Primary Care Screening for Abdominal Aortic Aneurysm: Updated Evidence Report and Systematic Review for the U.S. Preventive Services Task Force

Abstract

Importance.

Ruptured abdominal aortic aneurysms (AAA) have high mortality estimated at 81%.

Objective.

To systematically review the evidence on benefits and harms of AAA screening and small aneurysm treatment to inform the US Preventive Services Task Force.

Data Sources.

MEDLINE, PubMed (publisher supplied only), Database of Abstracts of Reviews of Effects, and Cochrane Central Register of Controlled Trials for relevant English-language studies published through September 2018. Surveillance continued through July 2019.

Study Selection.

Trials of AAA screening benefits and harms; trials and cohort studies of small (3.0cm to 5.4cm) AAA treatment benefits and harms.

Data Extraction and Synthesis.

Two investigators independently reviewed abstracts and full-text articles and extracted data. Peto’s method was used to pool odds ratios for AAA-related mortality, rupture, and operations; the DerSimonian and Laird random effects model was used to pool calculated risk ratios for all-cause mortality.

Main Outcomes and Measures:

AAA and all-cause mortality; AAA rupture; treatment complications.

Results.

Fifty studies (N=323,279) met inclusion criteria. Meta-analysis of population-based RCTs estimated that a screening invitation to men aged 65 years and older was associated with a 35% reduction in AAA-related mortality over 12 to 15 years (Peto OR 0.65 [95% CI, 0.57–0.74], 4 RCTs [n=124,926]), 38% reduction in AAA-related ruptures over 12 to 15 years (Peto OR 0.62 [95% CI, 0.55–0.70] 4 RCTs [n=124,929), and 43% reduction in emergency surgical procedures over 4 to 15 years (Peto OR 0.57 [95% CI, 0.48–0.68] 5 RCTS [n=175,085]). In contrast, no all-cause mortality benefit was seen at 12- to 15-year follow-up (RR 0.99 [95% CI 0.98–1.00] 4 RCTs [n=124,929]). One-time screening was associated with significantly more procedures over 4 to 15 years in the invited group compared with the control group (Peto OR 1.44 [95% confidence interval [CI], 1.34–1.55] 5 RCTs, N=175,085). Four trials (N=3,314) of small aneurysm surgical treatment demonstrated no significant difference in AAA-related mortality or all-cause mortality compared with surveillance over 1.7 to 12 years. These four early surgery trials showed a substantial increase in procedures in the early surgery group. For small aneurysm treatment, registry data (3 studies, N=14,424) showed that women had higher surgical complications and postoperative mortality compared with men.

Conclusions and Relevance:

One-time AAA screening in men aged 65 years and older was associated with decreased AAA-related mortality and rupture rates but no effect on all-cause mortality. Higher rates of elective surgery but no long-term differences in quality of life resulted from screening.

Abdominal aortic aneurysms (AAA) are often asymptomatic, with slow expansion until rupture. AAA screening to identify and treat aneurysms before rupture can potentially prevent a fatal outcome. To prevent rupture, AAA, defined as an aneurysm of 3.0 centimeters (cm) diameter or larger, is most commonly surgically repaired via open repair or endovascular aneurysm repair (EVAR) when it reaches a diameter of 5.5 cm.^1–3 The role of pharmacotherapy to slow aneurysm expansion has been uncertain.⁴

Reported AAA prevalence rates in those aged 60 or older have declined from 1.6% to 7.2% in the 1990s^5,6 to more contemporary estimates that range from 1.2% to 3.3%.^7,8 The most important risk factors for the development of AAA include advanced age,^9,10 male sex,^10,11 smoking,^4,11–13 and family history of AAA.^12–14

In 2014, the United States Preventive Services Task Force (USPSTF) recommended one-time screening for AAA by ultrasonography in asymptomatic men aged 65 to 75 years who have ever smoked (B Recommendation).¹⁵ The USPSTF concluded that the benefits of screening do not clearly outweigh the possible harms and recommended that clinicians selectively offer screening for AAA in men ages 65 to 75 years who have never smoked (C Recommendation).¹⁵ Also, the USPSTF recommended against routine screening for AAA in asymptomatic women who have never smoked (D Recommendation) and determined that there was insufficient evidence for screening women aged 65 to 75 years who have ever smoked (I statement).¹⁵ This review was prepared to inform an updated recommendation by the USPSTF on the evidence related to the effectiveness of one-time and repeat screening for AAA and possible related harms, as well as the effectiveness and related harms of treatment (pharmacotherapy or surgery) of small AAAs (3.0–4.5 cm diameter).

METHODS

Scope of Review

Five key questions (KQs) (Figure 1) were developed to identify the benefits (KQ1) and harms (KQ3) of one-time screening for AAA, the effects of rescreening for AAA on health outcomes or AAA incidence (KQ2), and the effectiveness (KQ4) and harms (KQ5) of treatment of small AAA (3.0 to 5.4 cm diameter). Additional methodological details are publicly available in the full evidence report http://www.uspreventiveservicestaskforce.org/Page/Document/UpdateSummaryFinal/abdominal-aortic-aneurysm-screening¹.

Figure 1. Analytic Framework.

AAA, abdominal aortic aneurysm; cm, centimeter

^a Evidence reviews for the US Preventive Services Task Force (USPSTF) use an analytic framework to visually display the key questions that the review will address to allow the USPSTF to evaluate the effectiveness and safety of a preventive service. The questions are depicted by linkages that relate interventions and outcomes. A dotted line depicts a health outcome that follows an intermediate outcome. Refer to USPSTF Procedure Manual for further details.¹⁷

Data Sources and Searches

To identify studies published since the previous review,¹⁶ literature searches were conducted from January 2013 through September 4, 2018, in MEDLINE, PubMed (for publisher-supplied records only), the Database of Abstracts of Reviews of Effects, and the Cochrane Central Register of Controlled Trials (eMethods). Additional studies were located by reviewing reference lists of other systematic reviews and through suggestions by experts. Ongoing surveillance was conducted after September 2018 through July 26, 2019, to identify newly published studies that may affect the findings of the review. This was accomplished through targeted searches of journals with a high-impact factor and journals relevant to the topic to identify major studies that might affect the conclusions or understanding of the evidence and therefore the related USPSTF recommendation. No additional articles were identified during the surveillance period.

Study Selection

Two reviewers independently evaluated articles from the previous review in addition to citations and full-text articles from the literature searches against specified inclusion criteria (eTable1 in the Supplement). Eligible screening studies used ultrasound as the screening modality for identifying AAA in asymptomatic adults >50 years of age. Only randomized clinical trials (RCTs) comparing one-time screening to no screening were used to evaluate the effectiveness of screening for AAA (KQ1). When assessing the benefits of repeated AAA screening and the harms of screening for AAA, RCTs and large cohort studies (n ≥1000) of asymptomatic adult populations were considered (KQ2 and KQ3). Studies of the effectiveness of treatment and related harms focused on individuals with small AAAs (3.0 to 5.4 cm diameter) because the vast majority of screen-detected aneurysms are small in size. The effectiveness of treating small AAAs (KQ4) was examined through RCTs evaluating surgical intervention or pharmacotherapeutic treatment compared with surveillance, usual care, or placebo. The criteria for assessing harms of treating small AAAs (KQ5) included RCTs, observational studies, and registry data related to surgical harms. The results for pharmacotherapy interventions for KQ4 and KQ5 are not reported in depth in this article but are presented in the appendix and available in the full report.

Data Extraction and Quality Assessment

Two reviewers applied USPSTF design-specific criteria¹⁷ to assess the methodological quality of all eligible studies. and were evaluated to be good- or fair-quality by using items from the Newcastle-Ottawa Scale¹⁸ and USPSTF quality rating standards.¹⁷ Each study was assigned a quality rating of “good,” “fair,” or “poor.” Discordant quality ratings were resolved by discussion or by a third reviewer and adjudicated as needed. Studies were rated as poor quality and excluded if there was a major flaw such as very high attrition (generally >40%), differential attrition between intervention groups (generally >20%); substantial lack of baseline comparability between groups without adjustment; or major concerns about the trial conduct, analysis, or reporting of results. Poor-quality observational studies had multiple threats to internal validity and were excluded from the review. One reviewer extracted data from all included studies rated as fair- or good quality directly into summary tables, and a second reviewer checked the data for accuracy.

Subpopulations of interest were selected a priori based on the previous review and recommendation statement, established characteristics associated with the development of AAA, and feedback received from three key informants during the scoping phase. The subpopulation approach described in Whitlock et al¹⁹ was followed to audit outcomes and rate the credibility of the subpopulation data provided by included studies.

Data Synthesis and Analysis

To evaluate the effectiveness of screening for AAA, all-cause mortality and AAA-related mortality, rupture, and emergency surgical procedures were examined in RCTs that compared screening versus no screening. The primary analysis for all-cause mortality pooled calculated risk ratios using the DerSimonian & Laird²⁰ (DL) random effects model since statistical heterogeneity was very low (I²=0%, τ² = 0.0). The Peto method was used to pool odds ratios for AAA-related mortality, rupture, and emergency surgical procedures because events were very rare and trials had a similar number of participants in both study groups.²¹ One study used a competing risk analysis for calculating an incident rate ratio for AAA mortality which produced comparable findings to the odds ratio calculated for this review.

Meta-analyses of the rescreening studies included in KQ2 were not conducted because of substantial differences in patient population, length of follow-up, and outcomes reported.

To analyze the harms of screening versus no screening in KQ3, 30-day mortality after elective surgery, 30-day mortality after emergency surgery, overall operations, elective operations, emergency operations, and quality of life (QOL) measures were examined. Only two trials reported 30-day mortality after elective surgery and 30-day mortality after emergency surgery outcomes; therefore, they were not pooled. The Peto method was used to pool overall operations, elective operations, and emergency operations, as described under KQ1. Because of the substantial difference in quality-of-life measurements and insufficient reporting of data (e.g., lack of variation parameters), these data were unable to be pooled in the studies of screening versus no screening.

All statistical testing was two-sided and 0.05 was considered as significant. Statistical heterogeneity was examined across trials with the I² statistic and chi-square test of heterogeneity. Stata version 15.1 (StataCorp LP, College Station, TX) was used for all analyses.

The effectiveness of early intervention (KQ4) and associated harms of treating small AAAs (KQ5) was evaluated by capturing AAA growth, all-cause mortality, AAA-related mortality, and aneurysm ruptures. The data were narratively described and presented in data tables. Meta-analyses were not conducted due to the small number of studies of each intervention type.

The strength of evidence was rated for each key question based on consistency (similarity of effect direction and size), precision (degree of certainty around an estimate), reporting bias (potential for bias related to publication, selective outcome reporting, or selective analysis reporting), and study quality (i.e., study limitations).

RESULTS

Two reviewers evaluated 3,946 citations and 137 full-text articles against inclusion criteria, and 33 studies (69 articles)^6,8,22–89 met inclusion criteria for this systematic review (Figure 2). Nine new studies were included (4 RCTs,^24,58,80,90 2 cohort studies,^68,81 and 3 registry studies^27,61,71) and 24 studies (13 RCTs, 8 cohort studies, 1 case control study, and 2 registry studies) were carried forward from the previous USPSTF report.

Figure 2. Literature Flow Diagram.

AAA, abdominal aortic aneurysm; Jan, January; Sep, September; USPSTF, U.S. Preventive Services Task Force

^a Reasons for exclusion:

Aim: Study aim was not relevant

Setting: Study was not conducted in a country relevant to U.S. practice, or not conducted in, recruited from, or feasible for primary care or a health system

Population: Study was not conducted in an included population

Outcomes: Study did not have relevant outcomes or had incomplete outcomes

Intervention: Intervention was out of scope

Study Design: Study did not use an included design

Comparative Effectiveness: Active comparator

Quality: Study was poor quality

Publication Type: Abstract-only, non-English publication

Benefits of Screening

Key Question 1. What are the effects of one-time screening for AAA on health outcomes in an asymptomatic population aged 50 years or older?

Two fair-^6,23 and two good-quality^55,84 population-based screening RCTs assessed AAA screening effectiveness on AAA-specific mortality and all-cause mortality (Table 1): the Multicentre Aneurysm Screening Study (MASS);^{22,45,46,84,85} the Chichester, United Kingdom, screening trial;^23,77,79,87 the Viborg County, Denmark, screening trial;^54–57,59 and the Western Australia screening trial. ^{6,66,67,82,89} The trials randomized participants to either an invitation to one-time ultrasound screening or a usual care control group. All trials defined AAA as an aortic diameter of ≥3.0 cm and AAA prevalence varied from 4% to 7.6%, with the majority of screen-detected AAAs being small in size (≤4 to 4.5 cm). Mean or median follow-up in these four population-based screening trials ranged from 12.8 to 15 years, with short-term results published at 3- to 5-year intervals.

Table 1.

Characteristics and outcomes of AAA screening trials

Trial name Quality	Participants randomized Country	Eligible Age (mean, years)	AAA prevalence, %	Mean follow-up, years	All-cause Mortality,a RR (95% CI)	AAA Mortality,a OR (95%CI)	AAA rupture,a OR (95%CI)	All Operations,a OR (95%CI)	Elective Operations,a OR (95%CI)	Emergency operationsa, OR (95%CI)	30-day mortality after elective surgery, RR (95% CI)	30-day mortality after emergency surgery, RR (95% CI)
Chichester23 Fair	6,040 men 9,342 women UK	65–80 (72.0b)	Men: 7.6 Women: 1.3	Men: 15.0b Women: 10.0b	Men: 1.0 (0.97–1.04)	Men: 0.88 (0.60–1.31)	Men: 0.87 (0.60–1.25)	Men: 1.45 (0.97–2.17)	Men: 2.13 (1.28–3.55)	Men: 0.77 (0.41–1.47)	NR	NR
Viborg59 Good	12,639 men Denmark	64–73 (67.7)	3.9	13.0	0.98 (0.95–1.02)	0.37 (0.24–0.59)	0.46 (0.27–0.79)	1.24 (0.93–1.64)	1.97 (1.40–2.78)	0.47 (0.29–0.77)	NR	NR
Western Australia6 Fair	41,000 men Australia	64–83 (72.6)	7.2	12.8	0.99 (0.97–1.01)	0.92 (0.69–1.22)	0.73 (0.54–0.98)	1.23 (1.09–1.40)	1.30 (1.14–1.48)	0.60 (0.37–0.95)	0.76 (0.40–1.45)	1.43 (0.90–2.25)
MASS84 Good	67,800 men UK	65–74 (69.2)	4.9	13.1	0.98 (0.96–1.00)	0.59 (0.50–0.70)	0.58 (0.50–0.67)	1.54 (1.37–1.73)	2.11 (1.85–2.41)	0.50 (0.39–0.64)	0.82 (0.43–1.57)	0.98 (0.68–1.43)
VIVA8,58 Fair	50,156 men Denmark	65–73 (69.0b)	3.3	4.4a	NRc	NRc	NR	1.87 (1.54, 2.26)	2.27 (1.84, 2.81)	0.82 (0.53, 1.27)	NR	NR
Pooled Estimate	-	-	-	-	0.99 (0.98–1.0)	0.65 (0.57–0.74)	0.62 (0.55–0.70)	1.44 (1.34–1.55)	1.75 (1.61–1.90)	0.57 (0.48–0.68)	Not calculated	NA
I2	-	-	-	-	0.0%	79.7%	53.3%	74.1%	88.5%	26.7%	Not calculated	NA

^aCalculated

^bMedian

^cDid not include the Viborg Vascular (VIVA) trial mortality data due to the inability to capture the independent contribution of AAA screening within the multicomponent screening program

Abbreviations: AAA = abdominal aortic aneurysm; CI = confidence interval; MASS = Multicenter Aneurysm Screening Study; NA = not applicable; NR = not reported; OR = odds ratio; RR = relative risk; UK = United Kingdom; VIVA = Viborg Vascular trial

One additional new population-based screening trial in Denmark (VIVA) was included solely for the outcome of number of operations.⁵⁸ VIVA randomized participants to a multicomponent screening versus no screening for hypertension, peripheral artery disease (PAD), and AAA. Participants with confirmed AAA or PAD were counseled to initiate preventive interventions, with aspirin and statin therapy prescribed to those meeting a total cholesterol threshold value.⁵⁸ The effects of AAA screening alone could not be independently assessed with respect to all-cause mortality or AAA mortality because multicomponent screening and cardiovascular disease (CVD)-prevention interventions were administered; however, the number of procedures was included in this review as they would almost exclusively be expected to be due to AAA screening.

AAA-specific mortality in men was the primary outcome of the four screening trials. A meta-analysis of the trials^6,23,59,84 (N=124,929) estimated a statistically significant reduction over 12 to 15 years of follow-up in AAA-specific mortality associated with an invitation to screening with high heterogeneity (Peto OR 0.65 [95% CI, 0.57–0.74], I²=80%, number needed to screen 305 men [95% CI, 248 to 411]) (Figure 3). A meta-analysis of all-cause mortality in men from the four screening trials^6,23,59,84 (N=124,929) did not reach statistical significance (RR 0.99 [95% CI, 0.98–1.00]; I²=0%) (Figure 4). Only the MASS trial reported a statistically significant reduction in all-cause mortality (HR 0.97 [95% CI, 0.95–0.99]).⁸⁴

Figure 3. Pooled analysis of AAA-related mortality and ruptures (male-only) for rupture in one-time screening trials.

CG, control group; CI, confidence interval; IG, intervention group; N, population size; n, sample size; OR, odds ratio

Figure 4. Pooled Analysis of all-cause mortality (male-only) in one-time screening trials.

CG, control group; CI, confidence interval; IG, intervention group; N, population size; n, sample size; RR, relative risk

In addition to mortality outcomes, the screening trials reported ruptures and emergency operations among primarily male study populations. Pooled results of these four trials^6,23,59,84 (N=124,929) showed a statistically significant reduction in AAA rupture associated with the an invitation to screening (Peto OR 0.62 [95% CI, 0.55–0.70]; I²=53%, number needed to screen 246 men [95% CI 207 to 311]) (Figure 3). An invitation to screening in the 5 trials^{6,23,58,59,84} (N=175,085) was also associated with a statistically significant reduction in emergency operations in the screening group (Peto OR 0.57 [95% CI, 0.48–0.68]; I²=27%) (Figure 5). This would reduce the number of emergency procedures by 2 per 1000 men screened (95% CI, 2 to 2).

Figure 5. Pooled analysis of operations (male-only) in one-time screening trials.

CG, control group; CI, confidence interval; IG, intervention group; N, population size; n, sample size; OR, odds ratio

Patients invited to participate in the screening trials were predominantly men. Only the Chichester^23,77 trial examined AAA screening in women (59% of participants were women; n=9,342 women), showing that women had a lower AAA prevalence compared with men (1.3% vs. 7.6%).^77,79 There was no significant difference for women in AAA-related or all-cause mortality at 5 years (AAA mortality: 0.06% vs. 0.04%; all-cause mortality: 10.7% vs. 10.2%) or AAA rupture rate at 10-year follow-up (0.2% in both groups) between the invited and control groups, but the trial was underpowered.

KQ2. What are the effects of rescreening for AAA on health outcomes or AAA incidence in a previously screened, asymptomatic population without AAA on initial screening?

No trial-level evidence examined the effectiveness of one-time screening plus rescreening compared with one-time screening alone. Seven cohort studies (5 fair-quality,^{32,34,78,81,83} 2 good-quality^{30,31,36,49,63,68}) and one fair-quality case-control study⁶⁰ recruited screen-negative participants (2.5 or 2.6 to 2.9 cm,^{32,34,60,68,81,83} less than 2.5 cm,⁸³ or less than or equal to 3 cm ^49,78,83) and administered various rescreening protocols (rescreening every 1 to 5 years with 1 to 6 repeated scans) reporting the proportion of initially screen-negative aortas that reached 5.0 or 5.5 cm at the repeat scan (eTables 2 and 3 in the Supplement). This group of heterogeneous studies reported that AAA-related mortality over 5 to 12 years was rare (<3%) among those with normal aortas (<3 cm) on the initial scan. Upon rescreening, few aortas grew to >5 cm (0 to 2.2%) at 5 years^32,34,49 and 0% to 15% had progressed at 10 years (eTable 3).³¹ Four studies reported no AAA ruptures or AAA-related deaths^32,49,81,83 at 4- to 5-years’ follow-up; one population screening program reported 2.4% ruptures at 7.9-years median follow-up (eTable 2).³¹ Overall, this heterogenous body of literature was too limited to make conclusions about the effectiveness of rescreening.

Harms of Screening

KQ3. What are the harms associated with one-time and repeated screening?

Two population-based screening trials reported no statistically significant difference in 30-day operative mortality from elective (RR 0.76 [95% CI 0.40=1.45]; RR 0.82 [95% CI 0.43–1.57]) and emergency surgical procedures (RR 1.43 [95% CI 0.90–2.25]; RR 0.98 [95% CI 0.68–1.43]) among those invited to screening compared with those in the control group at 12.8- to 13.1-year follow-up^6,84 (Table 1). All five screening trials reported more AAA-related operations in the invited group than the control group, with 1.1% to 2.9% of the screened group undergoing surgical repair compared with 0.6% to 2.4% of the control group.^{6,23,58,59,84} The pooled data estimated significantly more procedures in the invited group compared with the control group (Peto OR 1.44 [95% CI, 1.34–1.55]; I²=74%) (Figure 5). Implementing a screening program would increase the total number of operations per 1000 men by 6 (95% CI, 5 to 8). Elective operations were also consistently more common in the screened group (1.0% to 2.8%) than in the control group (0.4% to 2.2%) in all 5 trials,^{6,23,58,59,84} (Figure 5, Table 1). Pooled analysis of these trials confirmed a higher elective operation rate in the screened group than in the control group (Peto OR 1.75 [95% CI, 1.61–1.90]; I²=89%) (Figure 5, Table 1). This would increase the number of elective operations by 8 per 1000 men screened (95% CI, 6 to 9).

There were no RCTs assessing the harms of rescreening versus no rescreening in those with normal sized aortas (<3.0 cm) on initial screening. Six fair-quality cohort studies examined procedure rates in rescreened cohorts.^{32,49,60,68,81,83} Five of these studies showed a low procedure rate (0%- to 4%) at up to 5-year follow-up; ^{32,49,60,81,83} a single screening program reported a higher procedure rate of 10.9% at 7.8-years mean follow-up (eTable 2 in the Supplement).⁶⁸

Two subsamples of screening RCTs^6,22,82,84 and three small cohort studies^52,62,88 had mixed results but generally showed no substantial differences in QOL or mood scores between screen-positive and screen-negative participants at up to 12 months’ follow-up; one of these RCTs (MASS) reported lower QOL scores at 6 weeks but all scores were within age-matched population normal standards.^22,84

Benefits of Treatment for Small AAAs

KQ4: What are the effects of treatment on intermediate and health outcomes in an asymptomatic, screen-detected population with small AAAs (i.e., aortic diameter of 3.0 to 5.4 cm)?

Early surgical repair versus surveillance of 4–5.4 cm AAAs

Four trials evaluated the effectiveness of immediate surgical repair of small aneurysms (4 to 5.4 cm) versus surveillance every 3 to 6 months until the aneurysm reached 5.5 cm, rapidly expanded (>1 cm/year), or became symptomatic. The Aneurysm Detection and Management trial (ADAM) and the UK Small Aneurysm Trial (UKSAT) evaluated the effectiveness of early open surgery,^51,75 and Comparison of Surveillance versus Aortic Endografting for Small Aneurysm Repair (CAESAR)²⁹ and Positive Impact of endoVascular Options for Treating Aneurysm earLy (PIVOTAL) trial⁷⁰ evaluated EVAR interventions^29,70 (Table 2).

Table 2.

Characteristics and outcomes of Open and EVAR AAA Surgical trials

Trial name	Surgical Approach	Participants randomized	Country	Eligible Age (mean, years)	Female, %	AAA size for inclusion and surveillance, cm	Mean follow-up, years	All-cause Mortality,a RR (95% CI)	AAA Mortality,a RR (95%CI)	AAA Rupture,a RR (95%CI)	Participants undergoing AAA procedure, No. (%)
UKSAT75	Open Repair	1090	UK	60–76 (69.3)	17.5	4.0–5.5	12	0.96 (0.88–1.05)	0.67 (0.45–1.02)	0.51 (0.26–0.99)	IG: 528 (93.8) CG: 401 (76.1)
ADAM51	Open Repair	1136	US	50–79 (68.1)	0.8	4.0–5.4	4.9	1.17 (0.95–1.44)	1.13 (0.57–2.24)	0.18 (0.04–0.81)	IG:527 (92.6) CG: 349 (61.6)
CAESAR29	EVAR	360	Europe and Asia	50–80 (68.9)	4.2	4.1–5.4	2.7b	1.22 (0.49–3.03)	0.98 (0.06–15.52)	Not calculated	IG: 175 (96.2) CG: 85 (47.8)
PIVOTAL70	EVAR	728	US	40–90 (70.5)	13.4	4.0–5.0	1.7	0.99 (0.49–1.99)	1.98 (0.18–21.72)	Not calculated	IG: 326 (89.1) CG: 112 (30.9)

^aCalculated

^bMedian

Abbreviations: AAA = abdominal aortic aneurysm; ADAM = Abdominal aortic aneurysm Detection and Management study; CAESAR = Comparison of surveillance versus Aortic Endografting for Small Aneurysm Repair; CG = control group; cm = centimeter; CI = confidence interval; EVAR = endovascular aneurysm repair; IG = intervention group; NA = not applicable; No. = number; NR = not reported; PIVOTAL = Positive Impact of Endovascular Options for Treating Aneurysms Early; RR = relative risk; UK = United Kingdom; UKSAT = the UK Small Aneurysm Trial; US = United States

The CAESAR²⁹ and PIVOTAL trials⁷⁰ terminated recruitment early due to interim analysis concluding intervention futility, but they completed scheduled follow-up visits in those who had already been enrolled.

The 4 early surgical trials found no significant differences in all-cause or AAA-specific mortality at any follow-up time between those receiving early surgical repair versus those under surveillance (Table 2).^51,74,75 An individual patient data (IPD) analysis (n=2,226) of the two open repair trials additionally supported no survival benefit (adj HR 0.99, [95% CI, 0.83 to 1.18]).³⁷ Ruptures were rare events in all surgical trials; however, those who underwent early open repair had a significant reduction in the rate of rupture compared with those who underwent surveillance at each follow-up interval (RR 0.18 [95% CI 0.04–0.81] at 4.9 years n=1136; RR 0.33 [95% CI 0.13–0.83] at 4.6 years n=1090; 0.51 (95% CI 0.26–0.99] at 12 years n=1090) (Table 2).^51,74,75 There were only 3 ruptures in the EVAR trials, making comparisons challenging. Overall there were more surgical interventions in the early surgery groups than in the surveillance groups composed mostly of elective surgical procedures (Table 2).

Seven short-term drug trials (N= 1,553) of antibiotic, antihypertensive medications, and mast cell stabilizers showed no overall effect on AAA growth compared with placebo. Details are provided in the full evidence report and in eTable 4 of the Supplement.

KQ 5. What are the harms of treatment in an asymptomatic, screen-detected population with small AAAs (i.e., aortic diameter of 3.0 to 5.4 cm)?

The four trials of early surgery^29,51,70,75 and five registry publications reported complication rates for surgical patients with AAAs <5.5 cm^{27,40,61,71,72} (Tables 3–4).

Table 3.

Harms associated with Open AAA repair, trials and registry data

Trial or registry name	Study design	Mean follow-up, yrs	No. of participants analyzed	30-day mortality after elective repair, no. (%)	Reintervention rates, no. (%)	Readmission rates in 30 days, no. (%)	Major complications
ADAM51	RCT	4.9	Total n = 866 IG: 526 CG: 340	IG: 11 (2.1)a CG: 6 (1.8)a	IG: 9 (1.7) CG: 4 (1.2)	IG: 108 (20.5)d CG: 56 (16.5)d	IGd: Any major complication: 24 (4.6) f MI: 5 (1.0)f Stroke: 3 (0.6) Pulmonary embolism: 4 (0.8) CGd: Any major complication: 26 (7.6) f MI: 13 (3.8)f Stroke: 2 (0.6) Pulmonary embolism: 1 (0.3)
UKSAT75	RCT	12	Total n = 915 IG: 526 CG: 389	IG: 26 (5.0) CG: 25 (6.3)	NR	IG: 30 (6.3)e CG: NR	NR
ACS NSQIP71	Registry	NR	Total n = 5126	25 (3.5)b	64 (9.1)	44 (6.2)	MI: 24 (3.4) Stroke: 5 (0.7) Pulmonary embolism: 3 (0.4) Overall morbidity within 30 days of surgery: 69.4% Bleeding complications: 460 (65.2)
Vascunet27	Registry	NR	Total n = 12,610	391 (3.1)c	NR	NR	NR

^aOperative mortality associated with the repair of unruptured AAA

^bN analyzed for 30-day operative mortality; Open n=705

^cDefined as hospital death or death within 30 days of surgery

^dFollow-up timing NR

^eFrom 1 -year follow-up data; the use of bifurcated grafts (12/30, 40%) was associated with a 2-fold increase in the risk of reoperation, p=0.03

^fp< 0.05

Abbreviations: AAA = abdominal aortic aneurysm; ACS NSQIP: American College of Surgeons National Surgical Quality Improvement Program; ADAM = Abdominal aortic aneurysm Detection and Management study; CG = control group; CI = confidence interval; IG = intervention group; MI = myocardial infarction; No = number; NR = not reported; UKSAT = the UK Small Aneurysm Trial

Table 4.

Harms associated with EVAR AAA repair, trials and registry data

Trial or registry name	Study design	Mean follow-up, yrs	No. of participants analyzed	30-day mortality after elective repair, no. (%)	Reintervention rates, no. (%)	Readmission rates in 30 days	Endoleaks at 30 days, no. (%)	Endoleaks at 1 year, no. (%)	Major complications within 30 days, no. (%)
CAESAR29	RCT	2.6a	Total n = 260 IG: 175 CG: 85	IG: 1 (0.6) CG: 0 (0)	IG: 10 (5.7)d CG: 0 (0)d	IG: NR CG: NR	IG: 28 (16.0)h CG: 7 (8.2)h	IG: 21 (12.0)h CG: 2 (2.4)h	IG: Any major morbidity: 6 (3.4) CG: Any major morbidity: 4 (4.7)
PIVOTAL70	RCT	1.7	Total n = 431 IG: 322 CG: 109	IG: 1 (0.3) CG: 1 (0.9)	IG: 12 (3.7)e CG: 5 (4.6)e	20 (4.6)g	IG: 36 (11.9) CG: 10 (10.3)	IG: 72 (26.1) CG: 30 (35.1)	IG: Serious cardiac event: 17 (5.3) Serious pulmonary event: 4 (1.2) Serious renal event: 6 (1.9) CG: Serious cardiac event: 9 (8.3) Serious pulmonary event: 1 (0.9) Serious renal event: 1 (0.9)
ACS NSQIP71	Registry	NR	Total n = 705	31 (0.7)b	150 (3.4)	304 (6.8)	NR	NR	Acute renal failure: 15 (0.3) Sepsis: 20 (0.4) Septic shock: 6 (0.1) Cardiac arrest: 14 (0.3) MI: 46 (1.0) Pulmonary embolism: 5 (0.1) Stroke: 12 (0.3) Overall morbidity within 30 days of surgery: 11.4% Bleeding complications: 296 (6.6)
ASERNIP-S40	Registry	3.2a	Total n = 478	5 (1.1)	13 (3)f	NR	46 (9.6)	97 (20.3)i	Significant postop complications reported in 138 (29.0) 72 systemic complications were noted in 64 (13.4) of patients
EUROSTAR72	Registry	1.7	Total n = 1962	31 (1.6)	NR	NR	NR	NR	30-day systemic complications combined: 235 (12.0) Cardiace: 55 (2.8) Pulmonarye: 31 (1.6) Early procedure or device-relatede: 57 (2.9)
Vascular Study Group of New England61	Registry	1.0	Total n = 1336	Men: 7 (0.7) Women: 4 (1.1)	NR	NR	NR	NR	NR
Vascunet27	Registry	NR	Total n = 12,610	88 (0.7)c	NR	NR	NR	NR	NR

^aMedian

^bN analyzed for 30-day operative mortality; Open n=705, EVAR n=4471

^cDefined as hospital death or death within 30 days of surgery

^dp< 0.05

^eFollow-up timing NR

^fReinterventions ≤ 30 days after surgery; ≥ 30 days after surgery: 50 patients underwent 72 additional interventions by open (20 times in 16 patients [5 had an EVAR procedure]),

^gGroup not specified for readmission rates

^hOperative mortality associated with the repair of unruptured AAA

ⁱEndoleak on follow-up imaging ≥ 30 days after procedure

^jAt 4 years: Type I proximal: 5.3%; distal: 11.3%; Type III: 14.4%

Abbreviations: AAA = abdominal aortic aneurysm; ACS NSQIP: American College of Surgeons National Surgical Quality Improvement Program; ASERNIP-S = Australian Safety and Efficacy Register of New Interventional Procedures-Surgical; CAESAR = Comparison of surveillance versus Aortic Endografting for Small Aneurysm Repair; CG = control group; CI = confidence interval; EVAR = endovascular aneurysm repair; IG = intervention group; MI = myocardial infarction; No = number; NR = not reported; PIVOTAL = Positive Impact of Endovascular Options for Treating Aneurysms Early

Both the ADAM trial and UKSAT reported no significant difference in 30-day postoperative mortality rates in the early open repair and surveillance groups (2.1% vs. 1.8%; 5.0% vs 6.3%) (Table 3).^51,75 The two largest and most contemporary registries (2011 to 2015; 2005 to 2013) capturing open repairs of small aneurysms reported a 30-day operative mortality rate within the range reported in these trials^27,71 (3.1% and 3.5%) (Table 3). Thirty-day operative mortality after EVAR in both the CAESAR and PIVOTAL trials was rare (Table 4).^29,70 The two largest and most contemporary registries (2011 to 2015; 2005 to 2013) capturing EVAR of small aneurysms reported a 30-day operative mortality rate for EVAR of 0.7% (Vascunet and NSQIP).^27,71 The two oldest registries reported slightly higher mortality rates from EVAR at 1.1% (ASERNIP) and 1.6% (EUROSTAR) (Table 4).^40,72

Only the ADAM trial reported adverse event rates for the open repair intervention and control groups, and results were mixed. The rate of 30-day readmissions was not significantly different between surgery and surveillance groups, nor were the overall complication rates significantly different (Table 3).⁵¹ Further, the event rate for total major complications was higher in the surveillance group than the early treatment group (7.6% to 4.6%, no statistical testing reported), with a significantly higher risk of surgery-related myocardial infarction reported in the surveillance group (1.0% vs 3.8%, p=0.004). The NSQIP registry reported overall 30-day morbidity for open repair as approximately 69.4% at 30 days postintervention, with the most common complication being bleeding (Table 3).⁷¹

Complications were variably reported in the two EVAR trials.^29,70 In the CAESAR trial, the percentage of patients with any adverse events was significantly higher at 32.4-month follow-up in the early EVAR group compared with the surveillance group (19.1% vs 5.1%, p<0.01). In addition, the percentage of patients with any morbidity related to repair at 30 days was also higher in the EVAR group compared with the surveillance group (17.7% vs. 6.0%, p=0.01) (Table 4).²⁹ Rates of any major morbidity (3.4% vs 4.7%) and 30-day endoleaks (16% vs 9.9%) were not significantly different in the early EVAR compared with the surveillance groups, but the early EVAR group had significantly more endoleaks at 1 year (12% vs 2.8%, p=0.03) and significantly more reinterventions than the group undergoing surveillance (5.7% vs 0%, p=0.03). The PIVOTAL trial reported rates of endoleaks in the early intervention and surveillance groups at 30 days postintervention (11.9% vs 10.3%) and at 1 year (26.1% vs 35.1%), but statistical testing was not reported (Table 4).

Two EVAR registries (ASERNIP⁴⁰ and EUROSTAR⁷²) and the one registry of both open and EVAR (NSQIP)⁷¹ reported complication rates following intervention on small AAAs (Table 4). Registry data reported a composite of major or systemic complication rates for EVAR ranging from 12% to 29% at 30 days postintervention, which is consistent with trial data.^29,70–72 ASERNIP and NSQIP reported reintervention rates within 30 days of EVAR of approximately 3%;^40,71 these rates are likewise comparable to trial data.^29,70 The NSQIP reported readmission rates for small AAAs at 30-days postintervention as 6.8% following EVAR.⁷¹ Readmission rates were not reported in the trials, so these data cannot be compared with trial findings. ASERNIP and EUROSTAR reported the occurrence of endoleaks at 20% and 31%, respectively, at 3- to 4-year follow-up.^40,72

Eight short-term drug trials (N=1,598) reported high adverse event-related discontinuation rates with propranolol (38% and 60% of the propranolol groups withdrew from the trials); other medications (including other antihypertensive medications [ACE-inhibitors, Ca-channel blocker and antibiotics]) appeared to be well-tolerated based on few trial withdrawals were reported from 1 to 2 studies per drug class (eTable 5 in the Supplement).

Subpopulations

There was limited credible subpopulation information from the body of included studies. The Western Australia trial⁶ showed that smoking was associated with a higher risk of all-cause mortality (OR 1.59 [95% CI, 1.47 to 1.72]) and AAA-related mortality (OR 2.95 [95% CI, 1.04 to 8.43]) in the screened group, but no included study examined differential screening benefits by smoking status or family history. Viborg and Western Australia trials’ subgroup analyses suggested that there is no differential screening effect on mortality by age.^6,59 The Chichester trial,^23,77 recruited 9,342 women and showed a lower prevalence of AAA in women compared with men (1.3% vs. 7.6%) with most screen-detected AAAs measuring 3.0 to 3.9 cm.^77,79 There was no significant difference in AAA rupture rates among women at 10-year follow-up (0.2% in both the screened and unscreened group) or in AAA-related (0.06% vs. 0.04%) or all-cause mortality (10.7% vs. 10.2%) at 5 years between the invited and control groups; however, the trial was underpowered. Based on two open repair trials, there was no differential effect on all-cause mortality by sex (KQ4a).^51,74 Registry data, however, showed a higher rate of postoperative mortality following elective repair of small AAAs in women compared with men, regardless of the surgical technique (KQ5a).^27,40,61

DISCUSSION

Since the previous systematic review for the USPSTF,¹⁶ the following new data was included in this review: 1) the final long-term follow-up from the Western Australia trial was added to the meta-analysis confirming prior AAA mortality benefits of screening;⁶ 2) two new small rescreening cohort studies offered little additional information to a heterogeneous literature;^68,81 3) three additional pharmacotherapy trials showed no benefit in halting AAA growth;^24,80,90 4) one new population-based screening trial (VIVA) added to the meta-analysis on additional operations associated with screening confirmed previous results;⁵⁸ and 5) three contemporary registries^27,61,71 provided complication rates from EVAR and open repair generally comparable to those cited in the included trials.

A summary of the evidence by Key Question is provided in Table 5. The meta-analyses demonstrated that offering one-time screening to men ages 65 to 75 years was associated with lower AAA-related mortality, AAA rupture, and emergency surgical procedures over 13 to 15 years of follow-up (KQ1), but does not resolve the question of all-cause mortality benefit. In terms of harms, screening for AAA was shown to expose patients to more procedures, which was primarily driven by elective operations. Overdiagnosis and overtreatment were not addressed in these trials but may be important considerations given that most screen-detected aneurysms are small.

Table 5.

Summary of Evidence

Key Question	Subpopulation	No. of Studies (No. of Observations)	Summary of Findings	Consistency and Precision	Other Limitations	Strength of Evidence	Applicability
KQ1: What are the effects of one-time screening for abdominal aortic aneurysm (AAA) on health outcomes in an asymptomatic population age 50 years or older?	Entire study population	4 (124,929)	Invitation for one-time screening in men aged 65 years and older was associated with a 35% reduction in AAA-related mortality, 38% reduction in AAA rupture rate, and a 43% reduction in the number of emergency operations, but no statistically significant difference in all-cause mortality at 12 to 15-year follow-up.	Reasonably consistent and precise	--	Moderate to high	These population-based screening trials were set in the 1990s in mostly Caucasian males aged 65 to 75 years. Since this time, AAA prevalence has declined along with smoking prevalence, and medical management of CVD has changed.
KQ1a: Do the effects of one-time screening for AAA vary among subpopulations (i.e., by age, sex, smoking status, family history, or race/ethnicity)?	Age	2 (51,119)	The Viborg and Western Australia population-based screening trials reported subanalyses with substantial limitations suggesting that there is no differential screening effect based on age.	Consistent; imprecise	Subanalyses were not prespecified, stratified at randomization to ensure baseline comparability, adjusted for confounders, or powered to detect subgroup differences; no heterogeneity testing was performed.	Low	Both trials were conducted in mostly Caucasian older men.
	Sex	1 (9,342)	There was no significant difference in AAA rupture rate at 10-year follow-up, or in mortality (both AAA-related & all-cause) at 5 years between the invited and control groups. Most AAA ruptures occur ≥ age 80 years in women.	Consistency NA (1 study) imprecise	Underpowered for health outcomes	Low	Population was older Caucasian women in Chichester, UK.
	Smoking	1 (19,249)	Results showed that smoking was associated with a higher risk of all-cause mortality (OR 1.59 [95% CI, 1.47 to 1.72]) and AAA-related mortality (OR 2.95 [95% CI, 1.04 to 8.43]) in the screened group of men aged 64 to 83 years. This trend was more pronounced in the age group of 65 to 74-year olds, however no formal analysis was performed to explore if there is a differential screening effect based on smoking status.	Consistency NA (1 study); Imprecise	Subanalyses were not prespecified or powered to detect subgroup differences; no heterogeneity testing was performed to determine if there is modification in the effectiveness of screening among smokers.	Low	Trial was conducted in mostly Caucasian older men.
KQ2: What are the effects of rescreening for AAA on health outcomes or AAA incidence in a previously screened, asymptomatic population without AAA on initial screening?	Entire study population	8 (8,018)	Studies rescreened participants with various rescreening protocols (rescreening annually to 5 years with 1 to 6 repeated scans), demonstrating that AAA-related mortality over 5 to 12 years is rare (<3%) among those with normal aortas (< 3 cm) on the initial scan. Upon rescreening, few aortas grew to > 5 or ≥5.5 cm (0 to 2.2%) at 5 years and 0 to 15% had progressed at 10 years. Four studies reported no AAA ruptures or AAA-related deaths at 4 to 5-year follow-up; one study reported 2.4% ruptures at 7.9-year median follow-up.	Inconsistent; Imprecise	Heterogeneous rescreening protocols. A small number of participants with normal aortas were included in these studies; all but a single study was exclusively in men; there are no matched controls in most studies; the primary focus of most studies was growth rate as the follow-up time for most studies was 5 years, which is a timeframe too short to expect the development of AAA-related health outcomes.	Low	Mostly men (only one trial conducted in women). All but one trial conducted outside of the US.
KQ2a: Do the effects of rescreening for AAA vary among subpopulations (i.e., by age, sex, smoking status, family history, or race/ethnicity)?	Sex	1 (33) Remaining 7 KQ2 studies in men	One small study in women was too small to compare to other rescreening studies in men.	Consistency NA; Imprecise	Few studies reporting outcomes by subpopulation; most studies did not report if subgroup analyses were prespecified and studies tested numerous risk factors. Overall rescreening literature limited by lack of adequate reporting; heterogeneity of study rescreening protocols; and short follow-up times with focus on growth rates rather than health outcomes.	Low	Small study of women conducted outside of the US.
	Smoking	3 (4,706)	Two studies report multi-regression analyses suggesting that current smoking is an independent risk factor for the development of AAA at rescreening and another univariate analysis in a study of solely women shows a similar trend for smoking, however the number developing AAAs over the rescreening interval was small.	Consistent; Imprecise		Low	Studies were conducted in mostly men.
KQ3: What are the harms associated with one-time and repeated screening?	Entire study population	5 (175,085)	There were more procedures in the invited group compared to the control group (5 studies, Peto OR 1.44 [95%CI, 1.34 to 1.55]), largely driven by elective operations (Peto OR 1.75 [95% CI, 1.61 to 1.90]). There was no statistically significant difference in 30-day operational mortality rates in the invited versus control groups for either elective surgical procedures or emergency surgical procedures at the 12 to 15-year follow-up.	Consistent; reasonably precise	---	Moderate	Trials were conducted in mostly Caucasian older men.
KQ3a: Do the harms of one-time and repeated screening for AAA vary among subpopulations (i.e., by age, sex, smoking status, family history, or race/ethnicity)	Age	1 (19,571)	Single population-based screening trial reports no significant difference in the number of elective operations and lower 30-day operative mortality after elective and emergency surgery in the younger age subset (65 to 74 years) compared to entire trial population (64 to 83 years).	Consistent NA 1 trial; Imprecise	Subanalyses were not prespecified, stratified at randomization to ensure baseline comparability, adjusted for confounders, or powered to detect subgroup differences; no heterogeneity testing was performed.	Low	Trial was conducted in mostly Caucasian older men.
KQ4: What are the effects of treatment (pharmacotherapy or surgery) on intermediate and health outcomes in an asymptomatic, screen-detected population with small AAAs (i.e., aortic diameter of 3.0 to 5.4 cm)?	Entire study population: open versus surveillance	2 (2,226)	No significant difference in all-cause mortality, AAA-related mortality but reduction in rupture seen with early open surgery compared to surveillance for small AAAs.	Consistent; Imprecise	Only 2 studies. No differentiation of sizes between 4 to 5.4 cm	Moderate	Trails primarily recruited males with small AAAs (4 to 5.4 cm).
	Entire study population: EVAR versus surveillance	2 (1,088)	No significant difference in all-cause or AAA-related mortality with early EVAR compared to surveillance for small AAAs.	Consistent; Imprecise	Both trials were stopped early at interim analysis due to futility.	Moderate	Trails primarily recruited males with small AAAs (4 to 5.4 cm).
	Entire study population: Pharmacotherapy versus placebo	7 (1,553)	Drug trials of antibiotics, antihypertensive medications, a mast cell stabilizer showed no overall effect on AAA growth compared to placebo.	Consistency NA given different drug classes; Imprecise	1 to 2 trials for each medication, follow-up times too short to expect development of AAA-related events or changes in health outcomes (all-cause mortality, AAA-related mortality, rupture)	Low	Studies were predominantly conducted in males with small AAAs. All trials were conducted outside of the US.
KQ4a: Do the effects of treatment of small AAAs vary among subpopulations (i.e., by age, sex, smoking status, family history, or race/ethnicity)?	Open vs. surveillance by age & sex	2 (2,226)	IPD MA available pooling the two early open versus surveillance trials (ADAM and UKSAT) with 5 to 8-year follow-up reported no differential all-cause mortality effect by sex. The two trials reported no differential all-cause mortality treatment effect by age at 5 and 12-year follow-up. One of these RCTs (UKSAT) reports no significant difference in all-cause mortality by smoking status; this subanalysis did not address differential mortality effect of early surgery by smoking status. There is no data on family history or race.	Unknown consistency; Imprecise	Subanalyses by age: one trial did not report interaction testing results. Subanalyses by sex: both prespecified analyses, one trial did not adjust for confounders or report interaction testing	Low	Participants were men and women with small AAAs; analyses were separated by older and younger participants (reported in 5 to 10-year age strata), with and without a smoking history.
	EVAR vs. surveillance	No studies	---	---	---	---	---
	Pharmaco-therapy by age & smoking	Age: 1 (552) Smoking: 1 (32)	None of the pharmacotherapy trials report health outcomes by subgroup. One small doxycycline trial and one propranolol trial performed limited subgroup analyses which do not support a treatment effect modification by age or smoking history.	NA single trial for each medication/ imprecise	Too few studies. Available analyses would be considered exploratory, at best particularly given that the subgroup methodologies were of low quality (no prespecification of analysis, adjustment for confounders, or interaction testing) and overall trial results do not support a AAA growth benefit.	Low
KQ 5: What are the harms of treatment in an asymptomatic, screen-detected population with small AAAs (i.e., aortic diameter of 3.0 to 5.4cm)?	Open vs. surveillance	5 (21,298)	Two trials of early open repair vs. surveillance reported a 50% higher rate of procedures in the early intervention group with no significant difference in 30-day operative mortality. Readmission rates at 30 days were not significantly different; and major surgical complications were lower in the early intervention group. QOL results were mixed, but generally showed declines in both the early surgery and surveillance groups over time with no statistically significant difference between the groups up to 1 to 2 years; only the ADAM trial showed higher general health scores in the early repair group in the first 2 years that did not persist over time. One trial reported an increased incidence of impotence in the early repair group at up to 4 years. Registry harms data were generally comparable to the 2 trials with the exception of reintervention rates, which were higher in the registries compared to the ADAM trial.	Reasonably consistent for procedures, 30-day operative mortality; consistency NA for other complications as only reported in 1 trial Reasonably precise for procedures; 30-day mortality and complications imprecise	Surgical morbidity complications not well-reported	Moderate	Registry data from both national and international sources; contemporary.
	EVAR vs surveillance	7 (22,600)	Two trials of early EVAR vs. surveillance reported approximately 100% more procedures in the early intervention group and rare 30-day operative mortality rates between the groups. In the CAESAR trial, the early intervention group had a higher percentage of patients with any adverse events (19% vs. 5%, p<0.01), any morbidity related to repair at 30 days (18% vs. 6%, p=0.01), endoleaks at 1 year (12% vs. 3%, p=0.028), and reinterventions (6% vs. 0%, p=0.03), but rates of any major morbidity over the trial duration were not significantly different (3.3% vs. 2.8%, p=0.99). Conversely, PIVOTAL largely reported no significant difference in rates of adverse events at 30 days (12% vs. 10%) and 1 year (26% vs. 35%) and reinterventions (3.7% vs. 4.6%). Reported complication rates from registry data were generally comparable to those rates reported in the above trials for 30-day operative mortality and reinterventions.	Consistent for procedures, 30-day operative mortality, reinterventions, major morbidity; precise for procedures, 30-day mortality, reinterventions	Individual post-operative complications and major morbidities variably reported	Moderate	Registry data from both national and international sources; contemporary.
	Pharmaco-therapy vs. placebo	8 (1,598)	With the exception of the two propranolol trials reporting high adverse events related discontinuation rates (38% and 60% of the propranolol groups withdrew from the trials); other medications (including other antihypertensive medications [ACE-inhibitors, Ca-channel blocker] and antibiotics) appear to be well-tolerated based on few trial withdrawals reported from 1 to 2 studies per drug class.	Propranolol- consistent, imprecise; Doxycycline- consistent, imprecise; Other drugs-NA for consistency as 1 trial for each drug; imprecise	1 to 2 trials per drug class with limited harms reporting	Low	Predominantly male population with small AAAs. All trials conducted outside of the US.
KQ5a: Do the harms of treatment of small AAAs vary among subpopulations (i.e., by age, sex, smoking status, family history, or race/ethnicity)?	Sex	3 (14,424)	There is scant data on harms in subpopulations. No trial data are available to examine harms in subpopulations. Existing evidence shows higher 30-day operative mortality and secondary complications in women compared to men for both EVAR and open repair. No information available for other subpopulations.	Consistent; imprecise	Few registries & non-trial data, few outcomes (mostly 30-day operative mortality).	Moderate	Registry data from both national and international sources; contemporary.

Abbreviations: AAA = abdominal aortic aneurysm; AARDVARK = Aortic Aneurysmal Regression of Dilation: Value of ACE-Inhibition on RisK trial; ADAM = Abdominal aortic aneurysm Detection and Management study; CI = confidence interval; HR = hazard ratio; IQR = interquartile range; mm = millimeter(s); AORTA = the Anti-inflammatory Oral Treatment of AAA; CG = control group; CVD = cardiovascular disease; EVAR = endovascular aneurysm repair; IG = intervention group; MI = myocardial infarction; N = population size; NA = not applicable; NR = not reported; PAT = propranolol Aneurysm Trial; PVD = peripheral vascular disease; SAE = serious adverse events; vs = versus

The interest in a more targeted, high-risk approach to screening to enrich yield is particularly relevant given declines in AAA prevalence in men over the past two decades.^{68,91 92,93} However, limiting screen-eligible populations to only “high risk” populations inherently results in missed cases. Any attempt to expand screened populations (e.g., extending to all men regardless of smoking history, increasing upper age threshold, adding women) would invariably increase small aneurysm detection and would contribute to overdiagnosis and overtreatment. Based on US data showing that a substantial proportion of small aneurysms are repaired despite the lack of evidence of benefit over surveillance,⁹⁴ the number of procedures and consequent surgical harms that may ensue as a result of broadening the eligibility for screening remains a concern.

Because the population-based screening trials almost exclusively recruited white men aged 65 to 75 years and generally did not report outcomes by subpopulation, one critical question is whether these findings can be extrapolated to other populations. In the absence of trial data, assessing generalizability requires an understanding of contextual evidence about contemporary prevalence, natural history, and treatment effectiveness.

Indirect evidence in subpopulations (older age, female sex, smoking and family history) reveal a set of complex issues. A large proportion of AAA burden (prevalence and ruptures) occurs in older age groups.^95,96 While AAA prevalence rises with age, so do surgical complications, including mortality.⁹⁷ The prevalence of AAA in women has consistently been reported to be less than men.^98,99 However, small AAAs appear to have a higher risk of rupture,^25,100–102 rupture at a later age,^{77,101,103–107 101,108} and result in higher surgical complications—including 30-day postoperative mortality rates,^{104–106,109,110} in-hospital mortality,¹¹¹ major complications,^106,110,112 and readmissions¹⁰⁵ after elective open repair or EVAR—in women than in men. A recent model examining the effectiveness of screening women aged ≥65 years utilizing contemporary assumptions¹¹³ estimated that 3,900 screening invitations would be required to avoid one AAA-related death, which is higher than estimation in other models for men (the number needed to invite to screening was 700 in men).¹¹⁴

Smoking is the strongest predictor of AAA prevalence,^{10,99,102,115,116} growth, and rupture rates.¹⁰² Even with substantial declines since 1995–2002 when the screening trials were conducted,⁹¹ AAA prevalence in male smokers aged 65 to 75 years matches that of the population-based screening trials.¹¹⁷ Family history is associated with an increased risk of developing AAA (RR 1.9 [95% CI, 1.6 to 2.2]).^14,118,119 At this time, however, there is a lack of evidence to determine whether those with family histories exhibit differences in natural history or surgical success rates to alter the net screening benefits. Overall, because there is no direct trial evidence evaluating screening effectiveness in subpopulations and no externally validated risk assessment tools, decision analysis models populated with meta-analytic estimates of prevalence, yield, and surgical complication rates would be considered the best available evidence to date.

There are several limitations to the existing literature. The four, large population-based screening trials began recruiting participants during an era that predated the current widespread implementation of aggressive CVD risk-factor management and reductions in smoking. Thus, the contemporary AAA prevalence cited in Europe, and therefore the absolute benefit of screening, have declined over the intervening time. A general United States population-based estimate of contemporary AAA prevalence is lacking, particularly for subpopulations as a result of low AAA screening uptake in the United States. Further, existing trial literature does not address AAA screening’s potential effect on CVD mortality through identification of those at increased CVD risk and provision of aggressive CVD risk modification.^84,120–122 On the other hand, those identified with AAA would already be candidates for aggressive CVD risk management based on Atherosclerotic Cardiovascular Disease Risk Algorithm’s predicted 10-year risk of >7.5% or 10% as is standard contemporary guidance in the United States.^123,124

Limitations

Limitations of the current review and analysis include results that were limited to studies that met the USPSTF’s fair- or good-quality criteria per USPSTF methods.¹⁷ For three of the Key Questions (KQ2, KQ4, KQ5), there were too few studies or the studies were too clinically or statistically heterogeneous for pooling.²¹

Conclusions

One-time AAA screening in men aged 65 years and older was associated with decreased AAA-related mortality and rupture rates but no effect on all-cause mortality. Higher rates of elective surgery but no long-term differences in quality of life resulted from screening.