Reducing deaths from postpartum haemorrhage Justus Hofmeyr East. Cape Dept of Health, University of the Witwatersrand/ Fort Hare

Mdantsane, Eastern Cape (near East London, South Africa)

Labour ward, Cecilia Makiwane Hospital

Effective Care Research Unit

Reported maternal deaths in SA : n=2445 (MMR about 200/ );

Comparison of more developed and less developed provinces Maternal mortality ratio / Proportion of deaths due to obstetric haemorrhage Eastern Cape>20016% Western Cape506%

Effective Care Research Unit Research overview: Prevention of PPH 1980’s: Uterotonic effects of traditional herbal remedies 1989: First randomised trials of misoprostol in 3 rd stage of labour Systematic reviews 2003: First placebo-controlled trials of misoprostol in addition to routine care for treatment of PPH (SA and Gambia):

Cochrane reviews of randomised trials of PPH prevention: Blood loss >1000ml (%) >500ml 5 trials, 2305 women *=significant, (*)=borderline

Misoprostol for PPH treatment: published reports AuthorsNMisoprostolPPH stopped O’Brien rectal14 (100%) A-Aleem rectal16 (88%) Adekanmi uterine1 (100%) Ozan oral 2hrly2 (100%) Sojai rectal5 (100%) Oboro uterine1 (100%) Total4139 (95%)

Misoprostol for treatment of PPH: Systematic review (2003): One Trial: misoprostol rectally vs injectable oxytocics. Main outcome: failure to stop bleeding (assessed clinically, unblinded)

Pharmacokinetic Studies: Time to peak concentration (minutes)

Pharmacokinetic Studies: Peak concentration (pg/ml)

Misoprostol vs placebo (in addition to routine management): measured blood loss of 500 ml or more after enrolment

Høj L., Cardoso P., Nielsen B.B., Hvidman L., Nielsen J., Aaby P. Effect of sublingual misoprostol on severe postpartum haemorhhage in a primary health centre in Guinea-Bissau: randomized double blind controlled trial. BMJ 2005;331; One death in the misoprostol group

Results: Blood loss after delivery (Hoj 2005)

Inadequate reporting of safety issues from clinical trials in academic journals Susan C Morgan, ‘..it would appear reasonable to include adverse event to misoprostol as a possible cause for the death. It is known that misoprostol can cause hypotension which could be particularly hazardous in the context of blood loss and undiagnosed cardiovascular or cerebrovascular disease.’

Re: Inadequate reporting of safety issues from clinical trials in academic journals Lars Høj ‘.. the event was appropriately reported for inclusion in meta-analyses to come.’

Questions for authors regarding the effect of sublingual misoprostol on postpartum haemorrhage Nancy L. Sloan, Beverly Winikoff, Jennifer Blum ‘Finally, as with our study, the one maternal death occurred in the misoprostol group. Perhaps it is time to conduct a specific trial that measures the impact of administering misoprostol in the third stage of labour on maternal death as a primary endpoint.’

Review of misoprostol in the third stage of labour and maternal mortality Justus Hofmeyr, Metin Gülmezoglu Methods: We searched The Cochrane Controlled Trials Register and the electronic data base Pubmed/Medline on the word 'misoprostol' in December 2005 Randomised comparisons of misoprostol used for the prevention or treatment of postpartum haemorrhage, with placebo or alternative methods (usually injectable uterotonics).

Results: We identified 32 prevention and three treatment trials of postpartum haemorrhage where misoprostol was used as one of the interventions In 24 trials no mention of maternal deaths was made. 3 prevention and 1 treatment trial reported maternal deaths

Review: all misoprostol in 3 rd stage trials with maternal deaths

Deaths in non-misoprostol groups TrialDosageCircumstances of deaths WHO 2000Prevention 600 oral PPH (2 cases)

Deaths in misoprostol groups TrialDosageCircumstances of deaths WHO oralPPH (2 cases) Walraven 2005 Gambia Prevention 600 oral 1. PPH (2200ml) 2. DIC due to malaria (blood loss 300ml) Hoj 2005 Guinea-Bissau Prevention 600 subl. Blood loss 1417ml, apparently controlled. Died after 90 minutes ? Cause Hofmeyr 2004 South Africa Treatment 200 oral 400 subl. 400 rectal 1. Continued bleeding. Hysterectomy. Multiple transfusions. Died after 2 days 2. Bleeding form torn cervix, cardiac arrest 7.5 hours after birth 3. Previous caesarean section. Bled 500ml before and 425ml after enrolment. Hypotension and cardiorespiratory arrest 1 hour after enrolment.

Limitations of this review Numbers are small, with wide confidence limits Consistent, at the 95% confidence level, with anything between a modest reduction and a large increase in maternal mortality with misoprostol This was a post -hoc analysis in response to an observed clustering of maternal deaths, and the statistical calculations need to be interpreted with circumspection.

Discussion Blood loss is a proxy outcome – we assume that reduced blood loss will translate to reduced deaths Is it feasible for reduced blood loss to be associated with more deaths?

Example of falsely reassuring proxy outcomes Common cause of death from myocardial infarction is ventricular tachyarrhythmia Class 1 antiarrhythmics effective in reducing ventricular arrhythmias after myocardial infarction Assumed would reduce deaths and became routine for post-myocardial infarction care

Mortality following ventricular arrhythmia suppression by encainide, flecainide, and moricizine after myocardial infarction. The original design concept of the Cardiac Arrhythmia Suppression Trial (CAST). Epstein AE, et al. JAMA 1993; 270: OBJECTIVE--To test the hypothesis that in survivors of myocardial infarction, the suppression of ventricular premature depolarizations improves survival free of cardiac arrest and arrhythmic death. INTERVENTION--Administration of encainide, flecainide, moricizine, or placebo to suppress ventricular premature depolarizations.

(2) Mortality following ventricular arrhythmia suppression by encainide, flecainide, and moricizine after myocardial infarction. The original design concept of the Cardiac Arrhythmia Suppression Trial (CAST). Epstein AE, et al. JAMA 1993; 270: RESULTS--At 1 year from the time of randomization to blinded therapy, mortality of placebo-treated patients 5% vs 10% of active drug-treated patients (P =.0006). CONCLUSIONS--The suppression of asymptomatic or mildly symptomatic ventricular arrhythmias after myocardial infarction does not improve survival and can increase mortality. Treatment strategies designed solely to suppress these arrhythmias should no longer be followed. [Estimated deaths from use of anyi-arrythmics: 10,000]

Hypothesis Apart from its uterotonic effects, misoprostol has known pharmacologic effects on multiple systems, such as thermoregulation It is plausible that a drug with ubiquitous pharmacologic effects might have unexpected adverse effects on homoeostatic mechanisms, when used in the third stage of labour Blood loss is a proxy outcome – we assume that reduced blood loss will translate to reduced deaths

Conclusion: These results should not be construed as evidence of an adverse effect of misoprostol on maternal mortality We propose, for further investigation, the hypothesis that misoprostol may have an as yet unexplained adverse effect on maternal homoeostatic functions in the third stage of labour.

Way forward: Suggestions 1.Ongoing systematic review of maternal mortality in misoprostol trials 2.Continued research on the benefits and risks of 3 rd stage misoprostol 3.Monitoring of implementation programs 4.Further research to determine lowest effective dose and best route

UTERINE MASSAGE TO REDUCE POSTPARTUM HAEMORRHAGE: A RANDOMIZED TRIAL Ferreira S, Singata M, Mangesi L, Abdel-Aleem H*, Hofmeyr GJ. Effective Care Research Unit, Eastern Cape Dept of Health, University of the Witwatersrand/Fort Hare, *Department of Obstetrics and Gynecology, Assiut University Hospital, Assiut, Egypt

Introduction: Active management of the third stage of labour, including injectable uterotonics, is effective in reducing postpartum haemorrhage (PPH) Worldwide, most deaths from PPH occur in women delivering at home, without access to uterotonic drugs There is a need to investigate simple measures, which can be applied universally at community level to reduce the burden of postpartum haemorrhage. The International Federation of Gynaecologists and Obstetricians (2003) recommend routine massage of the uterus after delivery of the placenta. Massage is thought to stimulate uterine contraction, possibly through stimulation of local prostaglandin release and thus to reduce haemorrhage.

Introduction (2) There is very little empirical research to evaluate the effectiveness of uterine massage in the third stage of labour If sustained routine massage is shown to be effective, promotion of this simple method through educational campaigns would have the potential to reduce maternal mortality from postpartum haemorrhage on a very wide scale. If not effective, attention can be directed towards other methods of reducing postpartum haemorrhage.

Objective: To determine the effectiveness of sustained uterine massage to reduce postpartum haemorrhage Collaborative program: Departments of Obstetrics and Gynaecology, East London Hospital Complex and Assiut University, Egypt

Preliminary Study H Abdel-Aleem, G J Hofmeyr, M Shokry, E El-Sonoosy Conducted in the Dept. of O&G, Assiut University Hospital Pregnant women admitted in labor were invited to participate. If the labor progressed to spontaneous vaginal delivery, women were randomly allocated, to one of two groups: –(1) Combined routine active management (oxytocin 10 IU I.V or I.M., immediate cord clamping and controlled cord traction) plus intermittent uterine massage every 10 minutes for 60 minutes (manual stimulation of the whole surface of the uterus using steady repetitive movements, as firmly as can be achieved without causing distress to the mother, till the uterus became contracted). –(1) Routine active management. A plastic drape was placed under the woman’s buttocks to collect all blood loss after delivery of the baby for 30 and 60 minutes.

Uterine massage vs Control: Mean Blood Loss (ml)

Uterine massage vs Control: PPH and Use of additional uterotonics

Discussion Limitations: –small sample size –staff could not be blinded to the group allocation. The chance of bias with respect to blood loss assessment was minimized by using objective, direct measurement. However, the use of other interventions could have been influenced by knowledge of the group allocation. Larger studies are needed to ascertain with more precision the effect of uterine massage on more substantive outcomes, and also the effectiveness in the absence of availability of injectable uterotonics.

Large study Protocol Pregnant women expected to deliver normally are invited to participate If the labour progresses to spontaneous vaginal delivery, consented women are enrolled by drawing the next in a randomly allocated numbered series of opaque sealed envelopes, to one of three groups: 1. Routine management. Oxytocin 10u IMI 2. Sustained uterine massage: Uterine massage commenced shortly after delivery and sustained for 30 minutes. Oxytocin delayed until after the 30-minute period of massage, unless blood loss of 500ml is measured before that time. 3. Combined routine management plus sustained uterine massage.

Protocol (2) In all 3 groups controlled cord traction is carried out according to hospital policy. A low profile plastic “fracture” bedpan is placed under the woman’s buttocks to collect all blood loss after delivery of the baby for 30 minutes. There is no other interference in the routine care of the woman. The primary outcomes are: 1. Blood loss >300ml within 30 minutes; 2. Placenta delivered >30 minutes after birth.

Analysis: Data will be entered into an excel data spreadsheet and analysed using epi info 2002 Outcome data will be compared using relative risks with 95% confidence intervals and for comparisons of medians the Mann-Whitney 2 sample test. To show a reduction of blood loss >300ml in 30 minutes from 13.5% to 9% with sustained uterine massage will require 800 women per group (2400 women).

Conclusions Uterine massage has been recommended without evidence of effectiveness or guidelines as to how and for how long it should be performed Sustained uterine massage is a promising intervention for widespread use in both hospital and non-hospital settings Funding needed to complete a large, multicentre trial