviewed at 1 year after sterilization, 81% at 3 years, 73% at 5 years, and 57.7% 8-14 years (so some drop off, but that’s expected) 

  • At each follow up interval, younger women (age 18-27) had lower percentage of follow-up than older women 

  • Black, non Hispanic women also had lower rates of follow up compared to white non-Hispanic women 

• Sterilization failures

  • Out of 10,685 women in the analysis, only 143 were true sterilization failures = 1.3% failure rate  

    • 21 (14.7%) ended in SAB 

    • 26 (18.2%) were TABs 

    • 41 (28.7%) ended in delivery 

    • 47 (32.9%!!!) ended in ectopic pregnancies 

  • Another 34 women not included in analysis had luteal phase pregnancies  

  • 16 were from tubal anastomosis and IVF, and 5 were “unknown” classification

View fullsize

• Above table: lifetime accumulation of sterilization failure by method from 1-10 years per 1000 procedures and 95% CI (only showing years 1-4 because all the years made the table huge) 

  • We can see that for clip and interval partial salpingectomy, there seems to be a higher rate of lifetime pregnancies 

  • Lowest risk was postpartum partial salpingectomy 

• Also looked at 10-year cumulative probability of failure is affected by age at tubal sterilization 

  • Probability for failure in women <28 is greater than for women sterilized at ages >34 (makes sense … if you’re younger, you likely have more “fertile” years ahead of you) 

View fullsize

• After adjustment for age, race, and study site, interval partial salpingectomy, spring-clip application, and bipolar coagulation were more likely than postpartum partial salpingectomy to result in sterilization failure 

• After adjustment, black women were at higher risk than white women for sterilization failure 

• There were also interestingly differences between sites! 

So what did this all mean? 

• Sterilization failure rates 

  • Higher than previously thought! For all comers it was a little over 1% 

  • HIgher failure rates occurred after longer times (ie. more than 1-2 years, which was what other studies had looked at)

    • Failure rates between 5-10 years after procedure ranged from 1.2-8.3/1000 procedures depending on method 

  • Method failure rate also is affected by age, race, and also institution! (meaning how well or properly you do the procedure could affect effectiveness) 

  • Also, risk of ectopic increases with tubal ligation 

• What was the follow-up or impact of the CREST study? 

  • There was way more data collected than just this, and way more than just this study that was published from the CREST dataset 

  • Some other studies that were interesting: 

    • Risk of regret after tubal sterilization (1985) - 2% regretted after 1 year, 2.7% did so after 2 years 

      • Characteristics of those that had more regret: age <30 (regardless of parity), concurrent C/S

        • After 5 year follow up, risk of regret in those 20-24 was 4.3%, rate for those 30-34 was 2.4% 

        • Where do we get this 20% risk of regret from??? - different study from 1999 - in women <30 years of age 

          • In that same study for women >30, risk of regret was 5.9% 

          • Also, for women <30 the cumulative probability of regret decreased as time since birth of the youngest child increased

          • Risk of regret was actually lowest for women with no previous births!!

    • Unintended laparotomy associated with laparoscopic tubal sterilization: rate was: 51/5021, so about 1%

      • Increased risk: prior abdominal or pelvic surgeries  

    • Characteristics of those that sought tubal reanastomosis

      • 6.2% sought information for reanastomosis 

      • Women who were <30 were more likely to seek out this information 

      • Of those that actually had anastomosis, they were more likely to be white, have lower gravidity, and be younger, and to have experienced changes in marital status

• How does this change our practice? 

  • We are performing different procedures from the ones that were studied in the CREST procedure

    • Nevertheless, I still quote the findings from this study for patients when they want them: 

      • Risk of failure depends on method

      • Risk overall of failure is low, but can be as high as 1% overall, and even higher depending on age and type of procedure 

      • Risk of conversion to laparotomy from laparoscopy is overall low but increases with more surgeries in the belly 

      • Risk of regret is as high as 20% – I think I may now qualify this only for certain populations! 

    • We shouldn’t NOT perform sterilization procedures, however, just because of risk of regret 

      • Even if someone is nulliparous, young, and not married, if they are well counseled and still desire sterilization, we can perform it 

The Term Breech Trial

Actual title: Planned cesarean section versus planned vaginal birth for breech presentation at term: a randomized multicenter trial

Background:

• Who did the study, and who published it?

  • Another large collaborative group study – the Term Breech Trial Collaborative Group

    • Primary authors of the group based in Canada

    • Participating sites spanned Europe, Canada, Australia, Central/South America, Africa, Middle-East, India & Pakistan

    • Funded by Canadian Institutes of Health Research (like Canada’s NIH)

• Why was the study done?

  • As we covered in our ECV and breech birth podcasts, breech presentation affects about 3-4% of term pregnancies.

  • At the time this was conducted (published in 2000), it was a controversial question for delivery approach:

    • Cohort studies prior to this trial suggested that cesarean section was potentially better.

      • However, these cohort studies were potentially confounded by inclusion of pregnancies that would not be considered for breech birth (i.e., footling presentation) or by lack of experience with breech deliveries by the participating physicians. 

    • Two smaller RCTs prior to this trial and a meta-analysis did not find any substantial benefit to planned cesarean.

  • Thus this ambitious trial was carried out to determine which was truly better.

• What was the research question?

  • “To determine whether planned cesarean section was better than planned vaginal birth for selected fetuses in the breech presentation at term.” 

Methods:

• Who participated and when?

  • Enrollment between Jan 1997 and April 2000 at 121 centers in 26 countries.

  • Eligibility:

    • Breech presentation at or after 37 weeks → assigned to planned cesarean or planned vaginal delivery after consent by telephone system.

    • If assigned to CD group: planned CD was scheduled for 38+ weeks.

      • If patient presented in labor, CD was performed as soon as possible.

      • Presentation was confirmed again prior to cesarean and if cephalic, vaginal birth was then planned.

    • If assigned to VD group: management was expectant until spontaneous labor began, unless an indication to induce labor or a different reason for cesarean developed.

      • Labor management protocol was standardized (a LOT of standards):

        • Induction and amniotomy were allowed for standard OB indications

        • FHR monitoring by intermittent auscultation or cEFM

        • Augmentation with oxytocin was permitted, “so long as the clinician was confident that there was no evidence of cephalopelvic disproportion.”

        • Adequate labor progress in 1st stage was 0.5cm/h after onset of active labor, and in 2nd stage was descent of breech to pelvic floor within 2hr of full dilation, and delivery being imminent within 1hr of active pushing.

          • If fetal heart rate abnormalities were encountered or lack of progress in labor occurred, CD was performed.

        • Anesthesia choice was left to patient and providers.

        • Delivery means was recommended to have no intervention until spontaneous exit of the infant to the umbilicus, and minimum intervention thereafter with no traction on the body; and controlled delivery of the head using forceps or Mariceau-Smellie-Veit maneuver. 

          • Total breech extraction was not permitted.

      • Deliveries were to be attended by an “experienced clinician” – defined as someone who considered themselves to be skilled and experienced at vaginal breech delivery and confirmed by their department head at their institution. 

        • Each clinician was assigned a code number prior to study enrollment and there was info recorded on their qualifications and years of experience with breech delivery. 

• What were they looking for?

  • Primary outcome: perinatal or neonatal mortality at less than 28 days of age, or one or more serious neonatal morbidities such as:

    • Birth trauma, such as subdural hematoma, IVH, spinal cord injury, basal skull fracture, peripheral nerve injury, or clinically significant genital injury; 

    • Seizures at less than 24h of age or requiring 2+ drugs to control; 

    • Apgar of less than 4 at 5 mins

    • Cord base deficit of 15 or greater

    • Hypotonia for at least 2 hours

    • Stupor, decreased response to pain, or coma

    • Intubation and ventilation for at least 24h

    • Tube feeding for four days or more

    • NICU admission longer than 4 days

  • Secondary outcome: maternal mortality or serious maternal morbidity for up to six weeks postpartum, such as:

    • PPH > 1500 cc or need for transfusion

    • D&C for bleeding or retained tissue

    • Hysterectomy

    • Cervical laceration involving lower uterine segment (if SVD)

    • Vertical incision or serious hysterotomy extension (if CD)

    • Vulvar or perineal hematoma requiring evacuation

    • DVT / PE

    • Pneumonia, ARDS

    • Wound infection requiring prolonged hospital stay, wound breakdown

    • Bladder / ureter / bowel injury requiring repair, or development of fistula

    • Bowel obstruction

    • Other serious morbidities

  • They also used multiple logistic-regression analyses to test for interactions between demographic / baseline characteristics and outcomes:

    • I.e., looking at the perinatal outcomes for babies, looked at interaction with maternal age, parity, type of breech presentation, gestational age, labor vs induction, EFW by US vs Leopold’s, etc. 

  • They also looked at center’s standard of care in “usual care” vs “high standard of care” environments (as determined by a pre-study survey) and maternal/perinatal mortality rate in the center’s country

    • High standard of care was defined as:

      • Having ability to perform a cesarean within 10 mins (vs 60 mins)

      • Personnel to bag-mask a baby available immediately (vs 10 mins)

      • Personnel to intubate / provide PPV to baby within 10 mins (vs 30 mins)

      • Ability to ventilate a baby for >24h (vs need for transfer for this)

Results:

• Who was recruited?

  • 2088 pregnant patients were randomized, with 1043 assigned to CD and 1045 assigned to VD

    • Maternal outcomes were available for 1041 (CD) and 1042 (VD) groups

    • Neonatal outcomes were available for 1039 (CD) and 1039 (VD) groups

      • Overall very low loss to follow up!

  • Groups were overall very similar:

    • 52% in each group were nulliparous

    • EFW > 3kg in ~66% in each group (with 60% in each group estimated by ultrasound)

      • 21.9% in each group underwent attempted ECV

    • 35% of deliveries in each arm took place in “high standard of care” centers.

    • ~42% in each group presented in labor, and ~23% in each group had membranes ruptured on presentation.

    • Median GA at delivery was similar in each group (39w3d CD vs 39w6d VD)

  • In the planned CD group:

    • 90% had a CD, and 10% had a VD

  • In the planned VD group:

    • 43.3% had a CD, and 56.7% had a VD

      • CDs most commonly performed for:

        • Fetopelvic disproportion or abnormal progress in labor (50.1%)

        • Fetal heart rate abnormality (28.6%)

        • Footling breech presentation (15.3%)

        • Patient request (13.5%)

        • Medical or OB complication (10%)

        • Cord prolapse (2.7%)

    • ~15% of labor was induced, and ~50% of labor was augmented.

      • Only 9% had a protocol violation of labor management:

        • 3.6% for prolonged labor

        • 1.4% for footling or uncertain presentation of breech at delivery

        • 2.7% had no experienced clinician at delivery 

• Outcomes:

  • Primary

    • There was a significantly lower risk of perinatal/neonatal morbidity in the planned cesarean group:

      • 17/1039 (1.6%) CD vs 52/1039 (5.0%) VD (RR 0.33)

        • This held true in both countries with low perinatal mortality rates and high perinatal mortality rates.

        • This difference lessened (but remained statistically significant) with increasing experience of the attending provider – that is, there was less risk of morbidity when more experienced personnel were present.

        • This difference also lessened (but remained statistically significant) when excluding induced/augmented labor, deliveries without a skilled clinician, and those with footling/uncertain breech presentation 

          • Bottom line: there is higher perinatal morbidity with planned VD, despite increasingly optimal environments in subanalyses

    • There were 3 neonatal deaths in the cesarean group, and 13 in the vaginal delivery group.

      • One of the CD group deaths was actually a vaginal birth with difficult delivery.

      • Two of the vaginal birth deaths were likely prior to enrollment; two were of neonates discharged from the hospital in good condition otherwise; and two were suspected intrapartum demises before a cesarean section could be started.

    • The reduction in perinatal morbidity was much greater in countries with lower perinatal mortality rates, despite a higher likelihood of cesarean birth in the VD groups in those countries

      • I.e., in an area with low perinatal mortality and a higher likelihood to get cut intrapartum, the benefit of planned CD was even greater

    • There was more birth trauma, seizures, hypotonia, low Apgars, cord blood acidemia, need for mechanical ventilation, and longer NICU stays in the planned vaignal delivery group. 

    • Birth weight over 4000g was more likely in the VD group (5.8% vs 3.1%). 

  • Secondary

    • No major group differences in maternal morbidity/mortality. 

Impact

What was the impact of all of this, and what are we doing now?

• Well you kind of know already – how many planned breech births have you done?

  • For most of us, that answer is probably 0 “planned” and maybe 1 or 2 “unplanned.”

• This was absolutely a practice-changing paper – and has guided how we approach the patient with breech presentation at term ever since.

• What criticisms exist of this trial?

  • High number of patients were randomized in labor, and questionable counseling:

    • ~42% in each arm – this begs the question about how “planned” each method actually was, and how much counseling or selectivity may have been realistically applied to each patient.

    • Can also see this in the higher rate of >4kg babies in the VD arm, as well as only 20% in each arm had an attempt at ECV. 

  • Standard of care was not consistent

    • Continuous EFM (for better or worse) was not standard

    • Fetal weight and attitude of the head, while recorded for trial data, did not have to be assessed by US

    • Generalizability is hard to interpret with the pooled results:

      • I.e., would you take your chances on a breech in Australia with a well-counseled patient with CTG in a place where you can do a C-section in under 10 minutes and have highly qualified neonatal staff immediately available? Or in another location where you can’t do a c-section even by 60 minutes? 

        • The subgroup analyses by perinatal mortality rate of countries tried to answer that, but it is curious they didn’t choose to divide it up by their “high standard” versus “usual standard” centers.

      • Subsequent analysis has shown that 30% of the morbidity/mortality in the VD group can be attributed to the 6.7% of vaginal deliveries attended by no experienced provider – also calling into question the impact of experience. 

  • Enrollment scheme:

    • Units would call to centralized location to get allocation and this was not stratified by center – theoretically, a highly experienced obstetrician could call and end up with randomization to cesarean each time, and likewise an inexperienced one could call and randomize to vaginal delivery each time.

  • Numerous protocol violations and enrollment questions

    • Included initially in the trial were: 2 demised infants, 2 sets of twins, an anencephalic baby, a baby with spina bidfida, and another baby with suspected anomalies. 

    • 2 of the vaginal delivery group deaths were suspected to have been prior to randomization, yet counted in the perinatal mortality. 

      • And a number of other deaths in the VD arm may not have been related to the delivery itself – i.e., SIDS, GI issues, anomalies, inadequate respiratory resuscitation or resources.

What about doing breech births? Can we? / should we?

• We don’t recommend breech birth – that would be irresponsible to do outright!

• Well counseled, well-selected patients in centers with experience in breech delivery, ability to perform cesarean, and provide immediate neonatal resuscitation – it may be reasonable, though this trial doesn’t totally answer that question based on problems with generalizability.

  • In the wake of this trial, the PREMODA study was performed in France and Belgium – a prospective observational study allowing providers to select mode of delivery where breech delivery was still an option/standard.

    • ⅔ of women were allocated to cesarean delivery, but 71% of those undergoing planned vaginal breech birth were successful. 

    • There was no difference in perinatal mortality/morbidity was noted between groups in this study – suggesting that rigorous protocols and assessment by those with experience may make breech birth safer.

Interesting critique of trial: https://www.ogmagazine.org.au/14/2-14/term-breech-trial/