Abstract
Objective
Abdominal aortic aneurysm (AAA) is an important vascular disease in older adults but data on lifetime risk of AAA are sparse. We examined lifetime risk of AAA in a community-based cohort and prospectively assessed the association between mid-life cardiovascular risk factors and AAAs.
Approach and Results
In ARIC, 15,792 participants were recruited at Visit 1 in 1987–89 and followed-up through 2013. Longitudinal smoking status was defined using smoking behavior ascertained from Visit 1 (1987–89) to Visit 4 (1996–1998). We followed participants for incident, clinical AAAs using hospital discharge diagnoses, Medicare outpatient diagnoses, or death certificates through 2011 and identified 590 incident AAAs. An abdominal ultrasound was conducted in 2011–2013 in 5,911 surviving participants and 75 asymptomatic AAAs were identified. We estimated the lifetime risk of AAA from the index age 45 years through age 85. At age 45, the lifetime risk for AAA was 5.6% (95% confidence interval: 4.8–6.1) and was higher in men (8.2%) and current smokers (10.5%). Smokers who quit smoking between Visit 1 and Visit 4 had a 29% lower AAA lifetime risk compared to continuous smokers but had a higher risk than pre-Visit 1 quitters. The lifetime risk of rupture or medical intervention was 1.6% (95% CI: 1.2, 1.8). Smoking, white race, male gender, and greater height, LDL or total cholesterol were associated with an increased risk of clinical AAA and asymptomatic AAA.
Conclusions
At least one in 9 middle-aged current smokers developed AAA in their lifetime. Smoking cessation reduced the lifetime risk of AAA.
Keywords: abdominal aortic aneurysm, lifetime risk, risk factor, prospective cohort study
Subject codes: epidemiology, cardiovascular disease, risk factors
INTRODUCTION
Abdominal aortic aneurysm (AAA) is an important manifestation of vascular disease in older adults1–3 and rupture of an AAA is a life threatening condition.4 Traditional atherosclerotic disease risk factors, particularly age, male sex, smoking, and high total cholesterol, are known to contribute to the etiology of AAA.5–8 While risk factors for AAA incidence have been reported in several prospective, epidemiologic studies,5–15 data are scarce on the lifetime risk of AAA, which is defined as the cumulative incidence of AAA between an index age and death, as lifetime risk estimation requires substantial follow-up duration.14 In addition, information on AAA risk factors is constrained by sparse data for asymptomatic AAA,5, 10 because most prospective studies focus on clinical cases.6–9, 11–15 While smoking is the most important risk factor for AAA, there are limited data on how smoking cessation contributes to the reduction of AAA risk in smokers.5, 8, 14 The objectives of this study were three fold: first, to examine the overall lifetime risk of AAA in the Atherosclerosis Risk in Communities (ARIC) Study, a large community-based cohort, as well as in subgroups defined by age, race, gender, and smoking; second, to prospectively assess the association between mid-life atherosclerotic disease risk factors and clinical and asymptomatic AAAs; and third, to longitudinally investigate the harms and benefits for AAA risk associated with continuous smoking versus smoking cessation, respectively.
Materials and Methods
Materials and Methods are available in the online-only Data Supplement
RESULTS
At Visit 1 in 1987–89, there were 15,703 ARIC participants at risk of developing AAA and 5,778 of them attended the Visit 5 ultrasound exam in 2011–2013. Demographic characteristics and cardiovascular risk factors at baseline are shown in Table 1. Compared to those who did not attend the ultrasound screening, the ultrasound subsample had a healthier cardiovascular risk factor profile at baseline (Table 1), as expected for a survival cohort after long follow-up.
Table 1.
Baseline characteristics obtained at Visit 1 in the entire ARIC cohort and the subsamples by ultrasound exam attendance: mean ± SD or percentage
| Baseline characteristics | Entire ARIC cohort* (N=15,703) | Visit 5 ultrasound subsample† (N=5,778) | Did not attend the V5 ultrasound exam† (N=9,335) | P-value╪ |
|---|---|---|---|---|
| Age, years | 54.2±5.76 | 51.8±5.02 | 55.5±5.74 | <0.001 |
| Male, % | 44.8% | 41.1% | 45.4% | <0.001 |
| African Americans, % | 27.1% | 22.3% | 31.1% | <0.001 |
| BMI, kg/m2 | 27.7±5.38 | 27.1±4.77 | 28.2±5.74 | <0.001 |
| Height, cm | 168.5±9.29 | 168.5±9.26 | 168.2±9.3 | <0.001 |
| Smoking status | <0.001 | |||
| Current smoker, % | 26.1% | 16.7% | 30.4% | |
| Former smoker, % | 32.2% | 33.5% | 31.1% | |
| Never smoker, % | 41.7% | 49.8% | 38.5% | |
| Pack-years smoking (including non smokers) | 16.0±21.7 | 10.4±16.3 | 18.4±23.4 | <0.001 |
| Hypertension, % | 35.0% | 23.6% | 41.7% | <0.001 |
| PAD, % | 2.73% | 1.25% | 3.50% | <0.001 |
| Diabetes, % | 11.9% | 5.29% | 16.4% | <0.001 |
| Total cholesterol, mg/dL | 214.67±42.0 | 210.0±40.0 | 217.0±43.1 | <0.001 |
| LDL cholesterol, mg/dL | 137.8±39.4 | 133.4±37.8 | 139.6±40.1 | <0.001 |
| HDL cholesterol, mg/dL | 51.6±17.1 | 53.3±16.9 | 51.0±17.2 | <0.001 |
| Triglycerides, mg/dL | 131.72±90.47 | 119.9±78.2 | 137.8±96.5 | <0.001 |
| SBP, mm Hg | 121.3±19.0 | 116.2±15.7 | 124.3±20.0 | <0.001 |
| DBP, mm Hg | 73.8±11.3 | 72.8±10.3 | 74.3±11.8 | <0.001 |
Eligible for clinical AAA ascertainment;
Excluding clinical AAAs ascertained during the follow-up;
P-value for comparison between the ultrasound subsample and participants who did not attend the ultrasound exam after adjustment for age, sex, and race;
PAD = peripheral arterial disease; SBP=systolic blood pressure; DBP=diastolic blood pressure.
From Visit 1 through 2011, during a median of 22.5 years of follow-up, we identified 590 clinical AAA events over 310,448 person-years of follow-up, yielding a crude incidence rate of 1.90 (95% confidence interval (CI): 1.75, 2.06) per 1000 person-years. In the ultrasound subsample, we identified 75 additional asymptomatic AAAs based on the Visit 5 ultrasound exam. Of the 75 asymptomatic AAAs, 60 had a maximal diameter of 30–39 mm, 11 of 40–49 mm, and 4 of ≥ 50 mm.
Lifetime risk of clinical AAA
Conditional on survival to age 45, the lifetime risk of clinical AAA in the whole ARIC cohort was 5.6% (95% CI: 4.8–6.1) (Table 2), with a higher risk in men (lifetime risk of 8.2%), whites (6.5%), current smokers (10.5%), and those in the top two tertiles of smoking pack-years (9.0% and 11.1%). The lifetime risk and incidence rate of AAA stratified by race, gender, Visit 1 age, smoking status, and pack-years are shown in Tables 2 and 3, respectively. There was a dose-response or gradient relationship of AAA lifetime risk with smoking status and smoking pack-years in the entire cohort as well as in men, women, and whites. In African Americans, current smokers only had a slightly greater risk than former smokers, and participants in pooled pack-year tertiles 2 and 3 had a similar risk to those in tertile 1. However, the lifetime risk in African Americans stratified by gender and smoking status is less interpretable due to the limited number of AAAs in these subgroups. Conditional on survival to age 55, the lifetime risk (95% CI) of AAA by longitudinal smoking status from Visit 1 through Visit 4/3 was as follows: never smokers (1.4% (0.9%, 1.8%)), quitters before Visit 1 (4.1% (3.2%, 4.9%)), recent quitters (6.6% (2.3%, 9.7%)), and continuous smokers (9.4% (7.1%, 10.9%)). Those who had quit smoking after Visit 1 for at least 3–8 years (ie, recent quitters) had a 29% lower AAA lifetime risk compared to continuous smokers but still had a higher risk than participants who quit smoking before Visit 1.
Table 2.
Lifetime risk (95% confidence interval)* of clinical AAA in the whole ARIC cohort and by Visit 1 demographic and smoking variables, 1987 to 2011.
| Overall | African-Americans | Whites | |||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| All | Women | Men | All | Women | Men | All | Women | Men | |
| Overall | 5.6 (4.8, 6.1) | 3.2 (2.5, 3.8) | 8.2 (6.9, 9.0) | 3.2 (1.8, 3.9) | 2.6 (1.0, 3.6) | 4.1 (1.6, 5.3) | 6.5 (5.6, 7.1) | 3.5 (2.6, 4.2) | 9.5 (8.0, 10.5) |
| Age | |||||||||
| 45–54 years | 3.2 (2.2, 3.9) | 1.9 (0.6, 3.0) | 5.0 (3.7, 5.9) | 1.7 (0.3, 2.7) | 1.5 (0.0, 2.9) | 2.2 (0.6, 3.4) | 3.8 (2.6, 4.8) | 2.2 (0.4, 3.6) | 5.9 (4.3, 7.1) |
| 55–64 years† | 6.0 (5.2, 6.6) | 3.5 (2.6, 4.1) | 8.6 (7.1, 9.6) | 3.4 (1.9, 4.3) | 2.9 (1.1, 4.1) | 4.1 (1.4, 5.5) | 6.8 (5.9, 7.5) | 3.7 (2.7, 4.5) | 9.8 (8.1, 11.0) |
| Smoking status | |||||||||
| Never smoker | 2.0 (1.3, 2.5) | 1.2 (0.6, 1.7) | 3.9 (1.8, 5.2) | 1.0 (0.1, 1.7) | 1.1 (0.0, 1.9) | 0.7 (0.0, 1.5) | 2.4 (1.5, 3.1) | 1.3 (0.6, 1.9) | 4.7 (2.3, 6.4) |
| Former smoker | 6.3 (5.0, 7.2) | 2.9 (1.2, 4.2) | 8.1 (6.4, 9.3) | 4.1 (1.3, 5.7) | 1.1 (0.0, 2.3) | 6.2 (1.4, 8.8) | 6.8 (5.4, 7.9) | 3.5 (1.4, 5.1) | 8.6 (6.7, 9.9) |
| Current smoker | 10.5 (7.7, 11.7) | 8.2 (4.8, 9.9) | 12.8 (7.9, 14.7) | 5.9 (0.1, 7.8) | 7.7 (0.0, 11.5) | 4.3 (0.0, 6.2) | 12.5 (9.4, 14.2) | 8.5 (5.1, 10.5) | 17.0 (10.7, 20.0) |
| Pack-years of smoking | |||||||||
| None (0) | 2.0 (1.3, 2.5) | 1.2 (0.6, 1.7) | 3.8 (1.8, 5.2) | 1.0 (0.1, 1.6) | 1.1 (0.0, 1.9) | 0.7 (0.0, 1.5) | 2.4 (1.5, 3.1) | 1.3 (0.6, 1.9) | 4.7 (2.3, 6.3) |
| Tertile 1: 0.1 to 15.2 | 3.6 (2.2, 4.6) | 3.1 (1.0, 4.7) | 4.2 (2.4, 5.6) | 5.0 (1.6, 7.1) | 4.0 (0.0, 6.6) | 6.4 (0.6, 10.0) | 3.0 (1.5, 4.1) | 2.7 (0.3, 4.6) | 3.4 (1.8, 4.7) |
| Tertile 2: 15.3 to 33.9 | 9.0 (6.5, 10.4) | 7.7 (3.3, 10.3) | 10.2 (7.3, 12.1) | 5.2 (0.4, 6.8)‡ | 5.9 (0.0, 9.4)‡ | 4.9 (0.0, 6.7)‡ | 9.7 (7.2, 11.4) | 7.4 (3.0, 10.0) | 11.7 (8.6, 14.0) |
| Tertile 3: 34.0 to 243 | 11.1 (8.7, 12.4) | 7.4 (4.2, 9.4) | 12.7 (9.5, 14.4) | 12.9 (10.2, 14.4) | 8.6 (4.9, 10.9) | 14.6 (11.2, 16.6) | |||
Lifetime risk (%), adjusted for risk of death, calculated starting at index age 45 years over a follow-up of 40 years (maximum age 85 years) unless otherwise mentioned below;
Lifetime risk calculated starting at index age 55 years over a follow-up of 30 years (maximum age 85 years);
Tertiles 2 and 3 were combined due to small sample size in tertile 3 in African Americans.
Table 3.
Associations between baseline characteristics and incident, clinical AAAs in ARIC, 1987 to 2011
| Baseline characteristics | Person-years | N AAA | AAA Incidence (95% CI)* | Hazard Ratio (95% CI)** |
|---|---|---|---|---|
| Age, years (≤10 years follow-up) | ||||
| 45–49 | 41255 | 4 | 0.10 (0.04, 0.26) | 1 |
| 50–59 | 75777 | 54 | 0.71 (0.55, 0.93) | 4.10 (1.46, 11.5) |
| 60–64 | 32858 | 65 | 1.98 (1.55, 2.52) | 9.65 (3.44, 27.1) |
| P for linear trend | <0.0001 | |||
| Age, years (>10 years follow-up) | ||||
| 44–49 | 49350 | 60 | 1.22 (0.94, 1.57) | 1 |
| 50–59 | 81701 | 262 | 3.21 (2.84, 3.62) | 2.20 (1.62, 2.99) |
| 60–69 | 29508 | 145 | 4.91 (4.18, 5.78) | 3.33 (2.38, 4.64) |
| P for linear trend | <0.0001 | |||
| Race | ||||
| African American | 80851 | 84 | 1.04 (0.84, 1.29) | 1 |
| White | 229597 | 506 | 2.20 (2.02, 2.40) | 1.53 (1.17, 2.00) |
| P for difference | 0.002 | |||
| Sex | ||||
| Women | 177464 | 169 | 0.95 (0.82, 1.11) | 1 |
| Men | 132984 | 421 | 3.17 (2.88, 3.48) | 1.33 (1.00, 1.78) |
| P for difference | 0.05 | |||
| Smoking status | ||||
| Never smoker | 136759 | 72 | 0.53 (0.42, 0.66) | 1 |
| Former smoker | 100187 | 219 | 2.19 (1.91, 2.50) | 2.45 (1.85, 3.25) |
| Current smoker | 73191 | 298 | 4.07 (3.63, 4.56) | 7.59 (5.78, 10.0) |
| P for linear trend | <0.0001 | |||
| Packyears (≤10 years follow-up) | ||||
| 0 or never | 64223 | 3 | 0.05 (0.02, 0.14) | 1 |
| Packyears, T1 (0.1 – 15.0) | 28474 | 10 | 0.35 (0.19, 0.65) | 4.74 (1.22, 18.4) |
| Packyears, T2 (15.2 – 33.8) | 27855 | 28 | 1.01 (0.69, 1.46) | 11.8 (3.50, 39.7) |
| Packyears, T3 (34.0 – 242.6) | 26718 | 79 | 2.96 (2.38, 3.69) | 24.8 (7.68, 79.9) |
| P for linear trend | <0.0001 | |||
| Packyears (>10 years follow-up) | ||||
| 0 or never | 73880 | 69 | 0.93 (0.74, 1.18) | 1 |
| Packyears, T1 (0.1 – 15.0) | 32061 | 52 | 1.62 (1.24, 2.13) | 1.56 (1.07, 2.28) |
| Packyears, T2 (15.2 – 33.8) | 28873 | 139 | 4.81 (4.08, 5.68) | 4.13 (3.04, 5.62) |
| Packyears, T3 (34.0 – 188.0) | 22663 | 201 | 8.87 (7.72, 10.2) | 6.35 (4.72, 8.56) |
| P for linear trend | <0.0001 | |||
| Hypertension | ||||
| No | 208156 | 340 | 1.63 (1.47, 1.82) | 1 |
| Yes | 100782 | 245 | 2.43 (2.14, 2.76) | 1.44 (1.20, 1.74) |
| P for difference | 0.0001 | |||
| PAD | ||||
| No | 293185 | 537 | 1.83 (1.69, 1.99) | 1 |
| Yes | 6298 | 29 | 4.60 (3.20, 6.63) | 1.74 (1.16, 2.63) |
| P for difference | 0.008 | |||
| Diabetes | ||||
| No | 276727 | 546 | 1.97 (1.81, 2.15) | 1 |
| Yes | 30924 | 42 | 1.36 (1.01, 1.84) | 0.52 (0.36, 0.75) |
| P for difference | 0.0005 | |||
| Height, cm | ||||
| T1 (< 163) | 107058 | 105 | 0.98 (0.81, 1.19) | 1 |
| T2 (164 – 172) | 99590 | 158 | 1.59 (1.36, 1.85) | 1.28 (0.94, 1.72) |
| T3 (173 – 199) | 103458 | 327 | 3.16 (2.84, 3.52) | 1.64 (1.16, 2.31) |
| P for linear trend | 0.005 | |||
| HDL-C, mg/dL | ||||
| T1 (<42.4) | 99113 | 315 | 3.18 (2.85, 3.55) | 1 |
| T2 (42.4 – 55.9) | 101817 | 173 | 1.70 (1.46, 1.97) | 0.67 (0.55, 0.81) |
| T3 (56.0 – 163.0) | 104903 | 98 | 0.93 (0.77, 1.14) | 0.48 (0.37, 0.61) |
| P for linear trend | 0.0004 | |||
| LDL-C, mg/dL | ||||
| T1 (<119.6) | 101799 | 117 | 1.15 (0.96, 1.38) | 1 |
| T2 (119.6 – 151.8) | 100864 | 202 | 2.00 (1.74, 2.30) | 1.41 (1.10, 1.79) |
| T3 (151.9 – 504.6) | 99262 | 252 | 2.54 (2.24, 2.87) | 1.58 (1.25, 2.00) |
| P for linear trend | 0.0001 | |||
| Total cholesterol, mg/dL | ||||
| T1 (<194.7) | 103963 | 160 | 1.54 (1.32, 1.80) | 1 |
| T2 (195.7 – 228.6) | 102334 | 187 | 1.83 (1.58, 2.11) | 1.16 (0.93, 1.46) |
| T3 (229.6 – 593.1) | 99492 | 238 | 2.39 (2.11, 2.72) | 1.47 (1.18, 1.84) |
| P for linear trend | 0.0007 | |||
| Triglycerides, mg/dL | ||||
| T1 (<88) | 106402 | 135 | 1.27 (1.07, 1.50) | 1 |
| T2 (89 – 136) | 101483 | 193 | 1.90 (1.65, 2.19) | 0.92 (0.73, 1.18) |
| T3 (137 – 1926) | 97969 | 258 | 2.63 (2.33, 2.98) | 0.96 (0.74, 1.25) |
| P for linear trend | 0.77 | |||
CI=confidence interval, T1=tertile 1, T2=tertile 2, T3=tertile 3;
AAA incidence rate per 1,000 person-years of follow-up;
Hazard Ratio: adjusted for age, sex, race, height, pack-years of smoking, alcohol consumption (g/day), HDL-C, LDL-C, triglycerides, hypertension, PAD and diabetes; for total cholesterol and smoking status as risk factors, the association analysis did not adjust for LDL-C or smoking pack-years, respectively.
The clinical AAAs in the primary analysis included AAAs with rupture or corrective procedures, as well as medically documented, asymptomatic AAAs. Since estimates of the incidence rate and lifetime risk for more severe clinical cases are also potentially useful, we conducted a secondary analysis focusing on the symptomatic subgroup. After exclusion of asymptomatic hospital or CMS outpatient AAAs, 181 clinical AAAs with a rupture or medical intervention remained, yielding an incidence rate of 0.58 (95% CI: 0.50, 0.67) per 1000 person-years and a lifetime risk of 1.6% (95% CI: 1.2, 1.8). Among the 181 cases, 27 had ruptures and 175 had medical intervention, with 21 of them having both. The incidence rates (95% CI) / lifetime risks (95% CI) by major risk factor categories were as follows: women --0.19 (0.14, 0.27) / 0.6 (0.3, 0.9), men --1.11 (0.94, 1.30)/ 2.7 (2.0, 3.2), African Americans --0.28 (0.19, 0.43)/ 0.8 (0.1, 1.2), whites --0.69 (0.59, 0.80)/ 1.9 (1.5, 2.2), never smokers --0.15 (0.09, 0.23) / 0.6 (0.2, 0.9), former smokers --0.77 (0.61, 0.96)/ 2.0 (1.4, 2.5), and current smokers --1.13 (0.91, 1.41)/ 2.7 (1.3, 3.3).
Associations between Visit 1 risk factors and risk of clinical AAA
Table 3 and Supplementary Table I show the associations between Visit 1 risk factors and incident clinical AAA. For age and smoking pack-years, we performed separate analyses for the first 10 years of follow-up and for >10 years of follow-up, as the proportional hazard assumption was violated when using the entire follow-up time but held for each period separately. In the basic adjustment model (Supplementary Table I), age, male gender, white race, current and former smoking, pack-years of smoking, hypertension, peripheral arterial disease (PAD), height, triglycerides, and low-density lipoprotein (LDL) or total cholesterol were positively associated with incident clinical AAA, while diabetes and high-density lipoprotein cholesterol (HDL-C) were negatively associated with AAA. BMI and alcohol consumption were not associated with AAA (data not shown). With simultaneous adjustment (Table 3), the associations remained significant for most risk factors except triglycerides. The hazard ratios for age and smoking pack-years were stronger for earlier follow-up (when the cohort was younger) than later follow-up. Among the risk factors identified, age, smoking status, and pack-years showed the strongest associations with AAA, with hazard ratios approximately 3 to 9 fold higher in the highest age group, 7-fold higher in current smokers, and 6 to 24 fold higher in the highest smoking pack-year group, compared to the lowest ranked group for each of the risk factors.
Associations between Visit 1 risk factors and risk of asymptomatic AAA detected by ultrasound
Similar to the clinical AAA data, white race, male gender, smoking status and pack-years, greater height, and elevated LDL or total cholesterol at baseline were significantly and independently associated with the risk of asymptomatic AAA (Table 4). Baseline age, hypertension, PAD, diabetes, and HDL-C were not significantly associated with the risk of asymptomatic AAA (Table 4 and Supplementary Table II). Compared to the lowest tertiles, the odds ratios were more than 5 fold for current smokers and those in the top tertile groups for height or smoking pack-years at baseline.
Table 4.
Associations between baseline characteristics in 1987–89 and ultrasound AAAs* detected in 2011–2013 at Visit 5: the ARIC study
| Baseline characteristics | N Subjects | N AAA | AAA prevalence%† | Odds Ratio (95% CI)‡ |
|---|---|---|---|---|
| Age, years | ||||
| 45–49 | 2307 | 22 | 0.95 (0.63, 1.45) | 1 |
| 50- | 3493 | 53 | 1.52 (1.14, 2.04) | 1.26 (0.67, 2.37) |
| P for difference | 0.47 | |||
| Race | ||||
| African American | 1296 | 9 | 0.69 (0.36, 1.33) | 1 |
| White | 4504 | 66 | 1.47 (1.15, 1.87) | 2.44 (1.07, 5.54) |
| P for difference | 0.03 | |||
| Sex | ||||
| Women | 3415 | 15 | 0.44 (0.26, 0.73) | 1 |
| Men | 2385 | 60 | 2.52 (1.95, 3.24) | 2.33 (1.02, 5.34) |
| P for difference | 0.04 | |||
| Smoking status | ||||
| Never smoker | 2885 | 14 | 0.49 (0.29, 0.82) | 1 |
| Former smoker | 1939 | 29 | 1.50 (1.04, 2.15) | 2.34 (1.11, 4.95) |
| Current smoker | 969 | 32 | 3.30 (2.34, 4.67) | 6.86 (3.18, 14.8) |
| P for linear trend | <0.0001 | |||
| Packyears | ||||
| 0 or never | 2910 | 14 | 0.48 (0.28, 0.81) | 1 |
| Packyears, T1 (0.1 – 10.0) | 945 | 8 | 0.85 (0.42, 1.69) | 1.46 (0.56, 3.78) |
| Packyears, T2 (10.2 – 26.0) | 946 | 21 | 2.22 (1.45, 3.40) | 3.59 (1.64, 7.84) |
| Packyears, T3 (26.1 – 135.0) | 923 | 32 | 3.47 (2.45, 4.90) | 5.91 (2.76, 12.6) |
| P for linear trend | <0.0001 | |||
| Hypertension | ||||
| No | 4408 | 58 | 1.32 (0.78, 2.01) | 1 |
| Yes | 1361 | 17 | 1.25 (1.02, 1.70) | 0.88 (0.43, 1.79) |
| P for difference | 0.72 | |||
| PAD | ||||
| No | 5546 | 69 | 1.24 (0.98, 1.58) | 1 |
| Yes | 70 | 1 | 1.43 (0.20, 10.1) | 1.87 (0.23, 15.4) |
| P for difference | 0.56 | |||
| Diabetes | ||||
| No | 5457 | 72 | 1.32 (1.05, 1.66) | 1 |
| Yes | 305 | 3 | 0.98 (0.32, 3.05) | 1.08 (0.23, 5.06) |
| P for difference | 0.92 | |||
| Height, cm | ||||
| T1 (<163) | 1949 | 7 | 0.36 (0.17, 0.75) | 1 |
| T2 (164 – 172) | 1924 | 12 | 0.62 (0.35, 1.10) | 1.24 (0.44, 3.48) |
| T3 (173 – 199) | 1925 | 56 | 2.91 (2.24, 3.78) | 6.37 (2.26, 18.0) |
| P for linear trend | 0.0005 | |||
| HDL-C, mg/dL | ||||
| T1 (<44.0) | 1864 | 42 | 2.25 (1.67, 3.05) | 1 |
| T2 (44.3 – 58.0) | 1949 | 22 | 1.13 (0.74, 1.71) | 0.83 (0.40, 1.69) |
| T3 (58.7 – 163.0) | 1923 | 10 | 0.52 (0.28, 0.97) | 0.70 (0.28, 1.75) |
| P for linear trend | 0.45 | |||
| LDL-C, mg/dL | ||||
| T1 (<116.0 ) | 1893 | 10 | 0.53 (0.28, 0.98) | 1 |
| T2 (116.1 – 146.7) | 1893 | 16 | 0.85 (0.52, 1.38) | 1.51 (0.57, 3.98) |
| T3 (146.8 – 452.4) | 1893 | 47 | 2.48 (1.87, 3.30) | 3.53 (1.67, 7.48) |
| P for linear trend | 0.001 | |||
| Total cholesterol, mg/dL | ||||
| T1 (< 191.7) | 1906 | 12 | 0.63 (0.36, 1.11) | 1 |
| T2 (191.8 – 222.7) | 1900 | 23 | 1.21 (0.80, 1.82) | 2.73 (1.16, 6.44) |
| T3 (223.7 – 531.2) | 1929 | 39 | 2.02 (1.48, 2.77) | 3.55 (1.67, 7.53) |
| P for linear trend | 0.001 | |||
| Triglycerides, mg/dL | ||||
| T1 (<81) | 1914 | 9 | 0.47 (0.24, 0.90) | 1 |
| T2 (82 – 125) | 1909 | 35 | 1.83 (1.32, 2.55) | 4.43 (1.92, 10.2) |
| T3 (126 – 1562) | 1913 | 30 | 1.57 (1.10, 2.24) | 2.22 (0.94, 5.22) |
| P for linear trend | 0.07 | |||
CI=confidence interval, T1=tertile 1, T2=tertile 2, T3=tertile 3;
Excluding clinical AAAs detected by hospitalization/death or CMS records.
ultrasound AAA cases per 100 persons who attended the Visit 5 ultrasound exam;
Odds Ratio: adjusted for age, sex, race, smoking pack-years, height, alcohol consumption (g/day), triglycerides, and total cholesterol; for LDL cholesterol and smoking status as risk factors, the association analysis did not adjust for total cholesterol or smoking pack-years, respectively.
Associations between longitudinal smoking status and AAA risk
There was a significant association between the risk of both clinical and asymptomatic AAA and longitudinal smoking status that was defined prospectively based on information obtained from Visit 1 through Visit 4, with the highest risk being for continuous smokers, followed by recent quitters and then quitters before Visit 1 (p for trend < 0.01) (Table 5). The associations were independent of the other AAA risk factors and Visit 1 smoking pack-years, which was associated with longitudinal smoking status and thus could confound the associations. Compared with never smokers, those who had quit smoking after Visit 1 for at least 3–8 years (ie, recent quitters) still had approximately 2.6–3.5 fold greater risk for both clinical AAA and asymptomatic AAA in the next 15 years of follow-up, consistent with the lifetime risk data.
Table 5.
Association between longitudinal smoking status from Visit 1 through Visit 3/Visit 4* and AAA detected after Visit 3/Visit 4: the ARIC study
| Longitudinal smoking status | N Subject | Person-years | N AAA | AAA incidence (95% CI)† | HR1 (95% CI) | p1 | HR2 (95% CI) | p2 |
|---|---|---|---|---|---|---|---|---|
| Clinical AAA | ||||||||
| Never smoker | 4821 | 63469 | 54 | 0.85 (0.65, 1.11) | 1 | 1 | ||
| Quitter before V1 | 3941 | 49186 | 133 | 2.70 (2.28, 3.20) | 2.16 (1.56, 2.99) | <0.0001 | 1.83 (1.19, 2.81) | 0.006 |
| Recent quitter | 410 | 4742 | 20 | 4.22 (2.72, 6.54) | 3.94 (2.14, 7.27) | <0.0001 | 3.50 (1.53, 8.04) | 0.003 |
| Continuous_smk | 2207 | 24979 | 163 | 6.53 (5.60, 7.61) | 7.11 (4.70, 10.8) | <0.0001 | 6.41 (3.67, 11.2) | <0.0001 |
| P for linear trend | <0.0001 | <0.0001 | ||||||
| Asymptomatic AAA detected at Visit 5 ultrasound | ||||||||
| Longitudinal smoking status | AAA prevalence (95% CI)‡ | OR1 (95% CI) | OR2 (95% CI) | |||||
| Never smoker | 2341 | NA | 14 | 0.60 (0.35, 1.01) | 1 | 1 | ||
| Quitter before V1 | 1745 | NA | 25 | 1.43 (0.97, 2.12) | 1.40 (0.63, 3.12) | 0.41 | 1.34 (0.60, 2.99) | 0.47 |
| Recent quitter | 138 | NA | 3 | 2.17 (0.70, 6.74) | 2.65 (0.66, 10.6) | 0.17 | 2.66 (0.65, 10.9) | 0.17 |
| Continuous_smk | 629 | NA | 24 | 3.81 (2.56, 5.69) | 4.06 (1.51,10.9) | 0.005 | 3.93 (1.44, 10.7) | 0.008 |
| P for linear trend | 0.005 | 0.008 | ||||||
CI=confidence interval, HR=hazard ratio, V1=Visit 1, OR=odds ratio, Continuous_smk=continuous smoker;
the last visit was Visit 3 if smoking information in Visit 4 was missing;
Incidence rate per 1000 person-years for clinical AAA;
Number of asymptomatic AAA cases per 100 persons who attended the V5 ultrasound exam;
HR1/OR1 and p1: adjusted for age, sex, race at Visit 3/Visit 4, and pack-years of smoking at Visit 1;
HR2 and p2: for clinical AAA analysis--adjusted for age, sex, race, height, alcohol consumption (g/day), HDL-C, LDL-C, triglycerides, hypertension, diabetes, and PAD at Visit 3/Visit 4, and pack-years of smoking at Visit 1;
OR2 and p2: for asymptomatic AAA analysis--adjusted for age, sex, race, height, and LDL-C at Visit 3/Visit 4, and pack-years of smoking at Visit 1.
DISCUSSION
In this population-based, longitudinal study for a median 22.5 years, we reported AAA incidence; AAA lifetime risk; and quantified AAA lifetime risk by race, gender, smoking status/pack-years, and longitudinal smoking status. To the best of our knowledge, this is the first study to report the AAA lifetime risk in a community-based cohort with long follow-up. We also identified independent associations of clinical AAA incidence and asymptomatic AAA prevalence with traditional cardiovascular risk factors, including longitudinal smoking status.
Lifetime risk estimates are useful indicators of the public health burden of diseases in a population and various subgroups and thus are informative in guiding preventive and screening strategies 16. Lifetime risk estimates are also helpful in patient consultation and in improving public awareness of disease conditions because they are more readily understood by the general public than standard epidemiologic risk information (e.g., incidence rate).16, 17 The lifetime risk of AAA in our study was one in 17, and was higher in males (1 in 12), whites (1 in 15), current smokers (1 in 9), and those in the highest tertile of smoking pack-years (1 in 9). While the lifetime risk of AAA is lower than that of the major cardiovascular diseases such as coronary heart disease (1 in 2 to 3)18 and stroke (1 in 5 to 6),19 the lifetime risk of AAA in high-risk subgroups for AAA is similar to the lifetime risk of other chronic diseases such as breast cancer (1 in 8 females)20 and venous thromboembolism (1 in 12).21 There has been only one previous report of AAA lifetime risk that was based on electronic health records in England, with a median of 5.5 years of follow-up. That study reported a lifetime risk of AAA at 3.3%, 1.3 %, and 0.9 % for current smokers, ex-smokers, and never smokers by age 90 years.14 In that study, 1,937,360 patients were recruited from 225 family practices and their records were linked across various data sources to capture AAA diagnoses, including cases with or without leak, rupture or repair.14 Information on smoking was self-reported by patients at baseline. Detailed information on duration and amount of cigarette smoked was not available.
Smoking is the most important risk factor for AAA, as confirmed by prospective studies of clinical AAA6–9, 11, 14, 15 and asymptomatic AAA.5 Compared to never smokers, current smokers in our study had a 6–7 fold greater risk of clinical AAA and ultrasound detected asymptomatic AAA, and those in the highest tertile of smoking pack-years had more than 5 fold greater risk of ultrasound detected asymptomatic AAA and 6–25 fold greater risk of clinical AAA. The US Preventive Services Task Force (USPSTF) recommends 1-time ultrasound screening for AAA in male ever smokers aged 65 to 75 years and selective screening in male never smokers aged 65 to 75 years. The USPSTF made no recommendation for screening in female ever smokers of this age group because of insufficient evidence on the benefits and harms of screening. Our study, based on a long follow-up, showed that the lifetime risk of clinical AAA in women who were current smokers (8.2%) was similar to men who were former smokers (8.1%) and higher than in men who were never smokers (3.9%). Our unique data on lifetime risk of AAA expands the previous report of an increased risk of AAA for women smokers over a median follow-up of 4.9 years15 and another report of a higher incidence of AAA in women smokers than in men who never smoked during a mean of 12.7 years of follow-up.22 These data may be useful to policy makers in considering preventive strategies for female smokers in the future. Our data also indicate that physicians may need to pay similar attention in monitoring female current smokers for the occurrence of AAA as to male ever smokers after they pass middle age. By all means, the best preventive strategy to reduce AAA risk in smokers is to stop smoking.
In addition, our longitudinal smoking data indicate that those who had quit smoking for at least 3–8 years (ie, recent quitters) still had an approximately 2.6–3.5 fold increased risk for both clinical AAA and asymptomatic AAA in the next 15 years compared to never smokers, and their lifetime risk of AAA (6.6%) was higher than longer term quitters (ie, quitters before Visit 1) and never smokers (4.1% and 1.4%, respectively). The findings from our study add to the three previous studies that obtained smoking history at baseline and followed up participants for 5.5 years,14 7 years,5 and 16 years,8 respectively. These studies reported a 2.6–6.5 fold increased risk for clinical or asymptomatic AAA for people who had quit smoking for less than 10 years.5, 8, 14 To the best of our knowledge, there is no other study of AAA that quantified AAA risk, including lifetime risk, according to longitudinal smoking status.
Several prospective studies have assessed the associations between traditional cardiovascular risk factors and the risk of AAA.5–15 However, most of these studies focused on clinical cases without inclusion of asymptomatic AAA.6–9, 11–15 While the overall risk factor profile for asymptomatic and clinical cases might be similar, the magnitude of associations could be different. Only two prospective studies specifically investigated asymptomatic AAA detected by ultrasound.5, 10 The first identified 119 AAA cases from 4,345 participants during 7 years of follow-up.5 The second, including 35 cases and 140 controls, retrospectively investigated the relationship between asymptomatic AAA and lipids, which had been measured 12 years earlier.10 Our study replicated the previously published positive associations of clinical AAA with age, male gender, white race, smoking status and pack-years, hypertension, PAD, height, and total or LDL cholesterol, and negative associations with diabetes and HDL.6–9, 11, 13–15 Our study also replicated associations of asymptomatic AAA with male gender, smoking status and pack-years, LDL and total cholesterol.5, 10 Among these, the identification of white versus African American race and height as risk factors for ultrasound-detected, asymptomatic AAA is the first report for asymptomatic AAA from a prospective study. It is interesting to note that height showed a stronger association with asymptomatic AAA than clinical AAA in our study (risk ratio 6.37 vs 1.64 for tallest versus shortest tertiles), even though the distribution of baseline height in the ultrasound subset was similar to that of the whole ARIC cohort (Table 1). The biological mechanism linking height to increased risk of AAA is unknown. In a meta-analysis of adult height with vascular morbidity and death, height was associated with increased risk of death from ruptured aortic aneurysm.23 The authors postulated that the higher risk of rupturing associated with longer arteries might explain the association.23 Nevertheless, future studies are needed to evaluate if height predicts the risk of AAA using other AAA definitions based on an expected section of a patient’s aorta according to height. We did not identify a significant association of alcohol consumption with either clinical or asymptomatic AAA, which agrees with some other studies.5, 6, 9, 11 A study by Wong et al 8 reported a positive association between AAA and total alcohol consumption, but the association was attenuated after adjustment for hypertension. Another study by Stackelberg et al 12 found an inverse association between AAA and moderate consumption of alcohol, especially for consumption of beer in men and wine in women. Differences in sample size and beverage type might explain the discrepant findings across studies.
Some limitations of this study warrant consideration. There might have been proportionately more undiagnosed AAAs in individuals who were lost to follow-up, resulting in an underestimated incidence, and thus lifetime risk, of AAA. In addition, the lifetime risk of AAA was approximated as the cumulative incidence of clinical AAA through age 85. This approach might lead to an underestimation of lifetime risk of AAA because: 1) AAA may present beyond age 85 and 2) asymptomatic AAAs detected by our one-time ultrasound screening were not included in the calculation of lifetime risk due to a lack of information about the time of onset. The issue of a lack of information about the time of onset also applies to clinical AAAs for which we used the time of clinical presentation or the time of diagnosis to calculate the incidence rate and lifetime risk. This constraint could only be addressed in a longitudinal study with multiple ultrasound screenings. The missingness of undiagnosed AAAs would likely have been non-differential with regard to exposure level and thus likely diluted estimates of association between AAA and exposures. In the 2011–2013 ultrasound screening, the people who were evaluated had a healthier risk factor profile, which might have impacted the number of AAAs identified. In the analysis of asymptomatic AAAs with risk factors, we adjusted for the potential selection bias caused by differential participation in the ultrasound screening. Moreover, we defined hypertension and diabetes based on measurements of blood pressure and glucose or the use of corresponding medications. The risk of AAA might be different across treatment status as well as the extent of blood pressure or glucose normalization. Further studies are needed to evaluate whether treatment for hypertension and diabetes influences the risk of AAA. Finally, while we excluded participants with AAA repair prior to the baseline evaluation, we might have still included some prevalent AAAs in the cohort at baseline due to lack of a baseline ultrasound screening. However, the prevalence of AAA in the ARIC cohort at the baseline age of 45–64 years should have been low.
In conclusion, in this community-based prospective study with a long period of follow-up, at least one in 17 middle-aged adults and 1 in 9 current smokers developed AAA in their lifetime, while the lifetime risk of rupture or medical intervention was lower. Those who had quit smoking for at least 3–8 years before follow-up started showed a 29% lower AAA lifetime risk compared to continuous smokers. In addition to smoking status and pack-years, several other cardiovascular risk variables measured in middle-age, including white race, male gender, height, and LDL or total cholesterol, were associated with future risk for both clinical AAA and ultrasound-detected, asymptomatic AAA.
Supplementary Material
Highlights.
At age 45, lifetime risk for AAA in women, men, African Americans, and whites was 3.2%, 8.2%, 3.2%, and 6.5%, respectively.
Lifetime risk for AAA was highest in current smokers (10.5%) and those in the top two tertiles of smoking pack-years (9.0% and 11.1%).
Smokers who quit smoking between Visit 1 and Visit 4 had a 29% lower AAA lifetime risk compared to continuous smokers but still had higher risk than pre-Visit 1 quitters.
Female current smokers had similar lifetime risk of AAA (8.2%) to male former smokers (8.1%).
Both baseline and longitudinal smoking status, white race, male gender, greater height, and elevated LDL or total cholesterol were associated with increased risk of clinical AAA and asymptomatic AAA.
Acknowledgments
The authors thank the staff and participants of the ARIC Study for their important contributions.
Sources of Funding: The National Heart, Lung, and Blood Institute (NHLBI) supported this research through R01 HL103695 and ARIC contracts HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C.
Abbreviations
- AAA
abdominal aortic aneurysm
- ARIC
the Atherosclerosis Risk in Communities Study
- PAD
peripheral arterial disease
Footnotes
Disclosures: None.
References
- 1.Lederle FA, Johnson GR, Wilson SE, Littooy FN, Krupski WC, Bandyk D, Acher CW, Chute EP, Hye RJ, Gordon IL, Freischlag J, Averbook AW, Makaroun MS. Yield of repeated screening for abdominal aortic aneurysm after a 4-year interval. Aneurysm Detection and Management Veterans Affairs Cooperative Study Investigators. Arch Intern Med. 2000;160:1117–1121. doi: 10.1001/archinte.160.8.1117. [DOI] [PubMed] [Google Scholar]
- 2.Vardulaki KA, Prevost TC, Walker NM, Day NE, Wilmink AB, Quick CR, Ashton HA, Scott RA. Incidence among men of asymptomatic abdominal aortic aneurysms: estimates from 500 screen detected cases. J Med Screen. 1999;6:50–54. doi: 10.1136/jms.6.1.50. [DOI] [PubMed] [Google Scholar]
- 3.Singh K, Bonaa KH, Jacobsen BK, Bjork L, Solberg S. Prevalence of and risk factors for abdominal aortic aneurysms in a population-based study : The Tromso Study. Am J Epidemiol. 2001;154:236–244. doi: 10.1093/aje/154.3.236. [DOI] [PubMed] [Google Scholar]
- 4.Sakalihasan N, Limet R, Defawe OD. Abdominal aortic aneurysm. Lancet. 2005;365:1577–1589. doi: 10.1016/S0140-6736(05)66459-8. [DOI] [PubMed] [Google Scholar]
- 5.Forsdahl SH, Singh K, Solberg S, Jacobsen BK. Risk factors for abdominal aortic aneurysms: a 7-year prospective study: the Tromso Study, 1994–2001. Circulation. 2009;119:2202–2208. doi: 10.1161/CIRCULATIONAHA.108.817619. [DOI] [PubMed] [Google Scholar]
- 6.Iribarren C, Darbinian JA, Go AS, Fireman BH, Lee CD, Grey DP. Traditional and novel risk factors for clinically diagnosed abdominal aortic aneurysm: the Kaiser multiphasic health checkup cohort study. Ann Epidemiol. 2007;17:669–678. doi: 10.1016/j.annepidem.2007.02.004. [DOI] [PubMed] [Google Scholar]
- 7.Rodin MB, Daviglus ML, Wong GC, Liu K, Garside DB, Greenland P, Stamler J. Middle age cardiovascular risk factors and abdominal aortic aneurysm in older age. Hypertension. 2003;42:61–68. doi: 10.1161/01.HYP.0000078829.02288.98. [DOI] [PubMed] [Google Scholar]
- 8.Wong DR, Willett WC, Rimm EB. Smoking, hypertension, alcohol consumption, and risk of abdominal aortic aneurysm in men. Am J Epidemiol. 2007;165:838–845. doi: 10.1093/aje/kwk063. [DOI] [PubMed] [Google Scholar]
- 9.Tornwall ME, Virtamo J, Haukka JK, Albanes D, Huttunen JK. Life-style factors and risk for abdominal aortic aneurysm in a cohort of Finnish male smokers. Epidemiology. 2001;12:94–100. doi: 10.1097/00001648-200101000-00016. [DOI] [PubMed] [Google Scholar]
- 10.Wanhainen A, Bergqvist D, Boman K, Nilsson TK, Rutegard J, Bjorck M. Risk factors associated with abdominal aortic aneurysm: a population-based study with historical and current data. J Vasc Surg. 2005;41:390–396. doi: 10.1016/j.jvs.2005.01.002. [DOI] [PubMed] [Google Scholar]
- 11.Lederle FA, Larson JC, Margolis KL, Allison MA, Freiberg MS, Cochrane BB, Graettinger WF, Curb JD. Abdominal aortic aneurysm events in the women’s health initiative: cohort study. Bmj. 2008;337:a1724. doi: 10.1136/bmj.a1724. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Stackelberg O, Bjorck M, Larsson SC, Orsini N, Wolk A. Alcohol consumption, specific alcoholic beverages, and abdominal aortic aneurysm. Circulation. 2014;130:646–652. doi: 10.1161/CIRCULATIONAHA.113.008279. [DOI] [PubMed] [Google Scholar]
- 13.Rapsomaniki E, Timmis A, George J, Pujades-Rodriguez M, Shah AD, Denaxas S, White IR, Caulfield MJ, Deanfield JE, Smeeth L, Williams B, Hingorani A, Hemingway H. Blood pressure and incidence of twelve cardiovascular diseases: lifetime risks, healthy life-years lost, and age-specific associations in 1.25 million people. Lancet. 2014;383:1899–1911. doi: 10.1016/S0140-6736(14)60685-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pujades-Rodriguez M, George J, Shah AD, Rapsomaniki E, Denaxas S, West R, Smeeth L, Timmis A, Hemingway H. Heterogeneous associations between smoking and a wide range of initial presentations of cardiovascular disease in 1937360 people in England: lifetime risks and implications for risk prediction. Int J Epidemiol. 2015;44:129–141. doi: 10.1093/ije/dyu218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jahangir E, Lipworth L, Edwards TL, Kabagambe EK, Mumma MT, Mensah GA, Fazio S, Blot WJ, Sampson UK. Smoking, sex, risk factors and abdominal aortic aneurysms: a prospective study of 18 782 persons aged above 65 years in the Southern Community Cohort Study. J Epidemiol Community Health. 2015;69:481–488. doi: 10.1136/jech-2014-204920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sasieni PD, Adams J. Standardized lifetime risk. Am J Epidemiol. 1999;149:869–875. doi: 10.1093/oxfordjournals.aje.a009903. [DOI] [PubMed] [Google Scholar]
- 17.Fortin JM, Hirota LK, Bond BE, O’Connor AM, Col NF. Identifying patient preferences for communicating risk estimates: a descriptive pilot study. BMC Med Inform Decis Mak. 2001;1:2. doi: 10.1186/1472-6947-1-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lloyd-Jones DM, Larson MG, Beiser A, Levy D. Lifetime risk of developing coronary heart disease. Lancet. 1999;353:89–92. doi: 10.1016/S0140-6736(98)10279-9. [DOI] [PubMed] [Google Scholar]
- 19.Seshadri S, Beiser A, Kelly-Hayes M, Kase CS, Au R, Kannel WB, Wolf PA. The lifetime risk of stroke: estimates from the Framingham Study. Stroke. 2006;37:345–350. doi: 10.1161/01.STR.0000199613.38911.b2. [DOI] [PubMed] [Google Scholar]
- 20.Feuer EJ, Wun LM, Boring CC, Flanders WD, Timmel MJ, Tong T. The lifetime risk of developing breast cancer. J Natl Cancer Inst. 1993;85:892–897. doi: 10.1093/jnci/85.11.892. [DOI] [PubMed] [Google Scholar]
- 21.Bell EJ, Lutsey PL, Basu S, Cushman M, Heckbert SR, Lloyd-Jones DM, Folsom AR. Lifetime risk of venous thromboembolism in two cohort studies. Am J Med. 2015 doi: 10.1016/j.amjmed.2015.10.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Stackelberg O, Bjorck M, Larsson SC, Orsini N, Wolk A. Sex differences in the association between smoking and abdominal aortic aneurysm. Br J Surg. 2014;101:1230–1237. doi: 10.1002/bjs.9526. [DOI] [PubMed] [Google Scholar]
- 23.Adult height and the risk of cause-specific death and vascular morbidity in 1 million people: individual participant meta-analysis. Int J Epidemiol. 2012;41:1419–1433. doi: 10.1093/ije/dys086. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.