Neurourology and Urodynamics 30:692–699 (2011)

Modeling the Urinary Tract—Computational, Physical, and

Biological Methods
C.H. Fry,1* P. Sadananda,2 D.N. Wood,3 N. Thiruchelvam,4 R.I. Jabr,1 and R. Clayton5
1
Postgraduate Medical School, University of Surrey, Guildford, UK
2
Department of Physiology and Pharmacology, University of Bristol, Bristol, UK
3
Department of Urology, University College Hospital, London, UK
4
Department of Urology, Addenbrookes Hospital, Cambridge, UK
5
Department of Computer Science, University of Sheffield, Sheffield, UK

Models of the lower urinary tract are used to understand better the physiological and pathological functions of the tract
and to gain insight into the relative importance of different components. The key requirement of a model is described,
namely: to involve a continuous iteration with experiment; whereby experiments provide parameters and validation
for components of the model, which is then used to generate hypotheses, which are tested experimentally. Different
types of models are described: computational models that describe mathematically the whole urinary tract or com-
ponents; physical models useful especially in testing medical devices; and tissue-engineered models. The purpose of
modeling is first described in terms of the ability of models to predict the properties of the system of interest, using
components that have a physiological interpretation, and to gain insight into the relative importance of different com-
ponents. Examples are used to illustrate the use of modeling the urinary tract with reference to the different categories
listed above. Neurourol. Urodynam 30:692–699, 2011. ß 2011 Wiley-Liss, Inc.

Key words: encrustation; lower urinary tract; mathematical modeling; neurological models; tissue engineering

INTRODUCTION of the most well-known was the description by Hodgkin and

Huxley of electrical activity measured in the squid giant axon
The Purpose of Modeling to ‘‘put them into mathematical form and show that they will
account for conduction and excitation in quantitative terms.’’3
The fundamental behavior of the lower urinary tract (LUT)
This model was revolutionary because it not only provided
is well-understood: urine is propelled to the bladder, where it
completely new insight into the mechanisms underlying the
is stored and occasionally expelled. The proper function, and
action potential in excitable tissue, but also showed that tools
dysfunction, of this simple paradigm, however, involves
from physics and mathematics could be used to represent accu-
many physiological components that act at different time and
rately a detailed and multiscale physiological process. Signifi-
length scales, and predicting the contribution of any single
cantly, this model was modified 50 years ago by Denis Noble to
component in health and disease is very difficult. Moreover,
reconstruct the action potential of cardiac cells,4 and has under-
physical devices, such as stents and catheters are often used
pinned much of the recent activity in modeling the heart.5,6
to ameliorate the consequences of dysfunction but their
In the last two decades there has been a dramatic increase
proper composition and construction requires an understand-
in the availability of experimental data, with the advent of
ing of the physiological properties of the urinary tract. This
new techniques in molecular biology, genetics, and imaging.
complex and multiscale pattern of behavior is common to
There is a wealth and diversity of information at the cell and
many natural and man-made systems, and modeling is used
molecular scale. However, there is also a growing realization
widely in science and engineering to understand systems as
that the connection between these data and the function (and
diverse as cells,1 the early universe,2 as well as to design and
dysfunction) of systems at tissue and whole organ scales is
build systems such as aircraft. Broadly, a model is a simplified
extremely difficult to establish. Forging this connection is the
representation of an object, phenomenon, or process. Useful
vision of initiatives such as the virtual physiological human
models add value to existing knowledge because they are
(VPH), which aim to develop tools that combine system
quantitative, integrative, and predictive.
behaviors across multiple scales in tools that can be used by
The purpose of any biological model is to make predictions
basic scientists and clinicians for research, diagnosis, and
about the properties of the system of interest, using com-
guiding intervention.7
ponents that have a physiological interpretation, and to gain
The aims of this review are to consider models of the LUT, to
insight into the relative importance of different components.
identify knowledge gaps where a modeling approach may
The approach involves a reductionist and then constructionist
approach: to understand better the properties of isolated com-
ponents before reconstructing the original system. The most
Conflict of interest: none.
successful models involve a continuous iteration with exper- Christopher Chapple led the review process.
iment; where experiments provide parameters and validation *Correspondence to: C.H. Fry, Postgraduate Medical School, University of Surrey,
for components of the model, which is then used to generate Guildford, Surrey GU2 7WG, UK. E-mail: c.h.fry@surrey.ac.uk
new hypotheses, which are tested experimentally. Received 30 January 2011; Accepted 15 March 2011
Published online 15 June 2011 in Wiley Online Library
Students of physiology will be familiar with the previous (wileyonlinelibrary.com).
use of models to gain insight into the system of interest. One DOI 10.1002/nau.21131

ß 2011 Wiley-Liss, Inc.

 Modeling the Urinary Tract 693
yield new mechanistic insights, and to consider the challenges contractile properties of the bladder wall were added in a
in developing these new models. Animal models of the similar, global manner,14 and it was possible to estimate those
human system may be designed to mimic specific functions factors that were the principal determinants of pressure.15,16
or disorders, but these are discussed in a previous review8 and In this case the size and shape of the bladder, as well as its
not considered here. Our focus is on four types of LUT model: intrinsic contractile strength are significant and embodied in
such concepts as the Watts factor.17,18 These studies have per-
(i) Simplified, analytical models. For example, modeling the mitted nomograms and algorithms to be constructed that
urinary tract as a hydrodynamic system of fluid flow help, for example, to define the extent of functional bladder
through variable tubes, or reconstruction of urinary tract outflow obstruction and estimate detrusor contractile func-
function by adding together contributions from smaller tion.19,20 A benefit of such an approach is that certain
components. parameters may be defined to aid the clinician in describing
(ii) Detailed computational models. Reductionist models of the presence and extent of LUT dysfunction and propose strat-
the urinary tract, without the influence of confounding in egies of management. Attention may then focus on a more
vivo inputs. The examples used here are models of neuro- detailed description of component parts, such as the inter-
logical control of the LUT. action of contractile bladder function with the resistance of
(iii) Physical models using inert materials through which the urethra during normal voiding.21
urine flows at physiological rates. To test the physical An example of the approach is illustrated below where the
properties of materials used to construct stents or LUT is modeled as a contractile organ described by the Hill
catheters. equation22 contracting against an outflow resistance. The
(iv) Artificial biological models through cell cultures or organ relationship between the velocity, v, of muscle contraction
constructs. against a load F is



In setting out to generate and utilize a model that is useful F a a
to scientists and clinicians investigating the basic properties þ ðv þ bÞ ¼ 1 þ b 1
F0 F0 F0
or pathology of a system a number of principles are useful:
where F0 is the load at zero velocity and a,b are constants. For

Define the scope of the model: what are its aims and objec- a number of muscles the ratio a/F0 is about 0.2523 and replac-
tives—something specific or are there more generalized ing F by pressure, P and v by flow, Q24 yields the bladder
outcomes? output relation (BOR).

Ask an appropriate clinical question that may be modeled
with reliable readily-obtained data.



P0 Q 5

Make the model as simple as possible (Occam’s razor9), but Pþ Qþ 0 ¼ P0 Q 0 2
4 4 16
as complex as necessary to provide a heuristic in the devel-
opment of theoretical models. The simplest urethral pressure-flow function (urethral

Learn from the differences between the model output and resistance relationship, URR) is25:
real data to provide a way to improve the model (and
maybe question if the data collection is entirely correct). P ¼ Pop þ RQ 3

Use the output of the model to improve clinical understand-
ing and practice. where Pop is the urethral opening pressure and R is the outlet
resistance, assumed to be constant. Time-dependent acti-
vation of the bladder, (P(t)), and urethral relaxation, (R(t)), can
Simplified, Analytical Models be described by exponential functions with time constants t1,
t 2.
Urine flow through a complex structure such as the urinary
tract depends upon interacting physical properties of its com-




t t
ponent parts. The purpose of mathematical models is to recon- PðtÞ ¼ Pmax 1exp ; RðtÞ ¼ R0 exp þC 4
struct the function of the whole tract by summating in a t1 t2
linear or non-linear fashion the properties of the component
where Pmax is the maximum pressure during a contraction; R0,
parts so that the relative importance of different components
may be assessed. Rthe resistance at t ¼ 0 and C is a constant. Voided volume is
Qdt from which bladder volume is obtained by subtraction
Initial approaches were to treat urine flow through the uri-
from the starting value. The BOR and URR may be solved sim-
nary tract as a fluid dynamics problem. This was especially
ultaneously for increasing values of t and different variables
amenable with respect to the ureter and urethra that could be
plotted against each other and Figure 1 shows some of the
treated as collapsible (distensible) tubes surmounted by a var-
derived relationships. The results mirror some clinical obser-
iable pressure head, supported or not by peristaltic move-
vations, i.e., maximum flow rate increases with initial vol-
ments.10–12 From these studies basic conclusions were drawn;
ume.26 The simple model may be modified, such as the use of
e.g., in a closed tube the active and passive mechanical proper-
a distensible and curved urethra with a non-linear urethral
ties of the tube determine flow, whereas in open tubes outlet
pressure-flow relationship (P ¼ Pop þ RQ2).11,27,28
resistance dominates, and with peristalsis also present, the
magnitude and frequency of individual waves were also
crucial.
Similar models described the passive properties of the uri-
Detailed Computational Models
nary bladder during filling, without including a detailed
representation of structure and function.13 These models accu- The exponential increase of computing power with time
rately reproduced the pressure and volume response of the has enabled the development of more detailed models of LUT
bladder to both slow and stepwise filling. Subsequently, active function and neural control.

Neurourology and Urodynamics DOI 10.1002/nau

694 Fry et al.

Fig. 1. Output functions of a simple model pressure-flow relationships in the lower urinary tract: top, flow versus
time; bottom, pressure versus flow. Left: increasing initial bladder volume in 80 ml increments from an initial
value of 80 ml. Middle: increasing bladder stimulation rate with a time constant from 25 to 5.5 sec. Right:
decreasing bladder outflow resistance from 1.67 to 0.23 cm H2O ml1 sec1.

Models of LUT function. The VBN model of LUT function, simulates the electrical responses of motoneurones supplying
(named after authors of the original paper29) includes a the external urethral sphincter motoneurone after injection of
representation of the bladder that incorporates visco-elastic electric current. Results show a self-sustained discharge fol-
properties into the derivation of pressure changes from ten- lowing a depolarizing pulse, which is subsequently termi-
sion changes in the bladder wall, and takes account of the nated by a short hyperpolarizing pulse.
changing cross-sectional area and tortuosity of the urethra Moving up the hierarchy of bladder control, a flip-flop brain-
during voiding. Each physiological and physical change occur- stem model of bladder function, may be constructed [see36],
ring in LUT tissue during the micturition cycle as a function of where two physiologically important nuclei in the brainstem
time is incorporated into a numerical model. Volume-flow involved in filling and voiding, respectively, are considered as
curves produced by the VBN model are a good match to actual mutually inhibiting regions (Fig. 2A). The model proposes that
data recorded from patients, and enable reconstruction of one area inhibits a second during filling, thus maintaining
pathological conditions by making changes to physiologically- sphincter contraction. At void, the inhibition flips, such that
measured parameters and variables in the model.29,30 This the second area now inhibits the first, allowing bladder
type of model has also incorporated a description of the neu- contraction.
rophysiological control over the LUT29,31–33 – see below. More Simulations using this model are accurate when compared
recently a finite element model of the bladder-urethra during to biological experiments.36 For example, the model demon-
voiding has been devised.34 This model aims to reconstruct strates that in the absence of bladder efferents, voiding (but
the complex and irregularly-shaped anatomy of the LUT by not filling) was impaired either due to premature closure of
dividing it into a large number of discrete elements, each of the sphincter or due to incomplete detrusor contraction,
which is easier to analyze and whose solutions may be com- resulting in residual volume. Simulations using a flip-flop
bined to produce an overall output function. This reductionist model predict that bladder filling takes place without detrusor
viewpoint therefore divides the system to its smallest parts overactivity and that voiding is coordinated. The best anatom-
and micturition is modeled as a linear or non-linear sum of ical correlates of these areas are the L- and M-regions within
actions of individual constituents of the neurological circuitry. the brainstem. However, such mutual inhibition may not be
Models of neural control. Depending on the purpose of the physiologically realistic as these regions do not have anatom-
model, neural circuitry can be simplified to focus on specific ical interconnections, at least in the cat where such exper-
components or groups of components. For example, a model iments were performed.37
of the electrical behavior of a single external urethral sphinc- Some models take into consideration involvement of the
ter motoneurone has been developed, based on morphological cortex, such as one consisting of five ‘groups’ of paired
parameters such as soma size, dendritic diameter and con- neurones (Fig. 2B) to represent parasympathetic and sympath-
figuration, and several electrical parameters.35 The model etic neurons, the spinal cord, brainstem, (M-region) and the

Neurourology and Urodynamics DOI 10.1002/nau

 Modeling the Urinary Tract 695

Fig. 2. Simple neurological models of lower urinary tract control. A: A flip-flop model; left, Area 1 is active, result-
ing in inhibition of Area 2 and prevention of voiding; right, Area 2 is active, resulting in inhibition of Area 1 and
prevention of sphincter contraction. B: Architecture of a neural network, based on anatomical structures,
whereby the cortex and brain stem (M-region) ultimately relax (þ) or contract () the detrusor muscle. Positive
and negative feedback control modulates this activity. para: parasympathetic motoneurons; symp: sympathetic
motoneurons; inter: intermediolateral cell group of the spinal cord. Adapted from Ref. 35

cortex.35 This model studied detrusor muscle relaxation or Predictions made by this model are powerful and the
contraction as the sole outcome of simulations, ignoring any model has been validated by studying the actual properties of
actions on urethral function, and demonstrated the desired these neuronal populations. For example, the hypothesis
behavior of filling and voiding. implies that the firing rate of subthalamic neurones (STN)
Subsequent studies added additional anatomical and phys- should be proportional to the exponent of its input. Such a
iological information to permit spinal and supraspinal integ- relation is highly unusual in most neuronal populations. How-
ration and control to be included through addition of a novel ever, the response of STN has been extensively studied
‘‘group,’’ referred to as ‘‘urocontrol’’ (Fig. 3) although the ana- and conforms to this expectation. The novelty and elegance of
tomical substrate remains unclear.32 The ‘‘urocontrol’’ is likely this study lies in showing how a computation may be mod-
to contribute a minimal control strategy to realize a successful eled on a distinct network of neurons. Although not developed
micturition cycle. In addition, sacral and thoracolumbar (TH-L) specifically for bladder-related studies, the principles may
sensory information is postulated to reach the ‘‘urocontrol,’’ be applied to voluntary bladder control, where the voluntary
which is responsible for influencing directly the TH-L area, decision to void (based on bladder fullness, voiding opportu-
steering the M-region and relaying information to cortical nity, and access) can be broken down into a probability
structures. The ‘‘urocontrol,’’ L-region and sacral motoneurons function. Such a model can be used to predict the outcome
innervating the pelvic floor, receive cortical input. In this of pathology on bladder function. For example, in Parkinson’s
model, there is no interaction between the M and L regions. disease, where the basal ganglia are particularly affected,39
Study of the neural control of bladder has been generally patients have poor voluntary sphincter control,40 although
limited to autonomic control with little understanding about autonomic control remains intact. The pathology may be
voluntary aspects. However, fMRI studies have shown that mimicked by excluding the computational properties of
areas such as the insular cortex, thalamus, and basal ganglia specific groups of neurons to show the resulting effect on
play a role. Using a computational model of decision-making bladder function.
it was hypothesized that specific neuronal populations in the Models that describe the micturition cycle from the
basal ganglia involved in decision-making perform specific perspective of reductionist theories are made more useful by
functions, which can be computed to predict the final out- continuous update from new physiological and anatomical
come of the decision.38 Certain aspects of the neuronal circui- information.31 Although the reductionist approach is highly
try involved in decision-making were hypothesized to be successful, it also has drawbacks as its inappropriate use can
exactly those required to implement the computation defined limit our understanding of complex systems. Together with
by the Bayes probability theorem. This allows an end-behav- reducing a complex system as a method of its study, we must
ioral response (such as a voluntary desire to void) to be also find models to study how large scales of organization
studied by computing the individual functions of each set of influence smaller ones. Furthermore, we must study how
neurons. feedback loops, arising from constantly changing afferent

Neurourology and Urodynamics DOI 10.1002/nau

696 Fry et al.

Fig. 3. Spinal and supraspinal integration, and supraspinal control. The block ‘‘urocontrol’’ is postulated to serve
the control of the micturition cycle through the endpoints at the bottom of the flow diagram. A tenuous connec-
tion between the cortex and ‘‘urocontrol’’ may be added. Note the lack of connection between the L-region and M-
region in this model. Adapted from Ref. 32

inputs, might create structure at any given level. Thus both Artificial Biological Models
approaches, top-down and bottom-up, must be used to get a
full picture of the functioning of the organ system. Cellular and tissue construct models of the LUT serve two
potential purposes:
Physical Models
(1) Clinical applications for tissue replacement.
A practical requirement for the development of physical (2) The creation of a model to test functional hypotheses and
models is measurement of encrustation on catheters and interventions.
stents placed in the urinary tract. Clinical trials and animal
models are expensive, and the latter may not always be accu- With respect to clinical application, there have been several
rate as urine composition varies between animals, with com- attempts to apply tissue-engineering technology. Where the
ponents not always present in human urine. An alternative bladder is congenitally absent, incomplete, or suffers an
has been the development of inanimate models and an acquired insult, options for replacement are limited to the
example is a scaled glass model of the human urinary tract inclusion of a bowel segment in the urinary tract to create a
whereby the rate of encrustation of stents or catheters placed reservoir and/or conduit allowing the containment and sub-
in various regions of the model is measured, Figure 4.41 The sequent emptying of urine. These techniques have recognized
models are maintained in an environment of constant consequences, including the need for self-catheterization to
temperature and gas composition (usually 5% CO2 in air). Nov- empty, and complications including stones, urinary tract
el materials are generally tested against a standard construc- infection, metabolic derangement, and mucous production.46
tion made of silicon, which itself is resistant to encrustation.42 In addition, the in vivo functional inter-relationship of differ-
Shown here is its use to measure the rate of encrustation of ent tissues in the urinary tract remains unclear and artificial
catheters placed in the ‘‘ureter’’ or ‘‘bladder’’ portions. Urine tissue construct cultures offer a controlled way to study these
flowed through the system at constant rate and was siphoned processes, for example the influence of different cell types
away from the ‘‘bladder’’ portion when the volume had within the urothelium on detrusor function. Some of the
reached a defined value (300 ml). The catheters would be experimental approaches are outlined below:
removed from the system after a defined interval and ana- Attempts have been made to line de-epithelialized bowel
lyzed for the extent of encrustation and it composition (such and uterus with cultured urothelial cells – and results in a pig
as Ca and Mg salts, biofilm).43 Several media have been used model suggest potential for success.47 Cellular characteriz-
to induce encrustation, including freshly-voided or sterile ation of cultured, human urothelium suggested a high level of
urine as well as artificial urine. The latter has been used for phenotypic similarity to native tissue.48 Other animal models
convenience, although it may lack important constituents include a multilayered construct with tissue-engineered cells
such as Tamm-Horsfall protein.44 Testing materials for use in seeded onto a polymer scaffold49 and whilst successfully pro-
the LUT raises the additional problem of bacterial contami- ducing improved organ capacity there is little data about tis-
nation45 and thus the use of human urine in model systems, sue function. This has progressed to human implantation of a
artificially contaminated with organisms such as Proteus, has tissue-engineered bladder, optimized with an omental wrap,
been attempted. The key to such models is to ensure standard- and shows reasonable data for capacity and compliance with
ization of manufacture and use, so that there is real compara- follow-up data over a mean of 46 months.50 Although encour-
bility between different studies. aging much larger numbers are required before wider

Neurourology and Urodynamics DOI 10.1002/nau

 Modeling the Urinary Tract 697
contractile responses of muscle cells to physiological ago-
nists.56 Function may be further improved in a dynamic cul-
ture environment where muscle cells subjected to cyclical
strain over a 1–4 week period showed an enhancement of con-
tractile function.57 Cultured porcine urothelial cells exhibited
similar surface markers, such as UP IIIa, CK20, and ZO-1, and
physiological function, such as transepithelial resistance, to
native human urothelium58 lending weight to its suitability
as a potentially-useful tissue construct.
Injection of autologous myoblasts or stem cells to replace a
damaged or degenerate sphincter is an alternative, with good
initial results.59,60 This approach offers a relatively simple
solution, but injected cells are relied upon to differentiate sub-
sequently and long-term determination of their phenotype is
awaited. More recently it has been shown that systemic injec-
tion of mesenchymal stem cells can target the obstructed
bladder and are associated with a return to normal histologi-
cal and functional properties, possibly mediated by increased
tissue blood flow.61 The involvement of chemokines in stem
cell recruitment was hypothesized.
The exploration of LUT modeling using artificial tissue con-
structs is a novel and rapidly developing area. Initial excite-
ment about clinical application appears to have settled but
industrious pursuit of this important goal remains. The use of
single cell, single layer, and multilayer constructs as research
tools requires close definitions to achieve validity and compa-
rability with a clear understanding of cell environment and
cell-scaffold interactions. With each experimental model its
limitations must be explicitly stated in publications to allow
proper and comparable interpretation of data and offer the
potential to refine a model for both experimental testing and
clinical application.

DISCUSSION AND RECOMMENDATIONS

Despite the prevalence of LUTS, the diagnosis and manage-
ment of specific conditions remains difficult. In many
patients, symptoms do not relate directly to the underlying
condition because disease can produce a spectrum of patho-
physiological changes in more than one component of the
LUT. There is need to consider the LUT as an integrated sys-
tem, which interacts with the body as a whole, rather than an
Fig. 4. A glass model of the urinary tract for testing encrustation of isolated organ system.
catheters and stents for insertion into the urinary tract. A: The ‘‘bladder’’ This change in thinking is challenging because the LUT has
portion alone with a catheter in place. B: the ‘‘bladder’’ connected to a reser- a complex structure and function. From a macroscopic
voir by a glass ‘‘ureter.’’ The ‘‘bladder’’ drains automatically when the urine perspective the bladder can be viewed as a highly compliant
level reaches the top of the siphon. The dimensions of the system are balloon, which is able to contract and release urine through
similar to that of the adult human urinary tract. the urethra, which can be viewed as a collapsible tube. How-
ever, this large scale view overlooks important details that
include the features of the bladder wall that allow it to accom-
application may be justified on clinical and financial modate large changes in volume before a sensation of fullness
considerations. is generated, the way that relaxation of the bladder neck and
Culture of isolated cells gives products that are immunohis- sphincter is synchronized with contraction of the bladder
tochemically and histologically similar to native tissue.49,51 To muscle to enable emptying, and crucially how these behaviors
produce a foundation for safe and long-term clinical appli- are influenced by disease and lead to observable symptoms.
cation of this technology, artificially-generated cells and tis- In other organ systems, notably the heart and cardiovascu-
sues require characterization; i.e., what are the physiological lar system, multiscale modeling has provided tools that are
properties of these constructs and how have they been powerful because they allow our knowledge about processes
modified by exposure to an artificial culture environment? that operate at different spatial scales to be expressed. This is
For example, the developmental environment of the LUT influ- the vision of the virtual physiological human (VPH), which
ences final tissue phenotype.52 Furthermore, the mechanical aims to develop models that can be used to test and generate
properties of a support scaffold for cells determines cell hypotheses, and thus are capable of providing mechanistic
orientation and hence construct behavior53,54 and such insight into complex and multiscale physiological systems.
complex interactions vary between different support struc- Many of the tools and knowledge needed to build a multi-
tures.55 Functional outcomes of cells may be inferred by scale model of the human LUT already exist, but have been
examining their phenotype, for example measuring the developed piecemeal. What is missing is a co-ordinated

Neurourology and Urodynamics DOI 10.1002/nau

698 Fry et al.
interdisciplinary effort to bring these pieces together, and (vi) Although the reductionist approach of neuronal modeling
to undertake a systematic program of activity where new is highly successful, inappropriate use can limit our
modeling tools are evaluated and developed against exper- understanding of complex systems. Together with reduc-
imental and clinical data, generating anatomically and bio- ing a complex system as a method of its study, we must
physically-detailed models that can then be translated into also find models to study how large scales of organization
accessible computational tools for use by clinicians and influence smaller ones. Furthermore, we must study how
researchers. feedback loops, arising from constantly changing afferent
We reiterate the principles that stimulate the generation inputs, might create structure at any given level. Thus
and use of a clinically-useful model. both approaches, top-down and bottom-up, must be used
to get a full picture of the functioning of the organ

Define the scope of the model: its aims and objectives. system.

Define a clinical question that may be modeled with data
that is accurate and readily-obtained. ACKNOWLEDGMENTS

Make the model as simple as possible,9 but as complex as
We are grateful to Derek Griffiths and Frank Smith for dis-
necessary.
cussion regarding various aspects of this review and also for a

Use the differences between the model and real data to
preview of unpublished figures (DG). We acknowledge the
improve the model.
funding bodies supporting the work carried out by the differ-

Use the output of the model to improve clinical understand-
ent authors. The article was written whilst the corresponding
ing and practice.
author is in receipt of an EU FP7 INComb grant.
Based on the types of model reviewed here, we return to
Occam’s razor9 whereby models should start out simply and REFERENCES
then build in complexity as demanded from the disparity
1. Aldridge BB, Burke JM, Lauffenburger DA, et al. Physicochemical modelling
between actual data and model predictions. The ability of var- of cell signalling pathways. Nat Cell Biol 2006;8:1195–1203.
ious complexities to define better the outcome of any model 2. Springel V, Frenk CS, White SDM. The large-scale structure of the Universe.
then indicates the importance of such an inclusion. Nature 2006;440:1137–1144.
3. Hodgkin AL, Huxley AF. A quantitative description of membrane current
and its application to conduction and excitation in nerve. J Physiol 1952;
(i) The bladder is treated initially as a uniform sphere, 117:500–544.
with mechanical properties described by elastic, 4. Hutter OF, Noble D. Rectifying properties of heart muscle. Nature 1960;
contractile, and visco-elastic elements. This approach is 188:495.
relatively simple to implement, but does not represent 5. Fink M, Niederer SA, Cherry EM, et al. Cardiac cell modelling: observations
from the heart of the cardiac physiome project. Prog Biophys Mol Biol
well the bladder neck, nor normal shape changes in 2011;104:2–21.
the bladder. Replacing a spherical bladder with an 6. Clayton RH, Bernus O, Cherry EM, et al. Models of cardiac tissue electro-
oblate spheroid influences the compliance and strain physiology: progress, challenges and open questions. Prog Biophys Mol Biol
distribution in the bladder wall.62,63 The passive mechan- in press. 2011;104:22–48.
7. Hunter PJ, Coveney PV, de Bono B, et al. A vision and strategy for the virtual
ical properties of the bladder wall are also considered to physiological human in 2010 and beyond. Phil Trans R Soc A 2010;368:
be isotropic, although experimental data indicates 2595–2614.
anisotropy.64 8. Fry CH, Daneshgari F, Thor K, et al. Animal models and their use in under-
(ii) The different roles of detrusor muscle, lamina propria, standing lower urinary tract dysfunction. Neurourol Urodyn 2010;29:603–
608.
and urothelium during accommodation and stretch are 9. de Ockam G. Quaestiones et decisiones in IV libros Sententiarum cum centi-
not explicitly considered. Little is known about strain logio theologico. Lugduni: Johannes Trechsel; 1495.
distribution within the bladder wall during expansion, 10. Griffiths D. The pressure within a collapsed tube, with special reference to
and how this depends on the arrangement of muscle fiber urethral pressure. Phys Med Biol 1985;30:951–963.
11. Griffiths DJ. Dynamics of the upper urinary tract: I. Peristaltic flow through
bundles. These considerations may be very important in a distensible tube of limited length. Phys Med Biol 1987;32:813–822.
mediating the stretch response, which in turn may be sig- 12. Griffiths DJ, Constantinou CE, Mortensen J, et al. Dynamics of the upper
nificant for the underlying mechanism of over- and urinary tract: II. The effect of variations of peristaltic frequency and
under-active bladder behavior. bladder pressure on pyeloureteral pressure/flow relations. Phys Med Biol
1987;32:823–833.
(iii) Activation and force generation in the bladder is assumed 13. van Mastrigt R, Coolsaet BL, van Duyl WA. Passive properties of the urinary
to be uniform and isotropic. The role of bladder geometry, bladder in the collection phase. Med Biol Eng Comput 1978;16:471–482.
the layered structure of the wall, and the importance of 14. Griffiths DJ, van Mastrigt R, van Duyl WA, et al. Active mechanical proper-
supporting structures are not considered, yet may influ- ties of the smooth muscle of the urinary bladder. Med Biol Eng Comput
1979;17:281–290.
ence intravesical pressure and flow during emptying. 15. van Mastrigt R, Griffiths DJ. An evaluation of contractility parameters deter-
How these features are modified in patients with over- mined from isometric contractions and micturition studies. Urol Res
and under-active bladders is unknown. 1986;14:45–52.
(iv) The detailed structure of the bladder neck, trigone, and 16. Cucchi A, Quaglini S, Rovereto B. Proposal for a urodynamic redefinition of
detrusor underactivity. J Urol 2009;181:225–229.
external sphincter are not modeled explicitly, yet are 17. Griffiths DJ. Assessment of detrusor contraction strength or contractility.
likely to have an important effect on bladder emptying, Neurourol Urodyn 1991;10:1–18.
especially in the overactive bladder where remodeling of 18. Lecamwasam HS, Yalla SV, Cravalho EG, et al. The maximum Watts factor as
the bladder neck and detrusor muscle can take place. The a measure of detrusor contractility independent of outlet resistance. Neuro-
urol Urodyn 1998;17:621–635.
arrangement of muscle fiber bundles is presumably im- 19. Abrams PH, Griffiths DJ. The assessment of prostate obstruction from urody-
portant, as the relaxed state has to maintain continence, namic measurements and from residual urine. Br J Urol 1979;51:129–134.
but this has not been examined in the current generation 20. Tammela TL, Schäfer W, Barrett DM, et al. The Finasteride Urodynamics
of LUT models. Study Group. Repeated pressure-flow studies in the evaluation of bladder
outlet obstruction due to benign prostatic enlargement. Neurourol Urodyn
(v) Models of the urethra are not sufficiently detailed to 1999;18:17–24.
include possible effects of gender, specific urethral geom- 21. van Mastrigt R, Kranse M. Automated evaluation of urethral obstruction.
etry and catheter insertion on urine flow. Urology 1993;42:216–224.

Neurourology and Urodynamics DOI 10.1002/nau

 Modeling the Urinary Tract 699
22. Hill AV. The heat of shortening and the dynamic constants of muscle. Proc R 44. Tunney MM, Keane PF, Jones DS, et al. Comparative assessment of ureteral
Soc B 1938;126:136–195. stent biomaterial encrustation. Biomaterials 1996;17:1541–1546.
23. Van Mastrigt R. Estimation of contractility parameters for the urinary blad- 45. Choong S, Wood S, Fry C, et al. Catheter associated urinary tract infection
der. Neurourol Urodyn 1991;10:37–45. and encrustation. Int J Antimicrob Agents 2001;17:305–310.
24. Hübener U, Van Mastrigt R. Computer simulation of micturition. Urodina- 46. Greenwell TJ, Venn SN, Mundy AR. Augmentation cystoplasty. BJU Int
mica 1994;4:81–90. 2001;88:511–525.
25. Griffiths DJ, Rollema HJ. Urine flow curves of health males: a mathematical 47. Fraser M, Thomas DF, Pitt E, et al. A surgical model of composite cystoplasty
model of bladder and urethral function during micturition. Med Biol Eng with cultured urothelial cells: a controlled study of gross outcome and uro-
Comp 1979;17:291–300. thelial phenotype. BJU Int 2004;93:609–616.
26. Sjöberg B, Nyman CR. Hydrodynamics of micturition in healthy females: 48. Cross WR, Eardley I, Leese HJ, et al. A biomimetic tissue from cultured nor-
pressure and flow at different micturition volumes. Urol Int 1981;36:23–34. mal human urothelial cells: analysis of physiological function. Am J Physiol
27. Schäfer W. Urethral resistance: urodynamic concepts of physiological and 2005;289:F459–F468.
pathological bladder outlet function during voiding. Neuorurol Urodyn 49. Oberpenning F, Meng J, Yoo JJ, et al. De novo reconstitution of a functional
1985;4:161–201. mammalian urinary bladder by tissue engineering. Nat Biotechnol 1999;
28. Bush MB, Petros PE, Barrett-Lennard BR. On the flow through the human 17:149–155.
female urethra. J Biomech 1997;30:967–969. 50. Atala A, Bauer SB, Soker S, et al. Tissue-engineered autologous bladders for
29. Valentini FA, Besson GR, Nelson PP, et al. A mathematical micturition model patients needing cystoplasty. Lancet 2006;367:1241–1246.
to restore simple flow recordings in healthy and symptomatic individuals 51. De Filippo RE, Yoo JJ, Atala A. Urethral replacement using cell seeded tubu-
and enhance uroflow interpretation. Neurolurol Urodyn 2000;19:153–176. larized collagen matrices. J Urol 2002;168:1789–1792.
30. Valentini FA, Nelson PP, Besson GR, et al. Challenging the maximum flow 52. Nyirády P, Thiruchelvam N, Godley ML, et al. Contractile properties
rate: a new index of voiding dysfunction in men with benign prostatic of the developing fetal sheep bladder. Neurourol Urodyn 2005;24:276–
enlargement. BJU Int 2008;101:995–999. 281.
31. Bastiaanssen EH, van Leeuwen JL, Vanderschoot J, et al. A myocybernetic 53. Rohman G, Pettit JJ, Isaure F, et al. Influence of the physical properties of
model of the lower urinary tract. J Theor Biol 1996;178:113–133. two-dimensional polyester substrates on the growth of normal human uro-
32. Kinder MV, Bastiaanssen EH, Janknegt RA, et al. The neuronal control of the thelial and urinary smooth muscle cells in vitro. Biomaterials 2007;28:
lower urinary tract: a model of architecture and control mechanisms. Arch 2264–2274.
Physiol Biochem 1999;107:203–222. 54. Baker SC, Atkin N, Gunning PA, et al. Characterisation of electrospun poly-
33. Le Feber J, Van Asselt E, Van Mastrigt R. Neurophysiological modeling of styrene scaffolds for three-dimensional in vitro biological studies. Biomate-
voiding in rats: bladder pressure and postganglionic bladder nerve activity. rials 2006;27:3136–3146.
Am J Physiol 1997;272:R413–R421. 55. Baker SC, Rohman G, Southgate J, et al. The relationship between the mech-
34. Pel JJ, van Mastrigt R. Development of a CFD urethral model to study flow- anical properties and cell behaviour on PLGA and PCL scaffolds for bladder
generated vortices under different conditions of prostatic obstruction. Phys- tissue engineering. Biomaterials 2009;30:1321–1328.
iol Meas 2007;28:13–23. 56. Wood DN, Brown RA, Fry CH. Characterization of the control of intracellular
35. Bastiaanssen EH, Vanderschoot J, van Leeuwen JL. Learning procedure in a [Ca2þ] and the contractile phenotype of cultured human detrusor smooth
neural control model for the urinary bladder. Neurourol Urodyn 1993; muscle cells. J Urol 2004;172:753–757.
12:285–288. 57. Moon du G, Christ G, Stitzel JD, et al. Cyclic mechanical preconditioning
36. Fowler CJ, Griffiths DJ. A decade of functional brain imaging applied to blad- improves engineered muscle contraction. Tissue Eng A 2008;14:473–482.
der control. Neurourol Urodyn 2010;29:49–55. 58. Turner AM, Subramaniam R, Thomas DF, et al. Generation of a functional
37. Blok BF, Holstege G. Two pontine micturition centers in the cat are not inter- differentiated porcine urothelial tissue in vitro. Eur Urol 2008;54:1423–
connected directly: implications for the central organization of micturition. 1432.
J Comp Neurol 1999;403:209–218. 59. Mitterberger M, Pinggera GM, Marksteiner R, et al. Adult stem cell therapy
38. Bogacz R, Gurney K. The basal ganglia and cortex implement optimal of female stress urinary incontinence. Eur Urol 2008;53:169–175.
decision making between alternative actions. Neural Comput 2007;19:442– 60. Strasser H, Marksteiner R, Margreiter E, et al. Autologous myoblasts
477. and fibroblasts versus collagen for treatment of stress urinary incontinence
39. Gale JT, Shields DC, Jain FA, et al. Subthalamic nucleus discharge patterns in women: a randomised controlled trial. Lancet 2007;369:2179–2186.
during movement in the normal monkey and Parkinsonian patient. Brain 61. Woo LL, Tanaka ST, Anumanthan G, et al. Mesenchymal stem cell recruit-
Res 2009;1260:15–23. ment and improved bladder function after bladder outlet obstruction: pre-
40. Staskin DS, Vardi Y, Siroky MB. Post-prostatectomy continence in the parkin- liminary data. J Urol 2011;183:1132–1138.
sonian patient: the significance of poor voluntary sphincter control. J Urol 62. Damaser MS, Lehman SL. The effect of urinary bladder shape on its
1988;140:117–118. mechanics during filling. J Biomech 1995;28:725–732.
41. Choong SK, Wood S, Whitfield HN. A model to quantify encrustation on ure- 63. Damaser MS, Lehman SL. Two mathematical models predict the variation in
teric stents, urethral catheters and polymers intended for urological use. capacity and compliance of hypertrophied bladders. Adv Exp Med Biol
BJU Int 2000;86:414–421. 1995;385:255–256.
42. Morris NS, Stickler DJ, Winters C. Which indwelling urethral catheters resist 64. Korossis S, Bolland F, Southgate J, et al. Regional biomechanical and histo-
encrustation by Proteus mirabilis biofilms? Br J Urol 1997;80:58–63. logical characterisation of the passive porcine urinary bladder: implications
43. Shaw GL, Choong SK, Fry CH. Encrustation of biomaterials in the urinary for augmentation and tissue engineering strategies. Biomaterials 2009;
tract. Urol Res 2005;33:17–22. 30:266–275.

Neurourology and Urodynamics DOI 10.1002/nau