Abdominal Aortic Aneurysm

The aorta is the main artery of the body, which supplies oxygenated blood to all other parts. The aorta arises from the left ventricle (the main pumping chamber of the heart) and arches up over the heart before descending, behind it, through the chest cavity. It terminates in the abdomen by dividing into the two common iliac arteries of the legs. The aorta is thick-walled and has a large diameter to cope with the high pressure and large volume of blood passing through it.

Disorders of the aorta

The aorta, like other arteries, can become narrowed as a result of atherosclerosis (fat deposits on its walls), which may cause hypertension (high blood pressure). Coarctation of the aorta (in which the aorta is abnormally narrow at birth) and aortitis (inflammation of the aorta wall) are examples of aorta-specific disorders. Both aortitis and atherosclerosis may result in an aortic aneurysm (ballooning of the aorta wall), which may require surgery.

Aortic aneurysm in detail - technical

Introduction and definition

An aneurysm is by definition an abnormal dilatation of an artery or vein, and the application of this general principle to the abdominal aorta has seldom presented any problems in routine clinical practice. The universal use of abdominal ultrasonography as a basic diagnostic tool, and particularly the introduction of screening programmes for abdominal aortic aneurysms, has recently highlighted the need for a more precise definition to allow appropriate diagnosis of the many marginal aortic dilatations, or small aneurysms, that are now being discovered.

The diameter of the abdominal aorta, like other biological measurements, conforms to a normal population distribution curve. Median aortic diameter increases with age, is greater in men than in women, and is influenced by race, weight, height, and the prevalence of hypertension in the population studied. In men, after 60 years of age, the shape of the curve becomes increasingly skewed to the right as the prevalence of abdominal aortic aneurysm increases. A definition of aneurysm based on deviation from the mean aortic diameter for a population will therefore be of less value than one based on the median diameter.

In some patients with an abdominal aortic aneurysm, dilatation may also involve the suprarenal or thoracic aorta, leading to a diagnosis of thoracoabdominal aortic aneurysm or generalized arterial ectasia, depending on the amount and extent of the dilatation. Any definition that relies solely on comparison of the diameter of a suspected aneurysm with that of adjacent 'normal' aorta will fail in the 5 per cent of patients whose aneurysms are not confined to the infrarenal aorta.

Measurement of aneurysm diameter is commonly made from ultrasonograms, computed tomograms (CT) or magnetic resonance images (MRI); diameters can be recorded as anteroposterior, transverse, or the maximum diameter any plane. Since aneurysmal aortas elongate as well as dilate, obliquity of the long axis of the aneurysm is common. In consequence, transverse and any-plane diameters measured by MRI and CT are usually greater than anteroposterior diameters and always unreliable. Anteroposterior diameters measured by ultrasonography are usually less than those measured by CT and MRI, and correspond more closely to true diameter obtained by direct measurement at operation.

A modern definition of an aortic aneurysm would take into account the above facts and make due allowance for the inaccuracies of measurement inherent in even the most precise methods of diagnosis. The following definition is proposed: An aortic aneurysm is present when the maximum external anteroposterior diameter of the aorta either (1) is at least 4.0 cm or (2) exceeds the diameter of the adjacent aorta by at least 0.5 cm.

The imprecise but clinically useful term 'aortic ectasia' should be reserved for those cases where the aorta appears abnormally wide but is not aneurysmal by the above definition.

In practice an aneurysm is diagnosed by most clinicians when the aortic diameter is 3.0 cm or more. This pragmatic approach, which takes no account of sex, height, race, age, body morphology, modality or method of measurement, at least has the virtue of simplicity if not reproducibility or validity.

Epidemiology

Aortic aneurysms are very rare in people under the age of 55 years and before that age are virtually confined to patients with Marfan, Ehlers–Danlos, or arteria magna syndromes. The common idiopathic abdominal aortic aneurysm is largely a disease of elderly men. Comparison of fatalities from ruptured abdominal aortic aneurysm by age and sex show that deaths are 13 times more common in men than in women at age 60 to 65 years, but over 80 years of age are only four times more common in men than in women. At age 85 almost three times as many women as men are still alive, so among the very elderly the numbers of men and women presenting with ruptured aneurysms are similar. The changing pattern of presentation with age, combined with an increase in the number of elderly people in the populations of most wealthy nations, has led some surgeons to conclude erroneously that abdominal aortic aneurysm has increased in incidence disproportionately in women.

The annual risk of death from aortic aneurysm increases from 125/10 million for men aged 55 to 59 years to 2728/10 million at age 85 and over. At age 70 to 74 years, aortic aneurysms are responsible for 2.2 per cent of all deaths in men, and abdominal aortic aneurysms account for 77 per cent of these. The disease is a particularly common cause of unexpected deaths, and more than 5 per cent of sudden deaths investigated by autopsy in men over 50 years of age are found to be due to ruptured abdominal aortic aneurysm.

In the past 30 years there has been a linear increase in the number of recorded deaths from aortic aneurysm in England and Wales. In part this can be explained by the progressive growth in the number of elderly people in the population, but age-specific death rates for aortic aneurysm have also increased over the age of 60 years. Some of the increase might be real, but much of the apparent change is due to enhanced awareness of the disease, improved diagnosis, and altered referral patterns associated with the establishment of specialist vascular units. This issue is difficult to resolve because of the unreliability of national records of cause of death, which ultimately depend on the diagnostic acumen of the reporting doctor, seldom supported by autopsy evidence. In the United States of America a similar, but more rapid, increase in age-adjusted mortality for aortic aneurysm occurred between 1951 and 1968, but the number of recorded deaths then stabilized and, in white ('Caucasian') males, has declined since 1976, possibly as a consequence of the increasing impact of elective surgery for aneurysm.

Reports from the early part of the twentieth century illustrate that there has been a qualitative as well as a quantitative change in abdominal aortic aneurysmal disease. Before the introduction of effective antisyphilitic therapy, the majority of aortic aneurysms were a manifestation of tertiary syphilis, and the mean age of presentation was consequently lower than at present. In developed countries, syphilitic aneurysms are now rare and idiopathic abdominal aortic aneurysms of elderly people represent the vast majority of cases seen.

There has been debate about the influence of racial factors on abdominal aortic aneurysms, with most studies focusing on comparison of black and white populations in the United States and South Africa. There is no doubt that the disease is seen less often in people of African descent than in whites, but in both countries it is difficult to discover whether the differences simply reflect the lower life expectation and mean age of the black population and their poorer access to health care. The high prevalence of hypertension in black (African origin) people gives theoretical grounds for suspecting that, age for age, abdominal aortic aneurysm might well be more common than in whites. At present the evidence for racially determined differences in incidence must be regarded as suspect, and careful epidemiological studies will be required to resolve this issue.

Prevalence

The prevalence of any disease represents the total number of cases, both diagnosed and occult, present in the population at any given time, and should be distinguished from incidence, which is the number of new cases diagnosed over a specified period of time.

The majority of aortic aneurysms are asymptomatic and impalpable; consequently their prevalence in the community can be determined only by systematic screening. The results of many screening studies have been published, but most are fundamentally flawed as measures of prevalence. The incidence of aortic aneurysm increases rapidly with age and is much higher in men than in women. Prevalence studies must therefore differentiate each 5- or 10-year age group and men from women. Studies in patients with hypertension, atherosclerosis, or other diseases cannot produce prevalence data relevant to the whole population. Examination of the data does, however, allow a number of conclusions to be drawn.

Aetiology

The common abdominal aortic aneurysm of elderly men has been labelled as 'atherosclerotic'. This classification has little justification, has paralysed thinking, and needs to be re-examined. It is interesting to note that aneurysmal disease is encumbered by more than itsfair share of unhelpful, or frankly misleading, descriptive terms, among which are atheromatous, mycotic, inflammatory, dissecting, and arteriovenous aneurysms. In the elderly individual the aorta, in common with every other artery, will have obvious features of atherosclerosis but this is not enough evidence to make credible a pathological diagnosis that does not fit with many known facts about the disease.

Tilson has compared patients with abdominal aortic aneurysms and those with occlusive aortoiliac disease. He found that the patients with aneurysm were nine times more likely to be male; were, on average, 11 years older; and were much less likely to have had previous arterial surgery. Patients with occlusive disease were 16 times more likely to require reoperation after aortic surgery. In addition, patients with aneurysm were, on average, more than 5 cm (2 inches) taller than patients with occlusive arterial disease, and had a significantly greater body surface area. In our own experience, aneurysms in patients who have associated occlusive arterial disease are generally smaller and may be less likely to rupture than in patients without severe atherosclerosis. This view is supported by the observation that mean growth rates for small aneurysms are 50 per cent faster when there is no obvious occlusive arterial disease.

Genetic predisposition

Surgeons have been aware for some years of the occasional occurrence of several cases of abdominal aortic aneurysm within families, but proof of the familial pattern of the disease has been difficult to obtain because of the absence of symptoms and the advanced age of onset in most patients. Even carefully elicited family histories will often be unhelpful, since the majority of those with the aneurysmal diathesis will die from other causes, and many who die from rupture of the aneurysm will have the wrong diagnosis recorded unless an autopsy is carried out. The problem is compounded by the absence of a common name for aortic aneurysm, which is consequently unfamiliar to the general public.

Ultrasonographic screening studies have shown a prevalence of abdominal aortic aneurysm of 30 per cent in first-degree male relatives of patients with the disease. Because of the late age of onset, many of those with no evidence of an aneurysm at the time of screening could well develop the disease when they are older, so the lifetime prevalence in brothers and sons of patients with aneurysm may be as high as 50 per cent. The search for the gene or genes responsible is hindered by the absence of three-generation families with confirmed inheritance of aneurysm for genetic studies. It is likely that, with the present rapid strides in molecular biology, this problem will be solved in the next few years, using techniques such as paired-sibling analysis.

Attention to date has largely been focused on possible variations in collagen genes, particularly type III collagen, but no convincing link with the common idiopathic aortic aneurysm of elderly people has been established. Other target genes are those controlling production of metalloprotease or metalloprotease inhibitors. Given the very strong association of aortic aneurysm with tobacco smoking it is likely that in most patients any genetic predisposition requires an environmental potentiating factor to be present for the phenotype to be manifest. The very late age of onset, the fact that most cases are occult, the need for environmental potentiation, declining rates of tobacco smoking in developed countries, and the possibility of multigene predisposition all conspire to make the search for a genetic cause of aortic aneurysm among the more difficult in molecular biology.

Connective tissue degradation

Research efforts have concentrated on attempts to discover the mechanism of breakdown of collagen and elastin in the arterial wall of enlarging and ruptured abdominal aortic aneurysms. Several studies have shown the presence of proteolytic activity in tissue from aneurysmal aorta, but authentic collagenase has been shown to be present only when the aneurysm has ruptured. It is uncertain whether collagenolysis is the cause or a consequence of rupture. Similar uncertainties surround the detection of elastase in the aortic wall and serum of patients with aneurysm. The discriminant value of such analyses between patients with and without aneurysms has not always been confirmed, although recently a unique metalloprotease elastase has been found only in patients with aneurysm. In one family in which aneurysms occurred in several members at an early age, it has been shown that the disease is linked to a genetically determined defect in type III collagen. It is possible that other genetic variations in type III collagen may account for some, if not all, cases of abdominal aortic aneurysm. Such a finding, although at present speculative, would mirror the situation in osteogenesis imperfecta, where the disorder has been shown to be caused by a large number of different genetic variations in type I collagen.

Recently it has been shown in patients who have had arterial bypass of popliteal aneurysms with autogenous vein that the vein graft has a higher propensity to dilatation, ectasia, and aneurysm formation than comparable vein grafts used to bypass atherosclerotic arterial occlusive disease. Metalloprotease concentrations are also higher in patients with popliteal aneurysms, suggesting a systemic process of vascular collagen breakdown and vessel dilatation.

Environmental potentiation

Abdominal aortic aneurysms have been shown to be associated with:

male sex;
advancing age;
tobacco smoking;
hypertension;
chronic obstructive airways disease (irrespective of smoking history);
occlusive arterial disease affecting coronary, carotid, and limb arteries.

In addition, aortic aneurysm is most common in whites and those who are tall, but these are unlikely to be independent disease determinants and probably reflect 'racial' differences in population age structure and the relation between height, longevity, and socio-economic status.

The mechanisms by which environmental influences interact with underlying genetic predisposition to produce abdominal aortic aneurysm in an individual patient are at present uncertain, but the first three factors listed above have by far the greatest importance. It is interesting that, in common with other diseases, the marked protective effect of female sex is progressively lost with advancing age, although even in the very elderly person the risk of dying from aortic rupture is three times greater for men than for women.

Natural history

The great majority of abdominal aortic aneurysms are fusiform and are confined to the infrarenal segment. Small, saccular aneurysms are sometimes seen adjacent to atheromatous plaques in patients with predominant occlusive disease and rapidly growing, infective, 'mycotic' saccular aneurysms occasionally occur as a consequence of bacteraemia. Mycotic aneurysms are a local manifestation of systemic disease, require urgent medical and surgical treatment irrespective of size, and have a totally different natural history from that discussed here of the common idiopathic aortic aneurysms of the elderly person.

The median diameter of the infrarenal abdominal aorta increases with age in both men and women. An aortic aneurysm begins as a local accentuation of this normal ageing process. Physical laws predict that the rate of growth will increase with diameter, so once any local accentuation has started it can be expected to increase progressively with time. This explanation accounts for the three types of dilatation common seen:

a local aneurysm with normal adjacent arteries;
generalized arterial ectasia;
local dilatation within an ectatic arterial system.

Serial measurements have confirmed that growth rates increase as abdominal aortic aneurysms enlarge. The development of symptoms, risk of rupture, and clinical management of aortic aneurysms depend largely on their diameter, so it is convenient to discuss the natural history in relation to three somewhat arbitrary size ranges. It is important to remember, however, that the lifecycle of an individual aortic aneurysm is a continuous process from initial development to eventual rupture, the inevitability of which can be prevented only by elective surgery or prior death from some other disease. Looked at in this way, the description of an aortic aneurysm as a cancer of the artery is not quite so fanciful as it might seem.

Very small aneurysms (less than 4.0 cm diameter)

The prevalence of aneurysms of less than 4.0 cm in diameter has become apparent only with the introduction of screening programmes for the disease. Two-thirds of all aneurysms detected by population screening are of this size, the reasons for which are interesting and help in understanding some important features of the disease, as follows.

The longest part of the lifecycle of any aneurysm will be when it is small, since incremental growth rates increase as the aneurysm enlarges.
Large aneurysms are more likely to be detected and present in routine clinical practice.
The larger an aneurysm becomes, the more likely it is to rupture and remove the patient forever beyond the reach of screening programmes.

Very small aortic aneurysms generally enlarge much more slowly than the large aneurysms that present in routine clinical practice, and median growth rates of 0.2 cm per annum are usual. Clinical and autopsy evidence indicates that even these very small aneurysms do sometimes rupture, but there are insufficient data for the risk to be quantified accurately. Several clinical follow-up studies have shown no cases of rupture occurring in such patients while the aneurysms remained very small, but ruptures did occur as the aneurysms grew.

Small aneurysms (4.0–5.9 cm diameter)

Autopsy studies of patients with an abdominal aortic aneurysm show that one-third of aneurysms of less than 6.0 cm in diameter had ruptured and caused death. If we assume that the death rate in the two-thirds of patients who died with, rather than from, their aneurysm is the same as in the general population of the same age, these data suggest a rupture risk of around 2 per cent per annum. A complicating factor in accurate interpretation is that aortic diameter measured at autopsy is likely to have been less than the diameter of the same aneurysms in life. The actual risk of rupture for aneurysms of less than 6.0 cm diameter in life might therefore be substantially less than 2 per cent. Follow-up studies of patients managed conservatively with aneurysms of less than 6.0 cm diameter at the time of diagnosis have demonstrated a mean rupture rate of 6 per cent per annum over 3 years. Rupture rates are initially low but tend to increase progressively with the length of follow-up as the aneurysms continue to expand. For aneurysms of 4.0 to 4.9 cm diameter the mean expansion rate is 0.5 cm per annum, increasing to 0.7 cm per annum for aneurysms of 5.0 to 5.9 cm diameter.

Many of the patients managed conservatively have severe obstructive airways disease that would make operative intervention too hazardous. Obstructive airways disease is associated with more rapid expansion of aortic aneurysms and a higher risk of rupture. Consequently, data from follow-up studies of patients unfit for surgery cannot be used to estimate rupture risks for healthy patients. In addition, diameters measured 2 or 3 years before rupture occurs can give no useful information about rupture risk relative to current diameters.

Recently, data from the United Kingdom Small Aneurysm trial have revealed a rupture risk of not more than 1 per cent per annum for aneurysm of 4.0 to 5.5 cm maximum anteroposterior diameter measured by ultrasonography. Many of the ruptures in fact occurred in aneurysms that had already grown to more than 5.5 cm at the last measurement before rupture.

Large aneurysms (greater than 6.0 cm diameter)

Nowadays, patients with aneurysms of more than 6.0 cm diameter are invariably advised to have elective surgery. What we know of the natural history of large aneurysms comes from studies before operative treatment became possible in 1951, or from contemporary studies in patients too ill to undergo major surgery.

Studies from earlier in the twentieth century of clinically detected, and therefore presumably large and often symptomatic, abdominal aortic aneurysms report that most patients died within 3 years, and two-thirds of all deaths were from rupture of the aneurysm. Since the death rate from all causes in patients of the same age with no aneurysm was around 5 per cent per annum, the risk of rupture in these studies was therefore around 10 per cent per annum. Contemporary studies of patients with severe cardiac, respiratory, or other disease considered to make the risks of elective surgery unacceptably high show that aneurysm rupture accounts for half of all deaths, a rupture risk of 10 to 15 per cent per annum.

Clinical presentation

The majority of abdominal aortic aneurysms are asymptomatic and are often discovered incidentally. The patient may notice a pulsatile epigastric mass for the first time typically while lying relaxed in bed or bath. Large aneurysms in thin patients are readily detected on routine abdominal examination, but most are now discovered by ultrasonography or abdominal radiography performed to investigate unrelated symptoms. Urologists are a frequent source of referrals for many vascular surgeons, since prostatic hypertrophy and aortic aneurysm are both disorders of elderly men and detection of the aneurysm by abdominal palpation is easier during anaesthesia for prostatic resection. It is likely that much of the apparent increased incidence of abdominal aortic aneurysm over the past decade is attributable to general adoption of abdominal ultrasonography as the routine first-line investigation for abdominal symptoms.

In Britain, ruptured abdominal aortic aneurysm still accounts for around a third of all operations for the disease, but in the United States the figure for major vascular centres is currently between 5 and 20 per cent. Community studies have shown that 60 per cent of patients with ruptured abdominal aortic aneurysms do not reach hospital alive, while some of those who do are not operated upon. In Britain, rupture of the abdominal aortic aneurysm is sadly still the way in which more than half of all cases present. In both hemispheres there is a seasonal variation in the incidence of aortic rupture, with more cases occurring in the winter months. The reason for this pattern is unknown, but may be related to the similar observed seasonal variation in mean blood pressure.

Symptoms and signs of aortic rupture

Typically, rupture of an abdominal aortic aneurysm produces the sudden, unheralded onset of severe central abdominal and lumbar back pain. A pulsatile mass may be palpable in the epigastrium. Some patients may have experienced dull back pain of lesser severity for hours or days before, due to acute expansion of the aneurysm immediately before rupture. The lumbar pain may be worse on one side, commonly the left, because of the direction in which the retroperitoneal haematoma spreads. There may be a variable degree of psoas spasm, and sometimes pain in the lower limb, due to compression of lumbar or sciatic nerve roots. Rupture of an internal iliac (hypogastric) arterial aneurysm commonly produces maximal pain in the buttock and, rarely, blood may track with the sciatic nerve through the greater sciatic foramen to produce a gluteal haematoma.

Other early symptoms and signs depend on the volume of acute blood loss. Once the posterior peritoneum is breached, the patient will rapidly bleed to death into the peritoneal cavity, and most immediate deaths are due to intraperitoneal rupture. Survival after rupture depends on an intact posterior peritoneum, tissue tamponade, and early emergency surgery. When the connective-tissue tamponade provided by the retroperitoneum is very effective, or the leak is small, only modest haemorrhage may occur, and these patients can survive long journeys to hospital and several days before exsanguinating haemorrhage occurs. The self-selection of such patients for transfer to distant tertiary referral centres may be partly responsible for the superior results of some units. In most cases, tamponade is less effective and arrests acute haemorrhage only when assisted by hypotension secondary to blood loss. These patients exhibit pallor, sweating, tachycardia, and anuria; transfusion alone, by raising the blood pressure, will result in further haemorrhage. Immediate surgery to clamp the aorta above the site of rupture offers the only chance of survival.

Uncommon presentations

The great majority of abdominal aortic aneurysms will present as described above, but it is a common disease and any vascular surgeon will see several cases in their career, presenting in each of the following ways.

Aortic occlusion

Turbulent blood flow occurs in all aneurysms and slow transit of contrast medium is often seen on angiography. Turbulent flow contributes to the formation of the mural thrombus, which is present in most aneurysms. Sometimes the thrombotic process is more extensive and the aorta may occlude. Occlusion usually does not involve the origins of the renal arteries but is frequently accompanied by acute critical ischaemia of the lower limbs.

Distal embolization

Mural thrombus can become dislodged from within the aneurysm, perhaps as a consequence of direct abdominal trauma, and lodge as emboli in the arteries of the lower limb. One or two per cent of all emboli to the lower limb arise from this source.

Ureteric occlusion

Around 10 per cent of abdominal aortic aneurysms are of the 'inflammatory' type, with a variable degree of perianeurysmal fibrosis. One or both ureters can become encased in fibrous tissue and occluded, either by being drawn medially towards the aortic aneurysm or, more commonly, where they cross an 'inflammatory' aneurysm of the common iliac. The patient may present with hydronephrosis or anuria and renal failure.

Aortocaval fistula

This generally occurs in association with aortic rupture into the retroperitoneum, which consequently tends to dominate the clinical picture. In these circumstances the aortocaval fistula is usually only diagnosed peroperatively, when dramatic venous bleeding is seen on opening the aneurysm sac after aortic cross-clamping. Rarely, the aortic aneurysm may rupture only into the inferior vena cava and produce the characteristic clinical picture of venous engorgement and visible arterial pulsation in veins, accompanied by high-output cardiac failure.

Aortoenteric fistula

The majority of aortoenteric fistulas are seen as late complications of aortic surgery, and spontaneous fistulation into the gut from an aorta that has not been operated upon is extremely rare. Primary fistulas usually occur into the duodenum but fistulas in patients with an existing aortic graft tend to be into the first few centimetres of the jejunum and present with haematemesis and melaena. The treatment of this condition is one of the most difficult in vascular surgical practice, since in situ graft contamination and infection are inevitable.

Duodenal obstruction

The fourth part of the duodenum and duodenojejunal flexure is intimately adjacent to the abdominal aorta. A large infrarenal aortic aneurysm may therefore be a cause of external compression of the duodenum and high intestinal obstruction. The symptoms are those of duodenal distension with nausea and vomiting, which tends to be intermittent since the obstruction is incomplete.

Management

Symptomatic abdominal aortic aneurysms usually demand urgent or early treatment. The extent of preoperative investigation, assessment, and medical treatment may therefore need to be curtailed and the patient prepared for surgery as well as possible in the time available. The most immediate need for surgery arises in the patient with a ruptured aneurysm, and this is contrasted below with management of the asymptomatic patient. The management of other symptomatic presentations of the disease will fall somewhere between these two extremes, depending on how compelling is the need for surgery.

Ruptured abdominal aortic aneurysm

The key fact to remember is that these patients are in the process of bleeding to death from the moment rupture occurs. More than half will die within the hour from haemorrhage into the peritoneal cavity, and it is unlikely that these patients could ever be saved. The majority of patients arriving at front-line hospitals will be suffering from some degree of circulatory collapse with hypotension. In this condition, blood transfusion without arresting the haemorrhage is as futile as trying to fill a bucket with a hole in the bottom. The diagnosis should be made from the history and clinical examination. Investigations such as abdominal ultrasonography or radiography are unnecessary, time-consuming, and liable to cause fatal delay. The patient should be transferred immediately to the operating theatre, the only permissible investigation being the taking of a blood sample for cross-matching.

In the operating theatre all preparations for the operation are carried out before the induction of anaesthesia, which should take place only when the surgeon is poised to make the abdominal incision. Anaesthesia is liable to induce severe hypotension as the vasoconstrictor tone that has been maintaining circulation to vital organs is abolished. At this stage, transfusion is given to the extent necessary to sustain essential functions. Only when the aorta above the rupture has been controlled and securely clamped should full transfusion to restore normal blood pressure be given. In a number of patients with ruptured abdominal aortic aneurysm the haemorrhage is so well contained by the surrounding connective tissue that there are no obvious clinical signs of blood loss. Such individuals can survive long journeys to tertiary referral centres and may live for several days before the connective tissue finally gives way and fatal haemorrhage occurs. These patients are liable to be misdiagnosed as suffering from other conditions, of which the most common are ureteric colic, pancreatitis, and sciatica. To establish the diagnosis, ultrasonography or CT may be required. Once the diagnosis is certain, operation is required with appropriate urgency since fatal haemorrhage can occur at any time. It is particularly tragic to see a patient who arrived at the hospital in good condition transferred to the operating theatre in a collapsed state after prolonged delay.

Asymptomatic abdominal aortic aneurysm

The only substantial reason for treating the patient with an asymptomatic abdominal aortic aneurysm is to prevent their premature death at some indeterminate future date from its rupture. At present the only treatment known to reduce this risk is elective surgical replacement of the aneurysmal aorta. A decision to recommend treatment must therefore be based on balancing the operative mortality and morbidity against the risk of rupture. The limited information available on the natural history of abdominal aortic aneurysms shows a general relation between aneurysm diameter and rupture risk. For abdominal aortic aneurysms that remain 4.0 to 5.9 cm in diameter the risk of death from rupture is less than 2 per cent per annum and for diameters above 6.0 cm is of the order of 10 per cent per annum. Since the disease is unlikely to produce any distressing symptoms unless the aneurysm ruptures, it seems unreasonable to ask a patient to accept an immediate operative mortality risk greater than the annual expectation of death from the untreated disease. It is essential therefore that every patient should be carefully investigated and the individual risks of surgery assessed so that an informed judgement can be made in each case.

The United Kingdom small-aneurysm treatment trial has now confirmed that for patients whose aneurysms remain less than 5.5 cm in maximum anteroposterior diameter substantially more lives are lost from elective surgery than from aneurysm rupture. So far this is the only randomized, controlled study to have been completed but a similar study is currently being carried out in Veterans Administration Hospitals in the United States. The annual rupture rate for unoperated aneurysms in the United Kingdom study was less than 1 per cent and the mean elective operation mortality in 841 patients was 6.3 per cent. Operative mortality is 'front loaded', occurring within 30 days of operation, while an annual mortality risk from rupture of the aneurysm is spread over 12 months. In the light of such data a surgeon recommending elective resection for an asymptomatic abdominal aortic aneurysm of less than 5.5 cm in maximum anteroposterior diameter would have some difficulty defending the decision. Surgeons tempted to believe that their operative mortality figures are much less than 6.3 per cent should bear in mind that even if they have had not a single death in their last 50 cases there is only a 95 per cent probability that their overall operative mortality will be less than 6 per cent. A 1 per cent annual rupture risk and a 6 per cent elective operative mortality rate gives a number needed to kill (NNK) for this operation of 20 at 1 year after surgery.

'Inflammatory' abdominal aortic aneurysm

Inflammatory aneurysms comprise around 10 per cent of all abdominal aortic aneurysms encountered in clinical practice but since they are commonly symptomatic, this probably over-represents the prevalence of the inflammatory variant of aortic aneurysms in the entire population. The pathogenesis of this disorder is still the subject of debate, but the original suggestion that the inflammation is a response to leakage of blood from contained aortic rupture is no longer tenable. The macroscopic appearance at operation is of two types: (1) an angry, hyperaemic, periaortic inflammation, or (2) a chronic, fibrotic, icing-sugar aortic wall. Both types may be seen at different points on the same aneurysm. Histologically, the wall of all aortic aneurysms shows evidence of an inflammatory response and the difference between the macroscopically inflamed and non-inflamed aneurysm is quantitative rather than qualitative. The condition is best regarded as a chronic periaortitis and has much in common with idiopathic retroperitoneal fibrosis. Parums and Mitchinson in Cambridge, England, have advanced the theory that the periaortitis is an immune response to antigens, principally ceroid, leaking from atheromatous plaques into the aortic adventitia. It is unclear whether the liberation of lipoproteins from atherosclerotic plaques is simply a consequence of aortic dilatation or a contributory factor to aneurysm formation.

Inflammatory aneurysms may present with symptoms or signs suggestive of the diagnosis, or they may be discovered incidentally during investigation or at operation for an asymptomatic or ruptured abdominal aortic aneurysm. The belief that inflammatory aortic aneurysms are less likely to rupture is not supported by any evidence and should not weigh heavily in management decisions. Even the thickest aortic walls of inflammatory aneurysms tend to be thin posteriorly where they lie in contact with the vertebral bodies, and rupture at this point is not uncommon. The diagnosis of inflammatory abdominal aortic aneurysm should be suspected in patients presenting with a history of abdominal and back pain, and who have a tender but unruptured aneurysm. An elevated erythrocyte sedimentation rate will be present in half of those with an inflammatory aneurysm, and the diagnosis can be confirmed by demonstrating a thickened aortic wall on CT or MRI.

In some patients, one or both ureters may be obstructed by the periaortitis or, more commonly, where they cross an inflammatory iliac aneurysm. Rarely, such patients may first present with renal failure, and the diagnosis of inflammatory aortic aneurysm be made secondarily. Hydronephrosis due to ureteric obstruction is usually best treated before elective aortic surgery, either by ureteric stenting or nephrostomy.

The presence of a stent in the ureter has the additional advantage of providing a useful guide to identification at operation when the ureters are encased in dense fibrosis. In general, following replacement of the aortic aneurysm, the ureteric obstruction will resolve and operative dissection of the ureters to free them from the periaortitis is seldom necessary.

Operative replacement of an inflammatory abdominal aortic aneurysm is difficult but can be satisfactorily performed in most patients by modification of a standard operative technique, since, fortunately, in the majority of instances the neck of the aneurysm is relatively free of periaortitis. Rarely, an elective operation may be too hazardous to continue when the aorta above the aneurysm is also inflamed. In such patients a case can be made for abandoning the procedure and treating for 3 months with systemic steroids to suppress the periaortitis before a further attempt at aneurysm replacement. 'He who fights and runs away lives to fight another day.'

Investigation of the aneurysm

The purpose of these investigations is to determine accurately the size and extent of the aneurysm, to note the thickness of the aneurysm wall and the presence of any localized saccular dilatation, and, finally, to assess the importance of coexistent occlusive or aneurysmal arterial disease elsewhere.

Ultrasonography of the abdominal aorta is the first-line investigation; its advantages are that it is cheap, freely available, accurate, reliable, and reproducible. Being non-invasive it can be repeated as often as required, either to confirm the original findings or to monitor growth of the aneurysm. Its main disadvantages are that it is observer-dependent and the permanent images produced can be difficult for anyone, other than the person who performed the scan, to interpret. Visualization of the suprarenal aorta and iliac arteries is often difficult and the study may be impossible in the grossly obese or when large amounts of bowel gas are present. It remains, however, the most useful investigation for measuring the diameter of an infrarenal aortic aneurysm. The maximum anteroposterior diameter measured by ultrasonography is the reference standard for judging the rupture risk of an aneurysm and the need for elective aortic surgery. CT, MRI, and transverse-diameter measurements all tend to overestimate the maximum diameter of an aneurysm and patients should not be recommended for surgical treatment on the basis of such measurements alone.

CT produces excellent permanent records of cross-sectional anatomy that are easy to interpret. Its main uses are to discover the extent of any suprarenal aortic involvement and the thickness of the arterial wall, in order to detect 'inflammatory' aortic aneurysms. It can also be used to measure aortic diameter, but inaccuracies can occur if the section is not at right angles to the long axis of a tortuous aneurysmal aorta. An element of tortuosity is present in most aneurysms and transverse or oblique diameters are inherently unreliable. The anteroposterior diameter will usually be less and will more accurately reflect the true maximum diameter. Spiral CT allows a three-dimensional image of the aorta to be constructed so that a diameter at right angles to the long axis of the aneurysm can be measured. This technique overcomes some of the disadvantages of conventional CT but it has yet to be established that diameters measured by spiral CT correspond as closely to true diameters as those obtained by ultrasonography.

MRI is now available in most centres, and the quality and definition of the images produced by the newer machines is excellent. Since there is no radiation exposure it could well replace CT in many of its present uses.

Angiography is used mainly to discover the relation of the renal arteries to the aneurysm and, by outlining the kidneys, may reveal relevant abnormalities, such as horseshoe or pelvic kidneys. When the aneurysm is large and blood flow turbulent or sluggish, it is sometimes difficult to obtain high-quality angiograms of the limb vessels from aortic injection of contrast, but this information can be of help in planning the extent of any arterial surgery required. The presence of iliac or femoral aneurysms or occlusive arterial disease in iliac, femoral, or more distal arteries of the limbs may be revealed. Routine angiography rarely furnishes information that alters the management of patients with abdominal aortic aneurysm. In general the risks of the procedure now outweigh any benefit obtained and its routine use should be abandoned.

Investigation of the patient

The purpose of these investigations is to discover how well the patient is likely to tolerate the trauma of a major arterial operation. Of particular importance are cardiac and respiratory function, and the presence of carotid arterial or other coexistent disease.

Attention is paid in the history to symptoms of angina, breathlessness at rest or on exertion, and previous myocardial infarction. All patients should have routine monitoring of their blood pressure and an electrocardiogram. Hypertension should be controlled and, if the history or electrocardiogram suggests possible abnormalities of myocardial function or blood supply, further investigations are essential. Echocardiography is useful for detecting abnormalities of valve or heart-wall function, and the technique of multigated acquisition nuclear imaging allows a ventricular ejection fraction to be calculated at rest and after exercise. In some patients, coronary angiography will be indicated, and any coronary arterial disease discovered may need treatment by angioplasty or bypass grafting before aortic surgery can be safely undertaken. Evidence of recent myocardial infarction is an important reason to recommend delaying elective surgery for all but the largest aneurysms, since the chances of further myocardial infarction and death are substantially increased by operation within 6 months of the infarct. Other relative contraindications to surgery are a low ventricular ejection fraction at rest or one which falls markedly on exercise, indicating inadequate blood supply to the myocardium.

Routine measures of respiratory function, such as peak expiratory flow and spirometry, are simple to perform as an extension of the normal clinical examination and should be a standard part of the assessment of all patients. More complex measurements of gas exchange are rarely necessary but, when required, the services of a respiratory function laboratory may prove helpful. Although poor lung function may be a contraindication to elective surgery, the presence of chronic obstructive airways disease is an important risk factor for aneurysm rupture. With the assistance of a chest physician, most patients can be improved to the point where the risks of elective surgery become acceptable.

The presence of severe, asymptomatic, carotid arterial stenosis presents a more difficult problem to resolve, and debate still goes on about whether carotid endarterectomy should be performed before, during, or after surgery for aortic aneurysm, or indeed, whether it should be performed at all. Each case will need to be resolved on its merits, depending on the relative importance of the aortic aneurysm and carotid stenosis in the individual patient, but, in general, the patient with an asymptomatic stenosis of the internal carotid artery is not greatly at risk of having a stroke during aortic surgery and carotid endarterectomy in such patients is not supported by the evidence available.

Because the majority of patients with an abdominal aortic aneurysm are old, coexistent disease is common. Malignant disease discovered during investigation for the aneurysm presents a particular problem. It is difficult to be dogmatic about the treatment priority, but a rationale for therapy is outlined here. Generally, primary treatment for the cancer is given first, since delay is liable to progressively reduce the chances of cure while, on the other hand, provided rupture has not occurred, a large aneurysm is as easily replaced as a small aneurysm. If primary treatment of the cancer seems to have been successful, then the aortic aneurysm is replaced as soon as the patient has recovered from cancer surgery. If treatment of the cancer is definitely non-curative and life expectation is limited, aneurysm surgery is seldom advised, since in most patients death from rupture of the aneurysm is likely to preferable to death from carcinoma.

Operative mortality and morbidity

In the decade after the first recorded replacement of an abdominal aortic aneurysm in 1951, operative mortality was high, at around 15 per cent. Over the past 0 years, as a result of refinements in surgical and anaesthetic technique and better pre- and postoperative management, elective operative mortality has been consistently reduced to under 5 per cent in some vascular units. It would be a mistake to assume that the low mortality currently reported by some specialist units represents the common experience, and in many hospitals a figure of 10 per cent or more is not unusual. It has been shown that reported operative mortality is higher when data are collected prospectively rather than retrospectively, and higher still when prospective data are collected in the community rather than in the hospitals where the surgery is performed. Prospectively collected community data indicate that the mean 30-day operative mortality of elective surgery for aortic aneurysm is around 8 per cent. Approximately 1 in 20 abdominal aortic aneurysms extends close to or above the origin of the renal arteries. Surgery in these cases may involve clamping the aorta above the renal arteries and sometimes their reimplantation. Operating on suprarenal aneurysms requires special skill and techniques, and inevitably is associated with greater hazard than for the uncomplicated infrarenal aneurysm.

Elective aortic surgery remains a major operation and, even in the uncomplicated case, morbidity is considerable. Most patients can be discharged from hospital within 10 days of operation, but few will be restored to complete well-being in less than 2 months. Currently, there is debate about the long-term outlook for those who have undergone successful replacement for aortic aneurysm. Some follow-up studies have shown a similar life expectation to that of the general population of the same age. This conclusion is disputed by others and seems to be inherently improbable given the known association of abdominal aortic aneurysm with a number of other diseases that impair life expectancy. There is no doubt, however, that the patient who survives surgical replacement of their aneurysm has a greater life expectation than one whose aneurysm is left untreated.

Operative technique

As with any operation, there are many variations in technique used by individual surgeons for routine operations, together with specific variations which may be employed to deal with special situations encountered. The standard techniques used by the author successfully over many years are described below and brief notes on useful or alternative techniques are appended after the main account.

Elective replacement

Preparation for the operation begins days or sometimes weeks before, to ensure that all the clinical information required is obtained and the patient is in the best state of health achievable at the time of operation. The main hazard peroperatively is sudden change in circulatory haemodynamics as a consequence of clamping and unclamping of the aorta, or blood loss. It is therefore essential that adequate monitoring of cardiac function and intravascular volume is in place before surgery begins.

The patient is placed supine on the operating table and the skin is prepared with antiseptic from the nipples to the knees—particular care must be taken to ensure cleaning of the external genitalia. The operation field extends from the xiphisternum to mid-thighs, with the genitalia being securely excluded by towelling and the use of adherent skin drapes. A vertical, midline abdominal incision is made from sternum to symphysis pubis and the peritoneal cavity opened. A complete inspection of all the intra-abdominal organs is made to exclude other pathology. The small intestine is retracted to the right and draped from the operative field, usually being retained within the abdomen, but in an obese patient better exposure is obtained if the intestine is exteriorized within a plastic 'gut' bag.

Minimal dissection of the retroperitoneum is employed to limit bleeding and it is unnecessary to mobilize adherent duodenum. The neck of the aneurysm is identified by palpation, and the overlying peritoneum and fascia divided in the midline until the aorta is exposed. The inferior mesenteric vein is displaced to the left and seldom needs to be divided. Midline dissection continues until the left renal vein is identified, and fascial division is continued transversely at the lower border of the vein to free it and allow its retraction if required. Blunt dissection on both sides of the aorta in a strictly vertical plane continues until the vertebral body is encountered. Intravenous heparin is administered and the neck of the aorta clamped anteroposteriorly.

When the common iliac arteries are not aneurysmal, their dissection is easily accomplished by division of peritoneum and fascia over their anterior surfaces with blunt finger dissection down each side. The vessels are clamped anteroposteriorly. Common and internal iliac arteries may be aneurysmal and, in these circumstances, the external iliac arteries are clamped and back-bleeding from the internal iliac arteries controlled after the aortic aneurysm is opened.

The aortic aneurysm is inspected and the inferior mesenteric artery is oversewn with a transfixion suture at its origin. The aneurysm is then incised in the midline from its neck to the aortic bifurcation, and the mural thrombus evacuated. At each end of the incision, transverse scissors cuts are made so that half the circumference of the aorta is divided. Bleeding from lumbar vessels is controlled by direct pressure until permanently arrested by oversewing with transfixion sutures. At this stage the operative field should be bloodless and the ends of the aorta can be inspected and cleared of adherent thrombus.

In 80 per cent of cases a tube graft can be used, but where the iliac arteries are aneurysmal a bifurcated 'trouser' graft will be required. In the latter case it is preferable, and usually satisfactory, to anastomose each limb of the graft either to the termination of the common iliac artery or to the external iliac artery. It is desirable to retain circulation into at least one internal iliac artery to minimize the risk of ischaemia in the gut or spinal cord. The graft is stitched to the neck of the aneurysm, using an inlay technique and a monofilament prolene continuous suture. Exposure is improved by inserting a self-retaining rake retractor within the aneurysm sac. Three sutures are placed on each side of the midline of the aorta and graft posteriorly. and the graft is 'parachuted' into place. The suture is continued on each side to the midline anteriorly, particular care being taken to place accurately the corner sutures in the lateral walls of the aorta. The anastomosis is tested for leaks at this stage, since subsequent haemorrhage from the posterior wall is more difficult to deal with.

The tube graft is cut to the appropriate length and the distal anastomosis made in exactly the same fashion. A vital step in the procedure is to ensure that no particulate material, atheroma, thrombus, or tissue remains inside the graft or iliac arteries proximal to the arterial clamps. Graft and proximal iliac arteries are therefore irrigated thoroughly with saline before the anastomosis is completed. It is unnecessary to release the distal arterial clamps to achieve this end, and doing so may precipitate arterial thrombosis by exposing blood in the distal vessels to tissue thromboplastin from the operation site.

Five minutes before the anastomoses are completed the anaesthetist is warned that restoration of circulation to the legs is imminent so that circulatory volume can be appropriately and rapidly augmented when required. One distal clamp only is removed and the aortic clamp slowly released. This is a dangerous phase of the operation, and flow through the graft is titrated against the patient's blood pressure and heart filling pressure. Only when pressures are normal, with full restoration of blood flow to one limb, is the second distal clamp slowly released.

The operation is completed by meticulous haemostasis, and reversal of anticoagulation may be necessary to achieve this end.

Technical variations

Transverse incision

Many surgeons use a transverse, upper abdominal, dome-shaped incision for abdominal access. The muscles of the abdominal wall are cut in the line of the incision, which commences midway between the umbilicus and sternum and runs parallel to the costal margins. It is claimed that postoperative respiratory complications are reduced by this incision but it is more time-consuming to make and close, bleeding is greater, and access to the iliac arteries is difficult.

Retroperitoneal approach

The patient is positioned corkscrewed on the operating table with the pelvis horizontal and the shoulders vertical. An incision is made from the tip of the left twelfth rib to the midline below the umbilicus. Muscles of the abdominal wall are cut in the line of the incision. The extraperitoneal plane is identified and the peritoneum retracted medially to expose the aorta. The left kidney may be retracted with the peritoneum or allowed to remain lying on the psoas muscle. Advocates of this approach point to the facility of the technique in obese patients and claim reduced postoperative morbidity. The disadvantage is that exposure of the right common iliac artery cannot readily be achieved.

The perirenal aneurysm neck

Aneurysms extending substantially above the origins of the renal arteries are discussed in the section on thoracoabdominal aortic aneurysms, but aneurysms with a neck at, or immediately above, the origin of the renal arteries can be dealt with by slight modification of the operative approach to the common infrarenal aortic aneurysm. The problem may be encountered unexpectedly at operation in a patient who was thought on preoperative assessment to have a purely infrarenal aneurysm.

The aorta needs to be dissected and clamped above one or both renal arteries. To achieve this exposure the left renal vein must be freed from the aorta and retracted either superiorly or, occasionally, inferiorly. It will be necessary to divide either the left gonadal vein or sometimes the left adrenal vein to permit safe retraction, but division of the renal vein itself is rarely required. The anastomosis between graft and aorta is accomplished expeditiously, with the renal arteries being incorporated within the proximal aorta as a single or double, short tongue. After completion of the proximal anastomosis, the clamp is reapplied to the graft below the renal arteries and blood flow to the kidneys restored before attention is turned to the distal aortic anastomosis.

It is claimed that the retroperitoneal approach allows the perirenal aortic aneurysm to be dealt with as easily as the infrarenal aortic aneurysm and is part of the advocacy for general adoption of this approach.

Types of graft

In the early days of aortic surgery, aortic homografts were widely used but were abandoned because their use was inconvenient and they were prone to aneurysmal degeneration. They have been obsolete for 50 years, but have been reintroduced in the last 20 years and advocated for use in the presence of established infection of a synthetic graft. Synthetic grafts in common use are of woven or knitted Dacron or polytetrafluoroethylene. Woven Dacron is stiffer and less permeable than the knitted material.

The problem of blood leaking through the graft at operation can be overcome by coating the knitted graft in various ways, but this adds to the cost disadvantage compared with the woven material. Polytetrafluoroethylene grafts are impermeable but suffer from two disadvantages. First, the aortic body of a trouser graft must be cut to the correct length with minimal tolerance if the legs are to lie at an acceptable angle. Secondly, polytetrafluoroethylene is prone to leak at stitch holes for a prolonged time, a problem which is increased when a large needle is used.

If a trouser graft is essential, anastomosis of the limbs to the iliac arteries is preferable, since this avoids additional incisions in the groins and consequently reduces the risk of contamination of the graft and graft infection.

The external iliac arteries invariably and mysteriously remain uninvolved by aneurysmal change, although they may become occluded by atherosclerosis.

Clear indications for aortofemoral grafting are the presence of large femoral arterial aneurysms and stenosis or occlusion of the external iliac arteries.

Replacement of the ruptured abdominal aortic aneurysm

Two-thirds of ruptured aneurysms leak either directly or secondarily into the peritoneal cavity within an hour or so and cause death from exsanguination. These patients usually die at home or on their way to hospital and thus do not come to surgery. Successful surgery is possible only because of temporary tamponade of the rupture by tissue pressure in the retroperitoneum, assisted by hypotension. As described above, these patients are on the brink of death and sustain fatal haemorrhage when the peritoneum ruptures, tissue tamponade fails, or the blood pressure is increased by injudicious transfusion before the rupture is secured. Most lives are saved by immediate transfer from emergency room to the operating theatre, but operative mortality rates improve as the time between rupture and surgery increases. The time delay associated with interhospital transfer selects for treatment those with stable, contained ruptures. Those with poorly contained ruptures die before referral or in transit.

The patient is prepared for surgery while conscious on the operating table. Induction of anaesthesia should occur only when the surgeon is poised ready to make the incision. Blood pressure should be maintained at no more than 100 mmHg systolic until the aorta is controlled. This ideal may require rapid transfusion in order to maintain cerebral and myocardial perfusion on anaesthetic induction when compensatory vasomotor tone is suddenly relaxed. On opening the peritoneal cavity the posterior peritoneum is exposed by displacing the small intestine. The aorta above the aneurysm is identified by palpation and occluded by direct compression with the surgeon's left hand or a 'snake catcher' instrument against the vertebral bodies. Only when the aorta has been securely occluded by compression should the posterior peritoneum over the upper part of the aneurysm be incised. The wall of the aneurysm should be exposed by sharp dissection with scissors through the haematoma and connective tissue, and only when the aortic adventitia is clearly identified is it permissible to use blunt finger dissection in the periadventitial plane to clear the neck of the aneurysm. When the neck of the aneurysm is securely clamped the 'snake catcher' is removed. The operation then proceeds as for an elective replacement.

Death from ruptured aneurysm is a direct consequence of blood loss. Patients will rarely survive if additional blood loss is caused by extensive, unnecessary, hasty, or careless dissection. Iatrogenic blood loss is most likely to occur from tearing of the left gonadal, left adrenal, or left renal vein by attempts to dissect the aneurysm neck that are too hasty, too wide, and too high. Control of blood flow into the aneurysm must be achieved before any dissection is attempted.

High aneurysm rupture

When rupture of the aneurysm occurs close to its neck, it may be difficult to dissect the neck to clamp it. In this situation the aortic neck can be identified from within the lumen of the aneurysm and a large Foley urethral catheter inserted. The balloon of the catheter is inflated in the neck of the aneurysm and tamponade of the aorta against the vertebral bodies is slowly released to confirm that the balloon is securely impacted. Careful, external, two-handed dissection of the neck of the aneurysm can then be performed and the balloon catheter replaced by an aortic clamp.

Endoluminal surgery

The first endoluminal repair of an abdominal aortic aneurysm was reported in 1991. Since then the technique has been employed on several hundred patients throughout the world using a large number of different home-made, customized, and commercially produced devices. The feasibility of endoluminal surgery in carefully selected patients is unquestionable but technical failures are common and the effectiveness of the procedure in providing useful prophylaxis against rupture of an aortic aneurysm is unproved.

Types of endoluminal graft

Aortic tube graft

Eighty per cent of open, elective repairs of aortic aneurysms can be accomplished with tube grafting. The avoidance of bifurcated grafts whenever possible has contributed to the reduction in operative mortality and morbidity and chronic graft infection. The first commercial grafts for endoluminal insertion were designed with the hope of retaining these advantages. Unfortunately, endoluminal aortic tube grafts are suitable for only a very small minority of aortic aneurysms since a caudad, normal-diameter aortic cuff of at least 10 mm long above the origin of the common iliac arteries is required for their insertion. In the rare instances when a suitable cuff is present, graft insertion is often technically unsuccessful because of the need to select the exact length of graft required if distal graft fixation without kinking or iliac arterial occlusion is to be achieved.

Aorto-bi-iliac graft

Endografts from the aorta to both iliac arteries are either of a one-piece or modular type. With one-piece devices a guidewire attached to the contralateral iliac graft limb is first passed from the ipsilateral groin into the aorta, to be captured and withdrawn by a snare passed from the contralateral femoral artery. With modular devices the ipsilateral graft limb is a full-length trouser leg and the contralateral limb a short trouser leg. The short trouser leg is deployed inside the aortic aneurysm and a long trouser-leg extension is invaginated into it from the contralateral femoral artery.

Deployment of the distal end of the iliac graft limbs may be into the common or external iliac arteries. Usually the common iliac arteries are chosen to allow perfusion of the internal iliac arteries and minimize the risk of retrograde blood flow between the graft and lumen of the aneurysm.

Aorto-unilateral graft with extra-anatomic bypass

Aorto-bi-iliac graft insertion is often technically difficult and some patients have only unilateral iliac arteries suitable for endoluminal graft deployment. A customized technique for such patients is to insert a tapered aorto-uni-iliac graft from the most suitable limb. The common iliac artery in the contralateral limb is then occluded and blood flow re-established with a femorofemoral cross-over graft using a conventional, open surgical technique. An alternative procedure is to first occlude the ipsilateral internal iliac artery and insert a tapered aortofemoral endoluminal graft. The distal limb can then be appropriately tensioned, cut to length, and sewn directly to the femoral artery. The operation is similarly completed by occlusion of the contralateral common iliac artery and insertion of a femorofemoral cross-over graft.

Anatomical constraints on endoluminal surgery

The morphology of the non-aneurysmal aortic cuffs, the aortic and iliac aneurysms, and non-aneurysmal iliac and femoral arteries can all impose constraints on the use of endoluminal grafts. Some of the constraints are important enough to be regarded as absolute by most surgeons, others are relative contraindications that preclude the use of some types of graft or increase the risk of technical failure.

All endoluminal grafts for aortic aneurysms require an aortic neck below the origin of patent renal arteries. The neck should be free of mural thrombus, have a maximum diameter of 28 mm and ideally a length of at least 15 mm. The constraint of a maximum diameter was initially imposed by the calibre of deployment system that can usually be introduced through the femoral or iliac arteries. A more fundamental problem is that large-diameter aortic necks are already aneurysmal or ectatic and soon expand away from the graft fixation system, with the development of proximal leaks. Similarly, mural thrombus in the neck is indicative of pathological dilatation, and has the additional disadvantages that initial graft fixation is seldom possible and peripheral embolization of dislodged thrombus is likely. The neck-length restriction can be reduced if an uncovered graft-fixation system is deployed at or above the origin of the renal arteries. The diameter of the wire used for the fixation system is less than the diameter of most renal arterial orifices. The random chance of a narrow orifice coinciding with the soldered junction of two wires is such that renal infarction does not usually occur. The technique of suprarenal graft deployment is for surgeons who are prepared to play Russian roulette with their patients' renal function.

Aortic tube-graft insertion requires a distal aortic cuff of 28 mm maximum diameter and 10 mm minimum length. The presence of such a cuff in aneurysms of clinical significance is unusual and fewer than 5 per cent of endografts currently inserted are aortic tube grafts.

Aortoiliac grafts require that at least one common iliac artery should have a minimum diameter of 8 mm and maximum diameter of 14 mm, and be at least 25 mm long. If the deployment system is introduced from the groin, the minimum-diameter restrictions also apply to the femoral and external iliac arteries. An acute angle at the junction of common iliac artery with the aorta can often be increased by traction on the iliac artery. Persistent angulation of less than 90° can prevent introduction of the deployment system into the aorta.

Mortality and morbidity of endoluminal aortic graft insertion

Thirty-day mortality in published reports from individual series of endoluminal aortic aneurysm exclusion varies from none to 28 per cent with a mean of 5 per cent. It is certain that the true overall operative mortality is currently substantially higher because of reporting bias and the fact that most surgeons and radiologists have performed few procedures, and are on a learning curve for the technique.

National and international registries of endoluminal aneurysm surgery have been established and these have recorded operative mortality of around 10 per cent. Even this figure is more likely than not to under-represent the true mortality, since reporting of cases remains voluntary. It is likely that as case selection becomes more refined and technical expertise improves the operative mortality will be reduced.

As with all major surgery, patients undergoing endoluminal exclusion of aortic aneurysm are at risk of developing cardiac, respiratory, renal, and multiorgan failure. There are also specific problems that are more likely to be associated with endovascular aneurysm surgery. The most common of these appears to be persistent postoperative pyrexia, the cause of which is uncertain. Others are arterial access and instrumentation injuries, thromboembolic events, groin lymph leaks, and endograft or extra-anatomic graft infection.

Early and late technical failure

The reported initial technical success rate for endoluminal aneurysmal exclusion, defined as correct placement of the graft without perigraft blood flow (endoleakage, see next), death or graft occlusion within 30 days of implantation, ranges from 48 to 95 per cent. The most common technical problem is blood flow between the deployed endograft and the aneurysm, now commonly called an endoleak. Endoleaks occur at either the proximal or distal implantation sites or from patent lumbar, inferior mesenteric or iliac arteries. For an endoleak to be demonstrable and persist there must be both an inflow and an outflow for the leaking blood. Some small, peroperative endoleaks seal spontaneously. Others may be closed secondarily by further insertion of an endoluminal stent either at the time of initial surgery or subsequently.

Large peroperative endoleaks, failed graft deployment, or other major technical problems require conversion to open aneurysmal repair. The need to convert to conventional aneurysm surgery varies with the technical expertise of the operator and the difficulty of the procedure attempted. It is generally accompanied by a high operative mortality.

Some early endoleaks are either unrecognized at the time of operation or appear subsequently. It may be possible to seal them by further endoluminal graft insertion. A persistent endoleak means that the operation has failed, aneurysm expansion will continue to occur, and the patient remains at the same risk of aneurysm rupture as there would have been had the operation not been performed.

Late endoleaks occur either as a consequence of graft material failure or continued expansion of the aorta or iliac arteries at the site of graft fixation. Most graft failures to date have been of the metal fixation devices, particularly of fixation hooks, due to metal fatigue. The fabric of some endografts has been chosen as much for its low bulk as its strength and is likely to be less durable than conventional graft fabric. The most fundamental problem, however, is that of continued expansion of the aorta at the neck of the aneurysm. Patients with aortic aneurysms usually have generalized arterial ectasia and continued arterial expansion tends to occur with advancing age. Some surgeons believe that most, if not all, endovascular grafts will become too small as aortic dilatation progresses, with consequential late graft displacement or the development of endoleaks and continued aneurysm expansion.

Does endovascular aneurysm exclusion provide useful prophylaxis against rupture?

Substantially the sole purpose of elective surgery for aortic aneurysm is to provide prophylaxis against unnecessarily premature death from rupture. Net benefit in life years gained is achievable when the death rate from the natural history of the disease is high and when the operative mortality of an effective, durable, prophylactic operation is low. When the operative mortality is the same as the annual mortality from the natural history of aortic aneurysm rupture a net gain in life years will begin to accrue from 2 years after treatment if the operation provides total protection from rupture. Endovascular exclusion of aortic aneurysm is most likely to be technically successful in small aneurysms at little risk of rupture. When used for large aneurysms, early technical failure is common and delayed technical failure from endoleaks is probable. At present, endovascular surgery for aortic aneurysm is a technically challenging and exciting innovation of no established therapeutic value. It remains to be seen whether it will become an effective means of providing prophylaxis against rupture of an aneurysm in some patients.