Ratio of Apolipoprotein A-II/B Improves Risk Prediction of Postoperative Survival After Carotid Endarterectomy
Background and Purpose—Even in patients with high-grade carotid stenosis, cardiovascular morbidity causes more deaths than strokes do. Despite successful low-density lipoprotein (LDL) cholesterol lowering, a significant risk of atherosclerotic cardiovascular disease remains, eventually rendering other lipid or lipoprotein ratios more efficient treatment targets. This study aimed to investigate the predictive value of the ratio of serum apolipoprotein A-II/B for overall mortality (primary outcome) of carotid surgery patients.
Methods—This single-center, nonrandomized, prospective cohort study comprised 327 consecutive patients undergoing carotid endarterectomy for high-grade internal carotid artery stenosis. Baseline lipoprotein concentrations were measured, and patients were observed for the occurrence of the primary outcome until the census date (January, 2003 to January, 2012; median follow-up, 102.3 months).
Results—The ratio of apolipoprotein A-II/B (hazard ratio, 0.74 per SD; confidence interval, 0.60–0.91; P=0.004) showed the highest association with the primary outcome compared with other lipid-risk parameters, significantly improving a prognostic model based on major cardiovascular risk factors, including LDL, high-density lipoprotein, and triglycerides in terms of overall performance, calibration, and discrimination. This led to a significantly improved reclassification of 8.9% of all patients (net reclassification improvement, 0.137; P=0.006 and integrated discrimination improvement, 0.041; P<0.001) and of 13.6% of patients with a serum baseline concentration of <100 mg/dL LDL (net reclassification improvement, 0.270; P=0.030 and integrated discrimination improvement, 0.061; P=0.002).
Conclusions—Apolipoprotein A-II/B significantly improves risk prediction of overall survival, also in carotid surgery patients with lower LDL levels. Consequently, this ratio might provide an efficient diagnostic tool and eventually a treatment target for actual lipid-lowering therapies, which has to be addressed in future randomized controlled trials.
Cardiovascular disease remains the leading cause of death1 with hyperlipidemia as a major contributing factor.2 Even in patients with high-grade internal carotid artery stenosis (ICAS) at high risk of ischemic stroke, myocardial infarction is responsible for >50% of perioperative deaths and causes more postoperative deaths than strokes,3 thus reducing the benefit of carotid surgery, especially in patients with asymptomatic ICAS.
Accurate assessment of lipoprotein-associated cardiovascular risk is the basis for many screening and therapeutic efforts.4 Yet, despite successful low-density lipoprotein cholesterol (LDL) lowering, a significant cardiovascular risk remains,5 eventually rendering measurement of other lipids, apolipoproteins, or their ratios more effective targets for therapy.4 The role of apolipoprotein (Apo) A-II in atherosclerosis has long been controversial6 until in 2013 Wang et al7,8 reported for the first time a causative atheroprotective effect in hypercholesterolemic transgenic animals expressing human dimeric ApoA-II. The atherogenic role of LDL’s major ApoB has been widely acknowledged.9
This study aimed to investigate the predictive value of the ratio of ApoA-II/B serum concentrations for overall survival (primary outcome) after carotid endarterectomy.
Materials and Methods
An expanded Method section is available in the online-only Data Supplement.
This single-center, nonrandomized, prospective cohort study, approved by the local Ethics Committee (EC number, 04-067-0604), evaluated the association of the ratio of ApoA-II/B, compared to other established lipid-risk parameters, with postoperative mortality of 327 consecutive patients undergoing carotid surgery for ICAS from 2003 to 2012, median observation period of 102.3 months (8.5 years). The primary study outcome was death from any cause (overall death).
Sample Handling and Laboratory Measurements
Blood was drawn 0 to 7 days before carotid surgery. Lipoproteins were separated using a combined ultracentrifugation–precipitation method (β-quantification).10 The levels of total cholesterol and triglycerides were measured using enzymatic reagents (Wako Chemicals, Neuss, Germany). Lipids and apolipoproteins (A-I, A-II, and B) were determined by immunoturbidimetry (Greiner, Flacht, Germany). Lipoprotein analyses were performed on an Olympus AU640 (Olympus Diagnostika, Hamburg, Germany).
Quantitative associations of apo-, lipoproteins, and ratios with the primary outcome (overall mortality) were estimated in multivariable proportional hazard models (Cox regression). Prognostic indices (PI)11 with or without inclusion of the ApoA-I/B or ApoA-II/B ratio were compared with regard to general performance (R2), calibration (X2, Hosmer–Lemeshow test), discrimination, and reclassification.12 A 2-sided P value of <0.05 was considered to indicate statistical significance.
Demographs, baseline concentrations, and main outcomes of 327 patients are presented in Table I in the online-only Data Supplement. After 8.5 years of observation time, 139 patients (42.5%; women, 45 [13.8%]; men, 94 [28.7%]) had died. The ratio of ApoA-II/B best predicted overall survival (hazard ratio, 0.74 per SD; confidence interval, 0.60–0.91; P=0.004), presented in Tables 1 and 2, and cardio-, cerebro-, and vascular survival (hazard ratio, 0.62 per SD; confidence interval, 0.28–0.90; P=0.012). Pairwise comparison with ApoA-II/B in univariable and multivariable proportional hazard models predicting overall survival rendered all other parameters, including LDL/high-density lipoprotein and ApoA-I/B, insignificant (P>0.05). Moreover, ApoA-II/B showed the highest stratification, illustrated in Figure I in the online-only Data Supplement.
Prognostic Value of ApoA-II/B
To estimate the usefulness of ApoA-II/B as a prognostic marker, PI based on established major cardiovascular risk factors, including history of symptomatic ICAS and lipid risk parameters (LDL; triglycerides) with (PIincl ApoAII/B) or without (PIestablished), the ratio of ApoA-II/B, were computed. The PIincl ApoAII/B showed a better general performance (Cox–Snell, R2=0.240; Nagelkerke, R2=0.322) compared with the established model (Cox–Snell, R2=0.202; Nagelkerke, R2=0.271). Likewise, the new model’s calibration improved (Hosmer–Lemeshow, X2=10.11, P=0.257 instead of X2=3.62, P=0.890), and its discrimination increased significantly (P<0.001; Figure).
Most important, inclusion of ApoA-II/B in the prognostic model caused a significantly improved reclassification of 29 of 327 (8.9%) patients (net reclassification improvement, 0.137, P=0.006; integrated discrimination improvement, 0.041, P<0.001), presented in Table II in the Data Supplement, and also of 12 of 88 patients with low LDL (<100 mg/dL; 3.6%; net reclassification improvement, 0.270; P=0.041 and integrated discrimination improvement, 0.061; P=0.002), presented in Table III in the online-only Data Supplement.
In 2013, Wang et al7 reported significantly reduced atherosclerotic lesions in the aorta and coronary arteries of transgenic animals expressing, for the first time, human dimeric ApoA-II. Likewise, ApoB is the major apolipoprotein of LDL, but it is also present in intermediate-density lipoproteins and very low-density lipoprotein cholesterol, which makes its measurement a more accurate estimate of the total concentration of atherogenic particles and, thus, superior in cardiovascular risk prediction.9 Therefore, in this study, we investigated the ratio of ApoA-II/B, which emerged as a better predictor of lipoprotein-associated risk of cardiovascular and overall mortality compared with all other established lipoprotein parameters, including ApoA-I, ApoA-II, or ApoA-I/B in patients undergoing carotid endarterectomy for ICAS. Moreover, its inclusion in a prognostic model based on major cardiovascular risk factors13 and established lipoprotein risk factors allowed a significant improvement of prediction of postoperative overall mortality, also in patients at lower lipid risk, ie, with LDL serum concentrations of <100 mg/dL. Consequently, this ratio may provide an efficient treatment target for actual lipid-lowering therapies.
Eventual changes in patients’ dietary habits, which might bias this study’s findings, have not been documented.
This study identifies the ratio of apolipoprotein A-II/B as the best prognostic parameter for lipoprotein-associated risk of postoperative overall mortality compared with other established lipid parameters in patients undergoing carotid surgery for high-grade ICAS. This suggests that this ratio might provide an efficient diagnostic tool and treatment target of actual lipid-lowering therapies, which is supported by its retained predictive value within a low-LDL range.
N. Duschek had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. N. Duschek and A. Assadian contributed to the study concept and design; J. Strassegger, S. Ghai, J. Basic, J. Falkensammer, and N. Duschek contributed to the acquisition of data; N. Duschek, T. Stojakovic, H. Scharnagl, J. Falkensammer, K. Huber, and A. Assadian contributed to the analysis and interpretation of data; N. Duschek, J. Basic, J. Strassegger, S. Ghai, K. Huber, and A. Assadian contributed to the drafting of the article; N. Duschek, T. Stojakovic, H. Scharnagl, J. Basic, J. Falkensammer, K. Huber, and A. Assadian contributed to the critical revision of the article for important intellectual content. N. Duschek, T. Stojakovic, J. Ghai, H. Scharnagl, and J. Falkensammer contributed to statistical analysis; T. Stojakovic, H. Scharnagl, J. Strassegger, S. Ghai, and J. Basic contributed to Administrative, technical, or material support; N. Duschek and A. Assadian contributed to study supervision.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.009663/-/DC1.
- Received April 7, 2015.
- Revision received April 7, 2015.
- Accepted April 8, 2015.
- © 2015 American Heart Association, Inc.
- Go AS,
- Mozaffarian D,
- Roger VL,
- Benjamin EJ,
- Berry JD,
- Blaha MJ,
- et al
- McQueen MJ,
- Hawken S,
- Wang X,
- Ounpuu S,
- Sniderman A,
- Probstfield J,
- et al
- Wang Y,
- Niimi M,
- Nishijima K,
- Waqar AB,
- Yu Y,
- Koike T,
- et al
- Remaley AT
- Bachorik PS,
- Ross JW
- 11.↵IBM. Spss advanced statistics 21. ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/21.0/de/client/Manuals/IBM_SPSS_Advanced_Statistics.pdf. 2012. Accessed March 28, 2015.
- D’Agostino RB Sr,
- Vasan RS,
- Pencina MJ,
- Wolf PA,
- Cobain M,
- Massaro JM,
- et al