In a classical experiment in 1934 Goldblatt and his colleagues showed that a clip partially occluding one renal artery of a dog caused hypertension. It has since been demonstrated that reduced renal perfusion stimulates renin release from the affected kidney. This in turn increases the conversion of angiotensinogen to angiotensin I. Angiotensin converting enzyme converts angiotensin I to angiotensin II and this sequence stimulates the release of aldosterone from the adrenal cortex (Fig. 1) 329.

In fact it is probably the vasoconstrictor action of angiotensin II which is responsible for the initial stages of hypertension. Salt and water retained under the influence of aldosterone are excreted by the other kidney. However, if there is only one kidney, or if the other kidney has been damaged by prolonged hypertension, the body is no longer able to excrete excess salt and water; hypertension is then due to increased intravascular volume. Thus, in the early stages of this disease the affected kidney secretes high levels of renin and renin production from the other kidney is suppressed. Later in the disease salt and water retention suppress renin production in both kidneys.

In humans 70 per cent of renal artery stenoses are caused by atherosclerosis. Patients with this disease are usually elderly, often male, and frequently have other manifestations such as ischaemic heart disease and peripheral vascular disease.

The other major cause of renal artery stenosis, probably accounting for 20 per cent of cases, is fibromuscular dysplasia. There are three different forms of the disease, each affecting a particular layer of the arterial wall. The most common is medial dysplasia, characterized by hyperplasia of either the fibrous tissue or smooth muscle of the media (Fig. 2) 330. Up to 70 per cent of cases of medial dysplasia are caused by medial fibroplasia, in which thickened fibromuscular ridges alternate with attenuated media so that the lumen is alternately stenosed and aneurysmal. This produces the so-called ‘string of beads’ sign on angiography (Fig. 3) 331. Fibromuscular hyperplasia usually affects the distal two-thirds of the main renal artery and sometimes the primary branches. The entity occurs most often in young women. It commonly affects both renal arteries and occasionally the carotid and iliac arteries.

Other less common causes of renal artery stenosis include trauma, renal artery aneurysm (compressing the adjacent otherwise normal artery), arteriovenous fistulae (diverting blood from the kidney), neurofibromatosis, Takayasu's disease (an arteritis causing stenosis of arteries arising from the aorta and found mainly in young adults and children of Asian origin), and hypoplasia.

Stenosis of an artery reduces blood flow appreciably only when the luminal diameter is reduced by 70 per cent. Even demonstration of such a stenosis of the renal artery on duplex scanning or arteriography in a hypertensive patient is not sufficient to warrant a diagnosis of renovascular hypertension. The renal artery disease may be an incidental manifestation of atherosclerosis in a patient with essential hypertension. Many nephrectomies and renal artery reconstructions were performed in patients with hypertension and renal artery stenosis before it was realized that fewer than half benefited from the procedure. The ultimate proof of renovascular hypertension is its cure by correction of the stenosis, but several tests have been developed that allow patients more likely to respond to such intervention to be selected.

In the general hypertensive population the prevalence of renovascular hypertension is only 1 to 5 per cent, but it is one of the few curable forms of hypertension. There are particular groups of individuals who are more likely to have renovascular hypertension and these need to be identified for further investigation. In addition, young adults with hypertension, although not necessarily more likely to have a renovascular cause, have a lot to gain by avoiding lifelong treatment if renovascular hypertension is discovered and corrected.

Patients with atherosclerosis often have renal artery stenosis. In patients unselected for hypertension who are undergoing angiography for peripheral vascular disease the prevalence of renal artery stenosis is about 30 per cent; 12 per cent of these have bilateral disease. In hypertensive patients with evidence of peripheral vascular disease the prevalence of renovascular hypertension, as opposed to merely renal artery stenosis, is about 14 per cent.

Patients with severe hypertension are more likely to have renovascular hypertension: the prevalence is 30 per cent in those with a diastolic pressure over 125 mmHg and up to 45 per cent in patients with accelerated hypertension and renal failure.

Age is another important factor. Thirty per cent of patients who develop a diastolic pressure greater than 105 mmHg after the age of 50 years have renovascular hypertension, and it is probably also more common in the elderly with renal failure. Over 70 per cent of older patients with a unilateral atrophic kidney have renal artery stenosis, and nearly 80 per cent of these have significant contralateral stenosis. Many have incipient or overt renal failure as a result, not necessarily accompanied by marked hypertension. At the other end of the scale, renal artery stenosis ranks with coarctation of the aorta as a major cause of hypertension in children.

The presence of a bruit, audible in the upper abdomen or flank, increases the likelihood of renal artery stenosis in any patient.

Repeated angiograms taken over 4 to 5 years show increasing stenosis in about 50 per cent of patients with atherosclerotic disease but in only 20 per cent of those with fibromuscular disease. Parallel changes in renal size and serum creatinine can also be detected. When the diameter of the renal artery is reduced by 75 per cent, between 12 and 40 per cent of those with atheromatous stenoses progress to complete occlusion within 1 year. Even at this stage the situation may not be irremedial. Sufficient renal tissue may be maintained by collateral vessels around the capsule and hilum to enable restoration of function if renal artery flow is reinstated (Fig. 4) 332. One-third of patients with atheromatous disease present with bilateral renal artery stenosis; those with initially unilateral disease have a 40 per cent chance of contralateral renal artery stenosis developing within the next 4 years. Fibromuscular disease is even more likely to affect both renal arteries (60 per cent at presentation) although it is less likely to progress.

While one kidney is functioning normally, unaffected by either stenosis of its artery or hypertensive parenchymal damage, adequate excretion is maintained. However, if that kidney develops functional impairment the patient begins to slip into renal failure. Any kidney affected by renal artery stenosis tends to shrink, with loss of nephrons. This shrinkage can often be reversed by restoring blood supply, but once the kidney measures less than 8 cm in length function is unlikely to return. Thus intervention can be of benefit to patients with renal failure, provided it occurs before critical renal mass is lost. There is no correlation between renal function and blood pressure, either before or after treatment.

Acute pulmonary oedema occurs in up to 25 per cent of patients who have renal failure due to renal artery stenosis but it is reversed by revascularization of the kidney. Renovascular disease must be excluded in any patients presenting with pulmonary oedema.

Two initial requisites for the diagnosis of renovascular hypertension are hypertension and renal artery stenosis. Confirmation of the latter used to be left to the later stages of investigation because it required arteriography with its attendant risks. Two recent developments have changed that. Duplex ultrasonography is a non-invasive means of detecting renal artery stenosis both by visualization of the vessel and by measurement of the effect of stenosis on blood flow velocity and waveforms. In specialized centres prepared to devote time to each examination this test is both specific and sensitive. However many examinations are difficult to perform and interpret because of obesity or bowel gas: up to 50 per cent of examinations may have to be abandoned because of technical difficulties. The test also relies heavily on operator expertise. It is therefore unlikely to be useful as a screening test in its present form.

The other major advance is intravenous digital subtraction angiography. This technique requires injection of contrast medium into a vein rather than an artery, which considerably reduces associated morbidity. It relies on computer techniques to demonstrate the arteries when relatively low concentrations of contrast are present. There are several drawbacks to the method. A large amount of contrast has to be injected, which may further compromise renal function in patients with incipient or overt renal failure. Visualization of the vessels depends on cardiac output: output may be impaired in patients with ischaemic heart disease, and even in those with good cardiac output visualization of the renal arteries is often inadequate. Despite these reservations intravenous digital subtraction angiography is becoming a useful screening tool for renal artery stenosis in patients who have a high probability of renovascular hypertension on clinical grounds.

Once the diagnosis of renal artery stenosis in association with hypertension has been established, further investigations are required. The peripheral blood renin level is often elevated, but this is not always the case. Captopril, an angiotensin converting enzyme inhibitor, increases plasma renin levels in patients with renovascular hypertension and can be used to increase the accuracy of the test.

A more invasive test is measurement of renal vein renin by selective catheterization of each vessel via a femoral vein puncture. If unilateral disease is present the affected kidney should secrete high levels of renin and the contralateral kidney should have suppressed renin production. A ratio between the two kidneys of more than 1.5 has long been used as a highly specific (80–100 per cent) test for renovascular hypertension. Unfortunately the test lacks sensitivity: false negative rates of 20 to 50 per cent have been reported. Expressing the higher renin level as an increment relative to the inferior vena cava renin level is said to increase the sensitivity to 74 per cent. If the renal vein renin ratio is measured before and after administration of captopril, sensitivity of the test is increased, but at the expense of specificity.

Renin levels are influenced by fluid balance and by antihypertensive treatment. Ideally these tests should be performed when the patient has been off treatment for 2 to 3 weeks, with controlled sodium intake, and after 12 h of rest but it is often difficult to comply with such restrictions. Impaired renal function may also affect the test.

Radionuclide renal scanning is a less invasive means of assessing renal function. Labelled hippuran can be used to assess renal blood flow. However this alone offers no more than other measures of renal artery stenosis. A reduction in renal blood flow following exercise is said to predict a poor response to renal revascularization.

Labelled diethyltetraphenyl aminic acid is used more widely to provide a measure of glomerular filtration rate: in renal artery stenosis, uptake and excretion are delayed. Oral captopril administration produces a further fall in glomerular filtration rate in patients with renovascular hypertension, and makes this a relatively reliable test. Unfortunately most of these tests are only helpful in unilateral disease.

The ‘gold standard’ by which all these tests are assessed is the response of blood pressure to restoration of blood supply. There has been a suggestion that balloon dilatation of the renal artery at the time of intra-arterial angiography should be used as both test and treatment, in view of its low morbidity and mortality reported from certain centres. This approach is not in general use.

Intra-arterial aortography is necessary prior to intervention to demonstrate the lesion accurately. Intravenous studies do not provide adequate resolution to plan treatment. A flush aortogram needs to be performed first so that any accessory renal arteries, which may be diseased, are detected and the origins of all the renal arteries adequately shown.

The advent of powerful antihypertensive medications, such as &bgr;-blockers and, more recently, calcium channel blockers and angiotensin converting enzyme inhibitors, have brought blood pressure under satisfactory control in most patients. In patients with atheromatous disease the risks of any further intervention are increased because of their associated disease and age. If there is no evidence of renal failure and the blood pressure can be satisfactorily controlled by medication there is probably no justification for more aggressive intervention. However, a reduction in systemic pressure may reduce renal perfusion still further and lead to progressive renal failure. Renal function should be regularly monitored in these patients so that further intervention can be instituted before critical loss of renal substance. Angiotensin converting enzyme inhibitors can cause a reduction in glomerular filtration rate by removing the selective vasoconstrictive action of angiotensin II on the efferent arterioles and so precipitate renal failure. This is particularly likely to happen when there is a critical stenosis in a solitary kidney or bilateral disease.

Patients with fibromuscular hyperplasia are often diagnosed before or during middle age. The risks of intervention are lower and the outcome usually better than that in atherosclerotic patients. In addition, the advantage to the patient in terms of years free of treatment, or on reduced medication, are such that an attempt is usually made to restore blood supply.

Almost any child or young adult with hypertension due to renal artery stenosis should have the stenosis corrected; the results are good whatever the cause.

Percutaneous transluminal angioplasty of the renal artery is performed under angiographic screening. A guidewire is fed up to the renal artery from a femoral artery puncture site and then passed across the stenosis (or occasionally across an occlusion). Once in position a catheter bearing a balloon at its tip is slid over the guide wire. Inflation of the balloon disrupts plaque or stretches dysplastic vessel wall to remove the stenosis.

This technique works well in fibromuscular dysplasia: 50 per cent of patients are cured and maintain diastolic pressure below 90 mmHg without antihypertensive medication. Most of the remainder have improved blood pressure control, although they still require medication. Patients with atheromatous disease fare less well. The ‘cure’ rate is only about 20 per cent although a further 60 per cent are improved. Atheromatous disease of the renal artery is often an extension of aortic disease and limited to the renal artery ostium where it joins the aorta. In this situation satisfactory dilatation is usually unsuccessful because the balloon tends to bulge back into the aorta and only 25 per cent of patients have any long-term benefit.

Percutaneous transluminal angiography also has a place in the treatment of renal failure in this disease. Major centres are reporting a reduced or normal creatinine level in over 85 per cent of patients with renal failure due to either fibromuscular dysplasia or atherosclerosis.

Complications of percutaneous transluminal angiography probably arise in less than 10 per cent of cases, and range from a significant haematoma at the puncture site to renal artery thrombosis or dissection and segmental infarction of the kidney by distal embolization of fractured plaque or thrombus. Although facilities for emergency surgery used to be made available because of the potential risks of the procedure these risks are small; the chances of successful emergency surgical intervention are even smaller so this is probably no longer necessary. Restenosis occurs in fewer than 10 per cent of techically successful dilatations in patients with fibromuscular dysplasia but in at least 25 per cent of atheromatous lesions (15 per cent of non-ostial lesions). Mortality associated with the procedure ranges between 0 and 3 per cent but in most centres is less than 1 per cent.

Results vary slightly between different centres: those that have a more aggressive approach and use larger balloons tend to have better results without much increase in morbidity.

Expert anaesthesia for renal artery surgery is critical. Application and removal of an aortic clamp causes large fluid shifts and imposes significant cardiac stress. Adequate monitoring with good intravenous access is essential. Minimal requirements are intra-arterial and central venous lines for measuring pressure, and many would also advocate a Swan-Ganz catheter.

There are several different surgical approaches to revascularization of the kidney. Exposure of the renal arteries may be transperitoneal or retroperitoneal, although the latter method is less suitable for the right renal artery. A combined approach may also be used in which the peritoneal cavity is opened via a transperitoneal incision and then the spleen, pancreas, splenic flexure of the colon and descending colon are mobilized and swung to the right to expose the aorta, the renal and superior mesenteric arteries, and the coeliac axis (Fig. 5) 333.

Transaortic renal artery endarterectomy is often performed. The aorta is clamped and opened at the level of the renal arteries. A local endarterectomy of one or both vessels is performed, and can be extended to the aorta itself at this level if necessary (Fig. 6) 334. This technique is particularly designed for the common situation when the renal artery disease is an extension of aortic plaque, narrowing only the ostium of the renal artery. It cannot be used if the disease is more distal in the vessel.

The aorta is clamped, usually below the renal vessels to maintain flow in them as long as possible. A length of reversed saphenous vein is anastamosed to the aorta and then to the renal artery beyond the stenosis (Fig. 7) 335. Graft of a synthetic material, such as polytetrafluoroethylene, may be used instead of vein.

Patients with atheromatous disease sometimes have occlusive aortoiliac disease or an abdominal aortic aneurysm which also requires treatment. In such cases it may be justifiable to replace the lower abdominal aorta with a graft at the same time. Patients with such extensive disease often have bilateral renal artery stenosis. Grafts can be run off the aortic graft to one or both renal arteries (Fig. 8) 336.

Occasionally the aorta is so heavily calcified that aortorenal grafting would present serious technical problems. The iliac arteries may be relatively spared, especially on the anterior wall, and an iliorenal graft can be constructed using saphenous vein (Fig. 9) 337. Failing this, or if there is reluctance to clamp the aorta in a high-risk patient, the visceral circulation can be used to donate blood to the kidney. The splenic artery, provided it is relatively free of atheroma, can be transected and anastamosed to the left renal artery leaving the spleen in situ (Fig. 10) 338. The hepatic artery can be used to supply the right kidney. A vein graft is usually taken from the side of the hepatic artery just beyond the gastro-duodenal artery and run down to the right renal artery (Fig. 11) 339. Alternatively the gastroduodenal artery itself, the communication between the coelic axis and the superior mesenteric circulations, can be divided distally and anastamosed to the right renal artery if it is long enough. Reconstructions based on the visceral arteries should only be performed if significant coeliac axis disease (and in the case of gastroduodenal artery use, superior mesenteric artery stenosis) have been excluded on angiography.

In fibromuscular dyspasia particularly, there may be stenoses of smaller branches of the renal artery. Correction of these lesions requires microvascular techniques using vein graft; for this to be performed the kidney is usually removed from the body and perfused with cold fluid to protect aginst ischaemic damage. It is replaced as an autotransplant in the iliac fossa, anastomosing the artery and vein to the iliac vessels (Fig. 12) 340.

Occasionally the kidney has shrunk too far for function to be saved by reperfusion, but it is still a source of renin, causing hypertension. In other cases it proves impossible safely to reperfuse the kidney, either because of the extent of disease or the frailty of the patient, and the other kidney is functioning. In these circumstances nephrectomy may provide a simple and safe remedy for hypertension.

The results of surgery in fibromuscular dysplasia of the renal artery are reasonably good. About 60 per cent of patients are ‘cured’ of their hypertension, 30 per cent have improved blood pressure control, and only 10 per cent fail to respond. Very few patients with fibromuscular dysplasia have renal failure but failure is usually also improved, or at least stabilized, by surgical intervention.

Patients with atheromatous disease respond rather less well. Less than 40 per cent are probably ‘cured’ of their hypertension, although slightly more are ‘improved’ and about 20 per cent fail to respond.

Renal failure occurs in about 15 per cent of patients with renovascular hypertension due to atheromatous disease compared with only about 2 per cent of those with fibromuscular dysplasia. Overall, approximately 50 per cent of patients with renal failure are improved, 40 per cent stabilized, and 10 per cent show further deterioration. These results are in the face of declining renal function prior to surgery and so stabilization of renal function is an improvement attributable to surgery. The benefits of surgery decline with increasing degrees of renal failure. Only 20 per cent of those with a serum creatinine greater than 180 &mgr;mol/l are improved, about 60 per cent stabilize, and 20 per cent fail to respond. Of those with a serum creatinine greater than 270 &mgr;mol/l about 30 per cent fail to respond, and the perioperative mortality rate, in one series, increased from 3 per cent to 13 per cent. Occasionally, patients with severe renal failure respond dramatically well to revascularization of the kidney. They tend to have bilateral renal artery occlusions and potentially functional renal tissue which has been maintained via collaterals. It is obviously important to be reasonably sure that such tissue exists before embarking on complicated, relatively high-risk surgery. The angiogram may show filling of the distal renal artery via collaterals (Fig. 4) 332. Occasionally, but not always, a radioisotope scan demonstrates collateral blood flow reaching the kidney. Renal length on ultrasound is a fairly good guide: if less than 8 cm revascularization is unlikely to improve renal failure and may increase renovascular hypertension. Demonstration of viable glomeruli in a renal biopsy, taken either pre- or perioperatively, is also useful.

The mortality rate associated with renal artery surgery is low in patients with fibromuscular hypertension. In patients with atheromatous disease it is 3 to 10 per cent if surgery is confined to only one renal artery but this increases with extended surgery: aortic reconstruction with bilateral renal artery revascularization carries a perioperative mortality rate of up to 15 per cent. As might be expected the presence of significant atheromatous disease in the coronary and cerebral circulations increases mortality. It is probably also higher in those with renal failure.

A direct comparison between percutaneous transluminal arteriography and surgery in the treatment of renovascular hypertension cannot be made because the patient groups are dissimilar. The former is used mainly to treat young patients with fibromuscular dysplasia who are otherwise reasonably fit. It is not applicable to the patient with widespread atheroma affecting the aorta and extending into the renal arteries, who is also at much greater risk from surgery. The patients with fibromuscular dysplasia who come to surgery are those with complex branch lesions and those in whom percutaneous transluminal angioplasty has failed.

This disease encompasses a wide spectrum of patients. At one end is the man in his fifties with significant stenosis of one renal artery causing hypertension, with normal renal function, a slightly irregular aorta, mild asymptomatic peripheral vascular disease, and no evidence of significant heart disease. At the other extreme is the woman in her eighties with severe renal failure, hypertension, rest pain, and a history of several myocardial infarctions. In between are patients with any combination and severity of these problems.

If hypertension that can be controlled satisfactorily with medication is the main problem and the kidneys appear to be functioning adequately, no further intervention is required, provided renal function is monitored regularly. If hypertension is difficult to control or renal function starts to deteriorate and the lesion is a short stenosis or even occlusion, percutaneous transluminal angioplasty is indicated. However if the lesion is ostial, as it often is in these circumstances, this treatment may fail; if it succeeds the stenosis is likely to recur. In these circumstances many vascular surgeons would proceed directly to surgery.

In patients with more extensive atheromatous disease, who are usually older, both renal arteries and the aorta itself are involved. Thus the patient may have renal failure and lower limb ischaemia or an abdominal aortic aneurysm, both of which require treatment. However these patients also probably have significant myocardial ischaemia and will respond poorly to the stress of an aortic clamp; they also have cerebrovascular disease, which increases the risk of perioperative cerebrovascular accident. Hypertension may be controllable by medication, rest pain by analgesics, renal failure by dialysis, and the risks of aneurysm rupture can be ignored. However, this approach usually leads to a relatively poor quality of life. It is possible to replace the aorta and revascularize both kidneys, even in the elderly. Obviously the risks associated with such surgery depend on the patient's general health and the skill of the surgeon and anaesthetist. In major centres of such surgery the perioperative mortality has been reported as less than 6 per cent. Renal failure is improved or stabilized in all but 6 per cent, and hypertension improves in over 90 per cent. If the patient's health dictates it and the extent of associated disease allows, the kidneys may be reperfused via the visceral vessels to avoid the dangers of aortic clamping. Although surgical treatment of renovascular disease was previously restricted to the young and middle-aged, it is now being used in the treatment of the elderly, in whom medical treatment or percutaneous transluminal angiography prove unsatisfactory. The attendant risks to heart and brain can be reduced by careful monitoring, newer techniques, and even coronary artery bypass grafting or carotid endarterectomy prior to renal revascularization.

These patients are usually young, and have none of the stigmata of generalized vascular disease. They rarely develop renal failure, but they are at risk from cerebrovascular accidents or heart disease because of their hypertension: this must be detected and treated before the development of irreversible end-organ damage. Younger patients have a less satisfactory response to medical treatment, and surgical intervention is less hazardous. In general, fibromuscular renovascular hypertension responds well to either percutaneous transluminal angioplasty or surgical intervention, but the former is preferred since it is a relatively simple procedure with fewer associated risks. Most dysplastic lesions are amenable to percutaneous transluminal angioplasty, except for those in the smaller branches of the renal artery where the risks of dilatation are greater and the technical success rate diminishes. Surgery is usually only undertaken in patients in whom percutaneous transluminal angioplasty has failed or in whom the lesion is in the distal branches.

The most common cause of renovascular hypertension in children is fibromuscular dysplasia. Other causes include arteriovenous malformations, aneurysms, hypoplasia, Takayasu's disease, and neurofibromatosis. If the lesion is proximal it is occasionally possible to excise the affected artery and reanastamose the vessel to the aorta; however, a more complex procedure is usually required. Aortorenal bypass grafting is an option, but the use of saphenous vein in prepubertal children has led to aneurysmal dilatation or stricture formation in the graft at a later stage. This can be avoided by using autogenous artery, usually one of the internal iliac arteries. An alternative approach is to autotransplant the kidney. This approach is essential when microscopic bench repair of intrarenal abnormalities is required. Hypoplastic renal arteries are often associated with a low coarctation of the aorta, making this vessel unsuitable for bypass grafting. However the iliac arteries are usually spared, and either the kidney can be autotransplanted or an iliorenal graft constructed. Occasionally, especially in those patients with unilateral disease, the kidney is not worth salvaging and the only option is nephrectomy.

The results of surgery for renovascular hypertension in children are excellent, with a cure rate of 90 to 100 per cent.

Deceleration in either a vertical or a horizontal direction imposes stretching and shearing forces on the renal artery because it joins the relatively mobile kidney to the fixed aorta. These forces cause intimal tears, particularly in the middle third of the artery, which may then lead to dissection and thrombosis of the vessel. Occasionally the entire vessel wall is disrupted and a perirenal haematoma forms. There is often remarkably little blood loss (and few concomitant physical signs) because of renal artery spasm. Blunt trauma to the kidney usually causes parenchymal damage, but it may disrupt the renal pedicle.

Complete occlusion or disruption of the main stem of the renal artery leads to renal infarction. Most of these injuries occur in young adult males with previously normal renal arteries. Development of cortical collateral vessels that might maintain viable renal tissue is usually poor. Thus in most of these patients there is a very narrow window of opportunity, probably within 1 to 2 h of injury and almost certainly within 6 h, in which to revascularize the kidney and recover function. There are reports in the literature of successful revascularization several days after injury, but these are relatively rare. Late revascularization is usually associated with development of renovascular hypertension and a need for nephrectomy. The patient may develop renovascular hypertension even if no attempt is made at revascularization, since ischaemic renal tissue survives, maintained by a partially occluded renal artery or cortical collaterals. The incidence of hypertension following trauma has been cited as between 10 and 50 per cent, but it is difficult to assess: not all cases of renal trauma are detected, and follow-up is incomplete on those that are detected. It develops between days and years after the injury and in up to 80 per cent of cases appears to be transient.

Renal artery damage is often missed in the initial assessment since there are frequently few signs and often other injuries. It is more likely to be discovered at autopsy, during later urological assessment, or when the patient is evaluated for hypertension. At the initial presentation, patients may complain of unilateral flank pain, and occasionally of abdominal pain. Patients with severe disruption present with shock; those with lesser injuries have at least microscopic haematuria, some have loin tenderness and in a few a bruit can be heard in the upper abdomen or flank.

Any trauma patient with even microscopic haematuria must undergo intravenous urography. This will demonstrate any abnormality in renal function and, very importantly, will confirm that the patient has another functioning kidney. Patients in whom the intravenous urogram is normal can be treated conservatively; if the examination shows no function on one side and the patient is haemodynamically unstable, immediate laparotomy is required to deal with the renal artery damage. If the patient is stable but has no function demonstrable in one kidney angiography or CT scan with contrast to demonstrate the arteries may be helpful as a prelude to reconstructive surgery, provided it does not delay intervention unduly.

A patient who presents with hypertension some time after trauma should undergo angiography to delineate any renal artery defect, in the same way as any other patient with renovascular hypertension.

Most patients with renal trauma and normal renal function on intravenous urography can be managed conservatively. Those with severe disruption of the renal pedicle who present with profuse bleeding and absent renal function need emergency laparotomy. In these patients, extensive disruption often precludes reconstruction of the kidney, and nephrectomy is required.

The stable patient with absent renal function can either undergo surgery with a view to reconstruction or be treated conservatively. If more than 6 h have elapsed since the injury, reconstruction is unlikely to be successful and, unless both kidneys are non-functional, is probably not justified. The ischaemic kidney often shrivels and poses no further problem. If the other kidney is normal, renal function is maintained and relatively few patients develop troublesome renovascular hypertension. Laparotomy to remove an ischaemic kidney is unnecessary at this stage.

Reconstruction may be feasible if surgery is performed within 6 h of injury. Thrombectomy of the renal artery and excision of the affected segment with reanastomosis or vein interposition graft are usually required.

Patients known to have suffered trauma to the renal artery, especially those with non-functioning kidneys on intravenous urography should undergo regular blood pressure monitoring so that renovascular hypertension can be detected and treated before end-organ damage results.

Renal artery aneurysms appear to be uncommon. Retrospective studies of postmortem reports put the prevalence at about 0.01 per cent; however many are small and intrarenal, and might not be detected on routine postmortem examination. Angiographic studies detect such aneurysms in up to 1 per cent of individuals (Fig. 13) 341, although these are often selected for study on the basis of symptoms such as pain, haematuria, or hypertension, which might bias the findings. However, one prospective postmortem study found the prevalence to be 9.7 per cent in an apparently unselected series of individuals. Renal artery aneurysms occur with equal frequency in either sex and have been detected at all ages.

The majority of these aneurysms are less than 1 cm in diameter and most are saccular. They occur with equal frequency in the main stem artery, in the primary branches, and in the peripheral branches. They are bilateral in up to 10 per cent of patients, and multiple aneurysms occur within one kidney in another 13 per cent. Many of these aneurysms occur at branch points, and as they sometimes occur in children they may be due to congenital weakness of the arterial wall at these points. They are rarely associated with atherosclerosis, despite the fact that about 20 per cent are calcified.

Fibromuscular dysplasia, and in particular the most common form, medial fibroplasia, is associated with microaneurysm formation in the main stem artery or its primary branches. However, these are usually considered separately from other renal artery aneurysms.

The predominant risk is rupture. Rupture is most likely to occur during pregnancy, especially in the third trimester, but even then splenic artery aneurysm rupture is 4 to 5 times more common. It presents with abrupt onset of flank pain and hypotension and the mortality is high for both mother (56 per cent) and fetus (82 per cent). Few renal artery aneurysms which rupture in pregnancy are associated with medial hyperplasia, despite its occurrence mainly in young women. Prior to rupture they are usually asymptomatic, although a bruit may be audible.

Aneurysms may thrombose but, because they are saccular, rarely cause arterial occlusion. For the same reason they are unlikely to shed emboli into the distal renal vasculature. Occasionally they are associated with haematuria, presumably because of erosion into the renal pelvis.

Renal artery aneurysms have been associated with hypertension but are rarely the cause of it. Many of the patients in whom they are detected on angiography are investigated because of hypertension and the aneurysm is incidental: surgical treatment of the aneurysm does not usually cure the hypertension unless the aneurysm distorts and compresses the adjacent artery and causes a functional stenosis. Fibromuscular dysplasia may be associated with both hypertension and renal artery aneurysms but it is the stenotic component of the disease that is important in producing hypertension.

An aneurysm detected in pregnancy should be repaired before the third trimester, despite the risks to the fetus. Females of childbearing age should also have such an aneurysm repaired. The aneurysm can often be excised and the defect in the side of the artery closed with a vein patch. If the aneurysm is more extensive an aortorenal vein graft may be required. When multiple intrarenal aneurysms are present, ex-vivo reconstruction becomes necessary.

Unless a renal aneurysm is thought to be the cause of hypertension or troublesome haematuria from the renal artery, they rarely need treating in any other patients. Few patients with aneurysms experience problems, even over long periods of time.

Renal arteriovenous fistulae are uncommon. Approximately 30 per cent are congenital; nearly half of the acquired fistulae are secondary to renal biopsy and many follow renal surgery, such as nephrolithotomy or partial nephrectomy, or renal trauma. A few may arise from erosion of a renal artery aneurysm into an adjacent vein and they are sometimes seen in renal tumours. Fistulae occasionally develop in the renal pedicle following nephrectomy, probably due to mass ligature of the artery and vein. Fistula following renal biopsy seems particularly likely to occur if the patient is hypertensive prior to the procedure.

Congenital arteriovenous fistulae are often large, complex, cirsoid malformations with multiple communications, and occupy much of the kidney (Fig. 14) 342. Those which develop following biopsy are usually small and intrarenal (Fig. 15) 343. Those associated with trauma may affect the intra- or extrarenal vessels and sometimes even the inferior vena cava.

Small arteriovenous fistulae usually cause no problems. Larger fistulae are associated with two main problems. The first is due to the increased demands on cardiac output as blood is diverted through the low resistance fistula away from the general circulation. Systolic hypertension with a low diastolic pressure develops, and the patient may suffer high output cardiac failure. The second complication is the steal of blood from normal renal parenchyma to the fistula. Relative ischaemia of renal tissue stimulates renin production and leads to renovascular hypertension, so that diastolic pressure also rises. In a patient with pre-existing renal parenchymal disease, a fistula may also seriously compromise renal function. Rarely, an arteriovenous fistula ruptures to produce an intra- or extrarenal haematoma.

About 60 per cent of patients have microscopic or macroscopic haematuria, but unless they have symptoms associated with heart failure or hypertension there are usually no significant problems in the history. On examination the only sign of the fistula itself is a continuous bruit audible in the upper abdomen or flank.

An intravenous urogram may demonstrate reduced opacification of renal substance where blood has been diverted to the fistula, or distortion of the renal pelvis by an intrarenal fistula. In most cases, though, this investigation is of little help. Renal isotope scans may also demonstrate reduced perfusion but offer little more. The definitive investigation is the angiogram: this demonstrates the fistula, its feeding and draining vessels, and the extent of steal from the rest of the kidney.

Measurement of renal vein renin levels may not be helpful in the patient who has also developed renovascular hypertension, because the high flow of arterial blood through the renal vein dilutes renin washed out from the affected kidney.

Most small intrarenal arteriovenous fistulae do not require treatment. The majority of those that develop following renal biopsy or trauma close spontaneously within 18 months.

Single intrarenal fistulae which cause problems can often be treated by embolization under radiographic control. Metal coils are usually placed in the segmental feeding vessel to promote thrombosis, often with temporary balloon catheter occlusion of the vessel proximal to this point to prevent the coils being carried through into the vein and systemic circulation.

Congenital arteriovenous fistulae are often too complex to treat by embolization and require surgery. This usually involves partial or complete nephrectomy because of the extent of renal disease.

Fistulae of the pedicle vessels are usually treated surgically. The fistula is excised and standard repair of the artery and vein performed, either by simple suture, vein patch, or interposition graft.

Hypertension associated with renal arteriovenous fistulae is improved or cured in 60 per cent of patients in whom the fistula is closed or the kidney removed. If the fistula is associated with trauma the response rate increases to 85 per cent.

Goldblatt H, Lynch J, Hanzai RF, Summerville WW. Studies on experimental hypertension–I, The production of persistent elevation of systolic pressure by means of renal ischaemia. J Exp Med 1934; 59:347–79.

Jenkins AMc. Operations for renal ischaemia. In Bell PRP, Jamieson CW, Ruckley CV, eds. Surgical Management of Vascular Disease, 751–66.

Maldonado JE, Sheps SG, Bernatz PE, DeWeerd JH, Harrison EG. Renal arterio-venous fistula. Am J Med 1964; 37:499–513.

Morris PJ. Renovascular hypertension: the indications for and the results of surgery. In: Bell PRP, Jamieson CW, Ruckley CV, eds. Surgical Management of Vascular Disease, 739–50.

Novick AC. Surgical correction of renovascular hypertension. Surg Clin N Am 1988; 68:1007–25.

Peterson NE. Review article: traumatic bilateral renal infarction. J Trauma 1989: 29:158–67.

Sos TA. Angioplasty for the treatment of azotaemia and renovascular hypertension in atherosclerotic renal artery disease. Circulation 1991; 83(suppl I):I1162–II166.

Tegtmeyer CJ, Bayne Selby J, Hartwell GD, Ayers C, Tegtmeyer V. Results and complications of angioplasty in fibromuscular disease. Circulation 1991; 83(suppl I):I155–I161.

Tham G, Ekelund L, Herrlin K, Lindstedt EL, Olin T, Bergentz S-E. Renal artery aneurysms. Natural history and progress. Ann Surg 1983; 197:348–52.

van Bockel JH, van Schilfgaarde R, van Brummelen P, Terpstra JL. Renovascular hypertension. Surg Gynecol Obstet 1989; 169:467–78.