MINIMUM NUMBER OF READINGS NEEDED TO DETERMINE LONG-TERM VISIT-TO-VISIT SYSTOLIC BLOOD PRESSURE VARIABILITY FOR PREDICTING CARDIOVASCULAR OUTCOMES IN DIABETIC PATIENTS

초록

Objective: In clinical practice, office blood pressure (BP) measurements are used for assessing long-term, visit-to-visit variability (VVV) of BP and the consistency of BP control in patients treated for hypertension. We aimed to determine the reproducibility of VVV and the minimum number of BP measurements required to estimate VVV for predicting cardiovascular outcomes. Design and method: This multicenter, retrospective study used data from the Korea University Medical Center database built on electronic health records. A total of 10,475 diabetic patients with >9 readings of BP during a minimum of two years were identified. Patients were followed up for 3 years after determining the coefficient of variation (CV), and the primary outcome was major adverse cardiovascular events (MACE), including cardiovascular mortality, myocardial infarction, stroke, and hospitalization for heart failure. Results: In diabetic patients, the average age of the patients was 65.2 years; 53.2% were men; and 87.4% had hypertension. The mean CV of systolic BP (SBP) was 9.0 ± 3.6. The mean CV of multiple readings that best predicted 3-year MACE were 8.4 for the 3 SBP readings (area under the curve [AUC], 0.57), 9.5 for the 5 readings (AUC, 0.61), 9.0 for the 7 readings (AUC, 0.63), 9.0 for the 9 readings (AUC, 0.64), and 9.7 for all readings (AUC, 0.66). In univariable analysis, 3-year MACE was significantly associated with high VVV from all SBP readings (HR 2.44, p<0.001) and 5 SBP readings (HR 2.03, p<0.001). Hence, a minimum of 5 SBP readings (median time 252 days) is needed for determining intraindividual VVV. Conclusions: In clinical practice, long-term VVV is reliably reproducible in diabetic patients. We recommend a minimum of five BP measurements for the estimation of intraindividual VVV. It is important to consider the number and duration of BP measurements used to calculate VVV when predicting cardiovascular outcomes.