Predictive factors in pregnancies with reduced fetal movements: a pilot study

Abstract

Background

Pregnancies with reduced fetal movements (RFM) are at risk for poor neonatal outcomes and stillbirth.

Aim

To investigate whether Doppler measurements or angiogenic factors are good predictors of adverse neonatal outcomes in pregnancies with RFM.

Methods

This is a prospective pilot cohort study of 3243 women seeking care for RFM. Standard care was carried out in all cases. An extra Doppler examination was performed in 128 women to assess the flow in the middle cerebral artery, the umbilical artery, and the uterine artery. In 62/128 pregnancies, a maternal blood sample was obtained for angiogenic and antiangiogenic factors. The composite neonatal outcome of the study was one or more of the following factors: Apgar score <7 at 5′, arterial aPh in the umbilical cord ≤7.1, transfer to Neonatal Intensive Care (NICU), stillbirth, and small for gestational age (SGA).

Results

In 14.1% (18/128) of the Doppler group and 11.7% (365/3115) of the standard care group, there was an adverse neonatal outcome (p = .51). A higher intervention rate was found in the Doppler group (28% vs. 5.4%, p < .01). The predictive model of adverse neonatal outcomes in women with RFM with angiogenic factors was 0.73 (95% CI 0.54–0.92). The area under the curve improved to 0.89 (CI 95% 0.81–0.97) when parity was added to the model.

Conclusion

Angiogenic factors may have a place in the prediction of the neonatal outcome of RFM pregnancies. The prediction model’s capacity was driven by parity. The obstetrical intervention rate increased with additional Doppler examinations.

Keywords:

• Doppler
• reduced fetal movement
• ultrasound

Introduction

Four out of 1000 pregnancies end in stillbirth in Sweden, some of which might be preventable [1]. Pregnancies with reduced fetal movements (RFM) are considered a risk group for perinatal complications such as stillbirth and small for gestational age (SGA) [2–6]. Between 4% and 22% of pregnant women seek medical attention for RFM during their pregnancies [7,8].

Placental insufficiency is considered contributory to stillbirth in up to 65% of cases [9–11]. Doppler examination of the umbilical artery (UA) and the middle cerebral artery (MCA) with an estimation of the cerebroplacental ratio (CPR) could be a feasible method of detecting placental insufficiency. A large retrospective cohort study performed showed that a low CPR value within two weeks of delivery, regardless of fetal size, was associated with intrapartum fetal compromise and admission to the Neonatal Intensive Care Unit (NICU) [12]. The CPR evaluated in the third trimester could be an independent predictor of the risk of stillbirth and perinatal mortality in a mixed-risk population [13]. However, a recently published meta-analysis on Doppler parameters showed that the CPR has limited added value to the UA [14]. Furthermore, it is still unclear if the CPR applies to all pregnancies or is a good predictor for adverse neonatal outcomes in pregnancies with RFM.

Biomarkers of placental function can potentially aid the diagnosis and prediction of pregnancy complications. An imbalance between the concentration of angiogenic and antiangiogenic factors has previously been shown to be associated with pregnancy complications such as preeclampsia, SGA, and stillbirth [15,16]. In a recently published study, the ratio of angiogenic factors sFlt-1/PlGF was analyzed in a population of women seeking care for RFM. The results showed that adverse outcome rates in these pregnancies were low, and more studies are required [17,18].

Women in Sweden are encouraged to seek care when they experience RFM during pregnancy. This information is given to them by midwives at antenatal clinics, and previous studies show that there is good compliance with the recommendations [19]. The current management for women who seek health care for RFM after 24 weeks of gestation includes a CTG (cardiotocography), an US (ultrasound) to estimate the deepest pocket of amniotic fluid (AF), and fetal movements. However, the current standard management of RFM lacks evidence for its effectiveness.

This study aimed to investigate if Doppler measurements or angiogenic factors are good predictors of adverse neonatal outcomes in pregnancies with RFM.

Materials and methods

Study population

A prospective longitudinal cohort study was conducted between January 2016 and December 2017 at Soder Hospital, Stockholm. Soder Hospital is a secondary referral maternity clinic with 7900 deliveries per year.

Inclusion criteria were pregnant women seeking medical care due to RFM Sunday to Friday with singleton pregnancies ≥34 weeks. RFM was defined as the womans subjective experience of fewer or no fetal movments. Pregnancies with known chromosomal and congenital anomalies, multiple pregnancies, and women who did not speak Swedish or English were not included in the study.

Ultrasound examination

All pregnancies in Sweden are scheduled to be dated according to ultrasound measurements in the first or second trimester. The definition of SGA was estimated weight >2 standard deviations (SD) below expected weight for gestational age based on Swedish reference curves [20].

After standard care checkup procedures for RFM, the participants were offered an additional Doppler examination for assessing flow in the UA, the UtA, and the MCA. The US examination was performed within 24 h of the episode with RFM. Since it is considered unethical not to give appropriate care in the case of pathological Doppler flow, the obstetrician in charge of the patient was not blinded to the US examination results.

All the US examiners were obstetricians specialized in obstetrical ultrasound, and they were certified according to national guidelines. The pulsatility index (PI) in the UtA was estimated as previously described by Gomez et al. [21], and the CPR was calculated as the ratio between the MCA PI and the UA PI. Published reference values are used in standard praxis in Sweden to define the normal values [22]. In the case of abnormal Doppler US, biometry [20] was performed.

Doppler US measurements were recorded using a 4–8 or 2–7 MHz transabdominal transducer (Voluson E8 or E10, GE Medical Systems, Kretz Ultrasound, Zipf, Austria).

Recordings were obtained during the absence of fetal breathing and body movements and, if necessary, during a temporary maternal breath-hold. For accuracy, Doppler measurements were repeated three times. The best quality measurement was chosen. The insonation angle was kept as low as possible.

The MCA PI was measured by placing the 2-mm Doppler gate at the MCA’s proximal third, close to the circulus Willisii. For the UtA measurement, the 3-mm gate was placed as close to the iliac artery as possible. For the UA flow (UA PI), measurements were made from a free cord loop with a gate size of 3 mm. To adjust for gestational age, all Doppler values were converted into multiples of medians (MoM) [23].

Angiogenic and antiangiogenic factors analysis

A maternal venous blood sample was collected at the visit with RFM. The blood was centrifugated and stored at −80 °C within 30 min of sampling. Angiopoietin-1 (LXSAHM-01), Angiopoietin-2, PlGF, and VEGFR1/Flt-1 (LXSAHM-03) plasma levels were measured using the multiplex human immunoassay kit (R&D Systems/Biotechne, Abingdon, UK).

The samples were diluted two-fold with the assay diluent provided in the kit. The final dilution of the sample would be 4-fold, and 50 μL was added to the respective wells according to the 96-well plate layout.

To each microplate well, 50 μL of human magnetic premixed microparticle cocktail provided in the kit was added, followed by 50 μL of standards or samples per well. The microplate was incubated for two hours at room temperature on a horizontal orbital microplate shaker (0.12″ orbit) set at 800 ± 50 rpm. Next, the plate was placed on a designed magnetic device before being washed with 100 μL wash buffer three times following the manufacturer’s protocol. Later, 50 μL of diluted biotin antibody cocktail was added to each well and incubated for one hour at room temperature on the shaker set at 800 ± 50 rpm, followed by the washing step described earlier. Afterwards, the plate was again re-incubated for 30 min at room temperature on a shaker at 800 ± 50 rpm after 50 μL of diluted Streptavidin-PE was added to each well. The wash step was repeated, and finally 100 μL of wash buffer was added to each well before the plate was incubated for two minutes on a shaker at 800 ± 50 rpm. The plate was then read using a Luminex analyzer (Luminex Corporation, Austin, TX, USA).

Outcome and data collection

The midwife in charge assessed the Apgar scores at 1 and 5 min after delivery. A cord blood sample was taken after delivery before the newborn’s first cry, and acid-base values were analyzed within a few minutes (ABL 800 Bayer^®, Leverkusen, Germany). A composite neonatal outcome was constructed and defined as follows: 5 min Apgar score <7, arterial aPh in the umbilical cord ≤7.1, transfer to NICU, stillbirth, and SGA, defined according to Swedish national guidelines as −2 SD below expected weight for gestational age.

At least one of the above factors was required to ensure that the outcome would be noted as a composite neonatal outcome.

Data regarding maternal, pregnancy, and neonatal characteristics were collected from the medical records in Obstetrix (Cerner, North Kansas City, US)

Ethical approval

The study was approved by the regional Research Ethics Committee, Karolinska Institute, Stockholm, Sweden (2015/2043-31/2). The study complied with the Declaration of Helsinki. Written informed consent was obtained from all the women before inclusion in the study.

Statistical analysis

All the statistical analyses were performed using SPSS Statistics 24.0 (IBM Corp, Armonk, NY, USA). All the demographic data on the women and their pregnancies measured on a continuous scale were presented as medians. All the categorical variables were presented as numbers and percentages.

Data from participants and non-participants were compared using the Mann–Whitney test for the continuous variables and the Chi-squared test for categorical variables.

The frequencies of all potential predictors for composite neonatal outcome were compared between pregnancies with and without the outcome using the Mann–Whitney U test for the continuous variables and the Chi-squared test for the categorical variables.

The receiver operating characteristic curves (ROC) and areas under the curve (AUC) were calculated to predict the composite neonatal outcomes. The AUC summarizes a test’s performance in terms of sensitivity and specificity. An AUC of 0.5 suggests no discrimination, 0.7–0.8 is acceptable, 0.8–0.9 is excellent, and > 0.9 is outstanding. The final models for predicting the composite neonatal outcome included the predictive factors investigated in this study and the risk factors associated with the composite neonatal outcome. The prediction models were based on the estimates from the logistic regression analyses. Statistical differences were considered significant given a p-value < .05.

Results

Reduced fetal movement

From January 2016 to December 2017, 3243 pregnant women sought medical care due to RFM at the hospital. In this pilot study, from May 2016 to December 2017, 139 women agreed to participate in the study and had the extra Doppler examination performed. Eleven deliveries were excluded due to incomplete examinations or gestation age <34 weeks, leaving 128 complete investigations (Figure 1).

Figure 1. Flow chart.

Table 1 shows the women’s and newborns’ demographic data from the group attending health care at the clinic for RFM during the study period. The age of the women with an extra ultrasound examination was significantly higher (p < .001); they were more likely to be multipara (p = .04) and have fewer complications during pregnancy (p < .01) compared to the group with standard care. It was more common for the women with the extra Doppler examination to seek care for RFM only once (p < .01), and significantly more of them were induced to labor (36.9 vs. 23.1%, p < .01, Table 1). In 21 (0.7%) pregnancies, a stillbirth resulted. Fifteen of these were stillborn at the first RFM visit, one was a known trisomy, and one had a known IUGR (Intra Uterine Growth Restriction) with abnormalities in the umbilical flow but did not accept cesarean section delivery. No stillbirths occurred in the group in which an additional Doppler examination was conducted.

Table 1. Data of the women and new-born included in the project.

	No extra ultrasound		Extra Doppler
	n = 3115		n = 128		p-Value^a
Maternal characteristics
Age (years)	31	(15–54)	34	(19–45)	<.01^a
BMI	23.5	(15–58)	23.4	(18–40)	.80
Smokers (%)	401	13.4	13	10.2	.35
History of psychiatric disease (%)	521	18.0	30	23.4	.09
IVF (%)	218	7.0	12	9.4	.53
Nullipara (%)	1798	57.7	62	48.4	.04^a
Complications of pregnancy^a (%)	175	5.6	6	4.7	<.01^a
Number of visits for RFM
Single visit (%)	2610	83.4	87	68	<.01^a
Multiple visits (%)	519	16.6	41	32
Onset of labor
Spontaneous (%)	2038	65.0	70	54.9	<.01^a
Induced (%)	723	23.1	47	36.9
Planned cesarean (%)	375	12.0	10	7.9
Method of delivery
Spontaneous vaginal (%)	2291	73.1	94	73.7	.8
Vacuum (%)	196	6.3	8	6.4
Cesarean (ES + AS, %)	649	20.7	25	19.7
Fetal characteristics
Gestational age (days)	281	(164–303)	281	(255–296)	.39
Birth weight (g)	3570	(535–5590)	3630	(2050–4680)	.49
Length (cm)	50	(28–58)	50	(43–56)	.87
Female gender (%)	1610	49.6	66	52.0	.71
Poor neonatal outcome score
APGAR ≤ 7 at 5 min (%)	102	3.3	4	3.1	.99
Arterial pH ≤ 7.10 (%)	144	5.9	6	5.5	.99
Neonatal ward (%)	92	2.9	6	4.7	.43
Stillbirth (%)	21	0.7	0	0	.71
SGA (%)	121	3.9	6	4.7	.82
Composite neonatal outcome (%)^b	365	11.7	18	14.1	.51

Deliveries are divided into groups according to an extra Doppler examination or not. Values are numbers (%) or median (range).

^a p-Value < .05 is suggested as statistically significant.

^b Composite neonatal outcome: - - Apgar score ≤7, at 5’, arterial pH in cord blood ≤ 7.10, transfer to neonatal care unit, IUFD and small for gestational age (SGA) at delivery.

BMI: body mass index; IVF: in vitro fertilization; SGA: small for gestational age.

Despite being a high-risk group, when comparing the Doppler group (N = 128) with all deliveries without RFM at Soder Hospital 2016–2017 (N = 11,944), there were no statistically significant differences in the composite neonatal outcome or any of the subcomponents (Table 2).

Table 2. Neonatal outcomes for the group with reduced fetal movements (RFM) and additional Doppler examinations (N = 128) and the group with no RFM (N = 11,944).

	Pregnancies with RFM	Pregnancies without RFM
	N = 128	N = 11,944	p-Value^a
Poor neonatal outcome score
Apgar ≤7 at 5 min (%)	4 (3.1)	249 (2.1)	.66
Arterial pH^b ≤7.10 (%)	6 (5.5)	411 (4.4)	.88
Neonatal ward (%)	6 (4.7)	600 (5.0)	.86
Stillbirth (%)	0 (0)	13 (0.11)	.32
SGA (%)	6 (4.7)	387 (3.3)	.50
Composite outcome* (%)	18 (14.2)	1336 (11.2)	.37

^a p-Value significant if < .05.

^b Not available for all neonates (N = 120 in the RFM group and N = 9336 in the control group).

SGA: small for gestational age.

*Composite neonatal outcome: 5 min Apgar score ≤ 7 arterial pH in cord blood ≤ 7.1, transfer to neonatal care unit, intrauterine fetal death, and small for gestational age (SGA) at delivery.

Obstetrical interventions after Doppler examination

The number of interventions defined as follow-up, induction of labor, or hospitalization was higher in the group with extra Doppler examination compared to the standard care group at 28% (36/128) vs. 5.4% (7/128, p < .01). One explanation may be that the investigator was not blinded to the Doppler examination results and had to act on these according to the clinic’s guidelines on Doppler abnormalities. No statistically significant differences within the RFM group were found regarding the CPR MoM, the UtA PI MoM, and the MCA PI MoM when comparing those with an adverse composite neonatal outcome and those with a normal outcome (Table 3).

Table 3. Angiogenic factors and Doppler parameters for the study group with and without composite neonatal outcome. Figures are presented as medians (range).

	Total median values	Without composite neonatal outcome	With composite neonatal outcome	p-Values^a
		(N = 56)	(N = 6)
Angiogenic factors
Ang-1	28.40	28.80	36.18	.17
	(6.03–70.17)	(6.03–70.17)	(6.39–36.18)
PlGF	29.40	30.84	23.00	.16
	(2.68–484.80)	(2.68–484.80)	(3.80–152.00)
VEGF	3.47	3.40	19.8	.77
	(0.02–209.00)	(0.02–209)	(0.04–39.3)
Antiangiogenic factors
sFlt1	9.28	9.14	10.48	.46
	(0.72–47.88)	(0.72–47.88)	(2.84–16.40)
Ang-2	8.32	8.32	8.38	.93
	(0.004–70.00)	(0.004–70.0)	(1.12–68.48)
Ratio angiogenic/antiangiogenic factors
PlGF/sFlt1	3.21	3.38	1.74	.56
	(0.23–82.45)	(0.23–82.45)	(0.35–23.31)
Ang-1/Ang-2	3.10	3.31	2.34	.35
	(0.97–12,017.50)	(0.97–12,017.50)	(0.45–5.71)
Doppler parameters
CPR MoM	0.99 (0.53–1.98)	0.98 (0.53–1.98)	1.12 (0.73–1.73)	.49
UtA PI MoM	0.82 (0.41–2.82)	0.83 (0.48–2.82)	0.77 (0.46–1.43)	.56
MCA PI MoM	1.24 (0.58–2.29)	1.23 (0.58–2.29)	1.26 (0.54–1.01)	.79

^a p-values < .05 are considered as statistically significant assessed with the Mann–Whitney U test.

CPR: cerebroplacental ratio; MCA: middle cerebral artery; MoM: multiples of medians; UtA: uterine artery; PI: pulsatility index.

Angiogenic and antiangiogenic factors

In 62/128 (48.4%) of the Doppler group, a maternal blood sample was collected to analyze angiogenic factors (Angiopoietin-1, PlGF, and VEGF) and antiangiogenic factors (Angiopoietin-2 and sFlt1). Women with an adverse composite neonatal outcome did not differ from those without (Table 3).

The ROC and AUC were calculated to assess the predictive value of the angiogenic and the antiangiogenic factors in pregnancies with RFM. Each angiogenic factor was assessed individually, and a final model with all the angiogenic factors together was created. The best model, where all the angiogenic factors were analyzed together, had an AUC of 0.73 (95% CI 0.54–0.92). When adding parity as a predictor of the neonatal outcome, the model performed better, improving the AUC to 0.89 (CI 95% 0.81–0.97) for the angiogenic factors (Figure 2).

Figure 2. ROC curves and AUC for prediction of composite neonatal outcome.

Discussion

In our pilot study, we found that women with RFM surveilled with extra Doppler were at similar risk for adverse outcomes as the background population and the non-Doppler group with RFM. The angiogenic and antiangiogenic factors had a moderate predictive capacity for the composite outcome. However, the study was underpowered to draw too many conclusions on the usefulness of the CPR in RFM pregnancies. We recently reported that women in the RFM group had a higher risk for low Apgar, low pH, and a fivefold higher risk of stillbirth than the non-RFM group [7].

At Soder Hospital in Stockholm, approximately 2000 women seek care for RFM and approximately 40 children die intrauterine every year [7]. The most common scenario among these cases is that the woman attends health care for RFM, and the fetus is already dead.

Obstetrical interventions after doppler examination

Higgins et al. [17] reported a study where factors associated with placental dysfunctions were analyzed. They found that antenatal placental assessment may improve the identification of RFM pregnancies at the highest risk of adverse pregnancy outcomes, but further work is required to improve clinical utility. Our study shows that by adding an extra Doppler examination (CPR and UtA PI), there was a statistically significant higher rate of interventions without any improvement of adverse composite neonatal outcomes. This is in line with a recent large meta-analysis with individual participant data showing no predictive value for adverse outcomes. The authors concluded that their results do not support the use of the CPR outside a research setting [14]. Doppler measurements, mainly UA Doppler, have been shown to predict adverse outcomes among SGA fetuses, with increasing blood flow class changes in the UA being the best predictor of neurological handicap [24].

Our results are also in line with the results of the AFFIRM study [25], in which a care package for RFM significantly increased interventions without substantially reducing the stillbirth rate. One explanation for the high intervention rate in our study may be that the investigator was not blinded to the Doppler examination and could act on it. The Doppler group (including the CPR) did not have a lower incidence of adverse neonatal outcomes compared to standard care. There were no differences in the AUC or ROC when comparing standard care and the CPR.

Angiogenic and antiangiogenic factors

Placental insufficiency and ischemia induce an imbalance between the proangiogenic factors (VEGF-A, PlGF, and Angiopoietin-1) and the antiangiogenic factors (Angiopoietin-2 and s-Flt1) [26]. A previous study has shown that women with high levels of Angiopoietin-2 had an increased risk of having low placental weight, which is associated with poor neonatal outcomes. Similarly, having high levels of PlGF has been associated with a reduced risk of low placental weight and poor neonatal outcomes [27]. There were no differences in the median value of the proangiogenic factors or the antiangiogenic factors between the groups when comparing the neonatal outcomes in our study.

A randomized controlled trial was presented by Armstrong-Buisseret et al. [17], in which blood samples for analysis of angiogenic factors were sampled from women who sought care for RFM. Based on the results of an additional blood test for the sFlt-1/PlGF ratio (high or low), induction vs. standard care was offered to women presenting with RFM. A ratio above 38 was the cutoff value used. In the intervention arm, 8% of the infants had an adverse composite pregnancy outcome at delivery compared with 4% in the control arm. The authors concluded that the frequency of adverse outcomes was low, and a large sample size would be desired along with considering the most appropriate outcome measures. In our study, various angiogenic factors were analyzed. They reflect the placental dysfunction, which has a good association with RFM. For example, in previous publications, we have seen a strong association between intrauterine fetal death and SGA, which is strongly associated with placental dysfunction [7,8]. In our study, the sFlt-1/PlGF ratio was also analyzed. Only two women had a value >38, and these babies belonged to the normal composite outcome group after labor.

Interestingly, however, when a prediction with a ROC curve was made, a combination of the angiogenic factors showed an AUC > 0.70, which is considered an acceptable test. Nevertheless, this is a pilot study with a small sample size, and further studies with much larger sample sizes are needed. We also found more nulliparas in the group with adverse composite outcomes (92.9% vs. 42.5%, p < .01). When parity was added as a predictor of the neonatal outcome, all models improved to 0.89 (CI 95% 0.81–0.97) for the angiogenic factors. Therefore, we assume that parity is an important factor when dealing with women with RFM. These results are in line with our previous study, which showed that nulliparous women with RFM had an almost twofold increased risk of delivering a baby with a poor composite neonatal outcome compared with multiparous women.

This study’s strengths include its prospective design with all ultrasound examinations performed by a few ultrasound examiners only. High-quality data were registered from medical records. The whole group of RFM pregnancies received consistent standard management, and the angiogenic factors were used as predictors in a new context.

The study’s limitations include its small size, which limits possible generalizations. It is, however, clear that introducing the information on the CPR increased the number of interventions. Since the obstetrician handling the case was not blinded to the Doppler results, it is difficult to draw any conclusions regarding the predictive value of the CPR. The low number of patients included in the study may lead to an overfit of the predictive model in angiogenic factors analysis.

Conclusion

RFM is a common complaint when working in obstetrical health care. Women with RFM are urged to attend health care by medical staff who are concerned about the pregnancy outcome. In pregnancies where a Doppler examination was performed, the risk of developing the composite outcome was like the control group, but interventions increased. Angiogenic factors may have a place in the prediction of the neonatal outcome of RFM pregnancies.

Disclosure statement

No potential conflict of interest was reported by the author(s).