Elsevier

Computers in Biology and Medicine

Volume 65, 1 October 2015, Pages 269-278
Computers in Biology and Medicine

A study of human colonic motility in healthy and constipated subjects using the wireless capsule

https://doi.org/10.1016/j.compbiomed.2015.07.011Get rights and content

Abstract

Constipation is a common and distressing condition that has been linked to major morbidity, burdens the health care system, and impacts patients׳ quality of life. However, there is no perfect method for diagnosing and treating constipation. The purpose of this paper is to develop an automatic algorithm to identify patients with constipation from healthy subjects. Data from 12 healthy subjects and 10 patients with constipation were analyzed. The key challenges for data processing were data filtering, feature extraction, information evaluation, and providing the reference conclusion; these were resolved by employing the phase space reconstruction (PSR), independent component analysis (ICA), dynamic feature extraction algorithm, and the Wilcoxon rank sum test. The contractile frequency (Fr), motility index per unit time (MIU), average peak of peristaltic wave (Pave) and variance (Var) were extracted as dynamic parameters and analyzed. Results between groups were compared with the Wilcoxon rank sum test. There were statistically significant differences between healthy subjects and patients with constipation for Fr and MIU (P<0.05), whereas there was no statistically difference for Var. Moreover, the Fr and MIU of patients with normal transit constipation (NTC) are significantly lower compared to healthy subjects, whereas patients with slow transit constipation (STC) did not show significant differences. The proposed algorithms were able to differentiate between healthy subjects and patients with constipation based on the colonic motility profiles.

Introduction

Constipation is a very common symptom that is estimated to affect about 15% of adults and 9% of children [1], [2]. To facilitate diagnosis, constipation was divided into three types as described by American Gastroenterological Association: normal transit constipation (NTC), slow transit constipation (STC), and defecatory disorders (DD) [1]. STC refers to patients whose colonic transit time (CTT) exceeds 59 h, but the patients do not have a defecatory disorder [1], [3]. NTC refers to patients who have fewer high-amplitude propagated contractions and reduced phasic contractile responses to a meal or to pharmacologic stimuli, but the patients׳ CTT is normal [1]. DD are primarily characterized by impaired rectal evacuation [1]. Diagnosis of constipation and its subtypes is based on symptoms and various physiologic tests including ano-rectal manometry, defecography, rectal barostat test, measures of colonic transit, and colonic manometry [4], [5], [6], [7], [8], [9], [10], [11], [12]. Ano-rectal and colonic manometry can cause skin irritation or infections, thus contributing to deteriorating health conditions [13]. Defecography has drawbacks of radiation exposure, embarrassment, interobserver bias, and inconsistent methodology [14]. Barostat comprises of a highly compliant balloon that is placed in the rectum and connected to a computerized pressure-distending device, which can be used to assess rectal sensation, tone and compliance [14]. However, it is invasive and brings a lot of pain to patients. These problems have led to the development of novel noninvasive monitoring systems, such as the wireless power capsule (WPC) produced by the SmartPill Corporation and the multi-parameters medical information noninvasive detection system (MMI-NDS) developed by us [15]. The MMI-NDS can continuously measure pH, temperature, and pressure information from the human gastrointestinal (GI) tract under normal physiological conditions and transmit the information wirelessly via radio. Patients׳ bowels do not need to be cleared before inspection and patients can move freely during the examination. The MMI-NDS is much more tolerable for patients with constipation. In this paper, the GI data were obtained with the MMI-NDS.

Based on GI physiology parameters, the researchers tried a variety of methods to diagnose GI diseases. CTT has been used in adults and children to investigate colonic motility [16], [17], [18], [19], [20], but CTT could not distinguish healthy subjects from patients with NTC. Colonic manometry was used to distinguish normal colonic motor function from colonic neuromuscular disorders [21], [22], however, patients must clear their bowels before the test, and the data obtained from the manometer cannot characterize the gastrointestinal dynamics under normal physiological state. Kloetzer et al. analyzed the number of contractions (Ct), area under the pressure curve, and motility index (MI) of GI data from 71 healthy and 42 gastroparesis and found a significant difference between healthy and gastroparetic subjects for Ct and MI [23]; however, the data were not filtered before analysis which may lead to incorrect results to some extent.

The novelty and significance of this manuscript can be summarized as the following three points:

  • 1)

    To better diagnose human gastrointestinal disease, we developed the novel MMI-NDS and solved the designed challenges of reducing system size and power consumption, and improving the measurement accuracy of the sensors [15]. The detailed description of MMI-NDS can be found in Section 2.

  • 2)

    In addition to inability to diagnose patients with NTC and characterize the gastrointestinal dynamics under normal physiological state, all of the above studies were limited to STC and did not include NTC. Moreover, all of the above studies did not combine with data filtering, feature extraction, information evaluation to develop an automatic algorithm to identify constipation, so in this paper, constipated patients included not only patients with STC, but also patients with NTC. Besides that, we developed an automatic algorithm combining with phase space reconstruction (PSR), independent component analysis (ICA), dynamic feature extraction algorithm, and Wilcoxon rank sum test to identify constipation.

  • 3)

    Method validation was performed by comparing the analysis results to other literatures. According to the results of comparison analysis, the proposed methods can identify more patients from healthy subjects comparing to the algorithm presented in the other literature.

Section snippets

Pressure monitoring system

GI tract pressure data were obtained by the MMI-NDS that we developed recently. The MMI-NDS includes a wireless capsule, a recorder and a workstation, as shown in Fig. 1. The capsule measured 25 mm in length by 11 mm in diameter can continuously monitor pH, temperature, and pressure of human GI tract under normal physiological conditions at a sampling frequency of 0.83 Hz [15]. The recorder records the data transmitted by the capsule and stores them in the storage card. Then the recorder is

Data analysis methods

Because the raw data recording by the recorder was belong to the whole digestive tract, in order to obtain colonic pressure data, the GI pressure data were divided into three sections—gastric, small intestine, and colonic pressure data—according to GI pH information [24].

The key challenges for data analysis were data filtering, feature extraction, information evaluation, and providing the reference conclusion. In this paper, data filtering was accomplished through the following three steps: (a)

Threshold method

Because the raw data contained not only breathing noise, coughing noise but also electromagnetic noise, an adjusted threshold filtering method was used to reduce the influence of jamming signals. The results of data filtering are shown in Fig. 4. As can be seen from the graph, the abnormal values caused by coughing and electromagnetic noise in Fig. 4(a) were filtered by the threshold method. The results of the filtering are shown in Fig. 4(b). However, the breathing noise and the other

Discussion

Colonic motor disturbances in chronic constipation are heterogeneous and incompletely understood [39]; the differences in colonic motility between healthy subjects and patients with constipation were not clear. The aim of this paper was to analyze human colonic pressure signals and found differences between healthy subjects and patients with constipation to provide a reference for the diagnosis of constipation.

Data from 12 healthy subjects and 10 patients with constipation were analyzed. There

Conclusion

To better identify constipated patients, we developed an automatic algorithm combining with PSR, ICA, dynamic feature extraction algorithm, and Wilcoxon rank sum test. Colonic pressure data from 12 healthy subjects and 10 patients with constipation were analyzed. The following observations were obtained:

  • 1)

    There was a statistically significant difference for Fr and MIU between healthy subjects and patients with constipation. However, there was no statistically significant difference for Pave and

Conflict of interest statement

There is no conflict of interest.

Li Lu received the B.S. degree in physical engineering from HuangHuai University, Zhumadian, China, in 2008 and the M.S. degree in physical engineering from Zhenzhou University, Zhenzhou, China, in 2011. She is currently pursuing the Ph.D. degree at the Department of Instrument Science and Engineering, Shanghai Jiaotong University. Her research interests include wireless communication, design of the micro medical device, digital IC design, micro-sensor, and data processing.

References (41)

  • G.R. Locke et al.

    A new questionnaire for gastroesophageal reflux disease

    (1994)
  • I. Nejadgholi et al.

    Using phase space reconstruction for patient independent heartbeat classification in comparison with some benchmark methods

    Comput. Biol. Med.

    (2011)
  • R. Khatibi et al.

    Investigating chaos in river stage and discharge time series

    J. Hydrol.

    (2012)
  • M.A. Ghorbani et al.

    A probe into the chaotic nature of daily streamflow time series by correlation dimension and largest Lyapunov methods

    Appl. Math. Model.

    (2010)
  • H. Parastar et al.

    Is independent component analysis appropriate for multivariate resolution in analytical chemistry?

    TrAC Trends Anal. Chem.

    (2012)
  • Y.-K. Jan et al.

    Wavelet analysis of sacral skin blood flow oscillations to assess soft tissue viability in older adults

    Microvasc. Res.

    (2009)
  • K. Ravi et al.

    Phenotypic variation of colonic motor functions in chronic constipation

    Gastroenterology

    (2010)
  • B. Kuo et al.

    Generalized transit delay on wireless motility capsule testing in patients with clinical suspicion of gastroparesis, small intestinal dysmotility, or slow transit constipation

    Dig. Dis. Sci.

    (2011)
  • P. Bampton et al.

    High resolution colonic manometry–what have we learnt?-A review of the literature 2012

    Curr. Gastroenterol. Rep.

    (2013)
  • M. Camilleri et al.

    Wireless pH‐motility capsule for colonic transit: prospective comparison with radiopaque markers in chronic constipation

    Neurogastroenterol. Motil.

    (2010)
  • Cited by (3)

    • Minimally invasive wireless motility capsule to study canine gastrointestinal motility and pH

      2017, Veterinary Journal
      Citation Excerpt :

      The WMC is a 13 × 27 mm wireless sensor that measures pH, temperature, pressure and time simultaneously. In humans, the information obtained from the WMC aids in identifying and describing disorders of the GIT (Green et al., 2013; Lu et al., 2015). The aim of the study was to evaluate the WMC and assess the technique feasibility in non-hospitalized dogs.

    • A Flexible Implant for Multi-Day Monitoring of Colon Segment Activity

      2023, IEEE Transactions on Biomedical Circuits and Systems
    • Functional Disorders: Slow-Transit Constipation

      2017, Clinics in Colon and Rectal Surgery

    Li Lu received the B.S. degree in physical engineering from HuangHuai University, Zhumadian, China, in 2008 and the M.S. degree in physical engineering from Zhenzhou University, Zhenzhou, China, in 2011. She is currently pursuing the Ph.D. degree at the Department of Instrument Science and Engineering, Shanghai Jiaotong University. Her research interests include wireless communication, design of the micro medical device, digital IC design, micro-sensor, and data processing.

    Guozheng Yan received bachelor Ph.D. from Jilin University of Technology in 1993; received a postdoc position in Nanjing University of Aeronautics and Astronautics. Now he is a professor in Shanghai Jiaotong University. His research interest is medical precision instrument engineering.

    Jingjing Sun born in 1986, received the B.S. degree in communication engineering from JiangXi Normal University, Nanchang, China, in 2006 and the M.S. degree in Optics from JiangXi Normal University, Nanchang, China, in 2009, and received the Ph.D. degree at the Department of Physics, Shanghai Jiaotong University, China, in 2013. She is currently working at Shanghai Institute of Quality Inspection and Technical Research, Shanghai, China. Her research interests include biochemical sensor, magnetical fluid, testing of electronic appliances.

    View full text