Shunt Vessels Research Articles

USE OF TRANSCOLONIC 123I‐IODOAMPHETAMINE TO DIAGNOSE SPONTANEOUS PORTOSYSTEMIC SHUNTS IN 18 DOGS

Transcolonic 123I‐Iodoamphetamine is rapidly absorbed across the colonic mucosa and binds to amine receptors in the liver and lungs. During the first ten minutes following colonic administration, a simple ratio of lung counts to lung counts plus liver counts provides an accurate estimate of the fraction of portal blood that bypasses hepatic sinusoids in dogs with portosystemic shunts. Studies were performed on 24 dogs with suspect portosystemic shunt. Shunt fraction values for 18 dogs with surgically confirmed portosystemic shunt were obviously higher than published values for normal dogs, and also significantly higher than values for the other six dogs, later confirmed to lack shunts. Postoperative studies were repeated on ten dogs with single shunt vessels 1–2 days after shunt closure. Total shunt ligation resulted in normal postoperative shunt fraction, whereas partial shunt ligation resulted in persistant elevation of shunt fraction. Transcolonic iodoamphetamine scintigraphy is noninvasive, easy to perform, and provides an accurate method to diagnose dogs with portosystemic shunt.

Arteriovenous malformation of the spinal cord

We report a case of spinal arteriovenous malformation (AVM) in which detailed morphological examination of the intramedullary vascular lesion was carried out employing serial section studies. The relationship of the malformation to parenchymal lesions was evaluated. An abnormal vessel at the spinomedullary transition, whose lamina elastica was partially interrupted, was suggestive of a shunt vessel. The location of the intramedullary abnormal vessels and foci of softening of the spinal cord almost overlapped in distribution. Three-dimensional reconstruction of the intramedullary abnormal vessels demonstrated marked elongation and tortuosity of the original vascular architecture and they ended in the extramedullary abnormal vessels with thickened wall. Thus, our study showed that the intramedullary abnormal vessels were the result of secondary changes subsequent to pressure and volume overload inflicted upon the spinal venous system through AVM. Compression by tortuous elongation of the intramedullary vessels is the apparent cause of the parenchymal softening. A traumatic factor in the aetiology of the AVM was also discussed, since the patient had had two preceding episodes of traffic accidents with cranial and lumbar injury. Trauma seemed to be significant factor as a trigger of symptoms.

Hemodynamic effects of prostacyclin infusions in healthy volunteers.

The selective hemodynamic effects of an intravenous infusion of prostacyclin on skin and on forearm blood flow during increasing doses up to 6 ng.kg.-1 min.-1 were evaluated in 12 healthy volunteers. The highest dose raised total skin blood flow (p = 0.0001) as measured by laser Doppler with a concomitant rise in finger skin temperature (p = 0.02). Capillary perfusion as measured by transcutaneous oxygen monitoring remained relatively unaffected (p = 0.25). Cold-induced vasoconstriction and the recovery thereafter were not influenced. Diastolic blood pressure tended to decrease (p = 0.06), but there was no change in systolic blood pressure, forearm blood flow, and calculated forearm vascular resistance. The authors conclude that prostacyclin may have dissimilar vasoactive potencies in different vascular beds as shown by a pronounced increase in total skin blood flow, owing to vasodilation of shunt vessels, simultaneously with an insignificant change in nutritional skin blood flow and forearm muscle blood flow.

Analysis of gas exchange in the ophthalmic rete of pigeons.

Many birds have ophthalmic retia (OR), paired vascular networks in the head, that cool arterial blood flowing toward the brain. They thus independently regulate brain and core body temperatures (Bernstein et al., 1979; Kilgore et al., 1979). The system works because evaporation from nasal and oral membranes cools venous blood before it flows through the OR. This blood then takes up heat from warmer arterial blood flowing countercurrent to it (Midtgard 1983; Clair, 1985). The vascular and neural anatomy of the OR suggests that the sympathetic nervous system controls both total blood flow and the proportioning of flow between heat-exchange and shunt vessels (Midtgard, 1985). Birds routinely tolerate body temperatures of 45°C, indicating an extreme capacity for heat storage. This and the ability to cool the brain serves during heat exposure and exercise by reducing the need for evaporative cooling.

Optic nerve sheath meningioma: diagnostic features and therapeutic alternatives.

Twenty-two cases of optic nerve sheath meningioma were reviewed. The clinical features included slowly progressive visual loss in every case. Orbital signs of limited ocular movements and mild proptosis were present in only a third of cases; gaze-evoked amaurosis occurred in 3 cases. The disc was abnormal in every case, usually swollen if vision was 6/12 or better, atrophic if the vision worse than 6/12. Optociliary shunt vessels occurred in 5 patients and were of diagnostic significance. The diagnosis was made by a high-resolution CT scan of the orbits showing tubular expansion of the optic nerve sheath. Calcification of the optic nerve was present in 12 cases. Tubular expansion of the optic nerve sheath may occur in raised intracranial pressure, optic nerve glioma, granuloma, lymphoma or metastatic disease. In the absence of calcification these alternative diagnoses must be considered. Surgery, undertaken for the intracranial component of the tumour, did not halt progressive visual loss.

Gill microcirculation of the air-breathing climbing perch, Anabas testudineus (Bloch):relationships with the accessory respiratory organs and systemic circulation.

The general macrocirculation and branchial microcirculation of the air-breathing climbing perch, Anabas testudineus, was examined by light and scanning electron microscopy of vascular corrosion replicas. The ventral aorta arises from the heart as a short vessel that immediately bifurcates into a dorsal and a ventral branch. The ventral branch distributes blood to gill arches 1 and 2, the dorsal branch to arches 3 and 4. The vascular organization of arches 1 and 2 is similar to that described for aquatic breathing teleosts. The respiratory lamellae are well developed but lack a continuous inner marginal channel. The filaments contain an extensive nutritive and interlamellar network; the latter traverses the filament between, but in register with, the inner lamellar margins. Numerous small, tortuous vessels arise from the efferent filamental and branchial arteries and anastomose with each other to form the nutrient supply for the filament, adductor muscles, and arch supportive tissues. The efferent branchial arteries of arches 1 and 2 supply the accessory air-breathing organs. Arches 3 and 4 are modified to serve primarily as large-bore shunts between the dorsal branch of the ventral aorta and the dorsal aorta. In many filaments from arches 3 and 4, the respiratory lamellae are condensed and have only 1-3 large channels. In some instances in arch 4, shunt vessels arise from the afferent branchial artery and connect directly with the efferent filamental artery. The filamental nutrient and interlamellar systems are poorly developed or absent. The respiratory and systemic pathways in Anabas are arranged in parallel. Blood flows from the ventral branch of the ventral aorta, through gill arches 1 and 2, into the accessory respiratory organs, and then returns to the heart. Blood, after entering the dorsal branch of the ventral aorta, passes through gill arches 3 and 4 and proceeds to the systemic circulation. This arrangement optimizes oxygen delivery to the tissues and minimizes intravascular pressure in the branchial and air-breathing organs. The efficiency of this system is limited by the mixing of respiratory and systemic venous blood at the heart.

Adrenergic innervation of the gills, pulmonary arterial plexus, and dorsal aorta in the neotenic salamander, Ambystoma tigrinum.

The presence of adrenergic innervation was investigated in four different vascular segments of the neotenic tiger salamander, Ambystoma tigrinum, by histofluorescent staining for catecholamines. The segments were the respiratory section of the gill, the branchial shunt vessels, a vascular plexus in the pulmonary artery, and the dorsal aorta. No adrenergic fibers were detected in the respiratory section of the gill or the pulmonary arterial plexus. In contrast, the branchial shunt vessels contained both adrenergic varicosities and catecholamine-containing cell bodies. These cells resemble Type I cells of the mammalian carotid body and amphibian carotid labyrinth. Adrenergic innervation of the dorsal aorta was sparse and restricted to the adventitia. The results suggest that adrenergic nerves may directly regulate blood flow in the gill, and thus gas exchange, by controlling vascular resistance of the branchial shunts. The contractile state of the dorsal aorta may also be under adrenergic control. In addition, it is suggested that the adrenergic cells of the branchial shunts may serve a receptor function in being sensitive to arterial blood gases.

Vascular resistance and vasoactivity of gills and pulmonary artery of the salamander,Ambystoma tigrinum

In order to understand the blood flow patterns and their regulation in the gills and pulmonary artery ofAmbystoma tigrinum, the vascular resistance and vasoactivity of the two major branchial perfusion pathways and a vascular plexus in the pulmonary artery were investigated using an isolated-tissue perfusion method. Acetylcholine and epinephrine were both pressor agents in all three vascular segments. Angiotensin II also constricted the branchial respiratory vasculature. Norephinephrine was primarily a vasodilator in the branchial respiratory vasculature, however, it had no effect on the shunt vessels of the gill or the pulmonary arterial plexus. Both gill circulations were insensitive to alterations in CO2 and pH. Anoxia produced a slight vasodilation of the branchial respiratory vessels but had no effect on the shunt vasculature. Mild hypoxia had no effect on either branchial circulations. The results suggest that: (1) blood flow through the respiratory section of the gill may vary between 8 and 47% of total gill flow, (2) the major perfusion pathway to the lung is probably from the efferent artery of the third gill through the ductus arteriosus and then into the pulmonary artery, (3) O2, CO2 and pH exert no local control of branchial perfusion, (4) both cholinergic and adrenergic regulation of branchial and proximal pulmonary arterial vascular resistance is possible, (5) a rise in circulating norepinephrine should increase blood flow to the respiratory section of the gill.

Haemodynamics of pelvic nerve induced penile erection in the dog: possible mediation by vasoactive intestinal polypeptide.

The haemodynamics of erection were elucidated in the anaesthetized dog by analysing in quantitative terms the changes of penile arterial inflow, venous outflow and tissue volume during graded pelvic nerve stimulation. The study also provides information on possible neurotransmitter mechanisms of the erectile response. Erection evoked by pelvic nerve stimulation appeared to result from two main circulatory events: first, there was a prompt dilatation of the penile 'resistance vessels', causing a greatly increased arterial inflow which in the early phase bypassed the cavernous bodies and, hence, increased venous outflow to the same extent. Secondly, the erectile response proper began after a distinct delay (approximately equal to 20 s). This was apparently caused by sudden opening of low resistance 'shunt vessels' diverting part of the arterial inflow into the cavernous bodies, leading to rapid filling. During the filling phase arterial inflow greatly exceeded venous outflow, and returned to the venous outflow level again in the steady state of full erection. The initial dilator response seemed to ensure rapid erection by establishing a high pressure head from the arterial microvessels to the cavernous spaces. The threshold frequency for the penile vasodilator response to pelvic nerve stimulation was 1-2 Hz and was always higher for the erectile volume response, viz. 2-4 Hz. Maximal effects for both were obtained at 16 Hz, causing on the average a 25-fold increase in peak arterial inflow, a 17-fold increase in venous outflow and a 107% increase in penile volume. Muscarinic blockade by atropine caused no significant decrease in the blood flow response induced by pelvic nerve stimulation, but clearly curtailed the erectile response. This indicates that the dilatation of the penile 'resistance vessels' is mainly non-cholinergic in nature, whereas a cholinergic mechanism seems to contribute to the erectile volume response proper. Pelvic nerve stimulation caused a substantial output of vasoactive intestinal polypeptide (VIP) from the penis which was correlated in onset and duration to the vasodilator response. Intra-arterial (I.A.) infusion of VIP elicited moderate erection and a penile vasodilator response which resembled the neural response. Similar effects were evoked by I.A. infusion of substance P, but the output of this peptide from the penis during stimulation was poorly correlated to the vascular events. These in vivo observations indicate that VIP might be the neurotransmitter responsible for the non-cholinergic pelvic nerve induced penile vasodilatation.

Ventilation-perfusion relationships during halothane anaesthesia and mechanical ventilation. Effects of varying inspired oxygen concentration.

Ventilation-perfusion relationships (VA/Q), assessed by a multiple inert-gas elimination technique, were studied during halothane anaesthesia and mechanical ventilation at different inspiratory oxygen fractions (FIO2). All nine patients (mean age 65 years, five smokers) displayed unaltered VA/Q distributions with increasing FIO2 from a mean of 29% to 53%. A further increase in FIO2 to a mean of 85% caused an increase in true shunt (VA/Q = 0) from 7 to 10% of cardiac output (P less than 0.01), but no increase in "low" VA/Q (VA/Q less than 0.1). On the return to FIO2 of 29%, true shunt was reduced to the initial level. The findings may fit in with release of hypoxic vasoconstriction when FIO2 is increased to 85%, or the opening up of a certain population of shunt vessels.

Combined Clinical and Computed Tomographic Diagnosis of Orbital Glioma end Meningioma

The clinical information on 22 patients with orbital optic nerve gliomas and 47 patients with meningiomas was correlated with computed tomographic findings obtained in both axial and coronal studies. Most of the gliomas occurred in children, although 7 patients presented after 20 years of age. Among the patients with meningiomas, the majority were women in early middle age, although two tumors occurred in children less than 20 years of age. Low grades of proptosis (median, 2 mm for both tumors), frequent significant visual field obscurations with eye movements, and opto-ciliary shunt vessels pointed toward the diagnosis of an optic nerve tumor. Patients with gliomas generally manifested massively swollen fusiform optic nerves with clear-cut margins due to circumscription by an intact dura. Kinks and bucklings of the optic nerve as well as infarctive cysts distinguished the glioma CT-scan patterns from the meningiomas. Distinctive axial CT-scan features of the meningiomas not shared by the gliomas were narrowly and diffusely enlarged nerves with polar expansions either at the orbital apex or immediately behind the globe; calcification; irregular excrescent margins signifying extradural invasion into the orbital soft tissues; a negative optic nerve shadow running down the center of the lesion; and bone erosion near the orbital apex. Coronal studies often revealed irregular margins signifying transgression of the dura. A diffusely and narrowly enlarged optic nerve shadow with regular margins (intrasheath lesions) was the one morphologically overlapping pattern displayed by 11 meningiomas and three gliomas. In these cases there tended to be more profound visual loss in the gliomas compared with the meningiomas, as well as the more frequent presence of opto-ciliary vessels in the meningiomas. Arteriography may be helpful in this particular category by demonstrating a tumor blush for the meningiomas, whereas this finding is typically absent with optic nerve gliomas. Meningiomas may be very closely simulated by dural or intraneural inflammations.

Optic Nerves Sheath Meningiomas: Clinical Manifestations

A retrospective clinical study of optic nerve sheath meningiomas based on 22 patients showed that symptoms most commonly develop in women between the ages of 35 and 60 years. The most common presenting symptoms were decreased vision and transient visual obscurations. In the earliest stages, many patients presented with normal to mildly impaired acuity (despite subjectively decreased vision), optic disc edema and enlargement of the blind spot. Optic disc edema was frequently associated with refractile bodies indicative of chronic swelling. Optic disc edema preceded the development of optic atrophy. Another group of patients presented with a history of longstanding vision loss, visual acuity of 20/200 or worse and optic atrophy. Optociliary shunt vessels were late findings only seen in five patients. The most consistent visual field abnormality was peripheral constriction. Cecocentral scotomas were uncommon. Intracranial involvement was present in five patients. There were two patients with bilateral optic nerve sheath meningiomas without CT evidence of intracranial involvement. Computerized tomography was found to be indispensible in the diagnosis of optic nerve sheath meningiomas and the detection of intracranial involvement.

Blood flow in the ovary and adjacent structures of the non-pregnant sheep.

Blood flow of the ovary, vascular pedicle and oviduct was measured in anaesthetized non-cycling and cycling ewes by timed collection of ovarian venous blood. The degree of arterio-venous shunting across the ovary and pedicle was estimated both in vivo and in vitro by perfusing the tissues with 15 +/- 5 micrometers radioactive microspheres. The mean ovarian blood flow in non-cycling animals was 1.9 ml/min, which was 51% of blood flow in the ovarian vein. In cycling animals ovarian blood flow at midcycle was 2.9 ml/min (66% of ovarian venous flow) in non-luteal ovaries and 4.3 ml/min (79% of venous flow) in luteal ovaries. The degree of arterio-venous shunting was low in all stages of the cycle (1.0-2.6% across ovary + pedicle). The degree of shunting was also found to be very small in vitro (0.007-1.38%) in both non-luteal and luteal ovaries. A considerable number of microspheres was entrapped in the vascular pedicle of the ovary indicating the presence of an extensive capillary bed. There was an inverse relationship between blood flow in the ovary and flow in the vascular pedicle. Alterations in distribution of blood flow between the ovary and adjacent structures supplied by the ovarian artery may be of functional significance in allowing rapid changes in ovarian blood flow. The results of the present study indicate that changes in ovarian blood flow during the oestrus cycle are not caused by an action on arterio-venous shunt vessels.

The ultrastructure and connections of blood vessels supplying the rat carotid body and carotid sinus.

The morphology of blood vessels supplying the carotid body and carotid sinus was analysed in 41 rats by using a combination of light microscopic, transmission electron microscopic and scanning electron microscopic methods. We found that a large sphincter-like intimai cushion was located at the orifice of thecarotid body artery, where the vessel arose from the external carotid or occipital artery. The sphincter contained circumferential smooth muscle and constricted the diameter of the orifice to less than half. After reaching the carotid body, the carotid body artery typically divided into three or four first-order and five or more second-order branches. Usually three or foursecond-order branches supplied the carotid body, but all other branches continued on to such structures as the superior cervical ganglion, nodose ganglion, vagus nerve and carotid sinus.Third andfourth-order branches gave rise toterminal arterioles that supplied the glomus tissue. Vessels resemblingprecapillary sphincters were located at the junction of terminal arterioles and capillaries. Precapillary sphincters had a wall comprised of protruding endothelial cells surrounded by smooth muscle cells or pericytes. Most terminal arterioles gave rise to two types of capillaries.Type I capillaries penetrated a glomus cell cluster and had an intimate association with glomus cells of that cluster had a luminal diameter ranging from about 8 to over 20 μm, but varied in calibre along their length. These vessels followed a winding course, made one or more U-shaped turns, and usually had multiple connections with venules. Type I capillaries had a thin fenestrated endothelium, an incomplete covering of pericytes, and a thin basal lamina. By contrast,type II capillaries did not penetrate glomus cell clusters, had a uniform diameter of about 7 μm, and had both straight and curved regions. Both types of capillaries were bypassed byarteriovenous anastomoses formed by terminal arterioles that joined small venules directly.Venules of the carotid body were interconnected with one another and joined major veins of the neck via several routes. Arterioles derived from the carotid body artery also supplied an extensive network of vasa vasorum in the adventitia of the carotid sinus. Short capillaries and larger shunt vessels connected arterioles with the numerous venules in the sinus wall; and the venules in turn were connected to the venous plexus at the surface of the carotid body. We conclude that the arterial branching pattern, intimai cushions and precapillary sphincters participate in the control of carotid body blood flow and also may influence plasma skimming. However, the existence of arteriovenous anastomoses in addition to at least two types of capillaries indicates that blood flow to the chemoreceptive tissue can be regulated independently of total blood flow. Furthermore, the redundancy of venous connections may be related to the sensitivity of carotid body chemoreceptors to changes in venous pressure.

Laser-Doppler measurement of digital blood flow regulation in normals and in patients with Raynaud´s phenomenon

Finger blood flow was measured continuously by a laser-Doppler flowmeter in normal subjects and in patients suffering from primary or secondary Raynaud´s phenomenon. In normals, skin areas with and without arteriovenous anastomoses could be differentiated. Blood flow in areas with shunt vessels decreased after a deep breath but not during venous stasis or the Valsalva maneuvre, while in other skin areas a decrease in blood flow was observed after all three maneuvres, suggesting a dual innervation of nutritional and shunt vessels. In patients with Raynaud´s phenomenon (scleroderma) fingertip blood flow reacted in the same way as normal skin without shunt vessels. During direct and indirect cooling, finger blood flow in patients with secondary Raynaud´s phenomenon reacted with the same relative reduction as normals but resting blood flow in the patients was significantly decreased and the rewarming period was greatly prolonged. In patients with primary Raynaud´s phenomenon an even more prolonged decrease in blood flow was observed after direct or indirect cooling. Defective function of arteriovenous anastomoses is proposed as an explanation of the deviations from normal. A pathophysiological classification of Raynaud´s phenomenon may be possible on the basis of the function tests described.

Evaluation of Cutaneous Blood Flow Responses by 133Xenon Washout and a Laser-Doppler Flowmeter

A new method for noninvasive measurement of cutaneous blood flow is laser-Doppler flowmetry. The technique is based on the fact that laser light is back-scattered from the moving red blood cells, with Doppler-shifted frequencies; these impulses lead to photodetectors and are converted to flow signals. In this work we used a new system with a low noise level. Comparison was made between this technique and the atraumatic epicutaneous 133Xenon technique for measurement of cutaneous blood flow during reactive hyperemia and orthostatic pressure changes. The laser-Doppler flowmeter seems to measure blood flow in capillaries as well as in arteriovenous anastomoses, while the 133Xe method probably measures only capillary flow. A calibration of the laser-Doppler method against the 133Xe method would appear to be impossible in skin areas where arteriovenous anastomoses are present. The changes in blood flow during reactive hyperemia, orthostatic pressure changes, and venous stasis were found to be parallel as measured by the two methods in skin areas without shunt vessels. The laser-Doppler flowmeter would appear to be a useful supplement to the 133Xe washout method in cutaneous vascular physiology, but it is important to keep in mind that different parameters may be measured.

Erection and priapism: a new physiopathological concept.

For normal erection two mechanisms are essential: the first provides increased arterial inflow, most probably this is obtained due to the activity of the intimal cushions within the arteriae helicinae and within the shunt vessels. The second mechanism uses increased arterial blood flow; this could be obtained due to the activity of the trabecular muscle fibers of the corpora cavernosa. In consequence, we would deal with two different types of priapism: one type--high-flow priapism--occurs at the level of the arteriae helicinae and the intimal cushions and provokes a high-flow situation, and the second type, with blood stasis in the corpora cavernosa, occurs at the level of the trabeculae due to persistent contraction of the smooth muscle fibers. The prognosis of the second type is much less favorable and should be cured by surgery within the first 48 h.

Innervation sites of dog hindpaw series and parallel vascular circuits.

The neural control of the distribution of blood flow between the parallel nutritional and nonnutritional vascular circuits in the dog hindpaw has been assessed. The right hindpaw of anesthetized dogs was vascularly and neurally isolated and placed in a volume recorder sealed by the skin flap. The animals were heparinized. The autoperfused preparation was isovolumetric during the control period. The inflow tubing had side arms for measurement of perfusion pressure and for injections of indicators (131I-albumin, 86Rb and 85Sr-microspheres). Vascular volume changes were determined from venous time-concentration curves and by plethysmography. The peripheral cut ends of the superficial and deep fibular nerves and the tibial nerve were individually stimulated at 5 and 11 Hz. The permeability surface area product of 86Rb and the capillary filtration coefficient were determined and A-V shunt patency was assessed by the venous recovery of 85Sr 15 microns microspheres. Each nerve stimulation increased blood flow resistance. Superficial fibular nerve innervation included essentially uniform effects on the arterioles and A-V shunts, the upstream arteries and also the veins. Deep fibular nerve stimulation indicated that its innervation included the arteries and the A-V shunt nonexchange vessels. Tibial nerve stimulations indicated that its innervation included the A-V shunt vessels and veins with minimal influences on the arteries and arterioles.

Aortic regurgitation in tetrad of Fallot and pulmonary atresia

Data on 18 patients, aged 12 to 42 years, with documented aortic regurgitation and tetrad of Fallot or pulmonary atresia with ventricular septal defect were reviewed. In two patients, aortic regurgitation was caused by surgical repair but in the others it was present before operation. There was increased volume overload from a long-standing surgical shunt (8 to 30 years' duration) or congenital systemic collateral vessels in 14. Six patients with a history of infective endocarditis had aortic cusp perforations. Failure to detect the presence of aortic regurgitation before radical repair in six patients contributed to operative problems and postoperative morbidity and mortality. Aortic valve surgery was performed in 13 patients.Aortic regurgitation is an acquired complication that should be specifically excluded by routine retrograde ascending aortography in all adolescents or adults with tetrad of Fallot or pulmonary atresia with ventricular septal defect. Earlier radical repair in the first decade of life may prevent the complication.

Air embolism: possible role of surfactant on recompression.

The relationship between surface tension and surface area has been measured on each of three common pulmonary surfactants - dipalmitoyl lecithin (DPL), dipalmitoyl phosphatidylethanolamine, and sphyngomyelin-under simultaneously simulated physiological conditions. These are selected to simulate the state of any surfactant that has migrated onto the surface of venous bubbles filtered by the pulmonary vasculature. It is concluded that, in the absence of shunt vessels, only DPL could reduce surface tension enough to allow pulmonary gas emboli to escape into arterial blood and then only after compression. This finding is discussed in relation to the delay in any appearance of bubbles in arterial blood and the possible facilitation of the release of asymptomatic lung bubbles by recompression therapy. The suggestion is made to reconsider stopping recompression of a subject with peripheral decompression sickness (the bends) at the depth of relief rather than risk releasing pulmonary gas emboli by further recompression. It is also demonstrated how the introduction of 1-min stops into compression can avoid surface tension falling to the low values at which it is theoretically possible for venous bubbles to escape into arterial blood during pulmonary hypertension.