Cell salvage for postpartum haemorrhage during vaginal delivery: a case series

Abstract

Background

The safety and effectiveness of cell salvage for vaginal delivery is unknown. This case series aimed to assess the utility and adverse events related to the use of cell salvage for maternal haemorrhage during vaginal delivery.

Materials and methods

A cohort study design was chosen, focused on postpartum haemorrhages that occurred after vaginal delivery for which cell salvage equipment was requested to be set up in the labour and delivery room outside of a sterile operating room environment. Variables recorded included duration of stay in hospital, occurrence of wound infections, sepsis, thromboembolic events, and amniotic fluid embolism.

Results

Of 28 cases of postpartum haemorrhage during vaginal deliveries involving the setup or use of cell salvage equipment, ten were associated with successful re-infusion of salvaged shed blood. These ten cases were compared to the 18 cases in which cell salvage equipment was set up, but insufficient shed blood was salvaged for re-infusion. There were no instances of postpartum sepsis, wound infection, or thromboembolism associated with the use of cell salvage for vaginal delivery. Although one case of suspected amniotic fluid embolism occurred, severe symptoms began prior to the infusion of salvaged blood.

Discussion

Infusion of salvaged shed blood collected from a vaginal delivery field is feasible. The outcomes of these cases do not exclude an unacceptably high risk of infection or embolic events. Trials evaluating the safety and effectiveness associated with the use of cell salvage in vaginal deliveries are justified.

Introduction

The use of cell salvage in haemorrhage attenuates allogeneic blood product consumption^1,2. The acceptability of cell salvage in obstetric haemorrhage has increased over time. Historical fears about inducing amniotic fluid embolism by administering blood that contains amniotic fluid components have largely been assuaged by investigations that demonstrate high efficacy of foetal squamous cell and amniotic fluid-derived tissue factor removal, by using efficient washing processes in conjunction with modern leucocyte depletion filters^3–5. The washing system, coupled with these filters, effectively removes cellular components of amniotic fluid and is capable of removing 97–100% of micro-organisms encountered^5–8. To date, investigations on the safety of cell salvage in obstetrics have focused on its use in sterile operating room environments. The administration of salvaged shed blood from lower genitourinary tract bleeding is controversial^9,10. Teare et al. found that, after washing salvaged blood obtained during vaginal delivery, the measured number of colony-forming units (CFU) of bacterial species was similar to that found with cell salvage during Caesarean delivery. When the blood was washed, the investigators hypothesised that re-infusion would result in clinically insignificant bacteraemia, at levels similar to those that would be seen with common dental procedures (0.3–4.0 CFU/mL)⁹. Despite this evidence, there is a dearth of literature describing the use of cell salvage in vaginal delivery, and there are limitations that are inherent in the few existing observational studies on this topic. Thus, there remain unaddressed safety concerns that shed blood, contaminated by lower gastrointestinal or urogenital flora, is unsuitable for cell salvage.

We aimed to assess adverse events associated with cases of haemorrhage during vaginal delivery at our institution, for which cell salvage was utilised in the labour and delivery room.

Materials and methods

This database review was approved by the University of Pittsburgh Quality Improvement Review Board. The database was created as part of this institution’s blood management programme, and it tracks all cases in which standby cell salvage is set up and in which re-infusion of salvaged shed blood is performed during maternal haemorrhage. The creation and maintenance of information from this database is required by the hospital’s quality department as well as by the AABB (formerly, American Association of Blood Banks) peri-operative blood collection and re-infusion accreditation standards, for which this institution is accredited. Details about the database have been described in prior work by Milne et al.¹¹ which evaluated factors leading to successful intra-operative re-infusion of salvaged shed blood during obstetric haemorrhage.

For this study, cases of maternal haemorrhage during vaginal delivery over the period 2010–2017 with institution of cell salvage were included for review. Cases were categorised based on the successful re-infusion of salvaged blood: one group received re-infusion of salvaged shed blood that was acquired from the vaginal delivery field, and in a second group cell salvage equipment was set up, but inadequate blood was collected to process and re-infuse. The decision to use cell salvage was made at the discretion of the obstetric provider and was based on several factors including history of haemorrhage, atony or lacerations in the present haemorrhage, placental abruption or other antenatal bleeding associated with disseminated intravascular coagulopathy, retained products of conception, or bloodless medicine request (e.g. Jehovah’s Witness) (Table I). Blood was processed by one of two machines, either the Brat^® (Cobe Cardiovascular Inc., Arvada, CO, USA) or a Sorin Electa^® (DiaSorin SpA, Mirandola, Italy). Each event was reviewed independently by two reviewers (EK, JMZ). Data were abstracted for demographic characteristics, estimated blood loss, cell salvage volume returned, reason for blood recovery, duration of stay in hospital, and the occurrence of wound infection, sepsis, thromboembolic events, or suspected amniotic fluid embolism. Inconsistencies were resolved by consensus between the reviewers (EK, JMZ).

Table I.

Patients’ characteristics and obstetric details of cases for which cell salvage was implemented for postpartum haemorrhage during vaginal delivery.

	Salvaged lood transfused (n=10)	No salvaged blood transfused (n=18)	p-value
Age (years)	28.6±8.4	32.9±4.7	0.09
BMI (kg/m2)	32.6±7.0	32.2±5.4	0.86
Gravidity	2.1±1.5	2.8±2.9	0.50
Parity	0.7±1.0	1.3±2.9	0.51
EGA (weeks)	36.1±5.2	36.2±3.6	0.94
Race			0.19
White	5 (50)	12 (67)
Black or African American	4 (40)	2 (11)
Asian	1 (10)	4 (22)
GBS status			0.51
Positive	4 (33.3)	4 (50.0)
Negative	3 (33.3)	9 (25.0)
Unknown	3 (33.3)	5 (25.0)
Reason for cell salvage setup			0.47
Atony	6 (55.6)	5 (37.5)
Lacerations	1 (11.1)	4 (12.5)
Abruption/DIC	1 (11.1)	3 (31.25)
Retained placenta	1 (11.1)	1 (0)
Other*	1 (11.1)	5 (18.75)
Antibiotics received during vaginal delivery	8 (80)	10 (56)	0.25
EBL (mL)	1,731±1,136	1,409±1,073	0.49
Shed blood volume retrieved (mL)	560.2±415.5	127±267.4	0.00

Data are presented as mean±standard deviation or frequency (percentage).

BMI: body mass index; EGA: estimated gestational age; GBS: group B streptococcus; DIC: disseminated intravascular coagulation; EBL: estimated blood loss.

^*Other reasons for cell salvage setup include Jehovah’s Witness status or other bloodless medicine request, history of or current uterine inversion, or history of postpartum haemorrhage.

Descriptive statistics are presented as the mean and standard deviation for continuous data, and as a frequency and percentages for categorical data. The normality of the distribution of data was evaluated using histograms. Demographic features were compared using Fisher’s exact test, the Student’s t-test was used for normally distributed data, while the Wilcoxon’s rank sum test for data not normally distributed. Comparisons were made between groups for duration of stay in hospital, wound infection, sepsis, thromboembolic events, and amniotic fluid embolism. All statistical analyses were two-sided, and an α=0.05 was used to reject the null hypothesis. Analyses were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA, USA).

Results

Twenty-eight women with postpartum haemorrhage during vaginal delivery for whom cell salvage equipment was requested to be used were identified. Ten women successfully received infusion of shed blood salvaged from the vaginal delivery field. Eighteen women had cell salvage equipment set up, but did not successfully receive infusion of salvaged blood from the vaginal delivery field. The reasons for no infusion of salvaged blood included insufficient blood collected by cell salvage to permit re-infusion, and transfer to the operating room, where salvaged blood was collected and/or re-infused from a sterile abdominal field. Table I shows the baseline characteristics of the patients and the haemorrhages. The time spent in hospital was similar in the two groups. There were no instances of wound infection, sepsis, or thromboembolic events, and there were no differences between groups for allogeneic blood product transfusion (Table II).

Table II.

Outcomes associated with transfusion of salvaged shed blood during vaginal delivery.

	Salvaged blood transfused (n=10)	No salvaged blood transfused (n=18)	p-value
Duration of stay (days)	4.9±4.8	3.8±1.4	0.36
Wound infection	0 (0)	0 (0)	-
Sepsis	0 (0)	0 (0)	-
Thromboembolic events	0 (0)	0 (0)	-
Suspected amniotic fluid embolism	0 (0)	1 (5.5)	1.0
Allogeneic blood transfused, by product
PRBC (units)	2 [0–4]	0 [0–6]	0.11
FFP (units)	0 [0–2]	0 [0–6]	0.54
Cryoprecipitate (units)	0 [0–2]	0 [0–6]	0.92
Platelet (packs)	0 [0–2]	0 [0–1]	0.21

Data are presented as mean±standard deviation, frequency (percentage), or median [range].

PRBC: packed red blood cells; FFP: fresh-frozen plasma.

One case of amniotic fluid embolism was noted. This patient had a vacuum-assisted vaginal delivery. She had severe lower uterine segment atony and rapidly developed disseminated intravascular coagulation. She was transferred to the operating room, at which time cell salvage equipment was utilised, and 945 mL salvaged blood gathered from a sterile abdominal field were collected and re-infused. Prior to the intra-operative re-infusion of this blood, the patient had been intubated and hypotensive, requiring two vasopressor infusions. She had also received allogeneic transfusion with packed red cells, fresh-frozen plasma, cryoprecipitate and platelets. She developed acute right ventricular failure, cardiogenic shock, and multi-organ failure requiring renal replacement therapy. Extracorporeal membrane oxygenation was implemented for 4 days for severe acute respiratory distress syndrome. She was de-cannulated, extubated, and discharged with home healthcare on postpartum day 19. She was suspected of having had an amniotic fluid embolism or transfusion-related acute lung injury.

Discussion

Re-infusion of salvaged blood after vaginal delivery is feasible. Our data provide preliminary clinical information on outcomes associated with re-infusion of salvaged blood from vaginal delivery, supporting the future conduct of larger-scale trials on safety. Notably, the deliveries reported in this work took place at a tertiary care women’s hospital with a large delivery volume; thus, it may not be appropriate to generalise the economics and logistics of cell salvage to other clinical settings with smaller delivery volumes. Barriers to routine implementation of cell salvage for obstetric haemorrhage may include purchasing and maintenance costs for equipment as well as hospital-level costs for trained personnel to operate the equipment¹². The success of re-infusion of salvaged shed blood depends on multiple factors, including type of surgery, rate and mechanism of bleeding, surgical technique in control of bleeding, adequacy of capture of blood lost, and the quantity of haemolysis that occurs during processing¹³. It is worth noting that in the present study, ten of 28 cases of vaginal delivery (36%) successfully received re-infusion of shed blood, whereas Milne et al.¹¹ reported that 53% of vaginal delivery cases received intra-operative blood salvage re-infusion. The present findings suggest that cell salvage for vaginal deliveries managed outside of the operating room may be associated with reduced rates of successful re-infusion of shed blood. We suspect that these differences in rates of successful re-infusion may be related to the nature of the bleeding and capture of blood loss, where vaginal delivery haemorrhages managed outside of the operating room bleed differently and may not be as amenable to efficient cell salvage as those requiring operative management. The reduced rate of success in the more recent cohort may also reflect the increased acceptance of utilisation of cell salvage for vaginal delivery among our institution’s obstetricians over time, which has led to more cases of cell salvage equipment being set up because of the risk of bleeding rather than active and ongoing haemorrhage, thereby increasing the number of occurrences in which insufficient blood was shed for successful salvage and re-infusion.

Cell salvage is often used as part of a bloodless medicine strategy among patients who identify as Jehovah’s Witnesses. There were two women in this cohort who self-identified as Jehovah’s Witnesses, who declined transfusion of allogeneic blood products, but who accepted, in principle, re-infusion of shed blood salvaged during vaginal delivery. In neither case was sufficient blood salvaged to allow re-infusion. It is worth noting that the routine use of leucodepletion filters during cell salvage results in an unavoidable reduction of 20–30 mL of blood volume. Although this volume may be clinically insignificant, it may affect decision-making for Jehovah’s Witnesses who are weighing their conscientious decisions on bloodless medicine options. While elimination of the use of leucodepletion filters may be possible to circumvent this issue, removal may or may not pose additional risks of bacterial contamination, which could be of particular concern during the use of cell salvage in vaginal deliveries^4,9,14.

Conclusions

These ten cases did not result in infection, thromboembolism, or amniotic fluid embolism. However, given the rarity of such events, the likelihood of observing these events in our small population was low. Based on these preliminary findings, larger clinical evaluations of the use of cell salvage in vaginal delivery are warranted to determine the safety and effectiveness of this intervention.