a cautionary note from SPRINT Increased all-cause mortality with intensive blood-pressure control in patients with a baseline systolic blood pressure of ≥160 mmHg and a Lower Framingham risk score: a cautionary note from SPRINT Tzung-Dau Wang1, FESC, Hung-Ju Lin1, Wen-Jone Chen2, FESC, Te-Chang Weng3, Wen-Yi Shau3 1. Cardiovascular Center and Division of Cardiology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei City, Taiwan; 2. Department of Emergency Medicine, National Taiwan University Hospital, Taipei City, Taiwan; 3. Pfizer, Taipei, Taiwan

Potential conflicts of interest Speaker‘s name: Tzung-Dau Wang I do not have any potential conflict of interest

Greater BP reduction, smaller risk reduction!? Background SPRINT: Primary outcome SPRINT: Death from any cause J curve: vulnerability to absolute BP value or BP reduction? Universal or individualized BP target? Greater BP reduction, smaller risk reduction!? N Engl J Med 2015;373:2103-2116. Chiang CE, Wang TD, et al. Acta Cardiol Sin 2017;33:213-225.

Methods Access to the patient-level data of SPRINT through National Heart, Lung, and Blood Institute BioLINCC data repository after approval from the Institutional Review Board at National Taiwan University Hospital Outcomes: (1) Primary outcome (MI, non-MI ACS, stroke, acute decompensated HF, and CV death), (2) all-cause death, (3) primary outcome + all-cause death, and (4) non-CV death (all-cause death – CV death, including undetermined/not yet adjudicated cases) Due to small sample size in subgroup analyses, HR estimated from Cox model assuming common baseline hazard across clinic site

Results: Step 1, more detailed subgroup analysis based on baseline systolic BP stratified by clinically-oriented cut-off points Primary outcome All-cause death HR (95% CI) P for trend Systolic BP, mmHg Overall 0.5 1.0 2.0 Intensive Better Standard Better 0.75 (0.64-0.89) Overall 0.5 1.0 2.0 Intensive Better Standard Better 0.73 (0.60-0.90) (n=9361) <130 0.72 (0.51-1.03) 0.107 0.76 (0.49-1.17) 0.005 (n=2324) ≥130 to <140 0.63 (0.45-0.88) ≥130 to <140 0.61 (0.40-0.94) (n=2640) ≥140 to <150 0.92 (0.65-1.30) ≥140 to <150 0.67 (0.41-1.07) (n=2159) ≥150 to <160 0.74 (0.49-1.11) ≥150 to <160 0.69 (0.42-1.13) (n=1262) ≥160 0.86 (0.53-1.37) ≥160 1.16 (0.68-1.97) (n=976)

a baseline systolic BP of ≥160 mmHg versus <160 mmHg Results: Step 2, safety concerns with intensive BP control for patients with a baseline systolic BP of ≥160 mmHg versus <160 mmHg Systolic BP, mmHg Intensive Standard HR (95% CI) Pint no. of patients (%) % per year no. of patients (%) % per year Primary outcome <160 ≥160 211/4181 (5.0) 32/497 (6.4) 1.60 2.13 282/4204 (6.7) 37/479 (7.7) 2.16 2.49 0.579 0.74 (0.62-0.89) 0.86 (0.53-1.37) 0.2 1.0 5.0 Intensive Better Standard Better All-cause death <160 ≥160 126/4181 (3.0) 29/497 (5.8) 0.93 1.86 185/4204 (4.4) 25/479 (5.2) 1.37 1.62 0.072 0.68 (0.54-0.85) 1.16 (0.68-1.97) Primary outcome and all-cause death <160 ≥160 282/4181 (6.7) 50/497 (10.1) 2.14 3.32 377/4204 (9.0) 46/479 (9.6) 2.88 3.09 0.088 0.74 (0.64-0.87) 1.08 (0.72-1.61) Non-cardiovascular death <160 ≥160 95/4181 (2.3) 23/497 (4.6) 0.70 1.48 131/4204 (3.1) 14/479 (2.9) 0.97 0.91 0.025 0.72 (0.56-0.94) 1.64 (0.84-3.18)

of ≥160 mmHg (n=976)? sub-subgroup analyses Results: Step 3, heterogeneity among patients with a baseline systolic BP of ≥160 mmHg (n=976)? sub-subgroup analyses Primary outcome All-cause death Prior CVD No Yes Framingham 10-yr risk score, % ≤31.3 >31.3 Age, year <75 ≥75 Antihypertensive agents Prior CKD 0.343 0.94 (0.50-1.77) 0.62 (0.31-1.27) 0.888 0.88 (0.34-2.29) 0.84 (0.49-1.45) HR (95% CI) Pint 0.901 0.83 (0.42-1.64) 0.88 (0.46-1.69) 0.674 0.56 (0.10-3.03) 0.87 (0.53-1.43) 0.293 1.07 (0.58-1.99) 0.64 (0.30-1.36) Prior CVD No Yes Framingham 10-yr risk score, % ≤31.3 >31.3 Age, year <75 ≥75 Antihypertensive agents Prior CKD 0.093 1.56 (0.73-3.37) 0.66 (0.31-1.40) 0.044 2.99 (0.96-9.26) 0.81 (0.43-1.54) HR (95% CI) Pint 0.574 1.34 (0.64-2.84) 1.00 (0.46-2.15) 0.582 1.59 (0.43-5.96) 1.10 (0.61-1.98) 0.187 0.84 (0.40-1.76) 1.75 (0.79-3.85) 0.1 1.0 10.0 Intensive Better Standard Better 0.1 1.0 10.0 Intensive Better Standard Better

Results: Step 4, comparing patients with a baseline systolic BP of ≥160 mmHg and a Framingham 10-yr risk score of ≤31.3% to the rest of SPRINT participants SBP-Framinghan 10-yr risk score combination Intensive Standard HR (95% CI)* Pint† no. of patients (%) % per year no. of patients (%) % per year Primary outcome 0.1 1.0 10.0 Intensive Better Standard Better SBP ≥160/risk ≤31.3% All others 8/244 (3.3) 235/4434 (5.3) 1.06 1.68 9/236 (3.8) 310/4447 (7.0) 1.19 2.24 0.95 (0.37-2.46) 0.75 (0.63-0.89) 0.648 All-cause death SBP ≥160/risk ≤31.3% All others 12/244 (4.9) 143/4434 (3.2) 1.55 1.00 4/236 (1.7) 206/4447 (4.6) 0.52 1.44 3.12 (1.00-9.69) 0.69 (0.56-0.86) 0.009 Primary outcome and all-cause death SBP ≥160/risk ≤31.3% All others 16/244 (6.6) 316/4434 (7.1) 2.11 2.26 11/236 (4.7) 412/4447 (9.3) 1.45 2.98 1.53 (0.71-3.29) 0.76 (0.66-0.88) 0.075 Non-cardiovascular death SBP ≥160/risk ≤31.3% All others 10/244 (4.1) 108/4434 (2.4) 1.29 0.75 4/236 (1.7) 141/4447 (3.2) 0.52 0.99 2.60 (0.81-8.31) 0.76 (0.59-0.98) 0.036 *Adjusted for age (treated as quintile) and sex in the subgroup of SBP ≥160 mmHg and 10-yr risk score of ≤31.3% †Adjusted for age (treated as quintile) and sex and assuming common baseline hazard across clinic

Results: Step 4, comparing patients with a baseline systolic BP of ≥160 mmHg and a Framingham 10-yr risk score of ≤31.3% to the rest of SPRINT participants SPRINT-subgroup: Baseline SBP ≥160 mmHg & 10-yr Framingham risk score ≤31.3% SPRINT-original Death from Any Cause SBP 168 mmHg at baseline SBP 125 mmHg at Year 1 SBP 140 mmHg at Year 1 † Adjusted for age (treated as quintile) and sex, and assuming common baseline hazard across clinic site due to small sample size

Limitations Results from this post-hoc analysis are hypothesis-generating in nature The subset of hypertensive patients who had had increased mortality with intensive BP control is composed of only 5% (480/9361) of SPRINT participants -- Patients with stage 2 hypertension and a 10-year Framingham risk score of ≤30% (close to 31.3%) represent one-seventh hypertensive adults aged ≥50 years according to the National Health and Nutrition Examination Survey, 2007 to 2012 Bress AP, et al. J Am Coll Cardiol 2016;67:463-472.

Conclusions Among the SPRINT participants with a baseline systolic BP of ≥160 mmHg and a lower 10- ­year Framingham risk score (≤31.3%, median), targeting a systolic BP of <120 mmHg compared with <140 mmHg resulted in an approximate 3-fold risk of death from any cause Despite of the hypothesis-generating nature, it seems prudent to recommend targeting an SBP of <140 mmHg rather than <120 mmHg in patients with stage 2 hypertension and a 10-year Framingham risk score of ≤30% (close to 31.3%) There was an intricate interaction between each individual’s baseline blood pressure, their inherent cardiovascular risk, and their degree of blood pressure reduction. We have to consider all three of these elements in managing hypertensive patients