Patient Level Prediction of Clinical Outcomes and Cost-Effectiveness in SPRINT (Optimize-SPRINT)

Although safe and effective antihypertensive medications have been available for decades, hypertension remains the leading modifiable cause of cardiovascular disease (CVD) and life-years lost in U.S. and the optimal blood pressure goals for medication initiation and titration is unclear. Standard systolic blood pressure (SBP) treatment thresholds are defined by national hypertension guidelines (usually SBP < 140 mm Hg), but recent data from the Systolic Blood Pressure Intervention Trial (SPRINT) show that intensive treatment to a lower SBP goal (<120 mm Hg) reduces the risk of CVD events and all-cause mortality compared to standard SBP treatment in high CVD risk patients. Overall, risk of serious adverse events (SAEs) were similar between SPRINT arms, but select SAEs occurred more frequently in the intensive arm including hypotension, syncope, electrolyte abnormalities, and acute kidney injury. It has proved difficult to control U.S. hypertensive patients to current standard SBP goals (i.e., proportion controlled ~50%). To optimize implementation and public impact of SPRINT, it is critical to know which patients are the highest value candidates for intensive SBP treatment (i.e., combination of high absolute CVD risk reduction couple with low SAE risk; therefore most cost-effective). Intensive SBP treatment may cost more initially (e.g., increased office visits, laboratory tests, and medications). Without determining the cost-effectiveness of intensive treatment and who is most likely to derive benefit or harm, we will not realize the full public health benefits and societal value of intensive SBP treatment. Clinical guidelines rightly depend on high quality, randomized clinical trials like SPRINT as the “gold standard” evidence supporting medical decisions. Nonetheless, trial summary statistics and cost-effectiveness estimates based on them are aggregate measures that do not precisely inform benefit, risk, or cost-effectiveness to guide treatment decisions in i