Use of advanced statistical techniques to predict all-cause mortality in the Systolic Blood Pressure Intervention Trial

Abstract

Background

The Systolic Blood Pressure Intervention Trial (SPRINT) was conducted in patients with hypertension and additional risk for cardiovascular disease who were randomized to the intensive blood pressure group targeting systolic blood pressure (SBP) less than 120 mm Hg and to the standard group where the target was less than 140 mm Hg. Analyses were done in the matched group of participants with the same gender, same age (±2 years) and same SBP (±3 mm Hg) at three months of treatment regardless of initial randomization to intensive or standard group (shaded area in Figure 1).

Methods and results

During 3.26 years of follow-up, intensive group participants had 14.8 mm Hg lower SBP and received on average one more (2.8 vs. 1.8) blood pressure lowering medications. This was associated with lower all-cause mortality in the intensive treatment group (hazard ratio, 0.73; 95% CI, 0.60 to 0.90, p = 0.003). The effect on SBP was achieved at 3 months and remained unchanged thereafter. This paper addresses two questions with respect to all-cause mortality in SPRINT in the matched set. 1) What is the effect of receiving more than one drug on all-cause mortality. Conditional logistic regression for all-cause mortality with respect to number of drugs indicated that during the 3.26 years of follow-up persons who received more than one drug were more likely to die (coefficient = 0.5039, OR = 1.6552, p = 0.0322) than patients who received one drug. 2) Was there a U curve relationship between on treatment SBP and all-cause mortality? A U curve fitting a quadratic equation (parabola) of SBP and all-cause death was observed. This was seen in the patients randomized to the standard target group in unadjusted analyses as well as in analyses adjusted for demographics or all covariates (p < 0.001 for all). The U curves in the combined group and the intensive treatment group were less pronounced.

Conclusion

SPRINT participants who were matched for gender, age, and SBP at 3 months, and received more than one drug had higher all-cause mortality during the 3.26 years of follow-up. Those who were randomized to standard treatment target had a U curve relationship between SBP at three months and all-cause mortality. The U curves in the combined group and the intensive treatment group were less pronounced.