Cover Image

HARDBACK
$69.95



View/Hide Left Panel

SECTION I
MEASUREMENT



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention SECTION I MEASUREMENT

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention This page intentionally left blank.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention 2 Measurement of Fetal and Infant Maturity ABSTRACT The heterogeneity of the population born preterm is striking. Although pregnancy duration, fetal growth, and fetal or infant physical and neurological maturity are all interrelated and associated with neonatal mortality and morbidity, they are conceptually distinct entities that are differentially influenced by internal and external conditions. Progress in understanding the etiologies and mechanisms of preterm birth and its consequences requires the use of precise definitions, recognition of the limitations of the measures used, and an understanding of the relationships among them. This field requires a better classification of preterm infants into subgroups on the basis of pathogenic pathways, placental findings, genomic markers, and environmental exposures, as well as the recognition that any given individual infant has unique combinations of risk factors and exposures. Outcomes should be reported by gestational age categories, but birth weight for gestational age is an important indicator of the adequacy of fetal growth. Research on methods of quantifying fetal and infant maturity should be encouraged. Pre- and postnatal markers of organ system maturity and predictors of morbidity and functional outcomes that are more effective than birth weight or gestational age should be identified and developed.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Precise definitions of preterm birth are essential for comparing and interpreting the various studies that address the complex problems of preterm birth. These include: (1) studies of the etiologies and mechanisms of preterm birth, (2) trials of the safety and efficacy of strategies for prevention of preterm birth, (3) health and neurodevelopmental outcome studies of preterm infants, (4) trials of the safety and efficacy of medication and treatment strategies for preterm infants, and (5) regional and international comparisons of preterm birth rates. The concept of prematurity involves biological immaturity for extrauterine life. Maturation is the process of achieving full development or growth. The embryo and fetus matures in utero until organ systems are capable of supporting extrauterine life. Although full-term newborns (neonate) have basic needs (warmth, milk), they are generally capable of sustained breathing, crying when hungry, sucking from a nipple, digesting milk, and complex physiological functions, including gas exchange, blood pressure control, glucose metabolism, and regulation of body fluids. Infants born preterm have immature organ systems that often need additional support to survive. Neonatal intensive care has developed to attend to those needs. Degree of maturity, therefore, is the major determinant of mortality and morbidity (the short- and long-term complications) of preterm birth. Born too soon, preterm infants are more vulnerable to organ injury, death, chronic illness, and neurodevelopmental disability than fullterm newborns (see Chapters 10 and 11). Because there are not good direct measures of degree of maturity, gestational age denotes duration of the pregnancy and is used as a proxy measure of degree of maturity. Prematurity is not a defined disease or syndrome, and there is no one specific cause or fixed set of outcomes. Moreover, proximal antecedents of prematurity, such as preterm labor or preterm rupture of membranes, may be the cumulative effect of many environmental and genetic factors. Indeed, delivery before term may be required because of threats to the health of mother or fetus from complications of pregnancy. Some causes originate at or even before conception. In addition, as described in Chapter 11, the outcomes for preterm infants are influenced by factors that lead to preterm birth, organ immaturity, neonatal management, and the postnatal environment. Preterm birth is therefore a common, complex condition that results from multiple interactions between the maternal and the fetal genomes and conditions in the intrauterine environment, the mother’s body, and her external environment. While immaturity is the primary characteristic of preterm infants, degree of immaturity varies, even when controlling for duration of pregnancy. As with older children, there is a biologic continuum, and similar gestational ages and fetal sizes may not indicate similar levels of maturity. Just as a 10-year-old may be tall or short and mature or immature for his or her

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention age, an infant born 4 months early may be large or small and more or less mature for an infant born at 24 weeks gestation. This biologic individual variation in size and maturity is the result of different genotypes and different intrauterine and extrauterine environments and experiences. Thus, the complexity of preterm birth and its causes and complications makes it impossible at this time to predict the outcome for an individual preterm infant at delivery with any degree of certainty. Careful attention to definitions that distinguish between different methods of determining gestational age and recognition of the limitations of measurement methods are necessary to achieve an understanding of the complexities of preterm birth. In lieu of functional measures of fetal or infant maturity, accurate measures of gestational age are essential for clinical care as well as research on the causes, mechanisms, and outcomes of preterm birth. This chapter is devoted to clarifying definitions, describing methods of determining gestational age and their limitations, and demonstrating the implications of the use of precise definitions of the terms used. The reader is referred to Appendix B for further discussions of the definitions of preterm birth and the measurement of gestational age. DEFINITION OF PRETERM BIRTH The World Health Organization (WHO) has defined preterm birth as delivery before 37 completed weeks of gestation. By convention, gestational age is reported in terms of completed weeks (i.e., one never rounds gestational age up, so 36 weeks and 6 days of gestation is 36 weeks and not 37 weeks of gestation). This definition makes the distinction between being born early and being born too small. Determining when natural conception takes place is difficult (see below), so birth weight (not gestational age) was initially used as a proxy measure for maturity. Although some infants are both too small and born too early, small infants can be either fullterm or preterm (Figure 2-1). Although it has long been recognized that pregnancy lasts 9 months and that infants who were born before 9 months gestation or who were born small were at risk for death or disability, it was not until the end of the 19th century that systematic attention to the care of preterm infants began. Early efforts at defining prematurity relied on birth weight, with a birth weight less than 2,300 or 2,500 grams considered low birth weight (LBW). LBW was first used as a standard by Nikolaus T. Miller, physician-in-chief of the Moscow Foundling Hospital, and also in 1888 by Pierre Budin, an obstetrician who was a leader in the care of premature infants. The American Academy of Pediatrics adopted this standard in 1935 (Cone, 1985). In 1948, the WHO defined prematurity as a birth weight of 2,500 grams (5 pounds, 8 ounces) or less.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention FIGURE 2-1 Fetal growth curves (birth weight percentiles by gestational age), United States, 1999 to 2000. Percentile based on gestational age calculated from LMP (first day of the mother’s last menstrual period); SGA = small for gestational age. SOURCE: Provided by Greg Alexander, 2006. The primary problem with the use of birth weight as a proxy for prematurity is that it identifies a group of infants heterogeneous for fetal development and may miss many preterm infants. In the 1960s, Battaglia and Lubchenco (1967) used measurements from a large population of infants to develop norms for fetal growth. At any given gestational age, the distribution of birth weights was such that some infants appeared to be within the norm for their gestational age (defined as “appropriate for gestational age”), some were relatively light (less than the 10th percentile for gestational age, or “small for gestational age”), and others were quite heavy (greater than the 10th percentile for gestational age, or “large for gestational age”). Battaglia and Lubchenco (1967) demonstrated that these categorizations of growth for gestational age had implications for mortality and morbidity. Thus, many preterm infants are large for gestational age but have a normal birth weight, and the rates of mortality and morbidity for these infants differ from those for term infants of normal birth weight. Moreover, at any gestational age, infants who had grown less well (small for gestational age) had poorer outcomes than heavier infants at that gestational age (Lubchenco and Butterfield, 1983). Since then, a number of neonatal growth curves have been published, with the standard diagnosis of small for gestational

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention age based on a birth weight less than 10 percent of the birth weight norm (Alexander et al., 1996, 1999; Kramer et al., 2001a; Usher and McLean, 1969) (Figure 2-1). Because of the higher accuracy of measures of birth weight, until recently, most researchers have continued to use birth weight cutoffs to designate infant risk. These included very low birth weight (VLBW) infants, whose birth weights are less than 1,500 grams (3 pounds, 5 ounces), and extremely low birthweight (ELBW) infants, whose birth weights are less than 1,000 grams (2 pounds, 3 ounces). The metric system is preferred because an ounce is not small enough to denote significant differences in weights. Moreover, not all researchers used the same birth weight categories, with some using birth weight categories of birth weight less than 2,000, 1,700, 1,250, 800, 750 grams, or, most recently, 600 or 500 grams (Chapter 11). Some researchers who subdivided their study cohort of infants with birth weight less than 1,500 g into two more categories used the term VLBW to describe infants with birth weights of 1,000 to 1,499 grams. The lack of commonly used birth weight categories makes this literature difficult to summarize. Unfortunately, few studies report outcomes by gestational age category. It has only been in the last several decades that gestational age estimates have become more accurate, primarily because of the increase in use of prenatal ultrasounds. Finding 2-1: Birth weight is an incomplete surrogate for gestational age for determination of the risk of perinatal morbidity and mortality. MEASUREMENT OF GESTATIONAL AGE To operationalize the current definition of prematurity, accurate measures of the duration of pregnancy (i.e., gestational age) are needed. Several methods are used to determine gestational age, but many are based on prenatal ultrasounds, which have provided a window onto the fetus and allowed observation of fetal growth and development (Goldstein et al., 1988; Neilson, 2000; Nyberg et al., 2004; Timor-Tritsch et al., 1988; Warren et al., 1989). Most often, prenatal ultrasounds determine pregnancy duration with early measures of fetal size, when there is little individual variation in fetal growth. Individual variations in fetal growth increase with the duration of the pregnancy and become quite prominent by the third trimester. It is ironic that gestational age, which reflects time (duration of pregnancy), is in fact often operationally determined by measures by fetal growth. A number of fetal growth curves have been generated and used to monitor fetal growth during pregnancy.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Use of Date of Last Menstrual Period Except for women who have used assisted reproductive technologies (ARTs; e.g., in vitro fertilization), the timing of the initiation of a pregnancy is imputed from the first day of the mother’s last menstrual period (LMP). Obstetricians have traditionally confirmed pregnancy dating by combining information regarding the mother’s LMP, periodic measurements of the mother’s abdomen, and when fetal heart sounds and movement (i.e., quickening) are detected (Rawlings and Moore, 1970). If a mother’s menstrual cycle is regular (i.e., it is 28 to 29 days long) and she receives good prenatal care with no problems with the pregnancy, LMP can be used to estimate the duration of a pregnancy (Rossavik and Fishburne, 1989). However, wide biologic individual variations in the interval between the onset of LMP and conception (from 7 to more than 25 days) can be due to variations in the timing of menstrual cycles, ovulation, and implantation of the blastocyst. Changes in age, levels of physical activity, body mass index (BMI), nutrition, breast-feeding, interpregnancy interval, smoking, alcohol consumption, and stressful life events can influence the length of an individual woman’s menstrual cycle and can therefore influence accuracy of LMP in estimating the duration of a pregnancy (Kato et al., 1999; Liu et al., 2004; Munster et al., 1992; Rowland et al., 2002). In addition to biological variations in menstrual cycles, ovulation, and implantation, many other factors contribute to difficulties with the use of LMP for pregnancy dating. Irregular menses, first-trimester vaginal bleeding, unrecognized spontaneous abortions, oral contraceptive use, and recall errors contribute to errors in calculating the duration of a pregnancy from LMP. Mothers who are socioeconomically disadvantaged are more likely to receive late or no prenatal care and to have a poor recall of LMP (Campbell et al., 1985; Dubowitz and Goldberg, 1981; Buekens et al., 1984). As many as 25 to 50 percent of the women in some samples have had difficulty recalling LMP (Campbell et al., 1985). Determination of gestational age by the use of LMP or by the use of clinical estimates thus causes significant differences in the gestational age distributions and in the preterm and postterm birth rates for large populations (Alexander et al., 1995; Mustafa and David, 2001). LMP data are missing or incomplete on approximately 20 percent of certificates of live births in the United States, especially for women who are socioeconomically disadvantaged, who are most at risk for preterm birth and intrauterine growth restriction (IUGR). Uncertainty about actual dates contributes to the recording of a digit preference for LMP (e.g., the most common day for LMP is the 15th of the month) (Savitz et al., 2002a; Waller et al., 2000). When LMP is used to determine gestational age, 40 weeks is added to the LMP to calculate the estimated date of confinement (EDC; which is also referred to as the estimated due date); that is, the day when the infant is due

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention FIGURE 2-2 Age terminology during the perinatal period. SOURCE: CFN (2004, p. 1363). (Reprinted with permission from Pediatrics, Vol. 114, pg. 1363, © 2004 by the AAP.) to be born (Figure 2-2). This convention of defining gestational age in terms of LMP has been incorporated into the definition of gestational age and used for many years. For women who have used ARTs, EDC is determined from the date of egg retrieval (which is equivalent to the day of ovulation to give a true conceptional age), but gestational age is expressed by use of the conventional definition (thereby adding 2 weeks to the conceptional age, which is approximately the timing of ovulation in a natural cycle). Although it is confusing, gestational age as arbitrarily defined includes an estimated 2 weeks before the embryo is fertilized for the sake of convention and based on the historical use of the term gestational age. Neonatologists and pediatricians have also adopted this arbitrary historical convention from obstetricians regarding gestational age. Neonatologists anticipate the amount of resuscitation and support a preterm infant may need both in the delivery room and in the NICU. Chronological age is the age from the time of birth of an infant, whether the infant was born preterm or fullterm. Postmenstrual age, which suggests (but does not guarantee) a specific degree of maturity/immaturity, is the infant’s chronological age (from birth) plus the infant’s gestational age at birth (Table 2-1). For office visits following discharge from the NICU, the pediatrician calculates the infant’s chronological age (from birth) and the infant’s age corrected for degree of prematurity (corrected age, calculated from the infant’s due date or EDC.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention TABLE 2-1 Methods for Determining Gestational Age Category Timing Measure Type Prenatal First to third trimesters Last menstrual period Maternal recall Prenatal First to third trimesters Obstetric clinical estimate* Maternal exam Obstetric ultrasound First trimester Detection of gestational sac, crown-rump length Fetal size Obstetric ultrasound Second to third trimester Biparietal diameter, femur length, abdominal and chest circumference Fetal size Postnatal Birth (within first day) Anthropometric measurements: birthweight, length, head circumference, foot length Infant size Postnatal Birth to 7 days Three to four days after birth to 40 weeks External physical characteristics Infant exam Postnatal PMA Neurological assessment: Amiel-Tison Infant exam Postnatal Birth to 4 or 5 days Combination: Dubowitz, Ballard, New Ballard Score and others** Infant exam Postnatal Any time after birth Disappearance of pupillary membrane (Anterior Vascular Capsule of Lens) Infant eye exam NOTE: PMA = Postmenstrual age. *Includes onset of pregnancy symptoms, fundal height, time when fetal heartbeat first detected, time of quickening (maternal detection of fetal movement). **Other combinations have been proposed but are less well known and are used less frequently (See Allen, 2005a for references). The new Ballard Score is most accurate in infants with less than 26 weeks if performed before 12 h after birth (See Allen, 2005a for references). SOURCE: Allen (2005a). Reprinted with permission from Mental Retardation and Developmental Disabilities, Vol. 11, p. 23, © 2005 by Wiley-Liss, Inc. Use of Ultrasound to Measure Gestational Age Prenatal gestational age estimates, especially those obtained by early fetal ultrasound, have proven to be more reliable than postnatal estimates of gestational age (Alexander and Allen, 1996; Allen, 2005a; Wariyar et al.,

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention 1997). The accuracy of prenatal ultrasound estimates of gestational age has been confirmed in a number of studies with women who had used ARTs or whose ovulation was confirmed on the basis of basal temperature records (Kalish et al., 2004; Nyberg et al., 2004; Persson and Weldner, 1986; Rossavik and Fishburne, 1989; Saltvedt et al., 2004). The earlier the ultrasound in the pregnancy, the more accurate the dating of the pregnancy (Drey et al., 2005; Johnsen et al., 2005; Kalish et al., 2004; Neilson, 1998; Nyberg et al., 2004). Measurements of fetal length from head to buttocks (i.e., the crown-rump length) can be used during the first trimester, and it is accurate within 2 to 5 days (Hadlock et al., 1992; Kalish et al., 2004; Wisser et al., 2003). By 14 weeks gestation, a fetus flexes and other measurements are used (e.g., biparietal diameter of the fetal head; head, abdominal and chest circumference; and femur and foot lengths) (Hadlock et al., 1987; Nyberg et al., 2004). Between 14 and 18 weeks gestation, measurement of the biparietal diameter of the fetus’s head estimates gestational age to within 9 days (Wariyar et al., 1997), and head circumference estimates gestational age to within 4 days (Chervenak et al., 1998). The use of multiple fetal measurements on a second-trimester ultrasound improves the accuracy of estimation of the gestational age (Chervenak et al., 1998; Hadlock et al., 1987; Johnsen et al., 2005). Because of increasing individual and pathological variations in fetal growth, gestational age estimates based on fetal measurements from a third-trimester ultrasound are less accurate, especially as the fetus approaches term (Alexander et al., 1999; Altman and Chitty, 1994; Lubchenco et al., 1963; Nyberg et al., 2004). By the third trimester, fetal growth can be adversely influenced by many environmental factors, including uteroplacental insufficiency, maternal drugs or toxins, and congenital infections. Racial, ethnic, and gender variations in birth weight by gestational age are the most prominent during the third trimester (Alexander et al., 1999). In multiple gestations, intrauterine crowding and competition for resources often results in IUGR. Intrauterine growth in twin pregnancies begins to diverge from that in singleton pregnancies at as early as 28 to 30 weeks gestation, with significant differences detected by 35 weeks gestation (Alexander et al., 1998; Min et al., 2000). Intrauterine growth variation is further compromised with higher-order multiples (e.g., triplets and quadruplets) (Alexander et al., 1998; Luke, 1996). Early prenatal ultrasounds before 20 weeks gestation are more accurate (95 percent confidence interval = ± 3 to 5 days) than any other prenatal or postnatal estimate of pregnancy duration (Alexander et al., 1992; Chervenak et al., 1998; Nyberg et al., 2004; Wisser and Dirscheld, 1994). Studies that have compared ultrasound with LMP have found that more infants were born within 1 to 2 weeks of their due date if the due date was calculated by ultrasound rather than by LMP (Mongelli and Gardosi, 1996;

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Discussion of the management and survival of infants born at the lower limit of viability requires extreme precision and attention to how mortality and outcome rates are calculated. For example, data should be provided as to what proportion of infants were resuscitated at delivery and provided with neonatal intensive care. The proportion of infants with congenital anomalies, especially anomalies that contribute to the infant’s death, should be reported. Care should also be taken to describe the chronological age at which survival is ascertained. The conventional definition of the neonatal mortality rate excludes infants who survive past 28 days but who die before they leave the NICU. Likewise, studies that report survival to the time of NICU discharge may miss deaths later in the first year of life that would be captured by infant mortality rates. When reviewing mortality rates for infants born at the limit of viability, attention to the denominator used to calculate mortality rates at the limit of viability is important (Allen et al., 1993; Evans and Levene, 2001). Many tertiary-care NICUs report birth weight- and gestational age-specific mortality rates that use the number of infants admitted to the NICU as the denominator. However, many infants born at 22 to 25 weeks gestation die shortly after delivery, and are never admitted to an NICU. In addition, a large proportion of infants born at a gestational age of less than 23 weeks or with a birth weight of less than 500 grams are stillborn (60 to 89 percent and 68 to 77 percent, respectively) (Sauve et al., 1998; Wood et al., 2000). Although there are guidelines for distinguishing between a fetal death and live birth (Table 2-2), the clinical distinction is not as clear as one might think. A fetal death occurs before the fetus is completely delivered and excludes induced terminations. The clinician must distinguish between evidence of life (e.g., beating heart, pulsation of the umbilical cord, or the movement of voluntary muscles) from transient or fleeting cardiac contractions, gasps, or jerks of the limbs. This is most difficult in fetuses born at 21 to 24 weeks of gestation. How many of these infants have been categorized in the past as live births instead of fetal deaths is unknown, as is how this categorization varies from region to region and even among health care providers at the same institution. A willingness to resuscitate a very immature infant who has a transient heart beat or gasp at delivery changes the classification of that infant from a fetal death to an infant death. This type of change in how an infant is classified has only a small impact on the preterm birth rate (because so many more infants are born after 26 weeks gestation), but could contribute substantially to rising U.S. Infant Mortality Rates. The use of perinatal mortality rates (the number of deaths of infants with gestational ages greater than 20 weeks/1,000 total births) may be a more useful measure of the outcomes of very preterm infants, since it includes infants who are stillborn and infants who die immediately after birth.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention TABLE 2-2 Definitions of Spontaneous Abortion, Fetal Death, Stillbirth and Live Birth Spontaneous abortion Abortion occurring without medical or mechanical means to empty the uterus. Fetal death or stillbirth Death before the complete expulsion or extraction from its mother of a product of conception, irrespective of the duration of pregnancy. The death is indicated by the fact that after such separation, the fetus does not breathe or show any other evidence of life, such as beating of the heart, pulsation of the umbilical cord, or definite movement of voluntary muscles. A death that occurs at 20 or more weeks of gestation constitutes a fetal death and after 28 weeks it is considered a late fetal death. Live birth The term used to record a birth whenever the newborn at or sometime after birth breathes spontaneously, or shows any other sign of life such as a heartbeat or definite spontaneous movement of voluntary muscles. Heartbeats are to be distinguished from transient cardiac contractions, and respirations are to be distinguished from fleeting respiratory efforts or gasps. SOURCES: CDC (2004e), Cunningham et al. (2005). Marked regional variations in the management and the rates of survival of infants born at the lower limit of viability and variations in the methods used to estimate gestational age make it difficult to evaluate trends over time with respect to live birth rates by gestational age and fetal death rates (Costeloe et al., 2000; Lorenz et al., 2001; Sanders et al., 1998; Tyson et al., 1996). For infants born at the lower limit of viability, an accurate estimate of gestational age is essential for guiding discussions about the many decisions to be made during and after delivery, including the timing and mode of delivery and whether antenatal steroids and aggressive resuscitation should be used at the time of delivery. Discussions of management and outcomes should focus not on when survival is possible but on a working definition of the limit of viability, when chances of survival, or of survival without major disability, are substantial (for example, 50 percent). Furthermore, better measures of fetal and infant maturity have the potential to improve the clinical care that is provided, improve the ability to predict short- and long-term outcomes, and assist families and health care providers in making difficult care management decisions.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Late-Term or Near-Term Infants Born at the Upper Border of Prematurity At all times during a pregnancy, accurate dating of the pregnancy and accurate estimates of fetal maturity provide better information for decision making by the health care provider and the family. Although this is especially true for the high-risk pregnancies noted above, the information also assists the health care provider and the family with making decisions on how a threatened preterm delivery is managed and the optimal timing and mode of delivery as a pregnancy approaches fullterm. Accurate gestational age and maturity information facilitates better prenatal counseling about the anticipated chance of survival, complications, and the long-term health and neurodevelopmental outcomes of the preterm infant born near to fullterm. There is not a standard accepted definition of this category of infants. Studies have varied as to whether they included infants with gestational ages from 32, 33, or 34 completed weeks gestation to 36 completed weeks gestation (Amiel-Tison et al., 2002). These infants have been called nearterm infants, late-term infants, and macropremies, but “late-term infants” serves as a reminder that they are not yet fullterm. The majority of preterm infants are born at 33 to 36 weeks of gestation (Table 2-3) (Appendix B). From 1995 to 2000, 8.9 percent of all U.S. births were infants born at 33 to 36 weeks gestation, whereas only 3 percent were born at gestational ages of less than 33 weeks. As many as 34 percent of twins are born at 32 to 35 weeks gestation, 31 percent are born at 36 to 37 weeks of gestation, and only 24 percent are born after 37 weeks gestation (Min et al., 2000). Many near-term preterm infants have normal birth weights, and most receive routine care in well-baby nurseries (Amiel-Tison et al., 2004a; Wang ML et al., 2004). TABLE 2-3 Proportion of Births for Various Gestational Age Categories, United States,1985 to 1988 and 1995 to 2000 Gestational Age (weeks) Gestational Age Categories 1985–1988 1995–2000 ≤28 Extremely preterm 0.66 0.82 ≤32 Very preterm 1.9 2.2 33-36 Moderately preterm 7.7 8.9 <37 Preterm 9.7 11.2 42+ Postterm 11.9 7.0 SOURCE: Alexander (2006 [Appendix B]).

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Although their outcomes are better than outcomes of preterm infants with gestational ages of less than 32 or 33 weeks, late preterm infants remain vulnerable to the complications of prematurity. They are more likely than fullterm infants to experience cold stress, hypoglycemia, respiratory distress syndrome, jaundice, and sepsis, yet there are wide variations among hospitals in treatments and resource use for late preterm infants (Amiel-Tison et al., 2002; Laptook and Jackson, 2006; Lewis et al., 1996; McCormick et al., 2006; Wang ML et al., 2004). Despite a relative lack of information regarding long-term outcomes, retrospective studies of children with cerebral palsy report that 16 percent to 20 percent were born between 32 and 36 weeks gestation (Hagberg et al., 1996; MacGillivray and Campbell, 1995). Accurate estimates of gestational age and better measures of fetal and infant maturity would provide important information for clinical decision making. Recognizing the higher mortality and morbidity rates for late preterm infants than fullterm infants, health care providers and families need to weigh carefully the advantages of earlier delivery against the health, financial, and economic costs of preterm delivery. IMPLICATIONS FOR PUBLIC HEALTH AND RESEARCH Although the complex interplay between the duration of pregnancy, fetal and infant size and maturity, and how they are measured are sources of some confusion, evaluation of the interrelationships among these factors provides an opportunity to gain some insight into the factors contributing to preterm birth. For example, racial disparities in all aspects of health have long been recognized, and the causes of these disparities are poorly understood. Public health databases with data on births, health problems, and deaths for large populations are available for exploration; but it must be recognized that the definitions of these variables may have changed over time. Racial and Ethnic Disparities Although controversy exists over inclusion criteria for racial and ethnic subgroups, racial and ethnic disparities in preterm birth rates, birth weight distributions for gestational age, neonatal and infant mortality rates, and gestational age- and birth weight-specific neonatal mortality rates have been consistently reported (see Appendix B). In the United States in 2003, preterm birth rates were 10.5 for Asian and Pacific Islanders, 11.3 percent for whites, and 17.8 percent for African Americans (Chapter 1). In 1997, the birth rates for white, Hispanic, and African American infants with gestational ages less than 28 weeks were 0.35, 0.45, and 1.39 percent, respectively

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention (Alexander et al., 2003). Racial, ethnic, and gender differences in birth weight for gestational age become increasingly prominent as pregnancies approach term (Alexander et al., 1999). At 40 weeks of gestation, birth weight for African American infants tends to be lower than those for white, Hispanic, and Native American infants. Furthermore, despite overall dramatic reductions over the decades, racial and ethnic disparities in neonatal and infant mortality rates in the United States persist (Alexander et al., 2003). Although mortality rates are higher for full-term African American infants than for full-term white infants, the smaller and more preterm the infant is, the more of a survival advantage that preterm African American infants have over preterm white and preterm Hispanic infants (Alexander et al., 1999, 2003; Allen et al., 2000; Demissie et al., 2001). As gestational age-specific neonatal mortality has decreased over the last several decades, this gap has narrowed but still exists for the more immature or smaller preterm infants (Allen et al., 2000; Hamvas et al., 1996, Appendix B). Borrowing the pharmacologists’ concept of the 50 percent lethal dose (LD50), which is the point at which 50 percent of a population dies and 50 percent survives, that 50 percent point decreased from 26.8 weeks gestation in 1975–1979 to 24.5 weeks in 1990–1994 for white infants born in South Carolina (Allen et al., 2000). For African American infants, the gestational age at which 50 percent survive decreased from 25.2 to 23.9 weeks. The gap in survival between white and African American infants decreased during those two time periods, from a difference of 1.6 weeks to 0.5 week of gestation, respectively. Although racial disparities in preterm birth rates and mortality rates have been noted for many years, the reasons for these differences are complex and not well understood. Some have suggested that many of the obstetric and neonatal intensive care advances (e.g., surfactant and antenatal steroid use) disproportionately improved the rates of survival for white preterm infants (who had higher incidences of RDS within each gestational age category) (Hamvas et al., 1996). Gender disparities, with higher mortality and pulmonary morbidity rates in preterm male infants are presumed to be based on not yet elucidated biological mechanisms (Stevenson et al., 2005). Others are concerned that neonatal mortality rates are higher in hospitals that serve predominantly minority populations (Morales et al., 2005). The reported racial and ethnic differences in risk factors for and presentations of preterm birth suggest that the etiologies of preterm birth may play a role (Ananth et al., 2005; Reagan and Salsberry, 2005). From 1989 to 2000, the rate of preterm birth following preterm labor increased by 3 percent for whites but decreased by 27 percent for African Americans (Ananth et al., 2005). Medically indicated preterm birth increased for both groups, but at different rates: 55 and 32 percent for whites and African Americans, respectively. Preterm births following ruptured membranes de-

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention creased for both groups: 23 and 37 percent, respectively. Declines in neonatal mortality were associated with increases in medically indicated preterm births in white preterm infants but with declines in preterm birth following labor or ruptured membranes in black preterm infants. An important question is how racial disparities in preterm birth are influenced by the broad social context (see Chapter 4). An important consideration for all of these issues is that the method used to estimate gestational age may play a complex (and interfering) role. Postnatal estimates of gestational age, especially the Ballard and Dubowitz gestational age assessments, tend to overestimate gestational age in preterm infants, but the magnitude of the overestimation varies with race and ethnicity. Even when a variety of maternal sociodemographic variables, pregnancy complications, and the type of delivery are controlled for, African American infants had significantly higher mean postnatal gestational age estimates for each gestational age interval (Alexander et al., 1992). The Ballard gestational age estimate was higher for African Americans for each gestational age interval, whether gestational age was determined by LMP or prenatal ultrasound (for a subset of infants for whom these data were available). In this study, the proportion of preterm births changed, depending on whether gestational age was estimated from LMP or Ballard score, and the amount of change varied by race. Since the Ballard estimates gestational age from physical and neurological infant characteristics, these and other similar data raise the question as to whether African-American infants mature more rapidly than white infants. In response to missing data for gestational age as determined by LMP on many U.S. birth certificates, a location for insertion of the clinical estimate of the infant’s gestational age was added to U.S. birth certificates in 1989 and was intended to be a cross-check of the gestational age determined by LMP. It is likely that postnatal estimates factored into the clinical gestational age estimates recorded on the birth certificate, although the magnitude of this tendency is unknown. Any data obtained by using these clinical gestational age estimates are therefore difficult to interpret, especially when one is trying to discern true racial and ethnic differences from artifacts of data reporting (i.e., the use of postnatal gestational age estimates). This problem has implications for the use of the clinical gestational age estimates to evaluate differences in gestational age distributions, birth weights for gestational age, and gestational age-specific mortality data. Rather than reverting to a reliance only on birth weight data, this problem argues for moving toward the adoption of early ultrasound to establish or confirm the gestational age or EDC for all pregnancies. The earlier in gestation that the prenatal ultrasound is performed, the greater the validity that the gestational age estimates for all racial and ethnic groups becomes.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Public Health Databases Developed countries have set up complex public health systems to collect data on births, deaths, and other indicators of health for their populations. In the United States, the recording of vital events is the responsibility of each of the states and not the federal government. The National Vital Statistics System is a collaborative effort between state and territory agencies and, as federal representatives, the Centers for Disease Control and Prevention and the National Center for Health Statistics (NCHS) (Martin and Hoyert, 2002). NCHS develops standards for the uniform reporting of live births and fetal, neonatal, and infant deaths to national public health databases through cooperative agreements. Although birth certificates are intended to establish the date of birth, the citizenship, and the nationality of a newborn infant, they contain valuable public health information and are the only national source of birth weight and gestational age data. Large state and national population databases with birth and death certificate data have been used to plot gestational age distributions, birth weights for gestational age, and gestational age- and birth weight-specific neonatal mortality rates. The systematic and random misclassification of birth weight, gestational age, and other birth and death certificate data is a continuing problem with all population databases, although procedures have been established to clean the data to eliminate those that are implausible. A consistent finding in many sets of perinatal data is a bimodal distribution of birth weight for gestational age, which is often attributed to the occasional miscoding of gestational age in fullterm infants with normal birth weights (David, 1980; Platt, 2002). One study noted frequent errors when obstetrics estimates of gestational age were transferred to infants’ medical records: in 15 percent of the cases, the gestational age was wrong by at least a week (Wariyar et al., 1997). Some were systematic errors, such as recording of the gestational age at the time of maternal hospital admission and not the gestational age at birth and the lack of attention to high-quality antenatal data. Similar errors may be made on original birth and death certificates, depending on the recorder’s training, experience, and level of attention to detail. Gestational age is used to calculate a variety of statistical indicators used to monitor the health of the mothers and children in a population. These indicators include the proportion of infants born preterm (before 37 weeks gestation), very preterm (before 32 weeks gestation), with an LBW (birth weight less than 2,500 grams) or a VLBW (birth weight less than 1,500 grams), and small for gestational age. The gestational age of a fetus at the time of initiation of prenatal care is one factor included when the prenatal care index is computed. All of this information can be used to target public health interventions and monitor their effects.

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention Changes in how gestational age is estimated and recorded have implications for all fetal and neonatal health indicators and for time and regional comparisons of fetal and neonatal health. Although birth certificates have recorded the duration of pregnancy since 1939 (initially in months), LMP has been recorded only since 1968. Although the addition of a clinical estimate of gestational age was recommended in 1989, the method used to determine that estimate was not specified on the birth certificate. The proportion of clinicians who base the clinical gestational age estimate on postnatal assessments is unknown. The magnitude by which postnatal assessments overestimate gestational age varies with both gestational age and race or ethnicity (Alexander et al., 1990, 1992; Sanders et al., 1991; Shukla et al., 1987; Spinnato et al., 1984). As discussed earlier in this chapter, an additional problem with postnatal assessments of gestational age is that they are generally based on signs of maturity, and infants with IUGR and from complicated pregnancies can demonstrate accelerated maturation after 30 weeks of gestation (Amiel-Tison et al., 2004a,b). As a remedy, in 2003 NCHS recommended that the clinical estimate of gestational age be replaced with the best obstetric estimate of gestational age, which “should be determined by all perinatal factors and assessments such as ultrasound, but not the neonatal exam” (CDC, 2004b, p. 172). States will be gradually making this change over the next several years. Agencies and researchers will need to consider these changes when they analyze the trends and state-to-state comparisons using gestational age. Fortunately, the change in gestational age determination from the use of LMP to ultrasound-based data has less of an effect on the preterm birth rate than it does on the birth rate for postterm infants born after 41 weeks gestation. Studies agree that this shift decreases the postterm birth rates, but the increase in the preterm birth rate is less (Goldenberg et al., 1989; Kramer et al., 1988; Savitz et al., 2005; Yang et al., 2002c). The timing of the ultrasound assessment is important: ultrasounds early in pregnancy increase the number of births determined to be fullterm by LMP but reclassified as preterm by ultrasound assessment more than ultrasounds late in the pregnancy. Multiparous mothers and mothers with small stature, diabetes, and high prepregnancy BMIs and fetuses with chromosomal anomalies were more likely to have large (≥7-day) discrepancies between LMP- and early ultrasound-based gestational ages (Morin et al., 2005). Thus, a gradual shift toward the use of prenatal ultrasound to determine or confirm gestational age may be contributing to the rising preterm birth rates. Administrators and researchers working with data from these large databases need to recognize and specifically state how gestational age variables were calculated, used, and imputed, especially when the data are interpreted and used. Differences in gestational age variables interfere with and complicate comparisons among states and countries and over time. For

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention example, one study defined gestational age in completed weeks as estimated from LMP when it was available, imputed gestational age if the month and the year were recorded, but had to rely on clinical estimates in 4 to 5 percent of the cases and had to eliminate from the calculations birth certificates with missing data (2 percent of white infants, 2.7 percent of African American infants, and 3.6 percent of Hispanic infants) (Alexander et al., 2003). These proportions will change in the coming years as more states begin to record best obstetric estimates and the rate of clinical use of early ultrasound to date pregnancies increases. Clarifying Mortality Rates The decentralized system for the reporting of vital statistics in the United States has made it difficult to compare state-to-state variations in preterm birth, fetal death, and infant mortality rates (Martin and Hoyert, 2002). In addition to variations in the reporting of gestational age on birth certificates, state requirements for the reporting of fetal deaths vary. There are also regional differences in the rates of underreporting of fetal deaths and missing data on fetal deaths. As attention has shifted toward survival at the lower limits of viability, the definitions of a fetal death and a live birth require attention. How life and death are defined and how very immature and critically ill fetuses are managed at delivery may have important effects on a number of recent trends, including rising preterm birth, neonatal and infant mortality rates, and decreasing fetal death rates. Less attention has generally been paid to fetal deaths than to neonatal and infant deaths. Approximately 16 percent of all pregnancies end in the death of the fetus (Martin and Hoyert, 2002; Ventura et al., 2001). Fetal death generally includes spontaneous abortions, miscarriages, and stillbirths. The majority (more than 90 percent) of fetal deaths occur in the first 20 weeks pregnancy; 5 percent occur at 20 to 27 weeks gestation; and 2 percent occur late in pregnancy; that is, after 27 weeks gestation. The greatest decrease has been in fetal deaths after 27 weeks gestation. States have different requirements on the data on fetal deaths that must be reported; some require gestational age (gestational age at or beyond 16 weeks, 20 weeks, or 5 months), some require birth weight (birth weight at or above 350, 400, or 500 grams), and some require both gestational age and birth weight criteria. Missing data regarding initiation of prenatal care vary from 17 percent of the records of fetal deaths at 20 to 27 weeks of gestation to 11 percent of fetal deaths beyond 27 weeks of gestation and 2.8 percent of live births. The possibility exists that changing practices in the categorization and reporting of live births and fetal deaths have contributed to falling fetal death rates and rising preterm birth and infant mortality rates in the United

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention States (Martin and Hoyert, 2002). Earlier obstetric intervention when the fetus is not doing well, prompt aggressive resuscitation in the delivery room, and the initiation of neonatal intensive care may save the lives of a few fetuses that formerly died in utero. However, many die within a few days (or weeks) of birth. Such efforts may have a significant impact on infant mortality rates (CDC, 2004f; MacDorman et al., 2005). If more infants are born alive at the limit of viability but die within days of delivery, thereby contributing to infant deaths, then a small rise in the infant mortality rate should not be viewed with alarm. It might generate a discussion of relative costs (emotional as well as financial) and how limited health care resources should be used, but it is not an indicator of worsening child health. Similarly, intensive prenatal care of high-risk mothers facilitates the detection of fetuses whose adaptive systems become overwhelmed by adverse intrauterine circumstances. An indicated preterm delivery that prevents a fetal death is not an indicator of worsening infant health, even if it does contribute to a higher preterm birth rate Consideration should be given to using perinatal mortality rates as another child health indicator. Perinatal mortality rates include fetal death rates as well as neonatal mortality rates and would not be expected to change in the two scenarios presented above. However, calculating perinatal mortality rates as an indicator of child health requires that attention be given to how perinatal data are collected and reported, especially regarding fetal deaths. Attention to the quality of data regarding causes of death could provide insight into the mechanisms of preterm birth as well as the causes of fetal and early neonatal deaths. This approach would require clinicians to vigorously search for causes of death whenever a fetal or early neonatal death occurs. Petersson et al. (2004) found that 11.5 percent of fetal deaths that would have been characterized as unexplained were due to infections with parvovirus, cytomegalovirus, or enterovirus. Other possible causes that should be explored include thrombophilias (e.g., Factor V Leiden), fetomaternal hemorrhage, chorioamnionitis (which would include pathologic examination of the placenta and the umbilical cord), uterine anomalies, umbilical cord or placental anomalies, toxin or drug exposures, and maternal illness (e.g., diabetes, hypertension with or without preeclampsia, thyroid disease, and autoimmune diseases) (Gardosi et al., 2005). Because the most common condition associated with fetal deaths is IUGR (43 percent), research into the mechanisms of IUGR, how to better detect IUGR, and the effect of IUGR on fetal organ systems should be encouraged. Every avenue that might lead to a better understanding of the causes and mechanisms of preterm birth, fetal and infant mortality, complications of prematurity, health sequelae, and neurodevelopmental disabilities should

OCR for page 53
Preterm Birth: Causes, Consequences, and Prevention be explored. The costs to infants, their families, and society as a whole are too high to continue to ignore this lack of knowledge. CONCLUSIONS The use of caution with the terms used and attention to their definitions is essential in efforts to understand the causes and consequences of preterm birth. It is important to recognize the limitations and variations in the data collected and entered into large administrative and research public health databases. Every effort should be made to improve the quality of national vital records, especially data on the gestational ages of newborns and the rates of preterm births. Uniform data collection and reporting procedures facilitate comparisons among states, over time, and with data from other countries. The impact of early dating of gestational age by ultrasound on clinical factors such as labor, tocolysis, administration of steroids, timing of elective induction of labor, determination of the mode of delivery, in utero transport, delivery room resuscitation, and determination of adequacy of fetal growth should be evaluated. Professional societies should encourage the routine use of early (before 20 weeks gestation) ultrasound for the establishment of gestational age. Standards of practice and recommendations for training of personnel to improve the reliability and the quality of ultrasound data should be established.