A simplified cervix model in response to induction balloon in pre-labour
© Smith; licensee BioMed Central Ltd. 2013
Received: 8 January 2013
Accepted: 23 September 2013
Published: 26 September 2013
Induction of labour is poorly understood even though it is performed in 20% of births in the United States. One method of induction, the balloon dilator applied with traction to the interior os of the cervix, engages a softening process, permitting dilation and effacement to proceed until the beginning of active labour. The purpose of this work is to develop a simple model capable of reproducing the dilation and effacement effect in the presence of a balloon.
The cervix, anchored by the uterus and the endopelvic fascia was modelled in pre-labour. The spring-loaded, double sliding-joint, double pin-joint mechanism model was developed with a Modelica-compatible system, MapleSoft MapleSim 6.1, with a stiff Rosenbrock solver and 1E-4 absolute and relative tolerances. Total simulation time for pre-labour was seven hours and simulations ended at 4.50 cm dilation diameter and 2.25 cm effacement.
Three spring configurations were tested: one pin joint, one sliding joint and combined pin-joint-sliding-joint. Feedback, based on dilation speed modulated the spring values, permitting controlled dilation. Dilation diameter speed was maintained at 0.692 cm·hr-1 over the majority of the simulation time. In the sliding-joint-only mode the maximum spring constant value was 23800 N·m-1. In pin-joint-only the maximum spring constant value was 0.41 N·m·rad-1. With a sliding-joint-pin-joint pair the maximum spring constants are 2000 N·m-1 and 0.41 N·m·rad-1, respectively.
The model, a simplified one-quarter version of the cervix, is capable of maintaining near-constant dilation rates, similar to published clinical observations for pre-labour. Lowest spring constant values are achieved when two springs are used, but nearly identical tracking of dilation speed can be achieved with only a pin joint spring. Initial and final values for effacement and dilation also match published clinical observations. These results provide a framework for development of electro-mechanical phantoms for induction training, as well as dilator testing and development.
Even though labour is induced in over 20% of women in the United States  both general labour dynamics and the more specific dynamics of the cervix are poorly understood . Even under controlled conditions with static models, practitioners have poor accuracy in identifying dilation levels and compliance .
To promote systematic exploration and analysis of cervical dynamics a new model is proposed here, expressed in explicit analytic form, which strikes a balance between the intuitive qualitative descriptions typically used by clinicians , and numerically intensive finite-element models . This lays the foundation for a framework for modelling in simulation and with electromechanical phantoms.
In this study we will concentrate on pre-labour (, p 185–186) (or the latent stage of labour) and the application of balloon dilators during this phase to induce active labour. Labour induction techniques are varied, with balloons being one of the oldest contemporary techniques . The median duration can last from about 5 hours for a woman who has previously given birth to about 8 hours for a woman giving birth for the first time (, p. 186). The transition from pre-labour to active labour is typically assumed to have occurred when the cervical diameter is anywhere from 3 to 5 cm [5, 7].
It is generally admitted that while cervical dilation is a convenient measurand for tracking progress of labour, it is insufficient. Head-to-cervix force, uterine activity, effacement rates, and dilation rates all appear to play a role in birth mode or outcome . The exact and distinct nature of these measurands is unclear, leading to a wide range of labour induction intervention methods  and on-going efforts at comparison [10–12]. One of these methods, the indirect balloon dilator, holds advantages over other methods in that it appears to mimic processes for cervical dilation related to head-to-cervix forces. These processes tend to be slower than those seen in pharmacological approaches or direct radial dilators, but are potentially safer  if care is taken to minimize risk of infection inherent with insertion of foreign objects in the endocervical canal .
Tractive force: amount and duration
The amount of explicit external tractive force varies, from none , to "minimal" , to about 0.5 kg (approx. 5 N)  or more , but is often not directly specified . The tractive force most often appears to be achieved by taping the end of the lumen to the patient’s thigh, leaving few practical methods for systematic measuring of force measurement. Typically, one waits many hours  for the balloon to be expelled naturally: from half-an-hour  to ten hours , with suggestions of six hours being an acceptable upper limit before attempting other forms of intervention . Expulsion of the balloon occurs when dilation reaches approximately four centimetres. This is comparable to the non-induced pre-labour phase of labour .
In this study we simulate a 0.5 kg balloon-applied tractive force, directed to the inner os of the cervix, leading to the indirect dilation of the cervix through a softening of the tissue. Final dilation diameter (4.5 cm) and effacement (2.25 cm) and time scale (about 7 hours) are similar to what is seen in practice. The details of the model are presented below.
Laplace’s law, for vessels under pressure, has been suggested as a modelling framework [13, 27]. In the same vein, more numerically intensive finite-element models have recently been produced . Both of these approaches, however, did not examine the use of balloon dilators during pre-labour.
Cervical anatomy and dynamics are highly variable, with some correlation to gestational age and parity . The proposed model’s parameters can easily be adapted to other configurations and dynamic characteristics given the combination of schematic and analytic expressions. In the context of this manuscript we make the following assumptions. The subject is assumed to be at full term of her pregnancy, that is, 39 to 40 weeks. Rovas et al. showed that cervical width is relatively constant from weeks 31 through 41 and that is similar in women who have given birth before and those who have not. Cervical width is therefore assumed to be 4.5 cm . As less than 40% of women in Rovas et al.’s study had an open cervix at 39–40 weeks, we will assume that the endocervical canal is initially closed and no funnelling is present. Therefore, cervical wall thickness is 2.25 cm. Here, we assume that cervical length, at 39 to 40 weeks, is approximately half-way between that for women who have previously given birth (1.2 cm) and women giving birth for the first time (3 cm) . For convenience, we will assume that length equals wall thickness: 2.25 cm.
Cervical mass had to be estimated since no published data on cervical mass at term were found after a literature search and consultations with a number of health professionals. To estimate cervical mass, we assume that mass distribution is uniform throughout the uterus, including the cervix. The uterus, without its contents, is assumed to have a mass of 1 kg and to be a 30 cm by 23 cm by 20 cm ellipsoid . The uterine wall is assumed to be 0.36 cm thick . Assuming a cervix in the shape of a cylinder 4.5 cm wide by 2.25 cm high, the at term cervix will have a mass of 0.027 kg.
While the effect of cervix mass on dynamics with a timescale measured in hours is small, it is important to note that this model is also meant for application to electro-mechanical models in which the timescale can be greatly accelerated. For instance, in a current prototype, full dilation can be achieved in less than a minute. For training and device testing purposes it is important that the model be applicable at both short and long timescales, hence the inclusion of the cervix mass.
Given the restriction of the model to pre-labour, it is assumed that dilation will not exceed Rovas et al.’s 4.5 cm cervix width, at which point the balloon will exit. Therefore, the cervix is assumed to be anchored to the lower uterus and the endopelvic fascia in such a manner that the anchor point does not move. This permits the model to exclude the dynamic effects associated with movement of the anchor. This assumption stops being valid as the "active" phase of labour begins. Neither friction nor surface deformations are modelled in this manuscript, leaving such details to future work.
Simplification of the cervix through symmetry: modelling one quarter of the cervix
The "quarter model" presented here the contains the minimum number of degrees of freedom which permit dilation in a direction perpendicular to the tractive loading typical of Foley-type balloon dilators. As can be seen in the bottom portion of Figure 2, one pin joint and one sliding joint contain spring-damper pairs. One controls pivoting at the point where the cervix is anchored to the remainder of the uterus, as well as to the endopelvic fascia. The other, across a sliding-joint, approximates the compliance that spans laterally, across the endocervical canal.
The mathematical model
Assuming symmetric geometry and dynamics, the model can be simplified to a single, stretchable arm. The model is composed of two pin joints and two sliding joints. This is illustrated schematically in Figure 2.
where sinit is the initial radius of the opening of the cervix, P2. We assume that sinit is zero in the simulations which follow. The length, l, is the initial thickness of the cervical wall.
where l is the length of the link between the two pin joints, with a mass located in the geometric centre. The moment of inertia about the z-axis is calculated as a thin cylinder, . The mass, m, is one quarter of the total cervix mass.
where the first row describes the rectilinear force that develops across the endocervical canal along P2, due to the stretching of the spring-damper, K2 and B2. The traction force in the second row is assumed to be wholly due to the constant pulling action by a 0.5 kg mass along the gravity vector, supported by four equal sections of the cervix. Therefore, FTraction = 0.25·0.5·9.81 = 1.23 N. The third row describes the moment about the first pin joint, anchored by the uterus and endopelvic fascia, consisting of an angle-dependent mass term and the spring-damper stretching moment about R1 due to K1 and B1.
Simulation parameters for the quarter model
Cervical wall thickness
Initial endocervical canal opening
Moment of inertia
Pin joint damping
Sliding joint damping
Results & analysis
Feedback control has been implemented to control dilation rate throughout the pre-labour phase. Using the model outlined above three cases have been examined: feedback control through only a sliding joint spring, feedback control through only a pin joint spring and, finally, feedback control through both pin joint and sliding joint springs.
Effect of varying compliance values on cervical dilation and effacement
The model assumes that the two main spring constants will increase in compliance (decrease in stiffness) over time, in response to the presence of the balloon. As the compliance increases the traction applied to the balloon acts to deform the cervix, creating both dilation and effacement effects. The balloon descends during the combination of effacement and dilation, continuing to apply the traction force until the end pre-labour when effacement and dilation are complete. A continuum of possible compliance trajectories over time is possible. In the next section simulations are conducted on three scenarios, permitting dilation and effacement to be controlled through the changing of the two spring constants.
Time-varying compliance trials
Published cervicograms typically show a constant dilation rate during pre-labour . Sometimes the dilation rate is specified numerically (e.g. Peisner and Rosen specify between 1.2 and 1.5 cm·hr-1. ). Here, we examine a pre-labour duration of 7 hours (6.5 hours plus a half-an-hour to allow for ramping up of applied traction force and desired dilation to steady-state), resulting in a dilation diameter rate of 0.692 cm·hr-1. To avoid large transients in the response of the cervix model two ramping functions, shown in Figure 3, were introduced. The gradual application of the balloon’s traction force is mimicked through a smooth ramping function. The response of the cervix, in the form of the desired dilation velocity also smoothly transitions to the steady state value of 0.692 cm·hr-1.
Closed-loop feedback control is used to ensure that dilation rate is near-constant for the majority of the induction. Three scenarios were examined: (1) a controlled pin-joint spring with no sliding spring, (2) a controlled sliding spring with no pin-joint spring, and (3) both controlled pin-joint and controlled sliding springs. Simulations were conducted with MapleSim 6.1’s numeric solver set to Rosenbrock (stiff) with 1E-4 absolute and relative tolerances.
At the beginning of the labour process the torque produced by the traction force on pin joint R1 is very large because it is completely perpendicular to the moment arm. Either spring in this model can counter the traction force while maintaining a low error in the desired dilation speed. It takes a much smaller feedback gain on the pin joint’s spring to produce the necessary counter-torque necessary to have the cervix hold the balloon in place. A larger effort, in the form of a much larger feedback gain is required by the spring in the sliding joint to produce the same effect. This corresponds to the highest spring constant values, K1 and K2, shown in Figures 6b and 6c. In addition, when comparing the three scenarios, the lowest spring constant values are achieved when two springs are used, but nearly identical tracking of dilation speed can be achieved with only a pin joint spring. Worst performance is found when only the sliding joint spring is used.
Comparison with clinical data
There is little quantifiable clinical data that describes the dynamics of the interaction between balloon dilator and cervix. What quantified data is available has been used to develop this model. The amount of explicit external tractive force in the model has been set to 0.5 kg (approx. 5 N)  the most commonly cited quantified value. The typical time taken for a balloon to be expelled varies from half-an-hour  to ten hours , while the typically cited range for pre-labour without intervention is 5 to 8 hours. Therefore, 6.5 hours, a value in the mid portion of both ranges, was chosen for simulation.
Because cervicograms typically show dilation progressing at a constant rate it was important that the simulations result in constant dilation rates, as well. While Peisner and Rosen found rates of between 1.2 and 1.5 cm·hr-1 the rate used in this simulation (in order to ensure constant dilation throughout the duration of pre-labour) had to be set to 0.692 cm·hr-1. Had the Peisner and Rosen values been used with this model and had it been assumed that the dilation rate was constant throughout pre-labour, the simulated pre-labour would have lasted under four hours.
When comparing the effacement curves in Figures 4, 5 and 6 to the dilation curves one can see that the effacement effect is more noticeable earlier than the dilation effect. This corresponds to clinical observations that "cervical effacement precedes significant dilation" .
The simulation ends when the cervix model dilates to 4.5 cm, which is within the acceptable range for transition from pre-labour to active labour (e.g. 3 cm to 5 cm). The 100% effacement value used in simulation agrees with the upper bounds given in .
While this model permits any arbitrary trajectory of compliance values versus time to be programmed, it does not help to answer why the cervix responds to the presence of the balloon. It is the author’s perspective that answering this question is key to understanding when intervention should or should not be performed. This will be the objective of future work.
At the moment, there is no other mathematical model which examines the descent of an induction balloon as the cervix softens during pre-labour. The closest models are those by Gee [13, 27], using Laplace’s law for a spherical pressure vessel, and finite element analysis approach House et al.. Both of these examine the reaction of the cervix to pressure exerted by the uterus, and neither is in the context of induction of labour, nor during pre-labour, nor with respect to balloon dilation.
Unlike the models proposed by Gee or House, the structure of this model lends itself well to the inclusion of variable compliance mechatronic devices  in the development of cervix phantoms. These phantoms could potentially be used to explore the dynamics of the cervix during pre-labour, with or without intervention.
Future work towards a physical training simulator
As discussed earlier, work is proceeding on physical prototypes capable of controlled dilation and effacement. These results provide a framework for further development of electro-mechanical phantoms for physical training or dilator testing. It is envisioned that a three-dimensional physical prototype can be devised using four quarter models wrapped in a compliant covering. Small motors at the pin and sliding joints can be programmed to mimic springs and dampers through a proportional-derivative algorithm. This physical model will also be useful in the development of contact models, including characterization of friction between the balloon and the cervix.
A model has been proposed and evaluated to represent cervical dilation in response to the presence of a balloon dilator during the pre-labour stage of labour. Simulations have been run to mimic a 6.5 hour pre-labour phase (with an additional half-an-hour for initialization) in response to a small, fixed external balloon-applied traction force. The model contains feedback pathways which permit the dilation rate to be controlled in response to the presence of constant traction applied to the interior os via the balloon. Dilation diameter of 4.5 cm and effacement of 2.25 cm were achieved with a 0.692 cm·hr-1 dilation rate. It was shown that compliance could be controlled at either the pin joint located at the junction between the uterus, the endopelvic fascia and the cervix, or along the sliding joint that spans laterally across the endocervical canal. When comparing the three scenarios, the lowest spring constant values are achieved when two springs are used, but nearly identical tracking of dilation speed can be achieved with only a pin joint spring. Worst performance is found when only the sliding joint spring is used. These results can be applied to the development of new electromechanical cervix phantoms for the study of cervical dynamics during pre-labour.
a Note that the · operator is an explicit multiplication and not the dot product.
b Note that while the dilation rate is given for cervix diameter, in order to be consistent with existing clinical data, the radius dilation rate was controlled in the model. The radius dilation rate value is half the diameter dilation rate.
James Andrew Smith is an Assistant Professor in Electrical and Computer Engineering at Ryerson University. He is also the stream coordinator for Ryerson’s newly accredited biomedical engineering program. His PhD is in Mechanical Engineering from McGill University and he undertook a postdoctoral fellowship in Sports Science in Jena, Germany. His research interest is in the development of electromechanical surrogates / phantoms for research in biomechanics and obstetrics.
Many thanks for Ms. Bonnie Yue at MapleSoft for technical support, and Dr. Mary Sharpe, Dr. Deborah Robertson, Dr. Eric Sokol, Dr. Peter Soothill, Dr. Phyllis Leppert, Dr. Derek Lobb and Dr. Yunru Li for their input. Thank you to Ms. Kathryn Atwell, Ms. Anna Leshchenko, Ms Marcelle-Andrée Carneiro and Mr. Abdallah El-Falou for the discussions regarding electromechanical models of the cervix. Finally, I would like to thank the Editor-in-Chief, Dr. Paul S. Agutter, and the anonymous reviewer for their patience and constructive feedback, which helped improve this manuscript.
- Swamy GK: Current methods of labor induction. Semin Perinatol. 2012, 36 (5): 348-352. 10.1053/j.semperi.2012.04.018.View ArticlePubMedGoogle Scholar
- Buhimschi CS, Buhimschi IA, Malinow AM, Saade GR, Garfield RE, Weiner CP: The forces of labour. Fetal Matern Med Rev. 2003, 14 (4): 273-307. 10.1017/S0965539503001128.View ArticleGoogle Scholar
- Huhn KA, Brost BC: Accuracy of simulated cervical dilation and effacement measurements among practitioners. Am J Obstet Gynecol. 2004, 191: 1797-1799. 10.1016/j.ajog.2004.07.062.View ArticlePubMedGoogle Scholar
- House M, Feltovich H, Hall TJ, Stack T, Patels A, Socrate S: Three-dimensional, extended field-of-view ultrasound method for estimating large strain mechanical properties of the cervix during pregnancy. Ultrason Imaging. 2012, 34: 1-14. 10.1177/016173461203400101.PubMed CentralView ArticlePubMedGoogle Scholar
- The Cervix, second edition. Edited by: Jordan JA, Singer A. 2006, Malden: Blackwell Publishing LtdGoogle Scholar
- Crane J, Leduc L, Farine D, Hodges S, Reid GJ, Aerde JV: Induction of Labour at Term. SOGC Clinical Practice Guideline 107. 2001, Ottawa: SOGCGoogle Scholar
- Peisner D, Rosen M: Transition from latent to active labor. Obstet Gynecol. 1986, 68 (4): 448-451.PubMedGoogle Scholar
- Allman AC, Genevier ES, Johnson MR, Steer PJ: Head-to-cervix force: an important physiological variable in labour. 2. Peak active force, peak active pressure and mode of delivery. Br J Obstet Gynaecol. 1996, 103 (8): 769-775. 10.1111/j.1471-0528.1996.tb09871.x.View ArticlePubMedGoogle Scholar
- Moleti CA: Trends and controversies in labor induction. MCN, Am J Matern/Child Nurs. 2009, 34: 40-47. 10.1097/01.NMC.0000343864.49366.66.View ArticleGoogle Scholar
- Jabbar T, Faisal S, Imran F, Kauser R: Induction fo labor comparison of cervical Foley’s catheter and prostaglandin E-2 at Term. Prof Med J. 2011, 18 (2): 201-207.Google Scholar
- Atad J, Hallak M, Auslender R, Porat-Packer T, Zarfati D, Abramovici H: A randomized comparison of prostaglandin E2, oxytocin, and the double-balloon device in inducing labor. Obstet Gynecol. 1996, 87 (2): 223-227. 10.1016/0029-7844(95)00389-4.View ArticlePubMedGoogle Scholar
- Hofmeyr GJ: Dewhurst’s Textbook of Obstetrics & Gynaecology, Seventh edition. 2008, Oxford: Blackwell Publishing Ltd, chap. 23 induction and augmentation of labourGoogle Scholar
- Gee H, Taylor EW, Hancox R: A model for the generation of intra-uterine pressure in the human parturient uterus which demonstrates the critical role of the cervix. J Theor Biol. 1988, 133 (3): 281-291. 10.1016/S0022-5193(88)80322-9.View ArticlePubMedGoogle Scholar
- Mattei A: Essai sur l’accouchement physiologique. 1855, Paris, France: Victor MassonGoogle Scholar
- Storer HR: Thirtieth regular meeting march 15 1870. J Gynaecol Soc Boston. 1870, 3 (3): 153-165. [Reported by Horatio R. Storer, Secretary]Google Scholar
- Aveling JH: On dilatation of the Cervix Uteri by Bougie. Brit Med J. 1872, 2 (622): 577-578.PubMed CentralView ArticlePubMedGoogle Scholar
- Cromi A, Ghezzi F, Agosti M, Serati M, Uccella S, Arlant V, Bolis P: Is transcervical Foley catheter actually slower than prostaglandins in ripening the cervix? A randomized study. Am J Obstet Gynecol. 2011, 204 (4): 338.e1-338.e7. 10.1016/j.ajog.2010.11.029.View ArticleGoogle Scholar
- Guinn DA, Davies JK, Jones RO, Sullivan L, Wolf D: Labor induction in women with an unfavorable Bishop score: randomized controlled trial of intrauterine Foley catheter with concurrent oxytocin infusion versus Foley catheter with extra-amniotic saline infusion with concurrent oxytocin infusion. Am J Obstet Gynecol. 2004, 191: 225-229. 10.1016/j.ajog.2003.12.039.View ArticlePubMedGoogle Scholar
- Jagani N, Schulman H, Fleischer A, Mitchell J, Randolph G: Role of the cervix in the induction of labor. Obstet Gynecol. 1982, 59: 21-26.PubMedGoogle Scholar
- Nyengidiki TK: The incompetent cervix. Online. 2008, [http://www.scribd.com/doc/6587108/The-Incompetent-Cervix-2]Google Scholar
- Boulvain M, Kelly A, Lohse C, Stan C, Irion O: Mechanical methods for induction of labour (review). Cochrane Database Syst Rev. 2001, 4: 1-293.Google Scholar
- Phillips SJ: The induction of premature labour: its scope and present results. Lancet. 1920, 196 (5067): 741-743. 10.1016/S0140-6736(00)54948-4. [http://www.sciencedirect.com/science/article/pii/S0140673600549484]View ArticleGoogle Scholar
- Veyre JF, Laumosne J, Mavel A, Feldman JP, Pelikan P, Michiels Y: Le ballon de Boissard: antiquité ou moyen moderne de déclenchement. Revue Francaise de Gynecologie et d’obstetrique. 1974, 69 (10): 535-539.Google Scholar
- Mei-Dan E, Walfisch A, Suarez-Easton S, Hallak M: Comparison of two mechanical devices for cervical ripening: a prospective quasi-randomized trial. J Matern Fetal Neonatal Med. 2012, 25 (6): 723-727. 10.3109/14767058.2011.591459.View ArticlePubMedGoogle Scholar
- Tenore J: Methods for cervical ripening and induction of labor. Am Family Physician. 2003, 67 (10): 2123-2128.Google Scholar
- Friedman EA: Labor: Clinical Evaluation and Management, second edition. 1978, New York: Appleton-Century-CroftsGoogle Scholar
- Gee H: The Cervix, Second edition. 2006, Malden: Blackwell Publishing Ltd, chap. 13: Mechanics, biochemistry and pharmacology of the cervix and labourGoogle Scholar
- Rovas L, Sladkevicius P, Strobel E, Valentin L: Reference data representative of normal findings at two-dimensional and three-dimensional gray-scale ultrasound examination of the cervix from 17 to 41 weeks’ gestation. Ultrasound Obstet Gynecol. 2006, 27: 392-402. 10.1002/uog.2658.View ArticlePubMedGoogle Scholar
- Young A: Juta’s Manual of Nursing: Health Care Priorities, Volume 3. 2005, South Africa: Juta and CompanyGoogle Scholar
- Boutin A, Jastrow N, Girard M, Roberge S, Chaillet N, Brassard N, Bujold E: Reliability of two-dimensional transvaginal sonographic measurement of lower uterine segment thickness using video sequences. Am J Perinatol. 2012, 29 (7): 527-32.PubMedGoogle Scholar
- Danforth’s Obstetrics and Gynecology. Edited by: Scott JR, Gibbs RS, Karlan BY, Haney AF. 2003, Philidelphia: Lippincott Williams & WilkinsGoogle Scholar
- Rouse DJ, Weiner SJ, Bloom SL, Varner MW, Spong CY, Ramin SM, Caritis SN, Grobman WA, Sorokin Y, Sciscione A, Carpenter MW, Mercer BM, Thorp Jr JM, Malone FD, Harper M, Iams JD, Anderson GD: Failed labor induction: toward an objective diagnosis. Obstet Gynecol. 2011, 117 (2 Pt 1): 267-272.PubMed CentralView ArticlePubMedGoogle Scholar
- Vanderborght B, Verrelst B, Van Ham R, Van Damme M, Lefeber D, Duran BMY, Beyl P: Exploiting natural dynamics to reduce energy consumption by controlling the compliance of soft actuators. Int J Robot Res. 2006, 25 (4): 343-358. 10.1177/0278364906064566.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.