Correlation of Blood Pressure Variability as Measured By Clinic, Self-measurement at Home, and Ambulatory Blood Pressure Monitoring

BACKGROUND: Blood pressure variability (BPV) has been postulated as a potential predictor of cardiovascular outcomes. No agreement exists as to which measurement method is best for BPV estimation. We attempt to assess the correlation between BPV obtained at the doctor's office, self-measurement at home (SMBP) and ambulatory BP monitoring (ABPM). METHODS: Eight weekly clinic BP measurements, 2 SMBP series, and 1 24-hour ABPM recording were carried out in a sample of treated hypertensive patients. BPV was calculated using the SD, the "coefficient of variation" and the "average real variability." Determinants of short-, mid-, and long-term BPV (within each measurement method) were also calculated. The different BPV determinants were correlated "intramethod" and "intermethod" by linear regression test. RESULTS: For the 104 patients (66.5 ± 7.7 years, 58.7% males), the ABPM BPV (SD, systolic/diastolic: 14.5 ± 3.1/9.8 ± 2.5 mm Hg) was higher than the SMBP (12.2 ± 9.8/7.4 ± 5.8 mm Hg; P < 0.001) and clinic BPV (10 ± 8.9/5.9 ± 4.9 mm Hg; P = 0.001). The main BPV correlation between methods was weak, with a maximum R2 = 0.17 (P < 0.001) between clinic and SMBP systolic BPV. The "intramethod" correlation of BPV yielded a maximum R2 = 0.21 (P < 0.001) between morning diastolic SMBP intershift/intermeans variability. The "intermethod" correlation of short-, mid-, and long-term BPV determinants was weak (maximum R2 = 0.22, P < 0.001, between clinic intraday variability/SMBP morning intershift variability). CONCLUSIONS: The "intramethod" and "intermethod" correlation between BPV determinants was weak or nonexistent, even when comparing determinants reflecting the same type of temporal BPV. Our data suggest that BPV reflects a heterogeneous phenomenon that strongly depends on the estimation method and the time period evaluated.