Statins and Aspirin for Chemoprevention in Barrett’s Esophagus: Results of a Cost-Effectiveness Analysis

Abstract

Data suggest that aspirin, statins, or a combination of the two drugs may lower the progression of Barrett’s esophagus (BE) to esophageal adenocarcinoma (EAC). However, aspirin is associated with potential complications such as gastrointestinal bleeding and hemorrhagic stroke, and statins are associated with myopathy. We developed a simulation disease model to study the effectiveness and cost-effectiveness of aspirin and statin chemoprevention against EAC. A decision analytic Markov model was constructed to compare four strategies for BE management; all regimens included standard endoscopic surveillance regimens: 1) endoscopic surveillance alone, 2) aspirin therapy, 3) statin therapy, and 4) combination therapy of aspirin and statin. Endpoints evaluated were life expectancy, quality adjusted life-years (QALY), costs, and incremental cost-effectiveness ratios (ICER). Sensitivity analysis was performed to determine the impact of model input uncertainty on results. Assuming an annual progression rate of 0.33%/ year from BE to EAC, aspirin therapy was more effective and cost less than (dominated) endoscopic surveillance alone. When combination therapy was compared to aspirin therapy, the ICER was $158,000/QALY, which was above our willingness-to-pay threshold of $100,000/QALY. Statin therapy was dominated by combination therapy. When higher annual cancer progression rates were assumed in the model (0.5%/year), combination therapy was cost-effective compared to aspirin therapy, producing an ICER of $96,000/QALY. In conclusion, aspirin chemoprevention was both more effective and cost less than endoscopic surveillance alone. Combination therapy using both aspirin and statin is expensive but could be cost-effective in patients at higher risk of progression to EAC.

Introduction

While the incidence of esophageal adenocarcinoma (EAC) has increased by 500% over the past 40 years (1-3), the management of Barrett’s esophagus (BE), a precursor to EAC, has remained largely ineffective and controversial (4-7). Current strategies and practice employ endoscopic surveillance with biopsy at regular intervals for early detection of EAC and dysplastic states (8). The goal of this surveillance is to reduce mortality and morbidity from EAC by preventing cancer through dysplasia treatment or by identifying cancer before it becomes invasive. Due to the rise of EAC incidence in recent years, chemoprevention has received much attention as a method to reduce the progression from BE to EAC, with numerous studies demonstrating the effectiveness of chemoprevention in BE (9). The use of aspirin or other non-steroidal anti-inflammatory drugs (NSAIDs) and statins has been associated with reduced EAC incidence or progression (10-12). Increased expression of the enzyme cyclooxygenase (COX)-2 has been detected in patients with BE and EAC, and it has been hypothesized that NSAIDs and statins both could act as COX-2 inhibitors (11, 13, 14). A recently published prospective cohort study by Kastelein et al. found that combination use of NSAIDs and statins as chemoprevention was associated with a 78% reduction in EAC incidence in BE patients (15).

Our prior study used a microsimulation disease model of Barrett’s esophagus to determine the effectiveness and cost-effectiveness of aspirin as chemoprevention for esophageal cancer in patients with BE (16). That analysis found that aspirin therapy was both effective and cost-effective when compared to no therapy, either alone or when combined with endoscopic surveillance. While cost-effectiveness analyses have been performed for aspirin as an esophageal chemoprevention agent, to our knowledge, no analysis has investigated the effectiveness and cost-effectiveness of statins, or aspirin and statin combination therapy.

In an attempt to address the need for recommendations regarding the use of chemopreventive agents in patients with BE, our study aimed to analyze the effectiveness and cost-effectiveness of aspirin, statin, and combination chemoprevention for BE management.

Materials and Methods

Model Design

A decision analytic Markov state transition model was constructed in TreeAge Pro 2012 (TreeAge, Williamstown, MA). Four different strategies for the management of BE initially without dysplasia were implemented in the model. All strategies included endoscopic surveillance with esophagectomy performed when cancer was detected and included: 1) endoscopic surveillance alone, 2) aspirin chemoprevention, 3) statin chemoprevention, and 4) combination of aspirin and statin chemoprevention.

Health states in the model included Barrett’s esophagus (no dysplasia, ND), low-grade dysplasia (LGD), high-grade dysplasia (HGD), post successful esophagectomy for cancer, inoperable or incomplete resection of cancer, and death. Possible causes of death included age-related mortality, surgical mortality, EAC, and complications. The Markov cycle length or time between state transitions was 1 month. The simulation began with a hypothetical cohort of 50-year-old individuals who were followed until age 80 or death. In each cycle, the simulated patient could stay in the same state, progress to the next state or cancer, or die from age-related all-cause mortality. For model simplicity and transparency, all patients were assumed to have the correct diagnosis of BE at the start of the model simulation (17, 18).

Endoscopic Surveillance Alone

In the endoscopic surveillance alone strategy (see Figure 1 for simplified schematic), patients would undergo upper endoscopies with biopsies at intervals recommended by American Gastroenterological Association (AGA) guidelines (8). For BE without dysplasia, patients were followed by endoscopic surveillance every 3 years. If patients were found to have LGD, the surveillance continued at 6-month intervals for the first year from diagnosis of LGD and at 12-month intervals thereafter. For HGD patients, endoscopic surveillance continued at 3-month intervals. Esophageal cancers that would undergo surgery were modeled to be either surgically resectable or unresectable based on published rates (16, 19). After surgical esophagectomy was performed on eligible patients, annual endoscopic surveillance was continued.

Figure 1. Endoscopic surveillance alone strategy.

Abbreviations: BE=Barrett’s esophagus; ND=no dysplasia; LGD= low-grade dysplasia; HGD=high grade dysplasia

Legend: A cohort of Barrett’s esophagus patients spends time in a Markov state every cycle until death or age 80. If the patients are found to have cancer, surgery is performed for patients eligible for resection

Aspirin Strategy with Endoscopic Surveillance

The aspirin with endoscopic surveillance strategy was similar to endoscopic surveillance alone strategy, except that patients were simulated to take a 325-mg enteric-coated aspirin daily (see Figure 2 for simplified schematic). Patients who took aspirin daily were modeled to have 53% reduction in the incidence of esophageal adenocarcinoma based on results from a prospective cohort study of BE patients which served as our main chemoprevention source as the analysis examined both aspirin and statins in their cohort (15). These patients could have aspirin-associated complications, such as gastrointestinal or genitrourinary bleeds or hemorrhagic strokes. When patients developed aspirin-associated complications and survived, they were modeled to discontinue the aspirin and return to the endoscopic surveillance alone strategy with standard cancer progression rate. Patients were not modeled to receive any other benefit from the aspirin, such as a cardiac benefit or chemoprevention of other cancers. Adherence to treatment was assumed to be 100% in the absence of complications.

Figure 2. Chemoprevention Strategy.

Abbreviations: BE=Barrett’s esophagus; Chemo = Chemoprevention

Statin Strategy with Endoscopic Surveillance

The statin with endoscopic surveillance strategy was similar to the endoscopic surveillance alone strategy, except that the patients would take a statin daily (see Figure 2 for simplified schematic). Patients who took a statin daily were modeled to have 54% reduction in the incidence of esophageal adenocarcinoma based on results from a cohort study (15). However, these patients could have statin-associated complications, such as myopathy or elevation in tranasminases. When patients developed statin-associated adverse effects and survived, they were modeled to discontinue the statin and continue with the endoscopic surveillance alone strategy with the standard cancer progression rate. Patients were not modeled to receive any other benefit from the statin, and adherence to treatment was assumed to be 100% in the absence of previously described adverse events.

Combination Strategy with Endoscopic Surveillance

In the combination strategy with endoscopic surveillance, patients were modeled to take both aspirin and a statin daily and to follow the endosopic surveillance regimen described above in the prior strategies (see Figure 2 for simplified schematic). Patients who took both medications were modeled to have a 78% reduction in the incidence of esophageal adenocarcinoma based on results from the cohort study (15). However, these patients could have either statin or aspirin-associated complications as described above. When patients developed statin or aspirin-associated adverse effects and survived, they were modeled to discontinue combination therapy and continue with endoscopic surveillance alone strategy with the standard cancer progression rate. Patients were not modeled to receive any other benefit from both drugs, and adherence to treatment was assumed to be 100% in the absence of previously described complications.

Parameter Estimates

Model parameters or inputs were estimated from the published literature. Base-case values and ranges used in sensitivity analyses are summarized in Table 1. When published estimates were not available, an expert in the field was consulted to provide an estimate for the parameter. The base-case estimates for the effects of aspirin, statin, or both on cancer progression rates were based on a prospective cohort study of 570 Barrett’s esophagus patients in the Netherlands (15).

Table 1.

Model Inputs

Parameters	Estimate, Base Case	Source
Risk of neoplastic progression (Hazard ratio)
NSAID only	0.47	(15)
Statin only	0.46	(15)
NSAID and Statin	0.22	(15)
Aspirin characteristics
Annual Risk of Complication (per 100,000)
Noncerebral bleeds
Major bleeds	40	(16, 34)
Intermediate bleeds	180	(16, 34)
Mortality from noncerebral bleeds	4.4	(16, 35)
Hemorrhagic CVA
Disabling	11	(16, 36, 37)
Nondisabling	9	(16, 36, 37)
Mortality from CVA	6	(16, 35)
Statin characteristics
Annual Risk of Complication (per 100,000)
Myopathy
Without rhabdomyolysis	95	(38, 39)
With rhabdomyolysis	1.6	(38, 39)
Mortality from myopathy	0.16	(39, 40)
Hepatitis
Without liver failure	1.74	(38, 39)
With liver failure	0.5	(38, 39)
Mortality from liver failure	0	(39, 40)
Procedure characteristics
Operative candidate, cancer	0.80	(30)
Surgical resectability rate
Surveillance	0.80	(16, 23, 30, 41-46)
No Surveillance	0.33	(16, 19)
Complications
Complication rate from EGD	0.00013	(16, 47, 48)
Mortality from EGD complication	0.0016	(16, 47, 48)
Mortality from esophagectomy	0.05	(30, 46)
Discount rate	0.03	(32, 49)
Transition probabilities
HGD to cancer	0.024	(23)
LGD to HGD		See Methods; Calibrated to overall progression rate from BE to EAC
LGD to cancer
ND BE to LGD
ND BE to cancer
Costs (2011 USD)
Cost of cancer (annual)	49,385	(13, 50-52)
Cost surveillance EGD	930	(30, 53-56)
Cost of post surgery state (annual)	1,496	(30, 53, 54)
Cost of esophagectomy	25,882	(30, 53, 54, 56)
Aspirin related
Aspirin drug cost (annual)	19.56	(26)
CVA complication	51,822	(16, 57)
Major bleed	6,958	(16, 58)
Intermediate bleed	1,332	(16, 51)
Statin related
Statin drug cost (annual)	872	(26)
Myopathy without rhabdomyolysis	30	(38, 50)
Myopathy with rhabdomyolysis	12,343	(38, 52)
Hepatitis without liver failure	38	(38, 50)
Hepatitis with liver failure	16,529	(38, 52)

Abbreviation: EGD = esophagogastroduodenoscopy

Model Transition Probabilities and Calibration

The transition probabilities between the various BE states are critical to the model’s validity. However, there is a wide range of estimates and uncertainty regarding transition rates between specific BE sub-states (e.g. from ND to LGD or LGD to HGD). The best quality and amount of data exist for the overall transition rate from BE to EAC. Because of the pivotal nature of the BE to EAC progression rate and the newly published estimates, the transition probabilities between the BE sub-states were therefore calibrated to generate overall BE to EAC transition rates of 0.12%, 0.33%, and 0.5% per year, which encompass a wide range of values (i.e. low, intermediate and high values) (20-22). The progression rate of 0.33% was used as our base-case estimate. A published rate from HGD to EAC was used in conjunction with other transition estimates to generate the overall BE to EAC rate (23). The transition rate from HGD to cancer was calculated to be 2.4% based on the published 7.3-year probability of 16% (23). As an additional check, the transition rates derived from the calibration were compared to the ranges of transition probabilities that were used for a previously validated US population simulation model of esophageal adenocarcinoma (EACMo) (24) that was calibrated to National Cancer Institute Surveillance, Epidemiology and End Results (SEER) data.

Costs and Utilities

Base-case costs and ranges used in sensitivity analyses are summarized in Table 1. Medicare reimbursement rates were used to estimate direct costs (25). Drug costs were obtained from the 2003 Red Book average wholesale prices (26) The statin cost used in our base-case analysis was obtained by averaging the Red Book prices of simvastatin and lovastatin. Published estimates of costs from prior years were converted to 2011 year dollars using the Consumer Price Index (U.S Bureau of Labor Statistics).

Quality of life measures for various states in the model were adjusted to utility scores for the specific health states: cancer = 0.5 and post-esophagectomy = 0.97 (27-30). For the base-case analysis, all cost and expected life years were discounted at an annual rate of 3% to adjust for the relative value of present dollars or a present year of life (31).

Outcomes

The primary outcome of the analysis was the incremental cost-effectiveness ratio (ICER) per quality-adjusted-life years (QALY) between competing treatment strategies. The analysis is based on standard cost-effectiveness methods using a societal perspective, which uses Medicare reimbursement costs (32). A willingness to pay (WTP) of less than $100,000/QALY was used as a threshold to determine cost-effectiveness. This threshold was derived from an analysis that estimated the ICER of hemodialysis which was inflation-adjusted to 2011 dollars (33). A WTP of less than $150,000/QALY was also considered in a probabilistic sensitivity analysis. Other outcomes assessed included cost, QALYs, and unadjusted life-years (life expectancy).

Analyses performed

A base-case analysis using best estimates for all model parameters and transition probabilities was performed. One-way sensitivity analyses were performed to investigate the effects of changes in model parameters on outcomes across a wide range of values, including non-dysplastic BE to EAC progression rate, aspirin and statin chemoprevention efficacy, complication rates, and the cost of statins. The ranges were based on published data. Additionally, probabilistic sensitivity analysis was performed. Distributions for specific parameters or model input variables were assigned and 1000 iterations were performed to gain further insight into the optimal strategy under uncertain conditions within our defined willingness to pay threshold.

Results

Base-Case Results

In the base-case analysis (Table 2 and Figure 3), the aspirin strategy dominated the endoscopic surveillance alone strategy, with the former strategy resulting in 0.167 more QALYs and costing $6,900 less. When the statin and combination strategies were compared to the endoscopic surveillance alone strategy, these strategies appeared to be cost-effective with ICERs of $37,600 per QALY and $16,300 per QALY, respectively.

Table 2.

Base Case Results

Outcome	Endoscopic surveillance alone	Aspirin with surveillance	Statin with surveillance	Combination with surveillance
Outcomes from the strategies
Cost (USD)	19,315	12,392	26,203	23,159
QALYs	16.873	17.040	17.056	17.108
Unadjusted LYs	24.411	24.714	24.741	24.833
Comparison among strategies using ICER values (USD)
Reference strategy
Endoscopic  surveillance alone	-	Dominates	37,640	16,350
Aspirin  with surveillance	-	-	863,200	158,300
Statin  with surveillance	-	-	-	Dominates

Figure 3.

Cost-effectiveness Analysis

Assuming a willingness to pay (WTP) threshold of $100,000 per QALY, the combination strategy of aspirin and statin was not cost-effective when compared to the aspirin strategy, as the ICER was $158,300 per QALY, or above our WTP. Comparing the statin strategy with aspirin strategy, the ICER was $863,200, which was well above our WTP. However, when the statin strategy was compared to endoscopic surveillance alone, the ICER was $37,640 per QALY; this ICER may be relevant in a hypothetical group of patients who cannot tolerate aspirin therapy. The statin strategy was dominated by the combination strategy.

Sensitivity Analysis

The results of the key sensitivity analyses are summarized in Table 3. The ICERs calculated in the table compare the combination strategy of aspirin and statin to the aspirin strategy. When the BE to EAC cancer progression rate of 0.33% per year was assumed in the base-case, the combination strategy did not seem to be the preferred strategy over the aspirin strategy. However, sensitivity analysis on the overall annual progression rate found that the combination strategy may be cost-effective at a higher progression rate of 0.5% per year with the resulting ICER, $96,000 per QALY.

Table 3.

Sensitivity Analyses: ICERs for Combination vs. Aspirin

Parameter	Combination vs. Aspirin only (USD)	Results Change (WTP $100,000)
BEND to Cancer progression
0.12%	441,400	-
0.33%	158,300	-
0.5%	96,000	Combination Cost-effective
Aspirin Complication (of base rate)
25%	166,400	-
50%	163,400	-
200%	147,100	-
400%	129,100	-
Statin Complication (of base rate)
25%	148,200	-
50%	151,280	-
200%	172,080	-
400%	209,400	-
Aspirin + Statin Reduction in progression
95%	82,700	Combination Cost-effective
90%	97,800	Combination Cost-effective
85%	117,500	-
80%	144,200	-
75%	182,200	-
Statin Cost (of base case cost)
20%	Dominates	Combination dominates
30%	11,180	Combination cost-effective
40%	32,100	Combination cost-effective
50%	53,040	Combination cost-effective
60%	73,970	Combination cost-effective
70%	94,900	Combination cost-effective
80%	115,830	-

The results of the model were not substantially affected by varying the complication rates of aspirin or statins. If the reduction in cancer progression for combined therapy of aspirin and statin is greater than 90%, then the combination strategy may become cost-effective (ICER<$100K/QALY). Lower statin cost (<70% of base-case estimate, $872 per year) could also make the combination strategy cost-effective. When the statin is reduced by more than 20%, the combination strategy dominates the aspirin strategy.

A two-dimensional sensitivity analysis of aspirin efficacy (based on HRs) and the transition probability between BE and cancer (see Figure 4) found that the combination strategy becomes more desirable as the transitional probability and aspirin efficacy decreases. When the transition probability is less than 0.35 and the aspirin efficacy is low, endoscopic surveillance may be the preferred strategy. Changing the efficacy rates of the statin strategy and the combination strategy showed little change.

Figure 4.

Two-Dimensional Sensitivity Analysis

Probabilistic sensitivity analyses (see results in Figure 5) using a WTP threshold of $150,000 per QALY found that although the probability of being cost-effective decreased with increasing WTP, the aspirin strategy was the preferred strategy when the WTP was between $0 and $100,000 per QALY . At a WTP of $100,000, the aspirin strategy was optimal in the majority of 51% of trials, while the combination strategy was optimal in 47.6% and the statin strategy in 1.4%. Above a WTP of $100,000 per QALY, the combination strategy becomes more cost-effective. At a WTP of $150,000, the combination strategy was optimal in the majority of 55.9% of trials, while the aspirin strategy was optimal in 42% and the statin strategy in 2.1%. Endoscopic surveillance alone was never a preferred strategy at any WTP value.

Figure 5.

Probabilistic Sensitivity Analysis

Discussion

Our analysis finds that aspirin therapy with endoscopic surveillance may be the most cost-effective strategy. Endoscopic surveillance alone and statin therapy were both dominated by aspirin and combination strategies respectively.

When compared to the aspirin strategy, the combination strategy was not cost-effective, with an ICER above the WTP threshold of $100,000 per QALY. However, if progression rates to cancer were increased (0.5% per year), combination therapy was cost-effective suggesting that combination therapy could be appropriate for those patients at high risk of EAC. In a probabilistic sensitivity analysis, the combination strategy becomes the cost-effective strategy when the WTP threshold is above $100,000.

Additionally, although the statin therapy was dominated by the combination therapy, when statins were compared with the endoscopic surveillance alone, they appeared cost-effective, suggesting that the addition of statin as a chemoprevention agent may be beneficial for patients who are unable or unwilling to take aspirin.

We performed sensitivity analyses acknowledging the uncertainty of model inputs including the benefit of chemoprevention and the progression rate from BE to EAC. The ICER comparing combination therapy to aspirin therapy decreased as the effect of combination therapy on EAC progression improved, and was below our WTP at a reduction of 90% or more. The combination therapy also had an ICER below the threshold when statins made up <70% of the base cost, and dominated the aspirin strategy when the cost was 20% of the base cost.

We previously published a model to determine the effectiveness and cost-effectiveness of aspirin as chemoprevention for esophageal cancer in patients with BE. In our prior analysis we demonstrated that aspirin therapy was both effective and cost-effective when compared to no therapy, either alone or when combined with endoscopic surveillance. Our current analysis confirms and adds to our previous work by adding statins as a possible therapy, either alone or combined with aspirin. Additionally, we updated our parameters with new data to reflect current trends and 2011 costs.

Our results have several implications for clinical care and health policy The overall five-year survival rate for esophageal cancer remains low, at 17%, but can increase to over 30% in patients with the earliest stages of disease (1). A patient’s greatest benefit to improving survival is early diagnosis; however, the cancer often spreads before symptoms are detected. BE is a precursor to cancer, yet an ideal management strategy remains elusive, and the utility of current endoscopic surveillance is controversial and costly. Several published studies have suggested that aspirin and statin may help prevent esophageal cancer (10-12). Furthermore, although we did not incorporate these factors into our model and analysis, both drugs have benefits that go beyond esophageal cancer, such as the prevention of cardiac disease and other cancers. Our analysis suggests the benefits of adding aspirin to an endoscopic surveillance program outweighs potential risks and can provide a cost-effective management strategy even when limited to esophageal adenocarcinoma. For patients at risk for cardiac disease, taking aspirin or statins may be even more compelling.

Our analysis has several limitations. The primary study we referenced to estimate the additive protective effect of combination therapy in reducing EAC progression rate used statins and NSAIDs, not just aspirin. The additive protective effects of statins and just aspirin were not available, but prior studies and our referenced study provided comparable benefits of NSAIDs and aspirin in EAC prevention (9-12, 15). Radiofrequency ablation (RFA) was not incorporated into our model as a treatment strategy because the model parameter estimates used in our analysis were based on results from a cohort study where RFA was not a treatment option (15). Also, as described above, our model did not incorporate any additional potential benefits from aspirin and statins beyond the chemoprevention of esophageal cancer. Our analysis focused solely on a comparison of aspirin and statins benefits in the prevention of EAC. In addition, death from cardiac diseases accounts for a large percent of mortality in older adults. To model this without long-term aspirin and statin data would result in an oversimplification of all-cause mortality rates, leading to significant inaccuracies in the model. Due to the lack of randomized clinical trials on the use of statins and aspirin among patients with BE for the prevention of esophageal cancer, our model used parameters primarily from prospective cohort studies. A study centered in the United Kingdom, ASPECT [http://www.octo-oxford.org.uk/alltrials/infollowup/aspect.html], is evaluating the impact of low and high dose aspirin on BE progression rates to cancer; however, a similar trial that evaluates the impact of statins and the combination of a statin and aspirin is not currently ongoing. When data from this trial becomes available in the future, we will be able to incorporate any changes in the impact of aspirin on progression rates into our model and analysis. In addition, as with any disease model that incorporates natural history, the limited amount of data available increases the uncertainty in the model and raises concerns regarding validity, particularly as future projections are made. Although the team of investigators that participated in this analysis has extensive experience with disease models, including more complex versions, we chose to construct a model that was as simple as possible in order to maintain a high level of model transparency and minimize the “black box” phenomenon. We performed extensive sensitivity analyses and despite these uncertainties our results were robust.

In summary, our results suggest that among the four treatment strategies analyzed, aspirin therapy is the cost-effective chemoprevention strategy for patients with BE. A combination of aspirin and statins could potentially be cost-effective in those patients with BE at a higher risk of progression to EAC. While the benefits of aspirin generally outweigh the risks in this patient population, statin therapy may be potentially cost-effective in patients who are unable or unwilling to take aspirin. While the $100,000-150,000 QALY figure may seem high, emerging technologies, medications, and screening can provide future improvements to BE management and improve the cost-effectiveness of care. Future clinical trials studying the long-term use of aspirin and statins in BE patients would be beneficial to confirm our model results.

Abbreviations

BE

Barrett’s esophagus

EAC

Esophageal Adenocarcinoma