Imperial College London

Emeritus ProfessorColinCaro

Faculty of EngineeringDepartment of Bioengineering

Senior Research Investigator



+44 (0)20 7594 5180c.caro




Mrs Andrea Laine +44 (0)20 7594 5181




4.03Royal School of MinesSouth Kensington Campus





Find out more about my research and the Circulatory Fluid Mechanics Group.

I have qualifications in physiology and medicine and a strong interest in mechanics.  My overall objective has been to foster interaction between physiology/medicine and physical science/engineering.  In so doing, I have aimed at advancing understanding of the mechanics underlying some normal and disturbed physiological processes and of contributing to the detection/management of medical conditions.  In addition, I have aimed at contributing from physiology/medicine to physical science/engineering.

Respiratory mechanics

My early research interest was in respiratory mechanics.  I found that brief restriction of chest cage expansion in normal subjects reproduced in them several of the respiratory features of hunchback.  These included reduction of lung compliance, seemingly because of 'collapse' of lung units, and airway resistance.  I also reported - apparently the first account - dependence of airway resistance on lung elastic recoil.  I have recently returned to research in respiratory mechanics.  On recognizing that arterial geometry is commonly non-planar and associated with swirling flow, I investigated the geometry and flow of the larger airways.  The geometry has usually been considered planar, or two-dimensional.  Airway casts show the bifurcations to be approximately planar.  However, over about the first 10 bronchial generations, the plane of bifurcation rotates by approximately 90 degrees between successive generations.  Using a mass transport technique to assess wall shear rate, I found, with steady physiological-level 'inspiratory' flow in a non-planar airway model, swirling of the flow and Reynolds number dependence of the distribution of wall shear.  These new findings predict intra-bronchial mixing and greater uniformity of the distribution of wall shear than in an otherwise comparable planar airway system.  The findings have implications for all airway transport processes, including the distribution of inhaled particles/aerosols. 

Atherosclerosis and arterial wall shear

Atherosclerosis, through its causation of events such as heart attack and stroke, remains the leading source of mortality in the western world.  When I entered this field in the 1960s, the received wisdom, dating from the mid-nineteenth century and supplemented by contemporary work, was that the lesions/plaques resulted from damage to the arterial wall, inflicted by the flowing blood.  I was aware that wall shear rate could influence blood-arterial wall mass transport and would vary spatially at bends and branches, that there were biochemical reactions involving precursors and cholesterol in the arterial wall, and that atherosclerotic lesions were distributed patchily in the arteries.  I conjectured, on that basis, that cholesterol (effectively lipoprotein) would accumulate preferentially in low wall shear regions, which would be the preferred sites for atherosclerosis. 

Post-mortem studies we made of the distribution of early atherosclerosis in human arteries, together with studies of the distribution of wall shear in models, were consistent with that correlation.  Given the longstanding belief that atherosclerosis resulted from damage to the arterial wall, that proposal was not accepted immediately.  It is now, however, widely agreed that atherosclerosis in adults, whether early or advanced, or associated with various risk factors for the condition, preferentially affects low wall shear regions; the correlation has become known as the low wall shear theory for atherosclerosis.  Our initial proposal emphasized the contribution of wall shear rate.  We recognized, however, that wall shear stress might influence wall permeability and subsequently demonstrated that, while wall shear could influence the transport of macromolecules between the blood and the arterial wall, the process was controlled by the 'resistance' of the blood-wall interface, implying an effect of wall shear stress. 

These findings would seem to have facilitated the discovery that intimal hyperplasia (the main cause of failure of arterial bypass grafts and vascular access grafts - the grafts implanted in patients requiring dialysis for renal failure) also develops preferentially in low wall shear regions.  They have furthermore stimulated the undertaking of a very large number of studies.  Among these have been investigations of arterial fluid and wall mechanics, arterial wall mass transport, and vascular cellular and molecular biology, including the mechanism of transduction of mechanical stress.  We have demonstrated among other findings that: there is transport of material across the arterial wall; concentration polarization at the blood wall interface; a low distribution volume for macromolecules in the medial (smooth muscle-containing) layer of the arterial wall; and a strong effect of applied stress and wall smooth muscle tone on that volume.  A low porosity of the media, could account for the observed accumulation of lipid and the initiation of atherosclerosis at the intima-media interface.  We showed, in addition, that arterial distensibility is reduced by the risk factors for atherosclerosis, cigarette smoking and nicotine and increased by the vasodilator nitrate, implying respectively increase and decrease of medial smooth muscle tone.  Moreover, as we showed, smoking reduces and nitrate increases the pulsatility of arterial blood flow.

Non-planar vascular geometry and flow

On observing in a crude model that the introduction of a bend upstream and non-planar to an existing bend caused swirling of the flow and improved the clearance of material from the system, I further investigated arterial geometry and flow.  It has been usual in modeling arterial flow to assume that the geometry is planar, other than at a few sites such as the aortic arch.  However, casts of human and animal arteries and magnetic resonance imaging (MRI) studies we performed in volunteers, revealed that arterial branching and curvature is commonly non-planar and associated with asymmetric flow.  Model studies showed that, in addition to swirling, features of the flow included in-plane mixing, a relatively uniform distribution of axial velocity and wall shear, and inhibition of flow separation and flow instability.  These features, which contrast with those present in planar geometries, may offer new insight and opportunities. 

As already noted, arterial bypass grafts and vascular access grafts are prone to failure, because of the development (particularly in low wall shear regions) of intimal hyperplasia.  MRI studies we performed showed that lower limb arterial bypass grafts in patients were generally planar, whereas the foregoing implies the importance of non-planar geometry. As a means of ensuring non-planar geometry and swirling flow at arterial bypass grafts and vascular access grafts, we have developed within an Imperial College spin-out company (Veryan Medical Limited) of which I am the Founder and a Consultant, small amplitude helical technology (SMAHT) PTFE grafts.  These grafts have a circular cross-section and have been shown to generate swirling flow.  The grafts have been implanted as carotid artery-jugular vein shunts in a small porcine study, with conventional PTFE grafts implanted contralaterally as controls in each animal. At sacrifice at 2 months, the conventional grafts were found to be obstructed by IH and thrombus, whereas there was virtually no pathology in the SMAHT grafts [226].  Having obtained European Regulatory approval, the SMAHT grafts have been implanted as vascular access grafts, in a small non-randomized safety and efficacy study in patients requiring renal dialysis [233].  At 6 months, the primary patency was within the conventional range, but the secondary patency was reported as being unequalled. Following FDA 510K approval, a randomized trial of the SMAHT grafts as vascular access grafts has been launched in the US. Continuing this general approach, we are developing a helical endovascular stent.  Numerical and model studies show the device to generate swirling flow and in-plane mixing. They show in addition that the helical geometry renders the stent relatively resistant to kinking and can be expected to render it relatively resistant to fracture.  The stent has been implanted to date in the arteries of a small number of pigs and as shown in in vivo studies to generate swirling flow. 

Following the success of the helical-centreline stent in suppressing intimal hyperplasia in porcine studies and given appropriate ethical approval, Veryan Medical (the spin-out company I founded), has conducted the first randomised trial of the stent (now named the BioMimics 3D helical-centreline arterial stent) in patients with femoropopliteal atherosclerosis. The prospective, controlled, 2:1 trial in 75 patients, in which the 3D stent was compared primarily with the Bard Life Stent, has given distinctly encouraging results at 12 months after stent implantation.  The trial is however still ongoing. As indicated by ultrasound studies and angiography, the 3D stent caused swirling of local arterial flow, predicted to increase wall shear stress and blood-wall oxygen transport. 

There have been reports of fluid-phase-controlled diffusional transport of species including oxygen and ATP between the blood and artery wall, and local wall hypoxia has been proposed as causative of atherosclerosis.  However, attention in this work has largely been confined to effects of wall shear rate, with little consideration of secondary motion and convection-enhanced transport.  Typical diffusion coefficients and diffusion distances are 1-2 x 10-5 cm2/sec and 0.1 - 1.0 cm, respectively, whereas velocities associated with secondary motion in non-planar conduits can be in the range 1-5 cm/sec.  We have modeled blood-wall oxygen transport and found that physiological swirling can substantially reduce the thickness of the diffusion boundary layer and increase oxygen concentration at the wall (including in low wall shear regions) [232].  Given the non-planarity of the arterial system and likelihood of the presence of swirling flow and in-plane mixing, this work may return interest to the role of blood-wall mass transport in vascular biology/pathology.  Moreover, it may focus interest on attendant diagnostic and therapeutic procedures.

Industrial Applications

The possibility that the non-planar geometry of the arteries and larger airways and the associated flow has an evolutionary/survival basis, led me to consider whether the findings could have relevance beyond the vascular and respiratory systems in, for example, the design of general or specialized piping systems.  We observed, in experimental studies, less flow instability in a non-planar than planar bend, at Reynolds numbers up to 8,000 [180].  Moreover, we found computationally, less pressure loss across a helical than comparable planar bend at a Reynolds number of 107.  I founded in 2004 an Imperial College engineering spin-out company, of which I am research director, to explore these areas.  The work has included bench-type and medium-scale model experiments, as well as numerical studies.  We have shown that SMAHT tubes are effective static mixers, the mixing occurring without the need for any internal structure and with lower than conventional pressure losses.  In multiphase work, we have successfully separated gas bubbles from liquid with SMAHT tubes.  We have found moreover substantially less slugging with flow of gas/water in a SMAHT than conventional riser, without the need for flow reduction [231] - a finding potentially of strong interest to the petrochemical industry.  In addition, consideration is being given to the use of SMAHT tubes in furnace coil design [230]. 

Interdisciplinary Activity

My interest in mechanics dates from my student days and a fellowship at the University of Pennsylvania.  In the mid 1960s, I became interested in arterial disease and flow and met Professor (later Sir James) Lighthill.  We undertook a collaborative study on indicator dispersion in models of the circulation [20] and, shortly thereafter, Lighthill, then Royal Society Research Professor at Imperial College, persuaded Imperial College to set up as an experiment a physiological flow studies unit, with me as director.  The PFSU was at its inception almost unique worldwide, with workers coming from physiology, medicine, applied mathematics and engineering.  The Unit can I believe claim to have made a prodigious contribution to bioengineering and interdisciplinary research, nationally and internationally.  Academic visitors came from many parts of the world, including the USA, continental Europe, Japan, Australia and India, and frequently formed similar organizations on their return.  Many also came to occupy leading academic positions (including in mathematics, biomedical science, medicine and engineering), with several being elected to international scientific academies, and founding, or holding high office in, world federations of bioengineering.  The writer initiated, in the 1960s, the still ongoing physiological series of Euromech Colloquia.  He became moreover an adviser in countries including the US, Japan, India and France, and was a founder council member of the World Council of Biomechanics and later International Chairman of the World Congress of Biomechanics.  The PFSU became in 1989 the Imperial College Centre for Biological and Medical Systems, with the writer as first director.  It is now the Department of Bioengineering, a full and rapidly growing undergraduate and postgraduate department of Imperial College.