Author + information
- Received March 29, 2011
- Revision received May 11, 2011
- Accepted May 14, 2011
- Published online October 11, 2011.
- Bob J.H. van Kempen, MSc⁎,†,
- Sandra Spronk, PhD⁎,†,
- Michael T. Koller, MD‡,
- Suzette E. Elias-Smale, MD, MSc⁎,†,
- Kirsten E. Fleischmann, MD, MPH§,
- M. Arfan Ikram, MD, PhD⁎,†,
- Gabriel P. Krestin, MD, PhD†,
- Albert Hofman, MD, PhD⁎,
- Jacqueline C.M. Witteman, PhD⁎ and
- M.G. Myriam Hunink, MD, PhD⁎,†∥,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. M. G. Myriam Hunink, Room Ee 21-40a, Erasmus Medical Center, Dr. Molewaterplein 40, 3015 GD, Rotterdam, the Netherlands
Objectives The aim of this study was to assess the (cost-) effectiveness of screening asymptomatic individuals at intermediate risk of coronary heart disease (CHD) for coronary artery calcium with computed tomography (CT).
Background Coronary artery calcium on CT improves prediction of CHD.
Methods A Markov model was developed on the basis of the Rotterdam Study. Four strategies were evaluated: 1) current practice; 2) current prevention guidelines for cardiovascular disease; 3) CT screening for coronary calcium; and 4) statin therapy for all individuals. Asymptomatic individuals at intermediate risk of CHD were simulated over their remaining lifetime. Quality-adjusted life years (QALYs), costs, and incremental cost-effectiveness ratios were calculated.
Results In men, CT screening was more effective and more costly than the other 3 strategies (CT vs. current practice: +0.13 QALY [95% confidence interval (CI): 0.01 to 0.26], +$4,676 [95% CI: $3,126 to $6,339]; CT vs. statin therapy: +0.04 QALY [95% CI: −0.02 to 0.13], +$1,951 [95% CI: $1,170 to $2,754]; and CT vs. current guidelines: +0.02 QALY [95% CI: −0.04 to 0.09], +$44 [95% CI: −$441 to $486]). The incremental cost-effectiveness ratio of CT calcium screening was $48,800/QALY gained. In women, CT screening was more effective and more costly than current practice (+0.13 QALY [95% CI: 0.02 to 0.28], +$4,663 [95% CI: $3,120 to $6,277]) and statin therapy (+0.03 QALY [95% CI: −0.03 to 0.12], +$2,273 [95% CI: $1,475 to $3,109]). However, implementing current guidelines was more effective compared with CT screening (+0.02 QALY [95% CI: −0.03 to 0.07]), only a little more expensive (+$297 [95% CI: −$8 to $633]), and had a lower cost per additional QALY ($33,072/QALY vs. $35,869/QALY). Sensitivity analysis demonstrated robustness of results in women but considerable uncertainty in men.
Conclusions Screening for coronary artery calcium with CT in individuals at intermediate risk of CHD is probably cost-effective in men but is unlikely to be cost-effective in women.
In asymptomatic individuals, primary prevention of coronary heart disease (CHD) is often based on the predicted 10-year risk of a CHD event. The Framingham risk factors are widely adopted for this purpose (1,2). Guidelines on cardiovascular disease (CVD) prevention recommend advice on a healthy lifestyle (e.g., smoking cessation, regular physical activity) for individuals with a low CHD risk (<10%, 10-year risk), supplemented by statins, antihypertensive medication, and sometimes aspirin for individuals at high CHD risk (>20%, 10-year risk) (3–5). In individuals at intermediate risk (10% to 20%, 10-year risk), the decision to treat with drugs is generally only recommended when either serum cholesterol or blood pressure levels are above a defined threshold. In this group, performing a noninvasive test might afford identification of those who could benefit from more aggressive treatment. Coronary artery calcium on computed tomography (CT), quantified by the CT coronary calcium score, is such a test (6,7).
Recent studies have demonstrated that the CT calcium score is a strong predictor of CHD risk, independent of the Framingham risk factors (7–16). In fact, more than one-half of the individuals originally classified at intermediate risk, on the basis of the Framingham risk factors, are reclassified to the high-risk (>20%) or low-risk (<10%) category when the calcium score is taken into account (7,17). Accordingly, these individuals should be treated more aggressively (high risk) or less aggressively (low risk). The reclassification to another risk category suggests that using CT might be beneficial but reclassification by itself is insufficient evidence to justify implementation (18,19). Studies, ideally clinical trials, demonstrating comparative effectiveness and cost-effectiveness are necessary.
In the absence of clinical trials showing the benefit of CT screening, an extensive evaluation of CT coronary calcium scoring with observational data is warranted (20). The objective of this study was to assess the comparative effectiveness and cost-effectiveness of screening an asymptomatic elderly population at intermediate risk for CHD for coronary calcium with CT.
We developed a Markov decision model with TreeAge for Health Care (TreeAge Pro 2009, TreeAge Software, Williamstown, Massachusetts) to analyze relevant strategies in asymptomatic elderly individuals at intermediate risk for CHD. The model structure, model parameters, and data sources are briefly described here. Details of the modeling assumptions and parameter estimation are given in the Online Appendix.
The following 4 strategies were considered (Fig. 1):
1. “Current practice.” This strategy reflects the incidence of CHD and non-CHD events of individuals at intermediate risk without any additional preventive intervention, as observed in the Rotterdam Study, and is used as the reference strategy. Some individuals were treated at baseline with statins, antihypertensive medication, or aspirin by their general practitioners—considered to be reflected in the observed incidence of CHD and stroke.
2. “Current guidelines.” This strategy, based on fully implementing the most recent guidelines on primary prevention of CHD for individuals at intermediate risk for CHD, implies giving lifestyle advice to all, statin therapy when baseline low-density lipoprotein (LDL) cholesterol exceeds 130 mg/dl (3.37 mmol/l) (4), and antihypertensive medication when baseline systolic blood pressure exceeds 140 mm Hg (5). In a sensitivity analysis, we lowered the LDL threshold to 100 mg/dl (2.59 mmol/l).
3. “CT calcium screening.” In this strategy a CT scan was performed to determine the coronary calcium score, and the 10-year CHD risk was recalculated on the basis of the Framingham risk factors and the calcium score combined. Consequently, a number of individuals will be reclassified to the high-risk or low-risk category. Individuals reclassified to the low-risk category received lifestyle advice and pharmacological treatment if systolic blood pressure was above 140 mm Hg (21) and/or plasma LDL levels were >160 mg/dl (4.14 mmol/l) (4). Individuals who remained in the intermediate-risk category were treated as recommended for individuals at intermediate risk, similar to strategy 2. Individuals reclassified to the high-risk group received lifestyle advice, statin therapy, and antihypertensive medication, irrespective of their baseline cholesterol and blood pressure levels. In addition, men received low-dose aspirin (80 to 100 mg daily). For both the current guidelines and CT calcium screening strategy, we assumed that individuals who used any of the 3 drugs at baseline would continue to use them.
4. “Statin therapy.” For this strategy we assumed that everyone not currently taking a statin would receive a moderate-dose statin and would be otherwise managed according to “current practice.” Although initiating statins in all individuals is not always considered feasible in all situations, it puts the CT calcium screening strategy into a broader perspective, between the least aggressive strategy (“current practice”) and fairly aggressive strategy (“statin therapy”), providing a range of possibilities for an individual at intermediate risk of CHD (20). Conceptually, an even more aggressive strategy would be to treat everyone not only with statins but also with antihypertensive medication and aspirin (in men). In a sensitivity analysis, we substituted the statin therapy strategy with this “aggressive medical treatment” strategy.
For each of the 4 strategies, the model kept track of quality of life, costs, and time spent in 1 of the following health states: 1) well; 2) after CHD event; 3) after major bleeding; 4) after stroke event; 5) after stroke event and CHD event; 6) after stroke and major bleeding; 7) after CHD event and major bleeding; 8) after CHD event and stroke event and major bleeding; 9) CHD or stroke death; and 10) non-CHD or nonstroke death. Each simulated individual started out in the “well” state. Age- and sex-specific probabilities of non-CHD death, fatal and nonfatal myocardial infarction, fatal and nonfatal major bleeding due to aspirin use, fatal and nonfatal stroke, and lethal cancer due to radiation determined the transition to the other states during each annual cycle. The time horizon was the remaining lifetime of the simulated individuals.
A CHD event was defined as any of the following outcomes: nonfatal myocardial infarction, coronary artery bypass graft, percutaneous coronary intervention, and CHD mortality. In sensitivity analysis, we repeated the analysis with “hard” CHD events as outcome, consisting of nonfatal myocardial infarction and CHD mortality. Stroke was defined as ischemic, hemorrhagic, or undefined stroke on cerebral CT. Major bleeding due to aspirin therapy (defined as extra cranial hemorrhage leading to substantial disability) was modeled as a secondary event.
After a CHD event, stroke event, or a major bleeding episode, individuals moved to the “post-CHD-event,” “post-stroke event,” or “post-major bleeding” state, respectively, or the combined states if 2 or all 3 events occurred. After a major bleeding episode, we assumed that aspirin therapy would be discontinued. In the case of a nonfatal CHD or stroke, individuals would be allocated medical treatment for secondary CVD prevention. Non-CHD deaths included fatal cancer due to radiation associated with CT scanning.
Rotterdam Study and event rates
From 1997 onwards, 2,028 participants in the Rotterdam Study underwent CT to determine their coronary calcium score and were subsequently followed for 9.2 years (median) (12,17,35). Primary care physicians were blinded for the findings on CT. Interobserver and intraobserver agreement on calcium scoring has been found to be excellent (36). Two regression models were developed to predict the 10-year risk of CHD on the basis of the Framingham risk factors (prediction model 1) and on the basis of Framingham plus the coronary calcium score (prediction model 2) (17). The Framingham risk factors included were: age, systolic blood pressure, antihypertensive medication, total and high-density lipoprotein cholesterol, diabetes, and current smoking (4). More than 50% of the individuals classified as Framingham intermediate-risk patients were reclassified to either high- or low-risk when CT coronary calcium was added as risk factor and the C-statistic increased significantly from 0.72 to 0.76 (17). The net improvement in reclassification was found to be 0.14 (p < 0.01).
After excluding individuals who had a history of CHD or stroke before the CT coronary calcium scan, we used the baseline Rotterdam Study data and the 2 prediction models to: 1) determine the baseline characteristics of the target population; and 2) determine the proportion of Framingham intermediate-risk individuals reclassified to low and high risk when the coronary calcium score was added. Of all individuals reclassified to the low-, intermediate-, or high-risk categories, we observed how many of them actually suffered from a CHD or stroke event with survival analysis stratified by sex (Table 1).
Probabilities of having a non-CHD event were calculated on the basis of age- and sex-specific mortality rates from national life tables of the general population (37). Life expectancy was adjusted for quality-of-life with mean health-related quality-of-life weights on the basis of published data (Table 1) (32).
Effectiveness of treatment
The benefit of statin and antihypertensive treatment on CHD and stroke incidence was obtained from meta-analyses and considered equal for men and women (24,30). On the basis of a recent update, there is evidence that elderly men benefit from aspirin therapy in primary prevention of CHD. For elderly women there remains considerable controversy (3). Therefore, aspirin treatment for primary prevention was, when applicable, only modeled in men (38).
Treatment adherence is an important determinant of treatment benefit (39). Although we used intention-to-treat-based relative risk reductions on the basis of clinical trials, which take into account adherence, the adherence rate in a population-based intervention is less than that achieved in the controlled setting of a trial. We assumed, on the basis of expert opinion, adherence to treatment in our population to be 70% of the adherence in the original trials for the reference case analysis and explored a range of 20% to 100% in a sensitivity analysis.
For secondary prevention and primary prevention in high-risk individuals, statins, antihypertensive medication, and—in men—aspirin therapy, are combined. Wald and Law (40) estimated the effect of combining medication for CHD prevention, but their approach does not account for possible synergy or dyssynergy between the drugs (41). Instead, we estimated the effect of combining drugs by multiplying the individual relative risks and multiplying the product by a synergy factor, which we varied in sensitivity analyses between 0.9 and 1.10—0.9 implying synergy, 1.0 implying independent effects, and 1.10 implying dyssynergy (see Online Appendix for details). The range in the synergy factor was chosen such that a combination of drugs was at least as effective as a single component of the combination of the same drugs.
Because we considered a population at intermediate risk, we accounted for the fraction of individuals that used (a combination of) statins, aspirin, or antihypertensive medication at baseline. An individual using statins at baseline but with LDL cholesterol levels >160, >130, and >90 mg/dl for the low-, intermediate-, and high-risk categories, respectively, was assumed to switch to a higher dose or more potent statin. The same was assumed for an individual using antihypertensive medication at baseline and systolic blood pressure levels >140 mm Hg, >140 mm Hg, and >120 mm Hg for the 3 risk categories, respectively.
We assumed that all individuals in the Rotterdam Coronary Calcium Study received lifestyle advice consistent with current primary care practice and therefore that the observed CHD and stroke event rates reflected this intervention.
Hemorrhagic stroke due to aspirin therapy was accounted for in the odds ratios of net treatment benefit for stroke from the meta-analysis. Extracranial major bleeding due to aspirin therapy was modeled explicitly as a secondary event with probabilities on the basis of a recent meta-analysis (23). Myopathy and hepatitis were modeled on the basis of a meta-analysis of the adverse effects of statins (42). On the basis of a recent modeling study by Pletcher et al. (43), we calculated the expected costs and disutilities of a myopathy and hepatitis episode, including costs for associated complications such as hospital admission, workup, and mortality, weighted by the probability of complications.
Costs incorporated in the model included healthcare costs and non-healthcare costs and were assessed from the societal perspective for the United States (Table 1). All costs were converted to the year 2010 with the consumer price indexes.
Healthcare costs included costs of diagnostic procedures; costs for personnel, materials, and equipment; costs for medication; costs for healthcare resource use in subsequent years after an event; and costs for overhead. The costs for a noncontrast cardiac CT were based on healthcare reimbursement rates in 2009. Medication costs were based on pricing information from the 2009 Red Book (33), which were comparable with current prices for statins, antihypertensive medication, and aspirin. We assumed, on the basis of baseline LDL cholesterol, that 30% of our population would need a potent and more expensive statin, such as rosuvastatin or atorvastatin, and the remaining 70% could do with a generic statin such as simvastatin. For antihypertensive medication we assumed that everyone would need at least a thiazide, combined with an angiotensin II receptor blocker, angiotensin-converting enzyme inhibitor, or calcium channel blocker in 60% of individuals (44). Medication costs were only accounted for in adherent individuals. In a sensitivity analysis, we used generic prices for statins and antihypertensive medication, estimated to be $160 yearly for generic statins and $300 for antihypertensive medication (33). For both strategies 2 and 3, we accounted for the costs of obtaining the Framingham risk factors by a general practitioner, including laboratory costs. Event-related costs included the costs of hospital stay, diagnostic workup, interventions, and rehabilitation during the first year after an event and were assumed to reflect the average cost after a nonfatal myocardial infarction, coronary artery bypass graft, or percutaneous coronary intervention (31,34,45,46). Non-healthcare costs included travel costs and patient time costs.
All authors agreed on the model structure and data input before performing the analyses to ensure an objective and unbiased analysis.
Important baseline characteristics, such as lipid levels, blood pressure, and medication use, were determined for the cohort of individuals at intermediate risk, stratified by sex. The number of individuals using a statin, antihypertensive, or aspirin under each strategy was determined. Quality-adjusted life years (QALYs), lifetime costs, incremental cost-effectiveness ratios (ICER) (i.e., additional costs divided by QALYs gained), and net health benefit (QALYs minus [costs/willingness-to-pay]) were calculated for all strategies. Future costs and effectiveness were discounted, to take into account time preference, at the currently recommended U.S. discount rate of 3% for both costs and effectiveness (47,48). To take into account second-order uncertainty, 100,000 independent samples were drawn from each of the input parameter distributions, generating outcome distributions for QALYs and costs for each strategy. Calculations were done for men and women separately.
Strategies were first ordered according to increasing cost. A strategy was considered dominated if another strategy was both more effective and less costly. A strategy was considered extended dominated if another strategy achieved more effectiveness at a lower ICER. The ICERs were calculated, after eliminating dominated and extended dominated strategies, as the difference in mean lifetime costs divided by the difference in mean QALYs for each strategy compared with the next-best non-dominated strategy. We considered $50,000/QALY gained as a commonly accepted threshold for the societal willingness-to-pay threshold for primary prevention (49–51) and varied it between $15,000 and $100,000 in sensitivity analyses. For the reference case analysis, we analyzed the model with input parameters as given in Table 1.
Extensive 1-way, 2-way, multi-way, and probabilistic sensitivity analyses were performed with plausible ranges of the parameter values. In particular, we explored model sensitivity to drug costs, aspirin therapy in women, and the relative risk of an event with aspirin therapy. Because some clinicians would be reluctant to withhold therapy from an individual who starts out with a predicted risk of 11% (putting him originally at intermediate risk) and—after inclusion of coronary calcium—a revised risk of 9% (putting him at low risk), we explored the effect of an alternative assumption in which treating individuals reclassified to the 5% to 10% risk category as individuals with intermediate risk (10% to 20%) and checked whether the optimal decision would change. Reclassification probabilities for this assumption are presented in Online Table 4.
Because the 2004 guidelines on the initiation of statin therapy include an optional cutoff value of 100 mg/dl for individuals at intermediate risk, we did an additional analysis with this cutoff value in the “current guidelines” strategy and the “CT calcium screening” strategy for the individuals who remained in the intermediate-risk group.
Probabilistic sensitivity analysis was performed with the outcome distributions of 100,000 Monte Carlo simulations (52). We calculated the probability that CT screening was cost-effective, compared with current practice, current guidelines, and statin therapy strategies for varying willingness-to-pay thresholds, which yielded acceptability curves.
Reference case analysis
Review of the baseline characteristics of the cohort at intermediate risk demonstrated that women were older than men and had less favorable risk factor levels, apart from smoking and calcium scores (Table 2). In men, implementing current guidelines for all individuals at intermediate risk led to a steep increase in the number of statin and antihypertensive users (from 12% to 75% and 23% to 64%, respectively), compared with current practice (Online Table 5). In women, a similar pattern was observed (from 15% to 87% and 52% to 84%, respectively) (Online Table 6). Implementing the CT screening strategy results in slightly fewer statin users, compared with implementing current guidelines in both men (69% vs. 75%) and women (41% vs. 87%). In men, statin users with either current practice or CT screening had a higher expected 10-year risk of CHD compared with nonusers (Online Table 7). This difference disappeared between users and nonusers with current guidelines. In women, this was only the case for CT screening (Online Table 8). In men (Table 3), CT calcium screening was more effective and more costly compared with current practice (QALY gain: 0.13 [95% confidence interval (CI): 0.01 to 0.26], cost-increase: $4,676 [95% CI: $3,126 to $6,339]), more effective and more costly than statin therapy (QALY gain: 0.04 [95% CI: −0.02 to 0.13], cost increase: $1,951 [95% CI: $1,170 to $2,754]), and more effective but slightly more costly than current guidelines (QALY gain: 0.02 [95% CI: −0.04 to 0.09], cost increase: $44 [95% CI: −$441 to $486]). The cost-effective plane in Figure 2A shows that, in men, current guidelines are extended dominated by CT screening, because the latter leads to a higher expected quality-adjusted life expectancy against a lower incremental cost-effectiveness ratio. The incremental cost-effectiveness ratio of statin therapy is $30,278/QALY, and for CT calcium screening it is $48,800/QALY gained (Table 3). In women (Table 4), CT screening was more effective and more costly than current practice (QALY gain: 0.13 [95% CI: 0.02 to 0.28]; cost increase $4,663 [95% CI: $3,120 to $6,277]), more effective and more costly than statin therapy (QALY gain: 0.03 [95% CI: −0.03 to 0.12], cost increase: $2,273 [95% CI: $1,475 to $3,109]), and less expensive but also less effective compared with current guidelines (QALY loss: 0.02 [95% CI: −0.03 to 0.07], cost savings: $297 [95% CI: −$8 to $633]). The cost-effective plane in Figure 2B shows that, in women, CT screening is extended dominated by current guidelines, because the latter leads to a higher expected quality-adjusted life expectancy against a lower incremental cost-effectiveness ratio, and therefore, CT screening is not considered cost-effective in women.
In men, at a willingness-to-pay threshold of $50,000/QALY, a slight dyssynergy between drugs would change the optimal decision from CT screening to statin therapy (Table 5). This shift would also occur if treatment adherence dropped below 58%, the effect of aspirin therapy on CHD was less protective, the cost of a CT scan rose above $200, or the risk of radiation-induced cancer increased more than 10-fold. In women, the optimal strategy changed from “current guidelines” to statin therapy in case of a slight dyssynergy between drugs. Strong protective effects of aspirin on the incidence of CHD and/or stroke would change the optimal strategy to “CT screening” (Table 6). Using generic drug prices made the CT screening more cost-effective in men, with an ICER of $24,675/QALY, whereas in women current guidelines became more cost-effective, with an ICER of $21,140/QALY. Substituting the statin therapy strategy with the aggressive medical treatment strategy did not change the optimal decision in men. In women the optimal decision switched from current guidelines to aggressive medical treatment.
Probabilistic sensitivity analysis demonstrated that, in men, CT screening was cost-effective compared with current practice in the majority of simulations if the willingness-to-pay threshold was above $50,000 (Fig. 3A). In women, even at higher willingness-to-pay thresholds, CT calcium screening would be cost-effective in <20% of the simulations (Fig. 3B).
In this study we evaluated the comparative effectiveness and cost-effectiveness of CT coronary calcium screening within the framework of current CVD prevention guidelines. In men, the incremental cost-effectiveness ratio for CT screening was just below the willingness-to-pay threshold of $50,000/QALY, and small changes in assumptions changed CT screening from being cost-effective to not cost-effective. Some of the assumptions could be considered plausible, such as a slight dyssynergy between drugs or a treatment adherence lower than 60%, whereas others were more extreme (e.g., a more than 10-fold increase in radiation risk). The uncertainty in optimal decision was further illustrated by the acceptability curves, which showed that, in a minor but substantial proportion of the simulations, CT screening was not cost-effective. However, with generic drug prices the ICER for CT screening dropped, and the result was more robust in sensitivity analysis.
In women, CT screening was not found to be cost-effective, even after using a wide range of varying assumptions, which included assumptions more favorable to the CT calcium screening strategy by treating individuals in the higher end of “low risk” (5% to 10% risk) more aggressively and with more treat-prone LDL thresholds. The difference in the optimal decision between men and women can be explained by the fact that, compared with men, more women were reclassified to the low-risk group, leading to less aggressive treatment. Furthermore, within the low-risk group, the observed risk of CHD is higher in women than in men, so the foregone benefit with less aggressive treatment is higher in women. The benefit of CT screening is obtained in the high-risk group, where individuals are treated more aggressively compared with current guidelines for treatment of intermediate-risk individuals. Because fewer women were reclassified to high risk, the potential benefit of CT screening is lower than in men. The balance is further shifted because aspirin is prescribed in men at high risk but not in women, due to controversy with regard to its efficacy in primary prevention of CHD.
The Adult Treatment Panel (ATP)-IV guidelines, which will be published soon, are expected to recommend more aggressive statin treatment than the current statin treatment guidelines. Our statin therapy strategy can be considered quite aggressive and is likely to be similar to the anticipated ATP-IV recommendation, ensuring future applicability of our results. Of note, when we compared CT screening with an even more aggressive treatment strategy, as we did in the sensitivity analysis with the “medical treatment” strategy, CT screening remained cost-effective in men. This implies that CT screening does not simply put more individuals on treatment but allocates treatment to individuals who are expected to benefit most.
A number of cost-effectiveness reports on CT coronary calcium scoring have previously been published but differed from our study in the strategies or target population considered or in that they dichotomized the calcium score rather than including the score in a risk prediction. These studies found that cost-effectiveness of CT screening was highly sensitive to the population screened and downstream costs (53–55). The relatively high incremental cost-effectiveness ratio we found for CT screening in men is comparable to results of other cost-effectiveness studies on interventions for primary prevention of CHD, such as the study by Pletcher et al. (43). Generalizability of our findings is further supported by comparable reclassification data on coronary calcium found by Polonsky et al. (7) in the multi-ethnic study of atherosclerosis.
First, we focused on individuals at intermediate risk, which implied individuals were on average older than 69 years of age. Screening for coronary calcium could potentially have value in other subgroups, but we explicitly chose to investigate CT screening in the intermediate-risk group as advocated by recent guidelines and current consensus. Second, the time horizon in our analysis was the remaining lifetime. Therefore, we had to extrapolate the incidence of CHD beyond the available 10-year data, but few simulated individuals lived beyond 15 years. Finally, although we stratified by sex, further stratification by different combinations of baseline risk factors was not possible due to a limited sample size.
As with all models of screening and diagnostic tests, the differences between the 4 strategies in terms of quality-adjusted life expectancy were small. Even though, in women, the results seem robustly unfavorable for the CT coronary calcium screening strategy, the residual uncertainty reflected in the acceptability curves indicates that further research might be beneficial. In men, the results indicated that CT screening was cost-effective in the majority of simulations. Nevertheless, in a substantial proportion of simulations in men, current guidelines or statin therapy was optimal compared with CT screening, indicating that further research is necessary.
Screening for coronary artery calcium with CT is probably cost-effective in men at intermediate risk of CHD. For women at intermediate risk for CHD, CT screening does not seem to be cost-effective.
For supplementary figures, tables, and text, please see the online version of this article.
The study was funded by ZonMw, project number 62300047. The funding source had no role in the design and conduct of the study; collection, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Dr. Krestin has served as a consultant to GE Healthcare. All other authors have reported that they have no relationships relevant to the content of this paper to disclose.
- Abbreviations and Acronyms
- coronary heart disease
- confidence interval
- computed tomography
- cardiovascular disease
- incremental cost-effectiveness ratio
- low-density lipoprotein
- quality-adjusted life year
- systolic blood pressure
- Received March 29, 2011.
- Revision received May 11, 2011.
- Accepted May 14, 2011.
- American College of Cardiology Foundation
- Wilson P.W.,
- D'Agostino R.B.,
- Levy D.,
- Belanger A.M.,
- Silbershatz H.,
- Kannel W.B.
- D'Agostino R.B. Sr..,
- Vasan R.S.,
- Pencina M.J.,
- et al.
- Grundy S.M.,
- Cleeman J.I.,
- Merz C.N.,
- et al.
- Pearson T.A.,
- Blair S.N.,
- Daniels S.R.,
- et al.
- Kondos G.T.,
- Hoff J.A.,
- Sevrukov A.,
- et al.
- Arad Y.,
- Goodman K.J.,
- Roth M.,
- Newstein D.,
- Guerci A.D.
- Taylor A.J.,
- Bindeman J.,
- Feuerstein I.,
- Cao F.,
- Brazaitis M.,
- O'Malley P.G.
- Vliegenthart R.,
- Oudkerk M.,
- Hofman A.,
- et al.
- LaMonte M.J.,
- FitzGerald S.J.,
- Church T.S.,
- et al.
- Budoff M.J.,
- Shaw L.J.,
- Liu S.T.,
- et al.
- Elias-Smale S.,
- Proença R.V.,
- Koller M.T.,
- et al.
- Pletcher M.J.,
- Pignone M.
- Brugts J.J.,
- Yetgin T.,
- Hoeks S.E.,
- et al.
- Deckers J.W.,
- Goedhart D.M.,
- Boersma E.,
- et al.
- Greving J.P.,
- Buskens E.,
- Koffijberg H.,
- Algra A.
- Law M.R.,
- Morris J.K.,
- Wald N.J.
- Lightowlers S.,
- McGuire A.
- Rice T.
- World Health Organization
- Ho P.M.,
- Bryson C.L.,
- Rumsfeld J.S.
- Wald N.J.,
- Law M.R.
- Corso P.S.,
- Hammitt J.K.,
- Graham J.D.,
- Dicker R.C.,
- Goldie S.J.
- Hirth R.A.,
- Chernew M.E.,
- Miller E.,
- Fendrick A.M.,
- Weissert W.G.
- Lee K.K.,
- Cipriano L.E.,
- Owens D.K.,
- Go A.S.,
- Hlatky M.A.
- Myrium Hunink M.G.,
- Glasziou P.P.,
- Siegel J.E.,
- et al.