SEARCH

SEARCH BY CITATION

Keywords:

  • Lowrisk women;
  • live preterm birth;
  • singleton;
  • spontaneous onset of labour

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

Objectives  To describe a 10-year trend in preterm birth.

Design  Population-based study.

Setting  Australia.

Population  All women who gave birth during 1994–03.

Methods  The proportion of spontaneous preterm births (greater than or equal to 22 weeks of gestation and less than 37 completed weeks of gestation) was calculated by dividing the number of women who had a live spontaneous preterm birth (excluding elective caesarean section and induction of labour) by the total number of women who had a live birth after spontaneous onset of labour (excluding elective caesarean section and induction of labour). This method was repeated for the selected population of women at low risk.

Main outcome measure  Preterm birth rates among the overall population of women; preterm birth among all women with a spontaneous onset of labour; and preterm birth in a selected population of women who were either primiparous or multiparous, non-Indigenous; aged 20–40 years and who gave birth to a live singleton baby after the spontaneous onset of labour.

Results  Over the 10-year study period, the proportion of all women having a live preterm birth in Australia increased by 12.1% (from 5.9% in 1994 to 6.6% in 2003). Among women with a spontaneous onset of labour, there was an increase of 18.3% (from 5.7 to 6.7%). Among the selected population of low-risk women after the spontaneous onset of labour, the rate increased by 10.7% (from 5.6 to 6.2%) among first time mothers and by 19.2% (4.4–5.2%) among selected multiparous women.

Conclusions  Over the 10-year period of 1994–03, the rate of spontaneous preterm birth among low-risk women having a live singleton birth has risen in Australia.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

Preterm birth is a complex and unresolved public health problem1 that correlates strongly with poverty, socio-economic status2 and the level of education among women giving birth.3 It is associated with perinatal mortality, serious neonatal morbidity and moderate to severe childhood disability in wealthy countries.4,5 In addition, little is known about the actual costs of preterm birth beyond inpatient care and first hospitalisation.6 Although preterm births currently comprise between 10 and 12% of all births,4,5 they contribute to more than two-thirds of all perinatal deaths.4 An article published in the BMJ showed that a rise of 51% in spontaneous preterm birth in a population of low-risk primiparous women over a 10-year period was greater than the increase in the whole population over that time.7 In the editorial accompanying this article, two well-known obstetricians from the UK commented that it appears ‘Few interventions have improved outcome, and management remains an important challenge in modern obstetrics…If the trends are real, and confirmed in other countries, the impact for society is considerable.’8

The objective of our study was to replicate the study on spontaneous preterm birth in primiparous low-risk women by Langhoff-Roos et al.7 with Australian data to see if the trends in preterm birth were generalisable to another developed country setting over the same time period using a similarly constructed ‘standard’ or ‘selected’ population of primiparous women at low risk.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

Data were obtained from the National Perinatal Data Collection (NPDC) and the Assisted Conception Database (ACD) and Australian and New Zealand Assisted Reproduction Database (ANZARD). The ACD and ANZARD include data on assisted reproduction technology treatment and some pregnancy outcomes. Data from the ACD and ANZARD were included for all women who gave birth after assisted reproduction technology in Australia over the 9-year period between 1994 and 2002 to a liveborn baby greater than or equal to 22 weeks of gestation. The NPDC is a national cross-sectional database on all births in Australia collated from state and territory collections and validated by the Australian Institute of Health and Welfare National Perinatal Statistics Unit. The data include maternal demographic factors, pregnancy and labour characteristics and neonatal outcomes. Currently, the ANZARD and ACD databases are not linked to the NPDC.

The study population included all women who gave birth in Australia over the 10-year period between 1994 and 2003 to at least one liveborn baby greater than or equal to 22 weeks of gestation.

Preterm births are defined as births greater than or equal to 22 weeks of gestation and less than 37 completed weeks of gestation. The proportion of spontaneous preterm births was calculated by dividing the number of women who had a spontaneous preterm birth (excluding elective caesarean section and induction of labour) by the total number of women who gave birth in the population after spontaneous onset of labour.

To study the changes in preterm birth among women at low risk, we defined a ‘selected population’ of primiparous women considered to be at low risk of preterm birth and similar to those studied in the study by Langhoff-Roos et al.7 The selected population consisted of primiparous non-Indigenous women, aged between 20 and 40 years and who gave birth to a live singleton baby after the spontaneous onset of labour. We also studied the proportion of spontaneous preterm birth among a selected population of low-risk multiparous women.

Data were stratified by parity and adjusted for maternal age, multiple pregnancy, nonspontaneous onset of labour and year of birth. Using multivariate logistic regression, we calculated the risk of preterm birth (22–27 weeks, 28–31 weeks and 32–36 weeks) using term births as a reference point.

For outcomes where the adjusted odds ratio (AOR) for preterm birth was significantly raised at the 5% level, the population proportional attributable risk (PAR) was calculated according to the following formula:

  • image

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

Over the 10 years, 2 527 339 women gave birth to a baby between 22 and 45 weeks of gestation. We excluded 0.54% of women who had a stillbirth (fetal death); 0.48% of records where there were births with missing gestational age; and another 2.38% of the total population where other relevant fields were missing. This left a study population of 2 510 623 women who gave birth to 2 540 199 liveborn babies. The overall proportion of women who gave birth to a preterm baby at greater than or equal to 22 weeks of gestation increased by 12.1% (from 5.9% in 1994 to 6.6% in 2003) (Table 1). Increases occurred over the 10 years in each of the three preterm groups (Table 1).

Table 1.  The percentage of women who had one live birth at 22–45 weeks of gestation, Australia, 1994–03
Women who had one live baby1994199519961997199819992000200120022003Change
n = 254 409n = 253 694n = 251 033n = 251 182n = 249 490n = 251 459n = 251 242n = 248 296n = 249 056n = 250 7621994–03
  • *

    Extremely preterm, 22–27 completed weeks; very preterm, 28–31 weeks; moderately preterm, 32–36 weeks; term, ≥37 weeks.

  • **

    Data were excluded from one state (which supplies approximately 2% of the national data) for these fields.

  • ***

    Women having their first baby.

  • ****

    All births excluding ‘induced’ or ‘no labour’.

  • *****

    Women aged between 20–40 years, non-Indigenous, primiparous giving birth to a singleton live baby at greater than 22 weeks of gestation after a spontaneous onset of labour.

Gestational age at birth* (%)
Extremely preterm0.30.30.30.30.30.30.40.40.40.49.1
Very preterm0.60.60.60.60.60.70.70.60.70.63.2
Moderately preterm5.05.05.35.25.35.55.55.55.55.612.5
Term/post-term94.194.093.893.893.793.493.493.593.493.4−0.7
Maternal characteristics (%)
Non-Indigenous97.096.996.996.896.696.596.596.496.496.4−0.5
Indigenous Australian**2.93.13.13.23.43.53.53.53.63.621.9
Multiple birth1.41.41.41.51.51.51.61.61.71.721.0
Primiparous***39.739.739.840.140.040.741.141.041.141.85.2
Spontaneous onset of labour****68.167.467.064.762.862.161.759.258.157.5−15.5
Selected population*****,**23.022.722.722.021.221.321.620.520.320.8−9.5

The change in the rate of preterm birth after the spontaneous onset of labour for the total population was 18.3% (from 5.7 to 6.7%). (For this rate, the denominator was all women who had a spontaneous onset of labour and gave birth to a live baby at 22–45 weeks of gestation.) Among the selected population of low-risk primiparous women, the preterm rate increased by 10.7% (from 5.6 to 6.2%), and among a selected low-risk population of multiparous women, the percentage increase was 19.2% (from 4.4 to 5.2%) (Figure 1).

image

Figure 1. Time trends in preterm birth among women with spontaneous onset of labour, Australia 1994–03.

Download figure to PowerPoint

The percentage of women giving birth to a live baby using assisted reproduction technology in Australia as a percentage of all women having a live birth of 22 weeks or more gestation has increased from 0.8% in 1994 to 1.9% in 2002 (Table 2). We were unable to show the association between risk of preterm birth and assisted reproduction technology pregnancy because these data were not linked at this stage in Australia except for death data.

Table 2.  Women who gave birth after assisted reproduction technology to at least one liveborn of 22–45 weeks of gestation, Australia, 1994–02
Women who gave birth to a live baby199419951996199719981999200020012002Change 1994–02
  1. The ACD and ANZARD data on assisted reproduction technology are not linked to the NPDC. The Table presents the estimated rate of women who had singleton and multiple live births in Australia after assisted reproduction technology as a proportion of all women having a live birth following assisted reproduction technology.

Women using assisted reproduction technologyn = 1991n = 2226n = 2391n = 2624n = 2796n = 3211n = 3599n = 4045n = 4747 
Singleton birth (%)80.280.280.179.579.279.878.278.279.5−0.8
Multiple birth (%)19.819.819.920.520.820.221.821.820.53.3
As a percentage of all women who had a live birth (%)0.80.91.01.01.11.31.41.61.9137.0

Table 3 shows the adjusted odds of extremely preterm, very preterm and moderately preterm births compared with term and post-term births among women who gave birth to a live baby at greater than or equal to 22 weeks of gestation in the population. The model controlled for maternal age, parity, multiple pregnancy, nonspontaneous onset of labour and Indigenous status; however, we were not able to control for assisted reproduction techniques (women using assisted reproduction technology). Significant contributors to preterm birth included multiple births (AOR 17.82; 95% CI 17.44–18.20), Indigenous status (AOR 2.18; 95% CI 2.14–2.24), being younger than 20 years (AOR 1.16; 95% CI 1.13–1.18) or older than 40 years at the time of giving birth (AOR 1.52; 95% CI 1.46–1.58) and primiparity (AOR 1.27; 95% CI 1.25–1.28) (Table 3).

Table 3.  Adjusted odds of preterm birth among women who had a liveborn baby (22–45 weeks of gestation), 1994–03, Australia
 All pretermExtremely preterm*Very preterm*Moderately preterm*
n = 158 711 (%)AOR (95% CI)n = 8591 (%)AOR (95% CI)n =15959 (%)AOR (95% CI)n =134 161 (%)AOR (95% CI)
  • CI, confidence interval. All odds ratios were mutually adjusted for other variables in the Table. Percentages may not add up to 100 because of missing data. ∼P < 0.001.

  • *

    Extremely preterm, 22–27 completed weeks; very preterm, 28–31 weeks; moderately preterm, 32–36 weeks; term, ≥37 weeks.

  • **

    Odds ratio for a 1-year change in age.

  • ***

    Odds ratio for a 1-year change in calendar time.

Onset of labour
Spontaneous62.01.0028.11.0056.91.0061.91.00
Not spontaneous38.00.94 (0.93–0.95)71.90.59 (0.56–0.62)42.91.14 (1.10–1.17)38.10.95 (0.94–0.96)
Parity
Primiparous45.61.27 (1.25–1.28)44.51.21 (1.16–1.27)47.51.35 (1.31–1.40)45.61.26 (1.24–1.27)
Multiparous54.41.0055.41.0052.51.0054.51.00
Type of pregnancy
Singleton87.91.0086.11.0084.51.0088.41.00
Multiple12.117.82 (17.44–18.20)13.923.24 (21.77–24.80)15.523.04 (21.99–24.15)11.616.93 (16.55–17.31)
Maternal age (years)**
20–4091.61.0090.11.0090.31.0091.81.00
<206.51.16 (1.13–1.18)8.01.33 (1.22–1.45)7.61.38 (1.30–1.47)6.31.12 (1.09–1.14)
>401.91.52 (1.46–1.58)2.01.65 (1.41–1.94)2.11.69 (1.52–1.89)1.91.49 (1.43–1.56)
Year*** 1.01 (1.00–1.01) 1.01 (1.01–1.02) 1.00 (0.99–1.00) 1.00 (1.00–1.01)
Ethnicity
Non-Indigenous94.51.0088.31.0089.71.0090.71.00
Indigenous6.22.18 (2.14–2.24)8.52.88(2.66–3.11)6.82.46 (2.30–2.62)9.92.11 (2.06–2.17)

The population attributable risk percent (PAR) based on the prevalence of the various risk factors in the population was 9.3% for primiparity, 19.6% for multiple births and 1.6 % for nonspontaneous onset of labour (based on the AOR).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

The results of this study confirmed the findings of Langhoff-Roos et al.7 showing an increase in the rate of preterm birth among liveborn infants overall and for the selected population of low-risk women. In comparison, our results showed a less pronounced increase in the rate of preterm birth among low-risk primiparous women (10.7% compared with 51%) and a similar increase in the rate among low-risk multiparous women (19.2% compared with 20%).7 Overall, the rate of preterm birth among both low-risk women and among the Australian population in general is consistently higher than that reported in the Danish study.7 In the BMJ rapid responses to the Danish study,9 various causative factors for preterm birth were proposed; from a history of induced abortion,10 to an increased risk from licorice consumption11 and to the rising rates of Chlamydia.12 What is most apparent is the fact that very little can be done to prevent the rising rates of preterm birth, while the causative factors remain uncertain.5,6

There were several limitations in our study. The data presented on mothers who conceived using assisted reproduction technology were not linked to pregnancy data from the NPDC during the years of the study. It is well reported that both singleton and multiple births after assisted conception have a disproportionate effect on preterm births.4 However, we note that among the 2 million mothers analysed in this study, the percentage of mothers using assisted reproduction technology who gave birth to a live baby as a percentage of all women who gave birth to a live baby over the 9 years was less than 1% from 1994 to 1996 and less than 2% from 1997 to 2003. From the most recent Australasian assisted reproduction technology data for 2004, we learn that 10.9% of singleton births were preterm, and of these, 1.7% were born before 27 weeks of gestation. In addition to this, 44.9% of women using assisted reproduction technology with a singleton baby had a caesarean section for birth.13 Langhoff-Roos et al.7 were able to control for in vitro fertilization (IVF) status; however, Denmark has a higher percentage of births conceived by assisted reproduction technology at 3–9%.14 In addition to this, we were unable to link national population data in terms of congenital anomaly. It is reported that the presence of a birth defect, even one compatible with normal survival, remains a risk factor for preterm birth.4

A further limitation relates to the lack of consistent data on smoking collated within the Australian NPDC. Data from 2003 show that the risk of preterm birth was 60% higher in babies of mothers who smoked during pregnancy than that in babies of mothers who did not smoke (OR 1.61; 95% CI 1.56–1.65).15 Therefore, we can deduce that being able to control for smoking may have lowered the reported rates of preterm birth, although the ‘selected’ population model excluded Indigenous women from the analysis; 52.1% of whom reported smoking during pregnancy compared with 15.8% of non-Indigenous mothers.10

Both the study by Langhoff-Roos et al.7 and our study show a divergence from the widely held perception that the rate of preterm birth is associated almost exclusively with an increase in multiple births and in demographic risk factors of younger and older mothers and those experiencing assisted reproduction technology. We found a similar rising trend among women without these identified risk factors for preterm birth.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

The results of this study contribute to the international discussion proposed by Langhoff-Roos et al.7 in the hope that analysing the differences in proportions of preterm births between countries might increase our understanding of preterm birth. This study has also showed that a rise in preterm birth in Australia is also associated with factors other than those such as multiple births. The rate of preterm birth among low-risk women with a spontaneous onset of labour and singleton pregnancy showed a rise over the 10-year study period.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

S.K.T. is funded through a Population Health Capacity Building Grant from the National Health and Medical Research Council of Australia as a member of the Health Research and Outcomes Network (HERON). The HERON is a collaborative programme of the Institute for Health Research and The University of Sydney, The University of New South Wales, University of Technology Sydney, The Cancer Council NSW and NSW Health using population health data to improve health services, policy and planning. We acknowledge the State and Territory governments and those associated with collection of the perinatal collections for provision of the perinatal data and the IVF and gamete intrafallopian transfer (GIFT) Fertility Centres in Australia and New Zealand and Fertility Society of Australia for provision of assisted conception data for the ACD and ANZARD. We acknowledge the Australian Institute of Health and Welfare (AIHW) for funding the National Pennatal Data Collection.

Ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References

University of New South Wales Human Research and Ethics Committee (UNSW HREC), May 2006.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Ethics approval
  10. References