The potential of medical therapy to influence the courses and outcomes of patients with thoracic aortic aneurysms is not known. The aim of this study was to determine whether statin intake is associated with improved long-term outcomes in these patients. A total of 649 patients with thoracic aortic aneurysms were studied, of whom 147 were taking statins at their first presentation and 502 were not. After a median follow-up period of 3.6 years, 30 patients (20%) taking statins had died, compared with 167 patients (33%) not taking statins (hazard ratio 0.68, 95% confidence interval 0.46 to 1, p = 0.049); 87 patients (59%) taking statins reached the composite end point of death, rupture, dissection, or repair compared with 378 patients (75%) not taking statins (hazard ratio 0.72, 95% confidence interval 0.57 to 0.91, p = 0.006). After adjustments for co-morbidities, the association between statin therapy and the composite end point was driven mainly by a reduction in aneurysm repairs (hazard ratio 0.57 95% confidence interval 0.4 to 0.83, p = 0.003). On Kaplan-Meier analysis, the survival rate of patients taking statins was significantly better (p = 0.047). In conclusion, the intake of stains was associated with an improvement in long-term outcomes in this cohort of patients with thoracic aortic aneurysms. This was driven mainly by a reduction in aneurysm repairs.
Aortic aneurysms are among the leading causes of death in the United States, and the only definitive treatment for ruptured aortic aneurysms is (surgical) repair. Statins have become widely used drugs in patients with cardiovascular disease, and their benefit is probably not limited to their lipid-lowering effect. Statins can influence remodeling of the vascular and cardiac wall by influencing the activity of matrix metalloproteases and of plasminogen activators and their respective inhibitors, which play important roles in the development and the progression of aneurysms. Thus, statins may have a beneficial effect on the progression of abdominal aortic aneurysms. However, a recent population-based study in patients with abdominal aortic aneurysms showed no decreased rupture rate in patients taking lipid-lowering drugs. We investigated whether statins have a protective effect on the outcomes and progression of patients with thoracic aortic aneurysms.
Methods
We included patients who were entered into the thoracic aortic aneurysm database at the Yale Center for Thoracic Aortic Disease, a major referral center for New England, until April 2006. Patients entered before January 1988, the year when the first statin was approved by the United States Food and Drug Administration, were excluded. We used demographic and clinical (including medications and outcomes such as dissection, rupture, and operative repair) variables and serial aneurysm measurements from imaging studies from the computerized database maintained by the center as part of ongoing studies. The database has been used in the past to investigate the clinical course of thoracic aneurysm progression and for better defining criteria for surgical intervention. The study was approved by the Yale human investigation committee.
We identified patients who had recorded information about the intake of statins at the time of their first visits and who had ≥1 measurement of aneurysm size. Statin therapy was prescribed at the discretion of the primary physician. The recorded smoking status reflected whether the patient was smoking at the time of database entry. Hyperlipidemia was recorded if the patient had a known diagnosis of hyperlipidemia at database entry. All patients who had surgical repair underwent open repair of their aneurysms (the first stent grafts were approved by the Food and Drug Administration in 2005). The indication for surgery was usually evidence of rupture or dissection, symptoms caused by the aneurysm, or rapid aneurysm growth or aneurysm size >5 or 5.5 cm in the ascending aorta or >6 or 6.5 cm in the descending aorta in Marfan and non-Marfan patients, respectively. Follow-up was at Yale University and was done primarily as clinically indicated. Because this was a post hoc analysis, there was no prespecified follow-up protocol. The end points were adjudicated using the Social Security Death Index database and chart abstraction as previously described. Briefly, hospital chart review was conducted on each identified patient, and the data were entered into a computerized database. Data recovered from hospital records and computer files were cross-referenced with hospital discharge abstract data monitored by the Connecticut Hospital Association and the Connecticut State Mortality Records as well as the Social Security Death Index.
Chi-square and Student’s t tests were used to compare patient characteristics by statin intake status. The main outcomes were death and a composite end point of death, rupture, dissection, or (operative) repair. Hazard ratios (HR) were used to account for different lengths of follow-up. To study differences in survival, outcomes of death and a composite end point of death, rupture, dissection, or repair over time were estimated separately for statin users and nonusers using the Kaplan-Meier method. Survival was measured as the time from enrollment in the study until the time to the outcomes. If the subject was not identified with the outcome of interest at the time of the last visit, he or she was censored as of the last date of the last interval of available follow-up. We limited follow-up to 10 years because the number of patients who had follow-up beyond 10 years was very small, and we censored events that occurred >10 years after the date of entry into the database. Log-rank tests were applied to ascertain any survival differences. These relations were further evaluated using multivariate Cox proportional-hazards regression analyses, with adjustment for other covariates. This method accounts for varying duration of follow-up among subjects. We considered a broad range of potential covariates, such as age, gender, risk factors for coronary disease such as body mass index (weight in kilograms divided by the square of height in meters), diabetes, hypertension, peripheral vascular disease, congestive heart failure, starting (initial) diameter, site of aneurysm, and medication use, as well as date of entry into the database (the later to account for changes in indications for surgery, changes in surgical mortality, and changes in medical therapy patterns over time). Selection into the multivariate model was based on published associations for each potential covariate to determine whether they were appropriate for further analysis. Because the prevalence of peripheral vascular disease was low and its coexistence with coronary artery disease frequent, we used a consolidated variable of cardiovascular (coronary and/or vascular) disease for the multivariate analysis.
To minimize baseline biases such as the fact that statin intake at presentation may have been associated with more aggressive medical care, we also developed a propensity score for the likelihood of receiving statin therapy. The score included the following variables: age; gender; family history of thoracic aneurysm; intake of β blockers, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers; hypertension; diabetes; cardiovascular disease; heart failure; and entry date into the database. The propensity score was then used as an independent covariate in the multivariate analysis. We also tested whether patients taking statins had smaller aneurysms at baseline and/or a slower rate of aneurysm progression compared with those not taking statins. The rate of aneurysm progression was calculated using linear growth assumptions and using the instrumental variable estimation method in patients who had ≥2 imaging studies, as previously described. The statistical software used was Stata version 9.2 (StataCorp LP, College Station, Texas) and SAS (SAS Institute Inc., Cary, North Carolina).
Results
Six hundred forty-nine patients who had their statin intake status recorded at the time of entry into the database and who had ≥1 aneurysm measurement were analyzed. Of these, 147 patients were taking statins at entry into the database and 502 were not. The clinical characteristics of the patients are listed in Table 1 . The mean age was 70 years, and 63% of the patients were men. Patients taking statins and those not taking statins differed significantly in terms of intake of medicines, with patients taking statins taking more β blockers, more angiotensin-converting enzyme inhibitors, and more angiotensin receptor blockers. Also, more patients taking statins had histories of coronary artery disease and hypertension. The initial sizes of the aneurysms were similar between the 2 groups, without any significant difference.
| Characteristic | All | Statin | No Statin | p Value ⁎ |
|---|---|---|---|---|
| (n = 649) | (n = 147) | (n = 502) | ||
| Men | 62.70% | 65.30% | 62.00% | 0.46 |
| Age (years) | 70.3 ± 14 | 71.3 ± 11.2 | 69.9 ± 15.1 | 0.32 |
| BMI (kg/m 2 ) | 24.25 ± 8.5 | 23.92 ± 7 | 24.30 ± 8 | 0.21 |
| Medications | ||||
| β blockers | 47.70% | 65.10% | 43.10% | <0.001 |
| ACE inhibitors | 24.00% | 30.20% | 22.70% | 0.03 |
| ARBs | 5% | 11.40% | 3.50% | <0.001 |
| Family history of TAA | 20.65% | 21.77% | 20.32% | 0.71% |
| Marfan syndrome | 2.90% | 0.60% | 3.50% | 0.066 |
| Hypertension | 32% | 42.10% | 29% | 0.002 |
| CAD | 27.20% | 39.40% | 23.70% | 0.001 |
| PVD | 4.90% | 4.10% | 5.20% | 0.589 |
| Heart failure | 8.10% | 6.10% | 8.70% | 0.304 |
| Smokers | 37.60% | 32.50% | 38% | 0.66 |
| COPD | 26.10% | 28.70% | 25.20% | 0.419 |
| Diabetes mellitus | 6% | 6.80% | 5.70% | 0.645 |
| Chronic kidney disease | 11.90% | 12.50% | 11.70% | 0.811 |
| Hyperlipidemia | 26.80% | 78.40% | 12.10% | <0.001 |
| Aneurysm diameter (cm) | 5.16 ± 1.31 | 5.12 ± 1.22 | 5.17 ± 1.34 | 0.69 |
The aneurysms were located in the ascending aorta in almost 2/3 of the patients ( Table 2 ). The second most frequent location was the descending aorta, in 19% of patients, followed by aneurysms located in the aortic arch and thoracoabdominal aortic aneurysms. The differences between the frequencies of locations in patients who were taking statins versus patients not taking statins were not statistically significant.
| Location | All | Statin | No Statin | p Value ⁎ |
|---|---|---|---|---|
| (n = 649) | (n = 147) | (n = 502) | ||
| Ascending | 425 (65.4%) | 101 (68.7%) | 324 (64.5%) | 0.35 |
| Aortic arch | 60 (9.2%) | 13 (8.8%) | 47 (9.3%) | 0.84 |
| Descending | 129 (19.8%) | 26 (17.6%) | 103 (20.5%) | 0.44 |
| Thoracoabdominal | 61 (9.4%) | 16 (10.8%) | 45 (8.9%) | 0.48 |
After a median follow-up period of 3.6 years, 30 patients (20%) taking statins had died, compared to 167 patients (33%) not taking statins (HR 0.68, 95% confidence interval [CI] 0.46 to 1, p = 0.049); 87 patients (59%) taking statins had reached the composite end point of death, rupture, dissection, or repair compared to 378 patients (75%) not taking statins (HR 0.72, 95% CI 0.57 to 0.91, p = 0.006). Sixty-four of the patients (43.5%) taking statins underwent aneurysm repair compared to 269 of those (53.6%) not taking statins (HR 0.72, 95% CI 0.7 to 91, p = 0.006).
In univariate analysis, age, gender, chronic obstructive pulmonary disease, cardiovascular disease, statin use, family history, aneurysm location, and initial aneurysm size were all significantly associated with death and with the composite end point of death, rupture, or dissection in patients with known statin status. Interestingly, a family history of aortic aneurysm and the location in the ascending aorta were associated with lower rates of the 2 end points. A test for interaction of statins with aneurysm location did not reveal a significant interaction.
On multivariate analysis ( Table 3 ), after adjustment for other factors such as age, gender, chronic obstructive pulmonary disease, cardiovascular disease, family history, aneurysm location, and initial aneurysm size, compared to patients who were not taking statins, patients who were taking statins had HRs of 1.03 (95% CI 0.62 to 1.72, p = 0.89) for death and 0.58 (95% CI 0.42 to 0.8, p <0.001) for death, rupture, dissection, or repair. The intake of statins was associated with a decrease in rupture, dissection, and repair, and the crude as well as the adjusted HRs for (operative) repair were the ones primarily driving the results for the composite end point. The propensity score showed good discrimination between patients taking statins and patients not taking statins (c statistic = 0.72). When the propensity score for statin intake was used in the multivariate analysis, the results did not change significantly ( Table 4 ). When we repeated the analysis without Marfan patients, the results were similar: the crude HR for death was 0.68 (95% CI 0.46 to 1, p = 0.052) and for the composite of death, rupture dissection, and repair was 0.71 (95% CI 0.56 to 0.9, p = 0.005). The adjusted HR for death was 1.05 (95% CI 0.63 to 1.75, p = 0.84) and for the composite of death, rupture dissection, and repair was 0.59 (95% CI 0.42 to 0.81, p = 0.001). The Kaplan-Meier survival analyses ( Figures 1 and 2 ) yielded a significantly improved rate of freedom from death (p = 0.048) and a significantly improved rate of freedom from death, rupture dissection, and repair (p = 0.001) for patients taking statins compared to those not taking statins.
| Variable | Effect Size for Death (95% CI) | p Value | Effect Size for Death, Dissection, Rupture, or Repair (95% CI) | p Value |
|---|---|---|---|---|
| Age | 1.005 (0.99–1.01) | 0.48 | 0.99 (0.98–1.001) | 0.11 |
| Male gender | 0.68 (0.48–0.94) | 0.02 | 0.81 (0.66–1.01) | 0.06 |
| β blockers | 0.86 (0.62–1.2) | 0.39 | 0.99 (0.80–1.23) | 0.96 |
| Aspirin | 1.07 (0.57–2.03) | 0.81 | 1.69 (1.13–2.52) | 0.01 |
| ARBs | 0.85 (0.36–1.98) | 0.71 | 0.60 (0.33–1.08) | 0.09 |
| Hypertension | 1.09 (0.75–1.57) | 0.64 | 0.99 (0.78–1.25) | 0.96 |
| PVD/CAD | 1.21 (0.84–1.75) | 0.29 | 0.96 (0.76–1.23) | 0.78 |
| Diabetes mellitus | 1.73 (1.01–2.95) | 0.04 | 1.03 (0.68–1.55) | 0.88 |
| Family history of TAA | 0.69 (0.43–1.1) | 0.12 | 0.87 (0.67–1.12) | 0.28 |
| Heart failure | 1.1 (0.69–1.78) | 0.65 | 1.07 (0.76–1.51) | 0.68 |
| Hyperlipidemia | 0.76 (0.48–1.20) | 0.24 | 1.22 (0.91–1.62) | 0.17 |
| ACE inhibitors | 0.88 (0.60–1.3) | 0.53 | 1.01 (0.78–1.30) | 0.93 |
| Initial diameter | 1.13 (1.02–1.26) | 0.02 | 1.20 (1.11–1.29) | <0.0001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
