Preterm birth (PTB) is gaining increasing attention. During the four last years, several call for action and for research has been made throughout the systematic revue of worldwide incidence/prevalence and also throughout the series of Global Reports 1 2 3 4 .

The seven Global Reports on preterm birth and stillbirth, issues of Conference on Preterm birth and Stillbirth on may 2009 (USA), estimated that 13 million babies are born before 37 completed weeks of gestation annually. Later, in the Global Action Report on Preterm Birth: Born Too Soon (WHO 2012), the number of preterm birth increased to 15 million annually.

Rates (% births <37 weeks of gestation), are generally higher in low- and middle income countries than high income countries. In these countries is born 92% of all preterm birth and 99% of them die. Rates are also increasing in most of these countries 5 .

Multiple factors may be associated with increase in PTB, among them, socio demographics changes in maternal population such as increased maternal age and delayed primiparity 6 7 . Nevertheless, sociodemographic characteristics of women, as age, parity, marital status, age at first delivery and fertility rates, often have been the object of study by economists 8 and demographers 9 10 . Few studies has addressed their impact on obstetrical and perinatal outcomes, among them, researches in France (Unité 953 of INSERM) since 70’s, have reported that certain maternal conditions could affect maternal risk and could be associated with higher rates of preterm birth, with low birth weight and with fetal growth restriction 11 12 . Consequently, sociodemographic factors have been evaluated since 70’s through periodic surveys which monitor maternal and perinatal health 13 14 .

Globally, few countries have been assessed the maternal sociodemographic factors over time as Finland and Sweden who have monitored the maternal population through studies in birth cohorts 15 16 . Those countries have also demonstrated that certain maternal characteristics are associated with higher risk of PTB and they showed that these characteristics may vary significantly over time.

Chile is a middle income country where the PTB rates have significantly increased 17 . This increase has been recorded in a context of demographic and epidemiological transition (which began in the 1960s) 18 . This process has led to modifications in some characteristics of the population of women (similar to those observed in some more developed European countries in the 1970s) including a diminution of fecundity rate from: 2.2 children per woman in 1996 to 1.9 children per woman in 2004; a decrease rate of marriage from 19.6 per 1000 in 1996 to 15.1 per 1000 in 2004. During this process it has also been observed an increase of employment and higher education and finally delayed maternal age at first children 9 .

In Chile, maternal and perinatal health constitute a major focus of public health policies and PTB is a health priority 19 . However, the determinants of preterm birth in Chilean population are partially known. Therefore we cannot yet formulate primary and secondary prevention strategies based on specific characteristics of either women or mothers.

The objective of present study is firstly: to examine sociodemographics characteristics of maternal population in the past two decades, then, to analyze their association with the risk of PTB, and finally to identify groups at higher risk of giving birth to a premature child.

In all, 83 017 observations (1.85%) were removed from the initial set. Table 1 shows that the absence of these groups did not modify the observed trends, but it required a specific analysis. Some results from that specific group of births are: mothers with missing values of GA had more often the characteristics associated to PTB such as: not living with a partner: 54.9% versus 48.4% in all population, age >38 years old: 6.8% versus 4.3% in all population and grand multiparity: 9.1% versus 1.5% in all population, (data not showed). Also, births with missing values of BW had higher rates of PTB: (55.7%).

Chile, Singletons Live Births, 1991-2008.

Over an 18-year period, the rate of PTB has increased significantly in this population database. At the same time, the sociodemographic profile of mother’s population has changed substantially: parity has decreased and the percentage of younger, older and mothers not living with a partner rose.

Mothers following these trends had a considerable high risk of give birth to preterm child over time. Moreover, the number of women belonging to such groups at risk has also increased markedly (except for grand multiparas).

All and every of the studied factors are not enough to explain the increase of PTB because increase in the rates of PTB affected the entire population including those mothers without sociodemographic risk. Also, the more recent period contributed significantly to this increase independently of these factors.

Probably the last period reveals the effect of other possible factors, not available in this data base, such as life style changes, new patterns of sexual behavior or family structure. Some of this factors have already been observed by other studies in Chile 29 .

Furthermore, another factors reported in the literature, but not available in data base, should be determined in this population, such as the status and conditions of working mothers, exposure to environmental and occupational contaminants, consumption habits, and most generally, the health issues of women.

Moreover, these trends could be related, as well, to the increase in obstetric interventions (cesarean deliveries and induction of labor) established by public and private policy in order to reduce maternal and fetal risk. In fact, the rates of both cesarean deliveries and induction of labor have been increasing in Chile for three decades and are among the highest in South America 30 31 . Unfortunately, the modality of delivery is not available in data base and this is an important limitation for understanding the nature of PTB.

The observed trends over the past two decades may also be related to an increase number of birth of extremely preterm infants. This could in part be associated to obstetrical and perinatal interventions aimed at diminishing fetal mortality among them, the progressive use of systematic fetal USN of first trimester.

The results suggest that in Chile, early, accessible prenatal monitoring ensures the collection of good quality data for estimating gestational age. It has reported a steady level of agreement between ultrasounds performed during the first trimester and gestational dates estimated by last menstrual period (61.9% in 1994 and 65.6% in 2001).

The quality of GA might be expressed in this study by the low proportion of error of classification of GA throughout the observation period. These errors decreased over time, this is probably the principal effect of systematic fetal USN of first trimester since the second period. In this context, the contribution of USN to increase of PTB seems to be modest. In addition, the trend to increase of PTB is sustained and it remains beyond the systematization of USN.

In all, only 1.85% of births were excluded of the initial set. This proportion is lower than other studies in population data base 32 33 . But the excluded births did not have a random distribution, then the exclusion of these groups could have underestimated the association between maternal characteristic and PTB especially during the first period, nonetheless the effect on the rates seems to be insignificant since these exclusions did not affect the trends of PTB.

The Tukey’s rule, was also used by Arbuckle in Canada 34 , this method has been recommended for data base with few errors 23 .

Comparing with previous studies, the trends observed in the groups at highest risk are similar to those observed in studies of parity, marital status and older mothers in France 35 and England 11 .

Similar results have been observed In Belgium 36 , whit a marked increase in the frequency of known factors contributors to PTB such as: maternal age > 35, assisted reproduction, obstetric intervention. Also increased the number of women with one or more of these contributors, but rate of PTB had a significantly increase in both groups: women with and women without these factors. Thereby, the steady increase in preterm birth could not be explained by these factors.

In the same direction of these results, Haas in USA 37 has reported a modest contribution of sociodemographic factors to increase of PTB. He concludes that other factors could be more important such as health status and access to medical care both before and during pregnancy.

In Chile access to medical care during pregnancy is universal. Beyond this, the high — and still rising — proportion of adolescents mothers is the continuation of an unsolved public health problem in Chile 38 .

Countries such as France 14 and Finland 39 , which have implemented long-term policies aimed at preventing preterm birth via integral prenatal care and social policies to support motherhood, have succeeded in significantly reducing the risk of PTB in younger mothers. Chile has not done so. This result confirms the importance of preconception care, from a conceptual framework such as has been given in Born Too Soon 5 .

Among the groups at highest risk, the observed increase in maternal age is a new public health challenge in terms of maternal and perinatal results, as it requires a greater focus regarding the need for medical interventions during pregnancy and delivery, for assisted reproductive technologies, and for other highly specialized services 40 .

Chile reached in 2003 its lowest rate of maternal mortality 12.2/100.000 live births after years of successfully reducing its prevalence, nevertheless the prevalence has risen over the past five years to 19.8/100.000 live births in 2005. The authors claim that maternal age over 35 years is the principal explanatory factor of this result, due to concomitant chronic diseases affecting such group of mothers 41 .

Unfortunately, one of the principal limitations of the study is our inability to assess other important factors like clinical maternal and fetal condition. Moreover the absence of information available about the mode of delivery prevents us from being able to distinguish between spontaneous and medically indicated PTB 9 .

Concerning rates of multiple gestations in Chile, they are increasing; this trend is showed in a recent study made over the same period (1991–2008) 17 . This increase could also be associated to the maternal age and also to the increase of fertility treatments.

Concerning assisted reproductive technologies (ARTs), they are rising and it is related to the increase of maternal age and late primiparity. However, this group of births has not significantly contributed to the rate of prematurity. Indeed, the report including twenty years of ARTs in Chile shows that between 1991 and 2008 only 3910 children were born through ARTs. The contribution of this group to preterm birth in those twenty years is poor: 353 single preterm birth, 585 doubles preterm and 210 preterm births in triples 42 .

The principal reason is that these techniques are not yet a large coverage because they are very high cost.

Another challenge is the group of mothers living without partner. This is the only group for which the risk is probably due to sociodemographic context; nonetheless this group is still largely unknown. Countries that have succeeded in reducing this risk have implemented measures to support these mothers and to reduce their psychosocial and economic vulnerability 43 44 ; Chile, so far lacks policies in these respect.

Further studies may be suitable using the others available variables in data base, such as: socio demographics characteristics of father, maternal education level. This group of variables could guide the hypotheses about the socio economic condition as important contributor au PTB.

In this direction, the study of K Joseph gave an important orientation about the distribution of sociodemographics factor by socioeconomics levels.

At the same way, the evaluation of the risk associated with interaction between these factors appears useful in view of the multifactorial etiology of PTB and LBW. For the moment; little research exists in this domain, but Schempf, in the United States, observed an important increase in risk of PTB in both groups: primipara teenager and primipara older in Hispanic mothers 7 .

Preterm birth has increased over the past two decades in Chile. Some sociodemographic characteristics of mothers are steadily associated with the risk of preterm birth over time. Such association has allowed identifying five groups of mothers at higher risk of PTB. The epidemiologic importance of these groups is changing over time through an increase in the number of women belonging to groups at higher risk of preterm delivery.

Despite of these specifics results, sociodemographics factors, by themselves, are not enough to explain the marked trends to increase in rates of PTB affecting the overall population.

Further research could be define the role of others underlying factors over time, that are probably widely affecting the population of mothers, especially during the most recent period.

The authors declare that they have no conflicts of interest in regard to this research.

PL Designed the study, made possible the acquisition of data, analyzed and interpreted the data. GB Made substantial contributions to conception and design of this study, has been involved in drafting the manuscript and revised it, providing critical and important intellectual content. It has revised and approved the final version submitted for publication. Both authors read and approved the final manuscript.