Does fetal size affect maternal perception of fetal movements? Evidence from an individual participant data meta‐analysis

Abstract

Introduction

Maternal perception of fetal movements during pregnancy are reassuring; however, the perception of a reduction in movements are concerning to women and known to be associated with increased odds of late stillbirth. Prior to full term, little evidence exists to provide guidelines on how to proceed unless there is an immediate risk to the fetus. Increased strength of movement is the most commonly reported perception of women through to full term, but perception of movement is also hypothesized to be influenced by fetal size. The study aimed to assess the pattern of maternal perception of strength and frequency of fetal movement by gestation and customized birthweight quartile in ongoing pregnancies. A further aim was to assess the association of stillbirth to perception of fetal movements stratified by customized birthweight quartile.

Material and methods

This analysis was an individual participant data meta‐analyses of five case–control studies investigating factors associated with stillbirth. The dataset included 851 cases of women with late stillbirth (>28 weeks' gestation) and 2257 women with ongoing pregnancies who then had a liveborn infant.

Results

The frequency of prioritized fetal movement from 28 weeks' gestation showed a similar pattern for each quartile of birthweight with increased strength being the predominant perception of fetal movement through to full term. The odds of stillbirth associated with reduced fetal movements was increased in all quartiles of customized birthweight centiles but was notably greater in babies in the lowest two quartiles (Q1: adjusted OR: 9.34, 95% CI: 5.43, 16.06 and Q2: adjusted OR: 6.11, 95% CI: 3.11, 11.99). The decreased odds associated with increased strength of movement was present for all customized birthweight quartiles (adjusted OR range: 0.25–0.56).

Conclusions

Increased strength of fetal movements in late pregnancy is a positive finding irrespective of fetal size. However, reduced fetal movements are associated with stillbirth, and more so when the fetus is small.

Odds of stillbirth are increased for lower birthweight centiles, but are relatively greater for women not experiencing increased strength of movement.

Abbreviations

IPD

individual participant data

RFM

reduced fetal movements

Key message.

Increased strength of fetal movements in late pregnancy should be considered a good sign irrespective of fetal size. However, reduced fetal movements are associated with stillbirth, and even more so when the fetus is small.

1. INTRODUCTION

Perception of reduced fetal movements (RFM) in late pregnancy is a common concern for pregnant women and can be a sign of fetal distress or demise. ¹ Near term gestation, consensus‐based guidelines advise whether to continue the pregnancy or induce labor. However, at earlier gestations it is less clear how to interpret RFM, and often difficult to clearly define the appropriate clinical pathway.

Maternal perception of fetal movements is often attributed to maternal or fetal characteristics with little underlying evidence. ² For example, obesity is presumed to be associated with a reduced ability to perceive fetal movements; however, a systematic review concluded that there was limited data investigating the association and no evidence to support that women with obesity were less likely to perceive fetal movements. ³

Many guidelines suggest that at gestations beyond 32 weeks, strength and/or frequency of movements do not continue to increase but remain the same. ⁴ Physiologically this has been attributed to a decrease in gross body movements as a result of less intrauterine space, as well as fetal neurological maturation. ⁵ In support of this hypothesis, a magnetic resonance imaging study found that there was less proportion of imaging time spent moving as gestation and therefore fetal size increased. ⁶ Closer examination of this publication reports the number of movements decreased from 20 to 27 weeks, with little change in the number of movements from 28 weeks onwards. Other studies have reported that pregnant women noticed a change in fetal movement type but not frequency, ⁷ which have been described in studies as more rolling, sliding, pushing type movements. ⁸ , ⁹

If larger fetal size is associated with a reduction in perception of fetal movements, then this is important in understanding and interpreting RFM, notwithstanding that there is an acknowledged relationship of perception of RFM with small fetal size. If there is a reduction in the ability to perceive fetal movements as the fetus gets larger, this would coincide with the period at which RFM are most likely to be reported. Thus, understanding such a relationship is important to clinical decision‐making.

In our previous study, we described the most common pattern of fetal movements in late pregnancy, and their relationship of these patterns to stillbirth. ¹ The most commonly reported perception of fetal movements from 28 weeks of pregnancy in women with liveborn infants was a perception of increased strength of movements, though the proportion with no change in strength was similar to those who reported increased strength close to term. Compared to those for whom strength and frequency of movements remained the same the odds of stillbirth were reduced in those with increasing strength of movement and conversely was increased for women reporting a decrease in the frequency of fetal movements in the last 2 weeks. ¹

We hypothesized that increased strength of movements would be the most commonly reported perception of movements over gestation for all quartiles of birthweight, and second that smaller babies would be at increased likelihood of stillbirth, within each prioritized group of perception of fetal movements.

This analysis therefore had two aims:

1. To investigate the pattern of perceived fetal movements in the third trimester (28–41 weeks) in relation to fetal size at birth in normal pregnancies resulting in a livebirth.

2. To determine whether the relationship of perceived fetal movements to stillbirth is influenced by fetal size (approximated by birthweight quartile).

2. MATERIAL AND METHODS

The collaborative individual participant data (IPD) meta‐analysis of sleep and stillbirth study (CRIBSS) included data from the five studies identified according to the protocol ¹⁰ that investigated risk factors for stillbirth. The protocol was also registered prospectively in PROSPERO (February 16, 2017, ID: CRD42017047703). Included studies were: The Auckland Stillbirth Study (TASS, Auckland, New Zealand), ¹¹ The Multi‐Center Stillbirth Study (MCSS, Nationwide, New Zealand), ¹² The Midlands and North East of England Stillbirth Study (MINESS, various centers, UK), ¹³ The Sydney Stillbirth Study (SSS, Sydney, Australia) ¹⁴ and the Study of Trends and Associated Risks for Stillbirth (STARS, internet based study, with most participants from the USA). ¹⁵ The CRIBSS study methodology has previously been described in detail. ¹⁰ , ¹⁶

The CRIBBS dataset includes data on 851 cases of women with late stillbirth (>28 weeks gestation) and 2257 women with ongoing pregnancies who went on to have a liveborn infant (controls).

Data collected included maternal perception of fetal movements, which have previously been described in detail. ¹ Specifically, in the four in person case–control studies, both mothers of stillborn and control participants were interviewed by trained interviewers using a standardized questionnaire. Controls were interviewed at a gestational age randomly assigned, and cases as soon a reasonably possible, with respect of family circumstances, after the delivery of their stillborn infant. There was no specific timing or gestational age constraints on participants in the online study. The questions about perception of frequency and strength of movements were similar across studies and the specific questions and available responses are shown in Table S1. A previously defined prioritized variable ¹ was created based on the most common responses, with the following order, (1) increased strength of movement, (2) increased frequency of movement (if strength not increased), (3) reduced frequency of movement, (4) no change and (5) unsure.

Data were available for maternal height and weight at booking, gestational age at birth and birthweight. Customized percentiles were calculated using the GROW Global Bulk Centile Calculator: (version 8.0.1, 2018). ¹⁷ For the purposes of this analysis birthweight centiles were categorized into quartiles and are used as a proxy for estimated fetal weight, which is not available from the case–control studies that provide data to this analysis.

For the first aim of the study, only women with ongoing pregnancies (which resulted in a livebirth) were used to determine the proportion in each prioritized group of perceived fetal movement at each gestational age (completed weeks).

For the second aim, the analysis to assess the odds of stillbirth associated with prioritized perception of fetal movement used both the cases and the controls from the IPD. Analyses were stratified by birthweight quartiles using logistic regression to estimate the odds of stillbirth for each prioritized perceived fetal movement group. A multivariable model used in previous analyses of this IPD dataset ¹⁶ was used to adjust for potential confounders, namely maternal age, earliest pregnancy body mass index, maternal ethnicity, parity, maternal education level, marital status, pre‐existing hypertension or diabetes, maternal smoking, recreational drug use, supine going to sleep position, unusually vigorous fetal movements, and fetal hiccups. Finally, an interaction term was added to this model to assess whether the odds of stillbirth by prioritized perceived fetal movement group was different by birthweight quartile. Additionally, both univariable and multivariable logistic regression models included a strata term for the variable identifying the study each participant was from, to adjust for potential clustering within studies. To investigate the relationship to stillbirth of the prioritized perceived fetal movement groups by percentile, generalized additive models (Proc GAM) were created to show the shape of the relation of each fetal movement group by the continuous birthweight percentile. Analyses were carried out in SAS version 9.4 (SAS Institute), and statistical significance was defined at the 5% level.

2.1. Ethics statement

Each individual study included in the IPD obtained ethical approval from their research ethics committee or Institutional Review Board. Approval for the IPD meta‐analysis was obtained from the New Zealand Health and Disability Ethics Committee (NTX/06/05/054/AM06) on February 24, 2017.

3. RESULTS

The frequency of the prioritized categories of fetal movement from 28 weeks' gestation were stratified by quartiles of birthweight for the livebirths (Figure 1). These plots suggest a similar pattern for each quartile of birthweight. Increased strength is the predominant perception of fetal movement in all birthweight quartiles through to 37 weeks' gestation. Though the prevalence of increased strength decreases as gestation increases it remains the predominant response in comparison to the next most common group “the same” strength and frequency until near term.

FIGURE 1.

Prevalence of prioritized perception of fetal movements by gestational age and birthweight quartile for controls.

The univariable association of perception of fetal movements according to the prioritized fetal movement variable stratified by birthweight quartiles in relation to stillbirth is summarized in Table 1. The multivariable associations (Table 2) show some small changes (generally a shift towards unity) after controlling for the potential confounders but the same pattern of association Specifically, the odds of stillbirth associated with reduced fetal movements is increased in all quartiles of customized birthweight centiles but is notably greater in babies in the lowest two quartiles. The decreased odds associated with increased strength of movement is present for all birthweight quartiles though shows no pattern.

TABLE 1.

Univariable odds ratios for stillbirth and perception of fetal movements (prioritized variable) by customized birthweight centile quartile.

	Quartile 1 (0–24) (n = 409 SB, n = 502 cont) OR 95% (CI)	Quartile 2 (25–49) (n = 161 SB, n = 505 cont) OR 95% (CI)	Quartile 3 (50–74) (n = 127 SB, n = 474 cont) OR 95% (CI)	Quartile 4 (75+) (n = 145 SB, n = 498 cont) OR 95% (CI)
Increased strength (n = 124 SB – n = 997 cont)	0.49 (0.31–0.79) p = 0.0034	0.24 (0.14–0.42) p < 0.0001	0.42 (0.25–0.71) p = 0.0012	0.27 (0.15–0.47) p < 0.0001
Increased frequency (n = 25 SB – n = 90 cont)	1.92 (0.87–4.27) p = 0.11	0.76 (0.27–2.14) p = 0.60	0.24 (0.06–1.08) p = 0.06	0.73 (0.26–2.04) p = 0.54
Reduced frequency (n = 248 SB – n = 161 cont)	9.72 (5.99–15.79) p < 0.0001	4.81 (2.64–8.77) p < 0.0001	1.43 (0.76–2.71) p = 0.27	2.31 (1.30–4.08) p = 0.0041
Unsure (n = 57 SB – n = 86 cont)	3.69 (1.71–7.98) p = 0.0009	1.92 (0.84–4.40) p = 0.12	1.00 (0.39–2.55) p = 1.00	0.98 (0.36–2.64) p = 0.97
Same (n = 388 SB – n = 645 cont)	3.20 (2.12–4.83) p < 0.0001	1.41 (0.89–2.22) p = 0.14	1.35 (0.85–2.16) p = 0.21	Reference

Note: Bolded values indicate statistical significance reached at the 5% level.

Abbreviations: CI, confidence interval; cont, control; SB, stillbirth.

TABLE 2.

Multivariable odds ratios for stillbirth and perception of fetal movements (prioritized variable) by customized birthweight centile quartile.

	Quartile 1 (0–24) (n = 409 SB, n = 502 cont) aOR a 95% (CI)	Quartile 2 (25–49) (n = 161 SB, n = 505 cont) aOR a 95% (CI)	Quartile 3 (50–74) (n = 127 SB, n = 474 cont) aOR a 95% (CI)	Quartile 4 (75+) (n = 145 SB, n = 498 cont) aOR a 95% (CI)
Increased strength (n = 124 SB – n = 997 cont)	0.49 (0.29–0.84) p = 0.009	0.25 (0.13–0.46) p < 0.0001	0.56 (0.31–0.99) p = 0.05	0.33 (0.18–0.61) p = 0.0004
Increased frequency (n = 25 SB – n = 90 cont)	1.76 (0.68–4.56) p = 0.24	1.11 (0.37–3.39) p = 0.85	0.27 (0.05–1.40) p = 0.12	0.79 (0.26–2.46) p = 0.69
Reduced frequency (n = 248 SB – n = 161 cont)	9.34 (5.43–16.06) p < 0.0001	6.11 (3.11–11.99) p < 0.0001	2.08 (1.00–4.36) p = 0.05	2.10 (1.10–3.99) p = 0.02
Unsure (n = 57 SB – n = 86 cont)	3.13 (1.28–7.67) p = 0.01	1.91 (0.77–4.74) p = 0.17	0.91 (0.34–2.46) p = 0.85	1.06 (0.32–3.46) p = 0.93
Same (n = 388 SB – n = 645 cont)	3.28 (2.05–5.25) p < 0.0001	1.71 (1.02–2.87) p = 0.04	1.83 (1.09–3.09) p = 0.02	Reference

Note: Bolded values indicate statistical significance reached at the 5% level.

Abbreviations: CI, confidence interval; cont, control; SB, stillbirth.

^a Adjusted odds ratio – derived from multivariable model controlling for maternal age – earliest pregnancy BMI – maternal ethnicity – parity – maternal education level – marital status – pre‐existing hypertension or diabetes – maternal smoking – recreational drug use – supine going to sleep position – vigorous fetal movements – and fetal hiccups.

The univariable generalized additive model (Figure 2) shows that other than slightly increased odds in the bottom and top 10% of births the odds of stillbirth are relatively constant across birthweight centile for those in the increased strength group. For all other groups the odds are higher, particularly below the 20th percentile. The shape of the curves differs particularly below the 50th percentile, with differences in gradients of the curves particularly in the lower percentiles. A test for an interaction between birthweight quartile and prioritized fetal movement group was statistically significant in univariable analysis (χ ² (df = 12) = 24.71, p = 0.016) and remained so in the multivariable model (χ ² (df = 12) = 23.94, p = 0.02). An additional analysis using birthweight centile continuously also found this interaction to be statistically significant (p = 0.01 in both univariable and multivariable models).

FIGURE 2.

Univariable association of stillbirth and birthweight centile stratified by prioritized fetal movement group from generalized additive models.

4. DISCUSSION

The data from this large multiethnic sample from several countries provides evidence to support that maternal perception of strength and frequency of fetal movements in later gestations does not decrease and that fetal size does not alter maternal perception of fetal movements. The decreased odds of stillbirth associated with increased strength of movement irrespective of birthweight quartile, confirms the reassuring nature of this pattern. The odds of stillbirth associated with the prioritized fetal movement variable stratified by birthweight quartile are consistent with that seen in overall analyses, ¹ and are consistent with the observed association of RFM with fetal growth restriction, placental insufficiency ¹⁸ and abnormal uterine artery Doppler indices. ¹⁹ When RFM are considered, the odds of stillbirth was highest in infants in the lowest birthweight quartile, but was also elevated in the “same” group (those with a consistent perception of fetal movements); this highlights the increased risk of stillbirth for all fetuses that are small for gestational age.

A recent study suggested a relationship between the number of gross fetal movements and increasing birthweight, but a negative relationship with gestational age. ²⁰ However, we question the methodology used in that analysis. First, it predicts a fetal movement index from a model containing gestational age and birthweight resulting in coefficients in opposite directions, which seems likely to be due to the high degree of correlation (multicollinearity) between the two variables. Second, there is no adjustment of the repeated measures on individuals, making interpretation of the findings difficult. Another study of women presenting with RFM compared to those with no RFM, investigated the cerebroplacental ratio in relation to fetal movements and found significantly reduced birthweight centiles in the RFM group. ²¹ The relationship shown was not as strong when using estimated fetal weight, potentially reflecting inaccuracies associated with this measure.

The pattern of increased susceptibility for the smaller fetus where the mother perceives a reduction in fetal movements is similar to that previously reported from the CRIBBS IPD in relation to gestational age, where the odds of stillbirth were increased for those born preterm with maternal perception of RFM. ¹ There is some overlap of fetuses in these two groups (preterm and less than 25th percentile); however, there are many babies born preterm that are not in the lowest quartile and of many born small that are not preterm, highlighting the importance and vulnerability of a fetus whether it be preterm or small at term.

From a clinical perspective we believe that the reported perception of RFM in the presence of a known small fetus should highlight the need for other clinical investigations and ongoing monitoring of the pregnancy.

This study has some methodological limitations, including the fact the data on perceptions of strength and frequency of at each gestational age was not collected longitudinally on the same women. However, given the pattern of reporting of perception of fetal movements is consistent across the individual studies, we believe the pattern reported to be that perceived longitudinally and thus is unlikely to influence the results, and aligns with qualitative reports of fetal movement through pregnancy. ⁷

We have used birthweight centiles to create the categories of fetal size, this does not necessarily reflect the fetal weight centile at the time of reporting of the perception of fetal movements. However, estimated fetal weight was not available in these datasets and this measurement is known to be imprecise with only approximately 70%–75% within 10% of birthweight and less accurate at the extremes. ²² , ²³

There are potentially inherent biases present in case–control studies, most notably selection and recall bias and measurement error, at the same time it is well acknowledged that for rare outcomes case–control studies are generally the optimal study design. In four of the studies included, a structured interviewer administered questionnaire, was undertaken by trained interviewers to reduce the risk of measurement error and the potential risk of recall bias, with interviewer and interviewee not explicitly aware of study hypotheses. Similarly, the selection method for controls was developed to reduce potential issues of this type. The use of a strata variable in the analysis to account for clustering within studies also reduces the potential influence of bias in any individual study.

This was a large international collaborative IPD meta‐analysis, with data collected from the controls who were women with ongoing pregnancies and were blinded to the study hypotheses. We believe this dataset to be the largest repository relating to fetal movements in the third trimester and the findings are generalizable given the consistency of the findings across multiple countries. The IPD analytical approach allowed adjustment for potential confounding factors which cannot be conducted in aggregate data meta‐analysis.

5. CONCLUSION

This analysis suggests that size at birth (a proxy for fetal weight) has little impact on maternal perception of fetal movements in late pregnancy. Therefore, perception of reduced frequency of fetal movements near term should not be attributed to increased fetal size.

Increased strength of fetal movements in late pregnancy should be considered a sign of fetal wellbeing irrespective of fetal size. However, RFM are associated with stillbirth, and even more so when fetal size is small. Care for all women reporting with RFM should continue to follow existing local guidelines for RFM which should include ultrasound assessment of fetal size.

AUTHOR CONTRIBUTIONS

The conceptualisation of the IPD was undertaken by JMDT, ML, RSC, AG, CRG, TS, VC, LMOB, EAM, LMEMC and AEPH. Data curation was undertaken by JW and ML. Formal analysis was undertaken by JMDT and JW. Verification of the underlying data from each of the individual studies included in the IPD has been carried out by EAM, LMEMC, AG, LMOB and AEPH, and for the IPD analysis by AEPH. JMDT, AG, CRG, LMOB, EAM, LMEMC and AEPH were responsible for the acquisition of funding for either the individual studies comprising the IPD and the IPD itself. JMDT, RSC, AG, CRG, TS, LMOB, EAM, LMEMC and AEPH were involved in the investigation and methodology of the initial studies. Development of the methodology for the analysis of the IPD additionally included JW, BFB, ML, and LMA. Writing of the initial manuscript was undertaken by JMDT, with additional input from AEPH and BFB. All other authors provided constructive feedback and editing of the manuscript on at least one occasion and have approved the final submitted manuscript. The corresponding author attests that all listed authors meet the authorship criteria and that no others meeting the criteria have been omitted. All authors read and approved the final manuscript.

FUNDING INFORMATION

The initial IPD analyses were funded by a grant from Cure Kids.

CONFLICT OF INTEREST STATEMENT

None.

Supporting information

Table S1

Thompson JMD, Heazell AEP, Cronin RS, et al. Does fetal size affect maternal perception of fetal movements? Evidence from an individual participant data meta‐analysis. Acta Obstet Gynecol Scand. 2023;102:1586‐1592. doi: 10.1111/aogs.14652