Author + information
- Received September 19, 2012
- Revision received December 17, 2012
- Accepted December 26, 2012
- Published online March 26, 2013.
- Kevin F. Erickson, MD, MS⁎,†,⁎ (, )
- Sohan Japa, MBA†,
- Douglas K. Owens, MD, MS†,‡,
- Glenn M. Chertow, MD, MPH⁎,
- Alan M. Garber, MD, PhD§ and
- Jeremy D. Goldhaber-Fiebert, PhD†
- ↵⁎Reprint requests and correspondence:
Dr. Kevin F. Erickson, Division of Nephrology, Stanford University School of Medicine, 780 Welch Road, Suite 106, Palo Alto, California 94305
Objectives The authors sought to evaluate the cost-effectiveness of statins for primary prevention of myocardial infarction (MI) and stroke in patients with chronic kidney disease (CKD).
Background Patients with CKD have an elevated risk of MI and stroke. Although HMG Co-A reductase inhibitors (“statins”) may prevent cardiovascular events in patients with non–dialysis-requiring CKD, adverse drug effects and competing risks could materially influence net effects and clinical decision-making.
Methods We developed a decision-analytic model of CKD and cardiovascular disease (CVD) to determine the cost-effectiveness of low-cost generic statins for primary CVD prevention in men and women with hypertension and mild-to-moderate CKD. Outcomes included MI and stroke rates, discounted quality-adjusted life years (QALYs) and lifetime costs (2010 USD), and incremental cost-effectiveness ratios.
Results For 65-year-old men with moderate hypertension and mild-to-moderate CKD, statins reduced the combined rate of MI and stroke, yielded 0.10 QALYs, and increased costs by $1,800 ($18,000 per QALY gained). For patients with lower baseline cardiovascular risks, health and economic benefits were smaller; for 65-year-old women, statins yielded 0.06 QALYs and increased costs by $1,900 ($33,400 per QALY gained). Results were sensitive to rates of rhabdomyolysis and drug costs. Statins are less cost-effective when obtained at average retail prices, particularly in patients at lower CVD risk.
Conclusions Although statins reduce absolute CVD risk in patients with CKD, the increased risk of rhabdomyolysis, and competing risks associated with progressive CKD, partly offset these gains. Low-cost generic statins appear cost-effective for primary prevention of CVD in patients with mild-to-moderate CKD and hypertension.
Approximately 8% of U.S. adults are believed to have chronic kidney disease (CKD) (1). Because CKD is an independent risk factor for cardiovascular disease (CVD), patients with CKD face increased risks of cardiovascular (CV) morbidity and mortality (2–4). HMG-CoA reductase inhibitors (“statins”) are widely used to prevent CVD in the general population. Although clinical trials failed to demonstrate a benefit from statins in end-stage renal disease (ESRD) (5,6), subgroup analyses of several large trials (7–10) and 3 meta-analyses (11–13) demonstrate that statins reduce the relative risk of myocardial infarction (MI) and stroke in patients with mild-to-moderate CKD at a magnitude similar to reduction for non-CKD patients (7,14). Results from the SHARP (Study of Heart and Renal Protection) trial also suggest that statins prevent CV events in patients with advanced, non–dialysis-requiring CKD (15). Given the high CV risk for patients with non–dialysis-requiring CKD and the relatively low cost of generic statins, these drugs have the potential to improve health at an acceptable cost. Balanced against the potential benefits is the fact that patients with CKD experience decreased life expectancy, higher healthcare costs, and reduced quality of life. Additionally, statin-associated rhabdomyolysis (i.e., life-threatening muscle toxicity) is more common in persons with CKD than in the general population (16).
Although clinical trials support the efficacy of statins for patients with non–dialysis-requiring CKD, no consensus recommendations exist for their use in primary prevention. Several guidelines recommend that CKD be considered a “coronary heart disease risk equivalent” similar to diabetes when deciding whether to initiate statins (17,18). However, the Adult Treatment Panel (ATP) III guidelines do not include CKD as an independent risk factor for CV events (19). Treating CKD as a coronary heart disease risk equivalent could lead to statin treatment for up to 2 million additional persons with CKD in the United States (20).
We developed a decision-analytic model to assess the cost-effectiveness of statins for primary CVD prevention in non–dialysis requiring CKD. The analysis focuses on the sizable subset of the CKD population with hypertension, but with no other traditional CV risk factors, for whom the ATP III guidelines would not recommend statins.
We developed a Markov model of CVD and CKD to evaluate statin therapy for primary CV prevention in the following cohorts (Online Figs. S1a to S1c):
1. The base case: 65-year-old men and women with mild-to-moderate CKD and moderate hypertension (systolic blood pressure [SBP]: 130 to 140 mm Hg on treatment).
2. Additional cohorts of men and women defined by: a) age at initiation of statins; b) presence (and degree) of hypertension; c) CKD stage; and d) progression of CKD.
The model follows simulated patients over their lifetimes in 3-month intervals. Individuals can have an acute MI or stroke, die during or after surviving an MI or stroke, or die from non-CV causes. Those with prior nonfatal CV events have further elevated risks of mortality from both repeat CV events and CVD-associated complications (Online Fig. S1a).
Patients progress through CKD stages 3a (estimated glomerular filtration rate [eGFR] 45 to 59 ml/min/1.73m2), 3b (eGFR: 30 to 44 ml/min/1.73m2), 4 (eGFR: 15 to 29 ml/min/1.73m2), and 5 (eGFR <15 ml/min/1.73m2). Costs, quality of life, mortality, and rates of CV events vary by CKD stages until patients progress to stage 5 CKD, after which they experience mortality rates and costs equal to the averages of similarly aged U.S. patients with ESRD (Online Appendix: Mortality in ESRD; Online Fig. S1b). Outcomes included quality-adjusted life years (QALYs), life years, direct healthcare costs in 2010 U.S. dollars and incremental cost-effectiveness ratios discounted at 3% annually (21). We report costs and incremental cost-effectiveness ratios rounded to the nearest $100.
Model inputs were derived from published literature (Table 1). Rates and variability of CKD progression were derived from large observational cohorts of comparable patient populations via model calibration (Online Appendix: Modeling CKD Progression, Online Figs. S7 to S11). Baseline probabilities of MI and stroke were derived from age- and sex-based Framingham risk scores (Online Appendix: Determining Baseline Cardiovascular Risk). Baseline hazards of CV events estimated from the Framingham equations were multiplied by CKD stage–specific hazard ratios reflecting the increased risk of CV events independently associated with CKD stages 3a, 3b, and 4 (3,4). All-cause mortality rates for patients in each CKD stage were calculated by multiplying CKD stage–specific all-cause mortality hazards by age- and sex-specific mortality in the general population obtained from U.S. life tables (4,22). For each CKD stage, rates of noncardiovascular death were imputed from all-cause mortality rates by adjusting for competing cardiovascular risks (Online Appendix: Calibrating Rates of Noncardiovascular Death; Online Figs. S12 to S14).
Relative risk reductions for MI and stroke in stages 3a and 3b CKD were obtained from a meta-analysis (11) and were similar to the treatment effect observed in patients with CKD from the Pravastatin Pooling Project and in 2 other meta-analyses (9,12,13) (Online Table S5). Although lacking definitive evidence, evaluation of clinical trials suggests a trend toward decreased CVD risk reduction from statins for patients with stage 4 CKD (11,15). Consequently, we assumed the CV relative risk reduction equaled that reported in the SHARP trial for patients with stage 4 CKD (15). Additionally, we assumed that statins do not reduce CV risks for stage 5 CKD, that is, these patients did not incur added costs or receive added health benefits from statins (5,6). We assumed that all patients, unless statins are contraindicated, would be prescribed statins after experiencing a CV event and that published hazards of death and costs of medical care following CV events reflect the effects of statins. In our base case, we assumed statins do not influence the likelihood or rates of CKD progression (11).
Adverse effects of statins
Rates of muscle-related toxicity from statins are elevated in patients with CKD (16). We assumed patients with CKD on statins experience a 1-time risk of myalgias upon initiating statin therapy and an ongoing risk of rhabdomyolysis. Because previous analyses show that statins for primary prevention are not cost-effective if patients experience even minor quality-of-life decrements from myalgias (23), we assumed patients who experience myalgias discontinue statins. Patients who developed myalgias incurred a cost for checking muscle enzyme levels and a 3-month quality-of-life decrement due to the muscle pain. In the event of rhabdomyolysis, patients experience a risk of death from the acute illness, and for those surviving, statins were permanently discontinued.
Costs and quality of life
We included costs and quality-of-life decrements for each CKD stage, MI, stroke, and rhabdomyolysis, and for patients surviving an MI or stroke (Table 1; Online Appendix: Selected Model Assumptions). In the base case, statin costs included 40 mg of generic pravastatin daily available at discount retailers and integrated health systems in addition to laboratory monitoring and clinical follow-up (24–27).
Alternative assumptions about CKD and model validation
To understand how CKD progression affects the health benefit and cost-effectiveness of statins, we assessed outcomes in otherwise similar patient groups with nonprogressive CKD and without CKD. To verify that our assumptions accurately describe mortality in patients with CKD, we compared life expectancies produced from each patient group to U.S. life tables, demonstrating stepwise decreases following additions of hypertension, nonprogressive CKD, and progressive CKD (Online Tables S6 and S7).
All model inputs were varied in sensitivity analyses. The effect of statin costs was tested for a range of baseline cardiovascular risks. Probabilistic sensitivity analysis evaluated how the simultaneous uncertainties about model parameters might influence outcomes (Online Tables S8 and S9). We examined scenarios in which statins were assumed to slow the rate of CKD progression by up to 19% and scenarios in which statins increased the likelihood of developing either diabetes or memory loss (28–30).
Initiating statin therapy for patients with mild-to-moderate (stage 3a) CKD, moderate hypertension, and no other CV risk factors led to a modest improvement in health outcomes for 50- to 85-year-old men and women (Table 2). For 65-year-old men, statin therapy increased nondiscounted life expectancy by 50 days (from 11.65 to 11.78 years) and increased discounted QALYs by 36 days (from 7.21 to 7.31 QALYs). Statins reduced the risk of having at least 1 MI or stroke over a lifetime (or before progression to ESRD) from 39.8% to 34.7%. Relative benefits were smaller in patients with lower CV risk. For instance, for 65-year-old women, statin therapy increased nondiscounted life expectancy by 34 days (from 13.74 to 13.84 years) and increased QALYs by 20 discounted days (from 8.20 to 8.26 QALYs). The corresponding risk of at least 1 MI or stroke before development of ESRD was reduced from 24.1% to 20.6%.
Treatment with statins increased total lifetime costs for all patient groups (Table 2). Combining health benefits and costs, statin therapy for 65-year-old men with mild-to-moderate CKD and moderate hypertension cost $18,000 per QALY gained, and for 65-year-old women cost $33,400 per QALY gained, assuming the cost of statins at $4 per month (24,25).
Influence of CKD progression
Patients with advanced CKD and ESRD experience lower health-related quality of life, higher rates of costly morbidities, and premature mortality compared with patients without CKD. Although the increased CV risk in patients with CKD means that statins provide a larger absolute risk reduction in CV events compared with patients without CKD, higher “downstream” costs and worse health outcomes associated with progressive CKD blunted the cost-effectiveness of interventions to avert CV events in patients with less severe CKD. In 65-year-old men and women without CKD and moderate hypertension, statins cost $10,200 per QALY gained. In patients of similar age and hypertension severity with mild-to-moderate, nonprogressive CKD, statins cost $16,400 per QALY gained, slightly more than in patients without CKD. With CKD progression, the cost-effectiveness attenuated further, reflecting the greater costs and health decrement experienced due to progressive CKD for patients whose CV events were averted with statins. For a cohort of 65-year-olds with moderate hypertension, CKD, and the possibility of CKD progression, statins cost $25,800 per QALY gained (Online Fig. S15). In other words, statins will allow more patients with mild-to-moderate CKD to survive free from CV events who later experience more morbid and costly stages of advanced CKD and ultimately ESRD.
Age and baseline cardiovascular risk
Older individuals with CKD have substantially shorter life expectancies, higher ongoing medical costs, decreased likelihood of surviving long enough to progress to ESRD, and higher CV risks compared with younger individuals. In net, there was a trend toward statin use being more cost-effective for older patients with CKD (Table 2).
Statins were less cost-effective in groups with lower baseline CV risk. When varying the range of baseline CV risks (designated by the presence and severity of hypertension in conjunction with sex and age at statin initiation), statin therapy cost between $16,100 and $146,700 per QALY gained (Fig. 1).
When purchased at average retail prices, the cost-effectiveness of statin therapy compared favorably with other commonly accepted treatments in patients at higher CV risks (Fig. 2). If 40 mg of pravastatin is purchased at $47 per month, statin therapy cost between $51,700 and $87,700 per QALY gained in 50- to 85-year-old men with mild (SBP: 120 to 130 mm Hg on treatment) or moderate hypertension and $81,600 to $112,700 per QALY without hypertension; in women age 70 and older with mild or moderate hypertension, statin therapy cost between $74,300 and $110,900 per QALY gained. Due to their lower baseline CV risk, statins were less favorable in younger women with and without hypertension, where the cost was $117,700 to $746,700 per QALY gained (Online Tables S11 and S12; Online Figs. S16 to S18). Despite its greater potential efficacy (owing to more potent lowering of LDL cholesterol), brand-name rosuvastatin cost more per QALY gained, ranging from $80,500 per QALY gained in 65-year-old men to $390,500 per QALY in 50-year-old women.
Results were sensitive to the range of rhabdomyolysis risk in CKD, particularly in younger, lower-risk groups facing prolonged exposure to statins (Fig. 3). Due to higher healthcare costs, lower quality of life, and elevated risk of non-CV mortality in more advanced CKD, statins were less cost-effective when started in patients with lower GFR, despite higher CV risks (Online Fig. S24). Although not definitive, evidence suggests that statins may reduce the rate of CKD progression (28). If statins were to reduce CKD progression rates by 19%, they are cost-saving in 65-year old men and women (Online Fig. S25). When we considered the possibility of statins causing either diabetes or permanent memory loss, the cost-effectiveness of statins did not substantially change for CKD patients with higher CV risk (Online Figs. S26 and S27). However, for CKD patients with lower CV risks, statin therapy was substantially less cost-effective (or dominated by a strategy of no statins). In 65-year-old men and women (who are at higher CV risk), if the long-term risk of diabetes from statins is 2-fold higher in women and 6-fold higher in men than described, no statins becomes the preferred strategy because it yields longer quality-adjusted life expectancy at a lower cost (Fig. 4).
Probabilistic sensitivity analysis
Considering the simultaneous uncertainty in all model parameters, statins cost <$50,000 per QALY gained in over 99% of simulations for 50- and 65-year-old men, and in 94% of simulations for 65-year-old women. In 50-year-old women, statins cost <$50,000 per QALY gained in 38% of simulations, and cost <$100,000 per QALY gained in 90% of simulations (Online Figs. S28 to S31).
Statins for primary prevention of CVD in patients with moderate hypertension and mild-to-moderate CKD can increase life expectancy by 0.6 to 1.7 months and prevent MI or stroke for 2% to 5% of patients before developing ESRD. If generic statins were obtained at $4 per month (available from discount retail programs and integrated health systems) (24,25), they were cost-effective for patients with a wide range of CV risks, costing <$25,803 per QALY in men with and without hypertension and ranging from $20,152 to $60,043 per QALY in women with mild or moderate hypertension, comparing favorably with many commonly accepted treatments.
Statins are less cost-effective when their prices are higher and in patients with lower baseline CV risks. At average retail prices, statin therapy costs between $51,835 and $97,314 per QALY in men with mild or moderate hypertension and in women above age 70 with moderate hypertension. For patients with lower baseline CV risk (i.e., women below age 70 with mild or moderate hypertension, women of all ages without hypertension, and men below age 60 without hypertension), statin therapy cost between $100,818 and $746,741 per QALY.
In addition to baseline CV risk and statin price, the cost-effectiveness of statins was particularly sensitive to drug-related toxicity. This highlights the importance of exercising caution when prescribing statins for primary cardiovascular prevention to patients with CKD. An increase from the baseline incidence of rhabdomyolysis makes statins substantially less cost-effective when given to younger patients with lower CV risk. Because lower costs make statins affordable to larger segments of the population, patient education about signs and symptoms of muscle toxicity, as well as prompt discontinuation of therapy at the first sign of muscle injury, will be increasingly important. More knowledge about the relative risks of rhabdomyolysis from different statins in patients with CKD would be informative. Additionally, it will be important to verify that long-term statin risks such as diabetes and memory loss are low.
Our findings suggest that CKD may be different from traditional CV risk factors when considering statin cost-effectiveness. Current ATP III guidelines recommend initiating statin therapy in the general population according to a person's risk of future cardiovascular events as determined by a Framingham CV risk score (31). This recommendation is determined in part on the basis of the observation that the relative reduction in risk of CVD appears to be independent of the baseline risk (8,32). This observation has led to past findings that statins are more cost-effective in higher-risk populations where the absolute risk reduction is the largest (23,33,34).
Our analysis reaches a different conclusion for adults with CKD, hypertension, and no other traditional CV risk factors: Despite an increased risk of MI and stroke in patients with CKD, statins are not more cost-effective in patients with CKD than they are in the general population after considering traditional CV risk factors. This is because the presence of CKD introduces the following combination of offsetting health and economic influences: 1) patients with CKD are at an increased risk of CV disease and related mortality, increasing statins' net benefit; 2) patients with CKD are at an increased risk of muscle toxicity, reducing statins' net benefit; and 3) CKD progression leads to increased medical costs, decreased quality of life, increased non-CV mortality, and apparent ineffectiveness of statins once patients develop ESRD, reducing statins' net benefit. These partly offsetting health and economic influences are most pronounced when considering statin therapy in patients with lower baseline CV risks and when statins are obtained at higher prices.
First, it is unknown precisely how the benefits from statin therapy decline as CKD advances. Although clinical trials demonstrate that statins are effective for CV prevention in stage 3 CKD (7–10), statin trials in patients with ESRD have not shown a benefit (5,6). Because of these disparate observations—and evidence that CVD may have a different pathophysiologic basis in more advanced CKD (35)—we assume the treatment effect from statins diminishes in advanced CKD stages. Although this assumption reflects current observations, the assumed step-wise decrease in treatment effect has not been demonstrated definitively. Second, when patients with CKD progress to ESRD, medical care becomes substantially more expensive. By prolonging life in patients with CKD, a part of statins' incremental cost is due to the high downstream costs of dialysis. Manns et al. (36) discuss how including downstream healthcare costs in economic evaluations of patients on dialysis are “methodologically correct [but] may mitigate against the acceptance of interventions that are relatively inexpensive themselves but which improve patient survival.” This may pose challenges for policymakers because, as a society, we currently pay for the high cost of ESRD care. It is counterintuitive that a strategy aimed at prolonging life for patients with non–dialysis-requiring CKD would be subjected to a more stringent criterion; a higher cost-effectiveness threshold for statins in this population may therefore be justified. Understanding whether there is a direct benefit of statins on CKD progression would be particularly informative in this regard. Finally, the incidence of diabetes due to statins has been described for the general population, but is not well described in the CKD population. However, in our exploratory analysis, we found statins to be cost-effective in the base case despite using cost, quality-of-life, and hazard estimates that were weighted towards making diabetes more severe in its health and economic ramifications.
Although statins reduce absolute CVD risk in patients with CKD, increased risk of rhabdomyolysis, and competing risks associated with progressive CKD, partly offset these gains. Nevertheless, low-cost generic statins are cost-effective for primary prevention of CVD in patients with mild-to-moderate CKD and hypertension. At average retail prices, it could be argued that statin therapy for patients with CKD should be reserved for subgroups defined by higher cardiovascular risk and greater potential to benefit. Statins are only cost-effective in populations at lower CV risk if they have low risk of rare, but clinically important, side effects.
The authors thank Arjun Adhikari, MS, and Joshua Glucoft, MS, for their contribution to early model design and data collection.
For an expanded Methods section and supplementary figures and tables, please see the online version of this paper.
Supported by Agency for Healthcare Research and Quality grant F32 HS019178 (Dr. Erickson); DK085446 (Dr. Chertow); and AG037593 (Dr. Goldhaber-Fiebert). Dr. Owens was supported by the Department of Veterans Affairs. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Adult Treatment Panel
- chronic kidney disease
- cardiovascular disease
- estimated glomerular filtration rate
- end-stage renal disease
- myocardial infarction
- quality-adjusted life year
- Received September 19, 2012.
- Revision received December 17, 2012.
- Accepted December 26, 2012.
- American College of Cardiology Foundation
- Muntner P.,
- He J.,
- Hamm L.,
- Loria C.,
- Whelton P.K.
- Ridker P.M.,
- MacFadyen J.,
- Cressman M.,
- Glynn R.J.
- Tonelli M.,
- Isles C.,
- Curhan G.C.,
- et al.
- Shepherd J.,
- Kastelein J.J.P.,
- Bittner V.,
- et al.
- Navaneethan S.D.,
- Pansini F.,
- Perkovic V.,
- et al.
- Antman E.M.,
- Anbe D.T.,
- Armstrong P.W.,
- et al.
- Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults
- Lee K.K.,
- Cipriano L.E.,
- Owens D.K.,
- Go A.S.,
- Hlatky M.A.
- Walmart Pharmacy
- ZenRx Research
- Department of Health and Human Resources: Centers for Medicare & Medicaid Services
- Department of Health and Human Resources: Centers for Medicare & Medicaid Services
- Tonelli M.,
- Isles C.,
- Craven T.,
- et al.
- Grundy S.M.,
- Cleeman J.I.,
- Merz C.N.B.,
- et al.
- Pharoah P.D.,
- Hollingworth W.
- Schwarz U.,
- Buzello M.,
- Ritz E.,
- et al.
- Consumer Reports Health Best Buy Drugs
- United States Renal Data System
- Fryback D.G.,
- Dasbach E.J.,
- Klein R.,
- et al.