Abdominal Aortic Aneurysm

Bandolier has been asked by a GP whether there is enough evidence about the effectiveness of screening for abdominal aortic aneurysm (AAA) for it to be introduced into his practice. To try and answer this we searched MEDLINE from 1993 to the present to see if there were any new reports which helped. There was no single source of information which brought this subject together, but we did find some interesting papers.

RCT for AAA

A recently published English randomised trial in Chichester [1] has tried to examine the incidence of rupture following AAA screening. It was the most useful report and worth examining in some detail.

The setting was 15,775 men and women aged 65-80 years identified from GP registers and FHSA lists. These people were then randomised by computer to control and screening groups. Those randomised to screening were invited for abdominal ultrasonographic scanning by letter from their family practice; one reminder was sent if there was no reply.

Aneurysm was defined as maximum aortic diameter of 3 cm or more. Annual re-scanning took place if the diameter was 3 - 4.4 cm, and at three-monthly intervals if it was 4.5 - 5.9 cm. This protocol was continued until February 1994 or until the patient died, underwent surgery or declined further follow-up. Aortic diameters of 6 cm or more, and increase of diameter of 1 cm or more in a year, or developments of symptoms attributable to the aneurysm all constituted criteria for considering surgery.

Screening and acceptance

The population screened was:

	Men	Women
Screened	3205	4682
Control	3228	4660

Of those invited for screening, 5394 (68%) accepted. The acceptance rate declined with increasing age in both men and women, but refusal was always higher in women. About 30% of men and 38% of women in their eighth decade refused.

Prevalence of aneurysm

The prevalence of aneurysm in the screened population was higher in men than women, and highest (at about 9%) in men in their eighth decade.

The distribution of aortic diameter was about the same in men and women: about 68% were below 4 cm, 88% below 5 cm and about 8% were 6 cm in diameter or greater.

Outcome in controls

The 5-year mortality in the control population was 12.5%. Twenty men in the control group presented with aortic rupture, of whom 17 died within one year. Two women had rupture, and died within one year.

Seven patients (five men) had an AAA that was detected clinically and treated routinely by operation; none died within one year of the operation.

Death from ruptured aortic aneurysm according to registrar returns over five years in the same health authority are shown by age and sex in the figure below. They occur predominantly in men and women over 70 years of age.

Outcome in screened population

The 5-year mortality in the screened population was 13.1%. There were four ruptures in people who refused scanning, all of whom died without surgery. In the screened population there were 4 ruptures who died without surgery, and 31 people had surgery (3 emergency); 29 survived to one year.

Ruptured aortic aneurysm

The incidence of ruptured aortic aneurysm in men over the follow-up period of five years showed a fall of more than 50% (9 of 3205 in scanned plus refused scan versus 20 of 3228 in control). In women there was no difference (3 of 4682 in scanned plus refused scan versus 2 of 4660 in control).

Comment

This is a commendable study, and there is much more information than can be summarised here. The authors comment that the number of cases of ruptured AAA presenting for surgery fell from 15 cases in 1992/3 to 5 cases in 1994/5. They conclude (rightly) that although their study seems positive, a much larger multi-centre study would be needed to demonstrate unequivocal benefits or cost-effectiveness.

Natural history

In any screening programme proposal, the natural history of the disorder is a necessary background. The Chichester RCT provides that for controls. A Swedish study [2] followed 88 patients with AAA found on screening and followed prospectively by repeated ultrasonography.

An aneurysm here was defined as aortic diameter of >1.5 cm, though in 19 patients it exceeded 3.9 cm. In a follow-up period of up to five years, 38 of the 88 patients died: no patient died from ruptured AAA. Mortality in patients with AAA was higher than that in a control population.

Cost effectiveness of screening

Never an easy calculation, especially without good data on effectiveness. A group from Rochester and McMaster [3] did try and perform such an analysis using a computer model to simulate the costs and effectiveness of various screening protocols in men, covering a period of 20 years.

Using a protocol involving abdominal palpation as a "screen" with ultrasound confirmation was estimated to gain 20 life-years in a 10,000 man cohort (60-79 years) at a cost of US$28,700 per life-year. A single ultrasound screen gained 57 life-years at a cost of $41,500 per life-year (1993 estimates).

Comment

Screening for AAA would seem to fall clearly into the 'not proven' category.

References:

RAP Scott, NM Wilson, HA Ashton, DN Kay. Influence of screening on the incidence of ruptured abdominal aortic aneurysm: 5-year results of a randomized controlled study. British Journal of Surgery 1995 82: 1066-70.
H Bengtsson, P Nilsson, D Bergqvist. Natural history of abdominal aortic aneurysm detected by screening. British Journal of Surgery 1993 80: 718-20.
PS Frame, DG Fryback, C Patterson. Screening for abdominal aortic aneurysm in men ages 60 to 80 years: a cost-effectiveness analysis. Annals of Internal Medicine 1993 119: 411-6.

previous or next story in this issue