AMTSL trial

ACTIVE MANAGEMENT OF THE THIRD STAGE OF LABOUR WITHOUT CONTROLLED CORD TRACTION: A RANDOMIZED NON-INFERIORITY CONTROLLED TRIAL HRP Trial: A65554
Justus Hofmeyr, Presented on behalf of the AMTSL trial Group

Rationale

AMTSL reduces PPH by over 60% AMTSL components are: administration of oxytocin, delayed cord clamping, controlled cord traction (CCT), and uterine massage. Previous WHO guidelines recommended AMTSL, but Acknowledging the lack of evidence on the effectiveness of some individual components. CCT requires manual skills, carries risks (uterine inversion) If CCT does not have a meaningful impact on blood loss, then it could be omitted and a simplified package focusing mainly on the uterotonic could be recommended.

Uterine inversion being reduced with saline (OSullivans method)

Uterine replacement complete

Systematic review CCT vs no CCT Outcome: Blood loss>1000ml

WHO AMTSL Trial: Objective

Primary objective: to determine whether the simplified package of oxytocin 10 IU IM/IV, without CCT, is not less effective than the full AMTSL package with regard to reducing blood loss 1000 ml in the third stage of labour.

Hypothesis: non-inferiority within a risk ratio margin of 1.30.

Study design

Hospital-based, multicentre, randomized, non-inferiority controlled trial. Participating countries: Argentina, Egypt, India, Kenya, the Philippines, South Africa, Thailand, Uganda. (16 hospitals, 2 PHC centres) Sample size ~ 25,000 women Participants: - inclusion criteria: women expecting to deliver vaginally - exclusion criteria: advanced first stage of labour, women too distressed to give consent, minors without guardian, planned caesarean section, multiple pregnancies, birth considered abortion. Interventions: - experimental arm: "simplified package" (Placental delivery WITHOUT controlled cord traction. i.e. maternal effort, gravidity) - control arm: "full package" (Placental delivery WITH controlled cord traction )

Outcome measures

Primary outcome Severe PPH (blood loss 1000 ml or more) at one hour or up to 2 hours for women who continue to bleed beyond one hour
Secondary outcomes Blood loss 500 ml or more Blood transfusion Additional uterotonics Maternal death Manual removal of the placenta Additional surgical procedures Maternal death or severe morbidity Initiation of breastfeeding

Characteristics of women at trial entry and delivery (mITT population)

Simplified package n/N* (%) Age (mean, SD) Primigravida Gestational age 37 weeks Labour induced/augmented Spontaneous cephalic vaginal delivery Perineal trauma (episiotomy or tear requiring suture) Birth weight (mean, SD) Baby alive 253 (56) 5362/11823 (454) 10713/11815 (907) 5923/11815 (501) 11266/11818 (953) 7667/11816 (649) 3069 (538) 11661/11818 (987) Full AMTSL package n/N (%) 253 (56) 5476/11799 (464) 10620/11797 (903) 5796/11795 (491) 11240/11795 (953) 7616/11791 (646) 3054 (544) 11639/11797 (987)

Adherence to trial interventions and final delivery mode of the placenta

Simplified package n/N Adherence Cord management according to the protocol* Oxytocin administered Delayed cord clamping and cutting % Full AMTSL package n/N %

11253/11861 949 11779/11861 993 9321/11861 786

11406/11820 965 11751/11820 994 9242/11820 782

Final delivery mode of the placenta** Hands off/Maternal effort CCT Umbilical vein injection Manual removal of the placenta Curettage Hysterectomy Left in situ Placenta and baby delivered simultaneously 10668/11776 906 731/11777 10/11814 153/11814 165/11814 2/11814 3/11813 29/11814 621 008 13 14 002 003 025 833/11763 708

11389/11766 968 10/11794 105/11794 98/11794 6/11794 7/11792 23/11797 008 089 083 005 006 019

*Hands-off in the simplified package and CCT in the full package in the first 30 minutes
** More than one intervention may have been used for any one patient

Trial outcomes
Simplified package Outcome n/N Blood loss 1000ml Blood loss 500ml Blood loss (ml)(mean difference) Additional uterotonics Blood transfusion Manual removal of placenta Third stage duration (min.)(mean difference) Maternal death Additional surgical procedures Maternal death or severe morbidity Baby put to breast within 30 minutes 239/11621 % n/N % 188 206 219/11621 Full AMTSL package % Risk difference (95% CI) 017 (-019 to 053) 090 (003 to 178) 108 (47 to 169) 034 (-069 to 137) 006 (-012 to 024) 040 (014 to 067) 65 (62 to 68) 001 008 027 001 (-002 to 004) -006 (-011 to 000) -09 (-022 to 003) 014 (-062 to 091) 200 (018 to 220) 022 (005 to 103) 065 (037 to 113) 100 (099 to 101) 102 (097 to 107) 112 (078 to 162) 145 (114 to 186) Risk ratio (95% CI) 109 (091 to 131) 107 (100 to 114)

1598/11621 1375 1493/11621 1285 282 (11621) 271 (11621)

2434/11802 2062 2390/11783 2028 62/11814 153/11814 126 (11662) 2/11818 2/11814 20/11616 002 002 017 052 55/11790 047 089

130 105/11794 61 (11648) 1/11798 9/11790 31/11616

10565/11711 9021 10532/11693 9007

Trial outcomes
Simplified package Outcome n/N Blood loss 1000ml Blood loss 500ml Blood loss (ml)(mean difference) Additional uterotonics Blood transfusion Manual removal of placenta Third stage duration (min.)(mean difference) Maternal death Additional surgical procedures Maternal death or severe morbidity Baby put to breast within 30 minutes 239/11621 % n/N % 188 206 219/11621 Full AMTSL package % Risk difference (95% CI) 017 (-019 to 053) 090 (003 to 178) 108 (47 to 169) 034 (-069 to 137) 006 (-012 to 024) 040 (014 to 067) 65 (62 to 68) 001 008 027 001 (-002 to 004) -006 (-011 to 000) -09 (-022 to 003) 014 (-062 to 091) 200 (018 to 220) 022 (005 to 103) 065 (037 to 113) 100 (099 to 101) 102 (097 to 107) 112 (078 to 162) 145 (114 to 186) Risk ratio (95% CI) 109 (091 to 131) 107 (100 to 114)

1598/11621 1375 1493/11621 1285 282 (11621) 271 (11621)

2434/11802 2062 2390/11783 2028 62/11814 153/11814 126 (11662) 2/11818 2/11814 20/11616 002 002 017 052 55/11790 047 089

130 105/11794 61 (11648) 1/11798 9/11790 31/11616

10565/11711 9021 10532/11693 9007

The risk of severe postpartum haemorrhage is not increased

Trial outcomes
Simplified package Outcome n/N Blood loss 1000ml Blood loss 500ml Blood loss (ml)(mean difference) Additional uterotonics Blood transfusion Manual removal of placenta Third stage duration (min.)(mean difference) Maternal death Additional surgical procedures Maternal death or severe morbidity Baby put to breast within 30 minutes 239/11621 % n/N % 188 206 219/11621 Full AMTSL package % Risk difference (95% CI) 017 (-019 to 053) 090 (003 to 178) 108 (47 to 169) 034 (-069 to 137) 006 (-012 to 024) 040 (014 to 067) 65 (62 to 68) 001 008 027 001 (-002 to 004) -006 (-011 to 000) -09 (-022 to 003) 014 (-062 to 091) 200 (018 to 220) 022 (005 to 103) 065 (037 to 113) 100 (099 to 101) 102 (097 to 107) 112 (078 to 162) 145 (114 to 186) Risk ratio (95% CI) 109 (091 to 131) 107 (100 to 114)

1598/11621 1375 1493/11621 1285 282 (11621) 271 (11621)

2434/11802 2062 2390/11783 2028 62/11814 153/11814 126 (11662) 2/11818 2/11814 20/11616 002 002 017 052 55/11790 047 089

130 105/11794 61 (11648) 1/11798 9/11790 31/11616

10565/11711 9021 10532/11693 9007

The blood loss reduction with CCT is clinically not significant

Trial outcomes
Simplified package Outcome n/N Blood loss 1000ml Blood loss 500ml Blood loss (ml)(mean difference) Additional uterotonics Blood transfusion Manual removal of placenta Third stage duration (min.)(mean difference) Maternal death Additional surgical procedures Maternal death or severe morbidity Baby put to breast within 30 minutes 239/11621 1598/11621 282 (11621) 2434/11802 62/11814 153/11814 126 (11662) 2/11818 2/11814 20/11616 002 002 017 % 206 n/N 219/11621 % 188 Full AMTSL package % Risk difference (95% CI) 017 (-019 to 053) 090 (003 to 178) 108 (47 to 169) 034 (-069 to 137) 006 (-012 to 024) 040 (014 to 067) 65 (62 to 68) 001 008 027 001 (-002 to 004) -006 (-011 to 000) -09 (-022 to 003) 014 (-062 to 091) 200 (018 to 220) 022 (005 to 103) 065 (037 to 113) 100 (099 to 101) 102 (097 to 107) 112 (078 to 162) 145 (114 to 186) Risk ratio (95% CI) 109 (091 to 131) 107 (100 to 114)

1375 1493/11621 1285 271 (11621) 2062 2390/11783 2028 052 130 55/11790 105/11794 61 (11648) 1/11798 9/11790 31/11616 047 089

10565/11711 9021 10532/11693 9007

But, there was an increase in the need of manual removal of placenta

This effect is clustered in Philippines, where ergotamine is widely used for Prevention of PPH: excluding Philippines data, there is no additional risk of manual removal of placenta

Sensitivity analysis for the trial outcomes excluding Philippines

Simplified package Outcome n/N Blood loss 1000ml Blood loss 500ml Blood loss (ml) Additional uterotonics Blood transfusion 153/9411 987/9411 163 105 % Full AMTSL package n/N 140/9420 927/9420 % 149 984 014 (-021 to 049) 065 (-022 to 151) 109 (087 to 137) 107 (098 to 116)

% Risk difference (95% CI)

Risk ratio (95% CI)

266 (9403)
1026/9472 50/9484 62/9483 114 (9391) 1/9487 0/9484 12/9407 001 0 013 108 053 065

256 (9409)
962/9459 34/9463 64/9470 62 (9378) 1/9471 3/9463 15/9415 001 003 016 102 036 068

102 (39 to 164)

066 (-021 to 154) 017 (-002 to 036) -002 (-025 to 021) 52 (49 to 55) -000 (-003 to 003) -003 (-007 to 000) -003 (-014 to 008) 100 (006 to 160) -080 (037 to 171) 107 (098 to 116) 147 (095 to 227) 097 (068 to 137)

Manual removal of placenta

Third stage duration (min.) Maternal death Additional surgical procedures Maternal death or severe morbidity Baby put to breast within 30 minutes

8571/9468

905

8566/9461

905

-001 (-085 to 082)

100 (099 to 101)

Non-inferiority analysis graph showing where the 95%CI for the Risk Ratio for the primary outcome Blood loss1000ml lies with respect to the point of no difference and to the preset non-inferiority margin for the RR, =13

Simplified package better

Simplified package worse

1.09 (0.91, 1.31)

Modified ITT analysis

0.70

0.80

0.90

1.00

1.10

1.20

1.30
=1.3

1.40

RR

Systematic review CCT vs no CCT

Main findings

CCT has minimal added value in terms of reducing blood loss over and above the uterotonic Oxytocin 10 IU IM injection after delivery of the baby should be regarded as the primary intervention for prevention of PPH In settings where SBA are not available and oxytocin is used as routine uterotonic for prevention of PPH, CCT could be safely omitted during the third stage of labour In setting where SBA are available, and oxytocin is used as routine uterotonic for prevention of PPH, practising CCT may shorten the duration of the third stage of labour without additional harms or benefits Even when there is good adherence to a hands-off management of the umbilical cord and the placenta, about 6% of women will eventually require CCT. Thus, teaching of CCT for health professionals should continue

Implications for research

The findings of this trial strengthen the need to focus on strategies to scale up the use of oxytocin in peripheral levels of health system as the primary component of AMTSL.

It is timely to conduct a rigorous evaluation of the preventive and therapeutic use of uterine massage.

Non-inferiority analysis graph showing where the 95%CI for the Risk Difference for the primary outcome Blood loss1000ml lies with respect to point of no difference and to the preset non-inferiority margin for the Risk Difference, =045

Simplified package better

Simplified package worse

ITT

0.17 (-0.19, 0.53)

Modified ITT analysis

-0.50 -0.40 -0.30 -0.20 -0.10 0.00

0.10

0.20

0.30

0.40

0.50

0.60

Risk Difference

=0.45

Statistical approach and analysis

The objective of the trial was to assess non-inferiority of the simplified
package compared to the full package in terms of efficacy within a pre-stated non-inferiority margin() of 1.3 in RR To calculate the it was assumed: 1.5% risk of sPPH with full AMTSL package and 3.0% with simplified package 70% of the benefit of the full AMTSL package 80% power to show non-inferiority of the simplified package 0.45% of the full package sPPH rate Two-sided CI of 95% Alpha of 2.5% Estimated sample size with these assumptions was 22,908 For the secondary outcomes' interpretation, superiority hypothesis was used

Trial flow chart

Statistical approach and analysis: Intention to treat or per protocol?

In superiority trials, ITT analysis is best because non-compliance may dilute and
therefore underestimate the effect difference, but will not over-estimate giving a false positive result. Thus we can be confident that any benefit shown

is a true benefit. Per protocol analysis may be biased in any direction by

selective withdrawal , and we can have no certainty about the results. In a non-inferiority (equivalence) trial, ITT analysis is dangerous because by

underestimating the difference we may falsely assume equivalence.

Therefore many statisticians recommend using per protocol analysis, assuming that it will be less likely to underestimated the difference between groups.

However, if withdrawal is selective, per protocol analysis may in fact bias results
in either direction, making the results unreliable

Statistical approach and analysis: Intention to treat or per protocol?

For example, in this trial, women in the no-CCT group who bled excessively,
tended to have CCT. Exclusion of them for non-compliance would have biased the results towards fewer PPHs in the no-CCT group, falsely confirming

non-inferiority.
For this reason we used Intention to treat analysis as our primary analysis

Analysis populations

Two analysis populations were identified:

1. Intention-to-treat (ITT) population: all enrolled women considered within groups as randomized, regardless of adherence. 2. Modified intention-to-treat (mITT) population: as ITT population but excluding women having CS.

The trial's primary analysis and interpretation was based on mITT population.

Adherence to trial interventions and final delivery mode of the placenta

Simplified package n/N Adherence Cord management according to the protocol* Oxytocin administered Delayed cord clamping and cutting % Full AMTSL package n/N %

11253/11861 949 11779/11861 993 9321/11861 786

11406/11820 965 11751/11820 994 9242/11820 782

Final delivery mode of the placenta** Hands off/Maternal effort CCT Umbilical vein injection Manual removal of the placenta Curettage Hysterectomy Left in situ Placenta and baby delivered simultaneously 10668/11776 906 731/11777 10/11814 153/11814 165/11814 2/11814 3/11813 29/11814 621 008 13 14 002 003 025 833/11763 708

11389/11766 968 10/11794 105/11794 98/11794 6/11794 7/11792 23/11797 008 089 083 005 006 019

*Hands-off in the simplified package and CCT in the full package in the first 30 minutes
** More than one intervention may have been used for any one patient