Neurourology and Urodynamics
An Automated Hand-Held Elastometer for Quantifying the Passive Stiffness of the Levator Ani Muscle in Women Jennifer A. Kruger,1* Poul M.F. Nielsen,2 Stephanie C. Budgett,3 and Andrew J. Taberner2
2
1 Auckland Bioengineering Institute, The University of Auckland, Auckland, New Zealand Department of Engineering Science, Auckland Bioengineering Institute, The University of Auckland, Auckland, New Zealand 3 Department of Statistics, The University of Auckland, Auckland, New Zealand
Aim: Design and develop an automated, hand-held instrument (elastometer) to assess in vivo passive stiffness of the pelvic floor muscle. Materials and Methods: The elastometer system consisted of a hand piece, real-time controller, and laptop computer. A cable connected the hand-piece to the controller, which communicated with a laptop computer via an ethernet connection. Force sensitivity calibration and displacement accuracy were determined experimentally using a spring load and an Instron mechanical tester. A test re-test series quantified the in vivo repeatability (within a procedure) and reproducibility (between procedures after a 5 min delay) of passive stiffness in volunteers (n ¼ 20). Stiffness was determined from the gradient of the force–displacement curve for each cycle. Results: The force-aperture spring measurements from the elastometer showed consistent (r2 ¼ 1.0000) agreement with those measured by the Instron. The difference between spring stiffness as measured by the elastometer and the Instron (388.1 N/m cf. 388.5 N/m, respectively) was negligible. The intra-class correlation coefficient for repeatability within procedures was 0.986 95% CI (0.964–0.994) n ¼ 20, and reproducibility between procedures ICC 0.934 (95% CI 0.779–0.981) n ¼ 12. Bland–Altman analysis determined a bias of 0.3 and 18.5 N/m, for repeatability and reproducibility respectively. Neither bias is likely to be clinically significance. Conclusion: The elastometer demonstrated very good repeatability and accuracy in the measurement of force/displacement during in vitro testing. There was a high degree of repeatability and reproducibility in stiffness measurements in a test re-test series. Our results demonstrate the elastometer is accurate and reliable and # 2013 Wiley Periodicals, Inc. thereby suitable for larger clinical trials. Neurourol. Urodynam. Key words: avulsion; childbirth; levator ani muscles; passive stiffness; pelvic floor
INTRODUCTION
Childbirth-related injury to the muscles of the pelvic floor has been shown to affect 10–30% of women delivering their first child vaginally.1,2 The limited capacity of the pelvic floor muscles to stretch during vaginal delivery is likely to be related to this trauma. At present there is a paucity of data regarding the mechanical properties of the pelvic floor. We aim to redress this deficiency through the development of a novel ‘‘elastometer.’’ The muscles of the pelvic floor, otherwise known as the levator ani (LA) muscles, are intimately involved in the birth process and support the organs of the pelvis. Imaging modalities such as ultrasound and magnetic resonance imaging have confirmed that injury to the LA muscle occurs during vaginal birth and this injury (avulsion) is substantially implicated in the development of pelvic organ prolapse (POP).3 Epidemiological studies have also shown a significant association between parity and the risk of POP.4 Pelvic floor muscle dysfunction has been termed the ‘‘hidden epidemic’’ with 11–20% of women requiring surgery for POP within their lifetime.5,6 Developing measures to identify a priori, those who are most likely to suffer from injury during vaginal birth is of significant clinical importance. Previous instrumentation development for clinical use has focused on measuring active force production of the LA muscles.7–9 Current methods used to assess force production include measuring vaginal pressure10 (perineometry), the use of dynamometric speculum(s),8,9 and multidirectional probes with imbedded pressure sensors.11,12 All these devices are designed to measure active force during voluntary contraction of the LA muscle. This measurement has proved particularly #
2013 Wiley Periodicals, Inc.
useful to assess the effect of pelvic floor muscle physiotherapy in women with urinary stress incontinence, vulvodynia, and POP.13–15 To date, little attention has been directed at the passive/elastic properties of the muscle. Nonetheless, the elastic component of striated muscle is an important determining factor for the ability of any muscle to stretch, as during parturition muscle-strain can be as large as 300%.16 We had previously developed a novel elastometer designed to measure this elastic component.17 Despite the utility of this device, its modus operandi and appearance made it unsuitable for use in a large clinical study. In a later version of this instrument, we added rudimentary motorised control and data logging capabilities.18 Although this device was useful for brief tests, the control system and motor in this design were not robust enough to use over the long periods which are required in a clinical setting. In this paper, we detail the design and performance of a new automated, robust and precise elastometer for assessing the passive force and stiffness of the LA muscle. We describe its design, and measure its performance in a test re-test series, in vivo, to determine its acceptability for use in large clinical studies. Heinz Koelbl led the peer-review process as the Associate Editor responsible for the paper. Conflict of interest: none. Grant sponsor: Aotearoa Postdoctoral Research Fellowship Correspondence to: Jennifer A. Kruger, Ph.D., Auckland Bioengineering Institute, The University of Auckland, Auckland, New Zealand. E-mail:
[email protected] Received 3 July 2013; Accepted 25 September 2013 Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/nau.22537
2
Kruger et al. acquired. Data were filtered and acquired at a rate of 100 Hz, and presented as a force and displacement time-chart, and as a force–displacement plot. The laptop was operated from battery power during patient experiments. The speculum ends were wider at the tip (26 mm) than at the neck (18 mm) to reduce the effect of perineal muscles confounding the measurement of LA passive stiffness. The length of the speculum ensured that the speculum face was approximately adjacent to the LA muscle (typically located approximately 35 mm from the entrance of the vagina), in the coronal orientation. The speculum tips were magnetically attached to the arms, allowing easy removal for cleaning. Calibration
Fig. 1. Elastometer system (A) hand-piece, (B) controller, and (C) laptop with user-interface.
Fig. 2. Internal components of elastometer hand-piece. Aluminium covers not shown.
MATERIALS AND METHODS Elastometer
The elastometer system (Fig. 1) consisted of a hand-piece, realtime controller, and laptop computer. The hand-piece (Fig. 2) comprised two anodised aluminium arms (with detachable acetyl plastic speculum ends) actuated by a displacementcontrolled linear DC servo motor (Physik Instrumente M-227.25, with 0.1 mm resolution optical encoder) via symmetrical struts and a bi-directional force transducer (Omega LCM201-100N). The hand-piece communicated with a laptop computer via a wired ethernet connection. The controller was a compact reconfigurable data acquisition and control system (cRIO-9075, National Instruments, Austin, TX). A user interface (LabVIEW 2011) allowed the user to manually move the speculum, enter experimental protocol parameters and patient data, initiate measurement protocols, and view force and displacement measurements as they were
The non-linear relationship (Fig. 3A) between motor extension and speculum aperture is prescribed by the geometric arrangement of the arms and struts. The sensitivity of the force measurement (Fig. 3B) was determined at selected apertures by applying standard forces (0–10 N) between the speculum ends and measuring the output voltage of the force transducer bridge. The slope of the best-linear-fit of sensor voltage as a function of applied force quantified the sensitivity at a given aperture. This process was repeated at apertures ranging from 28 to 60 mm, in 5 mm increments. The resulting sensitivity function was fitted with a second-order polynomial function and interpolated in the control software (Figs. 3A and 3B). Force–Displacement Measurement Protocol
The development of this elastometer system was motivated by the desire to determine the dynamic passive stiffness of the LA muscle as a function of aperture in a cohort of up to 200 patients. The purpose of this study was to quantify the quasistatic relationship between force and displacement, while also recording (for possible later analysis) viscoelastic and stressrelaxation effects. Accordingly, the protocol for conducting this measurement consisted of a pre-programmed, highly repeatable sequence of force–displacement step measurements. The user inputs for generating the experimental sequence were the maximum aperture A, the number of evenly spaced measurement apertures at which to conduct force measurements n, the relaxation time at each aperture tr, and the opening rate v. When the experiment was initiated, the speculum opened at rate v to the first aperture set-point. Upon the speculum settling within 0.2 mm of the aperture setpoint, and after the relaxation time tr (typically 3 sec), force data were acquired over a period of 1 sec, and the average force was computed. The speculum tips were then actuated to the second measurement aperture, and the force measurement was
Fig. 3. A: Aperture as a function of motor extension. B: Force sensitivity as a function of aperture.
Neurourology and Urodynamics DOI 10.1002/nau
An Automated Elastometer to Quantify Passive Stiffness of the LA Muscle
3
Fig. 4. Front panel of user interface after the measurement of a nominal spring load.
repeated. After the final force measurement was conducted at the selected aperture limit A, the speculum was closed at a rate of 55 mm/sec (Fig. 4). The measured force and displacement data sets were recorded during the entire experiment (at a sample rate of 100 Hz) in a comma-separated-value text file on the laptop. Averaged force and displacement measurements (corresponding to each of the measurement apertures) were recorded in a second text file. Characterisation and Verification Experiments
To verify the force calibration of the elastometer, we constructed a spring-based apparatus to provide a repeatable stiffness load to the speculum. The apparatus stiffness was determined using an electromechanical testing device (Instron 5800). Subsequently, the apparatus was attached to the speculum ends, and the experiment repeated using the elastometer. The difference between the two measurements was computed at each measurement aperture, and expressed as a percentage of the mean force at that aperture. The linearity of the elastometer force measurement was quantified by fitting a straight line to the force-aperture measurements, and computing the residual from the best-fit line. The non-linearity was quantified by the ratio of the maximum deviation from linearity to the maximum force. The repeatability of the elastometer displacement-control and force-measurement was quantified by repeatedly (n ¼ 16) opening the speculum under a spring-load to an aperture of 40 mm. The aperture was recorded from the integrated displacement encoder, and compared to manual measurements from digital calipers (Mitutoyo). The repeatability was quantified by the standard deviation of aperture and force estimates (measured during a period of 1, 3 sec after the elastometer settled to within 0.1 of 40 mm). Clinical Experiments
We performed a series of experiments on volunteers in order to quantify the repeatability (repeated measures of stiffness, without removing the elastometer) and reproducibility (repeated measures of stiffness after removing and reinserting the elastometer) of in vivo measurements of pelvic floor stiffness. Neurourology and Urodynamics DOI 10.1002/nau
Participants for this study were recruited from a cohort of a larger study designed to determine the stiffness characteristics of the LA muscle pre- and post-vaginal birth. Ethical approval was granted by the Lower South Regional Ethics committee, and the Counties Manukau District Health Board (LRS/10/ 7029). Twelve women consented to participate in tests of the measurement reproducibility (where the elastometer was removed and then re-inserted 5 min later), and twenty consented to tests of repeatability (where the elastometer was held in situ for one procedure). Participants were examined supine, with knees comfortably flexed after voiding. Although the results are not reported in this paper, all participants are examined digitally prior to testing for elastometry as part of a larger study. The speculum ends of the elastometer were cleaned and soaked in a virucidal agent prior to use and each speculum bill was then covered with a condom and lubricated with hypoallergenic gel. The elastometer was introduced in the coronal orientation, following the natural axis of the vagina to the level of the LA muscle. The length of the speculum ensures a high likelihood that the displacement occurs where the LA muscle underlies the lateral walls of the vagina. This would correspond to the active component of the LA complex, often referred to as the pubovisceralis muscle, typically where subjective assessment of muscle strength and tone is gauged during palpation. Additionally, as a vaginal examination has already been performed, there was subjective verification that the elastometer felt in the same place as where the muscle has been palpated. Once the participant was comfortable, a force–displacement measurement was initiated with tr ¼ 3 sec, v ¼ 20 mm/sec and n ¼ 10. From our previous work17 and that of others19 we set the maximum aperture A to 50 mm. This measurement was repeated three times for each participant. The first of the three measurement cycles was to allow for tissue pre-conditioning, and to familiarise the patient with the measurement procedure. The two subsequent cycles were then conducted within 1 min. This procedure of three cycles was conducted once for 20 participants, and repeated on the same participant (lying in the same position) twice for 12 participants. When there were two
4
Kruger et al.
procedures, these were separated by a period of 5 min in order to allow the perturbed tissues to return to their resting states. The passive stiffness of the LA muscle was calculated from the slope of the force–displacement plots using the six data points that span apertures of 40–50 mm, for cycles two and three. Statistical Analysis
Intraclass correlation coefficients, (mixed model, absolute agreement), were used to determine the repeatability of the cycle 2 and 3 within, and between procedures (IBM SPSS V21; Armonk, NY: IBM Corp.). A Bland–Altman analysis was used to determine bias and the limits of agreement for corresponding cycles between procedures. This approach quantifies the level of agreement between subsequent measures of stiffness and the limits of that agreement, to determine the variability between the two procedures. RESULTS
The force-aperture spring measurements from the elastometer showed consistent agreement (Fig. 5) with those measured using the Instron electromechanical measurement device (r2 ¼ 1.0000). The maximum difference between the two force measurements at each aperture was 17 mN or 0.3% (at 32 mm aperture). The maximum non-linearity of the elastometer over this range was 0.15%. The stiffness of the spring as measured by the elastometer was 388.1 N/m, while the stiffness as measured by the Instron was 388.5 N/m. The instrument repeatability experiment resulted in the control system opening the speculum (under the load of a spring) to an aperture of 40.02 0.07 mm (mean SD). The aperture measured by the digital calipers was 39.82 0.15 mm. The mean and standard deviation of the corresponding force measurements was 8.241 0.011 N. An example of a test–retest repeatability procedure of six force–displacement measurements from a single patient is shown in Figure 6. ICC for repeatability between cycles within a procedure, and reproducibility between procedures showed excellent correlation; 0.986, 95% CI 0.964–0.994, and 0.934 95% CI 0.779–0.981, respectively. The Bland–Altman analysis comparing stiffness measurements from cycles 2 and 3 within one procedure showed minimal bias, with the mean difference at 0.3 N/m and the limits of agreement being narrow at 59.3 ! 60 N/m (Fig. 7). When comparing stiffness measurements from corresponding cycles between two procedures, the Bland–Altman limits of agreement were wider, ranging from 85 to 110 N/m, with a mean difference of 18.5 N/m.
Fig. 5. Instron and elastometer measurements of spring force-extension.
Neurourology and Urodynamics DOI 10.1002/nau
Fig. 6. Representative plot showing force–displacement cycles 2 and 3 for two procedures. Linear trend lines are fitted to measurements spanning apertures from 40 to 50 mm.
DISCUSSION
We have developed a reliable, portable pelvic-floor elastometer with the capability of executing a programmed measurement protocol and automated data collection. The device has proved repeatable for measurement of force with minimal divergence between that measured using the Instron machine and the device itself. The repeatability of measurements in vivo has also demonstrated negligible difference in gradient of the curve (stiffness) between cycles 2 and 3. Not surprisingly the reproducibility of the measurements between procedures was more variable than the repeatability of measurements during procedure one only. This is consistent with a previous test retest series of reproducibility using an earlier version of this device.18 Stiffness was calculated from the slope of the curve between 40 and 50 mm displacement, as we hypothesised that at this displacement the speculum is in close contact with the muscle underlying the soft vaginal tissue. Although it is well recognised that soft tissue is anisotropic and nonlinear, the fit of the line to the curve from this point is consistently more linear than at lower displacements, implying a relatively constant stiffness at these higher
Fig. 7. Bland–Altman plot showing the differences in stiffness against mean stiffness between cycles two and three for one procedure (n ¼ 20).
An Automated Elastometer to Quantify Passive Stiffness of the LA Muscle apertures. However, this may not be true for women who are not pregnant, as the vaginal tissue is less vascular and the passive resistance to displacement may be apparent at smaller displacements of the speculum tips. The difference in passive force between cycles 2 and 3 for one procedure (repeatability) is unlikely to be clinically significant, suggesting that determination of stiffness can be taken from either the second or third cycle of the procedure. The reproducibility of stiffness in this small sample was more uncertain, although the significance of a 100 N/m difference is unknown at this point. Research on the passive properties of the LA muscle in women is scarce; only a few studies have measured both the passive and active forces, in varying population groups. Verelst and Leivseth19 used an instrumented speculum to determine passive and active force development in the coronal plane in 20 parous women. In a test re-test series, these authors found a displacement of 40 mm provided the most reproducible measurements for active force development. The limits of agreement at this diameter were 3 N during a pelvic floor muscle contraction, which are comparable to our findings. Morin et al.20 described a new methodology for evaluating the passive properties of the pelvic floor muscle, using a modified intra-vaginal dynamometer. Measurements for this study were conducted in the anterior–posterior orientation of the vaginal opening, which precludes direct comparisons to our results. Nonetheless, these authors found a range of passive resistance from 1.03 1.12 N at minimal aperture of 15 mm of the device, to 6.18 2.84 N at maximum aperture of 30 mm. The mean passive resistance in our population of pregnant women at the same aperture (30 mm) was 1.72 0.86 N. Hormonal influences of pregnancy and orientating the elastometer in a coronal plane, are likely explanations for the differences in observed passive stiffness between these two population groups. It has been recommended by several researchers in the field,6,21,22 that defining causative mechanisms of pelvic floor muscle damage during vaginal childbirth is essential to identify those who are most at risk of sustaining injury. To the authors’ knowledge, this is the first time an instrument has been developed specifically in response to this need. Despite the knowledge that there is a significant correlation between POP and LA injury, little is known about the in vivo stiffness properties of the LA muscle. Modification of stiffness properties would be desirable for childbirth at least. Preliminary work has shown that, by mechanically stretching the LA muscles for several weeks prior to delivery, the incidence of LA trauma during vaginal birth may be reduced.23 Quantifying this change in stiffness, with the use of the elastometer, would bring new insight to the mechanics of the LA muscle, and determine whether these properties can be modified. Measurement of stiffness properties of other skeletal muscles is commonly used to estimate flexibility of a muscle group by measuring range of motion about a joint.24 An increased measured passive stiffness was positively correlated to a decrease in range of motion, when investigating flexibility of the hamstring muscles.25 Although there is no joint about which torque can be estimated in the pelvic floor muscle, an increased range of motion (or ability to stretch) would be a desirable characteristic for the pelvic floor muscles during parturition. Further work is ongoing and results from the clinical study will be available once sufficient numbers have been recruited. However, the ability to regain ‘‘tone’’ is necessary for the pelvic organs to remain in place after childbearing. Quantifying muscle stiffness may be of benefit therefore, not only as a Neurourology and Urodynamics DOI 10.1002/nau
5
predictor for vaginal birth, but as a measure of the state of the muscle prior to or following physiotherapy or surgical treatment for pelvic floor muscle dysfunction. Limitations
The choice of a 5 min time interval between procedures was a compromise between patient comfort and our desire for the tissues to fully relax to their resting state prior to subsequent measurements. Previous research on the passive stretch of the hamstring complex showed no variation in the visco-elastic properties of the muscle after three 45 sec cycles of passive stretch, and a return to baseline after 2 min.26 The speculum attachments are specifically designed to preferentially engage the LA muscles, and avoid contact with the perineal muscles. However, it is likely that our stiffness measurements include a component contributed by the contact of the perineal muscles with the narrower neck of the speculum. Nonetheless, our measurements may still be useful for the purposes of providing a potential risk assessment tool for pelvic floor muscle injury, as all these tissues will be stretched during the birth process. Our maximum aperture of 50 mm is not likely to be near the stretch limit of the muscle and is one which has been well tolerated by this population. ACKNOWLEDGMENTS
Peter Blythe constructed the mechanical components of this hand-piece, and contributed to their design. Nikolaj Nielsen, Tzu-chin Yu, Matthew Parker, Gemma Goodfellow and Mihailo Azhar helped to design, construct, calibrate and test the elastometer system. The participants who gave their time to the study. REFERENCES 1. Dietz H, Lanzarone V. Levator trauma after vaginal delivery. Obstet Gynecol 2005;106:707–12. 2. Kearney R, Miller JM, Ashton-Miller JA, et al. Obstetric factors associated with levator ani muscle injury after vaginal birth. Obstet Gynecol 2006;107:144–9. 3. Dietz HP, Simpson JM. Levator trauma is associated with pelvic organ prolapse. BJOG 2008;115:979–84. 4. Mant J, Painter R, Vessey M. Epidemiology of genital prolapse: Observations from the Oxford Family Planning Association Study. Br J Obstet Gynaecol 1997;104:579–85. 5. Smith FJ, Holman CD, Moorin RE, et al. Lifetime risk of undergoing surgery for pelvic organ prolapse. Obstet Gynecol 2010;116:1096–100. 6. DeLancey JO. The hidden epidemic of pelvic floor dysfunction: Achievable goals for improved prevention and treatment. Am J Obstet Gynecol 2005; 192:1488–95. 7. Whyte TD, McNally DS, James ED. Six-element sensor for measuring vaginal pressure profiles. Med Biol Eng Comput 1993;31:184–6. 8. Dumoulin C, Bourbonnais D, Lemieux MC. Development of a dynamometer for measuring the isometric force of the pelvic floor musculature. Neurourol Urodyn 2003;22:648–53. 9. Miller JM, Ashton-Miller JA, Perruchini D, et al. Test–retest reliability of an instrumented speculum for measuring vaginal closure force. Neurourol Urodyn 2007;26:858–63. 10. Sanches PR, Silva DP Jr, Muller AF, et al. Vaginal probe transducer: Characterization and measurement of pelvic-floor strength. J Biomech 2009; 42:2466–71. 11. Saleme CS, Rocha DN, Del Vecchio S, et al. Multidirectional pelvic floor muscle strength measurement. Ann Biomed Eng 2009;37:1594–600. 12. Constantinou CE, Omata S, Yoshimura Y, Peng Q. Evaluation of the dynamic responses of female pelvic floor using a novel vaginal probe. In: Elad D, Young RC, editors. Reproductive Biomechanics. Oxford: Blackwell Publishing; 2007. p. 297–315. 13. Bo K. Pelvic floor muscle training in treatment of female stress urinary incontinence, pelvic organ prolapse and sexual dysfunction. World J Urol 2012;30:437–43. 14. Bergeron S, Rosen NO, Morin M. Genital pain in women: Beyond interference with intercourse. Pain 2011;152:1223–5.
6
Kruger et al.
15. Dumoulin C, Lemieux MC, Bourbonnais D, et al. Physiotherapy for persistent postnatal stress urinary incontinence: A randomized controlled trial. Obstet Gynecol 2004;104:504–10. 16. Lien K, Mooney B, DeLancey JOL, et al. Levator ani muscle stretch induced by simulated vaginal birth. Obstet Gynecol 2004;103:31–40. 17. Kruger J, Murphy B, Dietz H, et al. Pelvic floor muscle compliance in Elite nulliparous Athletes. 38th Annual meeting of the International Continence Society 2008; 20th–24th October 2008. 18. Kruger J, Nielsen P, Dietz HP, et al. Test–retest reliability of an instrumented elastometer for measuring passive stiffness of the levator ani muscle. Neurourol Urodyn 2011;30:865–7. 19. Verelst M, Leivseth G. Force-length relationship in the pelvic floor muscles under transverse vaginal distention: A method study in healthy women. Neurourol Urodyn 2004;23:662–7. 20. Morin M, Gravel D, Bourbonnais D, et al. Application of a new method in the study of pelvic floor muscle passive properties in continent women. J Electromyogr Kinesiol 2010;20:795–803.
Neurourology and Urodynamics DOI 10.1002/nau
21. Abramowitch SD, Feola A, Jallah Z, et al. Tissue mechanics, animal models, and pelvic organ prolapse: A review. Eur J Obstet Gynecol Reprod Biol 2009;144:S146–58. 22. Dietz HP. The aetiology of prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2008;19:1323–9. 23. Shek KL, Chantarasorn V, Langer S, et al. Does the Epi-No birth trainer reduce levator trauma? A randomised controlled trial. Int Urogynecol J 2011;22: 1521–8. 24. Magnusson SP. Passive properties of human skeletal muscle during stretch maneuvers. Scand J Med Sci Sports 1998;8:65–77. 25. Magnusson SP, Aagaard P, Simonsen EB, et al. Passive tensile stress and energy of the human hamstring muscles in vivo. Scand J Med Sci Sports 2000;10:351–9. 26. Magnusson SP, Aagaard P, Nielson JJ. Passive energy return after repeated stretches of the hamstring muscle-tendon unit. Med Sci Sports Exerc 2000; 32:1160–4.
