Does sex and calibration influence cardiovascular risk prediction from central systolic blood pressure?

Background: The accuracy of central BP is usually improved from calibration with MAP and DBP (C2SBP) compared to calibration with SBP and DBP (C1SBP). While preliminary data suggest C2SBP may have the best accuracy in females, we aimed to assess whether this could translate into improved cardiovascular (CV) risk prediction. Methods: 12,927 participants free of baseline CV disease, with prospective follow-up from administrative databases and central BP measurements were included. C1SBP was estimated with SphygmoCor Px devices. C2SBP was derived from unprocessed radial pressure waveforms recalibrated with DBP and 40% form factor derived MAP. Participants with heart rate < 60 were excluded due to incomplete waveforms. Major adverse CV events (MACE) comprised myocardial infarction, stroke, heart failure with hospitalization and CV death. Multivariable Cox regressions, differences in area under the curve, net reclassification index and integrated discrimination index were calculated comparing C2SBP to C1SBP and to bSBP, with and without stratification for sex. Results: Over a median follow-up of 10.1 years (IQR 9.9–10.3), 2125 MACE (723/7013 females and 860/5934 males) occured. All BP parameters were significantly associated with MACE, regardless of sex. In the overall cohort, risk prediction metrics marginally favored C2SBP compared to bSBP, but were similar to C1SBP. No significant improvement of CV risk prediction was found in sex-stratified analyses (see Table). Conclusions: C2SBP marginally improved CV risk prediction when compared to bSBP but not to C1SBP in the overall cohort only. All three BP parameters were similarly predictive in both sex, although this analysis possibly lacked power. This may be related to the FFderived MAP (rather than oscillometric MAP), which is highly dependent on the brachial SBP.