Abdominal aorta aneurysm (AAA) screening of men aged 65
Health technology assessment
|Published
In this health technology assessment (HTA) we have assessed the clinical effectiveness of screening all 65-years old men for AAA in Norway by updating an earlier HTA report from the Norwegian Institute of Public Health. We have performed a health economic analysis by adapting into Norwegian setting an existing Markov model elaborated for Sweden.
Key message
Abdominal aortic aneurysm (AAA) is a weakening of the aortic wall resulting in an abnormal dilatation of the abdominal aortic artery. AAAs bears the risk of rupture, which is a dramatic emergency condition with a high risk of death. In Norway, about 1% of all deaths among men older than 65 years are caused by AAA rupture. The mortality rate due to AAA rupture is approximately 75%, with about half of the deaths occurring before the patient reaches the hospital. The larger the AAA, the higher is the risk of rupture. Prevalence of AAA among men varies across countries from approx. 1.5-3%. In Norway, it is estimated to be around 2.5%. Prevalence among women is one-sixth to one-fourth of that among men. Screening programs for AAA can help to identify patients with a high risk of AAA rupture. Those identified as suitable for repair, usually by ultrasound scan, are offered preventive (elective) surgery to reduce their individual risk of rupture. In Norway, the number of operations (urgent and elective) is approximately 700-800 per year.
AAA screening is considered a potentially beneficial healthcare intervention in several European countries. These countries have therefore introduced AAA screening programs into their public health service. Other countries, including Norway, currently have no systematic nationwide screening program.
In this health technology assessment (HTA) we have assessed the clinical effectiveness of screening all 65-years old men for AAA in Norway by updating an earlier HTA report from the Norwegian Institute of Public Health. We have performed a health economic analysis by adapting into Norwegian setting an existing Markov model elaborated for Sweden. Main conclusions are:
- AAA screening is likely to have a beneficial effect in men aged 65, as available evidence shows that AAA-related mortality can be reduced by half in both the short- and long-term.
- Evidence shows a non-significant reduction in overall mortality in the short term, however in the long-term AAA-screening may decrease overall mortality.
- Number of preventive operations increases by 2-3 fold, while number of urgent operations is reduced by half due to AAA-screening.
- The health economic analysis resulted in costs per gained life-year of about 154,000 NOK (ICER), and shows that AAA screening will lead to 62 avoided AAA related deaths per year.
Budget impact is estimated to be around 20 million NOK per cohort (approx. 28,000 men) per year.
Summary
Introduction
Abdominal aortic aneurysm (AAA) is a dilatation of the abdominal aortic artery, with a diameter enlarged by at least 50%, corresponding to a diameter of at least 30mm. AAAs bear the risk of rupture, which is a dramatic emergency condition with a high risk of death, and requires urgent surgery. The larger the AAA, the higher is the risk of rupture. AAA ruptures cause 1-2% of all deaths in the Western world. In Norway, approximately 1% of all deaths among men aged 65 and above result from AAA. The mortality rate due to AAA rupture is approximately 75%. About half of the patients die before reaching the hospital, while the death rate is 30-50% among those who reach the hospital. High age, male gender, smoking, hypertension, cardiovascular disease, and family history of AAA are factors associated with increased risk of developing AAA. Prevalence of AAA among men varies across countries from approximately 1.5-3%. In Norway, it is estimated to be around 2.5 % among men based on data from the Norwegian Vascular Surgery Registry (NORKAR). In Oslo, prevalence is 2.6-2.7% (1). Prevalence among women is one-sixth to one-fourth of that among men. Although AAA prevalence has been increasing in some countries, others have experienced declining rates.
Several European countries consider AAA screening to be a potentially beneficial healthcare intervention, and have introduced AAA screening programs in their public health services. Other countries, including Norway, currently have no systematic nationwide screening program. Individuals identified as having a high risk of AAA, usually through ultrasound examination, receive a recommendation of preventive surgery or so-called elective repair to reduce the risk of rupture. The repair can be performed using open or endovascular surgery (stent graft).
In Sweden, the national AAA screening program has, to date, identified 6,000 AAAs in men, of which 1,500 have had elective repair. The program saves approximately 100 men annually. Number needed to screen (NNS) and number needed to operate (NNO) to prevent 1 premature death are, respectively, 677 and 1.5. Screening is considered to be cost-effective in Sweden, with a gain of 577 quality-adjusted life years (QALY) and an incremental cost-effectiveness ratio (ICER) of 7,700 EURO per QALY (estimated for the year 2014).
Currently, Sweden, Germany, United Kingdom and the United States have national screening programs. Among individuals covered through Medicare in the United States, men, aged ≥ 65, who smoke, are offered an ultrasound examination. To our knowledge, Canada is considering implementation of a national AAA-screening program, and AAA screening has been initiated in several European countries including Denmark, the Netherlands, Spain and Estonia. Moreover, European and American guidelines recommend population AAA screening of men aged 65 (2-4).
Methods
We performed a literature search for ongoing and completed systematic reviews, HTAs and RCTs in nine databases. As we already knew about the systematic review by Ali et al. from 2016, for which the literature search was conducted in April 2015 (5), we limited our search to publications from 2015 and onwards. In assessing clinical effectiveness, we chose to update findings from an earlier report from the Norwegian Institute of Public Health on the same topic (6) by communicating results from the most recent systematic review identified and assessed to be of high methodological quality.
If we had not found a sufficiently updated systematic review of high quality, we planned to search for RCTs with results published after 2015. The search strategies were based on the following PICO (Population, Intervention, Comparator, Outcome): P = all men aged 65 years; I = population based systematic screening (with ultrasound) for abdominal aorta aneurysm (AAA); C = non-population based AAA screening/ no screening; O = AAA related mortality; overall (general or total) mortality; AAA rupture; AAA operations (urgent and elective); 30-days mortality due to AAA operation; quality of life. The strategies included topic and text terms for AAA combined with (AND) topic and text terms for screening. A filter for study design was also included for relevant databases.
To conduct a health economic analysis, we adapted an existing Swedish Markov model to reflect the Norwegian setting. The model used effect data from the Multicentre Aneurysm Screening Study (MASS) study (7), while epidemiologic information and costs were collected from Norwegian data sources.
Results
The literature search yielded 556 hits after removing duplicates. Selection of publications based on relevance (PICO), type of publication (systematic review) and publication date (the most recent) resulted in the inclusion of one systematic review from 2018 assessed to be of high quality (8). Ali et al. included four RCTs in meta-analyses with, in total, 125,576 participants to assess the effect of screening men, aged 64-83, for AAA.
Results from meta-analyses showed significantly reduced AAA-related mortality in screened men both at short-term follow-up (3-5 years), and long-term follow-up (13-15 years). At short-term follow-up RR was 0.57 (95% CI 0.44 to 0.72), while at long-term follow-up RR was 0.66 (95% CI 0.47 to 0.93). Confidence in these estimates was moderate according to the GRADE assessment. For overall morality, there was non-significant reduction in the short-term, and a significant reduction in the long-term. After 3-5 years RR was 0.94 (95% CI 0.88 to 1.02), while RR was 0.987 (95% CI 0.975 to 0.999) after 13-15 years. Confidence in these two estimates was respectively low and moderate according to the GRADE assessment.
Moreover, AAA screening led to a 48% decline in AAA ruptures after 3-5 years and a 35% decline after 13-15 years. After 3-5 years RR was 0.52 (95% CI 0.35 to 0.79), while RR was 0.65 (95% CI 0.51 to 0.82) after 13-15 years. Confidence in these estimates was moderate according to the GRADE assessment.
The number of AAA operations (urgent and elective) had doubled after 3-5 years, and had increased by 35% after 13-15 years. RR was 2.16 (95% CI 1.33 to 1.65) after 3-5 years follow-up and 1.35 (95% CI 1.17 to 1.57) after 13-15 years. There was a marked increase in the number of elective AAA operations (more than tripled after 3-5 years and almost doubled after 13-15 years). RR was 3.25 (95% CI 2.13 to 4.96) after 3-5 years follow-up, whereas RR was 1.83 (95% CI 1.29 to 2.59). The number of urgent operations decreased by half as a result of AAA screening. RR was 0.50 (95% CI 0.29 to 0.86) after 3-5 years, while RR was 0.52 (95% CI 0.42 to 0.63) after 13-15 years follow-up. Confidence in these estimates was moderate according to the GRADE assessment.
Thirty-day postoperative mortality due to AAA operation at both follow-up periods decreased significantly among the screened participants. The reduction was approximately 70% after 3-5 years and 45% after 13-15 years follow-up. RR was 0.31 (95% CI 0.20 to 0.48) after 3-5 years, while after 13-15 years RR was 0.55 (95% CI 0.39 to 0.80). According to the GRADE assessments done by Ali et al., the confidence in these estimates was moderate.
The health economic analysis estimated a cost per quality adjusted life-year gained of approximately 154,000 NOK, if AAA screening is implemented and reaches a stable phase. For a population of 28,000 men, AAA screening will contribute to 62 fewer AAA-related deaths. The estimated budget impact related to screening is between 15 and 20 million NOK per year. Approximately two million kroners per cohort in additional expenses, related to increased follow-up and treatment, are anticipated over time.
Discussion
The systematic review by Ali et al. on clinical effectiveness did not report any outcomes related to quality of life (8). In a more recent narrative review, the authors reported that quality of life, e.g., related to anxiety, was lower in men who had been diagnosed with AAA than in men without a diagnosis or among the general population, and that the diagnosed men felt their health had worsened over time (9). However, after surgery their quality of life returned to its pre-screening level. This evidence is nevertheless uncertain because the studies that have assessed quality of life are too heterogeneous to draw definite conclusions.
Sufficient participation is essential for obtaining the desired effect of a screening program. The majority of those invited to AAA screening choose to participate (75 – 85% of the invited men in Sweden). Swedish and international studies indicate that socioeconomic factors such as low income, low level of education, and single marital status are the main causes of lower participation. In Sweden, participation is also lower among newly arrived immigrants. While some studies may indicate an association between the location of the screening center and the level of participation, other studies have not found any association, and it does not seem that costs for patients affect their choice to participate or not.
A number of risk factors are associated with AAA, e.g., smoking (accounts for 75% of aneurysms ≥ 40mm), cerebrovascular diseases, family history with AAA and obesity. These factors should therefore be taken into account when considering the beneficial effects AAA screening may provide, especially since we are dealing with older individuals, where risks of comorbidity are higher.
Sensitivity analyses indicate that our results using the health economic model are quite robust when it comes to potential variations in the most central parameters. Uncertainty involving less important parameters, e.g. participation rate and travel cost, proportion of women in the underlying data, possibility of over-diagnosis, and stabilization time (“steady state”), could influence model results but are unlikely to change the conclusions.
Studies from United Kingdom and Sweden report that it takes 10 years to obtain the maximum effect of AAA screening programs for avoided deaths (10, 11). This means that one cannot expect to achieve the full results presented in our report until 10 years after screening has been introduced, but that the effect will gradually approach this level during the 10-year period. The probabilistic and deterministic sensitivity analyses show that the model is relatively robust when it comes to potential future changes in the model parameters.
The results from our health economic analysis show that cost per life-year gained is substantially higher when compared to Sweden, despite relying on the same model. This may reflect differences in healthcare sector costs in the two countries.
In the United Kingdom, USA and Australia, AAA prevalence has been declining; a trend generally attributed to lower rates of smoking, as smoking increases the growth rate of aneurysms and doubles the risk of rupture. Lower AAA prevalence results in a decrease in the benefit from screening. Nevertheless, screening is still considered cost effective in the United Kingdom at AAA prevalence rates as low as 0.35%, because untreated AAA is so deadly (11). In other countries, for example, Austria, Hungary, Romania and Denmark, AAA prevalence is increasing. In Sweden, the estimated incremental cost-effectiveness ratio (ICER) for AAA screening is 7,700 EURO per QALY. Sensitivity analyses indicate that screening is still cost effective at a AAA prevalence of 0.5% (10).
Conclusion
Evidence indicates that AAA screening is beneficial to men aged 65 years, as AAA related mortality is reduced by approximately 50% in the short- and long-term. The documentation shows that AAA screening may decrease overall mortality in the short-term, but the effect is not statistically significant. However, in the long-term the effect is statistically significant. AAA screening decreases the number of urgent operations, while the number of elective repairs increases.
Our health economic analysis of AAA screening for men, aged 65, estimates a cost of 154,000 NOK per quality adjusted-life gained (ICER). For each screened cohort, AAA screening will lead to a reduction of 62 AAA-related deaths annually. The estimated budget impact is approximately 20 million NOK per cohort (approx. 28,000 men) per year.
Introduction
Abdominal aortic aneurysm (AAA) is a dilatation of the abdominal aortic artery, with a diameter enlarged by at least 50%, corresponding to a diameter of at least 30mm. AAAs bear the risk of rupture, which is a dramatic emergency condition with a high risk of death, and requires urgent surgery. The larger the AAA, the higher is the risk of rupture. AAA ruptures cause 1-2% of all deaths in the Western world. In Norway, approximately 1% of all deaths among men aged 65 and above result from AAA. The mortality rate due to AAA rupture is approximately 75%. About half of the patients die before reaching the hospital, while the death rate is 30-50% among those who reach the hospital. High age, male gender, smoking, hypertension, cardiovascular disease, and family history of AAA are factors associated with increased risk of developing AAA. Prevalence of AAA among men varies across countries from approximately 1.5-3%. In Norway, it is estimated to be around 2.5 % among men based on data from the Norwegian Vascular Surgery Registry (NORKAR). In Oslo, prevalence is 2.6-2.7% (1). Prevalence among women is one-sixth to one-fourth of that among men. Although AAA prevalence has been increasing in some countries, others have experienced declining rates.
Several European countries consider AAA screening to be a potentially beneficial healthcare intervention, and have introduced AAA screening programs in their public health services. Other countries, including Norway, currently have no systematic nationwide screening program. Individuals identified as having a high risk of AAA, usually through ultrasound examination, receive a recommendation of preventive surgery or so-called elective repair to reduce the risk of rupture. The repair can be performed using open or endovascular surgery (stent graft).
In Sweden, the national AAA screening program has, to date, identified 6,000 AAAs in men, of which 1,500 have had elective repair. The program saves approximately 100 men annually. Number needed to screen (NNS) and number needed to operate (NNO) to prevent 1 premature death are, respectively, 677 and 1.5. Screening is considered to be cost-effective in Sweden, with a gain of 577 quality-adjusted life years (QALY) and an incremental cost-effectiveness ratio (ICER) of 7,700 EURO per QALY (estimated for the year 2014).
Currently, Sweden, Germany, United Kingdom and the United States have national screening programs. Among individuals covered through Medicare in the United States, men, aged ≥ 65, who smoke, are offered an ultrasound examination. To our knowledge, Canada is considering implementation of a national AAA-screening program, and AAA screening has been initiated in several European countries including Denmark, the Netherlands, Spain and Estonia. Moreover, European and American guidelines recommend population AAA screening of men aged 65 (2-4).
Methods
We performed a literature search for ongoing and completed systematic reviews, HTAs and RCTs in nine databases. As we already knew about the systematic review by Ali et al. from 2016, for which the literature search was conducted in April 2015 (5), we limited our search to publications from 2015 and onwards. In assessing clinical effectiveness, we chose to update findings from an earlier report from the Norwegian Institute of Public Health on the same topic (6) by communicating results from the most recent systematic review identified and assessed to be of high methodological quality.
If we had not found a sufficiently updated systematic review of high quality, we planned to search for RCTs with results published after 2015. The search strategies were based on the following PICO (Population, Intervention, Comparator, Outcome): P = all men aged 65 years; I = population based systematic screening (with ultrasound) for abdominal aorta aneurysm (AAA); C = non-population based AAA screening/ no screening; O = AAA related mortality; overall (general or total) mortality; AAA rupture; AAA operations (urgent and elective); 30-days mortality due to AAA operation; quality of life. The strategies included topic and text terms for AAA combined with (AND) topic and text terms for screening. A filter for study design was also included for relevant databases.
To conduct a health economic analysis, we adapted an existing Swedish Markov model to reflect the Norwegian setting. The model used effect data from the Multicentre Aneurysm Screening Study (MASS) study (7), while epidemiologic information and costs were collected from Norwegian data sources.
Results
The literature search yielded 556 hits after removing duplicates. Selection of publications based on relevance (PICO), type of publication (systematic review) and publication date (the most recent) resulted in the inclusion of one systematic review from 2018 assessed to be of high quality (8). Ali et al. included four RCTs in meta-analyses with, in total, 125,576 participants to assess the effect of screening men, aged 64-83, for AAA.
Results from meta-analyses showed significantly reduced AAA-related mortality in screened men both at short-term follow-up (3-5 years), and long-term follow-up (13-15 years). At short-term follow-up RR was 0.57 (95% CI 0.44 to 0.72), while at long-term follow-up RR was 0.66 (95% CI 0.47 to 0.93). Confidence in these estimates was moderate according to the GRADE assessment. For overall morality, there was non-significant reduction in the short-term, and a significant reduction in the long-term. After 3-5 years RR was 0.94 (95% CI 0.88 to 1.02), while RR was 0.987 (95% CI 0.975 to 0.999) after 13-15 years. Confidence in these two estimates was respectively low and moderate according to the GRADE assessment.
Moreover, AAA screening led to a 48% decline in AAA ruptures after 3-5 years and a 35% decline after 13-15 years. After 3-5 years RR was 0.52 (95% CI 0.35 to 0.79), while RR was 0.65 (95% CI 0.51 to 0.82) after 13-15 years. Confidence in these estimates was moderate according to the GRADE assessment.
The number of AAA operations (urgent and elective) had doubled after 3-5 years, and had increased by 35% after 13-15 years. RR was 2.16 (95% CI 1.33 to 1.65) after 3-5 years follow-up and 1.35 (95% CI 1.17 to 1.57) after 13-15 years. There was a marked increase in the number of elective AAA operations (more than tripled after 3-5 years and almost doubled after 13-15 years). RR was 3.25 (95% CI 2.13 to 4.96) after 3-5 years follow-up, whereas RR was 1.83 (95% CI 1.29 to 2.59). The number of urgent operations decreased by half as a result of AAA screening. RR was 0.50 (95% CI 0.29 to 0.86) after 3-5 years, while RR was 0.52 (95% CI 0.42 to 0.63) after 13-15 years follow-up. Confidence in these estimates was moderate according to the GRADE assessment.
Thirty-day postoperative mortality due to AAA operation at both follow-up periods decreased significantly among the screened participants. The reduction was approximately 70% after 3-5 years and 45% after 13-15 years follow-up. RR was 0.31 (95% CI 0.20 to 0.48) after 3-5 years, while after 13-15 years RR was 0.55 (95% CI 0.39 to 0.80). According to the GRADE assessments done by Ali et al., the confidence in these estimates was moderate.
The health economic analysis estimated a cost per quality adjusted life-year gained of approximately 154,000 NOK, if AAA screening is implemented and reaches a stable phase. For a population of 28,000 men, AAA screening will contribute to 62 fewer AAA-related deaths. The estimated budget impact related to screening is between 15 and 20 million NOK per year. Approximately two million kroners per cohort in additional expenses, related to increased follow-up and treatment, are anticipated over time.
Discussion
The systematic review by Ali et al. on clinical effectiveness did not report any outcomes related to quality of life (8). In a more recent narrative review, the authors reported that quality of life, e.g., related to anxiety, was lower in men who had been diagnosed with AAA than in men without a diagnosis or among the general population, and that the diagnosed men felt their health had worsened over time (9). However, after surgery their quality of life returned to its pre-screening level. This evidence is nevertheless uncertain because the studies that have assessed quality of life are too heterogeneous to draw definite conclusions.
Sufficient participation is essential for obtaining the desired effect of a screening program. The majority of those invited to AAA screening choose to participate (75 – 85% of the invited men in Sweden). Swedish and international studies indicate that socioeconomic factors such as low income, low level of education, and single marital status are the main causes of lower participation. In Sweden, participation is also lower among newly arrived immigrants. While some studies may indicate an association between the location of the screening center and the level of participation, other studies have not found any association, and it does not seem that costs for patients affect their choice to participate or not.
A number of risk factors are associated with AAA, e.g., smoking (accounts for 75% of aneurysms ≥ 40mm), cerebrovascular diseases, family history with AAA and obesity. These factors should therefore be taken into account when considering the beneficial effects AAA screening may provide, especially since we are dealing with older individuals, where risks of comorbidity are higher.
Sensitivity analyses indicate that our results using the health economic model are quite robust when it comes to potential variations in the most central parameters. Uncertainty involving less important parameters, e.g. participation rate and travel cost, proportion of women in the underlying data, possibility of over-diagnosis, and stabilization time (“steady state”), could influence model results but are unlikely to change the conclusions.
Studies from United Kingdom and Sweden report that it takes 10 years to obtain the maximum effect of AAA screening programs for avoided deaths (10, 11). This means that one cannot expect to achieve the full results presented in our report until 10 years after screening has been introduced, but that the effect will gradually approach this level during the 10-year period. The probabilistic and deterministic sensitivity analyses show that the model is relatively robust when it comes to potential future changes in the model parameters.
The results from our health economic analysis show that cost per life-year gained is substantially higher when compared to Sweden, despite relying on the same model. This may reflect differences in healthcare sector costs in the two countries.
In the United Kingdom, USA and Australia, AAA prevalence has been declining; a trend generally attributed to lower rates of smoking, as smoking increases the growth rate of aneurysms and doubles the risk of rupture. Lower AAA prevalence results in a decrease in the benefit from screening. Nevertheless, screening is still considered cost effective in the United Kingdom at AAA prevalence rates as low as 0.35%, because untreated AAA is so deadly (11). In other countries, for example, Austria, Hungary, Romania and Denmark, AAA prevalence is increasing. In Sweden, the estimated incremental cost-effectiveness ratio (ICER) for AAA screening is 7,700 EURO per QALY. Sensitivity analyses indicate that screening is still cost effective at a AAA prevalence of 0.5% (10).
Conclusion
Evidence indicates that AAA screening is beneficial to men aged 65 years, as AAA related mortality is reduced by approximately 50% in the short- and long-term. The documentation shows that AAA screening may decrease overall mortality in the short-term, but the effect is not statistically significant. However, in the long-term the effect is statistically significant. AAA screening decreases the number of urgent operations, while the number of elective repairs increases.
Our health economic analysis of AAA screening for men, aged 65, estimates a cost of 154,000 NOK per quality adjusted-life gained (ICER). For each screened cohort, AAA screening will lead to a reduction of 62 AAA-related deaths annually. The estimated budget impact is approximately 20 million NOK per cohort (approx. 28,000 men) per year.
