Estimating overdiagnosis in screening for abdominal aortic aneurysm: could a change in smoking habits and lowered aortic diameter tip the balance of screening towards harm?
BMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h825 (Published 04 March 2015) Cite this as: BMJ 2015;350:h825
All rapid responses
Authors’ response to rapid responses submitted 6-10th of March 2015:
We thank all responding authors for their interest in our paper.
Björk et al. argue that overdiagnosis does not exist in AAA screening and Sandiford suggests that screening does not affect AAA prevalence but merely “reveal cases that were previously undetected”. We acknowledge that defining overdiagnosis is complex and renders careful consideration.(1) However, we argue that a man who, due to screening, is diagnosed with an AAA that would never have caused any symptoms or death in the absence of screening should be defined as overdiagnosed. The definition of an AAA (i.e. an aortic diameter above 30 mm) is a man-made construct with the purpose to estimate a probability of future rupture of the abdominal aorta. However, it is an imperfect prognostic tool and many identified abnormalities are slow- or non-progressing. Men in the targeted age range (>65 years) often have important comorbidities and the vast majority will die from other causes than ruptured aortic aneurysms. Therefore AAA screening will inevitably result in overdetection of AAAs, i.e. overdiagnosis. The rejection of overdiagnosis, which is indeed an inherent part of AAA screening programmes, is recognisable from debates about other types of screening.(2)
While Sandiford correctly point out that a significant reduction in overall mortality has been reported by one individual study(3) and in a meta-analysis,(4) the US Preventive Services Task Force (USPSTF) conclude: ”An invitation to AAA screening was not associated with a statistically significant all-cause mortality benefit in any of the individual trials or in the pooled random-effects analysis at any time point up to 15 years”.(5) As explained in the USPSTF review, the reason that the conclusions differ is because different statistical methods were used.(5) Because of heterogeneity of the included studies, the correct method for pooling results from individual trials is random-effects analysis (as opposed to fixed-effects), which was therefore used by the USPSTF. Fixed-effect analysis does not sufficiently take the heterogeneity into account and statistically significant differences shown with this approach can be misleading. Neither the trials of breast cancer screening (>600,000 participants) nor the trials of colorectal cancer screening (>700,000 participants) had the power to, or indeed showed, an effect on total mortality. Given the total and disease specific mortality in the MASS trial, the sample size required to show an effect on total mortality from AAA screening with 80% power is 116,789 (our calculation), about twice as large as the actual sample size in the trial. The heterogeneity between the trials for total mortality identified by the USPSTF further strengthens the likelihood that the effect on total mortality in the MASS-trial was due to bias or random error, and not due to the reduction in disease-specific mortality, or reduced mortality following accompanying lifestyle changes. Indeed, as only the comparatively few who were diagnosed with AAA can be expected to undergo lifestyle-changes due to screening, an effect from this on total mortality for the whole study population seem very unlikely.
Sandiford, Earnshaw & Mackie and Björk et al. suggests that offering men with screen-detected AAAs general health advice and screening for cardiovascular risk factors is an additional benefit of AAA screening. Whether a screening programme for cardiovascular risk factors directed at a healthy population can affect cardiovascular morbidity and mortality has been extensively explored in randomised trials including >180,000 participants. These have been summarised in a Cochrane-review in 2012,(6) which showed that such a strategy does not measurably improve total or cardiovascular mortality, even in high-risk populations such as male industry workers. The conclusion of the Cochrane review was confirmed last summer by one of the largest randomised trial of screening for cardiovascular risk factors in a healthy population, the Inter99 study,(7) which included >60,000 individuals. There is thus substantial, though indirect, evidence that AAA screening will not provide such an additional, indirect benefit.
Sandiford argue that our estimate of overtreatment is biased since some men who underwent surgery for a screen-detected AAA might avoid AAA-rupture after the follow-up of 13-15 years. In the MASS study, the relative risk reduction of AAA death was 48% the first 10 years after screening and decreased to 20% between 10 to 13 years.(3) Thus, the vast majority of the screening effect had occurred at 13 years. Furthermore, at 13 years of follow-up the men in the MASS-study were aged between 78-87 years, while the life expectancy in the UK for 65-year old men is 83 years.(8)
Earnshaw & Mackie and Björk et al. argue that mortality after preventive surgery has improved and points to 30-day mortality rates of only 0.6% for the first 350 men operated due to screen-detected AAAs within the Swedish screening programme, and rates below 1% for the first 1000 men operated in the UK screening programme. We agree that improved mortality rates from preventive surgery would improve the benefit:harm ratio of AAA screening. However, changes in mortality rates from surgery are complex to interpret due to multiple confounding factors. In Sweden, more than 50% of preventive AAA-surgery is performed using endovascular techniques (EVAR).(9) EVAR has a substantially lower 30-day-mortality rate than open surgery, but has more long-term complications and mortality is similar to that for open surgery after two years.(10) Consequently, in cohorts with a high proportion of EVAR, 30-day mortality rates can be misleading. Mortality rates in screen-detected AAAs may also be lower because of lower prevalence of comorbidities and smaller average size of the lesions, which is partly due to overdiagnosis. Some studies have shown lower mortality rates for screen-detected AAAs than for non-screen-detected AAAs, however a recent study based on high-quality registry data indicated no difference.(11) Additionally, studies that use administrative data to appreciate mortality from preventive surgery for AAA appears to have “a systematic bias that reduces apparent 30-day mortality” partly because a higher proportion of patients with unreported data has adverse outcomes.(12) Furthermore, when a screening programme is implemented, the first men operated on will be of a younger mean age than men operated on in a screening programme that has been running for a long time. This is because many of the screen-detected AAAs eventually operated on will not proceed to a diameter above 54 mm (the threshold for preventive surgery) until 10-15 years after screening. This could affect the mortality rates referred to by Earnshaw & Mackie and Björk et al. because younger men are likely to have a lower postoperative mortality than older men.
Sandiford as well as Earnshaw & Mackie argue that AAA screening would be cost effective at an AAA prevalence substantially lower than today. However, none of the cost-effectiveness analyses previously published have taken the consequences of overdiagnosis and overtreatment into account. Furthermore, other interventions are more cost-effective raising the question of distributive justice and cost effectiveness does not guarantee an acceptable benefit:harm ratio from an ethical perspective.(13)
Earnshaw & Mackie point out that to properly investigate the potential benefits of a lowered cut-off of aortic diameter (from 30 to 25 mm) a new trial with at least 20 years of follow-up would be needed. We acknowledge that it may be warranted to change clinical practice even though data from large randomised controlled trials with long follow-up are lacking, but this require great certainty from other sources of evidence that the benefit is large and unambiguous and that the benefit-harm ratio is very favourable. Lowering the threshold in AAA screening would likely provide a small benefit, diminishing further with lower disease prevalence, and we have shown that the benefit-harm ratio with a lower threshold is questionable. Interventions that are based on inviting asymptomatic citizens to interventions they have not asked for increase the necessity of solid evidence of substantial benefit and acceptable harm. The USPSTF review concluded that the available studies investigating a lower threshold included “a small number of participants.., ..there are no matched controls in most studies, and most studies measured expansion rates rather than health outcomes” and recommended against including men with an aortic diameter of 25-29 mm in surveillance programmes.(5) Despite this, most counties in Sweden already include these men in their surveillance programmes.
Björk et al. state that ”a patient with a small AAA can live a completely normal life. It is important that not only the patient, but relatives and all health care personal are aware of the safety in this situation”. While small AAA’s may pose a small threat to health, there is no evidence that the affected man or his relatives perceive a small AAA as ”safe”. On the contrary, interview studies indicate that these men does not have a medically ”correct” understanding of the meaning of the enlarged size of their aorta.(14) This is in line with research on risk perception that shows that our understanding of risk relies primarily on emotions and that a cognitive comprehension plays only a minor role.(15,16) Indeed, it seem contradictory when Björk et al. argue that the AAA-diagnosis means that men can be “easily motivated” for lifestyle changes, while at the same time arguing that the diagnosis has little psychological impact.(13)
Quantitative research on psychological consequences of screening is complex and renders thorough methodological considerations to insure high validity and reliability. A recent review on psychological consequences of screening concluded that the available quantitative studies concerning AAA screening were insufficient,(17) which was also the conclusion in the Cochrane(18) and the USPSTF(5) reviews. The studies that Sandiford and Björk et al. claim indicate small psychological harmful effects have all used generic (standard) questionnaires (as opposed to diagnose-specific ones) which have a low validity in a screening context.(19) Additionally, none of the studies have accounted for the statistical psychometric properties of the included measures.
It is not possible to evaluate benefit:harm ratio of AAA screening without taking the aspect of overdiagnosis into account. The lack of recognition of the harms of established health care interventions among those responsible underlines the importance of the BMJ initiative to highlight overdiagnosis in a recent theme issue in their “Too Much Medicine” campaign (http://www.bmj.com/specialties/digital-theme-issue-overdiagnosis). Additionally, all screening programmes should be prepared for reassessment, and potentially termination, when called for due to changed conditions, such as changes in important risk factors and disease prevalence.
References:
1. Carter SM, Rogers W, Heath I, Degeling C, Doust J, Barratt A. The challenge of overdiagnosis begins with its definition. BMJ 2015;350:h869.
2. Barratt A. Overdiagnosis in mammograohy screening: a 45 year journey from shadowy idea to acknowledged reality. BMJ 2015;350:h867.
3. Thompson SG, Ashton HA, Gao L, Buxton MJ, Scott RAP. Final follow-up of the Multicentre Aneurysm Screening Study (MASS) randomized trial of abdominal aortic aneurysm screening. Br J Surg 2012;99:1649-1656.
4. Lindholt JS, Norman P. Screening for abdominal aortic aneurysm reduces overall mortality in men. A meta-analysis of the mid- and long-term effects of screening for abdominal aortic aneurysms. Eur J Vasc Endovasc Surg 2008;36:167-71.
5. Guirguis-Blake J M, Beil T L, Senger C A, Whitlock E P. Ultrasonography Screening for Abdominal Aortic Aneurysms: A Systematic Evidence Review for the U.S. Preventive Services Task Force. Annals of Internal Medicine 2014;160:321-9.
6. Krogsbøll LT, Jørgensen KJ, Grønhøj Larsen C, Gøtzsche PC. General health checks in adults for reducing morbidity and mortality from disease. Cochrane Database of Systematic Reviews 2012, Issue 10. Art. No.: CD009009.
7. Jørgensen T, Jacobsen RK, Toft U, Aadahl M, Glümer C, Pisinger C. Effect of screening and lifestyle counselling on incidence of ischaemic heart disease in general population: Inter99 randomised trial. BMJ 2014;348:g3617.
8. Office for National Statistics. UK Interim Life Tables, 1980-82 to 2008-10. Available at: http://www.ons.gov.uk/ons/rel/lifetables/interim-life-tables/2008-2010/s.... Accessed 27th of November 2014.
9. Swedish registry for vascular surgery (Swedvasc), available at: http://www.ucr.uu.se/swedvasc/index.php/arsrapporter. Accessed 18th of March 2015.
10. Stather PW, Sidloff D, Dattani N, Choke E, Bown MJ, Savers RD. Systematic review and meta-analysis of the early and late outcomes of open and endovascular reapir of abdominal aortic aneurysm. Br J Surg 2013;100(7):863-72.
11. Linne et al, Low post-operative mortality after surgery on patients with screening-detcted abdominal aortic aneurysms: a Swedvasc registry study, Eur J Vasc Endovasc Surg 2014;48:649-56..
12. Hussey K, Siddiqui T, Burton P, Welch GH, Stuart WP. Understanding Administrative Abdominal Aortic Aneurysm Mortality Data. Eur J Vasc Endovasc Surg 2015;49-277-82.
13. Johansson M, Jørgensen KJ, Brodersen J. Harms of screening for abdominal aortic aneurysm: there is more to life than a 0⋅46% disease-specific mortality reduction. Lancet In press.
14. Hansson A, Brodersen J, Reventlow S, Pettersson M. Opening Pandora’s box: The experiences of having an asymptomatic aortic aneurysm under surveillance. Health, Risk & Society 2012;14,4:341-59.
15. Loewenstein GF, Weber EU, Hsee CK, Welch N. Risk as feelings. Psychol Bull 2001;127(2):267-86.
16. Adelswärd V, Sachs L. The meaning of 6.8: numeracy and normality in health information talks. Soc Sci Med 1996;43:1179-87.
17. DeFrank JT, Barclay C, Sheridan S, Brewer NT, Gilliam M, Moon AM, Rearick W, Ziemer C, Harris R. The psychological harms of screening: the evidence we have versus the evidence we need. J Gen Intern Med 2015;30:242-48.
18. Cosford PA, Leng GC, Thomas J. Screening for abdominal aortic aneurysm. Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.:CD002945.
19. Brodersen J, McKenna SP, Doward LC, Thorsen H. Measuring the psychosocial consequences of screening. Health Qual Life Outcomes 2007;5:3.
Competing interests: No competing interests
Johansson et al raise an important issue in questioning whether overdiagnosis tips the balance of abdominal aortic aneurysm (AAA) screening towards harm. However there are a number of omissions and inaccuracies in their paper that should be taken into account.
1. “…there was no significant reduction in overall mortality”. This statement is in contradiction to two published studies of long term follow-up in AAA screening trials. Lindholt and Norman performed a meta-analysis to examine the pooled effects on AAA-related and total mortality.[1] They found that AAA screening resulted in a significant reduction in overall mortality (odds ratio 0.94, 95% confidence limits 0.92-0.97). In the final analysis of the MASS trial, Thompson et al reported an overall reduction in mortality of 3% (1%-5%).[2]
2. “[AAA Screening Programmes] assume that men who are currently invited for screening are at the same risk as those who participated in the trials of the 1980s and 1990s.” This is not strictly true. AAA screening programmes use results from the trials to estimate the cost-effectiveness of screening in any given situation, usually based on Markov models. The cost-effectiveness depends upon a number of parameters, some of which were derived from the trials, but others, such as the costs, the prevalence rate of aneurysms or the screening participation rate, can be put into the models separately to see if the programme remains cost-effective under different conditions. Two published studies suggest that AAA screening remains cost-effective at prevalence rates much lower than those experienced in the MASS trial.[3 4]
3. “Our calculations below, based on the largest randomised trial, indicate that screening almost doubles AAA prevalence.” For the objectivists among us screening does nothing to alter AAA prevalence (whatever aortic diameter cut-off point you might choose). All it does is to reveal cases that were previously undetected.
4. “…by capturing mostly small aneurysms, screening can be argued to change the AAA diagnosis from a life threatening condition to a risk factor.” Surely that is the whole point of screening?
5. “In the MASS study, 1334 men had screen detected AAAs but only 157 (11.8%) of them avoided death as a result. For the rest, the early diagnosis did not improve prognosis—that is, the men died from or survived their AAA despite screening.” This statement is not totally accurate. The final analysis of the MASS study found that screening prevented 157 AAA-related deaths. However, the total number of non-AAA-related deaths in the screening group was also lower than in the control group. Despite almost identical numbers in the screening and control groups, there were 119 fewer non-AAA related deaths in the screening group than the control group. Although this difference was only of borderline significance (p=0.08) an ‘intention to screen’ analysis would include them in the measure of effectiveness. There are plausible reasons why AAA screening may have had an impact on non-AAA related deaths. One of them is the potential bias that makes AAA more likely to have been given as the cause of death in the screened group than in the control group, particularly for those patients in whom an aortic aneurysm had been diagnosed.[1 5] This of course would also reduce the number of overdiagnoses.
Another related possibility, not considered by the authors, is that diagnosis of AAA by screening makes men more willing to adopt (or compliant with) cardiovascular risk reduction including smoking cessation and use of medications such as statins and antihypertensives. About half of the patients receiving a diagnosis of aortic aneurysm cease smoking and the growth rate of aneurysms in those patients is known to be significantly slower than in those who continue to smoke.[6] Although the evidence that aortic aneurysm growth is slowed by cardiovascular drugs is weak,[7] they are associated with lower mortality from non-AAA related causes, notably ischaemic heart disease. The likely impact of these screening associated-effects on aneurysm growth rates and non-AAA causes of death will tend to bias estimates of overdiagnosis upwards.
6. “In the MASS study, some asymptomatic men in the control group might have had AAA diagnosed and treated through opportunistic screening leading to an underestimate of overdiagnosis.” Although this may lead to an underestimate of overall AAA overdiagnosis, it does not cause any underestimation of the rate of overdiagnosis due to the formal population-based screening programme itself. It is the latter that is of primary concern to policy-makers.
7. “Quantitative studies on the psychosocial consequences of AAA screening are of insufficient quality to draw conclusions” This is a surprising statement from the authors, given that one of the main thrusts of their argument is that overdiagnosis is harmful, and may cause “appreciable anxiety throughout the remainder of their lives”. The authors seem to have missed several relevant studies, include one from Australia on the health-related quality of life impact of AAA screening which found that “Apart from physical functioning, screening was not associated with decreases in health and well-being. A high proportion of men rated their health over the year after screening as being either the same or improved, regardless of whether or not they were found to have an AAA”.[8] While Lindholt et al did find persistent anxiety among patients with conservatively-treated screen-diagnosed AAA, their findings are at odds with two other studies in addition to the one already cited,[9 10] and with broader research on the psychological impacts of individual health risk assessments.[11] Indeed, the authors have not given any consideration to the possibility that the apparently transient increased anxiety experienced by the small minority of screened men in whom AAA is diagnosed might be offset to a greater or lesser extent by the utility of reassurance gained by those with a negative screen result, particularly among those with relatives or friends who have died from AAA.
8. The estimate of 37 men having unnecessary preventive surgery with 1.6 deaths for every 10,000 enrolled patients is in my opinion the most important overdiagnosis concern raised by Johansson et al. The estimate is based on the assumption that the mortality benefit of screening disappears after 10 years of follow-up. However the MASS trial found that the benefit of being invited to screening continued to accumulate throughout the 13-year follow-up period and hence it is unrealistic to assume that it would suddenly fall to zero beyond that period. Far more likely is that the control group would have continued to experience a significantly higher AAA-related mortality rate compared with the screened population. Their unnecessary surgery and death estimates are highly sensitive to the possibility that more than 57 would have had a ruptured aneurysm (e.g. due to the detection bias mentioned above) because for every additional rupture included in this ‘necessary surgery’ estimate there is one less in ‘unnecessary surgery’. For the authors to argue that harms outweigh the benefits among those undergoing surgery, they effectively need to make a case that current practice in vascular surgery to operate on patients with aneurysms of 5.5 cm or greater is clinically unjustified, irrespective of whether these are detected by screening or not.
Overdiagnosis is inherent to clinical practice where one’s intention is to avoid unpredictable but catastrophic health outcomes, be they ruptured AAA, ruptured appendices or the development of incurable invasive cancer. It is entirely appropriate to question whether the harms associated with screening, including overdiagnosis, outweigh the benefits, provided that this is done in a fair and balanced manner. This does not mean, as the authors seem to believe, that overdiagnosis should be avoided at all costs.
1. Lindholt JS, Norman P. Screening for abdominal aortic aneurysm reduces overall mortality in men. A meta-analysis of the mid- and long-term effects of screening for abdominal aortic aneurysms. Eur. J. Vasc. Endovasc. Surg. 2008;36(2):167-71 doi: 10.1016/j.ejvs.2008.03.006[published Online First: Epub Date]|.
2. Thompson S, Ashton H, Gao L, Buxton M, Scott R. Final follow‐up of the Multicentre Aneurysm Screening Study (MASS) randomized trial of abdominal aortic aneurysm screening. Br. J. Surg. 2012;99(12):1649-56
3. Svensjö S, Mani K, Björck M, Lundkvist J, Wanhainen A. Screening for Abdominal Aortic Aneurysm in 65-Year-old Men Remains Cost-effective with Contemporary Epidemiology and Management. Eur. J. Vasc. Endovasc. Surg. 2014;47(4):357-65
4. Glover M, Kim L, Sweeting M, Thompson S, Buxton M. Cost‐effectiveness of the National Health Service abdominal aortic aneurysm screening programme in England. Br. J. Surg. 2014
5. Lindholt JS, Sørensen J, Søgaard R, Henneberg E. Long‐term benefit and cost‐effectiveness analysis of screening for abdominal aortic aneurysms from a randomized controlled trial. Br. J. Surg. 2010;97(6):826-34
6. Baxter BT, Terrin MC, Dalman RL. Medical management of small abdominal aortic aneurysms. Circulation 2008;117(14):1883-89
7. Sweeting MJ, Thompson SG, Brown LC, Powell JT, on behalf of the Rc. Meta-analysis of individual patient data to examine factors affecting growth and rupture of small abdominal aortic aneurysms. Br. J. Surg. 2012;99(5):655-65 doi: 10.1002/bjs.8707[published Online First: Epub Date]|.
8. Spencer CA, Norman PE, Jamrozik K, Tuohy R, Lawrence-Brown M. Is screening for abdominal aortic aneurysm bad for your health and well-being? ANZ journal of surgery 2004;74(12):1069-75
9. Lucarotti M, Heather B, Shaw E, Poskitt K. Psychological morbidity associated with abdominal aortic aneurysm screening. Eur. J. Vasc. Endovasc. Surg. 1997;14(6):499-501
10. Khaira H, Herbert L, Crowson M. Screening for abdominal aortic aneurysms does not increase psychological morbidity. Ann. R. Coll. Surg. Engl. 1998;80(5):341
11. Shaw C, Abrams K, Marteau TM. Psychological impact of predicting individuals’ risks of illness: a systematic review. Soc. Sci. Med. 1999;49(12):1571-98
Competing interests: I am Chair of the Waitemata Abdominal Aortic Aneurysm Screening Pilot Steering Committee, testing the feasibility and cost-effectiveness of AAA screening in Maori men and women.
The analysis done by Johansson and colleagues is welcome, since ongoing evaluation is important for all modern screening programmes. Their challenge to AAA screening, however, warrants scrutiny.
The NHS AAA Screening Programme (NAAASP) has been implemented fully in England. In the first five years, over 700,000 men have been screened, and around 10,000 diagnosed with an AAA. Over 1000 men with a large AAA have been referred for possible treatment, and the rest of these men remain in surveillance.
There is no doubt that in the effort to find important AAAs a number of men are found whose aneurysms will not develop or give them problems. That clear eyed examination of the balance of good and harm is explicit in the consideration of new screening programmes http://www.screening.nhs.uk/criteria. A clear and systematic way of assessing psychological harm accruing to screening programmes is awaited. In the meantime acceptability of programmes is an important criterion to be considered before a positive recommendation is made. The AAA programme has also focussed on efforts to ensure that men are offered information to assist them in making the right choice for them, and a decision aid has been developed and evaluated by the programme https://sdm.rightcare.nhs.uk/pda/aaa-screening. In addition, men who have small aneurysms are offered general health advice (including smoking cessation), and a recommendation is sent to their GP to consider antiplatelet and statin therapy, since they are known to be at higher risk of cardiovascular complications1.
Existing AAA screening programmes are based on trials designed 20 years ago, and clearly decisions about their value must include current data. For example, Johansson and colleagues rightly point out that preventive surgery carries risks. In the first 1000 men referred with a large AAA for planned treatment, perioperative mortality was below 1%. They, and others have observed reduced prevalence of AAA in 65 year old men (around 1.3% in NAAASP, compared to 4% in the randomised trials). Yet even using updated costs and figures, the NAAASP is cost effective using current willingness to pay thresholds, and will remain so down to a prevalence of 0.35%2. There is no evidence that the risk of rupture from a 5.5cm AAA has been affected by changing prevalence.
Finally, the authors are concerned about expanding the number of men in surveillance by potentially including men with an aorta just below the current threshold of 3cm. It is emerging that around 10% of these men will eventually get an AAA >5.4cm in diameter, though after 10-15 years. It is not possible to design a controlled trial that would answer the question in less than 20 years to determine whether a strategy of including these men in surveillance is in their interests. The NAAASP is currently assembling the evidence about whether their inclusion in some sort of modified surveillance, such as a rescan at aged 70 years might be cost effective, and will then seek advice from the UK National Screening Committee about future action.
References
1. Duncan JL, Harrild KA, Iverson L, Lee AJ, Godden DJ. Long term outcomes in men screened for abdominal aortic aneurysm: prospective cohort study. BMJ 2012: 344: e2958.
2. Glover M, Kim LG, Sweeting MJ, Thompson SG, Buxton MJ. Cost effectiveness of the NHS abdominal aortic aneurysm screening programme in England. Br J Surg 2014; 101: 976-982.
Jonothan Earnshaw, Clinical Director, NHS AAA Screening Programme
Anne Mackie, Director of Programmes, UK National Screening Committee
Competing interests: No competing interests
A sceptical approach towards screening is warranted, since discovering disease in a person feeling healthy is always associated with potential harm. It is also important, however, to have knowledge about the disease in question, and the patients affected. Dr Johansson and colleagues have written an article in last week’s edition of the BMJ where they use data from previously published investigations with the aim to evaluate the potential harm of a screening program for abdominal aortic aneurysm (AAA)(1). The authors have not performed any original within this research area themselves. From the authors' affiliations it can be assumed that they have no professional experience of treating or diagnosing patients with abdominal aortic aneurysms, and strangely enough there has been no research collaboration with the multiple research groups in Sweden which are investigating this area.
The authors “cherry-pick” data from other investigations, in particular from the Gloucestershire cohort study (2) and the UK MASS trial (3). By sometimes picking old data (such as mortality, that has decreased substantially over time), other times fresh data (such as on prevalence of disease at the age of 65, which also is decreasing over time, thanks to improving smoking habits), they paint a biased picture to serve their preconceptions. An example: only two of the first 350 screening detected patients who were operated on for AAA in Sweden died within 30 days, 0.6% (4), very different from the figures 3.9-4.5% given by the authors of this paper.
The fact that the authors have limited knowledge and experience of the patients and their disease is probably an explanation why there are so many misconceptions in the paper. The aim of this correspondence is to clarify some of those.
The main theme of the paper is that the authors try to show that a screening program results in “overdiagnosing” the disease, and they even claim that the prevalence of the disease increases by detection with screening (sic!). Much of the thinking of this paper emanates from screening for cancer, which is a much different situation. This is not a disease that spreads in the body, so there is no benefit to operate on the patient when the AAA is small. Large randomized trials have shown that there is no benefit to operate on asymptomatic patients if their AAA is less than 5.5 cm in diameter. A patient with a small AAA can live a completely normal life. It is important that not only the patient, but relatives and all health care personal are aware of the safety in this situation. This is also a situation when the patient can easily be motivated to life-style changes, including smoking cessation (5). The large MASS trial3 showed a significant decrease in all-cause mortality of 3% in those invited to screening (95% CI 1-5%), an extraordinary result in preventive medicine, one third of the reduction of mortality was non-aneurysm related, showing the power of promoting life-style changes, as well as of medical secondary prevention.
Furthermore, the authors claim that there has been a debate to change the definition of an AAA from 30 to 25 mm in the scientific community, which is simply not true. They write: “Diagnostic change— Sweden, the UK, and the US have initiated screening programmes for AAA. There are also proposals to change the aortic diameter for diagnosis from ≥30 mm to 25 mm”. No reference, which is natural since the statement is incorrect.
What has been debated, however, is that 65 year old men with a sub-aneurysmal aorta (a diameter of 25-29 mm) may benefit from rescreening at the age of 70 or 75 years, since by then many of them then have developed an AAA. The increased longevity, in combination with the minimally invasive surgical technique, makes it reasonable to believe that patients may benefit from detection of an AAA and preventive surgery even in their eighties.
In summary, there is no “overdiagnosing” of AAA, but the ambition of the screening program is to detect the aneurysm before it ruptures, and to treat it timely. In money-driven health care systems there is always a risk of overtreatment, but there are also often powerful mechanisms to control and reduce that kind of malpractice. To my knowledge there are no such problems in the UK or in Sweden.
The doctor who has met a patient with a ruptured AAA, have felt the pain and fright in the ambulance, the emergency or operating room, and the despair in the eyes of the relatives, knows what this is all about. This death, often in extreme pain, can be prevented with a simple and safe ultrasound examination, and with modern non-invasive low-risk surgery. This is why 80-90% of Swedish 65-year old men attend the AAA screening, when invited, although they pay the examination by themselves. Contrary to the selected arguments put forward by Dr Johansson and colleagues, the results from several studies evaluating the emotional impact of AAA screening suggest that AAA screening does not increase anxiety or depression among the screened subjects. Our data suggest that they feel very safe when under surveillance, and when necessary, during surgery (6).
Martin Björck, Anders Wanhainen and Sverker Svensjö
For the AAA screening program in Uppsala (started in 2006) and Dalarna (started in 2008), Department of Surgical Sciences, Uppsala University, Uppsala, Sweden
References:
1. Johansson M, Hansson A, Brodersen J. Estimating overdiagnosis in screening for abdominal aortic aneurysm: could a change in smoking habits and lowered aortic diameter tip the balance of screening towards harm? BMJ 2015;350:h825 doi: 10.1136/bmj.h825 (Published 3 March 2015)
2. Darwood R, Earnshaw JJ, Turton G, Shaw E, Whyman M, Poskitt K, et al. Twenty-year review of abdominal aortic aneurym screening in men in the county of Gloucestershire, United Kingdom. J Vasc Surg 2012;56:8-14.
3. Thompson SG, Ashton HA, Gao L, Buxton MJ, Scott RAP. Final follow-up of the Multicentre Aneurysm Screening Study (MASS) randomized trial of abdominal aortic aneurysm screening. Br J Surg 2012;99:1649-56
4. Linné A, Smidfelt K, Langenskiöld M, Hultgren R, Nordanstig J, Kragsterman B, Lindström D. Low Post-operative Mortality after Surgery on Patients with Screening-detected Abdominal Aortic Aneurysms: A Swedvasc Registry Study. Eur J Vasc Endovasc Surg 2014;48:649-656
5. Bohlin S, Fröjd C, Wanhainen A, Björck M. Change in smoking habits after having been screened for abdominal aortic aneurysm. Eur J Vasc Endovasc surg 2014;48:138-143
6. Khaira HS, Coddington T, Drew A, Roberts PN, Imray CH. Screening for abdominal aortic aneurysms does not increase psychological morbidity. Ann R Coll Surg Engl 1998;80:341-342
Competing interests: No competing interests
This paper gives a useful series of estimates of benefits and harms.
The sanguine response from a vascular surgeon is unsurprising, given that "A US Preventive Services Task Force review that included these trials estimated a 50% reduction in relative risk of AAA related mortality after 13-15 years’ follow-up (absolute risk reduction 0.46%)."
However, there was NO significant reduction in overall mortality !
This paper states that " if 10 000 men are invited to screening, 46 deaths from AAA can be prevented over 13-15 years." But since there is no (statistically significant) reduction in total mortality, I must assume that 46 more deaths would occur as a side-effect of screening, or surgery, or sheer fright, perhaps ?
Give me a sudden death, hopefully in my sleep.
Competing interests: No competing interests
I read with interest, as the authors wrestled with the problem of an imperfect tool used for aneurysm screening in national screening programs worldwide; if only we could, with any certainty, identify the individuals who will die from AAA. Nobody can of course, so, some societies have taken the view that attempting to identify those with increased risk, in order to give prophylactic treatment, will benefit society overall, despite a small degree of 'harm' attributed to receiving a diagnosis that may not require treatment, and harm from unnecessarily treating others. Those deliberations are detailed, evidence based, openly available and accounted for.
As far as screening programs go of course, AAA screening is one of the most effective. An easily diagnosed condition, with a fairly predictable course, a defined threshold for treatment, and a significant prevalence. In addition the disease has a durable, effective and relatively safe, genuinely curative treatment. From the publication, screening 10000 for AAA will prevent 46 deaths over a 15 year period, equating to a 'number needed to screen' to save one life at 250. This looks very favorable compared to, for example, breast screening with an equivalent figure of 1500, hemoccult tests for bowel cancer at 800, and serum cholesterol at 400.
The authors critique of AAA screening correctly identifies one of the most significant drawbacks of such programs, however, which is perhaps more relevant their to the justification and effectiveness . That is the tendency for them to be beneficial to the more prosperous classes of society who engage in them, whereas the majority of disease prevalence, and therefor greatest benefit, lies lower down in the socioeconomic spectrum. If a review of aneurysm screening is to take place, therefore, the focus should be on how to tap into and engage the at risk groups more, rather than tweaking the edges of the more reliable parts of it like aneurysm threshold measurement.
Competing interests: No competing interests
Re: Estimating overdiagnosis in screening for abdominal aortic aneurysm: could a change in smoking habits and lowered aortic diameter tip the balance of screening towards harm?
Screening for abdominal aortic aneurysm: Is it wise to conduct blanket screening of all 65+ males?
Johansson et al discuss over diagnosis in screening (1). I wish to draw attention to over screening for diagnosis of abdominal aortic aneurysm (AAA). The principal modifiable risk factors for AAA are essential hypertension and smoking which leads to reduction of medial-elastic-fibres through fragmentation and degeneration (2). There is depletion of medial-smooth-muscle-cells and overexpression of matrix-metalloproteinases 2 and 9, harmful cytokines and molecules (2). Patients with an ankle-brachial-pressures of below 0.87 have the highest mortality risk from AAA (3).
There is decreased prevalence of AAA in diabetes possibly because of increased medial-wall-stiffness due to advanced glycation end products, altered smooth-muscle-type and decreased matrix-metalloproteinases. These changes possibly are more marked in leg arteries hence ankle-blood-pressures increase more than brachial-blood-pressures resulting in higher ankle-brachial-index in diabetes (4).
Would it not be better to establish first if subjects have low, normal or high ankle-brachial-index and inviting only ones with either normal or low ankle-brachial-index? It is unlikely that major fragmentation/ degeneration process and major stiffening process are happening at the same time. By excluding subjects with ankle-brachial-index of >1.2, lot of monies will be saved by not screening them with ultrasound.
South-Asians (originally from India, Pakistan and Bangladesh) suffer from increased cardiovascular diseases as compared to Europeans. However, they have a very low prevalence of AAA possibly due to decreased prevalence of essential hypertension but increased prevalence of diabetes (5). Except in Bangladeshis the prevalence of smoking is also less in South-Asians. Moreover increase in ankle-blood-pressures with diabetes is greater and they have decreased prevalence of ankle-brachial-index of <0.9 (6) .
AAA is a disease of old age and South-Asians suffer from premature (<65years) cardiovascular disease. Therefore South-Asians included in screening will usually get a normal result which will provide false reassurance because at age 65+ they already might have an enormous burden of atherosclerosis in other vascular beds. In fact they might be already on medications for secondary prevention and they might change their attitude and behavior towards lifestyle-changes which they have already instituted in their daily-routine.
Acknowledgement: University of Leeds
Conflict of interest: None
References
(1) Johansson M, Hansson A, Brodersen J. Estimating overdiagnosis in screening for abdominal aortic aneurysm: could a change in smoking habits and lowered aortic diameter tip the balance of screening towards harm? BMJ 2015;350:h825.
(2) Annambhotla S, Bourgeois S, Wang X, Lin PH, Yao Q, Chen C. Recent advances in molecular mechanisms of abdominal aortic aneurysm formation. World J Surg 2008 June;32(6):976-86.
(3) Powell JT, Brady AR, Thompson SG, Fowkes FG, Greenhalgh RM. Are we ignoring the importance of ankle pressures in patients with abdominal aortic aneurysm? Eur J Vasc Endovasc Surg 2001 January;21(1):65-9.
(4) Aboyans V, Lacroix P, Tran MH, Salamagne C, Galinat S, Archambeaud F et al. The prognosis of diabetic patients with high ankle-brachial index depends on the coexistence of occlusive peripheral artery disease. J Vasc Surg 2011 April;53(4):984-91.
(5) Joshi P, Islam S, Pais P, Reddy S, Dorairaj P, Kazmi K et al. Risk factors for early myocardial infarction in South Asians compared with individuals in other countries. JAMA 2007 January 17;297(3):286-94.
(6) Kain K, Brockway M, Ishfaq T, Merrick M, Mahmood H, Ingoe JC et al. Ankle pressures in UK South Asians with diabetes mellitus: a case control study. Heart 2013 May;99(9):614-9.
Competing interests: No competing interests