Postpartum hemorrhage: a prospective, comparative study in Angola using a new disposable device for oxytocin administration
Abstract
Background. Postpartum hemorrhage (PPH) is the single most common cause of maternal death in the world, and oxytocin is known to be effective for its prevention and treatment. The use of syringes can be problematic in areas affected by HIV. The aim of this study was to introduce Uniject™ (a new disposable device for administration of 10 IU oxytocin) as part of active management of the third stage of labor (AMTL) and try to reduce PPH.
Methods. A prospective, comparative study was performed between March 1998 and May 2000 in Luanda. Seven hundred and eighty-two parturient women with physiological management were compared to 814 with AMTL. Postpartum lost blood was collected using a plastic sheet during labor and a bucket placed under a cholera bed for 2 h postpartum. Student's t-test and χ2 test were used.
Result. PPH was reduced from 40·4 to 8·2% and severe PPH (≥1000 ml) from 7·5 to 1% in the AMTL group (P < 0·001).
Conclusions. Uniject™ was well tolerated and offers an alternative for oxytocin administration. AMTL should be implemented also in resource-poor settings as a routine management to reduce PPH.
Abbreviations:
-
- AIDS
-
- aquired immune deficiency syndrome
-
- AML
-
- active management of the third stage of labor
-
- MMR
-
- maternal mortality ratio
-
- HIV
-
- human immunodeficiency virus
-
- IU
-
- international units
-
- PPH
-
- postpartum hemorrage.
Postpartum hemorrhage (PPH) is the most common single cause of maternal mortality, followed by sepsis, eclampsia, obstructed labor, and complications to illicit abortions (1,2).
It is estimated that, globally, around 150 000 maternal deaths may be caused by PPH each year (3). Field studies in low-income countries and in affluent settings on the prevalence of PPH have demonstrated figures in the range of 10–20% (4,5). Studies in Malawi and in the Philippines indicate a prevalence of slightly less than 10% (6,7), which is also estimated to be the global average figure (3).
PPH is defined by WHO as a bleeding equal to 500 ml from the genital tract during the first 24 h after delivery (8). This is an arbitrary limit, as a blood loss of 500 ml may be clinically insignificant in countries with low prevalence of pregnancy anaemia, while in many low-income countries, with a high prevalence of severe anaemia, the same blood loss may be life threatening.
Active management of the third stage of labor (AMTL) involves a combination of prophylactic administration of oxytocics, early cord clamping, and controlled traction of the cord. Most of the documentation from the advantages of AMTL has emanated from well-organized units in affluent societies. Data is limited regarding the applicability of this routine in settings with overburdened delivery units, scarcity of staff, and insufficient provision of drugs.
Angola is a country with a high maternal mortality. Demographic data is unreliable or missing to a large extent. The maternal mortality ratio (MMR) in 1990 was estimated at 1500 deaths per 100 000 live births and is one of the highest in the world (9). In the capital, Luanda, statistics are limited to institutional deliveries, which account for around half of the births in Luanda province. The institutional MMR for 1999 was 888 per 100 000 live births according to local statistics. In 1998, PPH was the cause of death in 21% of the 285 maternal deaths registered in the biggest maternity institution in the city, Maternidade Lucrécia Paím, a university hospital with about 17 000 births annually.
Due to scarcity of resources, no oxytocin was given as a routine in postpartum care. Against this background, AMTL represents a potentially valuable approach to improve postpartum care. Such an expected improvement might be further enhanced by the choice of method for oxytocin administration. The Program for Appropriate Technology in Health (PATH) has developed the UniJectTM device, which holds a prefilled dose of 1·0 ml (10 IU) oxytocin in a disposable, cushion-like package with a sterile needle attached. This device has been used in field trials for injectable contraceptives (10) and is a promising technology also for oxytocin administration. It has hitherto not been tested in any low-income country as an ingredient of AMTL for the prevention of PPH.
The objective of this study was to introduce the new device for administration of oxytocin to evaluate its efficacy and acceptability in clinical practice in an African setting. We wanted to elucidate whether this approach could help decrease the prevalence of PPH among women in obstetric care with extremely limited resources.
Subjects and methods
The study was carried out between March 1998 and May 2000 at the Maternidade Lucrécia Paím. The heavy flow of parturient women did not permit all women to be included. It proved necessary to define a time and an area not to interfere with the ordinary activity. This was done by limiting the study to the work hours between 0800 h and 1500 h and by studying women in only two rooms of the labor ward. Twelve midwives and four doctors were appointed for the study, taught the practical procedures, and instructed in filling in the study protocols.
Due to prevailing constraints in human and material resources and to the load of patients, it was judged impossible to design a randomized, controlled trial. Instead, we opted to perform a comparative study.
The first part comprised parturient women in the labor ward of MLP studied between March 3, 1998 and May 28, 1998 (n = 782). The women enrolled were in daytime spontaneous labor for vaginal delivery, presenting with a dilatation of the cervix of at least 3 cm at the first vaginal examination. Women with obstetric emergencies and multiple pregnancies were excluded. At the time of the study, parturient women in the third stage of labor did not routinely receive any oxytocic medication, and the umbilical cord was not actively pulled unless excessive bleeding called for action. Instead, the delivery of the placenta was passively observed, in what could be referred to as a physiological or expectant management of the third stage of labor. These women are referred to as ‘Group 1’ below.
Vaginal bleeding immediately after birth and during the first 2 h postpartum was collected and measured. This was done using a plastic sheet placed underneath the parturient woman directly after birth and allowing blood to be collected for measurement in a bucket while the women remained in a birthing position. The same bucket was then placed under the women while she moved to a so-called cholera bed, below which the bucket was located for a total of 2 h after delivery. Such a cholera bed has an opening in the middle, thus making it possible to collect the lost blood during this first period of rest without using swabs or sanitary pads.
The cut-off time of 2 h was selected for practical reasons taking into consideration the heavy work load of the labor ward, yet including the two most critical hours of PPH. Time interval between birth and expulsion of the placenta was registered as well as basic obstetrical data (type of presentation, episiotomy, birth weight, complications, and drugs given). Any bleeding beyond 2 h and other late complications were not possible to cover.
Among the women with PPH, we wanted to identify those with the most severe bleeding, who were in urgent need of treatment to reduce maternal morbidity/mortality. Severe PPH is here defined as a blood loss of 1000 ml.
Prolonged labor was calculated from portograms and was defined as a duration of labor which exceeded by more than 4 h the time limit set by an arbitrary standard of 1 h per cm cervical dilation and 2 h added for the expulsion phase.
The second part of the study consisted of an intensive, introductory course for midwives and doctors concerning AMTL encompassing routine administration of 10 IU oxytocin with the UniJect™ device, early cord clamping and gentle traction of the cord for delivering the placenta.
The third part included 814 women giving birth between November 8, 1999 and May 10, 2000, with the same inclusion criteria, methods, and protocol as in the first group. This second group was treated according to the principles of AMTL described above and is referred to as ‘Group 2’ below. The UnijectTM devices were stored in a refrigerator in the labor ward during the study period and were distributed daily to the midwives. These were trained to use UniJect™ for intramuscular injection into the thigh of the parturient women directly after delivery of the newborn as soon as multiple pregnancy had been clinically excluded. The duration of the third part was longer than the first part due to renovation work of the maternity ward during this period.
For comparison of the two groups in terms of postpartum blood loss and interval between birth and expulsion of the placenta, Student's t-test was utilized. For comparison between proportions, the χ2 test was used.
Ethical clearance has been given by the ethical committee at Karolinska Institutet, Stockholm, Sweden and by the Agostinho Neto University, Luanda, Angola.
Results
The results are summarized in Tables I–VI.
Blood loss | Group 1 (n = 782) (%) | Group 2 (n = 814) (%) | P value |
---|---|---|---|
<200 ml | 116 (15·0) | 399 (48·9) | <0·001 |
≥500 ml | 316 (40·4) | 67 (8·2) | <0·001 |
≥1000 ml | 58 (7·5) | 8 (1·0) | <0·001 |
Mean blood loss | 447 ml | 239 ml | <0·001 |
Parity | Prevalence of PPH group 1 (n = 782) (%) | Prevalence of PPH group 2 (n = 814) (%) | P value |
---|---|---|---|
1 | 100/262 (38·2) | 39/304 (9·5) | <0·001 |
2–3 | 104/278 (37·4) | 17/288 (5·9) | <0·001 |
≥4 | 112/241 (46·3) | 21/222 (9·5) | <0·001 |
Birth weight (g) | Prevalence of PPH group 1 (n = 778*) (%) | Prevalence of PPH group 2 (n = 807*) (%) | P value |
---|---|---|---|
1000 ≤ BW < 2500 | 44/113 (38·9) | 6/67 (9·0) | <0·001 |
2500 ≤ BW < 4000 | 262/654 (40·1) | 55/707 (7·8) | <0·001 |
BW ≥ 4000 | 9/11 (81·8) | 6/33 (18·2) | <0·001 |
- * Data missing and exclusion of BW < 1000 g.
Patient category | Group 1 (n = 750)* (%) | Group 2 (n = 810)† (%) | P value |
---|---|---|---|
Women with prolonged labor | 35 (4·7) | 42 (5·2) | 0·64 |
Prolonged labor and PPH | 19 (54·3) | 5 (11·9) | <0·005 |
Prolonged labor with oxytocine stimulation | 3 (8·6) | 6 (14·3) | 0·51 |
- * No data in 32 cases.
- † No data in 4 cases.
Episiotomy | No episiotomy | ||||
---|---|---|---|---|---|
No. of women | Blood loss (ml) [mean (SD)] | No. of women | Blood loss (ml) [mean (SD)] | Mean difference together with 95% CI (ml) | |
Group 1 (n = 750) | 219 | 444 (272) | 531 | 446 (312) | −2 (−47, 42)* |
Group 2 (n = 810) | 274 | 249 (205) | 536 | 234 (195) | 16 (−14, 45)* |
- * No significant difference in blood loss in episiotomy/no episiotomy subgroups before and after the introduction of AMTL.
Interval between birth and expulsion of the placenta (min) | Group 1 (n = 714) (%) | Group 2 (n = 768) (%) | |
---|---|---|---|
0–9 | 38 (5·3) | 687 (89·4) | |
10–19 | 463 (64·8) | 71 (9·2) | |
20–29 | 142 (19·9) | 8 (1·0) | |
30–39 | 63 (8·8) | 2 (0·3) | |
40–49 | 4 (0·6) | 0 (0·0) | |
50–59 | 3 (0·4) | 0 (0·0) | |
≥60 | 1 (0·1) | 0 (0·0) | |
Mean time (min) | 16·2 | 4·4 | (P < 0·001) |
There was no maternal death in the groups studied. The two groups did not differ concerning age (P = 0·50), proportion of primiparous women (P = 0·13), and grand multiparous women (P = 0·13).
Blood loss and interval between birth of the baby and delivery of the placenta decreased significantly with AMTL (Tables I and VI). The percentage of parturients with a small bleeding (≤200 ml) was significantly less in Group 1 than in Group 2 (P < 0·005). Severe PPH also showed a significant difference in the two groups and was detected among 59 women (7·5%) in Group 1. In seven of these, the cause of heavy bleeding was laceration of the cervix. In Group 2, severe PPH was detected in eight women (1%), four of those being caused by cervical lacerations (Table I).
The prevalence of PPH was significantly lower among both primiparous and grand multiparous women in Group 2. Within these groups, there was no difference among these two parity categories. The proportion of PPH was lowest among women with parity 2–3 within both groups (Table II).
Macrosomia of the baby (birthweight ≥ 4000 g) was detected in 11 cases in Group 1, of whom nine had PPH. In Group 2, there were 32 cases of macrosomia and six of those were associated with PPH. The prevalence of PPH was lower in Group 2 than in Group 1 among women giving birth to newborns weighing ≥ 4000 g (P < 0·005). The same holds true when comparing women with low birthweight newborns (Table III).
Prolonged labor occurred with similar prevalence in the two groups (around 5%). Few of these women were treated with oxytocin stimulation. The reduction of PPH after prolonged labor was significant when AMTL was used (Table IV).
Episiotomy is still used as a routine in primiparous women in Luanda, resulting in an overall prevalence of almost one third of all deliveries. The bleeding caused by episiotomy might contribute to the postpartum blood loss. We found a slight increase in blood loss among women having had episiotomy in Group 2 (borderline significance, P = 0·05), but no such difference in Group 1 (Table V).
Stillbirth occurred in 21 cases (2·7%) in Group 1 and in 14 cases (1·7%) in Group 2. Of these, 8 and 11, respectively, had no cardiac activity registered at the first examination after admission to the maternity ward and could therefore be regarded as prelabor fetal deaths. Of all stillbirths, one case in Group 1 and two cases in Group 2 had PPH.
The intervals between birth of the baby and expulsion of the placenta are shown in Table VI. In Group 2, the placenta was delivered in less than 10 min in 89·4% of the parturients as compared to only 5·4% in Group 1 (P < 0·001). The difference between the two groups was less marked after 30 min.
The prevalence of retention of the placenta depends on the definition of the normal time of the third stage of labor. Using the WHO criteria, we found only one case altogether. In Group 1, 71 patients (9·9%) had a placenta not delivered before 30 min, compared to 2 patients (0·3%) in Group 2 (P < 0·001).
Manual removal of the placenta was rare in both groups. Of the six cases in Group 1, five were associated with PPH (out of which two had severe PPH). In the single manual removal of the placenta recorded in Group 2, bleeding was normal.
Discussion
This study has shown that active management of the third stage of labor, using UniJect™, significantly reduces blood loss during the first 2 h postpartum compared to expectant management. It was also demonstrated that the duration of the third stage of labor was significantly reduced using this management.
Blood loss after delivery is usually not measured but more or less roughly estimated. This often leads to an underestimation of 20–40% (11) or even more when the bleeding exceeds 300 ml (12). The lack of accuracy is still more obvious in low-income countries, where time and staff constraints frequently impede accurate assessment of postpartum blood loss. Even when such loss is carefully measured, studies of the normal blood loss after delivery show a substantial variation (11,13).
In Group 1, the prevalence of PPH was more than 40% (a figure reduced to 29·5% when excluding women with an estimated bleeding of 500 ml). This prevalence is unexpectedly high. Contributing factors might comprise liver damage, secondary to the high prevalence of hepatitis A, B, and E found in Luanda (14). Such an association is beyond the scope of this paper but may suggest a field for future studies.
WHO has given attention to PPH in a number of documents. A Technical Working Group, convened by WHO in 1989, analyzed nine published reports of controlled trials in which oxytocic drugs were compared with either placebo or no routine prophylactic regimen. It was found that routinely used oxytocic drugs for active management of the third stage of labor reduced the risk of PPH by about 40% (15). The concluding recommendation from the group was that active management of the third stage of labor, where acceptable, should be promoted at the most peripheral level of maternal health care system. The same magnitude of hemorrhage reduction has later been confirmed in a Swedish study (16).
A meta-analysis of AMTL has been published in the Cochrane Database (including the additional components of early cord clamping and controlled cord traction) (17). In comparison with expectant management, AMTL was found to give a reduction of PPH of 62%. This suggests an additional effect of some 20% in reducing PPH due to the other components of active management apart from the oxytocin administration. The drug of choice to be used in AMTL in a low-income, often tropical, country is crucial. Current evidence demonstrates that oxytocin is more stable in warm climates than ergometrine or methylergometrine (18,19) and that an increased dose of oxytocin equals the effect of the standard dose of ergometrine (20). Oxytocin has also been shown to have some advantage in comparison with misoprostol (21), but the clinical significance of this has been questioned (22).
The normal interval between birth of the child and delivery of the placenta has been subject to different views with a range of 10–60 min (23). In a survey of 1300 normal deliveries, without the routine of AMTL, Cooms and Laros (24) found an average duration of 6 min for the third stage of labor and in only 3·3% the duration exceeded 30 min. The mean time for placenta delivery in Group 2 was 4·4 min in this study, while it was 16·4 min in Group 1, a reduction in agreement with other studies (16).
The prevalence of retention of the placenta varies in different reports from 0·6 to 5·5% depending on the definition used for the normal duration of the third stage of labor (23). It has been suggested that retained placenta could be more common in practice with routine administration of oxytocics (25,26). This risk was confirmed neither by the present study nor by a recent retrospective study from Ghana (27). There have also been concerns regarding the risk of uterine inversion with cord traction (26). We did not encounter any such case.
More importantly, we did not find any case of maternal death during this study. This could partially be explained by the fact that the most dramatic cases of antepartum bleedings have been excluded, as they were treated in the operation theatre on an emergency basis. We also believe that the study situation itself with more attention to the problem of bleeding and improved vigilance by the personnel helped in avoiding fatal complications.
This is the first study in which cholera beds have been used to collect postpartum blood loss. After comparison with a calorimetric measurement of hemoglobin, this simple method has been found appropriate and an alternative to more costly methods (28).
In a parallel study (Jangsten et al. unpublished) in the same setting, it was found that midwives and parturient women had a favorable attitude to AMTL using the new Uniject™ device. The midwives had no difficulties in learning the administration technique of Uniject™ and were satisfied with the decreased prevalence of hemorrhage complications among their patients. Two subpopulations of parturient women receiving the two kinds of treatment were interviewed blindly based on a semistructured questionnaire and a verbal rating scale to estimate pain and satisfaction of care. It was concluded that women subjected to AMTL experienced somewhat more postpartum pain than women with physiological management. This was further accentuated among breast-feeding and grand multiparous women. Spontaneous verbal reactions to pain and the duration of postpartum pain in relation to birth could not be differentiated in the two subpopulations. We interpret these findings as indications of satisfactory compliance to the use of Uniject™ as part of AMTL.
In conclusion, this study demonstrates a more dramatically beneficial effect of AMTL than in the quoted meta-analysis from more affluent countries. To our knowledge, this is the most extensive prospective study in Africa elucidating the role of AMTL. In most low-income countries, budgetary constraints in the health sector require simplified technology in public health. Parenteral oxcytocin is needed for AMTL. By using Uniject™, the logistical problems of syringes and needles are avoided. The device showed to be well appreciated by both midwives and parturient women. As Uniject™ cannot be reused, it does not add to the risk of transmission of blood-borne pathogens (29). Recently, attention has been drawn to the association of unsafe injections and the rapid spread of HIV/AIDS in Africa (30). The importance of this argument and the results of this study imply that production of large quantities of Uniject™ will be justified. This would offer a presumably less costly and less risky alternative than the prevailing routine.