Benefits and harms of screening men for abdominal aortic aneurysm in Sweden: a registry-based cohort study

Abstract

Background: Large reductions in the incidence of abdominal aortic aneurysm (AAA) and AAA-related mortality mean that results from randomised trials of screening for the disorder might be out-dated. The aim of this study was to estimate the effect of AAA screening in Sweden on disease-specific mortality, incidence, and surgery.

Methods: Individual data on the incidence of AAA, AAA mortality, and surgery for AAA in a cohort of men aged 65 years who were invited to screening between 2006 and 2009, were compared with data from an age-matched contemporaneous cohort of men who were not invited for AAA screening. We also analysed national data for all men aged 40-99 years between Jan 1, 1987, and Dec 31, 2015, to explore background trends. Adjustment for confounding was done by weighting the analyses with a propensity score obtained from a logistic regression model on cohort year, marital status, educational level, income, and whether the patient already had an AAA diagnosis at baseline. Adjustment for differential attrition was also done by weighting the analyses with the inverse probability of still being in the cohort 6 years after screening. Generalised estimating equations were used to adjust the variance for repeated measurement and in response to the weighting.

Findings: AAA mortality in Swedish men has decreased from 36 to ten deaths per 100 000 men aged 65-74 years between the early 2000s and 2015. Mortality decreased at similar rates in all Swedish counties, irrespective of whether AAA screening was offered. After 6 years with screening, we found a non-significant reduction in AAA mortality associated with screening (adjusted odds ratio [aOR] 0·76, 95% CI 0·38-1·51), which means that two men (95% CI -3 to 7) avoid death from AAA for every 10 000 men offered screening. Screening was associated with increased odds of AAA diagnosis (aOR 1·52, 95% CI 1·16-1·99; p=0·002) and an increased risk of elective surgery (aOR 1·59, 95% CI 1·20-2·10; p=0·001), such that for every 10 000 men offered screening, 49 men (95% CI 25-73) were likely to be overdiagnosed, 19 of whom (95% CI 1-37) had avoidable surgery that increased their risk of mortality and morbidity.

Interpretation: AAA screening in Sweden did not contribute substantially to the large observed reductions in AAA mortality. The reductions were mostly caused by other factors, probably reduced smoking. The small benefit and substantially less favourable benefit-to-harm balance call the continued justification of the intervention into question.

Funding: Research Unit and Section for General Practice, FoUU-centrum Fyrbodal, Sweden, and the region of Västra Götaland, Sweden.