Zhang et al.

BMC Pregnancy and Childbirth (2020) 20:490

https://doi.org/10.1186/s12884-020-03165-7

RESEARCH ARTICLE Open Access

Prediction of iatrogenic preterm birth in

patients with scarred uterus: a retrospective
cohort study in Northeast China
Liyang Zhang1, Hongtian Li2, Jiapo Li1, Yue Hou1, Buxuan Xu3, Na Li1, Tian Yang1, Caixia Liu1 and Chong Qiao1*

Abstract
Background: To build a novel and simple model to predict iatrogenic preterm birth in pregnant women with
scarred uteri.
Methods: In this retrospective, observational, single-centre cohort study, data from 2315 patients with scarred uteri
were collected. Multiple logistic regression analysis and mathematical modelling were used to develop a risk
evaluation tool for iatrogenic preterm birth. After modelling, the calibration and discrimination of the model along
with decision curve analysis were checked and performed to ensure clinical applicability.
Results: Among the 2315 patients, 417 (18.0%) had iatrogenic preterm births. The following variables were included
in the model: interpregnancy interval (0 to < 12 months, OR 5.33 (95% Cl 1.79–15.91), P = 0.003; 13 to < 24 months
(reference), 25 to < 60 months, OR 1.80 (95% CI 0.96–3.40), P = 0.068; ≥ 60 months, OR 1.60 (95% Cl 0.86–2.97), P =
0.14), height (OR 0.95, (95% CI 0.92–0.98), P = 0.003), parity (parity ≤1 (reference), parity = 2, OR 2.92 (95% CI 1.71–
4.96), P < 0.0001; parity ≥3, OR 8.26, (95% CI 2.29–29.76), P = 0.001), number of vaginal bleeding (OR 1.81, (95% Cl
1.36–2.41), P < 0.0001), hypertension in pregnancy (OR 9.52 (95% CI 6.46–14.03), P < 0.0001), and placenta previa (OR
4.21, (95% CI 2.85–6.22), P < 0.0001). Finally, a nomogram was developed.
Conclusions: In this study, we built a model to predict iatrogenic preterm birth for pregnant women with scarred
uteri. The nomogram we created can assist doctors in evaluating the risk of iatrogenic preterm birth and help in
making referrals; thus, better medical care can be given to improve the prognosis of patients and foetuses.
Keywords: Iatrogenic preterm birth, Scarred uterus, Prediction model

Background history of previous adverse events, maternal age [2],

Preterm birth (PTB) remains a major unsolved problem socio-economic factors [3], maternal obesity [4], pla-
in modern obstetrics and is significantly associated with centa previa [5], multiple gestations [6], cervical length
infant mortality, long-term morbidity and neurodevelop- [7], and other biomarkers [8–11]. However, most studies
mental impairment. The prevalence of preterm birth is have been specifically restricted to spontaneous preterm
5–18% of all live births worldwide [1]. To date, many birth. Over the past 24 years, spontaneous preterm birth
studies have been performed regarding the prediction of has declined by 25%, while the incidence of iatrogenic
preterm birth. The risk factors include an obstetric preterm birth has increased, representing nearly 30% of
all preterm births [12]. Efforts to study iatrogenic pre-
term birth have been ignored.
* Correspondence: [email protected]
1
Department of Obstetrics and Gynecology, Shengjing Hospital, China
China has one of the highest caesarean section rates in
Medical University, Shenyang, Liaoning Province, China the world. With new family planning policies emerging,
Full list of author information is available at the end of the article

© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the
data made available in this article, unless otherwise stated in a credit line to the data.
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 2 of 9

an increasing number of women have decided to have a maternal or foetal medical concerns; cases such as this
second child, which makes pregnancy with a scarred were still considered spontaneous preterm births. Deliv-
uterus an increasingly prominent problem. Pregnancy ery of the baby according to indications and decided by
with a scarred uterus has been considered risky and obstetricians was considered an iatrogenic preterm birth.
closely related to adverse pregnancy outcomes [13]. Methods of termination included CS and TOLAC.
Those who end up with iatrogenic preterm birth are at The other variables included in our study were as fol-
much higher risk of having the pregnancy terminated if lows: maternal age, height, weight before delivery, parity,
indications allow. At present, TOLAC (trial of labour interpregnancy interval (IPI), number of vaginal bleeding
after caesarean) is not a common practice in China; during the pregnancy, foetal position, myomectomy, dys-
most patients with scarred uteri have the pregnancy ter- menorrhea, regularity of prenatal examination, concep-
minated by caesarean section. Operation techniques and tion method, placenta previa, hypertension during
medical levels in rural areas still lag those in urban areas. pregnancy, and gestational diabetes mellitus. These data
In the hierarchical medical system of China, the evalu- were all collected from a computerized medical record
ation of high-risk patients is not precise; referral is not system including all data concerning basic individual in-
timely, and incorrect referral frequently occurs. Patients formation, medical and obstetric histories, and preg-
who will ultimately have iatrogenic preterm birth often nancy outcomes. Both maternal age and height were
fail to receive good medical care. Thus, we developed a collected at the first prenatal visit. Regarding weight, the
novel, simple prediction model of iatrogenic preterm data collected in our study are the weights before deliv-
birth for pregnant women with scarred uteri using a ery because of the loss of data on weight at the first pre-
Chinese patient database to precisely evaluate the risk of natal visit. Hypertension during pregnancy was defined
preterm birth and help in making referrals, which will according to the American College of Obstetrics and
further benefit the prognosis of pregnant women and Gynecology (ACOG) criteria [14]. Placenta previa was
foetuses. defined as a placenta overlying the internal cervical os
[15]. Gestational diabetes mellitus (GDM) was defined as
Methods “the type of glucose intolerance that develops in the sec-
Study design and participants ond and third trimesters of pregnancy, resulting in
This study included data from a large retrospective co- hyperglycemia of variable severity” [16] and was diag-
hort study in Northeast China from 2014 to 2017 at nosed used an oral glucose tolerance test (OGTT) be-
Shengjing Hospital, a regional tertiary medical centre. tween 24 and 28 weeks gestation by the International
The study was approved by the local ethics committee Association of the Diabetes and Pregnancy Study Groups
(ethics committee of China Medical University) in Shen- (IADPSG) criteria. Number of vaginal bleeding during
yang. The original cohort study recruited 8697 patients pregnancy were self-reported and defined as the total
with scarred uterus from all 69,931 deliveries made dur- number of vaginal bleeding experiences during the
ing the study period to evaluate the impact of scarred whole pregnancy.
uterus on the subsequent pregnancies. The inclusion cri-
teria for this study were all women with a scarred uterus Establishment of the model
from caesarean section or myomectomy AND singleton A total of 2315 cases were randomly split into a training
pregnancies at gestational week 20 or beyond. (including (n = 1566) and validation set (n = 749). Distributions of
cases of stillbirth). The patients were excluded if they continuous variables were assessed for normality using
had severe systematic disease, had a twin or greater the Kolmogorov-Smirnov test; none of the continuous
pregnancy, were unable to speak Chinese, or had no ac- variables were normally distributed in this study. Cat-
cess to a telephone. Finally, 2315 patients remained after egorical variables are presented as percentages, and con-
exclusion. tinuous variables are presented as the median (25 and
75% quantiles). Bivariate analyses were performed by the
Variable assessments Mann-Whitney U test or Fisher’s exact test.
Preterm deliveries are those that occur at less than 37 Simple and multiple logistic regression analyses were
weeks of gestational age. Our study recruited pregnant used to model risk factors for preterm birth. The
women who went into labour between the 20th and interaction between variables may lead to differences in
42nd weeks of gestation. Diagnostic records and actual the results of univariable and multivariable analysis. To
labour weeks were checked to determine whether the avoid missing important risk factors, variables with P <
patient had iatrogenic preterm birth. Patients with spon- 0.2 in the univariable analysis were included in the mul-
taneous preterm birth were also included in this study tivariable regression models using a forward stepwise al-
and treated as controls. A patient could start labour be- gorithm. Finally, 6 variables were included in our model:
fore 37 weeks of gestation and finish with CS due to height, parity, number of vaginal bleeding during
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 3 of 9

pregnancy, IPI, placenta previa, and hypertension during Multiple logistic regression model
pregnancy. In addition, odds ratios (ORs) and 95% confi- After multivariable analysis, parity, interpregnancy inter-
dence intervals (CIs) were calculated. The level of sig- val, hypertension in pregnancy, placenta previa, height
nificance for the P value was set as 0.05. and number of vaginal bleeding during pregnancy were
Other important assessments are outlined here. included in the final model. Among these factors, pa-
Discriminative ability was assessed using the AUC c- tients with hypertension during pregnancy had the high-
statistic. The Hosmer-Lemeshow test was used to deter- est risk of iatrogenic preterm birth (OR = 9.52, 95% CI
mine the adequacy of calibration, and a calibration plot 6.46–14.03), followed by a parity of more than 2 (OR =
was drawn. To evaluate multicollinearity, the variance 8.26, 95% CI 2.29–29.76) and an interpregnancy interval
inflation factor, tolerance, eigenvalue, and condition of 0 to < 12 months (OR = 5.33, 95% CI 1.79–15.91). In-
index were checked. Decision curve analysis was used to creased height could protect patients from iatrogenic
determine the clinical practicability. Finally, a nomogram preterm birth. Further details are shown in Table 2. The
was developed. VIF (variance inflation factor) of all the variables was ap-
proximately equal to 1 in our study, indicating low mul-
Statistical analysis ticollinearity in this model.
SPSS version 25.0 (IBM Corp, Armonk, NY, USA) was
used for statistical modelling. STATA Release 12 Soft- Discrimination and calibration of the model
ware (StataCorp, College Station, TX) was used to per- As mentioned in the Materials and Methods section, dis-
form the Hosmer-Lemeshow test and decision curve crimination of the model was assessed by the AUC c-
analysis. R (R Core Development Team) version 3.1.1 statistic. The AUC was 0.772 (95% Cl 0.739–0.804) in
was used to develop calibration tests and the nomogram. the training set and 0.776 (95% Cl 0.728–0.823) in the
whole validation set. The suggested cutoff for this model
Results was 0.15, and the false positive and false negative rates
Baseline characteristics for the given cutoff were 29.1 and 21.6%, respectively.
Of the 2315 enrolled patients, 417 (18.0%) had an iatro- The calibration plots for the two sets are shown in Fig. 2.
genic preterm birth, and 160 (6.9%) had a spontaneous The training set and validation set were well calibrated.
preterm birth. The median age was 32 (IQR 30–35) The p values from the Hosmer-Lemeshow test were
years. The median height was 162 (IQR 160–165) cm. 0.597 and 0.907 in the training and validation sets, re-
The majority of those enrolled had not experienced vagi- spectively. Overall, the model showed good discrimin-
nal bleeding during pregnancy (n = 2057, 88.9%) and had ation and calibration for both sets, and a nomogram was
less than 2 periods of labour (n = 2171, 93.8%). Over half developed (Fig. 3).
(n = 1212, 52.4%) of the patients had an IPI of more than
60 months, and the fewest proportion of patients hand Decision curve analysis
an IPI of 0 to < 12 months (n = 35, 1.5%). Hypertension Finally, to justify the clinical usefulness of the model, we
in pregnancy was reported for 9.1% of the patients, while assessed whether nomogram-assisted decisions would
placenta previa was present in 14.1% of all patients. Fur- improve patient outcomes by performing decision curve
ther details of the population characteristics are shown analysis. Decision curves can help us calculate the net
in Table 1. benefit of the use of our nomogram. The results of the
decision curve analysis are shown in Fig. 4. The decision
Maternal and foetal indications of iatrogenic preterm curve indicates that if the threshold probability for the
birth patient or doctor is between 15 and 60%, the use of our
Among the 417 patients who had iatrogenic preterm nomogram to predict iatrogenic preterm birth adds
birth, 318 (76.3%) were attributed to maternal reasons, more benefit than either a treat-all-patients scheme or a
43 (10.3%) were attributed to foetal reasons, and the treat-none scheme.
remaining 56 (13.4%) were associated with both mater-
nal and foetal reasons. Among the maternal indications, Discussion
placenta previa and hypertension during pregnancy In this study, six parameters were selected discreetly in
accounted were the most common reasons for iatrogenic the estimation of the overall risk of iatrogenic preterm
preterm birth. Other common indications included a birth: shorter maternal height, extremely low or ad-
threat of uterine rupture, GDM, and oligohydramnios. vanced IPI, greater vaginal number of vaginal bleeding
The maternal index of iatrogenic preterm birth (IPTB) is during pregnancy, higher parity, hypertension during
shown in Fig. 1. Meanwhile, the most common foetal in- pregnancy and placenta previa. By combining these fac-
dications were foetal distress (n = 41, 41.4%) and abnor- tors, the risk of iatrogenic preterm birth can be well
mal foetal position (n = 15, 15.2%). predicted.
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 4 of 9

Table 1 Characteristics of patients with or without iatrogenic preterm birth and univariable analysis in both datasets
Characteristic Training Dataset Validation Dataset
IPTB = 0(n = 1283) IPTB = 1(n = 283) P IPTB = 0(n = 615) IPTB = 1(n = 134)
Age, No. (%) 0.866*
≤ 35 991 (82.04) 217 (17.96) 477 (82.10) 104 (17.90)
>35 292 (81.56) 66 (18.44) 318 (82.14) 30 (17.86)
Parity, No. (%)
≤1 (reference) 1219 (83.04) 249 (16.96) 585 (83.21) 118 (16.79)
2 59 (67.82) 28 (32.18) < 0.0001* 28 (65.12) 15 (34.88)
≥3 5 (45.45) 6 (54.55) 0.004* 2 (66.67) 1 (33.3)
Myomectomy, No. (%) 0.866
No 1249 (81.90) 276 (18.10) 598 (81.98) 132 (18.02)
Yes 34 (82.93) 7 (17.07) 17 (89.47) 2 (10.53)
Dysmenorrhea, No. (%) 0.583
No 950 (81.62) 214 (18.38) 463 (82.53) 98 (17.47)
Yes 333 (82.84) 69 (17.16) 152 (80.85) 36 (19.15)
IPI, No. (%)
13to < 24 months (reference) 133 (90.48) 14 (9.52) 0.002* 54 (85.71) 9 (14.29)
0 to < 12 months 15 (65.22) 8 (34.78) 9 (75.00) 3 (25.00)
25 to < 60 months 489 (83.02) 100 (16.98) 0.028* 234 (86.99) 35 (13.01)
≥ 60 months 646 (80.05) 161 (19.95) 0.003* 318 (78.52) 87 (21.48)
Conception method, No.(%) 0.746
Normal 1276 (81.95) 281 (18.05) 614 (82.31) 132 (17.69)
IVF-ET 7 (77.78) 2 (22.22) 1 (33.33) 2 (66.67)
number of vaginal bleeding during pregnancy, No. (%) < 0.0001*
0 1182 (84.13) 223 (15.87) 559 (85.74) 93 (14.26)
1 85 (70.83) 35 (29.17) 44 (66.67) 22 (33.33)
2 13 (44.83) 16 (55.17) 9 (40.91) 13 (59.09)
≥3 3 (25.00) 9 (75.00) 3 (33.33) 6 (66.67)
Regularity of prenatal examination, No. (%) 0.024*
Yes 984 (83.18) 199 (16.82) 470 (82.75) 98 (17.25)
No 299 (78.07) 84 (21.93) 145 (80.11) 36 (19.89)
Foetal position, No. (%)
Cephalic presentation (reference) 1200 (83.22) 242 (16.78) 576 (83.24) 116 (16.76)
Breech presentation 68 (68.69) 31 (31.31) < 0.0001* 30 (83.33) 6 (16.67)
Transverse presentation 15 (60.00) 10 (40.00) 0.004* 9 (42.86) 12 (57.14)
GDM, No. (%) 0.412
No 1034 (82.32) 222 (17.68) 530 (82.94) 109 (17.06)
Yes 249 (80.32) 61 (19.68) 85 (77.27) 25 (22.73)
Hypertension in pregnancy, No. (%) < 0.0001*
No 1215 (85.56) 205 (14.44) 581 (84.94) 103 (15.06)
Yes 68 (46.58) 78 (53.42) 34 (52.31) 31 (47.69)
Placental previa, No. (%) < 0.0001*
No 1159 (85.54) 196 (14.46) 556 (87.70) 78 (12.30)
Yes 124 (58.77) 87 (41.23) 59 (51.30) 56 (48.70)
Height (interquartile range) 162 (160,165) 161 (160,165) 0.018* 162 (160.165) 162 (160.165)
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 5 of 9

Table 1 Characteristics of patients with or without iatrogenic preterm birth and univariable analysis in both datasets (Continued)
Characteristic Training Dataset Validation Dataset
IPTB = 0(n = 1283) IPTB = 1(n = 283) P IPTB = 0(n = 615) IPTB = 1(n = 134)
Weight (interquartile range) 72 (66,79) 70 (65,80) 0.585 71 (65,78) 70 (65,78)
Abbreviations: IPI interpregnancy interval, IPTB iatrogenic preterm birth, GDM gestational diabetes mellitus

Many studies have attempted to establish a simple way may be the underlying mechanism due to evolutionary
to predict preterm birth, but most focused on spontan- adaptation.
eous preterm birth. In our study, many variables were Placenta previa is a risk factor for preterm birth [5].
used to predict iatrogenic preterm birth. Most of the The formation of the lower uterine segment and cervical
underlying variables have been previously reported to dilation will cause a certain degree of spontaneous pla-
impact preterm birth, while myomectomy and dysmen- cental separation, which may result in severe haemor-
orrhea were included tentatively to assess their relation- rhage and can indicate preterm birth [19]. As a clinical
ship with iatrogenic preterm birth. Maternal and foetal indicator of iatrogenic preterm birth, placenta previa ac-
indications are the direct reasons for obstetricians to counts for 14.1% of all cases, which is much larger than
consider a patient at high risk of iatrogenic preterm the prevalence in China. To our knowledge, this is the
birth or to tend to terminate the pregnancy. Severe com- first report of the morbidity of placenta previa in a
plications during pregnancy, such as uterine rupture, are large-scale, Chinese, scarred uterus population.
likely to lead to iatrogenic preterm birth. Although the IPI is defined as the time from the most recent prior
prevalence of such complications is usually low, quick birth to conception of the index birth by Mckinney and
treatment is needed for these complications. Meanwhile, his coworkers [20]. An IPI of 0 to < 12 months
common foetal reasons for iatrogenic preterm birth, accounted for only a small portion of the data (n = 35,
such as foetal distress, also require quick treatment given 1.5%) because most of the women with IPIs less than 12
their sudden occurrence. Therefore, we mainly focused months were recommended for delivery of the baby.
on chronic pregnancy complications, such as placenta The 12 to < 24 months category was chosen as the refer-
previa and GDM, when we included complications into ence group based on Mckinney’s study, and an IPI of 12
the model. to < 24 months was associated with the lowest risk of
A shorter height has been associated with a progres- preterm birth in both Mckinney’s and our studies.
sive increase in the odds of having an infant born pre- Self-reported vaginal bleeding during pregnancy is pre-
term [17, 18]. Our study shows the same result by using dictive for preterm birth. The odds ratio of vaginal
a Chinese population, and a smaller maternal pelvic size bleeding was 2.7 (95% Cl 2.03–3.70) in our study, with

Fig. 1 Maternal indications of IPTB

Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 6 of 9

Table 2 Qualified risk factors for preterm birth in the multiple Women with a parity of less than two composed the
logistic regression model majority of our data (n = 2171, 93.8%) and in the general
Variables β P OR 95% CI population. This demographic characteristic is quite dif-
Parity ferent in China due to the singleton policy that had been
≤1 (reference) in place over the past years. Nulliparous and highly mul-
tiparous women are at higher risk of adverse pregnancy
2 1.070 < 0.0001 2.92 1.71–4.96
outcomes than those with low multiparity [23]. In our
≥3 2.112 0.001 8.26 2.29–29.76
study, advancing parity showed a higher risk of adverse
Interpregnancy interval pregnancy outcomes. However, the nulliparity included
13 to < 24 months (reference) in our study corresponded to women who had received
0 to < 12 months 1.674 0.003 5.33 1.79–15.91 myomectomy. Thus, there is no conflict between these
25 to < 60 months 0.590 0.068 1.80 0.96–3.40 two studies because of the different inclusion criteria.
Hypertension during pregnancy increases the risk of
≥ 60 months 0.467 0.140 1.60 0.86–2.97
preterm birth. A recent meta-analysis including 55 stud-
Hypertension in pregnancy
ies found that women with chronic hypertension had
No (reference) high pooled incidences of preterm birth [24]. Preeclamp-
Yes 2.253 < 0.0001 9.52 6.46–14.03 sia was also found to be associated with high rates of
Placenta previa preterm birth and puerperal complications, while gesta-
No (reference) tional hypertension was only found to be related to pre-
term birth [25].
Yes 1.438 < 0.0001 4.21 2.85–6.22
Apart from these six parameters, some factors were
Height −0.049 0.003 0.95 0.92–0.98
not included in our model and are thought to be import-
number of vaginal bleeding 0.593 < 0.0001 1.81 1.36–2.41 ant in the prediction of preterm birth. Age has always
Note: OR, odds ratio; 95% CI, 95% confidence intervals been considered a significant variable in preterm birth
prediction. We have tried many ways to categorize age,
including dividing it into three groups based on the re-
an incidence of 10.9%, smaller than that in other studies port of a U-shaped relationship with preterm birth [26].
[21, 22]. Our study used number of vaginal bleeding Sadly, none of these attempts give us a statistically sig-
during pregnancy instead of the presence of bleeding nificant result. Likewise, factors related to infection are
over the three trimesters or bleeding volume because we important variables related to preterm birth. However,
considered it to be easier for the patients to recall. screening for infection requires sequential tests,

Fig. 2 Calibration plot of the model

Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 7 of 9

Fig. 3 IPTB risk nomogram. Legend: Each predictor is assigned a score on each axis. The sum of all points for all predictors is computed and
denoted as the total score. The risk of IPTB for the total score was converted to a probability of GDM

Fig. 4 Decision curve analysis for IPTB. Legend: The decision curve analysis shows that if the threshold is between 0.15–0.6, use of the nomogram
in this study to predict IPTB adds more benefit than either a treat-all-patients scheme or a treat-none scheme
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 8 of 9

including vaginal secretion cultures and inflammatory institutions in rural areas of China that lack sufficient
indexes, which are not fully covered by medical insur- medical resources.
ance in China. Therefore, factors related to infection The shortcomings of this study are as follows: (1). The
were not included given their low cost-efficiency and influence of maternal weight on preterm birth is compli-
data integrity. cated. We decided to collect all the data on weight and
After screening the above six factors among all the weight gain for all trimesters at first. Due to data loss,
variables, a model was built and validated. The results of only weight before delivery was included in this study.
the discrimination and calibration tests are shown above. (2). A total of 2315 cases were divided into a training set
Overall, our model shows good discrimination and and a validation set; the nature of internal validation in-
calibration. However, discrimination and calibration dicates one of the weakness of our study. (3). Due to the
alone cannot capture the clinical consequences of a par- retrospective nature of this study, the data we collected,
ticular level of discrimination or degree of miscalibration except for age and height, were all acquired before deliv-
[27–29]. To justify the practical applicability of our ery. Therefore, the model is only valid for pregnant
model, decision curve analysis was applied in this study. women with more than 20 gestational weeks.
The results of the decision curve analysis indicate a
worth-expecting practice in the clinic. With the thresh- Conclusion
old probability between 15 and 60%, the use of a nomo- We built a model to predict iatrogenic preterm birth for
gram in our study to predict iatrogenic preterm birth pregnant women with scarred uteri. The variables in-
adds more benefit than either a treat-all-patients scheme cluded in the model were height, parity, IPI, vaginal
or a treat-none scheme. bleeding during pregnancy, placenta previa, and hyper-
The distribution of medical resources in China is not tension during pregnancy. The nomogram we developed
even. Many hospitals in rural areas lack neonatal inten- can assist doctors in evaluating the risk of iatrogenic
sive care units, which leads to adverse outcomes for the preterm birth and deciding whether these patients
newborns. Sometimes mothers with pregnancy compli- should be referred to an higher tier medical centre; thus,
cations cannot be treated effectively. Reasonable and better medical care can be provided to prevent adverse
efficient referral can improve this situation. Our model pregnancy outcomes and poor foetal conditions.
improves maternal and child outcomes by assessing the
risk of iatrogenic premature birth in patients, thus assist- Abbreviations
ing in referral-making and helping in the rational alloca- IPI: Interpregnancy interval; BMI: Body mass index; CI: Confidence interval;
OR: Odds ratio; IQR: Interquartile range; AUC: Area under curve;
tion of medical resources. A correct prediction can IPTB: Iatrogenic preterm birth; GDM: Gestational diabetes mellitus;
provide patients better medical care, thus improving the TOLAC: Trial of labour after caesarean
prognosis of the mother and foetus, while an incorrect
prediction would waste local medical resources or delay Acknowledgements
Not applicable.
the opportunity to treat patients. We recommend that
doctors use lower cutoff values in districts with rich Authors’ contributions
medical resources or high economic capability. Note that LCX and QC participated in the study design. YT, LJP, HY and LN conducted
the applicable population for this model is the same as the field study and collated the data. ZLY conducted the statistical analyses
and prepared the first draft of the manuscript. LHT and QC edited the
the inclusion criteria of this study, which means that this manuscript. XBX played a major role in illustrating and language editing. All
model is only suitable for pregnant women with more authors read and approved the final manuscript.
than 20 weeks of gestation. Meanwhile, new problems
may arise at any time, so we recommend using this Funding
The current study was supported by grants from the National Key R&D
model sequentially during pregnancy. Program of China (2016YFC1000404), the National Natural Science
Foundation of China (General Program; 81370735), the National Natural
Strengths and limitations Science Foundation of China (General Program; 81771610), and the
Outstanding Scientific Fund of Shengjing Hospital (201706). These funding
The prediction model we developed is novel. To our bodies played a role in protocol development and did not play any role in
knowledge, this is the first model to predict iatrogenic data collection, analysis, interpretation of data or writing the manuscript.
preterm birth in a scarred uterus population using the
population from Northeast China. In our study, we se- Availability of data and materials
The data analysed specifically for use in this study are not publicly available
lectively collected individual information that can be eas- due to their use in ongoing research, but reasonable requests for data can
ily acquired during consultation or through some basic be made to the corresponding author at the end of the research.
examinations. Thus, it is a convenient model that can be
easily practised in the clinic for the recognition of high- Ethics approval and consent to participate
The study was approved by the local ethics committee (ethics committee of
risk populations and for making referrals, which means China Medical University) in Shenyang. All participants agreed to participate
that it can be widely applied in primary health care in this study and signed an informed consent form.
Zhang et al. BMC Pregnancy and Childbirth (2020) 20:490 Page 9 of 9

Consent for publication height and risk for spontaneous preterm birth. Am J Obstet Gynecol.
Not applicable. 2015;213(5):700 e1–9.
19. Erez O, Novack L, Klaitman V, Erez-Weiss I, Beer-Weisel R, Dukler D, et al.
Competing interests Early preterm delivery due to placenta previa is an independent risk factor
The authors declare no competing interests. for a subsequent spontaneous preterm birth. BMC Pregnancy Childbirth.
2012;12(1):82.
Author details 20. McKinney D, House M, Chen A, Muglia L, DeFranco E. The influence of
1
Department of Obstetrics and Gynecology, Shengjing Hospital, China interpregnancy interval on infant mortality. Am J Obstet Gynecol. 2017;
Medical University, Shenyang, Liaoning Province, China. 2Institute of 216(3):316 e1–9.
Reproductive and Child Health/National Health Commission Key Laboratory 21. Ananth CV, Savitz DA. Vaginal bleeding and adverse reproductive outcomes:
of Reproductive Health, Peking University Health Science Center, No. 38 a meta-analysis. Paediatr Perinat Epidemiol. 1994;8(1):62–78.
Xueyuan Rd, Beijing 100191, China. 3China Medical University, Shenyang, 22. Yang J, Hartmann KE, Savitz DA, Herring AH, Dole N, Olshan AF, et al.
Liaoning Province, China. Vaginal bleeding during pregnancy and preterm birth. Am J Epidemiol.
2004;160(2):118–25.
Received: 17 March 2020 Accepted: 11 August 2020 23. Bai J, Wong FW, Bauman A, Mohsin M. Parity and pregnancy outcomes. Am
J Obstet Gynecol. 2002;186(2):274–8.
24. Bramham K, Parnell B, Nelson-Piercy C, Seed PT, Poston L, Chappell LC.
Chronic hypertension and pregnancy outcomes: systematic review and
References meta-analysis. BMJ. 2014;348:g2301.
1. Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R, et al. 25. Shen M, Smith GN, Rodger M, White RR, Walker MC, Wen SW. Comparison
National, regional, and worldwide estimates of preterm birth rates in the of risk factors and outcomes of gestational hypertension and pre-eclampsia.
year 2010 with time trends since 1990 for selected countries: a systematic PLoS One. 2017;12(4):e0175914.
analysis and implications. Lancet. 2012;379(9832):2162–72. 26. Fuchs F, Monet B, Ducruet T, Chaillet N, Audibert F. Effect of maternal age
2. Waldenstrom U, Cnattingius S, Vixner L, Norman M. Advanced maternal age on the risk of preterm birth: a large cohort study. PLoS One. 2018;13(1):
increases the risk of very preterm birth, irrespective of parity: a population- e0191002.
based register study. BJOG. 2017;124(8):1235–44. 27. Localio AR, Goodman S. Beyond the usual prediction accuracy metrics:
3. Joseph KS, Fahey J, Shankardass K, Allen VM, O'Campo P, Dodds L, et al. reporting results for clinical decision making. Ann Intern Med. 2012;
Effects of socioeconomic position and clinical risk factors on spontaneous 157(4):294–5.
and iatrogenic preterm birth. BMC Pregnancy Childbirth. 2014;14:117. 28. Van Calster B, Vickers AJ. Calibration of risk prediction models: impact on
4. Cnattingius S, Villamor E, Johansson S, Edstedt Bonamy AK, Persson M, decision-analytic performance. Med Decis Mak. 2015;35(2):162–9.
Wikstrom AK, et al. Maternal obesity and risk of preterm delivery. JAMA. 29. Van Calster B, Wynants L, Verbeek JFM, Verbakel JY, Christodoulou E, Vickers
2013;309(22):2362–70. AJ, et al. Reporting and interpreting decision curve analysis: a guide for
5. Ananth CVDK, Smulian JC, Vintzileos AM. Relationship among placenta investigators. Eur Urol. 2018;74(6):796–804.
previa, fetal growth restriction, and preterm birth. Obstet Gynecol. 2001;98:
299–306.
6. Fuchs F, Senat MV. Multiple gestations and preterm birth. Semin Fetal Publisher’s Note
Neonatal Med. 2016;21(2):113–20. Springer Nature remains neutral with regard to jurisdictional claims in
7. Bruijn MM, Kamphuis EI, Hoesli IM, Martinez de Tejada B, Loccufier AR, published maps and institutional affiliations.
Kuhnert M, et al. The predictive value of quantitative fibronectin testing in
combination with cervical length measurement in symptomatic women.
Am J Obstet Gynecol. 2016;215(6):793 e1–8.
8. Kekki M, Kurki T, Karkkainen T, Hiilesmaa V, Paavonen J, Rutanen EM. Insulin-
like growth factor-binding protein-1 in cervical secretion as a predictor of
preterm delivery. Acta Obstet Gynecol Scand. 2001;80(6):546–51.
9. Liong S, Di Quinzio MK, Fleming G, Permezel M, Rice GE, Georgiou HM. New
biomarkers for the prediction of spontaneous preterm labour in
symptomatic pregnant women: a comparison with fetal fibronectin. BJOG.
2015;122(3):370–9.
10. Vrachnis N, Grigoriadis C, Siristatidis C, Vlachadis N, Balakitsas N, Mastorakos
G, et al. The Janus face of maternal serum relaxin: a facilitator of birth,
might it also induce preterm birth? J Matern Fetal Neonatal Med. 2015;
28(18):2187–91.
11. Kim MA, Lee BS, Park YW, Seo K. Serum markers for prediction of
spontaneous preterm delivery in preterm labour. Eur J Clin Investig. 2011;
41(7):773–80.
12. Lucovnik M, Bregar AT, Steblovnik L, Verdenik I, Gersak K, Blickstein I, et al.
Changes in incidence of iatrogenic and spontaneous preterm births over
time: a population-based study. J Perinat Med. 2016;44(5):505–9.
13. Clark EA, Silver RM. Long-term maternal morbidity associated with repeat
cesarean delivery. Am J Obstet Gynecol. 2011;205(6 Suppl):S2–10.
14. American College of O, Gynecologists. ACOG Practice Bulletin No. 125:
Chronic hypertension in pregnancy. Obstet Gynecol. 2012;119(2 Pt 1):
396–407.
15. Silver RM. Abnormal placentation: placenta Previa, vasa Previa, and placenta
Accreta. Obstet Gynecol. 2015;126(3):654–68.
16. American Diabetes A. 2. Classification and Diagnosis of Diabetes. Diabetes
Care. 2016;39(Suppl 1):S13–22.
17. Derraik JG, Lundgren M, Cutfield WS, Ahlsson F. Maternal height and
preterm birth: a study on 192,432 Swedish women. PLoS One. 2016;11(4):
e0154304.
18. Shachar BZ, Mayo JA, Lee HC, Carmichael SL, Stevenson DK, Shaw GM,
et al. Effects of race/ethnicity and BMI on the association between