In 2009, the Global Alliance to Prevent Prematurity and Stillbirth Conference charged the authors to propose a new comprehensive, consistent, and uniform classification system for preterm birth. This first article reviews issues related to measurement of gestational age, clinical vs etiologic phenotypes, inclusion vs exclusion of multifetal and stillborn infants, and separation vs combination of pathways to preterm birth. The second article proposes answers to the questions raised here, and the third demonstrates how the proposed system might work in practice.
Preterm birth, defined by the World Health Organization (WHO) as birth prior to 37 complete weeks (259 days) after the first day of the last menstrual period preceding the pregnancy, is a major global public health problem. Recent systematic reviews by WHO and the March of Dimes have estimated that in 2005, 13 million infants were born preterm worldwide. The majority of these preterm births (85%) occur in Africa and Asia: the former because of high rates, the latter because of the large number of births. For several decades, preterm birth has been the focus of research and public health intervention in many developed countries because of the associated high risks of infant mortality and long-term neurocognitive, visual, and pulmonary sequelae and because rates have been stable or increasing.
See related editorial, page 99
In low- and middle-income countries, attention to preterm birth has been more recent, largely because gestational age (GA) at birth is not routinely recorded in noninstitutional deliveries and is often unknown. The longstanding emphasis on low birthweight, rather than preterm birth, naturally led to a primary emphasis on maternal under nutrition, which does not appear to be a major contributor to preterm birth.
In 2009, more than 200 participants attended the International Conference on Prematurity and Stillbirth convened by the Global Alliance to Prevent Prematurity and Stillbirth (GAPPS), hosted by Seattle Children’s Hospital with support from the Bill and Melinda Gates Foundation, March of Dimes, Save the Children, World Health Organization, United Nations Children’s Fund, and Program for Appropriate Technology in Health. A Global Action Agenda was developed (an agenda) that, in part, highlighted the need for a comprehensive, consistent, and uniform classification system for preterm birth.
The authors of this paper and the following two in the series were brought together as a direct result of the GAPPS meeting, with instructions to determine the need for such a classification system, to define the issues related to creating a preterm birth classification system, and to present a prototype classification system for general consideration.
Measurement issues
The earlier international definition of prematurity (birthweight ≤2500 g) did not distinguish between infants born early from those born small for their GA. For that reason, WHO changed the term from prematurity to preterm birth and defines the latter (preterm birth) based on GA only: a birth before 37 completed weeks (259 days) after the first day of the last menstrual period (LMP). That GA, a cutoff that was entirely arbitrary, however, is an issue that will be discussed later in this article.
First, it is important to consider how GA is measured. In the past, the GA estimate on a population-wide scale was primarily based on the LMP, an estimate that assumes that conception occurs on the same day of ovulation and that ovulation and conception occur 14 days after the onset of the LMP. Differences in the duration of the menstrual cycle, however, and particularly on the day of the menstrual cycle on which ovulation occurs, account for considerable variability in the day of conception vis-à-vis the first day of the LMP. In addition, LMP reporting is subject to error in recall and can be influenced by spotting or frank bleeding (which may reflect early miscarriages of clinically unrecognized pregnancies).
Historically, obstetricians and other prenatal care providers often used uterine fundal height as a check to validate the estimated duration of gestation, but more recently, clinical GA estimates have been based on ultrasound fetal measurements in the first half of the pregnancy. These are usually calculated from the biparietal diameter in the second trimester (13-20 weeks) or crown-rump length before 14 weeks of gestation.
Ultrasound-based estimates of GA have been shown to yield GA estimates that are 2-3 days earlier than LMP-based estimates (corresponding to ovulation on day 16-17 vs day 14), on average, and yield slightly higher rates of preterm birth. Ultrasound-based estimations may also have errors related to insufficient standardization and quality control of the operators, are based on equations derived from small samples (especially at early GAs) from different selected populations, do not provide variability around the GA estimates, and assume that all fetuses with the same measurement have the same GA (ie, they do not account for true differences in fetal growth in early gestation). Data entry errors can arise with either LMP or ultrasound GA estimates and can lead to an apparent bimodal or upwardly skewed distribution of birthweight at preterm GAs (especially at 27-32 weeks), owing to inclusion of term or near-term births.
In the United States and the United Kingdom, official estimates of preterm birth rates are based on LMP GA estimates. Despite the evidence from clinical and hospital-based study samples that ultrasound-based estimates are slightly shorter than LMP-based estimates, preterm birth rates based on the menstrual estimates in the United States have been shown to be considerably higher than those based on the clinical estimate, which, as in other countries, is increasingly based on ultrasound-derived GA estimates. US and Canadian preterm birth rates, for example, have been shown to be closer when the US rate is based on the clinical estimate. The higher US preterm birth rate based on the LMP GA estimate may reflect errors in recalling the date of the LMP, particularly among women with late onset of prenatal care.
The neonatal mortality rates among preterm births are very similar in Canada and the United States when the US rate is based on the clinical estimate but considerably lower when based on the menstrual (LMP) estimate, suggesting the inclusion of some term births among those classified as preterm based on LMP. Similarly, relative risks (vs infants born ≥37 weeks) are also very similar in Canada and the United States when based on the US clinical GA estimate but are considerably lower in the United States when based on the LMP-based estimate. Postterm rates (GA ≥42 weeks) are virtually identical in the United States (including both US whites and blacks) and Canada when the US rates are based on the clinical estimate but much higher in the United States when based on the menstrual estimate. In summary, these measurement differences can lead to substantial disparities in the GA distribution, and specifically the preterm birth rate, for different populations.
Finally, postnatal examination of the newborn infant has also been used to estimate GA, based on physical and neurological criteria. These estimates are far less satisfactory than early ultrasound-based prenatal estimates, however, because they are influenced by race, pregnancy complications, delivery complications, and birthweight for GA.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
