Risk of preterm birth

While there are many known and unknown causes of preterm birth, women identified as being at risk may benefit from advice about risk and protective factors.

23.1 Background

Preterm birth is defined as birth before 37 completed weeks of pregnancy WHO 2012. Sub-categories of preterm birth are based on weeks of gestational age: early preterm (<34 weeks), very preterm (28 to <32 weeks) and extremely preterm (<28 weeks). This section is concerned with spontaneous preterm birth as opposed to planned preterm birth.

23.1.1 Incidence of preterm birth

In Australia in 2014 AIHW 2016:

• overall, 8.6% of babies were born preterm, with most of these births occurring at gestational ages between 32 and 36 completed weeks
• the average gestational age for all preterm births was 33.3 weeks
• babies whose mothers smoked during pregnancy were more likely to be born preterm (13%) than those whose mothers did not smoke during pregnancy (8%).
• other characteristics associated with increased likelihood of preterm birth included:
  • babies born in multiple births: 63% of twins and all (100%) of other multiples (triplets and higher) were preterm, compared with 7% of singleton babies
  • babies born to mothers usually residing in more remote areas: 13% in very remote areas compared with 8% in major cities
  • babies of younger (<20 years) and older (≥40 years) mothers: 11% and 12% were preterm, compared with 8% of babies with mothers aged 20–39 years.

Nationally in 2014, approximately 14% of babies of Indigenous mothers were born preterm, compared with 8% of babies of non-Indigenous mothers AIHW 2016; similar rates were found in an earlier West Australian study (14.8 and 7.6%) Langridge et al 2010. However, a study in a Melbourne hospital found no significant difference in risk of preterm birth between Indigenous and non-Indigenous babies and mothers (Indigenous babies aOR 1.19, 95%CI 0.77 to 1.87, Indigenous mothers aOR 0.97 95%CI 0.52 to 1.80) Whish-Wilson et al 2016.

23.1.2 Risks associated with preterm birth

Preterm birth is associated with perinatal mortality, long-term neurological disability (including cerebral palsy), admission to neonatal intensive care, severe morbidity in the first weeks of life, prolonged hospital stay after birth, readmission to hospital in the first year of life and increased risk of chronic lung disease WHO 2012. Preterm birth can have a serious emotional impact on the family. In Australia in 2014 AIHW 2016:

• preterm babies were more likely to be admitted to a special care nursery or neonatal intensive care unit (72%) than babies born at term (10%) or post-term (13%)
• spontaneous preterm birth accounted for 14% of all perinatal deaths and one third (33%) of perinatal deaths of babies of Indigenous mothers.

23.2 Identifying women at increased risk of giving birth preterm

A range of risk and protective factors influence the likelihood of preterm birth. While many risk factors are not modifiable during a woman’s current pregnancy, addressing modifiable risk factors may reduce risk of preterm birth. It should also be noted that many women who experience preterm birth have no risk factors.

23.2.1 Significant risk factors

There is a significant association between preterm birth and:

• social disadvantage (OR 1.27, 95%CI: 1.16 to 1.39) Ncube et al 2016 and lower levels of maternal education (RR 1.48; 95%CI 1.29 to 1.69) Ruiz et al 2015
• previous preterm birth (absolute recurrence rate among women with a singleton pregnancy and previous preterm singleton birth 20%, 95% CI 19.9–20.6) Kazemier et al 2014
• pre-existing (p=0.002) Kock et al 2010 or gestational diabetes AIHW 2010
• current urogenital infections: eg chlamydia [OR 1.60; 90%CI 1.01 to 2.5] John Hopkins Study Team 1989, bacterial vaginosis [OR 1.85; 95%CI 1.62 to 2.11] Flynn et al 1999
• alcohol consumption (OR 1.34; 95%CI 1.28 to 1.41) Aliyu et al 2010, in a dose-response fashion Sokol et al 2007, Patra et al 2011
• smoking at the first antenatal visit (aOR 1.42, 95%CI 1.27 to 1.59) Bickerstaff et al 2012 and active smoking during pregnancy (aOR 1.53, 95%CI 1.05 to 2.21) Fantuzzi et al 2007, with risk further increased among women smoking more than 10 cigarettes a day compared to those smoking 1–9 cigarettes per day (aOR 1.69 vs 1.54) Fantuzzi et al 2007.

23.2.2 Other factors

Systematic reviews of RCTs found:

• women who were overweight and obese who participated in aerobic exercise for 30–60 minutes three to seven times per week had a lower risk of preterm birth <37weeks (RR 0.62, 95% CI 0.41 to 0.95) compared to controls Magro-Malosso et al 2016
• no clear difference in risk of preterm birth <37 weeks with treatment of periodontal disease (RR 0.87; 95%CI 0.70 to 1.10; low quality evidence) Iheozor-Ejiofor et al 2017.

Systematic reviews of observational studies show the following associations with preterm birth:

• country of origin/ethnicity: odds of very preterm birth among East African immigrants were higher than among Australian-born women (aOR 1.55, 95%CI 1.27 to 1.90) Belihu et al 2016 and higher among African American women than among Caucasian women (pooled OR 2.0; 95%CI 1.8 to 2.2), with no significant association for Asian or Hispanic ethnicity Schaaf et al 2013
• weight: risk was increased among women who were obese and gained more than the IOM recommendations (aOR 1.54; 95% CI 1.09 to 2.16) Faucher et al 2016
• emotional health and wellbeing: increased risk was associated with low social support compared to high social support (OR 1.22, 95%CI 0.84 to 1.76); stress (OR 1.52, 95%CI 1.18, to 1.97) Hetherington et al 2015; untreated depression (OR 1.56; 95%CI 1.25 to 1.94) Jarde et al 2016 and anxiety (RR 1.50, 95%CI 1.33 to 1.70) Ding et al 2014, (OR 1.70, 95%CI 1.33 to 2.18) Rose et al 2016 but not with maternal personality traits Chatzi et al 2013
• exposure to antidepressants: risk was increased among women exposed to antidepressants during pregnancy compared to women with depression but without antidepressant exposure (OR 1.17, 95%CI 1.10 to 1.25) Eke et al 2016, (RR 2.85, 95%CI 2.00 to 4.07) Huang et al 2014a; and risk was significantly increased with exposure in the third trimester (aOR 1.96, 95%CI 1.62 to 2.38) but not in the first trimester (aOR 1.16, 95%CI 0.92 to 1.45) Huybrechts et al 2014
• environmental factors: increased risk was associated with high environmental temperature Beltran et al 2013, especially heat stress Carolan-Olah & Frankowska 2014; exposure to passive smoke in any place (OR 1.20, 95%CI 1.07 to 1.34) or at home (OR 1.16, 95%CI 1.04 to 1.30) Cui et al 2016; risk associated with exposure to fine particulate matter was unclear due to significant heterogeneity between studies Sun et al 2015
• pre-existing conditions: risk of preterm birth was increased among women with hepatitis C (OR 1.62, 95%CI 1.48 to 1.76, P < 0.001) Huang et al 2015, human papilloma virus (OR 2.12, 95%CI 1.51 to 2.98, P<0.001) Huang et al 2014c, hypothyroidism (OR 1.19, 95%CI 1.12 to 1.26; P < 0.00001) and hyperthyroidism (OR, 1.24, 95%, CI 1.17- 1.31; P < .00001) Sheehan et al 2015 but not hepatitis B (OR 1.12, 95%CI 0.94 to 1.33) Huang et al 2014b.
• lifestyle factors: incidence of preterm birth (4.5% vs 4.4%; RR 1.01, 95%CI 0.68 to 1.50) was similar among women in the normal BMI category undertaking aerobic exercise during pregnancy and controls Di Mascio et al 2016; risk was increased among women with serum vitamin D levels lower than 50 nmol/L (OR 1.29, 95%CI 1.16 to 1.45) Qin et al 2016; and there was no clear or statistically significant relationship between preterm birth and shift work van Melick et al 2014, multivitamin use Johnston et al 2016 or influenza vaccination during pregnancy Fell et al 2015
• history of gynaecological procedures: risk was increased among women with a history of dilatation and curettage (D&C) (OR 1.29, 95% CI 1.17 to 1.42) or multiple D&Cs (OR 1.74, 95%CI 1.10 to 2.76) Lemmers et al 2016; surgically induced termination of pregnancy (OR 1.52, 95%CI 1.08 to 2.16); surgically managed miscarriage (OR 1.19, 95%CI 1.03 to 1.37) Saccone et al 2016; loop electrosurgical excision procedure compared to women with no history of cervical dysplasia (pooled RR 1.61, 95%CI 1.35 to 1.92) but not when compared to women with a history of cervical dysplasia but no cervical excision (pooled RR 1.08, 95%CI 0.88 to 1.33) Conner et al 2014; and treatment for cervical intraepithelial neoplasia before (OR 1.4, 95%CI 0.85 to 2.3) or during pregnancy (OR 6.5, 95%CI 1.1 to 37) Danhof et al 2015.

23.3 Prediction and prevention

23.3.1 Cervical length measurement

Cervical length measurement is an evolving area. There is currently uncertainty about the thresholds that identify high risk and the precision of measurement, particularly with transabdominal cervical length measurement.

Factors associated with short cervical length 

Factors associated with women being at increased risk of short cervical length at mid trimester include:

• previous spontaneous (Cho et al 2017) or induced (Miller et al 2015) preterm birth
• previous cervical excisional procedure (Cho et al 2017; Miller et al 2015)
• non-Caucasian ethnicity (Buck et al 2016; Miller et al 2015; van der Ven et al 2015).

Studies were consistent in finding no significant association between women’s short cervical length and maternal height (Cho et al 2017; van der Ven et al 2015) or assisted conception (Aboulghar et al 2009; Miller et al 2015). The evidence on the association between cervical length and maternal weight (Cho et al 2017; Kandil et al 2017; Palma-Dias et al 2004) or age (D'Agostini et al 2013; Miller et al 2015; van der Ven et al 2015) was inconsistent.

Accuracy of transvaginal cervical length measurement

Measuring women’s cervical length at the 18-20 week ultrasound using a threshold of 25 mm has the potential to predict preterm birth (Crane & Hutchens 2008; Domin et al 2010; Honest et al 2012) but is more accurate when combined with assessment of relevant maternal factors (eg nulliparity, previous preterm birth) (Celik et al 2008). No evidence on harms associated with cervical length measurement was identified.

Accuracy of transabdominal cervical length measurement

While transabdominal measurement of cervical length has been used as an initial screen prior to transvaginal ultrasound (Cho & Roh 2016; Friedman et al 2013a; Friedman et al 2013b; Kongwattanakul et al 2016; Saul et al 2008), the cervix may not be adequately visualised by this method in as many as 60% of women (Friedman et al 2013a; Friedman et al 2013b). The evidence on transabdominal ultrasound is inconsistent in terms of gestational age and cut-offs and universal transvaginal ultrasound is more cost-effective than including an initial transabdominal measurement (Miller & Grobman 2013). 

Universal versus targeted cervical length screening

In settings where universal transvaginal screening of women’s cervical length has been implemented:

• there has been a reduction in preterm births <37 weeks (aOR 0.82; 95% confidence interval [CI] 0.76 to 0.88), <34 weeks (aOR 0.74; 95%CI 0.64 to 0.85) and <32 weeks (aOR 0.74; 95%CI 0.62 to 0.90), with similar effect sizes in nulliparous and multiparous women with previous term births (Son et al 2016)
• in a United States study, women were more likely to decline cervical length screening if they were African American (aOR 2.17; 95%CI 1.93 to 2.44), obese (aOR 1.18; 95%CI 1.06 to 1.31), multiparous (aOR 1.45; 95%CI 1.29 to 1.64), aged <35 years (aOR 1.24; 95%CI 1.08 to 1.43), or smokers (aOR 1.42; 95%CI 1.20 to 1.68) (Temming et al 2016)
• after 6 months of implementation, there was no change in rates of acceptance of cervical length screening and rates of spontaneous preterm birth <28 weeks were higher among women who declined screening (aOR 2.01; 95%CI 1.33 to 3.02) (Temming et al 2016).

A study that calculated the number of women needed to screen (NNS) to prevent one early preterm birth found that, with a cut-off of ≤15 mm, the NNS in low-risk women would be 1,075 compared to 344 among nulliparous women and 167 among women with a previous preterm birth. At a cut-off of ≤20 mm, NNSs were 802, 221 and 97, respectively (Facco & Simhan 2013).

Another study found that targeted screening increased specificity compared to universal screening but sensitivity was reduced and nearly 40% of women with a short cervix were not identified (Miller et al 2015). 

Timing of transvaginal cervical length measurement among women at high risk of preterm birth

Evidence from observational studies suggests that cervical length measurement earlier than 20 weeks may predict cervical shortening and risk of early preterm birth in women at high risk of preterm birth (Banicevic et al 2014; Souka et al 2011; Vaisbuch et al 2010). However, a cervical length >25 mm does not preclude preterm birth in these women, with 16% to 21% experiencing preterm birth before 37 weeks (Caradeux et al 2017; Care et al 2014; Owen et al 2010). 

Implementation of universal transvaginal cervical length measurement

Studies in the United States and the United Kingdom suggest that universal cervical length measurement is cost-effective when women with a short cervix (≤25 mm) at 18-25 weeks receive vaginal progesterone (Cahill et al 2010; Crosby et al 2016; Einerson et al 2016; Jain et al 2016; Werner et al 2011). No Australian cost-effectiveness studies were identified.
Transvaginal measurement of cervical length does not significantly increase the time for completion of ultrasound examination and attitudes regarding discomfort or embarrassment did not differ between women who underwent no cervical length screening or transvaginal or transabdominal screening (Romero et al 2014).

However, a number of barriers may prevent or restrict the implementation of a universal cervical length screening program — cost, availability of vaginal progesterone and other treatment options, women’s acceptance of, or willingness to undergo, transvaginal ultrasound and the perceptions and beliefs of health professionals (Pedretti et al 2017). Issues of access to ultrasound services (eg due to remote location or language barriers) and availability of accredited trained professionals in some areas may also limit the availability of cervical measurement.

23.3.2 Holistic preventive strategies

Systematic reviews that evaluated holistic models of care and their effect on preterm birth found:

• a significant effect in reducing risk of preterm birth among women receiving midwifery-led care compared to other models of care for childbearing women and their infants (average RR 0.76, 95%CI 0.64 to 0.91; n=13,238; 8 studies; high quality) (Sandall et al 2016)
• no significant difference among:
  • women receiving group antenatal care compared to those receiving standard care (RR 0.87, 95%CI 0.70 to 1.09; 11 studies) (Carter et al 2016) and (RR 0.75, 95%CI 0.57 to 1.00; 3 3 studies; n=1,888, moderate quality) (Catling et al 2015)
  • women randomised to specialist preterm birth programs compared to those receiving standard care (RR 0.92, 95%CI 0.76 to 1.12; 15 RCTs) (Fernandez Turienzo et al 2016)
  • low risk women receiving a reduced number of antenatal visits (RR 1.02, 95%CI 0.94 to 1.11; 7 studies, n=53,661, moderate quality) (Dowswell et al 2015)
  • women receiving additional social support compared to those receiving standard care (RR 0.92, 95%CI 0.83 to 1.01; 11 RCTs; n=10,429) (Hodnett et al 2010), including adolescent women (RR 0.67; 95%CI 0.42 to 1.05; 4 studies; n=684) (Sukhato et al 2015)
  • women receiving telephone support during pregnancy compared to women receiving routine care or other support (RR 0.91, 95%CI 0.77 to 1.08, 4 RCTs; n=3,992) (Lavender et al 2013)
  • women in preterm labour using relaxation techniques compared to those not using relaxation techniques (RR 0.95; 95%CI 0.57 to 1.59; 11 RCTs; n=833) (Khianman et al 2012)
• successful approaches to increasing access to antenatal care and reducing preterm birth among Aboriginal and Torres Strait Islander women include community-based collaborative antenatal care and community-based support (Rumbold & Cunningham 2008) and partnership between Aboriginal grandmothers, Aboriginal Health Officers, midwives and existing antenatal care services (Bertilone & McEvoy 2015).

23.3.3 Progesterone treatment for women with a short cervix

A systematic review analysed the effectiveness of progesterone compared to placebo in women with short cervical length (without other risk factors for preterm birth or premature onset of labour). It found that, while preterm birth <34 weeks, <37 weeks and neonatal deaths were reduced in women overall, there was only a reduction of preterm birth <34 weeks in women with a short cervix (Jarde et al 2017). 

When studies specific to vaginal progesterone treatment in women with a short cervix were analysed separately, there were statistically significant effects on preterm birth <35 weeks (risk ratio [RR] 0.62; 95%CI 0.42 to 0.92; 1 RCT, n=458, moderate certainty) (Hassan et al 2011), preterm birth <34 weeks (RR 0.60; 95%CI 0.41 to 0.89; 2 RCTs, n=330, moderate certainty) (Fonseca et al 2007; van Os et al 2015), preterm birth <28 weeks (RR 0.55; 95%CI 0.25 to 0.97; 1 RCT, n=458, moderate certainty) (Hassan et al 2011) and respiratory distress syndrome (RR 0.51; 95%CI 0.31 to 0.86; 3 RCTs; n=788, moderate certainty) (Fonseca et al 2007; Hassan et al 2011; van Os et al 2015). Definitions of short cervix ranged from 10–20 mm to ≤30 mm. Treatment was commenced at 20 to 24 weeks gestation and was continued until 34–37 weeks. 

There were no statistically significant effects on preterm birth associated with intramuscular progesterone compared with placebo in women with a short cervix (1 RCT; n=657, low certainty) (Grobman et al 2012). 

A modelling study in the United Kingdom found universal cervical length screening and vaginal progesterone for women with a cervical length of ≤15 mm was cost-effective (Crosby et al 2016).

23.4 Discussing risk of giving birth preterm

When risk of preterm birth is increased, modifiable risk factors should be addressed Freak-Poli et al 2009, Kiran et al 2010, Carter et al 2011. Based on the evidence discussed in Section 23.2, discussion with women at risk of preterm birth can include the benefits of:

• having adequate social and emotional support
• quitting tobacco smoking and avoiding exposure to passive smoke
• not drinking alcohol during pregnancy
• having tests for urogenital infections
• participating in regular exercise, particularly if they are overweight or obese.

Women can also be advised that risk is not reduced by supplementing with Vitamins C or E Rumbold et al 2015a, Rumbold et al 2015b or probiotics Othman et al 2007, Hauth et al 2010.

A Cochrane review found no evidence to support or refute bed rest for prevention of preterm birth Sosa et al 2015. A subsequent cohort study found that, among women at high risk of preterm birth, activity restriction was associated with increased risk of preterm birth Levin et al 2017.

23.5 Practice summary: risk of preterm birth

When

A woman has identified risk factors for giving birth preterm

Who

• Midwife
• GP
• obstetrician
• Aboriginal and Torres Strait Islander Health Practitioner
• Aboriginal and Torres Strait Islander Health Worker
• multicultural health worker

What

Discuss lifestyle factors associated with preterm birth

• Explain that smoking during pregnancy makes it more likely that the baby will be born preterm and also causes other serious risks to the pregnancy.
• Explain that not drinking alcohol during pregnancy is the safest option.
• Offer testing for urogenital infection if the woman has risk factors for preterm birth. If results are positive, consider counselling, contact tracing, partner testing and treatment, and repeat testing.

Discuss protective factors

• Explain that moderate physical activity during pregnancy has a range of health benefits, particularly for women who are overweight or obese.

Take a holistic approach

• Provide information on relevant community supports (eg smoking cessation programs, drug and alcohol services, physical activity groups).
• Consider whether a woman may be at increased risk if she has recently arrived from a country with a high prevalence of preterm birth.
• Provide social and emotional support and access to continuity of carer, where possible