Discussion
Main findings
The aim of this study was to assess the risk of recurrent preterm birth following spontaneous preterm birth between 16+0-27+6 weeks. We found that, at all gestational ages, patients with previous sPTB from 16 weeks onwards are at high risk for recurrent preterm birth.
A short interpregnancy interval of 0-3 months was associated with an increased risk of subsequent preterm birth < 37 and < 32 weeks. Short intervals were more common in patients with prior sPTB < 24 weeks. Since in the Netherlands no active support is offered to neonates born before 24 weeks, these births inevitably end in perinatal death. Therefore, parents might pursue a subsequent pregnancy shortly after the immature birth.
Interpretation
Multiple studies assessed the association between obstetric history and the risk for subsequent PTB and found higher risks of PTB following recurrent miscarriage or prior PTB < 37 weeks.6,11–18 Limited research assessed the subsequent risk after birth between 16-24 weeks and a cohort comparison is complicated by international differences in terminology and registration. One study by Goldenberg et al. (‘93) found a recurrent PTB rate of 39% in women who gave birth between 13-24 weeks, which increased to 62% if the prior birth was between 19-22 weeks.12 Edlow et. al. (’07) found that women with prior birth between 14-24 weeks were 10.8 times more likely to experience subsequent second-trimester loss or PTB compared to those with previous full-term delivery.13 A third study from Denmark (’17) reported a recurrence rate of 7.3% following birth between 16 and 28 weeks, but this rate varied significantly depending on the characteristics of the previous birth (fetal anomaly, multiple gestation, or intrauterine fetal demise), complicating a comparison with our findings.11
Our results confirm that sPTB between 16-28 weeks is associated with a high risk for subsequent PTB, with an emphasis that the risk is also high in women with prior sPTB between 16+0-19+6 weeks. Women with prior birth at 16+0-19+6 weeks of gestation had a recurrent risk for PTB < 32 and < 37 weeks of respectively 5.8% and 11.7%, which is high compared to 1.0% and 5.5% in a general Dutch population of multiparous women with singleton and multiple pregnancies in 2021 (www.peristat.nl). Therefore, the obstetric history of women with prior birth between 16+0-19+6 weeks might deserve equal consideration in a risk assessment for subsequent PTB. Labelling spontaneous birth at this gestational age range as a miscarriage, may underestimate the risk for subsequent PTB. Using terminology that acknowledges the increased risk, by classifying birth between 16-20 weeks as PTB instead of miscarriage, could enhance the recognition, approach and preventive treatment of patients at risk.
The high recurrent risk after births at low gestational ages raises questions whether the subsequent risk may also be increased after birth at gestational ages below 16 weeks. If so, women with a previous miscarriages just below 16 weeks might be misidentified as patients at at-risk for subsequent PTB. Accurate national registration is vital to assess PTB risk following births at 13-15 weeks. All pregnant women in the Netherlands are advised to contact a midwife or general practitioner before 10 weeks of pregnancy, allowing for precise gestational age determination via ultrasound. Therefore, gestational age at which a (late) miscarriage might occur should be easy to determine. Registering these pregnancy outcomes will help PTB risk evaluation. If an increased PTB risk is found, further research is needed to assess whether and which preventive measures improve subsequent pregnancy outcomes.
Strengths and limitations
This study used data from the perinatal registry in the Netherlands, covering >97% of births.10 The large sample size with data from multiple consecutive years enabled a detailed assessment on subsequent PTB risk by gestational age and allowed for analysis on the interpregnancy interval. However, due to non-mandatory registration for births until 24 weeks, underrepresentation is likely for prior births between 16-24 weeks and also for the recurrence risk in that range.
In the index pregnancy selection, we excluded induced births, focusing on spontaneous and unknown start of labor. Excluding pregnancies complicated by congenital abnormalities or IUFD in the index pregnancy, likely removed inaccurately registered induced deliveries. However, we cannot rule out the possibility that the index pregnancy cohort might still include induced births, potentially underestimating the risk of subsequent PTB after sPTB. Our subsequent cohort lacks distinction for high initial PTB risk (e.g., multiples, congenital issues, IUFD). Therefore, the risk is most likely lower for uncomplicated singletons. Still, our PTB rates remain notably high, even compared to national figures encompassing all pregnancies.
No data were available regarding the use of preventive measures in the subsequent pregnancy. In the Netherlands, patients with previous sPTB < 34 weeks of gestation are typically offered preventive progesterone treatment, additional cervical length screening and potentially receive interventions such as a cervical cerclage.19 Therefore, it is plausible that a significant portion of our study population received preventive treatment in the subsequent pregnancy, which could underestimate the actual risk faced by patients. However, there may be limited awareness regarding the increased risk following PTB around 16 weeks, resulting in fewer or no preventive measures and therefore providing a representative risk estimate for this subgroup.
No core outcome set (COS) could be used in the design of this study because of limited availability of the required outcome measures in the national perinatal registry.
Out of 2,294 women with prior sPTB, we successfully linked 1,285 nulliparous women to a subsequent pregnancy in a primiparous cohort. No linkage could be established in 1009 women, possibly due to insufficient matching variables. For example, if the birth record of the subsequent pregnancy did not include the date of the prior birth and if ZIP code changed over time, there would insufficient matching variables to establish a linkage. Other reasons could include no subsequent pregnancy within the 5-year timeframe, cases where the only pregnancy within the 5-year timeframe resulted in a miscarriage or termination before 16 weeks, or misreported subsequent births as nulliparous births. Also, 172 patients from the index cohort were excluded due to the antenatal diagnosis of IUFD, which might involve cases of IUFD due to fetal distress from extreme preterm labor. Given its likelier occurrence before 24 weeks, the group of patients with prior birth between 16-24 weeks may not be entirely represented.