Volume 201, Supplement 1, December 2015, Pages S8-S14 https://doi.org/10.1016/S0167-5273(15)31027-5Get rights and content
Heart disease and stroke can be fatal, but they can also lead to serious illness, disability, and lower quality of life. Suffering a stroke may lead to significant disability, such as paralysis, speech difficulties, and emotional problems. Following a heart attack, individuals frequently suffer fatigue and depression, and they may find it more difficult to engage in physical activities. Key FactsTogether, heart disease and stroke are among the most widespread and costly health problems facing the nation today. On a personal level, families who experience heart disease or stroke have to deal with not only medical bills but also lost wages and the real potential of a decreased standard of living.
More InformationLearn more from the Centers for Disease Control and Prevention:
The population health effect and cost-effectiveness of implementing intensive blood pressure goals in high-cardiovascular disease (CVD) risk adults have not been described. Using the CVD Policy Model, CVD events, treatment costs, quality-adjusted life years, and drug and monitoring costs were simulated over 2016 to 2026 for hypertensive patients aged 35 to 74 years. We projected the effectiveness and costs of hypertension treatment according to the 2003 Joint National Committee (JNC)-7 or 2014 JNC8 guidelines, and then for adults aged ≥50 years, we assessed the cost-effectiveness of adding an intensive goal of systolic blood pressure <120 mm Hg for patients with CVD, chronic kidney disease, or 10-year CVD risk ≥15%. Incremental cost-effectiveness ratios <$50 000 per quality-adjusted life years gained were considered cost-effective. JNC7 strategies treat more patients and are more costly to implement compared with JNC8 strategies. Adding intensive systolic blood pressure goals for high-risk patients prevents an estimated 43 000 and 35 000 annual CVD events incremental to JNC8 and JNC7, respectively. Intensive strategies save costs in men and are cost-effective in women compared with JNC8 alone. At a willingness-to-pay threshold of $50 000 per quality-adjusted life years gained, JNC8+intensive had the highest probability of cost-effectiveness in women (82%) and JNC7+intensive the highest probability of cost-effectiveness in men (100%). Assuming higher drug and monitoring costs, adding intensive goals for high-risk patients remained consistently cost-effective in men, but not always in women. Among patients aged 35 to 74 years, adding intensive blood pressure goals for high-risk groups to current national hypertension treatment guidelines prevents additional CVD deaths while saving costs provided that medication costs are controlled. For ≈4 decades, the Joint National Committee (JNC) on the Detection, Evaluation and Treatment of High Blood Pressure (BP) supported formulation of US hypertension treatment guidelines. From 1977 to 2003 (JNC1 to JNC7), the guidelines progressively lowered diagnostic thresholds and treatment targets, effectively expanding the treatment-eligible population. The 2014 hypertension guidelines (referred to here as JNC8) recommended higher BP goals compared with JNC7, so that ≈5.8 million fewer adults were eligible for antihypertensive medication treatment.1,2 JNC8’s less intensive BP goal recommendations for patients aged ≥60 years and those with diabetes mellitus or chronic kidney disease (CKD) provoked controversy and uncertainty.3 More recently, the Systolic Blood Pressure Intervention Trial (SPRINT) found that targeting an intensive systolic BP (SBP) goal of 120 mm Hg in patients with high cardiovascular disease (CVD) risk and baseline SBP ≥130 mm Hg reduced CVD events by 25% and all-cause mortality by 27%, compared with a 140-mm Hg goal.4 The objective of this study was to project the potential value of adding intensive SBP goals in high-risk patients to the JNC7 or JNC8 guidelines in a contemporary population of untreated hypertensive individuals aged 35 to 74 years. We also assessed if the incremental cost-effectiveness of intensive BP goals remained sensitive to the costs of more frequent monitoring or high medication prices. Patients aged ≥75 years were excluded from this analysis because of uncertainty about the tradeoff of risks and benefits of antihypertensive therapy in that population. The CVD policy model is a computer-simulation, state-transition (Markov cohort) model of incidence, prevalence, mortality, and costs of CVD in US adults (Methods section in the online-only Data Supplement).5,6 Means or proportions and joint distributions of risk factors, including BP, cholesterol, hypertension medication use, smoking, diabetes mellitus, and CKD status were estimated from pooled National Health and Nutrition Examination Surveys (NHANES) 2007 to 2010. Default multivariate stroke and coronary heart disease incidence functions were estimated in original Framingham Heart Study analyses. The CVD policy model predicts life years, CVD events (myocardial infarction and stroke), coronary revascularization procedures, CVD mortality (stroke [International Classification of Diseases-10 codes I60–I69], coronary heart disease [I20–I25 and two thirds of I49, I50, and I51], hypertensive heart disease deaths [I11.0 and I11.9]), and non-CVD deaths (remainder of International Classification of Diseases codes). Reductions in heart failure deaths because of hypertension treatment were calculated by adding prevented ischemic heart failure deaths (I50 with coronary heart disease) and hypertensive heart disease deaths (I11.0 and I11.9; Methods section in the online-only Data Supplement). Default model input parameters were calibrated, so that 2010 coronary heart disease and stroke incidence predictions matched hospitalized myocardial infarction and stroke rates observed in the 2010 National Hospital Discharge Survey, and mortality predictions were within 1% of age-specific 2010 CVD vital statistics mortality rates. Age- and sex-specific SBP and diastolic BP β-coefficients from the Prospective Studies Collaboration7 were calibrated, so that CVD Policy Model age-weighted relative risks with BP reduction fell within the 95% confidence interval of the overall relative risk estimates for the same BP reduction observed in a large meta-analysis of randomized controlled hypertension treatment trials (Methods section in the online-only Data Supplement; Tables S1–S3 in the online-only Data Supplement).8 To test predictive validity, we populated the model with the Systolic Hypertension in the Elderly Program (SHEP) trial cohort and simulated the BP reduction achieved in the active treatment arm of the trial for 5-years of follow-up. Our estimates accurately reproduced the risk reduction observed in the original trial (Table 1; Methods section in the online-only Data Supplement; Table S4).9 Table 1. Main Assumptions for the Comparative Effectiveness Analysis of Adding Intensive Blood Pressure Goals for High-Risk Patients to Current US Hypertension Treatment Guidelines Model InputsJNC7 recommended a goal BP <130/80 mm Hg for diabetes mellitus or CKD and BP <140/90 mm Hg for all others. JNC8 recommended a goal <140/90 mm Hg for diabetes mellitus or CKD, diastolic BP <90 mm Hg if age is <60 years, and BP <150/90 mm Hg if age is ≥60 years and without diabetes mellitus or CKD. On the basis of SPRINT, intensive interventions were applied to adults aged ≥50 years with pretreatment SBP ≥130 mm Hg and either existing CVD, CKD, or 2013 American Heart Association/American College of Cardiology Pooled Cohorts 10-year CVD risk ≥15%. Using these categories and BP and treatment status information from NHANES, we estimated the number of currently untreated US adults eligible for treatment under JNC7 and JNC8 with and without the intensive intervention in selected high-CVD risk individuals (Table S5). BP change caused by antihypertensive medications was determined by pretreatment BP and the number of standard doses of medications needed to reach the guideline BP goal according to a trial-based formula.10 BP changes were calculated based on pretreatment BP, age, and sex. We assumed the same BP reduction per standard dose of the main drug classes and did not include non-BP–lowering benefits of specific agents (Table 1; Methods section in the online-only Data Supplement; Table S5).8,10 We expected that CVD risk is reduced log linearly in relation to BP reduction (mm Hg) down to SBP 120 mm Hg in high-CVD risk patients in intensive strategies, 130/80 mm Hg in select JNC7 groups (diabetes mellitus or CKD), and SBP 140 mm Hg in those aged 60 to 74 years but without diabetes mellitus or CKD.8,10,11 Hypertension treatment costs included monitoring, side effect, and averaged wholesale drug costs. Quality of life penalties were applied for side effects.10 A medication adherence rate of 75% estimated in a meta-analysis of clinical trials was assumed because it corresponded to risk reduction associated with treatment estimated in the same meta-analysis (Table 1).8 A status quo simulation projected CVD events, CVD deaths, heart failure deaths, costs, and quality-adjusted life years (QALYs) for adults aged 35 to 74 years with untreated hypertension from 2016 to 2026. Adults aged ≥75 years were excluded from this analysis because of variable medication-related adverse event risk in this group.12 Guideline simulations modeled treatment according to JNC7 or JNC8. Incremental to JNC7 or JNC8, intensive strategies targeted an SBP of 120 mm Hg goal in high-CVD risk patients, limiting to 5 antihypertensive drugs maximum.4 Incremental cost-effectiveness ratios (ICERs) were calculated as change in costs divided by incremental change in QALYs. ICERs <$50 000 per QALY gained were considered cost-effective, ≥$50 000 and <$150 000 of intermediate value, and ≥$150 000 of low value.13 All analyses were approached from a payer’s perspective. Future costs and QALYs were discounted at 3% per year. Subgroup and Sensitivity AnalysesJNC7 and JNC8 with and without intensive treatment in selected high-CVD risk individuals were compared within age groups. One-way sensitivity analyses assessed cost-effectiveness assuming lower and upper uncertainty boundaries of the main inputs, including increased monitoring costs for the intensified treatment strategies (Table 1). We also modeled medication adherence as low as 40%.14 Main analyses did not include patients with treated but uncontrolled hypertension because it was not clear what proportion of poor control was because of underuse of combination therapy, poor adherence, or resistant hypertension.6 Nonetheless, we repeated the analyses in the entire population with uncontrolled hypertension, including previously treated and uncontrolled hypertension. Probabilistic AnalysesProbabilistic (Monte Carlo) simulation sampled across uncertainty distributions of antihypertensive drug BP-lowering effectiveness, CVD relative risk reduction with treatment, quality of life penalties, costs related to side effects, and drug and monitoring costs. Uncertainty distributions were randomly sampled 1000×, and 95% uncertainty intervals were calculated for all model outputs. Cost-effectiveness acceptability curves were constructed to illustrate the probability that each hypertension treatment strategy would be cost-effective at different willingness-to-pay thresholds. ResultsMain and Probabilistic ResultsCompared with no treatment, JNC8 would increase the annual number of newly treated adults aged 35 to 74 years by ≈12 million and would avert ≈65 000 CVD events and 17 000 CVD deaths annually. Compared with JNC8, JNC7 would recommend treatment for nearly twice the number of untreated patients (21 million) and add substantial treatment costs but would avert 24 000 additional CVD events and 5000 additional CVD deaths annually (Table 2). Incremental to JNC8, JNC8 plus intensive treatment in selected high-risk groups (JNC8+intensive) would prevent 43 000 additional annual CVD events and 15 000 CVD deaths. Incremental to JNC7, JNC7+intensive would lead to 35 000 fewer annual CVD events and 14 000 fewer CVD deaths. Total annual heart failure deaths avoided ranged from ≈2000 under JNC8 alone to ≈4000 under JNC7+intensive (Table S6 in the online-only Data Supplement).
In men, implementing JNC7 in addition to JNC8 would be cost-effective (ICER, ≈$7000 per QALY gained; Table 2). Incremental to JNC8, JNC7+intensive and JNC8+intensive strategies would be cost saving in men aged 35 to 74 years. At a willingness to pay threshold of $50 000 per QALY gained, the probability JNC7+intensive was more cost-effective than any other strategy in men was 100% (Figure 1, cost-effectiveness acceptability curve). At a lower willingness to pay threshold of <$25 000, JNC8+intensive was more likely to be cost-effective than the JNC7+intensive strategy (>50%, probability more cost-effective). In women, JNC7 was borderline cost-effective compared with JNC8 (≈$52 000 per QALY gained). Adding intensive treatment of high-risk patients was cost-effective in women incremental to JNC8. At a willingness-to-pay threshold of $50 000 per QALY gained, the probability that JNC8+intensive was the most cost-effective strategy for women was 81.7%, whereas the probability that the JNC7+intensive strategy most cost-effective was 18.3% (Figure 2). Figure 1. Cost-effectiveness acceptability curves for the probability of selecting Joint National Committee (JNC)-7+intensive over JNC8+intensive treatment in high-risk men aged 35 to 74 years (high risk: ≥50 years old with one of the following: existing cardiovascular disease [CVD], 2013 American College of Cardiology/American Heart Association Pooled Cohorts 10-year CVD risk ≥15%, or chronic kidney disease). JNC7 alone and JNC8 alone were dominated and do not appear on the plot. Figure 2. Cost-effectiveness acceptability curves for the probability of selecting Joint National Committee (JNC)-7+intensive over JNC8+intensive treatment in high-risk women aged 35 to 74 years (high risk: ≥50 years old with one of the following: existing cardiovascular disease [CVD], 2013 American College of Cardiology/American Heart Pooled Cohorts 10-year CVD risk ≥15%, or chronic kidney disease). JNC7 alone and JNC8 alone were dominated and do not appear on the plot. Subgroup AnalysesIncremental to JNC8, JNC7 would be cost saving in men aged 60 to 74 years, cost-effective in men aged 45 to 59 years (ICER, ≈15 000 per QALY gained) and in women 60 to 74 years (ICER, ≈30 000 per QALY gained), but of intermediate and low value in men and women aged 35 to 44 years, respectively. Incremental to JNC8 alone, JNC8+intensive would be cost saving in all age groups, whereas JNC7+intensive would be cost saving in all men and in women 60 to 74 years old, but cost-effective in women aged 45 to 59 years (ICER, 44 000 per QALY gained; Table S7). Sensitivity AnalysesAssuming 20% less CVD risk reduction per BP change (in mm Hg), more frequent monitoring plus double the drug costs, or 40% medication adherence, adding JNC8+intensive or JNC7+intensive remained cost-saving or cost-effective in most instances (ICERs <$50 000; Table S8). High drug costs plus higher monitoring frequency or 40% adherence made JNC7+intensive of intermediate or low value in women. JNC7 alone was sensitive to high drug costs incremental to JNC8. Adding treatment of treated but uncontrolled hypertension would double the population eligible for treatment to BP control under all strategies and lead to 60 000 to 91 000 fewer CVD events with intensive strategies compared with JNC8 alone. ICERs for the comparison of JNC7 versus JNC8 with and without intensive strategies remained similar when the previously treated and uncontrolled group was added (Table S9). DiscussionWe projected that adding intensive strategies to JNC hypertension treatment guidelines would be cost saving in men and cost-effective in women aged 35 to 74 years, which held true even in the event of higher monitoring costs. From a payer’s perspective, JNC8+intensive would most likely be the highest value strategy in women, whereas JNC7+intensive would most likely be the highest value strategy for men. The committee appointed by the JNC8 recommended an SBP target of 150 mm Hg among individuals aged 60 years and older and a target of 140 mm Hg for patients with diabetes mellitus or CKD, based on selected hypertension medication treatment trials. SPRINT results were released after the JNC8 published its recommendations and suggested greater CVD benefit from an SBP goal of 120 mm Hg, as opposed to 140 mm Hg in patients at high CVD risk.4 SPRINT reinforced evidence favoring a lower BP goal in selected high-risk patients.15–18 Concerns about the risks of intensive treatment persist.19,20 The bulk of randomized trial evidence demonstrates reduction in major CVD events, renal outcomes, and retinopathy from BP lowering well below the 140/90 mm Hg threshold without clear effects on CVD or noncardiovascular death, and the size of these benefits is consistent with epidemiological associations.8,21,22 The more recent Heart Outcomes Prevention Evaluation (HOPE)-3 trial found that BP lowering conferred no appreciable benefit in intermediate risk patients (mean 10-year CVD risk, ≈10%), except for those with pretreatment systolic BP >144 mm Hg.23 Therefore, treatment of patients with pretreatment systolic BP 130 to 139 mm Hg and 10-year CVD risk <15% according to JNC7 remains controversial. Our study had several limitations. Hypertension treatment guideline effectiveness and cost-effectiveness may vary among specific population groups with higher hypertension prevalence, such as blacks16 and subgroups at high risk for CVD, in whom greater benefits may derive from hypertension treatment. Although we estimated the effect of hypertension treatment on ischemic heart failure hospitalizations and deaths, coronary heart disease hospitalizations and deaths involving heart failure are difficult to accurately measure based on International Classification of Diseases–coded data. We projected heart failure deaths prevented because of hypertension treatment, but we did not simulate heart failure incidence or capture heart failure states directly, and we may have underestimated reduced heart failure burden attributable with hypertension treatment. We did not account for non–blood pressure–lowering benefits of certain antihypertensive drug classes, such as angiotensin-converting enzyme inhibitors, in patients with heart failure or past myocardial infarction. We did assume that most CVD patients would require >1 medication to reach the BP goal, one of those being an angiotensin-converting enzyme inhibitor. We may also have underestimated monitoring costs, including personnel, technology, or additional office visits needed to achieve intensive goals. We followed the decision of the SPRINT trial and did not target an SBP 120 mm Hg goal in patients with diabetes mellitus. Uncertainty persists about benefits and risks of intensive BP lowering in these patients.24,25 Intensive BP lowering consistently lowered stroke risk in trials enrolling older patients with diabetes, but results for coronary heart disease were variable.23,24 Patients with stroke were excluded from SPRINT; our decision to target intensive BP goals in patients with stroke is supported by suggestion of a benefit from intensive treatment in patients with stroke enrolled in the Secondary Prevention of Small Subcortical Strokes (SPS3) trial.26 SPRINT included participants aged ≥75 years, but we excluded elderly patients from our analysis because of uncertainty about risks and benefits of intensive BP lowering in the frail elderly.27 JNC recommendations have increased hypertension awareness, treatment, and control in the US population and likely contributed to the decline in CVD mortality during the past 4 decades.28 Our results suggest that targeting an intensive goal of 120 mm Hg in selected high-CVD risk patients in addition to the standard JNC guidelines would be cost saving if high drug costs can be controlled. PerspectivesHypertension treatment is inexpensive, safe, and effective. Guidelines should not be applied blindly, without considering the balance between benefits and harms in individual patients. However, in robust otherwise healthy patients aged <75 years, targeting more intensive blood pressure treatment goals in high CVD risk patients would be cost saving if monitoring and drug costs could be contained. AcknowledgmentsN. Moise and A.E. Moran had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. L. Goldman, N. Moise, and A.E. Moran contributed to the study concept and design. A.E. Moran, C. Huang, P.G. Coxson, and N. Moise contributed to acquisition of data. N. Moise, A.E. Moran, C. Huang, C.N. Kohli-Lynch, and A. Rodgers contributed to analysis and interpretation of data. N. Moise, A.E. Moran, and A. Rodgers contributed to drafting of the article. N. Moise, A.E. Moran, P.G. Coxson, K. Bibbins-Domingo, A. Rodgers, and L. Goldman contributed to critical revision of article for important intellectual content. N. Moise, A.E. Moran, P.G. Coxson, and C. Huang contributed to statistical analysis. A.E. Moran obtained funding and supervised the study. This work was supported by funds from Health Resources and Services Administration (T32HP10260), National Heart, Lung, and Blood Institute (R01 HL107475-01), and the American Heart Association Founder’s Affiliate (10CRP4140089). This article was prepared using Framingham Cohort and Framingham Offspring Research Materials obtained from the US National Heart, Lung and Blood Institute (NHLBI) Biological Specimen and Data Repository Information Coordinating Center and does not necessarily reflect the opinions or views of the Framingham Cohort, Framingham Offspring, or the NHLBI. K. Bibbins-Domingo is a member of the US Preventive Services Task Force (USPSTF) and current co-Vice Chair. This work does not necessarily represent the views and policies of the USPSTF. FootnotesReferences
Adding intensive blood pressure goals for high-risk patients to current national hypertension treatment guidelines prevents additional cardiovascular disease deaths while saving costs provided that medication costs are controlled. |