Abstract: FR-PO537

Did Non-Standard Withdrawal of Antihypertensive Agents Exaggerate Treatment Effect in SPRINT?

Session Information

Category: Hypertension

  • 1104 Hypertension: Clinical and Translational - Salt and Hypertension

Authors

  • Judge, Conor S., HRB Clinical Research Facility, Galway, Ireland
  • Alvarez-Iglesias, Alberto, HRB Clinical Research Facility, Galway, Ireland
  • Ferguson, John P., HRB Clinical Research Facility, Galway, Ireland
  • Costello, Maria, HRB Clinical Research Facility, Galway, Ireland
  • Smyth, Andrew, HRB Clinical Research Facility, Galway, Ireland
  • O'donnell, Martin, NUI Galway, Galway, Ireland
Background

The control group in SPRINT targeted a blood pressure range of 130-140mmHg, which required down-titration of antihypertensive therapy when blood pressure was below 130mmHg on a single visit, or below 135mmHg on two consecutive visits. Such an approach would not be considered routine clinical care. We hypothesised that non-standard discontinuations of antihypertensive therapy in the control group, may have inflated the events rates and exaggerated the reported treatment effect.

Methods

Standard withdrawal of antihypertensive agents was defined as a withdrawal for systolic blood pressure less than 100mmHg at the current visit, or a reported, related adverse event occurring between the previous and current visit. We evaluated the association of antihypertensive withdrawal with CV events on follow-up using the Cox proportional-hazards regression. We repeated the primary analysis comparing the time to first occurrence of a primary outcome between treatment groups, adjusting for non-standard withdrawal in blood pressure medication, treated as time dependant covariates, to estimate the effect of intensive versus standard blood pressure control.

Results

Non-standard withdrawal of antihypertensive agents occurred in 9.3% of patient visits in the control group, compared with 5.1% in the treatment group (p < 0.001), and was associated with an increased risk of the composite outcome, which was significant for 2 follow-up periods (HR 1.65; 95% CI, 1.26-2.16 for initial 3 months, HR 1.47; 95% CI, 1.12-1.95 for 3 to 6-month period), which was independent of blood pressure effect. After adjusting for non-standard withdrawal/reduction of antihypertensive agents, the intensive-treatment group was associated with a lower risk of the composite outcome measure, compared to standard care (HR 0.81; 95% CI, 0.67 to 0.97).

Conclusion

Targeting a systolic BP range (130-140mmHg) in the control group of SPRINT, rather than a conventional blood pressure threshold (<140mmHg), resulted in withdrawals of antihypertensive medications that would not be considered routine care. An analysis that adjusted for non-standard withdrawal of blood pressure medications during the trial resulted in a significant, but diminished, treatment effect of intensive blood pressure control (HR 0.81 versus 0.75), and the effect on heart failure became non-significant.