Article Text

Download PDFPDF

Rates of very preterm birth in Europe and neonatal mortality rates
  1. D Field1,
  2. E S Draper1,
  3. A Fenton2,
  4. E Papiernik3,
  5. J Zeitlin4,
  6. B Blondel4,
  7. M Cuttini5,
  8. R F Maier6,
  9. T Weber7,
  10. M Carrapato8,
  11. L Kollée9,
  12. J Gadzin10,
  13. P Van Reempts11
  1. 1
    Department of Health Sciences, University of Leicester, Leicester, UK
  2. 2
    Department of Neonatology, Royal Victoria Infirmary, Newcastle, UK
  3. 3
    Université Paris V Réné Descartes et Maternité de Port-Royal, Assistance-Publique Hôpitaux de Paris, Paris, France
  4. 4
    INSERM, UMR S149, Epidemiological Research Unit on Perinatal and Women’s Health, Université Pierre et Marie Curie-Paris 6, Paris, France
  5. 5
    Unit of Epidemiology, Ospedale Pediatrico Bambino Gesù, Rome, Italy
  6. 6
    Department of Neonatology, University Hospital, Marburg, Germany
  7. 7
    Departments of Pediatrics and Obstetrics, Hvidovre University Hospital, University of Copenhagen, Hvidovre, Denmark
  8. 8
    Department of Paediatrics, Hospital de Sao Sebastiao de Santa Maria da Feira, Santa Maria da Feira, Portugal
  9. 9
    Department of Paediatrics, University Medical Centre Children’s Hospital, Nijmegen, The Netherlands
  10. 10
    Department of Neonatology, University of Medical Sciences, Poznan, Poland
  11. 11
    Department of Neonatology, University Hospital, Antwerp, Belgium and Study Centre for Perinatal Epidemiology, Brussels, Belgium
  1. Professor D Field, Neonatal Unit, Leicester Royal Infirmary, Infirmary Square, Leicester LE1 5WW, UK; david.field{at}


Objective: To estimate the influence of variation in the rate of very preterm delivery on the reported rate of neonatal death in 10 European regions.

Design: Comparison of 10 separate geographically defined European populations, from nine European countries, over a 1-year period (7 months in one region).

Participants: All births that occurred between 22+0 and 31+6 weeks of gestation in 2003.

Main outcome measure: Neonatal death rate adjusted for rate of delivery at this gestation.

Results: Rate of delivery of all births at 22+0–31+6 weeks of gestation and live births only were calculated for each region. Two regions had significantly higher rates of very preterm delivery per 1000 births: Trent UK (16.8, 95% CI 15.7 to 17.9) and Northern UK (17.1, 95% CI 15.6 to 18.6); group mean 13.2 (95% CI 12.9 to 13.5). Four regions had rates significantly below the group average: Portugal North (10.7, 95% CI 9.6 to 11.8), Eastern and Central Netherlands (10.6, 95% CI 9.7 to 11.6), Eastern Denmark (11.2, 95% CI 10.1 to 12.4) and Lazio in Italy (11.0, 95% CI 10.1 to 11.9). Similar trends were seen in live birth data. Published rates of neonatal death for each region were then adjusted by applying (a) a standardised rate of very preterm delivery and (b) the existing death rate for babies born at this gestation in the individual region. This produced much greater homogeneity in terms of neonatal mortality.

Conclusions: Variation in the rate of very preterm delivery has a major influence on reported neonatal death rates.

View Full Text

Statistics from

Information about public health and reproductive health derived from the monitoring of neonatal and infant mortality rates clearly differs between developed and developing countries. However, even in affluent western countries where these rates are at historical lows, the figures still command public interest not least because they show significant regional variation both within individual countries and between apparently similar countries.1 2 For example, the UK government recently published a review of this topic with the aim of providing insight into the variation in infant mortality rates seen across the UK.3 This review recognised immaturity as the major cause of infant mortality and highlighted our poor understanding of the public health influences that result in high or low rates of preterm delivery.

What is already known on this topic

  • Neonatal mortality shows marked variation between economically similar countries.

  • Such variation is commonly ascribed to differences in the availability and organisation of health care.

What this study adds

  • This is the first study to look prospectively at the influence of rates of very preterm delivery on neonatal mortality in nine European countries.

  • Rates of very preterm delivery show marked variation across Europe but seem particularly high in the UK.

  • High rates of very preterm delivery are associated with high rates of neonatal mortality in the UK.

In comparisons of national neonatal and infant mortality rates between developed countries, the UK typically is seen to have a relatively high rate. Despite the heterogeneous nature of these rates, they are not usually subdivided into more homogeneous categories (such as those deaths related to immaturity, congenital anomalies, sudden infant death, etc) and hence discussion of why differences exist between apparently similar countries tends to focus on very broad issues such as differences in heath care delivery. Detailed and comprehensive comparisons of the influence of underlying rates of preterm delivery are rare because of the difficulties in identifying all cases (ie, those born alive and those born dead) at these low gestations. However, there is some evidence that such variation does exist and that the rate of very preterm birth in the UK is relatively high.4 5

In this paper we examine specifically the influence of very preterm births on the official neonatal mortality rates of 10 separate geographical regions from nine separate EU countries using data from a prospective study.


Data were collected as part of the MOSAIC project in 10 European study regions. MOSAIC included a prospective cohort study of all births that occurred between 22+0 and 31+6 weeks of gestation in study regions during 2003. The study regions were: Flanders in Belgium, the Eastern region of Denmark, six of eight districts in the Ile-de-France region of France, Hesse in Germany, Lazio in Italy, the Central and Eastern regions of the Netherlands, Wielkopolska and Lubiskie in Poland, the Northern region of Portugal, and the Northern region and former Trent region of the UK. In general these regions were from European countries with comparable levels of infant health outcomes; infant mortality rates ranged between 4 and 5.5 per 1000 in 2003, with the exception of Poland, where the infant mortality rate was 7 per 1000.

The inclusion period was all of 2003 except in the French region of Ile-de-France where data were collected from 1 February to 31 August 2003. Within each study region the local team established a methodology to ensure that information was captured for all births, live and dead, at this gestation. The methodology has been described in more detail previously.6

We estimated the effect of underlying variation in the rate of very preterm live birth between the different study regions on published neonatal death rates. Neonatal death rates using officially reported statistics were obtained for each of the study regions. An adjusted neonatal death rate was then calculated for each region based on the overall preterm live birth rate for 22+0–31+6-week infants in the following way. Firstly, the ratio of the live birth rate for each study region compared to the overall rate (for the 10 regions together) was calculated. This was used to estimate how many 22+0–31+6-week live births there would have been within each study region if the overall rate for the 10 regions was applied. The next step was to calculate how many of this revised number of 22+0–31+6-week live births would be expected to be neonatal deaths by applying the official rate of neonatal death that occurred for this gestational group within each of the study regions. Finally, the adjusted neonatal death rate was calculated using these revised figures and 95% CIs were calculated for birth and death rates. Neonatal death rates for the 10 regions were ranked both before and after adjustment.


The full MOSAIC cohort included 7222 infants of whom 689 were terminations of pregnancy for congenital anomalies or other reasons. These terminations of pregnancy were excluded and subsequent calculations were based on a total of 6533 births of whom 5204 were born alive.8 After excluding terminations of pregnancy, the very preterm birth rate for the 10 regions was 13.2 per 1000 births (95% CI 12.9 to 13.5). However, the rate of delivery at this gestation showed considerable variation (table 1). The two UK regions had significantly higher rates than all other regions and were similar to each other (Trent 16.7, Northern 17.1 very preterm births/1000 births) despite being geographically distant and using different data collection teams/methods. It is interesting to note that the rates of delivery at this gestation were low in the regions of the two most southern MOSAIC countries: Northern Portugal (10.7, 95% CI 9.6 to 11.8) and Lazio in Italy (11.0, 95% CI 10.1 to 11.9). However, the rate was lowest in the Netherlands (10.6, 95% CI 9.7 to 11.6). Concentrating on live born infants alone, a very similar pattern is seen with, once again, the two UK regions having a significantly higher rate than the remaining European study regions apart from the German region of Hesse (table 1).

Table 1 Outcome of very preterm births in the 10 regions

The unadjusted official neonatal death rates are also shown in table 1. The effect of adjusting for variation in the underlying rate of very premature delivery is shown in table 2. Two main effects are seen:

Table 2 Effect of adjustment for variation in the rate of very preterm live birth on the observed rate of neonatal death
  1. The neonatal death rates of both UK regions fell and as a result neither was significantly different to the group mean (that for Northern region became the second lowest).

  2. The neonatal death rate for the Eastern and Central Netherlands study region became significantly higher although its relative position moved just one place down.

In considering both the adjusted and unadjusted rates of neonatal death, that for Hesse in Germany was the best despite a relatively high rate of very preterm birth.


This study is the first to examine rates of very preterm delivery in a group of geographically defined populations in developed countries. Previous studies of areas within one or two countries have suggested that the rate of delivery at this gestation in the UK is particularly high and these data confirm this finding.4 5 Differences between the study regions/countries in this study have emerged despite the fact that Organisation for Economic Co-operation and Development (OECD) data suggest that they are broadly similar in economic terms. Only Poland had significantly lower gross domestic product, but in fact the rate of increase in health spending at the time of this study was highest in Poland.7 The anomalous position of the UK is further emphasised by the close agreement between the rates in the two separate UK regions.

The changes in the ranking of the study regions seen after adjustment of the neonatal mortality rates shows how high rates of very premature delivery adversely affect the UK performance in particular. The study regions from Portugal and Italy moved down the ranking after adjustment as a result of their low inherent rate of very preterm delivery but slightly higher than average rate of death amongst live births at this gestation. The Netherlands has a high rate of neonatal death amongst live births at this gestation, presumably as the result of the national policy of not resuscitating babies born at ⩽25 weeks of gestation. Similarly, the situation in Poland is adversely affected by the policy of not terminating pregnancies with antenatally recognised major congenital anomalies, thus resulting in a greater proportion of babies with lethal malformations included within their live birth figures.8 In contrast, it is worth highlighting that in the German study region of Hesse, the rate of very premature delivery was amongst the highest but the rate of neonatal death (adjusted and unadjusted) was the lowest of the 10 regions. This strongly suggests that neonatal care strategies in Hesse were more effective in keeping babies alive; however, we have no data on rates of longer term disability at this stage and clearly these are equally important in determining optimum outcome.

The lack of similar studies in the literature is evidence of the fact that acquiring data of this type which includes “all births” down to 22 weeks is fraught with difficulties of data quality and case ascertainment. However, our data were acquired prospectively by a well established group of collaborators who came together to focus specifically on various health care issues relating to this gestational group of infants. Strategies to ensure completeness of the various cohorts were established during the preparatory work and hence we are confident that the rates reported here are accurate.

Given what is known about these countries, the observed differences are unlikely to be related to the availability of health care. In this study, we were not able to assess the influence of increasing numbers of “older mothers”, noted by others to carry a higher risk of preterm delivery.9 Artificial reproductive therapies are associated with higher rates of preterm delivery, but we have no data to explore the influence of such interventions in the countries involved in this study10 Other potential influences on the observed differences are exposure to poor diet, manual work during pregnancy, and a range of factors related to deprivation, which lead to increased stress in the pregnant woman.1114 Evidence is emerging that links each of the above to a potential biological mechanism that results in preterm birth, such as the presence of organisms in the genital tract.15 Other evidence highlights the potential importance of diet.1618 Further research related to these associations will be important if we wish to prevent preterm birth rather than deal with its consequences.

Prematurity has major implications for individuals, families and society and there is a need for developed countries to establish accurately their rates of preterm delivery. In countries like the UK where the rate of very preterm delivery is higher than in otherwise similar countries, understanding this difference should be a priority. The causes of excess preterm birth are clearly far wider than that of teenage pregnancy, which has been the focus of much attention in the UK.19 Establishing the biological and environmental links in the aetiology of preterm delivery represents a major challenge for developed societies but also presents a major opportunity for health gain.20


The authors would like to acknowledge the assistance of the personnel in the maternity and neonatal units in the regions participating in the MOSAIC project. The MOSAIC research group: Belgium, Flanders (E Martens, G Martens, A Bekaert, P Van Reempts); Denmark, Eastern Denmark (K Boerch, T Weber, B Peitersen); France, Ile-de-France (G Bréart, JL Chabernaud, D Delmas, E Papiernik); Germany, Hesse (L Gortner, W Künzel, R Maier, B Misselwitz, S Schmidt); Italy, Lazio (R Agostino, D Di Lallo, R Paesano); Netherlands, Eastern and Central (L den Ouden, L Kollée, G Visser, J Gerrits, R de Heus); Poland, Northern Wielkopolska and Lubuskie (G Breborowicz, J Gadzinowski, J Mazela); Portugal, Northern region (H Barros, I Campos, M Carrapato); UK, Trent region (E Draper, D Field, J Konje); UK, Northern region (A Fenton, D Milligan, S Sturgiss); INSERM U149, Paris (G Bréart, B Blondel, J Zeitlin); external contributors (M Cuttini, S Petrou). Steering committee (E Papiernik, project leader, J Zeitlin, research coordinator, G Bréart, ES Draper, L Kollée).


View Abstract


  • Competing interests: None.

  • Funding: This project was partially funded by a grant from the European Commission Research Directorate (QLG4-CT-2001-01907) and coordinated by Assistance-Publique Hôpitaux de Paris. The MOSAIC research group maintained complete independence from the funding body in relation to the research process and reporting of results.

  • Contributions: All authors were involved in the execution of the study and preparation of the manuscript.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.