Risk factors
9.1 Introduction
The previous chapters have all focused on adverse pregnancy and birth outcomes as risk factors for foetal loss and neonatal death. The present chapter takes the discussion and analysis one layer further to the underlying risk factors for the selected adverse pregnancy and birth outcomes themselves. These risk factors are more remote in the causal chain and include maternal factors, complications of the placenta, cord, and membranes, and birth complications. These have been referred to earlier, in Chapter 3, in the operationalisation of the conceptual model (Figure 3.6). Besides their direct effect on pregnancy/birth outcome, they are believed to affect foetal loss (i.e. spontaneous abortion and stillbirth), both directly and indirectly, and neonatal death indirectly through the intermediate outcomes (i.e. adverse pregnancy and birth outcomes).
As became clear in Chapter 3, many different causes and risk factors underlie the adverse pregnancy and birth outcomes, but what appear to be main risk factors? Are these risk factors also important in regions in transition and, more specifically, in Kerala? Moreover, does the relative importance of these risk factors change during the epidemiologic transition?
A region’s position within the transition could very well be reflected by its specific combination of underlying risk factors. Factors such as infections and malnutrition are likely to be important during the early of the transition, while others such as diabetes and advanced maternal age may become more significant during later stages. This could lead to a redefinition of the concepts of endogenous and exogenous mortality (see Chapter 2). For example, preterm birth, which is generally regarded as an endogenous or intrinsic cause, has been associated with maternal infection (exogenous or external) during pregnancy (see Section 3.4.2).
The current chapter identifies and discusses several main risk factors for the adverse pregnancy and birth outcomes. In Section 9.2, the following risk factors are selected for further analysis: hypertensive disorders of pregnancy, diabetes, maternal infections, smoking and alcohol consumption, anaemia, advanced maternal age, antepartum haemorrhage, and prolonged and obstructed labour. Before the results from the hospital survey in Kerala are presented, Section 9.2 defines and describes these risk factors on the basis of a brief literature review. In addition, some figures from secondary sources are presented on the frequency of these risk factors in developed countries, i.e. the EME region. Subsequently, Section 9.3 analyses data from the survey in SAT Hospital, Thiruvananthapuram, Kerala. Again, the relative importance is assessed at both the individual level and the population level. Finally, the results of the chapter are summarised and discussed in Section 9.4. Combining the results
with previous findings in Chapter 8 yields an understanding of important pathways to stillbirth and neonatal death in Kerala, a region in transition.
9.2 Risk factors
Risk factors for the selected adverse pregnancy and birth outcomes (congenital anomalies, LBW, preterm birth, IUGR/SGA, and birth asphyxia) that are repeatedly mentioned in the literature include: infections during pregnancy (a long list including rubella and cytomegalovirus), diabetes (diabetes mellitus and gestational diabetes), hypertensive disorders (e.g. pregnancy-induced hypertension, pre-eclampsia), small maternal size (weight and height), medication, alcohol and smoking, (mal)nutrition, placental problems (e.g. placenta praevia), and cord problems (e.g. cord entanglement). Furthermore, advanced maternal age is frequently seen as a cause of chromosomal anomalies, such as Down’s syndrome, while complications at birth (e.g. malpresentation, shoulder dystocia, cephalo-pelvic disproportion) increase the probability of birth asphyxia. Obviously, these factors do not only increase the risk of adverse pregnancy and birth outcome but are also the risk factors for foetal loss and neonatal death. Additional risk factors for mortality include maternal anaemia, antepartum haemorrhage, maternal epilepsy, maternal drug use, and rhesus immunisation. Risk factors that were indicated by local informants in Kerala and Karnataka during the feasibility study in 1998 (see Chapter 4 and Appendix D) included: (mal)nutrition, maternal hypertensive disorders, maternal anaemia, infections (maternal, antepartum, neonatal), antepartum haemorrhage, small maternal size (height and weight), and antepartum care and care during delivery. However, the informants did not believe that advanced maternal age, smoking, and alcohol consumption play a role in South India.
The present study focuses on several main risk factors. The following risk factors are selected for further analysis:
• hypertensive disorders of pregnancy (pre-existing and pregnancy-related),
• diabetes (diabetes mellitus and gestational diabetes),
• maternal infections during pregnancy,
• smoking and alcohol consumption,
• anaemia,
• advanced maternal age,
• antepartum haemorrhage (including placenta praevia and abruptio placentae), and
• prolonged and obstructed labour (i.e. birth complications).
These fall within the categories of maternal factors, complications of placenta, cord, and membranes, or birth complications. They are believed to have a direct effect on pregnancy/birth outcome and have been presented as such in the operationalisation of the conceptual model in Chapter 3 (see Figure 3.6). Out of the factors, maternal infections and anaemia are assumed to be more relevant to the earlier stages of the epidemiologic transition.
In this, anaemia is regarded as an indicator of a mother’s nutritional status. Conversely, hypertension (pre-existing), diabetes, smoking and alcohol consumption, and advanced maternal age gain importance during later stages of the transition. The relationship of the
remaining risk factors (i.e. antepartum haemorrhage, and prolonged and obstructed labour) to an epidemiologic or health transition may be unclear.
The sections below briefly define, describe, and discuss the selected risk factors. A large part of the information is derived from Cunningham et al. (1993), Treffers et al. (1995)1, Murray and Lopez (1998)2, and Symonds and Symonds (1998). Additional sources were:
Roberts et al. (2003) for hypertensive disorders; Moore (1999), and Colhoun and Chaturvedi (2002) for diabetes; Holzel (1993), Klein and Remington (1995), and Gotoff (1996) for infections; Surgeon General (1981) and Rall Chomitz et al. (1995) for smoking and alcohol consumption; Lao and Pun (1996) and Verster (1996) for anaemia; and Berendes and Forman (1991) and Den Ouden et al. (1997) for advanced maternal age. It should be noted that the figures found on the frequency of the risk factors are generally unclear as to whether they refer to prevalence or incidence. Therefore, the indications and denominators used below have generally been copied from the sources where available.
9.2.1 HYPERTENSIVE DISORDERS OF PREGNANCY
Hypertensive disorders of pregnancy (HDP) is a group of health problems that all refer to high blood pressure during pregnancy. The classification and definition of HDP are complicated due to lack of conformity between publications. An important distinction is between pregnancy-induced hypertension and pre-existing hypertension. Pregnancy-induced hypertension (PIH) occurs after the 20th week of gestation and regresses within several days after birth. The condition is sometimes also referred to as gestational hypertension. When the hypertension is accompanied by proteinuria (protein in urine) and/or oedema, it has progressed into pre-eclampsia. Pre-eclampsia is generally divided into subcategories on the basis of its severity. However, the condition may develop further into eclampsia. Eclampsia is diagnosed as pre-eclampsia accompanied by convulsions. Without prompt treatment, the condition may result in the death of the woman. In contrast to PIH, pre-existing hypertension is unrelated to pregnancy and is already present before the pregnancy and therefore before 20 weeks of gestation. It is sometimes referred to as chronic hypertension or essential hypertension. It is important to note that, in some cases, pre-existing hypertension is not recognised before pregnancy and that true diagnosis may be masked by the tendency of blood pressure to decrease in early pregnancy (Roberts et al. 2003). Pre-existing hypertension may be aggravated by pregnancy and may result in superimposed pre-eclampsia and/or superimposed eclampsia.
The diagnosis of HDP is based on blood pressure measurements, urine testing, and clinical observations of seizures and convulsions. Hypertension is generally diagnosed when systolic blood pressure is greater than or equal to 140 mmHg, or when diastolic pressure is 90 mmHg or above. These values should be measured on at least two occasions several hours apart. Some definitions also include an increase of 30 mmHg or more in the systolic pressure
1 To be more specific, Chapters 10, 11, and 13: by Bennebroek Gravenhorst et al. (1995), by Visser et al. (1995), and by Aarnoudse et al. (1995) respectively.
2 To be more specific, Chapters 4, 6, and 7: by AbouZahr (1998a), by AbouZahr and Guidotti (1998), and by AbouZahr (1998b) respectively.
values or an increase of 15 mmHg or more in the diastolic pressure values. Furthermore, proteinuria is generally defined as the presence of urinary protein in concentrations greater than 0.3 g/L in a 24-hour collection. It should also be noted that oedema is also common in otherwise uncomplicated pregnancies.
According to Symonds and Symonds (1998), 10 to 15% of all pregnancies in the UK are complicated by hypertension, but they believe that only 10% of these pregnancies develop proteinuria. Bennebroek Gravenhorst et al. (1995), in the Netherlands, estimate that only 5%
of pregnant women experience hypertensive disorders during pregnancy. In the USA, PIH is estimated to affect 7 to 10% of all pregnancies (Granger et al. 2001) but, in 2001, only 3.7%
of live births were to mothers with pregnancy-related hypertension (Martin et al. 2002). The incidence of pre-eclampsia is commonly believed to be about 5% (Cunningham et al. 1993) and the incidence of eclampsia has been estimated to be around 0.1% in developed countries (Duley 1991 cited by AbouZahr and Guidotti 1998). In developing countries, the frequency of eclampsia is higher at 0.5% of live births or about 0.47% of pregnancies (WHO 1991 cited by AbouZahr and Guidotti 1998). In developed countries, the impact of chronic hypertension is expected to increase and it has been suggested that in the near future, 5% of women who become pregnant will have pre-existing hypertension (Roberts et al. 2003). However, in 2001, only 0.8% of live births in the USA were reported to be to mothers with chronic hypertension (Martin et al. 2002).
The development of pregnancy-related hypertension, pre-eclampsia, and/or eclampsia is believed to be a result of maternal maladaptation to placental tissue and reduced placental perfusion. Research suggests an immunologic basis and effects of maternal factors. Women with the highest risks of these disorders include young women, older women, nulliparous women, and obese women. The increased risk among older women is believed to be related to an increased prevalence of chronic hypertension at advanced ages. Other risk factors include genetic predisposition, new paternity, diabetes, factors that predispose to cardiovascular disease, and a large amount of placental tissue (e.g. in the cases of twin pregnancy, and triploidy). The steady rise in the frequency of pregnancy-related hypertension during the 1990s in the USA (Martin et al. 2002) has been suggested to be the result of an increase in the number of older mothers and multiple births (Roberts et al. 2003). Since eclampsia is preventable, its occurrence is related to the availability, use, and quality of antenatal care. Pre- eclampsia and eclampsia have been associated with large placental infarcts, foetal growth retardation, foetal death, abruptio placentae, preterm birth, perinatal mortality and morbidity, and maternal mortality. Moreover, the presence of pre-eclampsia or eclampsia is an important reason for opting for the induction of labour. In the USA, 15% of all preterm births are indicated early deliveries because of pre-eclampsia (Roberts et al. 2003).
9.2.2 DIABETES MELLITUS
Diabetes in pregnancy can be the result of diabetes mellitus present before pregnancy, or of gestational diabetes. Gestational diabetes mellitus (GDM) may be induced by pregnancy, or it may be maturity-onset diabetes first recognised during pregnancy. Moreover, gestational diabetes is a strong risk marker for diabetes risk in later life. In some women with GDM the
disorder in glucose tolerance persists postpartum. Diabetes in pregnancy is diagnosed on the basis of glucose tolerance tests (GTT), although there is no international agreement on the diagnostic criteria and the exact test to be performed.
In the Netherlands, the general prevalence of diabetes mellitus ranges between 2.1 and 6.4 per 1,000 women (Bonsel and Van der Maas 1994). Gestational diabetes is estimated to occur in 1 to 2% of pregnancies. In the UK, diabetes affects 0.2 to 0.3% of pregnancies (this low figure probably refers to pre-existing diabetes) while another 1-2% of pregnant women are known to have an impaired glucose tolerance (Symonds and Symonds 1998). In the USA, 3.1% of live births in 2001 were to mothers with diabetes (pre-existing or gestational) (Martin et al. 2002). The prevalence of gestational diabetes in the USA is estimated to be between 1 and 3% (Cunningham et al. 1993). On the basis of figures from various studies that were compiled by Moore (1999), the prevalence of gestational diabetes appears to vary greatly between countries and between ethnic groups. The prevalence ranged from 1.5% among Caucasians in Minnesota (USA), to 3.5% in Pakistan, and as high as 8.3% among Native Americans (Cree).
The following factors have been associated with an increased risk of gestational diabetes: a family history of diabetes; age over 30; GDM in prior pregnancy or a prior heavy (> 4 kg), malformed, or stillborn infant; hypertension; cigarette smoking; and obesity. It is interesting to note that the recent obesity epidemic in low- and middle-income countries, as observed during the nutrition transition, has been associated with an increase in the prevalence of non-insulin-dependent diabetes mellitus (Popkin 1998). Adverse effects associated with diabetic pregnancy include: pre-eclampsia, hydramnios, macrosomia, major congenital anomalies (e.g. cardiovascular anomalies, anomalies of the central nervous system), unexplained late intrauterine death or stillbirth, complications at birth (e.g. shoulder dystocia), caesarean section, and neonatal morbidity. Perinatal mortality in diabetic pregnancies has decreased over the years because of the discovery of insulin in the early 1920s, the improved management of diabetes in pregnancy, and improved obstetric and neonatal care. Nevertheless, perinatal mortality rates are still somewhat higher among diabetics than non-diabetic mothers and this is believed to be largely due to congenital anomalies.
9.2.3 MATERNAL INFECTIONS DURING PREGNANCY
The majority of infections in pregnant women relate to the upper respiratory and gastrointestinal tracts and have no effect on the foetus. Nevertheless, a large number of viruses and bacteria can be passed on by so-called vertical transmission from mother to child during pregnancy or birth. During pregnancy, there are several modes of transmission from mother to foetus: transplacental, ascending from the maternal genital tract, and intrapartum.
Furthermore, invasive techniques for prenatal diagnosis and therapy, such as amniocentesis and intrauterine transfusion, are potential sources of infection. The transplacental route of transmission involves haematogenous spread by invasion of the mother’s bloodstream. As a consequence, placental infection and/or infection of the foetus may occur. The maternal infection could also affect the foetus indirectly, e.g. as a result of fever and metabolic
derangements. Ascending vaginal microflora may result in chorioamnionitis and may harm the foetus even when the membranes are still intact although the risk is more significant after rupture of the membranes, especially when delivery is delayed. In addition, aspiration or ingestion of infected amniotic fluid may result in foetal infection. During the delivery itself, the newborn may be infected while passing through the birth canal. Clearly, the risk of infection does not end once the neonate is born. Postnatally, a newborn may acquire infections from its mother or other persons in its environment.
Organisms and infections that are proven or strongly suspected to affect pregnancy outcome and foetal and neonatal health include: rubella, cytomegalovirus (CMV), herpes simplex, varicella-zoster-virus (chickenpox), parvovirus B19, HIV, hepatitis B, syphilis, Listeria monocytogenes, tuberculosis, Borrelia burgdorferi, group B Streptococcus, toxoplasmosis, and chlamydia. There are variations in the effects that maternal infection during pregnancy can have on the foetus and newborn. Consequences depend on the timing of in-utero exposure, the type of infection, the ‘dose’, and immune status. The adverse outcomes for foetus and neonate include: congenital anomalies (e.g. microcephaly, hydrocephalus, mental retardation), IUGR, perinatal asphyxia, foetal mortality, preterm birth, LBW, neonatal morbidity (e.g. congenital infection, neonatal infection, conjunctivitis, respiratory distress), and neonatal mortality.
The reported incidence of inflammation of the villi or villitis varies from 6 to 26% and evidence of chorioamniotis has been found in 24.4% of delivered placentas (Khong 1993). It is believed that about 4% of women acquire cytomegalovirus (CMV) infection during pregnancy (Symonds and Symonds 1998) and that 0.1 to 2% of live births are infected by cytomegalovirus (Pass 1987; Visser et al. 1995). In the USA, 0.9% of live births in 2001 were to mothers affected by genital herpes (Martin et al. 2002). In the UK, both listeriosis and toxoplasmosis affect 1 in 20,000 births (Symonds and Symonds 1998). For toxoplasmosis, the figures in other countries have been estimated to be higher: 0.1% of pregnancies in the Netherlands (Visser et al. 1995), 0.05% of births in France (Symonds and Symonds 1998), and 0.1% of live births in the USA (Pass 1987). Overall, Gotoff (1996) estimates that about 2% of foetuses are infected in utero, while up to 10% of infants are infected during delivery or in the first month of life.
9.2.4 SMOKING AND ALCOHOL CONSUMPTION
Harmful lifestyle behaviours of the mother, such as smoking and alcohol abuse, also have adverse effects on the health status of the foetus and newborn. In the case of cigarette smoking, not only smoking during pregnancy but also lifetime smoking histories and environmental cigarette smoking seem to have adverse effects. Tobacco chewing, which occurs more frequently in developing countries, also appears to have some adverse effects (Kramer 1987a). The relationship between smoking or alcohol consumption and adverse pregnancy and birth outcomes is complex and likely to be confounded, at least to some extent, by factors such as other negative lifestyle behaviours (e.g. drugs abuse), (mal)nutrition, psychosocial stress, and socioeconomic factors.
Smoking during pregnancy contributes to several adverse outcomes including: reduced weight and length, LBW, IUGR, abruptio placentae, placenta praevia, preterm birth, spontaneous abortion, stillbirth, and perinatal mortality. There appears to be a dose-response relationship between the number of cigarettes smoked and the occurrence of subsequent complications. For example, the percentage of LBW births increases with increasing number of cigarettes smoked during pregnancy. Moreover, quitting smoking at almost any point during pregnancy reduces the risk of LBW since most foetal growth takes place during the third trimester. On average, smoking reduces birth weight by 150 to 320 g and is believed to account for 20 to 30% of LBW births in the USA. In addition, smoking during pregnancy has been linked to 10% of foetal and infant deaths and to up to 14% of preterm births (Rall Chomitz et al. 1995). Explanations that have been suggested in the literature for the adverse effects of smoking include the following: carbon monoxide interfering with oxygen transfer by binding to haemoglobin, nicotine causing the narrowing of blood vessels and thereby reducing uteroplacental blood flow, and reduction in appetite and caloric intake.
Data on the prevalence of smoking just prior to, and during, pregnancy are available for the USA. The information is based on the Pregnancy Risk Assessment Monitoring System (PRAMS), a surveillance project of the Centers for Disease Control and Prevention (CDC), which includes data for 15 states in the USA. In 1998, the prevalence of smoking three months prior to pregnancy ranged from 23.6% in Alabama to 41.4% in West Virginia. The prevalence of smoking during the last three months of pregnancy was lower, and varied between 11.5% (New Mexico) and 27.7% (West Virginia) (Lipscomb et al. 2000). In 1999, the prevalence of smoking during the last three months of pregnancy ranged from 6.2% in Utah to 27.2% in West Virginia (Beck et al. 2002). According to data reported on US birth certificates, 12.3% of women giving birth in 1999 smoked during pregnancy (Mathews 2001).
Smoking during pregnancy has been associated with race/ethnicity, age, marital status, employment status, level of education, and socioeconomic status (Rall Chomitz et al. 1995;
Mathews 2001). It is believed that one-quarter of women who smoke prior to pregnancy quit during pregnancy while an additional one-third of these women reduce their number of cigarettes. However, about one-third of women who quit will take up smoking again before childbirth (Rall Chomitz et al. 1995).
Alcohol is a potential teratogen. The amount of alcohol, the frequency of intake, and the timing of drinking during pregnancy are likely to affect the consequences. Heavy drinking, especially early in pregnancy, may result in foetal alcohol syndrome. The features of this syndrome include: prenatal and postnatal growth retardation, craniofacial anomalies, mental retardation, and congenital heart defects. Heavy drinking is seen as 8 or more ‘units’
of alcohol per day. However, alcohol abuse of 4-6 units per day, and even more than one drink per day, have been associated with foetal growth retardation. The effects of more moderate drinking patterns during pregnancy are not so well established.
Data on alcohol use during pregnancy also come from the USA, from PRAMS (based on 15 states). In 1998, the prevalence of alcohol consumption during the last three months of pregnancy ranged from only 2.3% in West Virginia to 8.3% in Colorado. Use of alcohol during the months prior to pregnancy is much higher: 31.8% (Alabama) to 54.5% (Maine) of women reported using any alcohol in the three months prior to pregnancy (Lipscomb et al.
2000). Self-reported data on prevalence during the first trimester may be unreliable since women are likely to report their use on the time after they had come to know that they were pregnant. Therefore, it has been suggested that alcohol use just before pregnancy may be a more accurate measure (Day et al. 1993 cited by Lipscomb et al. 2000). Alcohol use and pattern of drinking (i.e. any use or heavy drinking) have been associated with race/ethnicity, age, level of education, employment, marital status, and parity (Rall Chomitz et al. 1995).
Overall, alcohol abstinence levels among pregnant women range from 16 to 53%. About 40 to 60% of pregnant women consume one alcoholic drink or less per day while the consumption of more than one drink per day is believed to occur in 3 to 13% of women (Rall Chomitz et al.
1995). The prevalence of foetal alcohol syndrome is 1 to 3 per 1,000 live births (Abel and Sokol 1987 cited by Rall Chomitz et al. 1995).
9.2.5 ANAEMIA
Anaemia refers to low concentrations of haemoglobin (Hb). Haemoglobin is necessary for the transport of oxygen in the blood. Normal haemoglobin levels differ among various groups, such as between women and men, and between pregnant and non-pregnant women. During pregnancy, the concentration of haemoglobin in the blood decreases and is lowest in the second trimester. However, the definitions of anaemia in pregnancy vary, and there is discussion about what is the appropriate haemoglobin level to indicate significant anaemia. A common definition of anaemia in pregnancy, that is applied by the WHO and UNICEF, is a haemoglobin level below 11 g/dl (UNICEF 1991; Lao and Pun 1996; Van den Broek et al.
1999). However, also frequently accepted is a haemoglobin level below 10 g/dl (Cunningham et al. 1993; Symonds and Symonds 1998). The definition by the Centers for Disease Control in the USA is more detailed: less than 11 g/dl in the first and third trimesters and less than 10.5 g/dl in the second trimester.
Anaemia is a quite common disorder in pregnancy. In developed countries, the incidence of anaemia in pregnancy is estimated to be around 5% (Whitfield 1995 cited by Lao and Pun 1996). In a study by Scott and Pritchard (1967 cited by Cunningham et al. 1993) among healthy women in a developed country, the mean Hb level was 11.5 g/dl at mid- pregnancy and 12.3 g/dl in late pregnancy. At mid-pregnancy, 29% had an Hb level of less than 11 g/dl and 4% of less than 10 g/dl. Later in pregnancy, the corresponding figures were only 6 and 1% respectively. More recently, in 2001, 2.5% of live births in the USA were to mothers who had a haemoglobin level of less than 10 g/dl during pregnancy (Martin et al.
2002). In a hospital-based study in Hong Kong, 7.5% of women who delivered during 1990- 1992 were diagnosed with antenatal anaemia, i.e. a lowest haemoglobin level below 10 g/dl at any time before delivery (Lao and Pun 1996). Compared to the developed regions, the frequency of anaemia is much higher in developing countries. For example, in India, the reported prevalence of anaemia (Hb < 11 g/dl) in pregnant women ranges from 40 to 50% in urban areas, and from 50 to as high as 90% in rural areas (UNICEF 1991).
Causes of anaemia include: iron deficiency, other nutrient deficiencies (i.e. folic acid, vitamin B12), loss of blood, chronic infections, malaria, intestinal parasites, and haemoglobinopathies. Haemoglobinopathies, such as sickle cell haemoglobinopathy and
thalassaemia, are genetically determined. Maternal anaemia has been said to lead to increased risks of IUGR, low birth weight, preterm birth, perinatal mortality, and maternal mortality.
However, it is also believed that the adverse effects of anaemia are limited unless the anaemia is severe and/or complicated by other conditions, e.g. pre-eclampsia and abruptio placentae.
Further, a high haemoglobin level of more than 13 g/dl has also been associated with adverse outcomes, such as preterm birth, low birth weight, perinatal mortality, and pre-eclampsia.
9.2.6 ADVANCED MATERNAL AGE
Frequently, an age of 35 years or over is applied to denote women at increased risk of experiencing complications, especially in reference to women who are pregnant for the first time. Other cut-off points used have been 30 years and 40 years of age. However, it is generally agreed that the risks increase gradually with advancing age and that there is no precise cut-off point. Indeed, set against increasing age, many risks show J- or U-shaped curves with the optimum lying between 20 to 24 years.
Advanced maternal age has been associated with many adverse outcomes for maternal or foetal health status. Adverse outcomes that are frequently mentioned in relation to advanced maternal age are: chromosomal anomalies, spontaneous abortion, stillbirth, gestational diabetes, placenta praevia, abruptio placentae, preterm birth, and perinatal mortality. However, for some of these outcomes, risk appears to be primarily increased among primiparous women. Moreover, various factors may confound the observed associations between maternal age and adverse pregnancy and birth outcome. Potential confounders include: parity, education, race, history of reproductive difficulties and complications, smoking and drinking behaviour, and pre-existing health problems.
In the Netherlands, 20.4% of live births in 2000 were to mothers aged 35 years or above, and as many as 62.6% involved mothers ≥ 30 years (CBS 2001). In the USA, maternal age appears to be lower. Vital statistics for the USA show that 35.5% of live births in 1998 occurred in women aged 30 years or above, 12.9% in women ≥ 35 years, and 2.2% in women aged ≥ 40 years (NCHS 2002). The highest maternal age was 54 years (6 live births). Ten years earlier, in 1988, singleton live births in California were distributed as follows: 31.4%
among mothers ≥ 30 years, 9.8% to mothers ≥ 35 years, and 1.5% to mothers ≥ 40 years (NCHS 1995). Here, the highest maternal age recorded was 49 years (13 live births).
9.2.7 ANTEPARTUM HAEMORRHAGE, PLACENTA PRAEVIA, AND ABRUPTIO PLACENTAE
Antepartum haemorrhage (APH) is defined as bleeding from the genital tract before labour but after early pregnancy. There appears to be no agreement on the gestational age limit that distinguishes APH from early pregnancy bleeding or threatened abortion. Suggested limits include the 24th and the 28th week, and have shown a tendency over the years to move to earlier pregnancy following the trend in the limit of viability. About half of the cases of APH are due to either placenta praevia or abruptio placentae, although other causes include uterine rupture and cervical lesions. Antepartum haemorrhage, placenta praevia, and abruptio
placentae are important causes of perinatal morbidity and mortality, and also of maternal morbidity and mortality. In addition, they are associated with preterm birth.
Placenta praevia is when the placenta is implanted in the lower uterine segment. The placenta may cover the outlet of the uterus completely or partially, or may be very near to it.
This means that during delivery the placenta lies in front of the presenting part of the foetus.
Placenta praevia is caused by implantation in the lower part of the uterus. Risk factors are believed to include higher parity, advanced maternal age, smoking, multiple gestation, prior caesarean section, and prior induced abortion. Abruptio placentae (or placental abruption) is the premature separation of the placenta, before the delivery of the foetus. Bleeding as a result of abruptio placentae may be concealed, external (i.e. revealed), or a combination of both. The exact causes of abruptio placentae remain unknown although it has been associated with pre- existing and pregnancy-induced hypertension, pre-eclampsia, cigarette smoking, advanced maternal age, high parity, IUGR, and dietary deficiencies. In some cases, maternal trauma and trauma to the uterus play a role.
Nijenhuis (1995) estimated that vaginal bleeding after 16-20 weeks of gestation occurs in 2 to 5% of all pregnancies. Similar frequency figures were found in the UK and USA about 30 years ago (AbouZahr 1998a). In 2001, only 0.6% of live births in the USA were to mothers who experienced uterine bleeding during the second or third trimester, prior to the onset of labour (Martin et al. 2002). Placenta praevia is believed to complicate about 0.2 to 0.5% of deliveries (Cunningham et al. 1993; Khong 1993; Bennebroek Gravenhorst et al. 1995;
AbouZahr 1998a). Symonds and Symonds (1998) present a higher estimation of approximately 1% of all pregnancies. Fraser and Watson (1989 cited by AbouZahr 1998a) combined data from various developed countries and estimated the overall incidence of placenta praevia to be only 0.55% of pregnancies, ranging from 0.29 to 1.24%. The reported frequency of abruptio placentae varies but is on average about 0.7% of deliveries according to Cunningham et al. (1993). In the UK, placental abruption has been found to occur in approximately 1% of pregnancies (Chamberlain et al. 1978 cited by AbouZahr 1998a). In the USA, 0.5% of live births in 2001 were affected by abruptio placentae and 0.3% by placenta praevia (Martin et al. 2002).
9.2.8 PROLONGED AND OBSTRUCTED LABOUR
Prolonged and obstructed labour refer to an abnormally slow progress of labour. Other terms that are used to denote difficult delivery include dystocia and failure to progress. However, what is considered to be ‘abnormal’ varies, and there is a difference in this respect between nulliparas and multiparas. Causes of prolonged and/or obstructed labour include: cephalo- pelvic disproportion (CPD), malpresentation or malposition of the foetus, foetal abnormalities, insufficient expulsive forces, and abnormalities of the reproductive tract.
Combinations of the above are also possible.
Cephalo-pelvic disproportion is a disproportion between the size of the maternal pelvis and the size of the foetal head. The size of the head may be too large, the pelvis too narrow, or a combination of both. A high risk of CPD is found in very young women and in women of small stature. CPD is commonly the result of physical immaturity or stunting in the mother or
distortion of her pelvis through disease or malnutrition (AbouZahr 1998b, p.244). In those regions of the world where malnutrition, disease, and early marriage are common, CPD is said to be the main contributor to obstructed labour. Abnormalities in presentation or position of the foetus include: breech presentation, face presentation, brow presentation, transverse lie, shoulder presentation, compound presentation, and cord presentation. Factors associated with malpresentation or malposition include placenta praevia, abnormalities of the uterus, polyhydramnios, oligohydramnios, anomalies of the foetal head, high parity, weakening of the uterus, multiple pregnancy, and preterm delivery.
Prolonged or obstructed labour normally requires some type of intervention since it may result in serious perinatal and maternal morbidity as well as the death of both mother and child. Adverse outcomes for the newborn include asphyxia, meconium aspiration, neonatal infection and pneumonia, intracranial haemorrhage, brain damage, and birth injuries or trauma. Interventions include manipulative techniques, instrumental delivery (e.g. forceps delivery, vacuum extraction) and operative delivery (i.e. caesarean section).
Prolonged labour is a relatively common complication. AbouZahr (1998b) believes that the incidence of dystocia varies between 3 and 6% with the lower figure occurring in more developed regions while the higher figure is believed to occur in Africa, the Middle East, South West Asia, and India. Global incidence is estimated at 5.1% of all live births, where live births are used as a proxy for deliveries (AbouZahr 1998b). In the USA, only 1.6%
of live births in 2001 were complicated by CPD (Martin et al. 2002). In Dublin (Ireland), Malone et al. (1996) studied nulliparas who delivered during 1990-1994 with spontaneous onset of labour, singleton gestation, cephalic presentation, and labour after 37 weeks. Overall, 1.6% of nulliparas had prolonged labour, defined as lasting more than 12 hours from the time of admission to the delivery ward until delivery of the infant. Prolonged labour was due to inefficient uterine action in 65% of cases and to CPD in 11% of cases.
Breech presentation occurs in 2 to 4% of all term deliveries in developed countries (Cunningham et al. 1993; Aarnoudse et al. 1995; Moore 1986; Symonds and Symonds 1998).
Earlier in gestation, breech presentation is more common. At a gestational age of 30 weeks, about 25% of foetuses lie in a breech presentation (Aarnoudse et al. 1995). Other abnormal presentations or positions occur much less frequently. Transverse lie occurs in only about 0.2 to 0.3% of deliveries, face presentation in about 0.2%, and brow presentation and compound presentation are rare (Cunningham et al. 1993; Symonds and Symonds 1998). Overall, of the live births in 2001 in the USA, 3.8% were complicated by breech or other malpresentation (Martin et al. 2002).
Finally, the proportion of instrumental and operative deliveries could be regarded as an indicator of the frequency of prolonged and obstructed labour. However, the proportion of such deliveries and the reasons to perform them vary widely between countries and cultures (AbouZahr 1998b). Recent figures indicate a global ‘epidemic’ of caesarean sections, with proportions from around 25% up to as high as 80% among hospital births in some countries (Dosa 2001; Brundtland 2002). The World Health Organization (1985 cited by Dumont et al.
2001) has estimated the optimum frequency of caesarean sections to be 5-15%.
9.3 Kerala and S.A.T. Hospital
This section describes the situation in Kerala with regard to the selected risk factors. Data on 1,001 births come from a hospital survey that was conducted in Sri Avittom Thirunal (SAT) Hospital, in Thiruvananthapuram. A description of the survey and the resultant data set was provided in Chapter 4. In chapter 8, some of the data on pregnancy and birth outcome, stillbirth, and neonatal death were already analysed. For a brief introduction to Kerala, please refer to Section 2.4.2.
The relative importance of the risk factors is assessed at both the individual level and the population level. The next section, Section 9.3.1, presents figures on frequency (cf.
prevalence) of the risk factors in the population. The following two sections, Sections 9.3.2 and 9.3.3, calculate the relative risks and etiologic fractions of adverse outcomes. Lastly, Section 9.3.4 discusses the observed associations and interrelationships between the risk factors themselves.
9.3.1 FREQUENCY OF RISK FACTORS
The current section presents figures on frequency (cf. prevalence) of the selected risk factors in the population. Results of the hospital survey are compared to, and complemented by, data for Kerala from other, secondary sources where if available. It is important to remember that SAT Hospital is a referral hospital – 20% of women in the survey had been referred – since this may increase the observed frequency of pregnancy complications in the population. Table 9.1 presents observed frequency of the risk factors in the population of the SATH survey.
Hypertensive disorders of pregnancy
Table 9.1 shows that the percentage of women in the survey with pre-existing hypertension was low at only 0.2%. However, the number of pregnancies affected by hypertension and hypertensive disorders increased during pregnancy. Pregnancy-related hypertensive disorders (including pre-existing hypertension aggravated by pregnancy) were diagnosed in 15.2% of pregnant women. The majority of these cases were pregnancy-induced hypertension (13.4%
of pregnant women), some were pre-eclampsia and/or impending eclampsia, whereas only one case developed into eclampsia. The results are based on the diagnoses in the hospital records. However, our pro forma contained the category ‘impending eclampsia’ but, regrettably, not the category ‘pre-eclampsia’. Therefore, pre-eclampsia had to be assessed retrospectively as impending eclampsia or pregnancy-induced hypertension (PIH) in combination with albumin in the most recent urine test (1.8% of pregnant women). This may underestimate the actual number of cases of pre-eclampsia since urine was usually tested only once in pregnancy and the gestational week of this test was not known in most cases. Overall, 15.3% of pregnant women in the sample were assessed as having pre-existing hypertension and/or pregnancy-related hypertensive disorders.
In comparison to the figures in Section 9.2.1 for countries in the EME region, the prevalence of pre-existing hypertension is somewhat low (0.2% vs. 0.8%) while the frequencies of PIH and hypertensive disorders in general in pregnancy appear to be comparatively high (13-15% vs. 4-15%). The frequency of eclampsia is similar (0.1%).
Table 9.1: Frequency of selected risk factors in pregnant* women or births*, SAT Hospital, Kerala, October 2000
Risk factor Frequency (%)
in pregnant* women / births*
Hypertensive disorders
pre-existing hypertension 0.2
pregnancy-related hypertensive disorders (all) 15.2
pre-eclampsia 1.8
eclampsia 0.1
all hypertensive disorders 15.3
Diabetes
pre-existing diabetes 0.7
gestational diabetes 3.1
all diabetes 3.8
Infections
positive HBsAg-test 2.9
maternal fever during labour 0.9
other infections in hospital records 0.9
Anaemia
anaemia in hospital records 0.6
Hb < 11 g/dl 38.6
Hb < 10 g/dl 7.7
Advanced maternal age
age at delivery >= 30 years 14.3
age at delivery >= 35 years 2.1
APH and placental problems
APH 1.1
placenta praevia 1.5
abruptio placentae 0.5
Prolonged and obstructed labour
mal-presentation or -position 5.0
instrumental or operative delivery 28.6
obstructed labour 12.5
failure of secondary powers 1.8
Note: *>= 28 weeks of gestation
Diabetes mellitus
The prevalence of pre-existing diabetes in the SATH survey was 0.7% (see Table 9.1).
Gestational diabetes was diagnosed in as many as 3.1% of pregnant women. The assessment was based on the hospital records. In total, 3.8% of pregnant women in the survey were affected by either pre-existing diabetes or gestational diabetes. In comparison to the figures for the EME region (see Section 9.2.2), the proportions from the SATH survey are similar but possibly slightly higher (4% vs. 3%). However, the prevalence of gestational diabetes in Pakistan is similar at 3.5% (Rizvi et al. 1992 cited by Moore 1999).
Maternal infections during pregnancy
With regard to maternal infections, local informants suspect that in South India many infections in pregnancy remain unregistered and are not mentioned in case records.
Nevertheless, in SAT Hospital two tests for infectious diseases are routinely performed: an HBsAg test for the hepatitis B virus, and a Venereal Disease Research Laboratory (VDRL) test for syphilis. In the SATH survey, the results of an HBsAg test were known for 78.7% of women. Of these women, 2.9% tested positive. Results of a VDRL test were known for 88.9% of pregnant women, and none of them had tested positive. In addition to the above results, 0.9% of deliveries were complicated by maternal fever, which suggests the presence of infection. Lastly, the case records of 9 women (0.9%) contained reference to an infectious disease in pregnancy. These infections were respiratory infection (4 cases), herpes genitalis (2 cases), acute bronchitis (1 case), diarrhoea (1 case), and urinary tract infection (1 case). Table 9.1 presents the results for the various infectious diseases.
Smoking and alcohol consumption
Information on maternal smoking and alcohol consumption was not included in the records of SAT Hospital. However, these factors are believed to be of very limited importance in Kerala and India. Local informants in the feasibility study in 1998 stated that cigarette smoking, the intake of alcohol, and even tobacco chewing are rare among Kerala women of reproductive age (Den Draak 2000). This seems to be supported by data from NFHS-2 (IIPS and ORC Macro 2001b). NFHS-2, in 1998-1999, included several questions on so-called lifestyle indicators. In Kerala, none of the women aged 15-29, and only 0.2% of women aged 30-39 reported drinking alcohol. As for smoking, none of the women aged 20-29 years were currently smoking or had ever smoked. In the age group 15-19, only 0.1% were currently smoking and the corresponding figure was 0.3% in the 30-39 age group. Chewing of paan masala or tobacco occurred more often but was still relatively infrequent: only 0.6% of women aged 20-24, 0.9% of those aged between 25-29, and 5.0% of those aged 30-39.
Teenage girls aged 15-19 years did not report any chewing of paan masala or tobacco (IIPS and ORC Macro 2001b).
Anaemia
Anaemia is believed to be common in developing countries. In the SATH survey, iron supplements were known to have been prescribed to 98% of women. In addition, haemoglobin (Hb) level was measured at least once in the vast majority of cases. Results of Hb measurements were available for 983 women (98.2%) in the survey. On the basis of the lowest Hb level known to have been measured, 379 pregnant women or 38.6% had anaemia as defined by an Hb level of less than 11 g/dl. In 7.7% (76 cases) the Hb level was below 10 g/dl, in 1.7% (17 cases) below 9 g/dl, and only one woman had severe anaemia of below 7 g/dl. The average Hb level was 11.1 g/dl (± 1.0 g/dl). The Hb measurements were taken at various gestational stages, ranging from gestational week 7 to gestational week 41. Despite the large number of anaemic women using WHO guidelines, anaemia was only diagnosed in 0.6% (6 cases) of pregnant women. It is said that, in practice, SAT Hospital uses a definition based on an Hb level of less than 9 g/dl. However, if this was true, then the prevalence should
have been higher at 1.7% (see above). Table 9.1 presents the results on anaemia in the SATH survey.
Other studies in Kerala have also presented findings on anaemia in pregnancy.
Gopalakrishnan and Syamalan (1994) found that ten years before the present study, in 1990, 8.5% of women who delivered their baby in SAT Hospital had an Hb level of less than 10 g/dl. This is slightly higher than the current findings. Vijayakumar et al. (1997) studied Hb level in women receiving antenatal care in primary health centres and government hospitals in and around Trivandrum (Thiruvananthapuram). The majority of these women were receiving iron-folate supplements. About 30% of pregnant women – 29.0% among those who received iron supplements and 33.8% among those who did not – were anaemic as defined by an Hb level below 11 g/dl. 11.2% of these had an Hb level below 10 g/dl (Vijayakumar et al. 1997).
More recently, NFHS-2 measured haemoglobin levels in women in 1998-1999. Among the 140 pregnant women in Kerala who were tested, 20.3% had an Hb level below 11 g/dl and 8.7% of less than 10 g/dl (IIPS and ORC Macro 2001b). What causes the differences in the prevalence of low Hb levels between NFHS-2 and the SATH survey, in particular with regard to Hb < 11 g/dl, is not known. However, the results from NFHS-2 were based on a relatively small number of 140 pregnant women at various gestational ages.
Compared to the EME region (see Section 9.2.5), the prevalence of anaemia or low Hb level in pregnancy seems to be somewhat higher in Kerala, though not as high as figures for all-India (40-90%).
Advanced maternal age
In the SATH survey, the mean age of women at the time of delivery was 24.7 years (± 4.0 years). Ages ranged from a minimum of 18 years to a maximum of 38 years. In 3 cases, the maternal age was missing or unknown. The vast majority of women (78.5%) in the survey were between 20 and 29 years of age. As Table 9.1 shows, 14.3% (143 cases) of women were 30 years or older but only 2.1% (21 cases) of births occurred to women ≥ 35 years. It is interesting to note that for three-quarters of the women ≥ 30 years, it was only their first or second birth (75.5%). The other births in this group were third or fourth order births (21.7%
and 2.8% respectively).
The prevalence of advanced maternal age in pregnancies and births thus appears to be lower in Kerala than in the EME region (see Section 9.2.6). Other data from SAT Hospital support this finding. Also in 1990, 14.3% of women who delivered in SAT Hospital were 30 years or older, and 2.5% were ≥ 35 years (Gopalakrishnan and Syamalan 1994).
Antepartum haemorrhage, placenta praevia, and abruptio placentae
In SAT Hospital, vaginal bleeding after about 16 weeks of gestation is regarded as antepartum haemorrhage (APH). On the basis of the hospital records, the SATH survey contained 11 cases (1.1%) of APH. The causes were placenta praevia (5 cases), abruptio placentae (4 cases), velamentous insertion of the cord (1 case), and in the other case the cause of APH was unknown. In total, 1.5% of pregnancies (15 cases) were complicated by placenta praevia and 0.5% (5 cases) were affected by abruptio placentae (see Table 9.1).
In comparison to the figures presented in Section 9.2.7 for the EME region, the incidences of APH and abruptio placentae in SAT Hospital are similar although perhaps slightly lower (cf. 0.6-5.0% and 0.5-1.0% respectively), while the occurrence of placenta praevia may be slightly higher (cf. 0.3-1.2%). Figures from 1990 seem to agree with the findings of the current survey. In 1990, APH was seen in 1.0% of cases and placenta praevia complicated 1.2% of deliveries (Gopalakrishnan and Syamalan 1994). However, it should be remembered that these figures all pertain to a referral hospital.
Prolonged and obstructed labour
In the SATH survey, lie and/or presentation at birth differed from normal delivery, i.e.
longitudinal lie and vertex presentation, in 5.0% (50 cases) of births (see Table 9.1). The most common type of malpresentation was breech (40 cases) while other complications were transverse lie (4 cases), face presentation (3 cases), brow presentation (1 case), compound presentation (head and hand) (1 case), and oblique lie (1 case). Breech occurred in 3.7% of term births and in 6.6% of preterm births.
In total, 28.6% (286 cases) of births involved instrumental or operative delivery. The vast majority of these were by caesarean section (258 cases), while only 27 were vacuum extractions and one by forceps delivery. The indications for instrumental or operative delivery included frequently obstructed labour by cephalo-pelvic disproportion (CPD) or foeto-pelvic disproportion (FPD), as in 121 cases or 42.3% of instrumental or operative deliveries in the survey. Other indications for operative delivery were cervical dystocia (1 case) and deep transverse arrest (1 case). Among vaginal deliveries, two cases were complicated by shoulder dystocia. Combined, these figures imply that as many as 12.5% of all births (125 cases) were complicated by obstruction or potential obstruction of labour. The indications for instrumental or operative delivery further included failure of secondary powers (18 cases) and a protracted active phase (2 cases of CPD and 4 unspecified cases). Table 9.1 presents the overall results with regard to these labour complications.
In comparison, in 1990, 5.2% of births in SAT Hospital were breech (Gopalakrishnan and Syamalan 1994). Overall, malpresentation or malposition occurred in 6.0% of births in 1990. Prolonged labour was recorded in 0.8% of births and obstructed labour in only 0.5%.
However, 4.8% of births were reported as being complicated by CPD. The percentage of caesarean sections in SAT Hospital was lower in 1990, at 21.4% (Gopalakrishnan and Syamalan 1994).
Although SAT Hospital is a referral hospital, and the proportion of caesarean sections in our SATH survey is high (i.e. 25.8% of births), NFHS-2 suggests even higher figures for Kerala as a whole. During the three years preceding NFHS-2, 29.3% of all births, 27.4% of births in rural areas, and as many as 37.5% in urban areas, were by caesarean section (IIPS and ORC Macro 2001b).
9.3.2 RELATIVE RISK OF ADVERSE OUTCOME
Below, data from the SATH survey are analysed to see whether the selected risk factors show an elevated risk of adverse outcome. The relative risk (RR) and its 95% confidence interval were calculated using SPSS. The RR compares the incidence of the adverse outcome in
affected persons to the incidence among those who are not affected by the risk factor. The accompanying confidence interval indicates the range for which there is a 95% probability that the value of the RR falls within that range. When the confidence interval includes the value of 1.0, the null hypothesis, which poses that the incidence of adverse outcome among those affected equals the incidence among non-affected persons (i.e. RR = 1), should not be rejected (see also Chapter 4). The confidence interval is to an extent affected by a small sample size.
Additional analysis looking for associations was on the basis of Pearson’s Chi-square and its p-value (see also Chapter 4). However, the generally small numbers in the SATH survey require caution and limit the formulation of strong statements.
It should be noted that the objective of the analysis was not to test whether the results could be considered representative for all of Thiruvananthapuram or Kerala. The analysis merely explores the possible existence of associations between the risk factors and the adverse outcomes within the population of the SATH survey. The resultant figures are not controlled for confounding factors. A discussion of the associations that are observed between the risk factors themselves is provided in Section 9.3.4.
Adverse pregnancy and birth outcome
Table 9.2 presents the relative risks of some of the adverse intermediate outcomes for maternal health problems. Almost all the relative risks in the table indicate an elevated risk as compared to non-affected foetuses. Only for low Hb level (< 10 g/dl), are the relative risks generally below one. In addition, the relative risks of SGA are less than one for certain risk factors. For example, the relative risk of SGA for diabetes (pre-existing and gestational diabetes) is 0.3, which means that neonates from diabetic mothers are less likely to be small- for-gestational-age. The null hypothesis, however, could not be rejected (i.e. 95% CI includes 1.0), which is probably to a large extent due to the small number of cases in the study.
However, it is generally acknowledged that children born to diabetic mothers have higher birth weights and increased risks of macrosomia (see Section 9.2.2).
Overall, the majority of relative risks in Table 9.2 include the value of 1.0 in their associated confidence intervals. However, for hypertensive disorders, maternal age ≥ 30 years or ≥ 35 years, antepartum haemorrhage (APH), and placental problems (placenta praevia and abruptio placentae), the risk of one or more adverse intermediate outcomes can be assumed to be elevated as compared with the incidence in non-affected infants. Table 9.2 shows that children born to mothers with hypertensive disorders (mostly pregnancy-related) run increased risks of: birth weight < 2,500 g (RR = 2.1, CI 1.6-2.7) or < 2,000 g (RR = 5.1, CI 3.0-8.9), preterm birth (RR = 3.9, CI 2.4-6.3), and SGA (RR = 1.8, CI 1.1-3.0). Further associations were found with birth weight < 1,500 g (RR = 10.1, CI 3.4-29.6), and very preterm birth < 32 weeks (RR = 7.2, CI 3.2-17.5). The associations are probably to a large extent explained by labour induction. Indeed, in the SATH survey, the presence of a hypertensive disorder is positively associated with preterm labour induction (p < 0.01). Of the 25 preterm births among women with hypertensive disorders, 13 (52.0%) were induced.
Additional analysis reveals that relative risks are generally higher for pre-existing
Table 9.2: Relative risks of adverse pregnancy/birth outcomes, by risk factor, SAT Hospital, Kerala, October 2000 LBWLBWPretermSGA, weightSGA, weightSGA, weightCongenital < 2.5 kg< 2.0 kg< 37 wksMathai 5thMathai 10thYudkin 3rdanomalies RR (95% CI)RR (95% CI)RR (95% CI)RR (95% CI)RR (95% CI)RR (95% CI)RR (95% CI) Hypertensive disorder2.15.13.91.81.4*1.5*1.8* (1.6-2.7)(3.0-8.9)(2.4-6.3)(1.1-3.0)(1.0-2.0)(0.9-2.3)(0.7-5.0) Diabetes1.1*1.8*2.2*0.3*0.3*0.3*2.8* (0.6-2.1)(0.6-5.4)(0.9-5.2)(0.0-2.2)(0.1-1.2)(0.0-1.9)(0.7-11.7) HBsAg - positive1.0*NCNC1.6*1.4*1.3*2.1* (0.4-2.5)--(0.6-4.8)(0.7-3.2)(0.5-3.9)(0.3-15.0) Other infections during pregnancy1.3*2.7*NC1.7*1.6*1.2*NC (0.4-4.4)(0.4-17.6)-(0.3-11.3)(0.5-5.2)(0.2-7.5)- Anaemia (in hospital records) 1.7*NCNCNC1.9*3.2*NC (0.6-5.5)---(0.6-5.8)(1.0-10.2)- Hb < 10 g/dl0.8*0.6*0.7*0.6*0.6*0.9*1.3* (0.4-1.3)(0.2-2.5)(0.2-2.2)(0.2-1.7)(0.3-1.2)(0.4-1.9)(0.3-5.6) Maternal age >= 30 years1.0*2.41.8*0.40.50.5*1.1* (0.7-1.4)(1.3-4.5)(1.0-3.2)(0.2-0.9)(0.3-0.9)(0.2-1.0)(0.3-3.6) Maternal age >= 35 years1.3*3.33.30.6*0.6*0.5*5.2 (0.6-2.8)(1.1-9.8)(1.3-8.1)(0.1-4.2)(0.2-2.3)(0.1-3.6)(1.3-20.9) Antepartum haemorrhage2.92.0*9.61.2*1.1*1.0*NC (1.7-5.1)(0.3-13.2)(5.3-17.4)(0.2-8.0)(0.3-3.9)(0.1-6.2)- Placental problems**1.9*2.2*4.50.6*0.9*0.5*NC (1.0-3.4)(0.6-8.5)(2.0-10.0)(0.1-4.4)(0.3-2.6)(0.1-3.5)- Note: NC - No cases in SATH survey. *95% confidence interval includes the value of 1.0; **Placenta praevia and abruptio placentae.
hypertension (N.B. only two cases) than for pregnancy-related hypertensive disorders. For example, the relative risk of very preterm birth is 25.3 (CI 5.9-108.5) for pre-existing hypertension but ‘only’ 7.6 (CI 3.3-17.7) for pregnancy-related hypertensive disorders. The RR of birth weight < 1,500 is as high as 40.8 (CI 9.2-182.2) among pregnant women with pre- existing hypertension.
In addition to the associations above, analysis shows that the highest relative risks of SGA are found in women with pre-eclampsia or eclampsia. For such women, the relative risk of SGA is 3.1 (CI 1.2-8.3) when defined by the 5th percentile of Mathai et al. (1996; Mathai, personal communication in 2000) and 2.3 (CI 1.2-4.6) when defined by their 10th percentile.
This is in line with the belief that foetal growth retardation is a result of problems in placental blood circulation and inadequate maternal-foetal supply of oxygen (see Chapter 3). Women affected by pre-eclampsia or eclampsia are also at increased risk of giving birth to an anomalous child (RR = 5.7, CI 1.4-22.9). However, this observed relationship should be seriously questioned. The two cases of congenital anomalies among eclamptic mothers showed indications of congenital heart disease (CHD). One of the babies was term and suspected of having CHD. The other was only 31 weeks and affected by patent ductus arteriosus (PDA), which is regarded as being physiological among very preterm infants (see Section 8.3.1). This suggests that the observed association between pre-eclampsia/eclampsia and ‘congenital anomalies’ is probably an association between pre-eclampsia/eclampsia and very preterm birth (see also Chapter 8). Indeed, after the exclusion of PDA, the RR was no longer statistically significant.
A maternal age ≥ 30 years or ≥ 35 years increases the risk of a birth weight < 2,000 g (RR = 2.4, CI 1.3-4.5, and RR = 3.3, CI 1.1-9.8, respectively). In addition, children born to mothers aged 35 or over are more likely to be anomalous (RR = 5.2, CI 1.3-20.9), preterm (RR = 3.3, CI 1.3-8.1), and also very preterm (RR = 4.8, CI 1.2-19.3), than children born to younger mothers. Further study of the SATH data shows that the congenital anomalies found among children born to women ≥ 35 years were: one case of multiple anomalies with suspected chromosomal origin and one case of PDA. Again, this suggests that the observed association with ‘congenital anomalies’ is probably explained by an association with very preterm birth. Indeed, after the exclusion of PDA, the RR was no longer statistically significant. Further, it is interesting to note that advanced maternal age (≥ 30 years) is associated with a reduced risk of SGA (see Table 9.2). This is probably explained by the observation that the older women in the survey are more likely to have diabetes (see Section 9.3.4).
Antepartum haemorrhage (APH), increases the risks of LBW (RR = 2.9, CI 1.7-5.1) and preterm birth (RR = 9.6, CI 5.3-17.4). However, no correlation was found with very preterm birth (< 32 weeks) and birth weights < 2,000 g or < 1,500 g. The risk of preterm birth is also elevated for cases with placental problems (RR = 4.5, CI 2.0-10.0). However, placenta praevia appears to be far less important than abruptio placentae. For abruptio placentae, the relative risks of LBW (RR = 5.3, CI 4.7-6.1), preterm birth (RR = 13.7, CI 8.2-22.7), birth weight < 1,500 g (RR = 15.3, CI 2.4-95.6), and very preterm birth (RR = 9.7, CI 1.6-59.3) are increased.
Table 9.3: Relative risks of birth asphyxia and low Apgar score, by risk factor, SAT Hospital, Kerala, October 2000
Birth Apgar Apgar
asphyxia 1 min. =< 3 5 min. =< 6 RR (95% CI) RR (95% CI) RR (95% CI)
Hypertensive disorder 1.0* 3.3* 5.5
(0.3-3.5) (1.0-11.2) (1.6-18.9)
Diabetes NC NC NC
- - -
HBsAg - positive 5.1 4.1* 11.1
(1.2-21.2) (0.5-31.2) (2.4-52.0)
Other infections during pregnancy NC NC NC
- - -
Maternal fever during labour 12.9 NC NC
(3.5-47.8) - -
Anaemia (in hospital records) NC NC NC
- - -
Hb < 10 g/dl 1.4* 1.3* NC
(0.3-5.9) (0.2-10.2) -
Maternal age >= 30 years 2.8 5.0 6.0
(1.1-7.2) (1.6-16.3) (1.8-20.4)
Maternal age >= 35 years 5.4 10.6 1.6
(1.3-22.1) (2.5-46.1) (2.6-51.5)
Antepartum haemorrhage NC 10.8 NC
- (1.5-75.5) -
Placental problems** NC 13.4 NC
- (3.1-57.2) -
Mal-presentation or -position NC NC NC
- - -
Obstructed labour 0.4* NC NC
(0.1-2.9) - -
Note: NC - No cases in SATH survey. *95% confidence interval includes the value of 1.0;
**Placenta praevia and abruptio placentae.
In addition to the above, Table 9.2 also indicates that diabetes (all types), HBsAg- positive (i.e. hepatitis B virus), other infections during pregnancy, anaemia (as based on hospital records), and Hb level < 10 g/dl could not be related in the SATH survey to significantly increased, or reduced, risks of adverse intermediate outcomes. This was also observed for gestational diabetes, Hb level < 11 g/dl, and Hb level < 9 g/dl. Nevertheless, pre- existing diabetes was found to be associated with elevated risks of birth weight < 2,000 g (RR
= 6.5, CI 1.9-21.7) and < 1,500 g (RR = 11.6, CI 1.7-77.6), despite its general relationship with macrosomia. In addition, pre-existing diabetes can be assumed to result in increased risks of preterm birth (RR = 7.2, CI 3.0-17.6), very preterm birth (RR = 7.2, CI 1.11-46.6), and congenital anomalies (RR = 7.8, CI 1.2-50.5; after excluding PDA: RR = 8.2, CI 1.3-53.6).
