Return of the intestinal loop to the abdominal coelom after physiological umbilical herniation in the early fetal period

Abstract

The intestine elongates during the early fetal period, herniates into the extraembryonic coelom, and subsequently returns to the abdominal coelom. The manner of herniation is well‐known; however, the process by which the intestinal loop returns to the abdomen is not clear. Thus, the present study was designed to document and measure intestinal movements in the early fetal period in three dimensions to elucidate the intestinal loop return process. Magnetic resonance images from human fetuses whose intestinal loops herniated (herniated phase; n = 5) while returning to the abdominal coelom [transition phase; n = 3, crown–rump length (CRL)] 37, 41, and 43 mm] and those whose intestinal loops returned to the abdominal coelom normally (return phase; n = 12) were selected from the Kyoto Collection. Intestinal return began from proximal to distal in samples with CRL of 37 mm. Only the ileum ends were observed in the extraembryonic coelom in samples with CRLs of 41 and 43 mm, whereas the ceca were already located in the abdominal coeloms. The entire intestinal tract had returned to the abdominal coelom in samples with CRL > 43 mm. The intestinal length increased almost linearly with fetal growth irrespective of the phase (R ² = 0.90). The ratio of the intestinal length in the extraembryonic coelom to the entire intestinal length was maximal in samples with CRLs of 32 mm (77%). This ratio rapidly decreased in three of the samples that were in the transition phase. The abdominal volumes increased exponentially (to the third power) during development. The intestinal volumes accounted for 33–41% of the abdominal volumes among samples in the herniated phase. The proportion of the intestine in the abdominal cavity increased, whereas that in the liver decreased, both without any break or plateau. The amount of space available for the intestine by the end of the transition phase was approximately 200 mm³. The amount of space available for the intestine in the abdominal coelom appeared to be sufficient at the beginning of the return phase in samples with CRLs of approximately 43 mm compared with the maximum intestinal volume available for the extraembryonic coelom in the herniated phase, which was 25.8 mm³ in samples with CRLs of 32 mm. A rapid increase in the space available for the intestine in the abdominal coelom that exceeded the intestinal volume in the extraembryonic coelom generated an inward force, leading to a ‘sucked back’ mechanism acting as the driving force. The height of the hernia tip increased to 8.9 mm at a maximum fetal CRL of 37 mm. The height of the umbilical ring increased in a stepwise manner between the transition and return phases and its height in the return phase was comparable to or higher than that of the hernia tip during the herniation phase. We surmised that the space was generated in the aforementioned manner to accommodate the herniated portion of the intestine, much like the intestine wrapping into the abdominal coelom as the height of the umbilical ring increased.

Introduction

The intestine elongates considerably during the embryonic period, involving two drastic phenomena: (1) herniation into the extraembryonic coelom and (2) return into the abdominal coelom (Meckel, 1817; Mall, 1898). The timing of herniation has been analyzed in several reports (Mall, 1898; Frazer & Robbins, 1915; Snyder & Chaffin, 1954; O'Rahilly & Müller, 2001). Through a histological analysis, Kim et al. (2003) reported that intestinal loop formation begins during Carnegie stage (CS) 14, with herniation beginning around CS16 and reaching its maximum in CS23. We also precisely analyzed the herniation and rotation of the intestinal loop (Ueda et al. 2016), as physiological umbilical hernias are noticeable after CS17. Soffers et al. (2015) developed hierarchical models of the midgut, superior mesenteric artery (SMA), and mesentery. Their models differed from the classical en bloc rotation model, which described the herniation and return of the intestine into the abdominal coelom.

The processes that promote the return of the intestine into the abdominal coelom are not fully known. There is some consensus regarding the timing of its return into the abdominal coelom (Mall, 1899; Snyder & Chaffin, 1952; Kim et al. 2003), which has been reported to occur when the crown–rump length (CRL) of the fetus is approximately 40 mm. Previous sonographic studies have shown that fetuses with a crown–rump length (CRL) of 38 mm or less (< 11 weeks’ gestation according to the last menstrual period) demonstrated intestinal herniation, but this was not evident in fetuses with CRLs > 44 mm (> 11.1 weeks’ gestational age) (Cyr et al. 1986; Bowerman, 1993). However, very few samples in the transition phase, during which the intestine exists in both the extraembryonic coelom and the abdominal coelom, have been studied. Hence, the mechanism in that phase remains unclear. To use Mall's term, herniated intestines are suddenly ‘sucked back’ into the abdominal coelom (Mall, 1898; Frazer & Robbins, 1915) during the transition. Although it is somewhat easier to understand how the intestine exits the abdominal coelom to enter the extraembryonic coelom, it is extremely difficult to grasp how and why it returns.

An understanding of intestinal development is undoubtedly important because failure of this process can result in malformations such as omphalocele, non‐rotation, malrotation, and sub‐hepatic cecum. Quantitative descriptions of these processes that are based on three‐dimensional (3D) images are needed. Thus, the present study was designed to document and measure the intestinal movements during the intestinal return process using in 3D fetal images obtained with magnetic resonance (MR) microscopy.

Methods

Human embryo specimens

The Kyoto Collection, comprising approximately 44 000 human embryos, is stored in the Congenital Anomaly Research Center of Kyoto University (Nishimura et al. 1968; Shiota et al. 2007). In most cases, the embryos were received from pregnancies that were terminated during the first trimester for socioeconomic reasons under the Maternity Protection Law of Japan. Parents provided their verbal informed consent to have the specimens deposited in the collection, and consent was documented in each medical record. Written consent was not obtained from all parents. The samples were collected from 1963 to 1995 according to the regulations of each time period. The present study, including the consent procedure, was approved by the Committee of Medical Ethics of Kyoto University Graduate School of Medicine, Kyoto, Japan (E986, R0316). The samples were anonymized and de‐identified. Aborted embryos brought to the laboratory were measured, examined, and staged using the O'Rahilly & Müller method (1987, 2010). The Kyoto Collection primarily comprises samples in the embryonic period between CS13 and CS23 (approximately 6–10 weeks’ gestation, according to the last menstrual period). Approximately 100 samples with CRLs of 30–69 mm (approximately 10–13 weeks’ gestation, according to the last menstrual period) were stored intact without prior dissections. We selected 17 samples that had been artificially aborted and exhibited no obvious damage or anomalies, especially in the umbilicus. Three samples with CRLs < 30 mm (CS22 and CS23) were selected from those used in a previous study (Ueda et al. 2016). Physiological herniation was observed in eight samples. Five samples were in the herniation phase (CRL 22–32 mm), during which most of the intestine is in the extraembryonic coelom, and three samples were in the transition phase, during which the intestine is in the process of returning to the abdominal coelom (CRL 37, 41, and 43 mm). There was no evidence of physiological herniation in 12 of the samples (return phase; CRL 44–69 mm).

Image acquisition

Magnetic resonance images were acquired using a 7‐T MR system (BioSpec 70/20 USR; Bruker Biospin MRI GmbH, Ettlingen, Germany) with 35‐ and 72‐mm‐diameter ¹H quadrature transmit‐receive volume coils (T9988 and T9562; Bruker Biospin MRI GmbH) and a 19‐mm‐diameter transmit‐receive solenoid coil (Takashima Seisakusyo Co., Ltd, Tokyo, Japan). The 3D T1‐weighted images were acquired using a fast low‐angle shot‐pulse sequence with the following parameters: repetition time, 30 ms; echo time, 3.637–7.002 ms; flip angle, 30–45°; field of view, 22.5 × 15.0 × 15.0 to 54.0 × 36.0 × 36.0 mm³; matrix size, 636 × 424 × 424 to 768 × 512 × 512; isotropic spatial resolution, 35.4–70.3 μm.

Image analysis, anatomical landmarks, and intestinal position

The intestinal segments [duodenum, small intestine, colon (large intestine), liver, and abdominal coelom] were reconstructed in each image set using amira Software Suite, Version 5.5.0 (Visage Imaging, Berlin, Germany) (Fig. 1A). The amiraskel software module was used to determine the center‐line length of the intestine. The 3D coordinates were initially assigned using amira to examine the positions of particular voxels in the 3D images.

Figure 1.

Orthogonal coordinates used in the present study. (A) Left lateral view showing the x‐ and z‐axes and the origin (Th8). (B) Ventral view indicating the x‐ and y‐axes and the origin (Th8). (C) Midsagittal view of an MR image showing the abdominal morphometry. (D) Transverse view of an MR image showing the abdominal morphometry. The vertebral column was used to determine the median plane and the x‐axis. Lines corresponding to the landmarks on the ventral cranial site at the 8th thoracic and 5th lumbar vertebrae (Th8 and L5, respectively) are defined as the x‐axis. Th8 is defined as the origin. up1: the abdomen at the cranial site of the umbilical ring; up2: the abdomen at the caudal site of the umbilical ring; Hup1: height of up1 in the x‐y plane; Hup2: height of up2 in the x–y plane; MR, magnetic resonance.

The volumes of the abdominal coelom (Vabd), liver (Vliver), and extraembryonic intestine were calculated from a 3D‐reconstructed image using the amira software. The volume of the intestine in the abdominal coelom (Vintest) was estimated by subtracting Vliver from Vabd. The width (greatest latero‐lateral length, or Awidth) and height (greatest anteroposterior length, or Aheight) of the 3D‐reconstructed image of the abdominal coelom was measured. The volume of the entire intestine was calculated as the sum of the Vintest and the volume of the extraembryonic intestine.

The vertebral column was used to determine the median plane and x‐axis (Fig. 1A,B).

Lines corresponding to the landmarks on the ventral cranial site at the 8th thoracic and 5th lumbar vertebrae (Th8 and L5, respectively) were defined as the x‐axis. Th8 was defined as the origin. The normal vector from the abdomen at the cranial site of the umbilical ring (up1) was defined as the z‐axis. The inner product of the x‐ and z‐axes was defined as the y‐axis.

The umbilical ring up2 was defined as the caudal site of the umbilical ring in the midsagittal plane. The heights of up1 and up2 (Hup1 and Hup2, respectively) in the x–y plane were measured and compared with the height of the intestinal loop hernia tip (Fig. 1C). Uwidth was defined as the greatest latero‐lateral length in the section parallel to the y–z plane, including up1 of the abdominal coelom.

The small intestine between the pyloric antrum and the cecum was divided into five equal parts colored red, yellow, green, blue, and purple from the oral to the anal sides, in that order. The large intestine from the cecum down was colored black.

Results

Intestinal length and volume in the extraembryonic and abdominal coelom

The entire intestinal length increased almost linearly with fetal growth, irrespective of the phase of the samples (R ² = 0.90) (Fig. 2A). Herniation into the extraembryonic coelom was observed in eight samples, five that (CRL 22–32 mm) were in the herniation phase and three that were in the transition phase. The intestinal length in the extraembryonic coelom and the percentage of the total intestinal length that was in the extraembryonic coelom were 22.9 mm and 65%, respectively, in samples with CRLs of 22 mm. These values reached their maximum in samples with CRLs of 32 mm (63.5 mm and 77%, respectively). The intestinal length in the extraembryonic coelom rapidly decreased in the three samples in the transition phase. Herniation was not detected in 12 samples (CRL 44–69 mm) in the return phase.

Figure 2.

Length of the intestine with fetal growth. (A) Lengths of the entire intestine (blue) and the intestine in the extraembryonic coelom (red). Data from the transition phase are indicated by an asterisk. (B) Volumes of the entire intestine (blue) and the intestine in the extraembryonic coelom (red). Data from the transition phase are indicated by an asterisk. The volume of the entire intestine is calculated as the sum of Vintest and the volume of the intestine in the extraembryonic coelom.

The volume of the entire intestine increased exponentially (to the third power) during development (Fig. 2B). The volume of the entire intestine at the end of transition phase (CRL 43 mm) reached 199.2 mm³. The intestinal volume in the embryonic coelom and the percentage of the entire intestinal volume that was in the extraembryonic coelom were 8.0 mm³ and 20%, respectively, in samples with CRLs of 22 mm. These values reached their maximum in samples with CRLs of 32 mm (25.8 mm³ and 33%, respectively). The intestinal volume in the extraembryonic coelom decreased in the three samples in the transition phase.

Growth of the abdominal coelom and change in the intestinal volume

The growth of the abdominal coelom was estimated by Awidth, Aheight, and volume. (Fig. 3A,B). Awidth and Aheight increased almost linearly (R ² = 0.83 and 0.83), irrespective of phase. Vabd increased exponentially (to the third power) during development. Both Vliver and Vintest increased in a similar fashion. Vliver was greater than Vintest in most samples, except in four large samples (CRL 56.5, 57, 65, and 60 mm).

Figure 3.

Volume of the liver and intestine in the abdominal coelom with fetal growth. (A) Morphometry of the abdominal coelom. Aheight: height of the abdominal coelom; Awidth: width of the abdominal coelom. (B) Volume of the abdominal coelom (Vabd), liver (Vliver), and intestine (Vintest). (C) Percentage of Vliver/Vabd and Vintest/Vabd.

The liver occupied a large volume in the abdominal coelom, while the remaining volume of the intestine accounted for 33–41% of the abdominal volume among the samples in the herniated phase. The proportion of the liver in the abdominal cavity decreased linearly, whereas that of the intestine increased linearly; neither showed any break or plateau (Fig. 3C).

Heights of the hernia tip and umbilical ring

Hup1 ranged from 3.3 to 5.4 mm among samples in both the herniation and transition phases, and from 6.1 to 10.7 mm among samples in the return phase (Fig. 4A). Height increased in a stepwise manner between the transition and return phases. Hup2 changed in a manner similar to that of Hup1. The height of the hernia tip was 6.2 mm in samples with CRLs of 22 mm and increased to 8.9 mm in samples with CRLs of 37 mm at maximum. Notably, the height of the umbilical ring (Hup1 and Hup2) in the return phase was comparable to or higher than that of the hernia tip in the herniation phase (Fig. 4A).

Figure 4.

Heights of the hernia tip and the umbilical ring with fetal growth. (A) Height of up1 (Hup1) (blue), up2 (Hup2) (red), and hernia tip (triangle) (green). Data from the transition phase are indicated by an asterisk. (B) Percentage of Hup2/Hup1. (C) Percentage of Hup1/Aheight and Hup2/Aheight. (D) Uwidth with fetal growth. (E) Percentage of Uwidth/Awidth. Aheight: height of abdominal coelom; Uwidth: the greatest latero‐lateral length in the section parallel to the y–z plane and including up1 of the abdominal coelom.

Hup2 was 67–78% of Hup1 in the smaller samples (CRL 22–27 mm), which implies that the umbilical ring was tilted caudally (Fig. 4B). The distal segments became comparable in height in the larger samples, although the heights of the cranial segment of the umbilical ring were larger than those of the caudal segment in almost all samples. The Hup1/Aheight and Hup2/Aheight ratios increased in a stepwise manner between the transition and return phases (Fig. 4C). Uwidth increased almost linearly with fetal growth, irrespective of the phase (Fig. 4D). The Uwidth/Awidth ratio was 50–95%, irrespective of the phase (Fig. 4E).

Distribution of the intestines in the extraembryonic and abdominal coeloms

The 3D anatomy of the intestinal tract reconstructed using amira is shown in Fig. 5 and Supporting Information Movies [Link], [Link], [Link], [Link] The intestine is highlighted from the oral side to the anal side in the following order: red, yellow, green, blue, purple, and black lines. The end of the duodenum was attached to the posterior abdominal cavity wall. From there, the intestinal loop was elevated, allowing it to enter the extraembryonic coelom. Portions of the red and black regions were identified in the abdominal coelom, whereas the remaining sections were seen in the extraembryonic coelom in samples with CRLs of 32 mm. Portions of the yellow region and the whole red region were located in the abdominal coelom of samples with CRLs of 37 mm, which indicates that intestinal return had begun in those samples. Only the purple region was observed in the extraembryonic coelom of samples with CRLs of 41 and 43 mm, which indicates that the transition phases continued. Note that the cecum was already located in the abdominal coelom in samples with CRLs of 41 and 43 mm. The entire intestinal tract had returned to the abdominal coelom in samples with CRLs exceeding 43 mm.

Figure 5.

Lateral view of the MR image and a corresponding drawing indicating the distribution of the intestine. The small intestine is highlighted from the pyloric antrum (red circle) to the cecum (yellow circle) in the following order: red, yellow, green, blue, and purple. The dotted line indicates the border between the abdominal coelom and the umbilical cavity (umbilical ring). The end of the duodenum (arrowhead) is attached to the posterior abdominal cavity wall. Next, the intestinal loop elevates (arrow), allowing it to enter the extraembryonic coelom. MR: magnetic resonance.

Discussion

In the present study, we used 3D reconstruction to observe 3D changes in the intestinal tracts of fetal samples with CRLs between 22 and 69 mm. These CRL measurements corresponded to herniation (22–32 mm), transition (37, 41, and 43 mm), and return (44–69 mm) phases. Observation of the intestine in the transition phase was limited in previous studies (Table 1). Mall (1899) described for the first time that return occurs rapidly and completely in fetuses with CRLs of approximately 40 mm. Thus far, this has been generally accepted (Mall, 1899; Frazer & Robbins, 1915; Cyr et al. 1986; Bowerman, 1993; Kim et al. 2003). Previous studies described that in cases with intestines in the return phase, the CRLs were approximately 40 mm (range: 32–44 mm) (Mall, 1899; Pernkopf, 1925; Snyder & Chaffin, 1952; Kim et al. 2003; Soffers et al. 2015) (Table 1). In our study, the extraembryonic coelomic intestinal length and the ratio of the extraembryonic coelomic intestinal length to the entire intestinal length reached their maximum values in samples with CRLs of 32 mm. Three of those samples (CRLs 37, 41, and 43 mm) were progressing into the transition phase. Our findings are consistent with those of previous studies and the current consensus. However, we observed that the cecum returned before the distal end of the small intestine (ileum). In our study, we observed that the height of the umbilical ring made stepwise increases from the transition to the return phase. We consider that the intra‐abdominal space was generated so that it could accommodate the herniated portion of the intestine, similar to the intestine wrapping into the abdominal coelom as the height of the umbilical ring increases.

Table 1.

Samples showing the intestinal loops in the transition phase

Authors (year)	CRL (mm)	Observations
Mall (1899)	32	Loop 6 (distal end of the ileum) and the cecum are still within the cord
Pernkopf (1925)	35	The distal ileum and appendix return last
	40	The intestines return to the peritoneal cavity, and the peritoneum protrudes into the umbilicus
Snyder & Chaffin (1952)	37	Cecum, ascending colon, half of the small intestine remain
Kim et al. (2003)	37 and 44	A small portion of the midgut remains
Soffers et al. (2015)	[9.0 weeks]	The domains contain the territorial loops that slide in a proximal‐distal direction back into the abdomen, with the distal ileum and appendix returning last

Transition phase: intestinal loop blocks the return from the extraembryonic coelom to the abdominal coelom.

Some studies reported that the cecum (appendix) is the last intestinal segment to return (Mall, 1899; Frazer & Robbins, 1915), whereas others reported that the end of the small intestine (ileum) and cecum were the last segments to return (Pernkopf, 1925; Snyder & Chaffin, 1952; Soffers et al. 2015). Kluth et al. (1995, 2003) observed a similar phenomenon in rat samples, specifically, the distal end of the intestine returned last, but not the cecum. It may be natural that the cecum returns earlier than the end of the ileum. The cecum is consistently located in the proximal portion of the extraembryonic coelom, whereas the small intestine is distributed more toward the distal tip of the extraembryonic coelom (Ueda et al. 2016).

Mall (1899) was the first to describe the intestinal loop being ‘sucked back’ into the abdominal coelom. Specifically, the lower portion of the body grows rapidly so the peritoneal cavity becomes too large for the organs; therefore, the intestine is ‘sucked back’ to fill the space. Frazer & Robbins (1915) described that the intestine was believed to return to the abdominal coelom, with each segment slipping back in, one after another, from the proximal segments of the intestine to the distal ones, followed by the cecum and colon (the so‐called rope model). The rope model can explain the observation that the intestine returns from proximal to distal, followed by the cecum and colon. This model implies that there is a strong traction inside the abdominal cavity or some pushing force in the extraembryonic coelom; otherwise, the intestine could not return precisely and in such a short period of time to the abdominal cavity. Frazer & Robbins (1915) assumed that the amniotic pressure on the umbilical sac may have been the driving force in addition to a ‘suck‐back’ mechanism. They considered that the narrow umbilical orifice hindered an en masse return of the intestine. Soffers et al. (2015) recently advocated hierarchical models, in which the intestinal loops slide domain by domain with the tertiary loops, which develop within the secondary loops, sliding in a proximal‐distal manner back into the abdomen. They showed that the umbilical ring remained large enough to allow the joint passage of the intestinal coils, artery, and mesentery, with the mesenteric rod becoming shorter during the same period.

Our volume measurement data support the possibility of a ‘suck‐back’ mechanism being the driving force. This would indicate that the liver is gradually displaced by a continuous increase in the volume of the abdominal coelom and liver, growth of the intestine, and changes in the intestinal fraction of the abdominal cavity. Those increases and changes in fraction continue, independent of the samples’ intestinal conditions (phases). The intestinal volume of the extraembryonic coelom may be a convenient space for the fetus in the herniated phase. The intestinal volume of the extraembryonic coelom and the percentage of the total intestinal volume in the extraembryonic coelom were a maximum of 25.8 mm³ and 33%, respectively, in samples with CRLs of 32 mm. Meanwhile, the volume of the entire intestine at the end of the transition phase (CRL 43 mm) was approximately 200 mm³. Thus, the amount of space for the intestine in the abdomen should already be large enough at the beginning of the return phase, considering the maximum intestinal volume of the extraembryonic coelom. Such a rapid increase in intestinal space in the abdominal coelom, which exceeded the intestinal volume of the extraembryonic coelom, could have generated an inward force, which supports the ‘suck‐back’ mechanism being the driving force.

In our study, the height of the umbilical ring and the Hup2/Aheight ratio increased in a stepwise manner from the transition phase to the return phase. The heights of the umbilical ring (Hup1 and Hup2) in the sample in the return phase reached the same height as the hernia tip in samples in the herniation phase. Such morphometry may also be explained by the fact that the abdominal space increased faster than the intestinal volume. We speculate that the space is generated in a manner that accommodates the herniated portion of the intestine, much like the intestine wrapping into the abdominal coelom as the height of the umbilical ring increases (Fig. 6).

Figure 6.

Models explaining how the intestinal loop returns to the abdominal coelom. (A) Classical ‘rope model’: the intestinal loops slip back into the abdominal coelom, one after another, starting with the proximal segments of the intestine, followed by the distal segments, and then the cecum and colon. (B) The intestinal loop is wrapped into the abdominal coelom as the height of the umbilical ring increases. The proximal segment of the umbilical cord may incorporate into the abdominal coelom, which would rapidly increase the abdominal wall area.

The abdominal wall should grow sufficiently before it returns, to prepare for rapid abdominal expansion (Mall, 1898). The elasticity of the abdomen may be limited in the craniocaudal and latero‐lateral directions, as the abdominal muscles such as the rectus and external and internal oblique muscles develop before the return phase (Mekonen et al. 2015; Warmbrunn et al. 2018). However, the bilateral rectus muscles are still separated in the midsagittal region, from which the umbilical cord arises. As such, the elasticity in the anteroposterior direction may be maintained. The trigger and mechanism of the rapid increase in height of the proximal portion of the umbilicus remain unknown. One possibility is that the proximal part of the umbilical cord may be incorporated into the abdominal coelom, which would rapidly increase the abdominal wall area. The MR images obtained in the present study could not provide detailed information about the abdominal wall structures. Further study using histological sections will be necessary to analyze the changes in those structures during the transition phase.

Magnetic resonance images were advantageous in that the 3D anatomy and anatomical relations could be observed and maintained. Based on the anatomical distribution of the intestinal tract in the extraembryonic and abdominal coeloms, data on the volume of the extraembryonic and abdominal coeloms and the size of the umbilicus in each sample may be reliable. These data support the Soffers hierarchical models for the most part, i.e. that the hierarchical domains (composed of secondary loops and developing tertiary loops within them) could be wrapped into the abdominal coelom. Soffers et al. (2015) determined the position of four hierarchical domains in the abdomens of samples in the return phase using information on the distribution of the SMA and mesentery to the domains using histological sections. These could not be observed via MR imaging in the present study. Further detailed discussions of the relation between our findings and their model, and further evaluation of their model are thus far difficult. Analyses of the histological sections of the samples in the transition phase samples will be the next step in determining more details of the mechanism.

Conclusion

We analyzed the morphological and morphometrical changes from herniation to transition to return phases in the umbilical herniation of the intestines. The transition phase was observed in three samples with CRL of approximately 40 mm, as previously described. The height of the umbilical ring increased in a stepwise manner between the transition and return phases. We surmised that the space was generated in a manner that could accommodate the herniated portion of the intestine, much like the intestine wrapping into the abdominal coelom as the height of the umbilical ring increases.

Conflict of interest

There are no conflicts of interest to declare.

Author contributions

Akari Nagata, Shinnosuke Hatta, Xiang Ji, Aoi Ishikawa, Rino Sakamoto: data analysis/interpretation. Shigehito Yamada, Hirohiko Imai, Tetsuya Matsuda: acquisition of data. Tetsuya Takakuwa: concept/design, drafting of the manuscript.

Supporting information

Movie S1. Movie showing the distribution of the intestine (CRL 32 mm ID 51397).

Movie S2. Movie showing the distribution of the intestine (CRL 37 mm ID 52002).

Movie S3. Movie showing the distribution of the intestine (CRL 41 mm ID 37011).

Movie S4. Movie showing the distribution of the intestine (CRL 43 mm ID F2307).