Abstract
Background The aim of this study was to evaluate the influence of statins on the growth of small abdominal aortic aneurysms (AAAs).
Methods We retrospectively examined AAA diameter in 211 patients who had undergone serial imaging surveillance.
Results Patients treated with and without statins were similar regarding age, initial aneurysm size, diagnosis of diabetes and hypertension, and smoking history. Patients receiving statins had a decreased aneurysm growth rate compared with those patients not receiving statins (0.9 mm/y [interquartile range, −1.0 to +1.0] vs 3.2 mm/y [interquartile range, 2.0–4.9], P < 0.0001). This difference in the rate of growth was maintained after adjusting for potential confounding factors.
Conclusions To date, this is the one of the largest retrospective studies demonstrating an association between statin use and decreased growth rate of AAA.
Abdominal aortic aneurysm (AAA) is common, with a prevalence of 4% to 9% in men and 1% in women in population-based screening studies from various countries.1–4Patients with AAA have a high mortality, primarily due to rupture, and AAA is among the top 15 leading causes of death in the United States.5The US Preventive Services Task Force recommends a 1-time screening for AAA by ultrasonography in men aged 65 to 75 years who have ever smoked.
Surgical or endovascular repair is recommended for those AAAs 5.5 cm or larger in diameter because of increased risk of catastrophic rupture. In contrast, small AAAs (<5.5 cm) have a low risk of rupture, and there is currently no validated treatment strategy to limit progression. During routine abdominal screening of asymptomatic patients, 90% of AAAs identified are less than 5.5 cm in diameter.6Two large randomized trials reported that survival was not improved by elective surgical repair of small AAAs.7Therefore, the development of noninvasive therapeutic approaches to slow the rate of AAA growth is of significant clinical relevance.
Statins, 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors, are widely prescribed for their lipid-lowering effects. There are conflicting studies reporting an association between AAA risk and hypercholesterolemia, with most, but not all, indicating an association between AAA and increased serum level of cholesterol.8–12Reducing cholesterol concentrations in patients is therefore desirable, but there is no clear relationship between serum cholesterol level and AAA growth.13,14However, recent advances have led to a better understanding of the molecular mechanisms involved in AAA formation, and many of the pleiotropic effects of statins, including anti-inflammatory, antioxidative, and the reduction of matrix metalloproteinases (MMPs) secretion, may prevent the progression of AAA.
To date, there have been no randomized controlled studies in patients to assess the efficacy of statins in reducing the growth of AAA. Numerous studies have shown an association of coronary heart disease and peripheral atherosclerosis with AAA.12,15It is unknown whether this association between AAA and atherosclerosis is causal or due to common risk factors.16However, it is currently advised by American Heart Association guidelines that AAA be considered an atherosclerotic equivalent.17This makes conducting a randomized controlled study with and without statin therapy unethical. Current human studies are limited to small observational studies and are inconclusive. Herein, we report the results of one of the largest retrospective studies evaluating the effect of statins on AAA growth.
MATERIALS AND METHODS
We retrospectively identified patients under surveillance for AAA at University of Iowa Hospitals and Clinics between January 2001 and January 2005. The study was approved by the University of Iowa’s institutional review board. Patients’ data were collected from electronic medical records, and all patients with a diagnosis of AAA were screened for eligibility of enrollment. We included only those patients with an AAA of at least 3 cm and who had undergone repeat imaging follow-up of at least 1 year. A total of 211 patients were included in this study, and their clinical data are summarized in Table 1. Statin users were identified as patients who were on any statin therapy at the initial imaging study. Non–statin users were identified as patients not on any statin therapy during the entire study period. A medication history was obtained at each visit associated with AAA imaging. Patients who were found at follow-up to have had a change in statin therapy (initiation or discontinuation) compared with the initial imaging study of AAA were excluded from the analysis. The imaging modalities used included ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI). Ultrasound was the imaging modality of choice in 40% of the patients; CT, in 37%; and MRI, in 23%. Of all patients, 90% had 2 imaging examinations completed, 7% had 3 examinations, and 3% had 4 examinations. The AAA diameter was defined by the maximum diameter.
Linear aneurysm growth rate (change in aneurysm size from initial imaging to follow-up per year of follow-up) was initially compared between patients who were on statin therapy and patients who did not use statins using Wilcoxon rank sum test. The demographic and clinical variables of these 2 groups of patients were also compared using Pearson χ2 test for the categorical variables and 2-sample t test or Wilcoxon rank sum test for the continuous variables.
Because there are other factors that may be associated with change in aneurysm size, the effect of statin use on change in aneurysm size was also tested, adjusting for the effect of other covariates in a multivariable linear regression model. In addition to statin use, the other independent variables included in this model were the demographic and clinical variables with P < 0.10 from tests comparing between the statin and nonstatin groups and from tests of association with change in aneurysm size.
For the linear regression analysis, the dependent variable was assumed to have a normal distribution. However, test of normality of change in aneurysm size showed that the data were not normally distributed. As an alternative, the change in the natural log (ln) of aneurysm size was used in the regression analysis. By using the difference in ln(aneurysm size), the analysis was based on the relative change in aneurysm size. The estimated mean change in ln(aneurysm size) from the fitted regression model, after back transformation, is expressed as the percentage change from initial aneurysm size. Also, to account for the possibility that follow-up differed among the subjects, ln(duration) was included as a covariate in the regression model.
RESULTS
The study population included 136 patients in the statin and 75 patients in the nonstatin group. The mean follow-up duration was 1 year for both groups. The prevalence of coronary artery disease, hypertension, diabetes, cerebrovascular disease, peripheral vascular disease, chronic obstructive pulmonary disease, and tobacco use was similarly high in both groups. The diagnosis of hyperlipidemia (HLP) was 82% in the statin users and 71% in the nonstatin users (P = 0.052). The concomitant use of β-blockers, angiotensin-converting enzyme inhibitors (ACEIs), and aspirin was also higher in the statin group, whereas the use of calcium channel blockers (CCBs) was greater in the nonstatin patients (Table 2).
The size of the AAA on presentation was similar in both groups (4.1 cm; P = 0.96). Patients treated with statin had a decreased linear aneurysm growth rate, with a median of 0.9 mm/y (interquartile range [IQR], −1.0 to +1.0 mm/y) compared with 3.2 mm/y (IQR, 2.0–4.9 mm/y) for those not receiving statins (P < 0.0001). After adjusting for the observed differences in HLP, use of β-blockers, ACEIs, CCB, and aspirin (Tables 1 and 2), the significant difference in the rate of aneurysm growth between statin and nonstatin groups was maintained (P < 0.0001). The estimated mean relative changes in aneurysm size computed from the fitted regression model for the statin and nonstatin groups and adjusted for the other covariates are shown in Table 3. Using the median initial aneurysm size of 4.1 cm, this corresponds to a 1-year mean change in aneurysm size of −0.3 mm (95% confidence interval [CI], −1.1 to 0.6 mm) for the statin group and 3.1 mm (95% CI, 2.2–3.9 mm) for the nonstatin group. The subgroup analysis was performed for patients who had the same imaging modality throughout the study period (117 statin and 65 nonstatin patients, respectively). The median linear aneurysm growth rate was 0.9 mm/y (IQR, −1.0 to +1.0 mm/y) for the statin group compared with 2.9 mm/y (IQR, 1.9–4.4 mm/y) for those not receiving statins (P < 0.0001).
DISCUSSION
The pathogenesis of AAA formation is complex and not completely understood, partly explaining the absence of an effective medical therapy. Formation of AAA is a product of the complex interaction among inflammation, vascular smooth muscle cell (VSMC) apoptosis, extracellular matrix degradation, and remodeling.18Studies of human AAA tissue have shown extensive inflammatory infiltrates containing macrophages and lymphocytes in both the media and the adventitia.19Activated macrophages secrete various proteinases, leading to an imbalance between the synthesis and the degradation of connective tissue proteins. Various extracellular proteinases participate in the process of the destruction of the human aortic wall; in particular, MMP-2 and MMP-9 have attracted interest in this process.20–23Activation of MMP directly contributes to AAA formation in murine models.24,25
Reactive oxygen species (ROS) are increased in the aneurysm wall compared with the normal aorta and adjacent nonaneurysmal aortic wall.26The infiltrated inflammatory cells are the primary source of ROS production through the activity of nicotinamide adenine dinucleotide phosphate oxidase.26Excess ROS generation increases the expression of MMPs and induces apoptosis of VSMC in the aneurysm wall.27–29Antioxidant therapy or the inhibition of nicotinamide adenine dinucleotide phosphate oxidase activity reduced the expression of MMPs and protected from AAA development in murine models.30–32
Statins possess several pleiotropic effects that target the pathophysiologic mechanisms of AAA formation. They are expected to prevent AAA development through anti-inflammatory effects, antioxidative effects, inhibition of proteases, and up-regulation of synthesis of extracellular matrix proteins.33Statins inhibit various inflammatory mediators and other key molecules, including MMPs produced by VSMCs and macrophages.34,35These effects are attributed in part to interference with protein isoprenylation, which plays a major role in inflammatory signaling pathways.36,37Simvastatin suppressed AAA progression in a mouse model and was accompanied by a reduction of MMP-9.38In an ex vivo human organ culture system, the application of cerivastatin reduced the tissue level of MMP-9 in a concentration-dependent manner, accompanied by the inhibition of the activation of infiltrated inflammatory cells.31Experimental studies have also shown that statins have multiple other beneficial effects including improvement of endothelial function through increase in endothelial nitric oxide synthase, suppression of medial VSMC apoptosis, and a reduction in recruitment of macrophages into the vascular wall.39,40
The clinical evidence for the efficacy of statins for AAA is sparse. In a prospective study, 37 patients were randomized to a 3-week course of simvastatin versus placebo before open aneurysm repair.41The excised aneurysm tissue from the simvastatin group had decreased MMP-9 levels consistent with animal studies described previously. A recent meta-analysis identified 7 high-quality clinical controlled studies that examined the effect of statin therapy on AAA expansion rate.42Of these, 6 studies were retrospective cohorts and 1 was a prospective cohort study, whereas only 4 contained more than 200 patients. There was a significant reduction in AAA expansion in patients taking a statin in 3 of the studies25,43,44; however, no significant difference was found in AAA growth between patients taking statins and control patients in 4 of the studies.45–48After the meta-analysis of these 7 studies (1006 patients taking statin and 3191 control patients), the sensitivity analysis showed no difference in AAA expansion rate with statin therapy.17Because of the different methods used in the estimation of AAA growth in the different studies, the validity of using a meta-analysis in these data is unclear and may introduce bias.49A separate recent meta-analysis reached the contrasting conclusion that statin therapy is associated with reduced AAA growth but did not include some of the larger more recently published studies.50
Our study shows an association between statin use and decreased growth rate of AAA. The mean growth rate for the nonstatin group was 3.2 mm/y, which is consistent with the expected rate for small AAAs (2.6–3.2 mm/y).51In contrast, the mean growth rate in the statin group was 0.9 mm/y. Known risk factors that affect the AAA growth rate, including age, sex, aneurysm diameter at first presentation, smoking, and diabetes, are similar in both groups. In addition, we did not find a relationship between HLP and the growth of AAA. The use of β-blockers and ACEIs was significantly higher in the statin group. There is evidence that the use of these medications might slow the development of AAA.52–54In our study, neither of these classes of medication affected the growth rate of the AAA; however, this conclusion is limited by the possibility of a type 2 error. On the other hand, our data do show an association of the use of CCB with a slower rate of aneurysm growth (Table 3). However, because the nonstatin users had the greater use of CCB (Table 2), this variable did not contribute to our primary finding of the protective effect of statins on aneurysm growth. The potential protective effect of CCB is consistent with animal studies showing inhibition in progression of experimental AAA by CCB, in part through suppression of MMP activity55,56; however, the protective role of CCB has not been found in human AAAs.57
This study has several methodological limitations. The study was performed without a prestudy sample size calculation and thus may be underpowered. The compliance of patients with daily statin therapy throughout the study period is not known. It is important to note that the patients in our study had a single repeat imaging study approximately 1 year after the initial study, resulting in a mean follow-up period of only a little longer than 1 year. This mean follow-up is short for a research study on aneurysm growth. The rate of AAA growth is likely to be more accurate in the 10% of patients having more than 2 measurements of AAA size over a couple of years. In addition, this study includes patients who had different imaging modalities (ultrasound, CT, and MRI), which introduce potential errors by increasing the variability of the measurements. Interobserver and intraobserver variabilities in the measurement of the AAA diameter is also a variable with the potential to influence the outcome. And finally, the growth of AAA within a patient does not occur at a linear rate. Instead, the change in aortic diameter is characterized by periods of rapid growth and quiescence. As a result, simple estimates of growth usually overestimate the rate of progression.
In conclusion, our study contributes to the limited human data available regarding the beneficial effects of statin therapy in asymptomatic AAA and demonstrates an association between statin use and decreased growth rate of AAA.
ACKNOWLEDGMENT
The authors thank Dr. Miriam Bridget Zimmerman for statistical consultation.