Clinical analysis of umbilical cord torsion-related prenatal stillbirth

Abstract

Objective

This study aims to investigate the high-risk factors and ultrasonic examination of umbilical cord torsion (UCT)-related prenatal stillbirth, with a specific focus on diagnostic challenges and clinical implications.

Method

We retrospectively analyzed the electronic data of 108 pregnant women(after 1:2 age propensity score matching from an initial 424 cases) with UCT who delivered from January 2013 to February 2024 in Tongji Hospital, Tongji Medical College, University of Science and Technology. To mitigate confounding and address group imbalance, a 1:2 matching principle of the same age was performed (36 stillbirths vs. 72 live fetuses). Univariate and multivariate logistic regression analyses were used to identify independent predictors. Model performance was assessed by the area under the receiver operating characteristic curve (AUC), Hosmer-Lemeshow test, and Adjusted R². The diagnostic utility of prenatal ultrasound was comprehensively evaluated.

Results

Analysis of the matched cohort confirmed balance in baseline characteristics and revealed seven significant associations. The multivariate model demonstrated excellent performance (AUC = 0.943, 95% CI: 0.895–0.991; Hosmer-Lemeshow P = 0.842). History of decreased or vanished fetal movement (OR = 64.337, 95% CI: 8.835-468.531, P < 0.001), umbilical cord root torsion at the fetal umbilical insertion site (OR = 24.426, 95% CI: 2.500-238.637, P = 0.006), and fetal growth restriction (FGR; OR = 13.292, 95% CI: 2.422–72.934, P = 0.003)were identified as the strongest independent predictors of UCT-related prenatal stillbirth. Other independent predictors (all P < 0.05) included: hydramnios (OR = 12.146, 95% CI: 1.684–87.620, P = 0.013); abnormal placental cord insertions (APCIs; OR = 0.050, P = 0.038, 95% CI: 0.003–0.841). Prenatal ultrasound exhibited a high specificity (93.1%) while a relatively weak sensitivity of 7.4% (95% CI: 3.3–14.2%), failing to detect 92.6% of UCT cases antenatally.

Conclusion

We developed and validated a highly accurate prediction model for UCT-related prenatal stillbirth, identifying a history of decreased or vanished fetal movement, umbilical cord root torsion at the fetal umbilical insertion site, and FGR as the paramount risk factors. Given the profound limitations of conventional ultrasound in prenatal detection, clinical management must undergo a critical shift: from reactive reliance on imaging to proactive management based on maternal symptomatology. Thus, decreased fetal movement should be the primary indicator for urgent assessment, while any sonographic suggestion of umbilical root torsion and FGR should trigger intensified surveillance and delivery planning.

Introduction

The umbilical cord is the only connection and vital conduit for nutrient exchange between the fetus and mother [1]. Intrauterine fetal death remains a challenging complication in obstetrics, with an incidence of approximately 1.2% [2], and abnormal umbilical cord accounting for approximately 5% of stillbirth causes [3]. Among these, umbilical cord torsion (UCT) is an important yet understudied abnormality. Despite its low incidence [4], UCT significantly influences perinatal outcomes and is notably difficult to diagnose before delivery [2, 5–7]. Moreover, UCT is often an accidental event that is challenging to prevent [2].

UCT is characterized by excessive spiral coiling beyond the physiological range, which can lead to vascular compromise through mechanical compression and thrombosis. Although detecting UCT before birth could help reduce associated perinatal mortality, its diagnosis remains challenging due to the inherent limitations of prenatal imaging and the dynamic changes in cord morphology throughout gestation.

However, current literature on UCT-related prenatal stillbirth is predominantly limited to case reports [3, 6], which insufficiently support robust risk stratification. Furthermore, considerable controversy exists regarding the diagnostic utility of prenatal ultrasound for UCT, complicating clinical decision-making and hindering the development of evidence-based management protocols.

This study aims to identify high-risk factors associated with UCT-related prenatal stillbirth and assess the diagnostic performance of prenatal ultrasound, intending to inform clinical surveillance strategies and guide preventive interventions to reduce incidence.

Methods

Study design

This retrospective cohort study aimed to identify independent risk factors for UCT-related prenatal stillbirth through multivariate modeling and evaluate the diagnostic performance of prenatal ultrasound in detecting UCT. A detailed flow chart of the study design is presented in Fig. 1.

Fig. 1.

The flow chart of the retrospective cohort study design

Study population

This retrospective cohort study adhered to STROBE guidelines for observational research. We screened all delivery records from January 2013 to February 2024 using ICD-10 codes for stillbirth (O36.74) and umbilical cord abnormalities (O69.2-O69.5), from which we identified a final cohort of 424 cases confirmed with UCT. Among them, 36 stillbirths were regarded as the stillbirth group, and among the remaining 388 live fetuses, 72 were matched at an age ratio of 1:2 as the live fetus group.

Inclusion criteria included

1. Age ≥ 20 years;

2. Gestational age ≥ 20 weeks;

3. Complete related information.

4. UCT ≥ 12 coils, or prenatal umbilical coiling index (UCI) > 0.36.

Exclusion criteria included

1. Incomplete related information;

2. Multifetal pregnancy, chromosome abnormalities, major fetal malformations, iatrogenic termination for non-medical indications;

3. Intrapartum stillbirth;

4. Patients with psychiatric diseases, cognitive disorders, or serious comorbidities;

5. The same pregnant woman with multiple UCTs.

Ethics statement

The study was reviewed and approved by the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (TJ-IRB202403027). Written informed consent was not required for the study because retrospective design and the use of fully anonymized data; all the information was obtained from the hospital’s electronic database. All procedures strictly adhered to the ethical principles of the Declaration of Helsinki.

Data collection

In our study, the data included maternal age, gravida, parity, IVF-ET, multipara, maternal complications (GDM, gestational diabetes mellitus; HDP, hypertensive disorders of pregnancy), gestational age at birth, fetal gender, fetal growth, umbilical artery blood flow, fetal movement, fetal heart rate, oligohydramnios, hydramnios, abnormal placental cord insertions (APCIs), umbilical cord around the neck, excessively long cord, excessively short cord, single umbilical artery, thin umbilical cord, umbilical thrombosis, true umbilical cord knot, number of UCT coils, the position of UCT, fetal gender.

Sonographic examinations

All sonographic examinations were performed by certified maternal-fetal medicine specialists with ≥ 5 years of experience. Blinded Assessment: Sonographers were unaware of pregnancy outcomes during image acquisition and analysis. Sonographic examinations were performed by physicians with a 5–9 MHz curved array volume transducer(GE VolusonS8, GE Voluson E8, and GE VolusonE10). The fetus and its appendages (umbilical cord, amniotic fluid, and placenta) were examined by two-dimensional ultrasound (2DUS), three-dimensional ultrasound (3DUS), and power Doppler ultrasound. The specific steps for prenatal ultrasound examination are as follows: (1) A Routine prenatal examination was performed to measure fetal size, check for fetal abnormalities, and record amniotic fluid. (2) The position where the umbilical cord was inserted into the placenta was identified, and an abnormal insertion was recorded. (3) Careful scanning was conducted along the umbilical cord to observe the umbilical cord entanglement, especially the entanglement of the umbilical cord root at the fetal umbilical insertion site. (4) The pitch values (along one side of the umbilical cord from theinner edge of the arterial or venous wall to the outer edge of the next coil) of three different umbilical cord segments(one near the fetal insertion of the umbilical cord, one near the placental insertion, and one anywhere between the two) were measured and recorded separately, and the average spiral length of the three segments was calculated. (5) The ratio of the systolic maximum blood flow velocity of the umbilical artery (S) to the end-diastolic blood flow velocity (D) was measured.

Relevant definitions and diagnostic criteria

Stillbirth is defined as the death of the fetus in utero after 20 weeks of gestation [8]and confirmed by the absence of fetal heartbeat through ultrasonography. Torsion of the umbilical cord was defined as ≥ 12coils according to a Chinese perinatology study [9] and Strong’s study [10] (torsion from 6 to 11 coils was considered as physiological), or prenatal umbilical coiling index (UCI) >0.36 [9]. Umbilical cord root torsion at the fetal umbilical insertion site is defined as the torsion where the umbilical cord is inserted into the fetus’s abdomen.

The UCI was the ratio of the total number of coils on the umbilical cord to the length of the cord, as described by Strong et al. [10]. The UCI was measured by ultrasound before delivery and calculated according to the method proposed by Sharma et al. We measured the distance between the coils, along one side of the umbilical cord, from the inner edge of the arterial or venous wall to the outer edge of the next coil. The reciprocal of this distance in cm was the ultrasonological umbilical cord coiling index [9]. The final value was the average of the three different umbilical segments(one near the fetal insertion of the umbilical cord, one near the placental insertion, and one anywhere between the two) [9]. Confirmation Criteria: All cases for diagnosis were confirmed by direct visualization after delivery. Excessively long cord was defined as an umbilical cord with a length of more than 80 cm [11]. Excessively short cord was defined as an umbilical cord with a length of less than 30 cm. FGR was defined as an estimated fetal weight (EFW) below the 10th percentile for gestational age by the Society for Maternal-Fetal Medicine (SMFM). APCIs included velamentous cord insertion and battledore placenta.

Statistical analysis

Data collection involved coding and entering participant information into IBM SPSS Statistics version 22.0. Statistical analyses were performed using both SPSS and GraphPad software. The measurement data of normal distribution were presented as mean ± standard deviation (M ± SD), and the T-test was used for comparison. The statistical data were expressed as numbers(%), and the Chi-square (χ²) test was used for comparison. The odds ratios (OR) and their 95% confidence interval (CI) of UCT-related prenatal stillbirth were computed by a multivariate logistic regression model. Significance was accepted at P < 0.05, but the multiple comparisons between groups were performed with p-value adjustments according to Bonferroni. Post-hoc power analysis determined 85% power to detect OR ≥ 3.0 (α = 0.05, two-tailed). Statistical significance was set at P < 0.05 with Bonferroni correction for multiple comparisons. Model Fit Statistics were assessed using Area Under the ROC Curve (AUC), Hosmer-Lemeshow Goodness-of-Fit Test, Adjusted R², and Events Per Variable (EPV). The diagnostic performance of prenatal ultrasound for detecting UCT was assessed in the matched cohort using sensitivity, specificity, positive and negative predictive values, and the missed diagnosis rate.

Results

Pregnancy-related characteristics of the stillbirth group

The 36 women in the stillbirth group had an average age of 24–40 years, among whom 31 (86.1%) were 24–34 years and 5 (13.9%) were ≥ 35 years. The Stillbirth group comprised 28 primiparas (77.8%) and 8 multiparas (22.2%). In the stillbirth group, there were 31 cases (86.1%) with regular birth examinations and 5 cases (13.9%) without regular birth examinations. The weeks of stillbirth ranged from 21^+ 6 to 40^+ 2 weeks, 7 cases (19.4%) from 21^+ 6 to 27^+ 6 weeks, 12 cases (33.3%) from 28 to 33^+ 6 weeks, and 17 cases (47.2%) from 34 to 40^+ 2 weeks. 25 cases had not undergone fetal heart rate monitoring in the week before the ultrasound detection of stillbirth.

Propensity score matching and cohort characteristics

The initial cohort comprised 424 pregnancies with confirmed UCT(36 stillbirths vs. 388 live fetuses). To mitigate potential confounding and address the significant imbalance in group sizes, a 1:2 age propensity score matching was performed. This process successfully created a well-balanced analytical cohort of 36 stillbirths and 72 live fetuses (Table 1). As shown in Table 1, after matching, all measured maternal baseline characteristics (including age, gravidity, parity, mode of conception, and complications) showed no statistically significant differences between the stillbirth and live fetus groups (all P > 0.05), confirming the effectiveness of the matching procedure in minimizing selection bias. Despite this balance in baseline characteristics, univariate analysis on the matched cohort revealed several significant associations. The incidence ofumbilical cord root torsion at the fetal umbilical insertion site (30.6% vs. 2.8%, P < 0.001), history of decreased or vanished fetal movement (61.1% vs. 2.8%, P < 0.001), FGR(41.7% vs. 13.9%, P = 0.001), history of fetal bradycardia(11.1% vs. 1.4%, P = 0.023), and abnormal amniotic fluid volume (oligohydramnios and hydramnios, P < 0.05) remained substantially higher in the stillbirth group. On the other hand, the incidence of APCIs(5.6% vs. 22.2%, P = 0.029) in the stillbirth group was significantly lower than that in the live fetus group (P < 0.05). There were no significant differences in the incidence of maternal factors, male fetus, history of absent or reversed end-diastolic umbilical artery blood flow, umbilical cord around the neck, excessively long cord, excessively short cord, single umbilical artery, thin umbilical cord, umbilical thrombosis, true umbilical cord knot, and UCT ≥ 30 coils between the two groups (P > 0.05).

Table 1.

Analysis of risk factors for prenatal stillbirth caused by UCT

Characteristic	The stillborn	The live fetus	t/X2	P
	Group (n = 36)	Group (n = 72)
Maternal factor
Advanced age	5 (13.9)	10(13.9)	0.000	1.000
Age	30.17 ± 3.982	30.17 ± 4.014	0.000	1.000
Gravida	1.944 ± 1.194	2.014 ± 1.369	0.271	0.787
Parity	1.278 ± 0.513	1.347 ± 0.508	0.665	0.508
IVF-ET	1 (2.8)	4(5.6)	0.419	0.517
Multipara	8 (22.2)	24(33.3)	1.421	0.233
No regular birth check-up	5 (13.9)	9(12.5)	0.041	0.840
GDM	5 (13.9)	12(16.7)	0.140	0.709
HDP	4 (11.1)	13(18.1)	0.873	0.350
Fetal factor
Male fetus	17 (47.2)	42(58.3)	1.195	0.274
FGR	15 (41.7)	10(13.9)	10.410	0.001
History of absent or reversed end-diastolic umbilical artery blood flow	2 (5.6)	1(1.4)	1.543	0.214
History of decreased or vanished fetal movement	22 (61.1)	2(2.8)	47.250	< 0.001
History of fetal bradycardia	4 (11.1)	1(1.4)	5.138	0.023
Amniotic fluid factor
Oligohydramnios	7 (19.4)	3(4.2)	6.667	0.010
Hydramnios	12 (33.3)	6(8.3)	10.800	0.001
Umbilical cord factor
APCIs	1 (2.8)	16(22.2)	4.800	0.029
Umbilical cord around the neck	7 (19.4)	18(25.0)	0.416	0.519
Excessively long cord	1 (2.8)	4(5.6)	0.419	0.517
Excessively short cord	2 (5.6)	1(1.4)	1.543	0.214
Single umbilical artery	3 (8.3)	2(2.8)	1.678	0.195
Thin umbilical cord	5 (13.9)	6(8.3)	0.810	0.368
Umbilical thrombosis	1 (2.8)	2(2.8)	0.000	1.000
True umbilical cord knot	1 (2.8)	3(4.2)	0.130	0.719
UCT ≥ 30 coils	8 (22.2)	8(11.1)	2.348	0.126
Umbilical cord root torsion at the fetal umbilical insertion site	11 (30.6)	2(2.8)	17.490	< 0.001

IVF-ET in vitro fertilization and embryo transfer, GDM gestational diabetes mellitus, HDP hypertensive disorders of pregnancy, FGR fetal growth restriction, APCIs Abnormal placental cord insertions, UCT umbilical cord torsion

Multivariate logistic regression analysis

After adjusting for confounding factors, multivariate logistic regression analysis on the propensity-matched cohort was performed to identify independent risk factors (Table 2). Multivariate logistic regression revealed that FGR, history of decreased or vanished fetal movement, hydramnios, and umbilical cord root torsion at the fetal umbilical insertion site were independent risk factors for UCT-related prenatal stillbirth (P < 0.05, OR > 1). The association between APCIs and UCT-related prenatal stillbirth was inverse(P < 0.05, OR < 1).

Table 2.

Multivariate logistic regression analysis of prenatal stillbirth with UCT

Characteristic	B	SE	Wald	P	OR	95% CI
FGR	2.587	0.869	8.872	0.003	13.292	2.422–72.934
History of decreased or vanished fetal movement	4.164	1.013	16.897	<0.001	64.337	8.835-468.531
History of fetal bradycardia	2.247	1.631	1.898	0.168	9.461	0.387-231.353
Oligohydramnios	1.978	1.109	3.182	0.074	7.230	0.823–63.548
Hydramnios	2.497	1.008	6.134	0.013	12.146	1.684–87.620
APCIs	-3.005	1.445	4.327	0.038	0.050	0.003–0.841
Umbilical cord root torsion at the fetal umbilical insertion site	3.196	1.163	7.551	0.006	24.426	2.500-238.637

Model fit: Hosmer-Lemeshow Goodness-of-Fit test : χ² = 2.731, P = 0.842; AUC = 0.943 (95% CI:0.895–0.991); Adjusted R² : 0.600; Events Per Variable (EPV): 5.14

The final model, which included seven variables, exhibited outstanding performance. It showed excellent discrimination between stillbirth and live fetus cases, with an AUC of 0.943 (95% CI: 0.895–0.991) (Fig. 2). The model also demonstrated excellent calibration, as indicated by a non-significant Hosmer-Lemeshow goodness-of-fit test result (χ² = 2.731, P = 0.842), meaning its predictions were not statistically different from observed outcomes. The model explained a substantial proportion of the variance (Adjusted R² = 0.600).

Fig. 2.

The ROC curve of multivariate Logistic regression analysis of prenatal stillbirth with UCT. The red line represents predicted probability, AUC = 0.943 (95% CI: 0.895–0.991). FGR: AUC = 0.639 (95% CI:0.523–0.755); History of decreased or vanished fetal movement: AUC = 0.792(95% CI:0.688–0.895); History of slow fetal heart rate: AUC = 0.549 (95% CI:0.430–0.667); Oligohydramnios: AUC = 0.576 (95% CI:0.457–0.695); Hydramnios: AUC = 0.625 (95% CI:0.507–0.743); APCIs: AUC = 0.424 (95% CI:0.313–0.534); Umbilical cord root torsion at the fetal umbilical insertion site: AUC = 0.639 (95% CI:0.520–0.758)

Diagnostic performance of prenatal ultrasound

The comprehensive diagnostic performance of prenatal ultrasound is detailed in Table 3. The sensitivity was critically low in both the stillbirth (8.3%, 95% CI: 1.8–22.5%) and live birth (6.9%, 95% CI: 2.3–15.5%) groups, resulting in an overall missed diagnosis rate of 92.6%. The specificity, however, was high (93.1%, 95% CI: 84.5–97.7%), and the positive predictive value was 100%, indicating that when UCT was suspected on ultrasound, it was always confirmed postnatally. The negative predictive value was 40.2%, highlighting the test’s profound inability to rule out the condition. Ultrasound diagnosed UCT (Fig. 3a, b and c). Ultrasound missed UCT (Fig. 4).

Table 3.

Diagnostic performance of prenatal ultrasound for detecting UCT in the matched cohort

Parameter	The stillborn group (n = 36)	The live fetus group (n = 72)	Overall (n = 108)
True Positive (TP)	3	5	8
True Negative (TN)	N/A	67	67
False Negative (FN)	33	67	100
False Positive (FP)	0	0	0
Sensitivity (95% CI)	8.3% (1.8–22.5%)	6.9% (2.3–15.5%)	7.4% (3.3–14.2%)
Specificity (95% CI)	N/A	93.1% (84.5–97.7%)	93.1% (84.5–97.7%)
Positive Predictive Value (95% CI)	100% (29.2–100.0%)	100% (47.8–100.0%)	100% (63.1–100.0%)
Negative Predictive Value (95% CI)	8.3% (6.8–10.1%)	50.0% (43.9–56.1%)	40.2% (34.5–46.1%)
Missed Diagnosis Rate	91.7% (33/36)	93.1% (67/72)	92.6% (100/108)
Accuracy	8.3% (3/36)	100% (72/72)	69.4% (75/108)

All values are presented with clear denominators to specify the sample size for each calculation. In the stillbirth group (n = 36), all cases were confirmed to have UCT postnatally; thus, no true negative cases exist, and specificity and NPV are not applicable (N/A). The NPV value shown (8.3%) is an arithmetical result of 1-prevalence in this cohort and should not be interpreted as a valid NPV. For the live fetus group(n = 72), true negatives (n = 67) were defined as cases without a prenatal ultrasound diagnosis of UCT and without postnatal confirmation, allowing valid estimation of specificity and NPV. No false positive cases occurred in this study; all prenatally suspected UCT cases were confirmed after delivery

Fig. 3.

Live fetus with UCT at 37⁺⁶ weeks. UCT images by two-dimensional ultrasound (a) and color Doppler ultrasound diagnosis (b), postnatal stereogram of UCT (c)

Fig. 4.

UCT-related prenatal stillbirth at 27⁺⁴ weeks. Postnatal stereogram of UCT, including umbilical root torsion at the fetal umbilical insertion site (arrow)

Discussion

The umbilical cord has one umbilical vein and two umbilical arteries. The umbilical artery is longer than the umbilical vein, while the lumen diameter of the umbilical vein is larger than that of the umbilical artery, and its vascular wall is thinner than that of the umbilical artery. The umbilical artery forms a physiological helix around the umbilical vein. Three blood vessels pass through the length of the umbilical cord in a spiral or coiled fashion. The spiral fusion of these umbilical blood vessels is called the spiral course [12]. The umbilical cord helix is one of the crucial characteristics of the umbilical cord [13], as it protects and supports the blood vessels of the umbilical cord. The causes of umbilical coil formation are unknown, and these hypotheses include: active or passive torsion of the embryo, movement and rotation of the fetus along the long axis of the umbilical cord, disproportionate growth of umbilical blood vessels, fetal hemodynamics, and arrangement of muscle fibers in the umbilical artery wall [14, 15]. The physiological coil of the umbilical cord spans approximately 6–11 weeks with no adverse effects on the fetus. The UCT is formed if the umbilical cord coils for more than 11 weeks. The underlying mechanism of UCT remains unclear [2]. It may be associated with umbilical cord dysplasia, the sporadic thickness of Wharton’s jelly, and excessive fetal movements [16]. The umbilical cord is rich in Wharton’s jelly, which comprises a substantial amount of collagen and elastic fibers, which can protect and support blood vessels, and prevent the umbilical cord from being compressed, ensuring continuous blood flow [17]. The weakness is often distorted and narrowed in the absence of Watton’s jelly locally [17]. UCT can cause twisting, narrowing, and occlusion of umbilical vessels, interruption of blood flow, and thrombosis [16–18]. It can also result in chronic hypoxia with severely reduced blood flow, oligohydramnios, and fetal growth retardation [3]. Complete blockade of the umbilical cord can also severely obstruct fetal-placental circulation with subsequent fetal death [2, 5, 18].

Herein, UCI was used to assess the extent of UCT. UCI can directly indicate the number of weeks of UCT and indirectly indicate the density of the umbilical cord helix. UCI was measured by ultrasound before delivery because postpartum UCI lacks the segment of the umbilical cord closest to the fetus, which tends to be more twisted than the part near the placental insertion [19]. In addition, antenatally the cord is more filled with blood, making the helix denser due to the intrinsic twist in the vessels [19]. After the umbilical cord is severed postpartum, the blood in the umbilical cord decreases, the umbilical cord contracts, and the umbilical cord helix becomes less dense than before birth. Prenatal diagnosis of UC currently has shortcomings. The diagnosis is often confirmed after birth [18].

The main condition in the group of UCT-related prenatal stillbirth was as follows: the rate of older pregnant women was lower than that of pregnant women of appropriate age, and the rate of multipara was lower than that of primipara. Unscheduled birth check-ups were fewer compared to regular birth check-ups, and even regular birth check-ups could not prevent stillbirth, accounting for 86.1%. The rate of stillbirth at 34–40^+ 2 weeks was the highest, reaching 47.2%. Pregnancy outcomes can be improved if such pregnant women can be timely admitted to a hospital and terminate their pregnancy.

Research reports indicate that old age is associated with UCT [9, 20]. In the present study, the stillbirth group and the live fetus group were matched at a ratio of 1:2 for the same age, and there was no significant correlation between the two. Research will be conducted in the future with a larger sample size. Studies have reported a more relaxed abdominal wall of multipara and a wider uterine cavity, which is conducive to the free movement of the fetus and the formation of UCT [9]. The rate of multipara in the stillbirth group was lower than that in the live fetus group, and no significant correlation was noted between the two. Stillbirth can occur in pregnant women without regular birth check-ups. The rate of non-regular birth check-ups in the stillbirth group was higher than that in the live birth group; however, the two were not statistically significant. Regular prenatal check-ups cannot completely prevent the occurrence of UCT-related prenatal stillbirth, and even stillbirth can occur within 3 days of prenatal examination, which is the current contradiction between doctors and patients. GDM is a key risk factor for UCT, which has a damaging effect on umbilical cord blood vessels and Wharton’s jelly [21, 22]. The incidence of GDM in the stillbirth group was lower than that in the live birth group, and no significant correlation was noted between the two. No relevant studies have been found in previous literature, and the sample size needs to be expanded for further confirmation.

A few studies argue that male fetuses are active and more prone to UCT. Here, the rate of male fetuses in the stillbirth group was lower than that in the live fetus group, and no statistical difference was observed between the two groups. Notably, FGR is one of the most prevalent complications in obstetrics and is often associated with premature birth and stillbirth. The essence of FGR is that the fetus is in a chronic state of hypoxia and malnutrition for a long time. FGR is mostly caused by uterine-placental insufficiency. When a fetus with FGR encounters a UCT event, it lacks sufficient metabolic reserves to cope with the momentary hypoxia, and the compensatory mechanism collapses subsequently. Acute hypoxia can rapidly lead to fetal acidosis, inhibit myocardial contractility, cause severe bradycardia, and eventually develop into irreversible brain damage and intrauterine fetal death.

Decreased or vanished fetal movement and slow fetal heart rate are early warning signs of intrauterine abnormalities, which may be caused by abnormal fetal blood supply. In this case, reduced fetal movement occurs even if the umbilical artery blood flow is not fully blocked [23]; the next step is interrupted blood flow, which results in stillbirth. History of decreased or vanished fetal movement was the strongest predictor in our model (OR = 64.337, P < 0.001). This result held immense statistical power even after adjusting for all other sonographic and clinical confounders. While decreased fetal movement was a non-specific symptom associated with various pathologies, its immense effect size and independent significance in our analysis underscore its critical role as a cardinal warning sign of acute fetal compromise. In some cases, it might be the paramount clinical sentinel. It might be the fetus’s most direct and timely alarm signal. Fetal movement can predict fetal distress or near-death earlier than fetal heart rate. Fetal movement can be reduced to disappear, which may last a few days to approximately one week; however, it may also quickly disappear. Nevertheless, the interval between the disappearance of fetal movement and that of the fetal heart can be up to 12–48 h. The incidence of stillbirth reduces if the pregnancy is terminated when the fetal heart rate remains good. In the Logistic regression analysis of this study, there was no statistical significance in the history of fetal bradycardia between the two groups. This might be related to the fact that among the stillbirth group, 25 cases had not undergone fetal heart rate monitoring in the week before the ultrasound detection of stillbirth, which largely led to the missed diagnosis of cases with a history of fetal bradycardia. Therefore, obstetricians need to take seriously and respond immediately to pregnant women’s complaints of reduced fetal movement. When the fetal movement or the fetal heart rate is abnormal, the fetal situation should first be established, a timely ultrasound examination should be performed, and the umbilical cord should be carefully observed to improve the pregnancy outcome as far as possible.

Amniotic fluid is an indispensable substance in the uterus that ensures normal fetal development. UCT obstructs umbilical blood flow and decreases the circulation of the placenta; the fetal blood circulation is redistributed to maintain blood supply to the brain and heart, whereas the renal blood flow and the fetal urine production are reduced, causing oligohydramnios, which also weakens the buffering effect of amniotic fluid. The periuterine pressure directly acts on the fetal body when the uterine wall is close to the fetus and the uterus contracts, increasing the chance of umbilical cord compression, thereby influencing the fetal placental circulation, and further causing fetal distress or even death. Although the association between oligohydramnios and stillbirth did not reach statistical significance in our model (P = 0.074), the magnitude of the effect (OR > 7) merits clinical attention. This finding was biologically plausible, as oligohydramnios could be both a cause and a consequence of worsened umbilical blood flow in the setting of UCT. The wide confidence interval likely reflects limited statistical power for this outcome. Nonetheless, the presence of oligohydramnios in pregnancies with suspected UCT should raise clinical concern and justify intensified surveillance, despite the borderline significance. Future studies with larger samples are needed to confirm this association.

In the case of hydramnios, the fetus had more room for movement, and its movements are freer, and the range of motion was greater. For a fetus that has already experienced UCT, a larger space for movement means that it may be more likely to perform actions that exacerbate the torsion (such as rolling and spinning), thereby increasing the number of twists of the umbilical cord or twisting it tighter, further exacerbating the blood circulation disorder. At the same time, UCT causes the umbilical cord to compress and narrow, thus obstructing the venous return, further causing increased leakage and hydramnios. Pilliod RA et al.reported an increased incidence of stillbirth in pregnancies with hydramnios. Although the underlying cause is unknown, this will be the focus of future research [24].

Studies have shown that excessively long cord is associated with UCT [9]. We found that the incidence of excessively long cord was lower in the stillbirth group than in the live fetus group; however, the two was not statistically significant. Univariate analysis revealed that the incidence of excessively short cord in the stillbirth group was higher than that in the live fetus group, whereas multivariate logistic regression showed no statistical difference. Further study should be conducted with larger sample sizes in the later stage. Studies have shown that a single umbilical artery is related to UCT [15]. The incidence of a single umbilical artery was higher in the stillbirth group than in the live fetus group; however, there was no statistical significance between the two groups. It has been reported that when the umbilical cord is twisted ≥ 30 coils, the stillbirth rate is high. The rate of UCT ≥ 30 coils was higher in the stillbirth group than in the live fetus group; nonetheless, the two were not statistically significant.

Umbilical cord root torsion at the fetal umbilical insertion site can result in cord root thinning and then cord-like necrosis, causing blood vessel occlusion or thrombosis, hence blocking the cord blood flow; the fetus eventually dies due to blood flow interruption. Other studies have revealed occurrences of intrauterine fetal death due to umbilical cord root torsion at the fetal umbilical insertion site [2, 16]. The significant finding of our study is the identification of umbilical cord root torsion at the fetal umbilical insertion site as a critical independent risk factor (OR = 24.426, P = 0.006). This location-specific vulnerability is mechanistically plausible. The fetal end of the cord, particularly within 5 cm of the abdominal wall, is known to have a relative paucity of Wharton’s jelly, the protective gelatinous tissue that cushions the umbilical vessels. This anatomical peculiarity renders this segment highly susceptible to the mechanical forces of excessive torsion, leading to vascular compression, thrombosis, and acute interruption of fetoplacental circulation. Our findings align with pathological studies that have identified cord root torsion as a frequent culprit in cases of sudden, unexplained stillbirth. This result shifts the clinical focus from the mere number of coils to their location, emphasizing the necessity of meticulous sonographic examination of the umbilical cord root torsion at the fetal umbilical insertion site, despite the technical challenges involved.

Therefore, torsion location has a greater impact on the fetus than the number of twists. A careful examination of the umbilical cord in the fetal umbilical insertion site(where the umbilical cord enters the abdomen of the fetus) is necessary to prevent excessive twisting of the umbilical cord in such fetuses that could result in umbilical vessel blockage or even rupture. Besides, there is a need to identify the right time to terminate the pregnancy to prevent the risk of intrauterine fetal death [13].

Unexpectedly, a significant inverse association was observed between APCIs and UCT-related prenatal stillbirth (OR = 0.050, P = 0.038). We hypothesize that this is less likely a direct biological protective effect and more likely a potential consequence of surveillance bias. This is plausible given that APCIs are readily detectable on prenatal ultrasound, typically leading to classification as high-risk pregnancies. Therefore, these patients likely received more frequent and stricter prenatal monitoring (for example, increased ultrasound surveillance, fetal heart rate monitoring, and even possibly earlier hospitalization for delivery). This enhanced medical intervention might have successfully prevented prenatal stillbirths that could have been caused by UCT, thereby making the ultimate stillbirth rate in this group lower than that of pregnant women who were not identified as high-risk but actually had problems of UCT. Although no comparable studies are available in the literature to directly support this observation, our results underscore the potential value and clinical importance of intensified surveillance. This surveillance bias, which may be caused by enhanced monitoring, requires further exploration and verification.

The exceptional performance of our multivariate model (AUC = 0.943, Hosmer-Lemeshow test P = 0.842) confirms the robustness of our findings. An AUC exceeding 0.90 is considered outstanding in clinical prediction models, indicating an excellent ability to discriminate between pregnancies that will and will not result in UCT-related stillbirth. Furthermore, the excellent calibration (P = 0.842) indicates that the model’s predicted probabilities closely align with the observed outcomes. This validation, achieved with a balanced cohortmitigates concerns about overfitting and reinforces the credibility of the identified risk factors for clinical application, particularly the paramount importance of umbilical cord root torsion at the fetal umbilical insertion site, history of decreased or vanished fetal movement, and FGR.

Our study quantifies the profound inadequacy of current prenatal ultrasound for detecting UCT. The overall sensitivity was a mere 7.4%, with a missed diagnosis rate in stillbirths of 91.7% (Table 3), consistent with literature reports of 85–95%.

The main reasons for missed diagnosis of UCT in ultrasonography include: (1) Prenatal ultrasonography hardly displays the whole and complete umbilical cord [16], particularly in the third trimesterdue to the large gestational age and fetal occlusion; (2)Influenced by amniotic fluid volume, fetal position, anterior wall placenta, and abdominal wall fat layer of pregnant women, part of the umbilical cord is difficult to display, especially the umbilical cord root at the fetal umbilical insertion site. (3) After intrauterine death, the blood in the umbilical cord stagnates, and ultrasound examination of umbilical blood flow shows no blood flow signal, and the difficulty of ultrasound observation of UCT increases. This creates a critical clinical paradox: we can identify who is at high risk, but we cannot reliably visualize the culprit lesion itself. Thus, there is a need to continuously improve the scanning skills during the actual examination, and perform multi-section and multi-angle scanning, especially the umbilical root at the fetal umbilical insertion site, to improve the accuracy of ultrasonic diagnosis of UCT.

Given the low sensitivity of conventional ultrasound, future research should explore focused Doppler interrogation at the fetal umbilical insertion site. Measuring Doppler indices such as the systolic/diastolic ratio, resistance index, and pulsatility index in this region may help identify hemodynamic alterations suggestive of vascular narrowing, occlusion, or flow interruption prior to overt morphological changes, potentially improving antenatal diagnosis and enabling timely intervention [23].

Strengths and limitations

This study is the first retrospective study with a large sample size to explore the high-risk factors of UCT-related prenatal stillbirth and ultrasound examination. Admittedly, our research provides valuable insights into UCT-related stillbirth. However, this study has some limitations as follows: (1) Although the data from our hospital over 11 years were reviewed, the number of research subjects in the stillbirth group was still relatively small, which might affect the reliability of statistical comparisons, especially for rare outcomes. In addition, a large proportion of pregnant women only received a diagnosis of stillbirth at our hospital’s outpatient department and ultimately chose to be hospitalized at their local hospitals for induction of labor. This might lead to the omission of cases of UCT among such stillbirths. However, we included all hospitalized positive cases during the study period without any loss in follow-up and exceeding the estimated sample size for sufficient representation. (2) Because of the retrospective nature, some interesting data might not be available or might be less reliable. The retrospective nature of this study introduced the potential for unmeasured confounding and selection bias. Even if we employed propensity score matching to balance key covariates (maternal age) and multivariate adjustment. (3) Although we excluded major confounding factors, such as medical conditions and multifetal pregnancies. Residual confounding from unrecorded variables might still persist; other potential confounders, including socioeconomic background, cultural factors, and education, were not adjusted for in a multivariate analysis. (4) This was a single-center study involving cases from one hospital; the generalizability of our findings may be limited. But it could provide a foundation for future research. (5) The high missed diagnosis rate of UCT-related prenatal stillbirth was also a limitation. Notwithstanding these limitations, we employed robust statistical methodologies to mitigate their impact by propensity score matching. The exceptional performance of our final model (AUC = 0.943) indicated that we had accurately identified the key factors that cause UCT-related prenatal stillbirth. Future prospective multicenter studies with standardized protocols for fetal movement monitoring and advanced ultrasound techniques are warranted to validate our findings and further refine the prediction model.

Conclusion

In conclusion, ultrasound doctors should focus on observing the degree of umbilical cord helix in pregnant women with the above high-risk factors, particularly the umbilical cord root at the fetal umbilical insertion site particularly to improve the diagnosis of UCT. UCT-related stillbirth remains largely unpredictable with existing diagnostic tools, highlighting the urgent need for technological innovations in prenatal cord visualization. Most affected pregnancies exhibit no substantial hemodynamic compromise even if the umbilical cord is twisted, and can achieve normal delivery without severe neonatal complications [15, 25]. Nonetheless, such pregnant women warrant high-risk management emphasizing fetal movement monitoring. Early risk recognition, heightened maternal awareness of fetal activity, and well-timed delivery are essential to reducing prenatal stillbirth incidence.

Abbreviations

IVF-ET

In vitro fertilization and embryo transfer

GDM

Gestational diabetes mellitus

HDP

Hypertensive disorders of pregnancy

FGR

Fetal growth restriction

ACPIs

Abnormal placental cord insertions

UCT

Umbilical cord torsion

OR

Odds ratio

US

Ultrasound

ROC

Receiver Operating Characteristic

AUC

Area Under Curve

EPV

Events Per Variable

SD

Standard deviation

CI

Confidence interval

TP

True Positive

TN

True Negative

FN

False Negative

FP

False Positive

Authors’ contributions

X.Z.W. contributed to conceptualization, proposal development, acquisition of data, data validation, data analysis，manuscript writing; D. Z. and X.W.S. contributed to acquisition of data, data validation，acquisition of images, final approval; X.Y.X. contributed to data analysis, formal analysis，manuscript revision, fund support，final approval; G.A. contributed to data analysis, formal analysis, manuscript revision, final approval; All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Key Research and Development Plan Projects in Hubei Province(2022BCA041).

Data availability

Data from this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (Research ID: TJ-IRB202403027, Date of Approval 29 March 2024).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.