8th Postgraduate Course for Training in Reproductive Medicine and Reproductive Biology

Maternal morbidity

M. Boulvain
Department of Obstetrics and Gynecology
Geneva University Hospital

1. How to define maternal morbidity?

2. How to evaluate the prevalence of MM?

3. Evaluation of risk factors for MM

4. Evaluation of interventions to reduce the risk of MM

Introduction

What is a maternal morbidity?

Any departure, subjective or objective, from a state of physiological or psychological well-being (Last, A dictionary of epidemiology, 1995)

(during pregnancy, childbirth and the postpartum period up to 42 days or 1 year).

Conditions or risk factors for subsequent handicaps (sequelae)

• Prolonged labour
• Haemorrhage
• Sepsis
• Preeclampsia

Resulting handicaps (sequelae)

• Uterine rupture and uterine scar
• Infertility
• Perineal or low abdominal pain
• Anemia
• Uterine prolapse
• Fecal and urinary incontinence

Maternal morbidity (MM) is difficult to measure, for several reasons. Because definition of what is MM could be very different according to authors, comparisons between studies are limited. Criteria to diagnose the diseases could also vary. Despite the fact that maternal mortality is a clear-cut condition, surveys are not easy to perform. Surveys to estimate the prevalence of MM are even more difficult to conduct. Several MM are difficult to diagnose, and require an (pelvic) examination, which is seldom possible in surveys (privacy, shyness about sexual and reproductive matters). Thus, very little information is available on morbidity, especially in the developing world.

Where the estimate comes from ?

The most frequently quoted estimate of maternal morbidity is that morbidity is 16 times more frequent than mortality, during pregnancy.

This is based on a small cohort study conducted in India in 1980. A total of 390 pregnancies, among 290 women, were included.

• there were 2 maternal deaths in the cohort (both post-partum haemorrhage).
• 33 episodes of illnesses related to the pregnancy were recorded.

How the authors define morbidity ? Any illness reported by the women. The authors do not present details.

Message

• always read the original report, as reviews and quotations are often interpretations not supported by the original data

Anaemia

Definition(s) of anaemia

Anaemia is defined during pregnancy as an haemoglobin (Hb) level below 110gr/L (WHO, 1992). During pregnancy, the Hb level is lower than outside pregnancy, and varies according to gestational age. Most women with Hb levels below this limit have normal pregnancies. Using the above definition, 20 to 50% of women, and even more in some areas, are considered as anaemic. Thus, anaemia could be either a very frequent disease, or a variation of the normal state. High levels of Hb and haematocrit are associated with preeclampsia. As a consequence, there is a need for a consensus on what is a clinically meaningful definition of anaemia during pregnancy.

Magnitude of the problem

Region	% of women Hb <110	Number in thousands
World	51	58 270
Developing	56	55 750
Developed	18	2 520
Africa	52	11 450
Asia (except Japan...)	60	40 140
Latin America	39	4 030
North America	17	570
Europe	17	920

Is moderate anaemia a risk factor for adverse outcomes ?

Several studies have suggested an association between anaemia (80 to 110 g/L) and preterm delivery or intrauterine growth retardation. However, some of these studies are not valid, and provide an estimate biased towards a greater risk associated with anaemia. Why ?

The variable of interest (anaemia) is associated with another factor (change in Hb with gestational age), which is associated with the disease (preterm delivery); thus the observed association with the disease is not a genuine one. This is called a confounder.

Haemoglobin level varies with gestational age. Studies where Hb level is measured at delivery show a lower mean level of Hb in women delivering preterm. The association found is confounded by gestational age at Hb sampling.

Routine iron supplementation during pregnancy

There is a rationale behind routine supplementation with iron, if the target disease (anaemia) is so frequent during pregnancy. However, to be recommended, such policy must have been shown effective in reducing clinically important adverse outcomes.

• Systematic review of randomised trials.

Population: In most trials, inclusion of women with Hb >100 g/L, before 28 weeks.

Intervention: usually, 100 mg iron (element) daily, versus placebo or no iron.

Outcome measures which were reported in RCTs: low pre-delivery Hb, low Hb 6 weeks post-partum, hypertension, caesarean section. Authors usually do not report on preterm delivery, IUGR, perinatal death, need for transfusion, fatigue which would be meaningful outcome measures.

• RCTs demonstrate that routine supplementation during pregnancy reduce the risk of having an Hb level below a certain level (e.g. <100 g/L) at the delivery or 6 weeks postpartum. Generally, all haematological parameters are improved by such a policy.
• However, RCTs do not demonstrate that routine supplementation during pregnancy reduce the risk of other outcomes (e.g. preterm delivery, IUGR)
• One trial has compared routine versus selective supplementation when Hb <100 g/L (actually, this is the real clinical question). This trial has shown a reduction of the risk of caesarean section, of transfusion, and an increase in miscarriage and stillbirth. These results are difficult to interprete and need to be replicated.
• A policy of routine supplementation with iron, especially in area where anaemia is prevalent, is probably beneficial. However, the evidence for such policy is not very strong.
• What is the most effective form of iron supplementation ? Oral iron is poorly absorbed. Iron is also associated with side effects, so most women are not compliant with the treatment (<30%). Weekly oral administration may have a similar effect as daily administration with fewer side-effects. Another alternative could be IV or IM administration. There are problems associated with intra-muscular administration of iron (iron deposits, allergy..). This form of administration might be more reliable, but this has not been evaluated by a RCT.
• Effectiveness of treatment for infections associated with anaemia, as malaria and worms. It seems logical to prescribe a prevention in regions where these infections are prevalent. However, there is no evidence, at the present time, clearly demonstrating the effectiveness of these interventions.

Interventions to prevent postpartum anaemia

Prevention of postpartum haemorrhage is one of the more effective interventions to reduce maternal mortality. A 500 ml blood loss is considered as normal and will not cause significant symptoms in women not anaemic before the delivery. RCTs have demonstrated that oxytocin or ergometrin administration is an effective intervention to prevent postpartum bleeding and the need for transfusion

Unfortunately these drugs (especially Ergometrin) are not stable in tropical conditions. This was shown by dosage of the drug in tablets and vials found in pharmacies, and by serial dosages in batches transported to tropical countries, in routine conditions (by air and by sea) Feasibility of this prevention is therefore questionable in some settings. Also, both drugs need IV or IM administration (ergometrin tablets are less effective and very sensitive to heat and humidity).

A large randomised controlled trial to assess Misoprostol as a prevention of post-partum haemorrhage is currently conducted. This drug is stable in tropical conditions, and can be administrated orally. If effectiveness is equivalent to Oxytocin, the advantages of Misoprostol might improve the feasibility and acceptability of prevention of post-partum haemorrhage.

Obstetrical fistulas

Obstetric fistulas are one of the most frequent and devastating forms of maternal morbidity in some developing countries. Obstetric fistulas are prevalent in countries where the general level of development is less advanced. The distribution of cases is not uniform within countries: women living in rural areas are more afflicted than those living in urban areas, and obstetric fistulas are more prevalent among women of specific ethnic groups. The major cause is the lack of availability of caesarean section in the case of cephalo-pelvic disproportion. In Europe and North America, obstetric fistulas represented a major source of maternal morbidity until caesarean section became available to women presenting with obstructed labour. A fistula hospital was build in New York at the end of last century to treat those patients.

Magnitude of the problem

The frequency of obstructed labour is reported to be in the range of 1 to 5% of deliveries, depending the author and on the definition used. Some hospitals, acting as referral units, report a proportion of admissions for obstructed labour as high as 10 to 20% of deliveries. Following obstructed labour, the risk of fistula is 5 to 10%. Combining the above probabilities, the risk in the general obstetric population may fall in the range from 0.5 to 5/1000 deliveries.

The precise incidence and prevalence of obstetric fistula are unknown, however there have been some attempts to assess the magnitude of this problem (table 1). Nearly 800 new cases of fistulas are admitted each year in Addis Ababa Fistula Hospital. When related to the total births in Ethiopia, an incidence of 0.55/1000 births can be computed. A similar estimate can be derived from the number of admissions in four major town's hospitals in Chad. As this condition primarily affects women following their first pregnancy, the risk for nulliparous women in these countries lies in the range of 1 to 5 per 1000. The above estimates only take into account women who present to selected hospitals after a vesicovaginal fistula has developed, relative to the total number of births in the country, thus underestimating the real incidence and prevalence of the condition.

Population-based estimates would be more representative of the real frequency of obstetric fistulas; unfortunately this information is currently not available. However, studies to estimate the real incidence and prevalence of this morbidity would be very difficult to conduct.

Causes of obstetric fistulas

Obstructed labour, and obstetric interventions performed in that context (i.e.: symphysiotomy, forceps delivery) are the principal causes (80 to 100%) in developing countries. In other countries, the main causes include cervical cancer and accidents during pelvic surgery.

During obstructed labour, bladder, rectum, and vaginal tissues are damaged by prolonged pressure of the head of the baby on the pelvic bones. Post-partum, these tissues become necrotic and slough. A false passage between the bladder and vagina is created, resulting in a constant leak of urine. The most severe dystocias may result in fistulas between the rectum and vagina, resulting in a leak of faeces.

Natural history of obstetric fistula

During a prolonged labour, the soft tissues of the pelvis are compressed between the fetal skull and mother's bony pelvis. When caesarean section is delayed, a prolonged period of impaction may result in an area of ischemia on the surrounding tissues. Anteriorly, this may cause a fistula between the bladder and the vagina (vesico-vaginal), and on the posterior aspect, between the rectum and the vagina (recto-vaginal fistula). Vesicovaginal fistulas alone account for 80% of obstetric fistulas, and the remaining 20% are either a rectovaginal fistula alone, or a combination of both. Most of prolonged labours are conducted home, however more than half of the women finally deliver in a health centre.

After delivery, either vaginally or by caesarean section, the vaginal mucosa may initially appear normal. A few days post-partum, the area of ischemia becomes purple or black, and the area begins to slough. Incontinence secondary to fistula may appear at that time. Over the following days, the necrotic areas continue to slough, and infection may further impair tissue healing. The bladder and vagina mucosa are the site of an inflammatory process. Subsequently, fibrotic tissue forms around the defect. Scarring results in a defect which is usually smaller than the one that existed immediately after the injury. Some small fistulas are reported to heal completely by this process. Several weeks later (10-12 weeks), as the inflammatory process subsides, surgical repair becomes feasible.

Consequences of obstetric fistulas

The medical, social and economic consequences of these fistulas are dramatic. The offensive odour results in social ostracization. Abandonment and divorce are frequent, and rejection by other family members may also occur as the condition persists for a period of time.

Most labours which result in fistula also are associated with stillbirth, and subsequent infertility is frequent.

How to address the problem ?

Two strategies addressing this problem have received the most attention: primary prevention of obstructed labour and surgical treatment of established fistulas. Prevention of obstructed labour is the most desirable approach, as it will dramatically reduce the incidence of obstetric fistulas, as observed in developed countries during the last century. Surgical treatment is certainly effective, achieving cure rates approaching 90% in specialised centres. However, both strategies have their limitations.

To prevent obstructed labour, high quality obstetrical services must be available to all pregnant women; women must be willing to go to hospital for delivery if a complication arises; transportation must be available day and night in case of emergency; and women (and/or the family or the community) must have the capacity to pay for the needed services. One or more of these conditions are absent in many areas. Most of the conditions are linked to global development and improvement of the social condition of women in developing countries. Comprehensive programs aimed at improving medical services are difficult to implement and their impact on obstetric mortality and morbidity may not be noted for several decades.

Surgical treatment of fistula is very difficult. This type of surgery needs highly skilled surgical and nursing care. Few units are capable of providing treatment for these women and many women do not have access to these specialised facilities.

Examples of such units are in Ethiopia (Fistula Hospital in Addis Ababa), Sudan (Dr Abbo), Nigeria (Kees Waldijk in Katsina, Nothern Nigeria; Ann Ward in South-East), Mali (Médecins du Monde in Mopti).

Potential method of prevention

Several factors may influence the post-partum probability of developing a fistula. Subsequent to the initial ischemia, the damaged tissues are extremely fragile. The pressure exerted by the fetal head on the trigone and urethra is likely to cause oedema and urinary retention after the delivery. This places a strain on the fragile area of the bladder and may further decrease the blood supply by stretching of the vessels. The necrotic material may become infected, thus impairing the process of healing. On the other hand, both tissues (bladder and vagina) are prone to regenerate if local conditions are favourable.

Several authors report healing of small fistulas following a 4-6 weeks period of bladder catheterization. In light of these reports and the arguments of biological plausibility presented above, bladder drainage following obstructed labour would appear to be a logical method to improve bladder and vaginal mucosa healing after ischemia. This may reduce the risk of vesicovaginal fistula following an obstructed labour.

Little information is available as to current hospital policies regarding bladder drainage following an obstructed labour. Some experienced authors advise to routinely catheterise the bladder in such situation, in order to decrease the pressure on the ischemic tissues (Chassar Moir, Lawson, Phillpott, Technical working group sponsored by WHO). Harrisson, in Zaria (Nigeria), has implemented a policy of six days bladder drainage following prolonged labour. This resulted in a frequency of vesicovaginal fistula of 5% after prolonged labour. However, comparison with the outcome in women not submitted to bladder drainage following delivery is not available, as no controlled studies have been conducted.

Pressure exerted by the catheter and the balloon may decrease the capacity to heal of bladder mucosa. The cost involved with the intervention, and the longer hospital stay may be a major obstacle for the implementation of this method of prevention. Possible side effects of prolonged bladder drainage, including urinary tract infection, must be balanced against potential benefits.

The present lack of evidence of effectiveness, and the potential for serious side effects prevents the implementation of this method of prevention on a large scale. A comprehensive evaluation of this intervention by a randomised controlled trial is needed before implementing a systematic policy of bladder drainage after obstructed labour.