Postpartum Hemorrhage in Humanitarian Settings: Heat-Stable Carbetocin and Tranexamic Acid Implementation Study in Uganda

Background of the Study

Postpartum hemorrhage (PPH) is defined as excessive blood loss (≥ 500 mL after vaginal birth or ≥ 1000 mL after cesarean birth within 24 hours). This condition contributes to over 25% of maternal deaths worldwide, primarily in low and middle-income countries (LMIC) affected by fragility and humanitarian disasters.^[1,2] Active management of the third stage of labor (AMTSL) can prevent most PPH deaths.^[3] It includes prophylactic uterotonic use, delayed cord clamping, and controlled cord traction (if skilled birth attendants are available).^[4] Oxytocin (10 IU, IM/IV) is the standard uterotonic recommended in settings with multiple options but requires cold-chain storage.^[1,5] Heat-stable carbetocin (HSC), a long-acting oxytocin analog, is recommended for PPH prevention but does not require cold-chain storage. Studies, including a 2018 trial by the World Health Organization (WHO), have shown HSC effectiveness and noninferiority to oxytocin in preventing PPH in low-income countries.^[6-8] Furthermore, it was found cost-effective in reducing the number of PPH events and deaths, as well as public healthcare system costs in India.^[6-8]

The WHO recommends prompt action in a multidisciplinary approach for treating PPH, including fluid replacement and additional uterotonic administration (following the initial dose for PPH prevention) as the primary interventions.^[1,5,9] Tranexamic acid (TXA), an antifibrinolytic agent, reduces bleeding risk in surgery and trauma.^[10,11] In obstetric care, TXA effectively controls bleeding in women with PPH regardless of the underlying cause.^[12] It does not require cold chain systems, is affordable ($1.30 per 1g ampoule), and is readily available in hospitals.^[7]

In 2022, two scoping literature reviews examined the existing implementation experience of HSC and TXA in low-resource countries, emphasizing their suitability in environments with limited cold-chain storage facilities.^[13,14] The reviews found that most research originated from hospital-based studies and lower-middle-income countries with limited findings from low-resource or humanitarian settings, including rural or basic emergency obstetric facilities. On August 21, 2023, we searched PubMed using the term “randomized controlled trial” combined with “postpartum,” “hemorrhage,” “vaginal,” and “carbetocin” or “tranexamic” for publications in English, focusing on trials in LMIC (as defined by the World Bank) over the past ten years.^[15] We identified one trial since the 2022 scoping literature reviews, the E-MOTIVE trial (Early Detection of Postpartum Hemorrhage and Treatment), which took place in hospitals—and not in lower maternity care facilities— in Kenya, Nigeria, South Africa, and Tanzania. It found that detecting PPH early using a calibrated blood-collection drape and implementing bundled treatments (uterine massage, oxytocic drugs, tranexamic acid, intravenous fluids, examination, and intervention escalation) resulted in a reduced risk of severe PPH, laparotomy for bleeding, or bleeding-related death, compared to standard care in women undergoing vaginal delivery.

Consequently, a gap exists within the extant body of evidence that describes strategies for implementing and building capacity to introduce HSC and TXA in humanitarian and fragile settings, and specifically in non-hospital contexts.

Objectives of the Study

Against this background, the overall objective of our implementation research consisted of examining the impact of an HSC and TXA-inclusive PPH capacity-strengthening package on the utilization of uterotonics (including HSC) for PPH prevention, PPH detection, and utilization of uterotonics and TXA for PPH treatment in low-resource cold-chain-challenged Basic Emergency Obstetric and Newborn Care (BEmONC) facilities. These facilities are commonly found in humanitarian settings. The intervention package is detailed under Methods.

Country Context

Uganda has witnessed a notable decline in its maternal mortality ratio over the past two decades. According to recent data, this ratio has decreased from 461 deaths per 100,000 live births in 2010 to 336 and 284 deaths per 100,000 live births in 2016 and 2020, respectively.^[2] In a study conducted by Ononge et al. in 2016, the overall incidence of PPH was found to be 9%.^[16] Another study revealed that merely 34% of women received all three AMTSL components, namely prophylactic uterotonics, delayed cord clamping, and controlled cord traction.^[17]

As of June 2023, Uganda, a country with a population of approximately 46 million inhabitants, provided shelter to 1.5 million refugees and 48,953 asylum seekers, with women and girls constituting 51% of this population.^[18] Approximately 60% of these individuals originate from South Sudan and have settled in the West Nile region, while 30% come from the Democratic Republic of Congo and have found refuge in Western Uganda.

Study Design and Participants

Our study utilized implementation research to thoroughly examine the effectiveness and mechanisms of the intervention in real-world scenarios, aiming to identify avenues for improvement.^[19] Our research study adopted a pragmatic, multi-intervention, multi-stepped, cluster-based design, which was non-blinded, non-randomized, but controlled. Clustering was appropriate since the PPH capacity-strengthening strategy focused on healthcare facilities rather than individual women. The study encompassed six facilities—Bidibidi Health Center (HC) III, Barakala HCIII, Uriama HCIII, Yinga HCIII, Ucea HCIII, and Nakivale HCIII—located in West Nile and the Western Region of Uganda. Facilities were included if classified as BEmONC-only and had recent cold storage disruptions, expressing willingness to participate. Facilities offering Comprehensive Emergency Obstetric and Newborn Care (CEmONC) were excluded. All birthing women were eligible without specific exclusions and were invited to participate in the study.

Stepwise Approach

A standardized sequence [Figure 1] was followed across the six sites. Step 1 (T1) encompassed 6 weeks of routine care. This was followed by a one-month period to prepare for step 2 (T2), which involved a 6-week period without HSC and TXA. Step 3 (T3) featured a 6-week interval with HSC but no TXA. Step 4 (T4) comprised 6 weeks with both HSC and TXA.

Close

Figure 1:

The four steps of the study. HSC: Heat-stable carbetocin, TXA: Tranexamic acid.

Export to PPT

Intervention

The PPH package incorporated three interventions for provider enhancement. First, refresher training focused on PPH skills through a low-dose, high-frequency method: 2-week self-study/webinars followed by a 2-day workshop. Monthly facility-based skills drills and reviews were used to bolster learning. Second, a WhatsApp-based community of practice promoted provider communication and collaboration. Third, PPH algorithm posters in delivery rooms aided adherence to protocols. Routine care included oxytocin, misoprostol, HSC, but not TXA, following national guidelines, although HSC was not yet available nationally at the time of the study.

Outcomes

The evaluation of primary outcomes encompassed four indicators. The first indicator was the proportion of women who received a uterotonic for PPH prevention by type of uterotonic utilized. Second was the proportion of women visually diagnosed with PPH. Third was the proportion of women diagnosed with PPH who received oxytocin for PPH treatment. Finally, was the proportion of women diagnosed with clinical PPH who received TXA for PPH treatment. Due to time and resource constraints in this implementation research study, the collection and management of quantitative data were streamlined. For example, instead of utilizing specifically designed case-report forms at the individual participant level, maternity registers were employed. The national research coordinator thoroughly examined all submitted documents, identified discrepancies, and worked with facility providers to resolve them before sending the documents to the data manager. Furthermore, quantitative outcome variables were not collected to evaluate the level of implementation of the intervention package, as recommended by Peters et al.^[19] Instead, qualitative research methodologies were employed to assess the perspectives of healthcare providers in this regard, constituting the study’s secondary outcomes, which are reported elsewhere.

Statistical Analysis

For robustness, we estimated the sample size in our quasi-experimental research based on uterotonic usage assumptions for PPH prophylaxis. Starting from an initial 30% uterotonic usage (T1),^[17] we conservatively anticipated 60% in T2, persisting in T3 and T4. Hence, ≥504 participants (6 facilities * 4 phases * 21) were required. Statistical power calculations employed R’s pwr package, yielding 80% power to detect a 30% difference (60%–30%) at a 5% two-sided significance level with a design effect of 1.5.^[20] The use of the design effect was necessary due to the cluster-based sampling approach employed in the study.

Data were analyzed using IBM SPSS Statistics (v25.0) (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.) and WINPEPI (v11.50). Participants served as assignment and analysis units. Considering patient sequential attendance, proportions satisfying key outcomes per phase were calculated, applying a clustering-based design effect of 1.5. Two-wise proportion comparisons were computed with WINPEPI. Result comparability was ensured by standardizing T2–T4 prevalences with T1 as the reference, accounting for variables like age, gestational age, and parity. Missing data were not imputed. Findings adhere to TREND and CONSORT reporting guidelines.^[21,22]

Ethical Approval

Ethical approval was granted by the Uganda National Council for Science and Technology (reference HS2456ES).

Sociodemographic Characteristics

During the period spanning from April 10, 2022 to April 4, 2023, a total of 2299 women from six healthcare facilities were included in the study. The participants were distributed as follows: 643 in T1, 570 in T2, 580 in T3, and 506 in T4. Noteworthy variations in baseline characteristics of the participants, including age, gravidity, parity, gestational age, delivery mode, and alive status upon discharge, were not observed across the different phases of the study, as indicated in Table 1.

Main Outcome Results

The primary outcomes of the study are presented in Table 2. The administration of uterotonics to prevent PPH was observed in nearly all participants, with no significant differences noted across the different phases. Notably, a significant change was observed in the types of uterotonic agents administered before and after the introduction of HSC. In the two initial phases (T1 and T2), oxytocin was predominantly used in 580 (93%) of 622 in T1 and 521 (92%) of 558 in T2), whereas HSC alone was the primary choice in the subsequent phases (421 (74%) of 560 in T3 and 270 (54%) of 494). Additionally, combined administration of uterotonics was recorded, with oxytocin and misoprostol being commonly used during the initial two phases (6% in T1 and T2, 2% in T3, and 1% in T4, unadjusted). Furthermore, during the later phases, HSC was administered in combination with either oxytocin (2% in T3 and 8% in T4, unadjusted) or misoprostol (1% in T3 and 3% in T4, unadjusted). There was no reported use of HSC for treatment or of TXA for prevention.

The incidence of visually diagnosed PPH in women ranged from 3% to 4%, with a notable decrease to 1% in T3. In terms of PPH treatment, oxytocin was administered universally in T1 (24 of 24) and T2 (25 of 25). However, its usage decreased to 50% in T3 (4 of 8) and further declined to 33% in T4 (5 of 15; T4–T1 95% CI: –100.0 to –30.9). TXA use varied across phases, but the differences were not statistically significant: 0% (0 of 24) in T1, 8% (9 of 16) in T2, 0% (0 of 8) in T3, and 33% (5 of 15) in T4. No serious adverse events related to the study were reported, and all participants were discharged alive.

Strengths and Limitations of the Study

While our study merits attention for its focus on humanitarian and fragile settings, it has limitations. First, data collection relied on providers’ reports in maternity registers, potentially introducing errors. To mitigate this, weekly data reviews were conducted. Medication timing was not recorded due to its absence in registers. Second, blinding was not feasible in pragmatic facility-based implementation research, posing assessment bias risk. However, the observed HSC uptake could not be solely attributed to the Hawthorne effect, reporting errors, or bias. Third, our implementation research was not randomized and had six clusters, raising the risk of dissimilarities between the groups. However, each cluster had sizable samples, followed similar phased approaches, and included controlled phases. After adjustment for factors like age, gestational age, and parity, and accounting for clustering, results remained largely unchanged.

Key Messages

• A postpartum hemorrhage (PPH) intervention package, enhancing routine obstetric services, can effectively increase the trends and types of uterotonics employed for PPH prevention [replacing oxytocin with heat-stable carbetocin (HSC)] while enabling tranexamic acid (TXA) administration for one-third of PPH cases.

• The gradual integration of HSC and TXA in humanitarian and fragile contexts is feasible and acceptable to clinicians.

• Continuous capacity enhancement and vigilant oversight is needed to avert and rectify potential confusion between established and novel pharmaceutical agents.